EPA Staff Technical Report:
Cost and Effectiveness Estimates of
Technologies Used to Reduce Light-duty
Vehicle Carbon Dioxide Emissions
United States
Environmental Protection
Agency
-------
EPA Staff Technical Report:
Cost and Effectiveness Estimates of
Technologies Used to Reduce Light-duty
Vehicle Carbon Dioxide Emissions
v>EPA
NOTICE
This technical report does not necessarily represent final EPA decisions or
positions. It is intended to present technical analysis of issues using data
that are currently available. The purpose in the release of such reports is to
facilitate the exchange of technical information and to inform the public of
technical developments.
United States EPA420-R-08-008
Environmental Protection .. , „„„
Agency March 2008
-------
Table of Contents
Executive Summary 1
1 Methodology 2
1.1 Fuel Economy Certification Data 2
1.2 Reports and Papers in the Literature 3
1.3 Confidential Data from Vehicle Manufacturers and Component Suppliers 4
2 CO2 Reduction for Individual Vehicle Technologies 5
2.1 Baseline Definition and Vehicle Classes 5
2.2 Summary of Estimates 6
2.3 Engine Technologies 8
2.3.1 Low-Friction Lubricants 8
2.3.2 Engine Friction Reduction 8
2.3.3 Variable Valve Timing Systems 8
2.3.3.1 Intake Camshaft Phasing (ICP) 8
2.3.3.2 Coupled Camshaft Phasing (CCP) 9
2.3.3.3 Dual (Independent) Camshaft Phasing (DCP) 9
2.3.4 Engine Cylinder Deactivation 9
2.3.5 Variable Valve Lift Systems 10
2.3.5.1 Discrete Variable Valve Lift 10
2.3.5.2 Continuous Variable Valve Lift 10
2.3.6 Camless Valve Actuation Systems 10
2.3.7 Stoichiometric Gasoline Direct Injection Technology 11
2.3.8 Lean-Burn Gasoline Direct Injection Technology 11
2.3.9 Gasoline Homogeneous Charge Compression Ignition 12
2.3.10 Gasoline Turbocharging and Downsizing 13
2.3.11 Diesel Engine 13
2.3.11.1 LeanNOx Trap Catalyst aftertreatment 14
2.3.11.2 Selective Catalytic Reduction NOx Aftertreatment 15
2.3.12 E20-E30 Optimized Ethanol Engines 15
2.4 Transmission Technologies 15
2.4.1 Automatic 5-speed Transmissions 15
2.4.2 Aggressive Shift Logic 16
2.4.3 Early Torque Converter Lockup 16
2.4.4 Automatic 6-, 7- and 8-speed Transmissions 17
2.4.5 Automated (shift) Manual Transmissions 17
2.4.6 Continuously Variable Transmissions 18
2.4.7 Manual (clutch shifted) 6-, 7-, and 8-speed Transmissions 19
2.5 Hybrid Vehicle Technologies 19
2.5.1 Integrated Starter Generator w/Idle-Off 20
2.5.2 Integrated Motor Assist (IMA)/Integrated Starter-Alternator-Dampener (ISAD) Hybrid.... 20
2.5.3 Power-Split Hybrids 22
2.5.4 Two-Mode Hybrids 25
2.5.5 Full-Series Hydraulic Hybrids 26
2.5.6 Plug-in Hybrid Electric Vehicles 26
2.6 Full Electric Vehicles 28
2.7 Vehicle Accessories 29
2.7.1 Electric Accessories and High Efficiency Alternator 29
2.7.2 Electric Power Steering for 12V and 42V systems 30
-------
2.7.3 Upgrade Electrical Systems to 42V 30
2.8 Other Vehicle Technologies 30
2.8.1 Aerodynamic Drag Force Reduction 30
2.8.2 Low Rolling Resistance Tires 30
2.8.3 Low Drag Brakes 31
2.8.4 Secondary Axle Disconnect (front axle for ladder frame and rear axle for unibody frame). 31
2.8.5 Weight Reduction 31
Synergistic Effects of Combining Multiple CO2 Reducing Technologies 33
3.1 EPA's Lumped Parameter Approach for Determining Effectiveness Synergies 34
3.2 Ricardo's Vehicle Simulation 36
3.2.1 Description of Ricardo's Report 37
3.2.1.1 Determination of representative vehicle classes 38
3.2.1.2 Description of Baseline Vehicle Models 39
3.2.1.3 Technologies Considered by EPA and Ricardo in the Vehicle Simulation 39
3.2.1.4 Choice of Technology Packages 41
3.2.1.5 Simulation Results 42
3.3 Comparison of Lumped-Parameter Results to Modeling Results 43
3.4 Using the Lumped-Parameter Technique to Determine Synergies in a Technology Application
Flowpath (Identifying "Technology Pairs" to account for synergies) 44
Costs for Technologies 46
4.1 Methodology for Estimating Variable Piece Costs 46
4.2 Piece Costs Assigned to CO2 Reduction technologies 46
4.2.1 Piece Costs Associated with Engine Technologies 54
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
4.2.
.1 Low-Friction Lubricants 54
.2 Engine Friction Reduction 54
.3 Variable Valve Timing Systems 55
.4 Engine Cylinder Deactivation 56
.5 Variable Valve Lift Systems 56
.6 Camless Valve Actuation Systems 57
.7 Stoichiometric Gasoline Direct Injection Technology 58
.8 Lean-Burn Gasoline Direct Injection Technology 58
.9 Homogeneous Charge Compression Ignition 58
.10 Gasoline Turbocharging and Engine Downsizing 59
.11 Diesel Systems 59
.12 E20-E30 Optimized Ethanol Engines 60
4.2.2 Piece Costs Associated with Transmission and Hybrid Technologies 60
4.2.2.1 Automatic 5-speed Transmissions 60
4.2.2.2 Aggressive Shift Logic 60
4.2.2.3 Early Torque Converter Lockup 61
4.2.2.4 Automatic 6-, 7- and 8-speed Transmissions 61
4.2.2.5 Automated (shift) Manual Transmissions 61
4.2.2.6 Continuously Variable Transmissions 62
4.2.2.7 Manual (clutch shifted) 6-, 7-, and 8-speed Transmissions 62
4.2.2.8 Hybrid Systems 62
4.2.3 Piece Costs Associated with Accessory Technologies 65
4.2.3.1 High Efficiency Alternators, electric water pumps and electrification of other
accessories for 12 Volt systems 65
4.2.3.2 Electric Power Steering for 12 Volt and 42 Volt systems 65
4.2.3.3 Upgrade Electrical Systems to 42V 65
4.2.4 Piece Costs Associated with Other Vehicle Technologies 65
11
-------
4.2.4.1 Aerodynamic Drag Reduction through reduced drag coefficient and reduced frontal
area 65
4.2.4.2 Low Rolling Resistance Tires 66
4.2.43 Low Drag Brakes 66
4.2.4.4 Secondary Axle Disconnect (front axle for ladder frame and rear axle for unibody
frame) 66
4.3 Estimates of Indirect Costs and the Use of Markup Factors 66
4.3.1 Current Methodology 68
4.3.2 Limitations and Uncertainties with Current Methodologies 70
4.3.2.1 Outdated Research 70
4.3.2.2 Studies Limited to U.S. Domestic Manufacturers 70
4.3.2.3 Treatment of Outsourced vs. Internally-Developed Technologies 70
4.3.2.4 Short-term and Long-term Adjustment Factors 71
Appendix 4.A Producer Price Index Adjustments 72
References 74
in
-------
Executive Summary
The National Research Council's Committee on the Assessment of Technologies for
Improving Light-Duty Vehicle Fuel Economy is tasked with providing updated estimates of the
costs and potential efficiency improvements that might be employed to improve fuel economy.
On March 4, 2008 the NRC Committee requested that EPA provide it with EPA's technical
analysis on the control of greenhouse gas emissions from light-duty vehicles, to aid the
committee in its work.A
This report presents EPA technical staff current assessment of the costs and effectiveness
from a broad range of technologies which can be applied to cars and light-duty trucks. The
report is divided into four major sections. In Section 1, we discuss the methodology used to
develop cost and effectiveness estimates, including what data sources we relied upon. In Section
2, we present our estimates of the carbon dioxide (CO2) reduction potential of nearly 40
individual technologies covering five broad categories: engines, transmissions, hybrids,
accessories, and others (e.g., aerodynamic improvements). These estimates are for individual
technologies compared to a baseline vehicle, and the estimated effectiveness cannot simply be
added up when considering a combination of technologies. This issue is addressed in Section 3
of the report, which discusses the synergistic effects of combining multiple technologies and
provides an estimate of the magnitude of this impact on CC>2 reduction effectiveness. Finally,
Section 4 provides an estimate of the direct costs associated with each of the technologies, as
well as a discussion of estimating indirect costs and the potential for future cost reductions.
The majority of the technologies discussed in this report are in production and available on
vehicles today, either in the United States, Japan, or Europe. A number of the technologies are
commonly available, while others have only recently been introduced into the market. In a few
cases, we provide estimates on technologies which are not currently in production, but are
expected to be so in the next few years.
In general, we believe these estimates we present are conservative. They rely on data sources
from the past one to six years, which in some cases are relatively old. The automotive industry is
a technology-driven industry, and new technologies are developed and introduced quickly. A
number of technologies which have only recently been introduced or will be within the next year
are likely to see improvements in their effectiveness and cost reductions beyond what we
estimate today. Nevertheless, we believe that the estimates presented in this report are defensible
and reasonable predictions for the next few years.
This report is based on our assessment of currently available data, much of which is in the
public domain. Based on this assessment, EPA technical staff concludes there are a large
number of technologies which can be applied to cars and trucks that are capable of achieving
significant reductions in greenhouse gas emissions, and improve vehicle fuel economy, at
reasonable costs.
See Attachment 1, which includes the request letter from the Committee.
-------
1 Methodology
In estimating the cost and effectiveness for vehicle CC>2 reduction technologies, we relied
upon a number of sources for technical information. We utilized those sources of information
which were determined to be credible for projecting the CC>2 reduction effectiveness of
individual vehicle technologies which are either currently available, or which we project will be
available in the next two to ten years. These data sources included: vehicle fuel economy
certification data; peer reviewed or publicly commented reports; peer reviewed technical journal
articles and technical papers available in the literature; and confidential data submissions from
vehicle manufacturers and automotive industry component suppliers. The following summarizes
our use of the most commonly utilized data sources. The discussion of each individual
technology in Section 2 (CC>2 reduction effectiveness) and Section 4 (technology cost estimates)
includes the citation to the source(s) of information we used for evaluating CC>2 reduction
potential of that specific technology.
EPA has conducted on-going technical analysis on the control of greenhouse gases
considered control of carbon dioxide (CO2), methane, nitrous oxide, and hydroflourocarbons. In
addition, the technical work on control of CC>2 emissions includes, but is not limited to,
emissions measured over EPA's traditional test cycles for measuring fuel economy. As EPA
acknowledged in a regulatory action completed in 2006, these test cycles, which are more than
20 years old, do not accurately represent the true fuel economy values which today's vehicles
will typically achieve on the road. These traditional tests also cannot account for the CC>2 or fuel
economy reduction potential of all available technologies, in particular efficiency improvements
for vehicle air conditioning systems. Given the charge to the NAS Committee, this technical
report has been limited to information on the effectiveness and cost of technologies to reduce
CC>2 emissions over the test cycles used for measuring fuel economy, and our estimates of
percentage improvements in CC>2 reduction are based on the traditional fuel economy test cycles.
1.1 Fuel Economy Certification Data
Where available, we considered data from recent model years from EPA's Fuel Economy
Certification Data. These CO2 reduction estimates were estimated from EPA's fuel economy
database on the two-cycle (FTP city & highway) fuel economy test results. During the standard
fuel economy test cycles, direct measurements of CC>2 emissions are made. This data, along with
other measurements, are then used to calculate the estimated fuel economy performance in
gallons of fuel consumed per mile. Vehicle certification data are an obviously reliable source for
determination of the CC>2 reduction potential when a directly comparable vehicle was offered
both with and without the specific CC>2 reducing technology, because a comparison between the
emissions data between the two vehicles directly reflect the application of the technologies on
the vehicle test cycles. Technology-specific effectiveness numbers were extracted for vehicles
where only the specific technology would be changed from a reference vehicle, in order to
eliminate any confounding of values across several technologies. In some hybrid vehicle cases,
the exact same vehicle may not be offered, and we selected a similar vehicle for comparison.
Examples of the use of such vehicle certification data are:
-------
• Honda Civic Hybrid compared to a directly comparable Honda Civic conventional
drive;
• GM's 2008 model year ("2-mode") hybrid full-size SUV powered by a V8 with
cylinder deactivation compared to a similar vehicle without a hybrid system but with
cylinder deactivation; and
• Honda's 2005 model year Odyssey V6 minivan with and without cylinder
deactivation.
1.2 Reports and Papers in the Literature
A large number of technical reports and papers are available which contain data and
estimates of the CC>2 reduction potential of various vehicle technologies. In addition to specific
peer-reviewed papers respecting individual technologies, we also utilized a number of recent
reports which had been utilized by various State and Federal Agencies and which were
specifically undertaken for the purpose of estimating future vehicle CC>2 reduction effectiveness
or improvements in fuel economy. The reports we utilized most frequently were:
• 2002 National Academy of Science (NAS) report titled "Effectiveness and Impact of
Corporate Average Fuel Economy Standards". At the time it was published, the NAS
report was considered by many to be the most comprehensive summary of current
and future fuel efficiencies improvements which could be obtained by the application
of individual technologies. The focus of this report was fuel economy, which can be
directly correlated with CC>2 emissions. In many cases, more recent information has
become available on the CC>2 reduction effectiveness of individual technologies. We
therefore assessed and, where reliable, utilized the updated information. For those
technologies for which we were not able to determine more recent, credible data than
were reported in the 2002 NAS report, we utilized the NAS report information. In
addition, the 2002 NAS report contains effectiveness estimates for ten different
vehicle classifications (small car, mid-SUV, large truck, etc), but did not differentiate
these effectiveness values across the classes. Where other sources or engineering
principles indicated that a differentiation was warranted, we utilized the 2002 NAS
effectiveness estimates as a starting point and further refined the estimate to one of
five vehicle classes using engineering judgment or by consulting additional reliable
sources.
• 2004 Northeast States Center for a Clean Air Future (NESCCAF) report "Reducing
Greenhouse Gas Emissions from Light-Duty Motor Vehicles". This report, which
was utilized by the California Air Resources Board for their 2004 regulatory action
on vehicle CC>2 emissions, includes a comprehensive vehicle simulation study
undertaken by AVL, a world-recognized leader in automotive technology and
engineering. In addition, the report included cost estimates developed by the Martec
Group, a market-based research and consulting firm which provides services to the
automotive industry. The NESCCAF report considered a number of technologies not
examined in the 2002 NAS report. In addition, through the use of vehicle simulation
modeling, the 2004 NESCCAF report provides a scientifically rigorous estimation of
-------
the synergistic impacts of applying multiple CC>2 reduction technologies to a given
vehicle.
1.3 Confidential Data from Vehicle Manufacturers and Component Suppliers
We also evaluated confidential data from a number of vehicle manufacturers as well as a
number of technology component suppliers. Over the past several years, EPA has met numerous
times with the worlds leading automotive companies as well as many of the major automotive
supply firms. During these meetings, EPA has received confidential briefings regarding
companies near and long-term plans for future technologies which can reduce criteria pollutants
(e.g., oxides of nitrogen and paniculate matter), reduce CC>2 emissions, and improve vehicle fuel
economy. EPA reviewed this information in the development of this report.
In February of 2007, the National Highway Traffic Safety Administration published a
detailed Request for Comment (RFC) in the Federal Register. This RFC included, among other
items, a request for information from automotive manufacturers and the public on the fuel
economy improvement potential of a large number of vehicle technologies. EPA has been
furnished this information by NHTSA (or in some cases directly from the vehicle manufacturer)
pursuant to a Memorandum of Understanding between NHTSA and EPA which conforms to the
Confidential Business Information rules of both agencies. The manufacturer's submissions to
this RFC were supplemented by confidential briefing and data provided by vehicle component
suppliers, who for many of the technologies considered are the actual manufacturers of the
specific technology and often undertake their own development and testing efforts to investigate
the CC>2 reduction potential of their products.
In considering the confidential data from vehicle manufacturers and component suppliers, it
was sometimes the case where the specific basis of a technology's effectiveness was not
described (e.g., was the estimate based on production vehicle testing, vehicle simulation
modeling, engineering judgment, or some other basis). In addition, it was also sometimes the
case where a manufacturers projection of the effectiveness of a technology for fuel economy
improvement (or CC>2 reduction) was also coupled with improvements in other vehicle attributes,
such as an increase in vehicle weight, or an increase in engine power or torque - making it more
difficult to distinguish the technology's impact only on CC>2 reduction effectiveness. For these
reasons, we tended to rely less solely on the manufacturers estimates for technology
effectiveness than on other data sources. Nevertheless, the confidential submissions from
automotive companies and technology suppliers were utilized in some cases, and were often used
to validate the estimates from other sources.
-------
2 CO2 Reduction for Individual Vehicle Technologies
The following sections detail the CC>2 reduction effectiveness of individual technologies. In
order to estimate CC>2 reduction effectiveness, it is necessary to clearly define the baseline
technology from which the new technology is being estimated, this is discussed in Section 2.1.
As also discussed in Section 2.1, we have developed CC>2 reduction effectiveness estimates for
five broad categories of vehicles in order to represent the range of products available in the light-
duty vehicle fleet. These five categories are labeled small car, large car, minivan, small truck
and large truck. The technologies are organized by six broad categories: engine technologies
(Section 2.3), transmission technologies (Section 2.4), hybrid technologies (Section 2.5), electric
vehicles (Section 2.6), accessory technologies (Section 2.7) and other vehicle technologies
(Section 2.8). A summary of our estimates for all technologies for the five vehicle classes is
presented in Section 2.2
Please note that the estimated CC>2 reduction effectiveness discussed in Section 2 do not
account for synergistic impacts between certain technologies. These synergistic impacts are
discussed in Section 3 of this report.
2.1 Baseline Definition and Vehicle Classes
In order to estimate both technology costs and CC>2 reduction estimates, it is necessary to
describe the vehicle characteristics baseline from which the estimates can be compared. For this
report, unless noted elsewhere, the baseline vehicle is defined as a vehicle with a port-fueled
injected, naturally aspirated gasoline engine with two intake and two exhaust valves and fixed
valve timing and lift. The baseline transmission is a 4-speed automatic, and the vehicle has no
hybrid systems. Our assessment attempted to maintain vehicle performance equivalent to todays
products, that is, we have provided estimates of the ability of various technologies to reduce CC>2
emissions without improving or reducing other vehicle performance characteristics when
compared to today's vehicles.
It is well known that both the costs and the effectiveness of any given CC>2 reduction
technology will not be the same on every vehicle, given the wide range of characteristics of
vehicle sizes and performance being offered today.
Existing reports in the literature typically group vehicles into categories or classes based on
those characteristics which can have a discernable impact on the application of a given
technology. For example, the 2002 NAS report divided the car and light-truck fleet into ten
different vehicle classes. However, the 2002 NAS provided the same range of estimates for fuel
consumption improvement and incremental costs for each of the ten vehicle classes discussed in
the report. The 2004 NESCCAF report provided five vehicle classes, and, where the authors
deemed it appropriate, different estimated values were provided for the five different vehicle
classes.
In this report, we provide effectiveness and cost estimates for five classes of vehicles, which
are generally intended to represent broad groupings of a wide variety of products offered in the
US car and light-duty truck market. In some cases, we have provided further differentiation
within a given class (e.g., between unibody and ladder-frame constructed vehicles for axle
-------
disconnect technologies). Where the data sources we reviewed provided sufficient detail, we
refined estimates to provide further effectiveness refinement within sub-division of the classes.
The names used to distinguish these five classes are for ease of reference in the report; they
are not intended to be viewed narrowly. We use these five categories to represent the following
types of vehicles.
• Small car: a subcompact or compact car typically powered by an inline 4 cylinder
engine.
• Large car: a midsize or large passenger car typically powered by a V6 cylinder engine.
• Minivan: a minivan or large cross-over unibody constructed vehicle with a large frontal
area, typically powered by a V6 engine, capable of carrying ~ 6 or more passengers.
• Small truck: small or mid-sized sports-utility and cross-over vehicles, or a small pick-up
truck, typically powered by a 6-cylinder engine.
• Large truck: large sports-utility vehicles and large pickup trucks, typically a ladder-on-
frame construction, and typically powered by a V8 engine.
2.2 Summary of Estimates
Table 2.2-1 through Table 2.2-5 summarize our estimates for the CC>2 reduction estimates of
various technologies which can be applied to cars and light-duty trucks. A similar summary of
costs are provided in Table 4.2-1. Each of these estimates is discussed in more detail in Sections
2.3 through 2.8.
Table 2.2-1 Engine Technology Effectiveness
Technology
Low friction lubricants - incremental to base engine
Engine friction reduction - incremental to base engine
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
0.5
1-3
Large
Car
0.5
1-3
Minivan
0.5
1-3
Small
Truck
0.5
1-3
Large
Truck
0.5
1-3
Overhead Cam Branch
VVT - intake cam phasing
VVT - coupled cam phasing
VVT - dual cam phasing
Cylinder deactivation (includes imp. oil pump, if applicable)
Discrete VVLT
Continuous VVLT
2
3
3
n.a.
4
5
1
4
4
6
3
6
1
2
2
6
3
4
1
3
2
6
4
5
2
4
4
6
4
5
Overhead Valve Branch
Cylinder deactivation (includes imp. oil pump, if applicable)
VVT - coupled cam phasing
Discrete VVLT
Continuous VVLT (includes conversion to Overhead Cam)
n.a.
3
4
5
6
4
4
6
6
2
3
4
6
3
4
5
6
4
4
5
Camless valvetrain (electromagnetic)
Gasoline Direct Injection-stoichiometric (GDI-S)
Gasoline Direct Injection-lean burn (incremental to GDI-S)
Gasoline HCCI dual-mode (incremental to GDI-S)
Turbo+downsize (incremental to GDI-S)
Diesel- Lean NOxtrap[]*
Diesel - urea SCR []*
5-15
1-2
8-10
10-12
5-7
15-26
[25-35]
15-26
[25-35]
5-15
1-2
9-12
10-12
5-7
21-32
[30-40]
21-32
[30-40]
5-15
1-2
9-12
10-12
5-7
21-32
[30-40]
21-32
[30-40]
5-15
1-2
9-12
10-12
5-7
21-32
[30-40]
21-32
[30-40]
5-15
1-2
10-14
10-12
5-7
21-32
[30-40]
21-32
[30-40]
: Note: estimates for % reduction in fuel consumption are presented in brackets.
-------
Table 2.2-2 Transmission Technology Effectiveness
Technology
5-speed automatic (from 4-speed auto)
Aggressive shift logic
Early torque converter lockup
6-speed automatic (from 4-speed auto)
6-speed AMI (from 4-speed auto)
6-speed manual (from 5-speed manual)
CVT (from 4-speed auto)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
6
Large
Car
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
6
Minivan
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
6
Small
Truck
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
n.a.
Large
Truck
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
n.a.
Table 2.2-3 Hybrid Technology Effectiveness
Technology
Stop-Start with 42 volt system
IMA/ISA/BSG (includes engine downsize)
2-Mode hybrid electric vehicle
Power-split hybrid electric vehicle
Full-Series hydraulic hybrid
Plug-in hybrid electric vehicle
Full electric vehicle (EV)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
7.5
30
n.a.
35
40
58
100
Large
Car
7.5
25
40
35
40
58
100
Minivan
7.5
20
40
35
40
58
n.a.
Small
Truck
7.5
20
40
35
40
58
n.a.
Large
Truck
7.5
20
25
n.a.
30
47
n.a.
Table 2.2-4 Accessory Technology Effectiveness
Technology
Improved high efficiency alternator & electrification of accessories
(12 volt)
Electric power steering (12 or 42 volt)
Improved high efficiency alternator & electrification of accessories
(42 volt)
Absolute CC<2 Reduction (% from baseline vehicle)
Small
Car
1-2
1.5
2-4
Large
Car
1-2
1.5-2
2-4
Minivan
1-2
2
2-4
Small
Truck
1-2
2
2-4
Large
Truck
1-2
2
2-4
Table 2.2-5 Other Vehicle Technology Effectiveness
Technology
Aero drag reduction (20% on cars, 10% on trucks)
Low rolling resistance tires (10%)
Low drag brakes (ladder frame only)
Secondary axle disconnect (unibody only)
Front axle disconnect (ladder frame only)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
3
1-2
n.a.
1
n.a.
Large
Car
3
1-2
n.a.
1
n.a.
Minivan
3
1-2
n.a.
1
n.a.
Small
Truck
2
1-2
1
1
1.5
Large
Truck
2
n.a.
1
n.a.
1.5
-------
2.3 Engine Technologies
Unless noted otherwise, the baseline engine for the engine technologies described in this
section is a port-fuel injected, spark-ignition, naturally aspirated, 4-valve per cylinder engine
with fixed intake and exhaust valve timing.
2.3.1 Low-Friction Lubricants
More advanced multi-viscosity engine and transmission oils are now available with improved
performance in a wider temperature band, with better lubricating properties. Manufacturers are
moving from 5W-30 to 5W-20 and even OW-20 engine oils to reduce cold start friction. This
may directionally benefit the fuel economy improvements of valvetrain technologies such as
cylinder deactivation, which rely on a minimum oil temperature (viscosity) for operation.
Confidential manufacturer data submitted by vehicle manufacturers in response to NHTSA's
February 2007 Request for Comment (2/2007 RFC) suggests that low-friction lubricants could
reduce CO2 emissions by 0.5 percent for all vehicle types.
2.3.2 Engine Friction Reduction
All reciprocating and rotating components in the engine are candidates for friction reduction,
and minute improvements in several components can add to a measurable fuel economy
improvement. Several friction reduction opportunities (piston surfaces and rings, crankshaft
design, improved material coatings, roller cam followers, etc.) have been identified that are still
available to a significant number of engine designs. Additionally, as computer-aided modeling
software continues to improve, more opportunities for incremental friction reduction might
become apparent. Confidential manufacturer data provided in response to the 2/2007 RFC
indicates that the CC>2 reduction potential ranges from 1 to 3 percent for engine friction reduction
technologies.
2.3.3 Variable Valve Timing Systems
Variable valve timing has been available in the market for quite a while. By the early 1990s,
VVT had made a significant market penetration with the arrival of Honda's "VTEC" line of
engines. VVT has now become a widely adopted technology: for the 2006 model year, over half
of all new cars and light trucks have engines with some method of variable valve timing.1
Therefore, the degree of further improvement across the fleet is limited to vehicles that have not
already implemented this technology.
Manufacturers are currently using many different types of variable valve timing mechanisms,
which have a variety of different names and methods. The major types of VVT are listed below.
2.3.3.1 Intake Camshaft Phasing (ICP)
Valvetrains with ICP - the simplest subset of cam phasing - can modify the timing of the
intake valve while the exhaust valve timing remains fixed. We estimate that ICP designs may
-------
enable a 1 to 2 percent reduction in CCh compared to fixed-valve engines. This estimate is based
directly on the work detailed in the 2004 NESCCAF report, which relied on vehicle simulation
modeling to predict the CO2 reduction (by vehicle class) of a long list of vehicle technologies
including VVT.2
2.3.3.2 Coupled Camshaft Phasing (CCP)
Coupled (or coordinated) cam phasing is a design in which both the intake and exhaust valve
timing are varied with the same cam phaser. The 2004 NESCCAF report indicates CCP designs
may enable a 2-4% reduction in CC>2 emissions above fixed-cam valvetrains. Vehicles with
higher power-to-weight ratios (large cars and large trucks) are at the high of this range, while
minivans are at the low end.
2.3.3.3 Dual (Independent) Camshaft Phasing (DCP)
The most flexible VVT design is dual cam phasing, where the intake and exhaust valve
opening and closing events are controlled independently. This design allows the option of
controlling valve overlap, which can be used as an internal exhaust gas recirculation (EGR)
strategy. EPA estimates that DCP designs may enable a 2-4% reduction in CO2 emissions
compared to fixed-valve engines, similar to estimates for CCP.3
2.3.4 Engine Cylinder Deactivation
In implementing cylinder deactivation, some (usually half) of the cylinders are "shut down"
during light load operation - the valves are kept closed, and no fuel is injected - as a result, the
trapped air within the deactivated cylinders is simply compressed and expanded as an air spring,
with reduced friction and heat losses. The active cylinders combust at almost double the load
required if all of the cylinders were operating. Pumping losses are significantly reduced as long
as the engine is operated in this "part-cylinder" mode. The theoretical engine operating region
for cylinder deactivation is limited to no more than roughly 50% of peak power at any given
engine speed. In practice, however, cylinder deactivation is employed primarily at lower engine
cruising loads and speeds, where the transitions in and out of deactivation mode are less apparent
to the operator and where the noise and vibration (NVH) associated with fewer firing cylinders
may be less of an issue. Manufacturers are exploring the possibilities of increasing the amount
of time that part-cylinder mode might be suitable to a vehicle with more refined powertrain and
NVH treatment strategies.
Cylinder deactivation has seen a recent resurgence thanks to better valvetrain designs and
engine controls (it was first offered in the 1980s on the Cadillac 8-6-4, but was discontinued
because of reliability problems). General Motors and Chrysler Group have incorporated cylinder
deactivation across a substantial portion of their V8-powered lineups. Honda (Odyssey, Pilot)
and General Motors (Impala, Monte Carlo) offer V6 models with cylinder deactivation.
Fuel economy improvement potential scales 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.
On its own accord, cylinder deactivation can reduce CO2 emissions by 6%, at minimum, for
applicable vehicles - those with engines of 6 or more cylinders (with vehicles with higher engine
displacement-to-weight ratios seeing more of a potential improvement). This number is
supported by official fuel economy test data on a model year 2005 V6 Honda Odyssey with
-------
cylinder deactivation compared to the same vehicle (and engine displacement) without cylinder
deactivation.
2.3.5 Variable Valve Lift Systems
Controlling the lift height of the valves provides additional flexibility and potential for
further reduction in CC>2 emissions. By reducing the valve lift, engines can decrease the
volumetric flow at lower operating loads, improving fuel-air mixing and in-cylinder mixture
motion which results in improved thermodynamic efficiency and also potentially reduced overall
valvetrain friction. Also, 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. At the same time, such systems may also incur increased parasitic losses associated
with their actuation mechanisms. A number of manufacturers have already implemented VVLT
into their fleets (Toyota, Honda, BMW) but overall this technology is still available for most of
the fleet. There are two major classifications of variable valve lift, described below:
2.3.5.1 Discrete Variable Valve Lift
Discrete variable valve lift (DVVL) is a method in which the valvetrain switches between
multiple cam profiles, usually 2 or 3, for each valve. 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 estimated to provide an additional 4% CC>2 emissions
reduction above that realized by VVT systems (with the exception of minivans at 3%, due to
their traditionally lower power-to-weight ratios).2
2.3.5.2 Continuous Variable Valve Lift
Continuous variable valve lift (CVVL) valvetrains are mechanically more complicated than
DVVL designs. Currently, only BMW has implemented this type of system (in its Valvetronic
engines, which also incorporates fully flexible valve timing), in which an extra set of rocker arms
are used to vary the valve lift height. This design is limited to overhead cam engines. The
contribution of CVVL, independent of any improvement of a VVT system, has been estimated to
potentially reduce CC>2 emissions by 4% (minivans) up to 6% (large cars) over VVT systems
according to simulation results in the NESCCAF report.2
2.3.6 Camless Valve Actuation Systems
Camless valve actuation relies on electromechanical actuators instead of camshafts to open
and close the cylinder valves. An engine valvetrain that operates independently of any
mechanical means provides the ultimate in flexibility for intake and exhaust timing and lift
optimization. With it comes infinite valve overlap variability, the rapid response required to
change between combustion modes (such as HCCI and spark ignition), intake valve throttling,
cylinder deactivation, and elimination of the camshafts (reduced friction). Camless valvetrains
have been under research for many decades due to the design flexibility and the attractive fuel
economy improvement potential they might provide.
Despite the promising features of camless valvetrains, significant challenges remain. High
costs and design complexity have reduced manufacturers' enthusiasm for camless engines in
light of other competing valvetrain technologies. The advances in WT, VVLT, and cylinder
10
-------
deactivation systems demonstrated in recent years have reduced the potential efficiency
advantage of camless valvetrains.
There is a broad range of opinion on the potential CO2 emissions reduction advantage of
camless systems, depending on the level of parasitic loads required to operate the actuators.
EPA projects that the potential net CO2 emissions reductions might range from 5-15% (over a
fixed cam-driven valvetrain) depending on the integration and optimization of a future camless
system.4 We are projecting the camless valve systems will not be widely available in high
volume light-duty vehicles within the next 10 years.
2.3.7 Stoichiometric Gasoline Direct Injection Technology
Gasoline direct injection (GDI, or SIDI) engines inject fuel at high pressure directly into the
combustion chamber (rather than the intake port in port fuel injection). Direct injection
improves cooling of the air/fuel charge within the cylinder, which allows for higher compression
ratios and increased thermodynamic efficiency. Injector design advances and increases in fuel
pressure have promoted better mixing of the air and fuel, enhancing combustion rates, increasing
exhaust gas tolerance and improving cold start emissions. GDI engines achieve higher power
density and match well with other technologies, such as boosting and variable valvetrain designs.
Several manufacturers have recently released GDI engines: besides Audi's lineup of GDI
engines, Volkswagen and BMW have GDI offerings; Toyota (Lexus IS 350) and General Motors
(Chevy Impala 3.6L) are already in production or about to be introduced. In addition, BMW and
GM have announced their plans to dramatically increase the number of GDI engines in their
portfolios.
On its own, Stoichiometric GDI does not bring with it the promise of CO2 emissions
reductions much beyond 2%, but combined with other technologies (boosting, downsizing) it
could enable a significant reduction in consumption compared to engines of similar power output
(as discussed in Section 2.3.10). Confidential data from multiple manufacturers agree with this
CO2 reduction estimate.
2.3.8 Lean-Burn Gasoline Direct Injection Technology
Direct injection, especially with diesel-like "spray-guided" injection systems, enables
operation with excess air in a stratified or partially-stratified fuel-air mixture, as a way of
reducing the amount of intake throttling. Also, with higher-pressure fuel injection systems, the
fuel may be added late enough during the compression stroke so as to delay the onset of
autoignition, even with higher engine compression ratios. Taken together, an optimized "lean-
burn" direct injection gasoline engine may achieve high engine thermal efficiency. European
gasoline direct-injection engines have achieved some success with this concept, although at far
higher NOx emissions levels than are allowed at today's Tier 2 emissions standards. To date, no
manufacturers have sold a light-duty lean-burn GDI engine in the US market due to the higher
cost of lean NOx catalyst systems relative to three-way catalysts, coupled with the corresponding
need for low-sulfur gasoline.
However, several injector suppliers are optimistic about the potential of lean-burn GDI
engines in the near future. Fuel system improvements, changes in combustion chamber design,
and repositioning of the injectors have allowed for better air/fuel mixing and combustion
11
-------
efficiency. There is currently a shift from wall-guided injection to spray guided injection, which
improves injection precision and targeting towards the spark plug, increasing lean combustion
stability. The increased combustion stability allows for recirculated exhaust gas (EGR) rates of
up to 40% which can significantly reduce in-cylinder NOx emissions. Combined with advances
in NOx aftertreatment (commensurate with diesel progress), lean-burn GDI engines may be a
possibility in North America.
As noted above, a key technical requirement for lean-burn GDI engines to meet Tier 2 NOx
emissions levels is the availability of gasoline with sulfur levels commensurate with ultra-low
sulfur diesel, for durability of the lean NOx catalyst systems. Without the availability of ultra
low sulfur gasoline, it does not appear that lean-burn GDI engines can be expected to penetrate
the light-duty market anytime soon.
The most recent CO2 reduction estimates for lean-burn GDI engines range from 8-10% for
small cars to 10-14% for large trucks, compared to a port-fueled (stoichiometric) engine. These
estimates are based on the 2004 NESCCAF report and are supported by confidential
manufacturer and supplier estimates.
2.3.9 Gasoline Homogeneous Charge Compression Ignition
Gasoline homogeneous charge compression ignition (HCCI), also referred to as controlled
autoignition (CAI), is an alternate engine operating mode that does not rely on a spark event to
initiate combustion. The principles are more closely aligned with a diesel combustion cycle, in
which the compressed charge exceeds a temperature and pressure necessary for spontaneous
autoignition. The subsequent combustion event is much shorter in duration with higher thermal
efficiency.
An HCCI engine has inherent advantages in its overall efficiency for two main reasons:
• The engine is operated with a higher compression ratio, and with a shorter
combustion duration, resulting in a higher thermodynamic efficiency, and
• The engine can be operated virtually unthrottled, even at light loads,
Combined, these effects have shown an increase in engine brake efficiency (typically 25-28%) to
greater than 35% at the high end of the HCCI operating range.5
Criteria pollutant emissions are very favorable during HCCI operation. Lower peak in-
cylinder temperatures (due to high dilution) keep engine-out NOx emissions to a minimum -
realistically below Tier 2 levels without aftertreatment - and particulates are low due to the
homogeneous nature of the premixed charge.
Due to the inherent difficulty in maintaining combustion stability without encountering
engine knock, HCCI is difficult to control, requiring feedback from in-cylinder pressure sensors
and rapid engine control logic to optimize combustion timing, especially considering the
transient nature of operating conditions seen in a vehicle. Due to the highly lean and/or dilute
conditions under which HCCI combustion is stable, the range of engine loads achievable in a
naturally-aspirated engine is somewhat limited. Because of this, it is likely that any commercial
application would operate in a "dual-mode" strategy between HCCI and spark ignition
combustion modes, in which HCCI would be utilized for best efficiency at light engine loads and
spark ignition would be used at higher loads and at idle. This type of dual-mode strategy has
12
-------
already been employed in a few diesel-HCCl engines in Europe and Asia (notably the Toyota
Avensis D-Cat and the Nissan light-duty "MK" combustion diesels).
Until recently, gasoline-HCCI technology was considered to still be in the research phase.
However, most manufacturers have made public statements about the viability of incorporating
HCCI into light-duty passenger vehicles, and have significant vehicle demonstration programs
aimed at producing a viable product within the next 5-10 years.
There is widespread opinion as to the CC>2 reduction potential for HCCI in the literature.
Based on confidential manufacturer information, EPA believes that a gasoline HCCI / GDI dual-
mode engine might achieve 10-12% reduction in CC>2, compared to a comparable SI engine.
Despite its promise, application of HCCI in light duty vehicles is not yet ready for the market,
and will remain so for at least a few more years. It is not anticipated to be seen in volume for at
least the next 5-10 years, which is concurrent with many manufacturers' public estimates.
2.3.10 Gasoline Turbocharging and Downsizing
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 size
while maintaining performance. This effectively reduces the pumping losses at lighter loads in
comparison to a larger, naturally aspirated engine, while at the same time reducing net friction
losses.
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, it
has considerable fuel economy potential when the engine displacement is 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 conservatively be
downsized roughly 30% to achieve 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 speeds.
The 2002 NAS report suggests that a downsized turbocharged engine at equivalent
performance levels would offer (CO2) reductions of 5 to 7%, which is supported by confidential
manufacturer data. EPA considers this 5 to 7% reduction achievable over a naturally-aspirated
stoichiometric GDI engine of comparable performance. This technology is available today.
2.3.11 Diesel Engine
Diesel engines have several characteristics that give them superior fuel efficiency to
conventional gasoline, spark-ignited engines:
• Pumping losses are greatly reduced due to lack of (or greatly reduced) throttling.
13
-------
• The diesel combustion cycle operates at a higher compression ratio, with a very
lean overall air/fuel mixture, both of which contribute to higher thermal
efficiency.
• Turbocharged light-duty diesels typically achieve much higher torque levels at
lower engine speeds than equivalent-displacement naturally-aspirated gasoline
engines.
Additionally, diesel fuel has a higher energy content per gallon; all of these effects combine
for dramatically lower CO2 emissions. However, diesel engines have emissions characteristics
that present challenges very different from gasoline engines to meet Tier 2 emissions.
Criteria pollutant emissions compliance strategies are expected to include a combination of
combustion improvements and aftertreatment. Several key advances in diesel technology have
made it possible to reduce emissions coming from the engine (prior to aftertreatment). These
technologies include:
• Improved fuel systems (higher pressures and more responsive injectors)
• Advanced controls and sensors to optimize combustion and emissions
performance
• Higher EGR levels to reduce NOx
• Lower compression ratios (still much higher than gasoline SI engines)
• Advanced turbocharging systems
For aftertreatment, the traditional 3-way catalyst found on gasoline-powered vehicles is
ineffective due to the lean-burn combustion of a diesel. All diesels will require a particulate
filter, an oxidation catalyst, and a NOx reduction strategy to comply with Tier 2 emissions
standards. The NOx reduction strategies most common are outlined below:
2.3.11.1 Lean NOx Trap Catalyst aftertreatment
A lean NOx trap (LNT) operates, in principle, by storing NOx (NO and NO2) when the
engine is running in its normal (lean) state. When the control system determines (via a
predictive model or an NOx sensor) that the trap is saturated with NOx, it switches to a rich
operating mode. This rich mode produces excess hydrocarbons that act as a reducing agent to
convert the stored NOx to N2 and water, thereby "regenerating" the LNT and opening up more
locations for NOx to be stored. LNTs are sensitive to sulfur deposits which can reduce catalytic
performance, but periodically undergo a desulfation engine operating mode to clean it of sulfur
buildup. Tailpipe CO2 reduction estimates for an LNT-based diesel car range from 15 to 32
percent compared to a fixed valvetrain, port-fueled gasoline engine. This estimate translates into
a corresponding tailpipe fuel consumption reduction estimate of 25 to 40 percent. These
estimates are based on the RIA supporting NHTSA's Light Truck CAFE Rule.
While there is already evidence of LNT-based diesels in production worldwide, they are all
certified at higher NOx emission levels than U.S. Tier 2 Bin 5 levels. However, EPA projects
that T2B5 LNT-based diesel engines will be available in the US within the next year or two,
based on announcements from Mercedes, Volkswagen, and Honda.
14
-------
2.3.11.2 Selective Catalytic Reduction NOx Aftertreatment
SCR uses a reductant (typically, ammonia derived from urea) continuously injected into the
exhaust stream ahead of the SCR catalyst. Ammonia combines with NOX in the SCR catalyst to
form N2 and water. The hardware configuration for an SCR system is more complicated that that
of an LNT, due to the onboard urea storage and delivery system (which requires a urea pump and
injector into the exhaust stream). While there is no required rich engine operating mode
prescribed for NOX reduction, the urea is typically injected at a rate 3-4% that of fuel consumed.
Manufacturers designing SCR systems are intending to align urea tank refills with standard
maintenance practices such as oil changes. CC>2 reduction estimates for diesel engines with an
SCR system range from 21 to 32 percent over conventional, port-fueled gasoline engines (which
translates to a fuel consumption reduction of approximately 30 to 40 percent).
As is the case with LNT-based diesels, EPA projects that SCR-based diesel engines will be
available within the next couple of years. Mercedes-Benz has recently announced two new
vehicles which have received US EPA certificates for Model Year 2009, the Mercedes Benz
R320 and GL320, both of which achieved Tier 2, Bin 5 emissions. Based on public
announcements from several other companies, we expect a large number of product offerings
from multiple companies over the next few years.
2.3.12 E20-E30 Optimized Ethanol Engines
Ethanol has many favorable combustion qualities that increase knock tolerance, provide for
more stable and faster combustion, and cool the charge air down much more than gasoline; taken
together, these properties may be leveraged in such a manner as to increase the engine's thermal
efficiency. For example, ethanol's high octane number permits an engine's compression ratio to
be increased, while still allowing spark advance to be further optimized. Moreover, its faster rate
of combustion allows for a higher rate of exhaust gas recirculation, thereby reducing pumping
losses.
Based on internal EPA work, we estimate that optimizing an engine to operate on E20 to E30
could increase fuel efficiency and reduce tailpipe CC>2 emissions by 7-10% relative to a port-
fueled, fixed cam gasoline engine. However, this EPA work has not been peer reviewed or
published, and therefore we consider this estimate to be preliminary. For this reason, we have
not included this preliminary estimate in the summary tables in Section 2.2. Note that this
technology would be applicable for a vehicle specifically optimized for E20-E30.
2.4 Transmission Technologies
2.4.1 Automatic 5-speed Transmissions
As automatic transmissions have been developed, more forward speeds have been added to
improve fuel efficiency, performance, and to improve a vehicle's market position. Increasing the
number of available ratios provides the opportunity to operate an engine at more optimized
conditions over a wider variety of vehicle speeds and load conditions. Also, additional ratios can
allow greater overdrive (where the output shaft of the transmission is turning at a higher speed
than the input shaft) which can lower the engine speed at a given road speed (provided the
engine has sufficient torque reserve at the lower rpm point) to reduce pumping losses. However,
in some cases, additional gears can add weight, rotating mass, and friction providing some offset
15
-------
to the efficiency advantage. Nevertheless, manufacturers are increasingly adding 5-speed
automatic transmissions to replace 3-, and 4-speed automatics.
Some 4-speed automatic transmission designs are capable of offering five (or more) ratios by
modifying the hydraulic control system (valvebody) and the electronic controls. This is much
less expensive than developing a new transmission, but the available ratios may not be ideally
spaced for optimizing fuel economy.
We estimate a 5-speed automatic transmission offers a CC>2 reduction of 2.5% (relative to a
4-speed automatic transmission). This estimate is based on the 2002 NAS report. The
effectiveness of this technology was well understood at the time of the NAS report and the 2.5%
value is also confirmed by CBI information from manufacturers. A 5-speed automatic
transmission is applicable to all vehicle types.
2.4.2 Aggressive Shift Logic
During vehicle operation, an automatic transmission's controller decides when to upshift or
downshift based on a variety of inputs such as vehicle speed and throttle position according to
programmed logic. This logic can be biased towards maximizing fuel efficiency by upshifting
earlier and inhibiting downshifts under some conditions. Additional adaptive algorithms can be
employed to maintain performance feel while improving fuel economy under most driving
conditions.
The 2002 NAS report states that aggressive shift logic can reduce fuel consumption by 1-3%
in a 5-speed automatic transmission. The 2004 NESCCAF report states that the benefit is 1.5%.
Information from manufacturers suggests that the benefit is in the lower end of the NAS range,
so we estimate the benefit to be between 1% and 2%. Aggressive shift logic is applicable to all
vehicle types with automatic transmissions, and since in most cases it would require no
significant hardware modifications, it can be adopted during vehicle redesign or refresh or even
in the middle of a vehicle's product cycle. The application of this technology does, however,
require a manufacturer to confirm that driveability, durability, and noise, vibration, and
harshness (NVH) are not significantly degraded.
2.4.3 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. If the torque
converter cannot be fully locked up for maximum efficiency, a partial lockup strategy can be
employed to reduce slippage.
The 2002 NAS report did not address this particular technology, but the 2004 NESCCAF
report used a number of literature sources to determine that early lockup can provide a 0.5% CC>2
16
-------
benefit. NESCCAF states that this estimate is conservative in order to protect shift quality and
driveability. This value is within the range of CBI information submitted by manufacturers, so
we believe 0.5% is an appropriate estimate for this technology.
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 driveability, performance, durability and NVH characteristics is required to
successfully implement this technology.
2.4.4 Automatic 6-, 7- and 8-speed Transmissions
In addition to 5-speed automatic transmissions, manufacturers can also choose to utilize 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 gearsets 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-speed automatics with 6-speed automatics, and 7-, and 8-
speed automatics have also entered production, albeit in lower-volume applications in luxury
cars.
The 2002 NAS report states that relative to a 4 speed automatic, a 6 speed automatic can
reduce fuel consumption by 3% to 5%. At the time of the NAS report, 6-speed automatics were
not in use, although some were in development. More current knowledge and information
provided by manufacturers suggests that the CC>2 reduction potential of 6-speed automatics is
more like 4.5% to 6.5% which is the range of effectiveness we believe is appropriate. 7-speed
and 8-speed automatics are just entering production in small numbers, so there is not a lot of
experience with them. Although they may be slightly more efficient than 6-speed automatics, we
group them together with 6-speeds. As more data becomes available and more manufacturers
gain experience with these transmissions, their effectiveness can be independently estimated. 6-,
7-, and 8-speed automatic transmissions are applicable to all vehicle types.
2.4.5 Automated (shift) Manual Transmissions
An Automated Manual Transmission (AMT) is mechanically similar to a conventional
manual transmission, but shifting and launch functions are controlled by the vehicle. There are
two basic types of AMTs, single-clutch and dual-clutch. A single-clutch AMT is essentially a
manual transmission with automated clutch and shifting. Because there are some shift quality
issues with single-clutch designs, dual clutch AMTs will likely be far more common in the U.S.
and are the basis of our estimates. A dual-clutch AMT uses separate clutches (and separate gear
shafts) for the even number 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 it comes time to shift, the controller disengages the odd-gear clutch while simultaneously
engaging the even-gear clutch, thus affecting a smooth shift. If, on the other hand, the driver
17
-------
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.
Overall, AMTs likely offer the greatest potential for CC>2 reduction 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 computer control. 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 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 an AMT will most likely be paired with an engine that offers enough
torque in the low-RPM range to allow for adequate launch performance.
The 2002 NAS report listed automated manual transmissions under emerging transmission
technologies and assigned a fuel consumption reduction potential of 3%-5% over a 4-speed
automatic. As these transmissions have entered production, it has become clear that the benefits
are larger. We estimate these transmissions offer a CO2 reduction potential of 9.5%-14.5%
over a 4-speed automatic transmission. This estimate was developed from an aggregation of
information from auto manufacturers and suppliers. AMTs can be used in all vehicle types.
2.4.6 Continuously Variable Transmissions
A Continuously Variable Transmission (CVT) is unique in that it does not use gears to
provide ratios for operation. Instead, the most common CVT design uses two V-shaped pulleys
connected by a metal belt. Each pulley is split in half and a hydraulic actuator moves the pulley
halves together or apart. This causes the belt to ride on either a larger or smaller diameter
section of the pulley which changes the effective ratio of the input to the output shafts.
Advantages of the CVT are that the engine can operate at its most efficient speed-load point
more of the time, since there are no fixed ratios. Also, CVTs often have a wider range of ratios
compared to conventional automatic transmissions which can provide for more options in engine
optimization. While CVTs by definition are fully continuous, some automakers choose to
emulate conventional stepped automatic operation because some drivers are not used to the
sensation of the engine speed operating independently of vehicle speed.
The 2002 NAS report shows a relatively wide range of fuel consumption reduction of 3-8%
compared to a 5-speed automatic. The 2004 NESCCAF report's estimate is much lower at 3-4%
better than a four-speed automatic. Based on an aggregation of manufacturers' information, we
estimate a CVT benefit of about 6% over a 4-speed automatic. This is above the NESCCAF
value, but in the range of NAS. We assume that it is only practical to apply CVTs to small cars,
large cars, and minivans because they are currently used mainly in lower-torque applications.
While a high-torque CVT could be developed for small trucks and large trucks, it would likely
have to be treated separately in terms of effectiveness. We do not see development in the area of
high-torque CVTs and therefore did not include this type in our analysis.
18
-------
2.4.7 Manual (clutch shifted) 6-, 7-, and 8-speed Transmissions
As with automatic transmissions, 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 at a given speed. Typically, this is achieved through adding additional
overdrive ratios to reduce engine speed (which saves fuel through reduced pumping losses) Six-
speed manual transmissions have already achieved significant market penetration, but for those
vehicles with five-speed manual transmissions, an upgrade to a six-speed offers a benefit of 0.5%
according to an aggregation of manufacturer-supplied information. 6-speed manual
transmissions were not addressed in either the NAS or NESCCAF reports. These transmissions
are applicable to all vehicle types with manual transmissions.
2.5 Hybrid Vehicle Technologies
A Hybrid 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 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 CC>2
emissions. The effectiveness of 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 to not 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 Honda Accord
Hybrid (now discontinued), it is more likely to be used for vehicles like trucks where towing
and/or hauling is a integral part of their performance envelope. In these cases, 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 vehicle with significantly diminished towing performance
with a low battery.
Different hybrid concepts utilize these mechanisms differently, so they are treated separately
for the purposes of this analysis. Below is a discussion of the major hybrid concepts judged to
be available in the near term.
19
-------
2.5.1
Integrated Starter Generator w/ Idle-Off
Integrated Starter-Generator (ISG) systems are the most basic of hybrid systems and offer
mainly idle-stop capability. They offer the least power assist and regeneration capability of the
hybrid approaches, but their low cost and easy adaptability to existing powertrains and platforms
can make them attractive for some applications. ISG systems generally operate at around 42
volts and so have smaller electric motors and less battery capacity than other HEV designs
because of their lower power demand.
Most ISG systems replace the conventional belt-driven alternator with a belt-driven, higher
power starter-alternator (see Figure 2.5-1). The starter-alternator starts the engine during idle-
stop operation, but often a conventional 12V gear-reduction starter is retained to ensure cold-
weather startability. Also, during idle-stop, some functions such as power steering and automatic
transmission hydraulic pressure are lost with conventional arrangements, so electric power
steering and an auxiliary transmission pump are added. These components are similar to those
that would be used in other hybrid designs. An ISG system could be capable of providing some
launch assist, but it would be limited in comparison to other hybrid concepts.
The 2002 NAS report states that the potential fuel consumption reduction of an ISG system
with idle-stop-only functionality is 4%-7%. Adding some regeneration and power assist can
yield a total of 5%-10% fuel consumption reduction. The 2004 NESCCAF report states a 4%-
10% benefit is possible. We chose 7.5% as the benefit based on the midrange of the NESCCAF
estimate and the fact that ISG systems in production today in the Saturn Vue and Aura do offer
some regeneration and power assist capability.
Figure 2.5-1 Schematic of ISG System [Husted, 2003]
Motor/
Generator
Controller
Inverter
Battery
2.5.2 Integrated Motor Assist (IMA)/Integrated Starter-Alternator-Dampener (ISAD)
Hybrid
Integrated Motor Assist (IMA) and Integrated Starter-Alternator-Dampener (ISAD) are
similar systems developed and marketed by Honda and Continental, respectively. Honda's
Integrated Motor Assist (IMA) utilizes a thin axial electric motor bolted to the engine's
crankshaft and connected to the transmission through a torque converter or clutch (see Figure
2.5-2). This electric motor acts as both a motor for helping to launch the vehicle and a generator
for recovering energy while slowing down. It also acts as the starter for the engine and the
20
-------
electrical system's main generator. Since it is rigidly fixed to the engine, if the motor turns, the
engine must turn also, but combustion does not necessarily need to occur. The Civic Hybrid uses
cylinder deactivation on all four cylinders for decelerations and some cruise conditions. The
Accord Hybrid also has cylinder deactivation, but it is on one bank of the V-6 engine and
activates during cruise conditions as well as decelerations. This system does not launch the
vehicle electric power alone, although on the Civic, the vehicle can cruise on electric power
during some conditions.
Figure 2.5-2 Schematic of Honda IMA System [Husted, 2003]
<
Controller
Inverter
Motor/
Generator
/'
Trans
Final Drive
Another application of this type of technology has been developed by Daimler for hybrid and
plug-in hybrid versions of the Sprinter delivery van. The major driveline difference between
Honda's IMA system and the Daimler system is a clutch between the engine and electric motor
on the Daimler system. (This is largely enabled by the longitudinal arrangement of the
powertrain in the Sprinter vs. the transverse arrangement in the front-wheel-drive Hondas.) The
clutch allows for some extra efficiency by completely decoupling the engine from the electric
motor and driveline under conditions where the engine is not running.
Since hybrids in general were relatively new technology at the time of the 2002 NAS report,
we relied on a combination of certification data (comparing vehicles available with and without a
hybrid system and backing out other components where appropriate—see Tables 2.5-1 and 2.5-2).
and manufacturer-supplied information to determine that the effectiveness of these systems in
terms of CC>2 reduction is 30% for small cars, 25% for large cars, and 20% for minivans and
small trucks, This 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.
21
-------
Table 2.5-1 Small Car IMA Certification Data
Civic Sedan 1.8L5-auto
Civic HEV 1.3LCVT
Raw % difference
CVT (on HEV only)
Net difference
Tailpipe CO2
City
296
181
-39%
Hwy
222
174
-22%
55/45 comb.
269
178
-34%
3.5%
-29%
Table 2.5-2 Large Car IMA Certification Data
Accord Sedan 3.0L5-auto
Accord HEV 3.0L, cyl.
deac, 5 -auto
Net % difference
Tailpipe COi
City
444
317
Hwy
306
254
55/45 comb.
386
286
-26%
For large trucks, there is no certification data to use for analysis, however there have been
several concept versions of Sprinter van hybrid and plug-in hybrid conversions that use this type
of hybrid drive system. Published reports from these hybrid concepts indicates that these
vehicles can achieve a 10% to 50% decrease in fuel consumption.6 Although the Sprinter is
capable of hauling loads similar to other large trucks sold in the U.S., it is not a directly
comparable vehicle due to its different construction. The Sprinter does not have the same towing
capacity of other large trucks sold in the U.S., and it is not designed for off-road use.
Nevertheless, we estimate that an IMA-type hybrid system in a large truck can yield a CO2
reduction of 20% based on the published Daimler information and the known performance of
this type of system in the other vehicle classes.
2.5.:
Power-Split Hybrids
Power-Split hybrids are currently marketed by Ford, Nissan, and Toyota. They are
significantly different than other hybrid designs because they do not use a conventional
transmission. The Power Split system replaces the vehicle's transmission with a single planetary
gear and a motor/generator. A second, more powerful motor/generator is permanently connected
to the vehicle's final drive and always turns with the wheels (see Figure 2.5-3). The planetary
gear splits the engine's torque between the first motor/generator and the drive motor. The first
motor/generator uses its torque input to either charge the battery or supply additional power to
the drive motor. The speed of the first motor-generator determines the relative speed of the
engine to the wheels. In this way, the planetary gear allows the engine to operate completely
independently of vehicle speed, much like a CVT.
22
-------
Figure 2.5-3 Schematic of Aisin/Toyota Power Split System [Husted, 2003]
Planetary
Wheel
The Power Split system allows for outstanding fuel economy in city driving. The vehicle
also avoids the cost of a conventional transmission, replacing it with a much simpler single
planetary and motor/generator. However, its highway efficiency is not optimized due to the
requirement that the first motor/generator must be constantly spinning at a relatively high speed
to maintain the correct ratio of engine speed to final drive speed. Also, load capacity is limited
to the first motor/generator's capacity to resist the reaction torque of the drive train. Newer-
generation Power Split systems, however, are reducing these limitations.
We believe Power Split hybrids will be used mainly on small cars, large cars, minivans, and
small trucks. We did not analyze the Power Split system on large trucks because of a lack of
certification data to extract an effectiveness and the fact that there does not seem to be any real
movement by manufacturers to introduce this particular technology on large trucks (although this
technology is scalable to large trucks). We 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 35% for the classes to which
it is applied. (See Table 2.5-3 and Table 2.5-4) Future generations of this technology will
certainly significantly improve on this technology to achieve greater CC>2 reductions, but this
analysis does not take these into account. We did not rely on NAS or NESCCAF because NAS
did not cover the technology in depth and NESCCAF used a comparison of certification data as
well.
23
-------
Table 2.5-3 Large Car Power Split Certification Data
Nissan Altima
3.5LCVT
HEV2.5LPS
Net % difference
Toyota Camry
3. OL 5 -auto
HEV2.4LPS
Net % difference
Lexus GS
4.3L 6-auto
HEV3.5LPS
Net % difference
Tailpipe CO2
City
444
317
404
222
493
355
Hwy
306
254
286
234
355
317
55/45 comb.
386
286
-26%
355
228
-36%
423
341
-19%
Table 2.5-4 Small Truck Power Split Certification Data
Ford Escape 4X4
3.0L4-auto
HEV2.3LPS
Net % difference
Ford Escape 4X2
3.0L4-auto
HEV2.3LPS
Net % difference
Toyota Highlander 4X4
3.3L5-auto
HEV3.3LPS
Net % difference
Tailpipe COi
City
467
277
444
247
493
286
Hwy
386
306
370
286
370
329
55/45 comb.
423
286
-32%
404
261
-35%
423
306
-28%
24
-------
2.5.4
Two-Mode Hybrids
GM , Chrylser, Daimler and BMW have formed a joint venture to develop a new HEV
system based on HEV transmission technology originally developed by GM's Allison
Transmission Division for heavy-duty vehicles like city buses. This technology uses an
adaptation of a conventional stepped-ratio automatic transmission by replacing some of the
transmission clutches with two electric motors, which makes the transmission act like a CVT.
Like Toyota's Power Split design, these motors control the ratio of engine speed to vehicle
speed. But unlike the Power Split system, clutches allow the motors to be bypassed, which
improves both the transmission's torque capacity for heavy-duty applications and fuel economy
at highway speeds. (See Figure 2.5-4)
Figure 2.5-4 Schematic of GM-DCX HEV System [Hargitt 2002]
Erfaine
Clutch
There is considerably less information about the effectiveness of Two-Mode hybrid systems
than for Power-Split systems; however it is expected that the effectiveness will be slightly higher
than Power-Split because of the higher efficiency on the highway. We assume this technology is
not applicable to small cars, but on large cars, minivans and small trucks, the effectiveness is
40%.
Large trucks, on the other hand, are being developed by GM and Chrysler. During its
development GM stated that they expected at least a 25% fuel economy increase (20% CC>2
decrease) from this system on a large SUV. This lower value compared to the other vehicle
classes is due mainly to the lack of engine downsizing in a large truck in order to maintain full
towing capability even in situations with low battery charge. It is difficult to directly compare
data for the Tahoe Hybrid to a conventional version, but preliminary data suggests that a 25%
fuel consumption reduction is appropriate. We therefore chose a value of 25% CC>2 decrease for
the Two Mode system in a large truck.
25
-------
2.5.5
Full-Series Hydraulic Hybrids
A Full Series Hydraulic Hybrid Vehicle (HHV) is somewhat similar in concept to a full-
series electric hybrid vehicle, except that the energy is stored in the form of compressed nitrogen
gas and the power is transmitted in the form of hydraulic fluid (See Figure 2.5-5).
Figure 2.5-5 Schematic of Series Hydraulic Hybrid System
High Pressure
Accumulator
Wheel
Hydraulic
Engine Pump
Low Pressure
Accumulator
Hydraulic Drive
PumtAMotor
Series HHV technology currently under development by EPA is capable of a 40% decrease
in tailpipe CC>2 emissions in the small car, large car, minivan, and small truck classes. In the
large truck class, a 30% CO2 reduction is possible. The large truck benefit is somewhat lower
than the other classes because it is assumed that a large truck requires a larger engine to maintain
towing and hauling performance after the energy in the high pressure hydraulic accumulator is
exhausted. This technology is still under development and not yet commercialized, however
there are technology demonstration vehicles in service with UPS in daily package delivery
service.
2.5.6
Plug-in Hybrid Electric Vehicles
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 (usually 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 2.5-5 below, illustrates how PHEVs compare functionally to both hybrid electric
vehicles (HEV) and battery electric vehicles (BEV). These characteristics can change
significantly within each class, 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 conventional vehicles on one end and pure electric vehicles on the other.
26
-------
Table 2.5-5 Conventional, HEVs, PHEVs, and BEVs Compared
Increasing Electrification 7
Attribute
Drive Power
Engine Size
Electric Range
Battery
Charging
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
BEV
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.8
PHEVs also provide electric utilities the possibility to increase electric generation during "off-
peak" periods overnight when there is excess generation capacity and electricity prices are lower.
Utilities like to increase this "base load" because it increases overall system efficiency and
lowers average costs. PHEVs can lower localized emissions of criteria pollutants and air toxics
especially in urban areas by operating on electric power. The emissions from the power
generation occur outside the urban area at the power generation plant which provides health
benefits for residents of the more densely populated urban areas. Unlike most other alternative
fuel technologies, PHEVs can use existing infrastructure for fueling with gasoline and electricity
so large investments in fueling infrastructure are not required.
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 EPA is studying this issue on an on-going basis, upstream CO2 emissions are not
unique to grid-connected technologies and so are not included in this analysis of tailpipe CC>2
emissions.
PHEVs will be considerably more costly than conventional vehicles and some other
advanced technologies. To take advantage of their capability, consumers would have to be
willing to charge the vehicles nightly, 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 vs. on-peak (although the oil savings will still be the
same in this case assuming the vehicle can be charged fully).
B For PHEVs herein, we define electric range as the sum of all the electrified miles in charge-depleting mode (before
the battery reaches a minimum state-of-charge and the vehicle reverts to charge-sustaining mode). Charge-depleting
mode may be interrupted by periods of engine-on operation, but is not necessarily ended by the engine turning on.
27
-------
The CC>2 reduction potential of PHEVs depends on many factors, the most important being
the electrical capacity designed into the battery pack. To estimate the tailpipe CC>2 reduction
potential of PHEVs, EPA has developed an in-house vehicle energy model (PEREGRIN) to
estimate the CC>2 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 model0.
We modeled the PHEV small car, large car, minivan and small trucks 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. We designed each PHEV with
enough battery capacity for a 20-mile-equivalent all-electric range 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, 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.0
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 CO2 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 2.5-6)
Table 2.5-6 Sample Calculation of PHEV Gasoline-Equivalent CO2 Reduction
EV energy comb (0.55 city / 0.45 hwy)
EV range (from PEREGRIN)
SAE J1711 utility factor
HEV mode comb FE (0.55 city / 0.45 hwy)
Total IJF-adjusted FE (UF*FCEv + (l-UF)*FCHEv)
Baseline FE
Percent FE gain
Percent CC>2 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 tailpipe CC>2 reduction based on model outputs and the Utility Factor
calculations, results in a 58% CO2 reduction for small cars, large cars, minivans, and small
trucks. For large trucks, the result is a 47% reduction. The lower improvement is due to less
engine downsizing in the large truck class.
2.6 Full Electric Vehicles
The recent intense interest in Hybrid vehicles and the development of Hybrid vehicle battery
and motor technology has helped make Electric Vehicle technology more viable than it has ever
been. Electric Vehicles require much larger batteries than either HEVs or PHEVs, but the
c PERE can be downloaded at http://www.epa.gov/otaq/models/ngm/pere.zip
D General Motors is developing one of their PHEVs, the Volt, to have a 40 mile range. This vehicle uses a series
hybrid arrangement with the electric drive as the primary motive source and is of a very different design than the
PHEV concept studied in this report.
28
-------
batteries must be of a high-energy and lower-power design to deliver an appropriate amount of
power over the useful charge of the battery. These high-energy batteries are generally less
expensive per kilowatt-hour than high-power batteries required for hybrids, but the size of the
battery pack still incurs a considerable cost.
Electric motor and power electronics designs are very similar to HEV and PHEV designs, but
they must be larger, more powerful, and more robust since they provide the only motive power
for the vehicle. On the other hand, the internal combustion engine, fuel system, and possibly the
transmission can all be removed for significant weight, complexity and cost savings.
As for PHEVs, we modeled two full electric vehicles, a small car and a large car using the
same model (PEREGRIN) and similar assumptions. Full EVs are only considered for these two
classes because the larger, heavier vehicles would require too much battery capacity to be
practical in the short-to-mid term and we do not see any serious development activities in these
vehicle types in the market.
We chose to model the full EVs with a range of 150 miles on the urban driving cycle because
this range offers a good compromise in capability and battery cost, weight and size with expected
technology in the near- to mid-term. Using the same methodology as used for PHEVs to
calculate gasoline-equivalent fuel consumption, we obtained the results shown in Table 2.6-1,
below.
Table 2.6-1 Full Electric Vehicle Gasoline-Equivalent CO2 Reduction
EV energy comb (0.55 city / 0.45 hwy)
City cycle EV range
Highway cycle EV range
Baseline FE
Tailpipe CC>2 reduction
Small Car
0.202 kwh/mi
150 miles
166 miles
35.5
100%
Large Car
0.244 kwh/mi
150 miles
162 miles
25.3 mpg
100%
2.7 Vehicle Accessories
2.7.1
Electric Accessories and High Efficiency Alternator
The accessories on an engine - for example, the alternator, coolant and oil pumps - are
traditionally driven by the accessory belt, or directly off of the crankshaft. Direct benefit may be
obtained by improving their efficiency, or by driving them electrically (12V) only when needed
("on-demand"), and thereby reducing the accessory load relative to mechanically-driven systems.
Examples would be electric water or oil pumps, and mechanical fans on some large trucks.
Indirect benefit may be obtained by reducing the flow from the water pump electrically during
the engine warmup 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.
The estimated CC>2 reduction for combined accessory improvements is between 1 and 2
percent based on the NAS report. This estimate is also supported by confidential manufacturer
information. Air conditioning and power steering are other candidates for accessory load
29
-------
reduction, but they are addressed separately below and not included in this efficiency estimate.
Improved accessories as described above are available today.
2.7.2 Electric Power Steering for 12V and 42V systems
Electric power steering (EPS) is advantageous over hydraulic steering in that it only draws
power when the wheels are being turned, which is only a small percentage of a vehicle's
operating time. This eliminates the parasitics associated with belt-driven power steering pumps
in open-center steering systems, 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
may be implemented on many vehicles with a standard 12V system; however for heavier
vehicles, a 42V system may be required for compactness and reliability (which adds cost and
complexity). CC>2 reduction estimates for EPS range from 1.5 to 2 percent over a hydraulically
driven power steering system based on the 2002 NAS report. This range is in agreement with the
estimates provided by manufacturers. Electric power steering is available today.
2.7.3 Upgrade Electrical Systems to 42V
Most vehicles today (aside from hybrids) operate on 12 V electrical systems. At higher
voltages, the power density of motors, solenoids, and other electrical components increases to the
point that new and more efficient systems, such as electric A/C compressors and electric power
steering (for heavier trucks) and may be feasible. A 42-volt system also acts as an enabler for an
integrated starter generator. In addition to enabling other technologies, greater CC>2 reductions
are possible for improved accessories on a 42V system, of 1 to 2 percent incrementally (over
12V improved accessories) based on the higher voltage alone. When combined with 12-volt
improved accessories, these estimates are consistent with the NAS report.
2.8 Other Vehicle Technologies
2.8.1 Aerodynami c Drag F orce Reducti on
A vehicle's size and shape determine the amount of power needed to push the vehicle
through the air at different speeds. Changes in vehicle shape or frontal area can therefore reduce
CC>2 emissions. Areas for potential aerodynamic drag improvements include skirts, air dams,
underbody covers, and more aerodynamic side view mirrors. EPA estimates a fleet average of
20% total aerodynamic drag reduction is attainable for passenger cars, whereas a fleet average of
10% reduction is more realistic for trucks (with a caveat for "high-performance" vehicles,
described below). These drag reductions equate to CC>2 reductions of 2% and 3% for trucks and
cars, respectively. These numbers are in agreement with the technical literature and supported
by confidential manufacturer information.
Aerodynamic drag reduction technologies are readily available today, although the phase-in
time required to distribute over a manufacturer's fleet is relatively long (6 years or so).
2.8.2 Low Rolling Resistance Tires
Tire characteristics (e.g., materials, construction, and tread design) influence durability,
traction control, vehicle handling, and comfort. They also influence rolling resistance - the
30
-------
frictional losses associated mainly with the energy dissipated in the deformation of the tires
under load - and therefore, CC>2 emissions. This technology is applicable to all vehicles, except
for body-on-frame light trucks and performance vehicles (described in the next section).
Based on a 2006 NAS/NRC report, a 10% rolling resistance reduction would provide a CC>2
emissions reduction of 1 to 2 percent - and at this level the tires would maintain similar traction
and handling characteristics. Lower rolling resistance tires are widely available today.
2.8.3 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 shoes are pulled away from the rotating disc. While most
passenger cars have already adopted this technology, there are indications that this technology is
still available for body-on-frame trucks. Manufacturers have indicated that low drag brakes
could reduce CC>2 emissions up to 1 percent for these trucks. Low drag brakes are available
today.
2.8.4 Secondary Axle Disconnect (front axle for ladder frame and rear axle for
unibody frame)
To provide shift-on-the-fly capabilities, many part-time four-wheel drive systems use some
type of front axle disconnect. 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. Confidential manufacturer information suggests that front axle
disconnect for 4WD vehicles can reduce CC>2 emissions by 1.5 percent.
We are not aware of any manufacturer offering this technology in the US today on unibody
frame vehicles; however, we see no reasons why this technology could not be introduced by
manufacturers within the next one to two years.
2.8.5 Weight Reduction
While certainly an effective option for reducing CC>2 emissions, reducing the weight of
vehicles is a controversial topic. In the past, conventional wisdom held that a heavier vehicle is
more safe than a light one. Recently, however, some studies have challenged the weight-safety
connection, notably a June 2007 ICCT study, Sipping Fuel and Saving Lives: Increasing Fuel
Economy Without Sacrificing Safety (available at http://www.theicct.org/reports_live.cfm).
Also, during the public comment period for NHTSA's 2006 light truck CAFE rule, some auto
manufacturers, notably Volkswagen and Honda challenged the traditional weight-safety
-------
connection. This traditional position was also challenged by the Aluminum Association which
said that a 10% weight reduction is possible without affecting safety.7
The most promising way to reduce the weight of vehicles while maintaining vehicle size and
performance is through material substitution. Examples of materials substitution include using
higher-strength steel alloys, or even aluminum, magnesium or other light metals, in place of
conventional steel structural components. Additionally, other materials can be replaced with
lower density materials in other vehicle components, such as replacing plastics with lighter
weight plastics.
In addition to materials substitution, components and systems can be redesigned to reduce
weight, even while improving performance and reliability and lowering cost. An example would
be redesigning a subsystem replacing multiple components and mounting hardware with a
simpler system using advanced materials and a more integrated design.
Although EPA is not in a position today to provide estimates of the effectiveness or costs of
materials or strategies to reduce vehicle weight, we believe they will play an increasingly
important role in future efforts to reduce CC>2 emissions. Because of the importance of this
emerging field, EPA intends to study weight reduction technologies in depth in the near future.
32
-------
3 Synergistic Effects of Combining Multiple CO2 Reducing
Technologies
In Section 2 of this report, we present CC>2 reduction effectiveness estimates for a large
number of individual technologies. When considering a combination of technologies to reduce
CC>2 emissions, for the reasons discussed below, simply adding up the individual effectiveness
values of a package of technologies will lead to an incorrect result, which in most cases will be
an over prediction of the benefit of the combined technologies.
In estimating the aggregate effectiveness of combinations of multiple technologies, it is
important to recognize technologies that address the same categories of efficiency losses, such
that their combined effectiveness is appropriately accounted for, as opposed to using a simple
sum or product of their individual benefits. For example, a variable valvetrain system and a six-
speed automatic transmission both act to shift the engine operating points to a portion of the
engine speed/load map where pumping losses are less significant, and it is therefore reasonable
to anticipate a negative synergy, or dis-synergy, between such technologies. On the other hand,
a vehicle technology that reduces road loads at highway speeds (e.g., lower aerodynamic drag or
low rolling resistance tires) may extend the vehicle operating range over which cylinder
deactivation may be employed, and so such a combination may be expected to have a slightly
positive synergy. As the complexity of the technology combinations is increased, and the
number of interacting technologies grows accordingly, it becomes increasingly important to
account for these synergies.
There are two methods EPA has used in order to account for the impact of combining
multiple technologies: lumped-parameter analysis, and full-scale vehicle simulation modeling.
In this Section, we discuss both of these techniques and how they can account for the impact of
combining multiple technologies into a "package" and the associated synergistic impacts
between technologies. We also discuss how full-scale vehicle simulation modeling, which while
generally more robust is also more resource intensive, can be used to validate results from the
lumped-parameter approach.
Full-scale vehicle simulation modeling is one of the most accurate and robust means for
determining synergies between technologies. In order to assess these synergies, EPA
commissioned rigorous, detailed vehicle simulation work with Ricardo, Inc. Ricardo is a global
leader in automotive design, engineering and simulation, and their software is used by the
automotive industry in the design of engines, transmissions and vehicles. The results of their
simulation work were analyzed and used to validate synergy estimates generated from a first-
order "lumped parameter" analysis. The lumped parameter analysis was a tool used by EPA for
the purpose of estimating technology synergies, and is based upon vehicle efficiency
characteristics published in the technical literature. The lumped parameter analysis method
(Section 3.1) and the Ricardo vehicle simulation modeling (Section 3.2) are described below.
Section 3.3 discusses a comparison between the two methods, and finally Section 3.4 describes
how the lumped parameter analysis can be used to estimate synergy pairs when technologies are
applied in a pre-defined flow path order.
33
-------
3.1 EPA's Lumped Parameter Approach for Determining Effectiveness Synergies
EPA engineers reviewed existing tools that could be used to develop estimates of the
technology synergies, including the NEMS model8. However, the synergies in the NEMS
model depend heavily upon an assumed technology application flow path; those technologies
that the model would apply first would be expected to have fewer synergies than those applied
later on. For this reason, and because this report includes many new technologies not available
in NEMS, it was necessary for EPA to develop its own set of estimates. EPA used a well-
documented engineering approach known as a lumped-parameter technique to determine values
for synergies. At the same time, however, EPA recognized the availability of more robust
methods for determining the synergistic impacts of multiple technologies on vehicle CC>2
emissions than the lumped-parameter approach, particularly with regard to applying synergy
effects differentiated across different vehicle classes, and therefore augmented this approach with
the detailed vehicle simulation modeling described in Section 3.2.
The basis for EPA's lumped parameter analysis is a first-principles energy balance that
estimates the manner in which the chemical energy of the fuel is converted into various forms of
thermal and mechanical energy on the vehicle. The analysis accounts for the dissipation of
energy into the different categories of energy losses, including each of the following:
• Second law losses (thermodynamic losses inherent in the combustion of fuel),
• Heat lost from the combustion process to the exhaust and coolant,
• Pumping losses, i.e., work performed by the engine during the intake and exhaust
strokes,
• Friction losses in the engine,
• Transmission losses, associated with friction and other parasitic losses,
• Accessory losses, related directly to the parasitics associated with the engine
accessories and indirectly to the fuel efficiency losses related to engine warmup,
• Vehicle road load (tire and aerodynamic) losses;
with the remaining energy 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 categorized into the major types of engine losses it reduces, so that
interactions between multiple technologies applied to the vehicle may be determined. When a
technology is applied, its effects are estimated 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. Figure
3.1-1 below is an example spreadsheet used by EPA to estimate the synergistic impacts of a
technology package for a standard-size car.
34
-------
Figure 3.1-1. Sample Lumped Parameter Spreadsheet
EPA Staff Deliberative Materials-Do Not Quote or Cite
Vehicle Energy Effects Estimator
Vehicle type: Standard Car
Family
Description: Technology picklist
Package: Z
Baseline
% of fuel
Reduction
% of original fuel
Baseline
New
Indicated
Efficiency
Indicated Energy
Brake Energy
Vehicle
Mass
Inertia
Load
13.0%
0%
13.0%
Mech
Efficiency
38.0% 71.1%
38.2% 82.5%
Road Loads
Drag
Aero
Load
4.0%
16%
3.4%
Brake
Efficiency
27.0%
31.5%
Tires
Rolling
Load
Parasitics
Access
Losses
4.0% 1.8%
8% 64%
3.7% 0.8%
Drivetrain
Efficiency
Fuel
Efficiency
77.8% 21.0%
87.2% 27.5%
Gearbox,
T.C.
Trans
33%
Road
Loads
100.0%
95.4%
Engine Friction
Friction Pumping
Losses Losses
Heat
Lost To
Exhaust &
Coolant
IndEff
Losses
Second
Law
6.6% 4.4% 32.0% 30.0%
5.6% 1.1% 31.8% 30%
Current Results
72.9% Fuel Consumption
27.1% FC Reduction
37.2% FE Improvement
N/A Diesel FC Reduction
Original friction/brake ratio
Based on PMEP/IMEP »»
(GM study)
PMEP Brake
Losses Efficiency
11% 27%
71.1% mech efficiency
Technology
Aero Drag Reduction
Rolling Resistance Reduction
Low Fric Lubes
EF Reduction
ICP
DCP
CCP
Deac
DVVL
CVVL
Camless
GDI
Turbo/Dnsize
5-spd
CVT
ASL
Agg TC Lockup
6-spd auto
AMT
42V S-S
12V acc + Imp alt
EPS
42V ace + imp alt
HCCI dual-mode
GDI (lean)
Diesel - LNT
Diesel - SCR
Opt. E25
Independent
FC Estimate
3.0%
1.5%
0.5%
2.0%
2.0%
3.0% total VVT
3.0% total VVT
6.0%
4.0%
5.0%
10.0%
1.5%
6.0%
2.5%
6.0%
1.5%
0.5%
5.5%
6.5%
7.5%
1.5%
1.5%
3.0%
11.0%
10.5%
30.0% over gas
30.0% over gas
8.5%
Loss Category
Aero
Rolling
Friction
Friction
Pumping
Pumping
Pumping
Pumping, friction
Pumping
Pumping
Pumping
IndEff
Pumping
Trans, pumping
Trans, pumping
Pumping
Trans
Trans, pumping
Trans
F, P, A
Access
Access
Access
Ind. Eff, pumping
Ind. Eff, pumping
Ind Eff, pumping
Ind Eff, pumping
Ind. Eff, pumping
User Picklist
Implementation into estimator Include? (0/1)
16% aero (cars), 10.5% aero (trucks)
8% rolling
2% friction
8.5% friction
12% pumping, 38.2% IE, -2% flic
18.5% pumping, 38.2% IE, -2% flic
18.5% pumping, 38.2% IE, -2% flic
39% pumping
30% pumping, -3% friction
37% pumping, -3% friction
76% pumping, -5% friction
38. 6% Ind Eff
39% pumping
22% pumping, -5% trans
46% pumping, -5% trans
9.5% pumping
2. 5% trans
42% pumping, -5% trans
35% trans (increment)
13% friction, 19% pumping, 38% access
18% access
18% access
36% access
41% IE, 25% pumping
40% IE, 38% pumping
48% IE, 85% pumping, -13% friction
46% IE, 80% pumping, -13% friction
39% IE, 40% pumping
1
1
1
1
0
0
1
0
1
0
0
0
0
0
0
1
1
1
1
1
0
1
1
0
0
0
0
0
Table 3.1-1 below lists the technologies considered in this example, their corresponding
individual technology effectiveness values, and a comparison of the gross combined package
CC>2 reduction (i.e. disregarding synergies) to the lumped parameter results. The difference is
the implied synergistic effects of these technologies combined on a package.
35
-------
Table 3.1-1 Comparison of Lumped Parameter Analysis with Standard Car Package
Technology
Aero Drag Reduction
Rolling Resistance Reduction
Low Friction Lubricants
Engine Friction Reduction
VVT - coordinated cam phasing
VVL - discrete variable lift
Aggressive shift logic
Early torque converter lockup
6-speed automatic trans
AMI (6-speed)
Stop-Start with 42 volt system
Electric power steering
42V ace + improved alternator
Individual
C02
Reduction
3.0%
1.5%
0.5%
2.0%
3.0%
4.0%
1.5%
0.5%
5.5%
6.5%
7.5%
1.5%
3.0%
Gross combined effectiveness
Lumped parameter estimate
Estimated synergistic effects
33.6%
27.1%
-6.5%
Cumulative
C02
Reduction
3.0%
4.5%
4.9%
6.8%
9.6%
13.2%
14.5%
15.0%
19.6%
24.9%
30.5%
31.5%
33.6%
The synergy estimates obtained using the lumped parameter technique were subsequently
compared to the results from the vehicle simulation work. EPA will continue to use the lumped
parameter approach as an analytical tool, and (using the output data from the vehicle simulation
as a basis) may adjust the synergies as necessary in the future.
3.2 Ricardo's Vehicle Simulation
Vehicle simulation modeling was performed by Ricardo, Inc. The simulation work
addressed gaps in existing synergy modeling tools, and served to both supplement and update the
earlier vehicle simulation work published by NESCCAF. Using a physics-based, second-by-
second model of each individual technology applied to various baseline vehicles, the Ricardo
model was able to estimate the effectiveness of the technologies acting either individually or in
combination. This information could then be used to estimate the synergies of these technology
combinations, and also to differentiate the synergies across different vehicle classes.
In total, Ricardo modeled five baseline vehicles and twenty-six distinct technology
combinations, covering the full range of gasoline and diesel powertrain technologies used in the
Volpe model, with the exception of the powersplit, plug-in and two-mode hybrid vehicle
technologies. The five generalized vehicle classes modeled were a standard car, a full-size car, a
small multi-purpose vehicle (MPV), a large MPV and a large truck. The complete list of
vehicles and technology packages is given below in this section, along with a detailed
explanation of the selection criteria.
36
-------
Each technology package was modeled under a constraint of "equivalent performance" to the
baseline vehicle. To quantify the performance, a reasonably comprehensive, objective set of
vehicle performance criteria were used as a basis to compare with the baseline vehicle,
characterizing the launch acceleration, passing performance and grade capability that a vehicle
buyer might expect when considering a technology package. The main metrics used to compare
vehicle performance are listed below in Table 3.2-1.
Table 3.2-1 Performance Metrics Used as Basis for "Equivalent Performance"
Characteristic
Overall
Performance
Launch
Acceleration
Passing
Performance
Grade
Capability
Performance Metric
Time to accelerate from 0-60 mph*
Time to accelerate from 0-30 mph
Vehicle speed and distance after a 3-second
acceleration from rest
Time to accelerate from 30 to 50 mph
Time to accelerate from 50 to 70 mph
Maximum % grade at 70 mph
(standard car, large car, small MPV and large MPV)
Maximum % grade at 60 mph at GCVWR (large truck)*
Notes:
All accelerations are assumes at WOT (wide-open throttle) condition
GCVWR = EPA Gross Combined Vehicle Weight Rating
A summary of the vehicle simulation results is given below in Section 3.2.1.5, including the CC>2
emissions reduction effectiveness for each technology package. The full Ricardo vehicle
simulation results, including the acceleration performance data, may be found in Ricardo's final
report posted publicly at EPA's website.9
3.2.1 Description of Ricardo's Report
In this section, the structure, methodology and results from the Ricardo vehicle simulation
report are summarized. EPA worked closely with Ricardo to develop baseline models of five
generalized vehicle classes that could be validated against EPA certification data, and then used
as a platform upon which to add various technology packages. The vehicle simulation modeling
results generated by Ricardo consist of the following:
• Baseline vehicle characterization, to determine the baseline fuel consumption and CC>2
emissions over the EPA combined cycle federal test procedure (FTP) for five baseline
vehicles, for validation with EPA certification data.
• Simulation of the vehicle technology combinations (applied to the baseline vehicles)
• Incremental technology effectiveness estimates, to examine the effect of adding
technologies one-by-one. These could then be used more directly to validate synergies
estimated using the lumped parameter method.
This section describes the selection process for each of the baseline vehicles and the
technology packages, and summarizes the results of the vehicle simulation.
37
-------
3.2.1.1 Determination of representative vehicle classes
In an effort to establish a reasonable scope for the vehicle simulation work and to update the
earlier simulation done by NESCCAF, EPA chose five representative vehicle classes as the basis
for evaluating technology benefits and synergies, representing the vehicle attributes of the
projected highest-volume light-duty car and truck sales segments. These five classes covered a
broad range of powertrain and vehicle characteristics, over which the effectiveness and synergies
of each of the technologies could be evaluated. The main distinguishing attributes of the five
vehicle classes considered by EPA and Ricardo are given below in Table 3.2-2.
Table 3.2-2 Attributes of the Five Generalized Vehicle Classes Considered by Ricardo
Vehicle Class
EPA Vehicle Types
Included
Curb Weight
Range
Engine Type
Drivetrain
Body Type
Towing Capability
Example Vehicles
Standard Car
Compact, Mid-
size
2800-3600 Ibs
14
FWD
Unibody
None
Toyota Camry,
Chevy Malibu,
Honda Accord
Large Car
Large car
>3600 Ibs
V6
RWD/AWD
Unibody
None
Chrysler 300,
Ford 500 /
Taurus
Small MPV
Small SUV,
Small Pickup
3600-4200
Ibs
14
FWD
Unibody
Partial
Saturn VUE,
Ford Escape,
Honda CR-V
Large MPV
Minivans, Mid-
SUVs
4200-4800 Ibs
V6
FWD/AWD
Unibody
Partial
Dodge Grand
Caravan,
CMC Acadia,
Ford Flex
Large Truck
Large SUVs,
Large Pickups
>4800 Ibs
V8
4WD
Ladder Frame
Full
Ford F-150,
Chevy Silverado
1500, Dodge
Ram
EPA then selected representative vehicle models for each of these classes, based on three main
criteria:
• The vehicle should possess major attributes and technology characteristics that are near
the average of its class, including engine type and displacement, transmission type, body
type, weight rating, footprint size and fuel economy rating.
• It should be among the sales volume leaders in its class, or where there is not a clearly-
established volume leader, the model should share attributes consistent with major
sellers.
• The vehicle should have undergone a recent update or redesign, such that the technology
in the baseline model could be considered representative of vehicles sold at the beginning
of the proposed regulatory timeframe.
Consideration was also given to include the sales-leading vehicle manufacturers among the
baseline models. Hence, the U. S. domestic manufacturers account for four of the five models
(Chrysler 300, GM/Saturn Vue, Chrysler/Dodge Caravan, and the Ford F-150), while import
manufacturers are represented in their strongest sales segment, the standard car class, by the
Toyota Camry.
38
-------
3.2.1.2 Description of Baseline Vehicle Models
The baseline vehicles selected to represent their respective vehicle classes are described
below in Table 3.2-3, listed with the critical attributes that EPA used as selection criteria. While
each attribute for these baseline vehicles does not match the precise average for its class, each of
these baselines is an actual vehicle platform that allows validation of the simulation data with
"real world" certification data.
Table 3.2-3 Description of Baseline Vehicles
Vehicle Class
Baseline Vehicle
CO2 Emissions* (g/mi)
in
4-*
D
.Q
'^
32
<
L •*-*
£«
O ns
ao
Base Engine
Displacement (L)
Rated Power (HP)
Torque (ft-lbs)
Valvetrain Type
Valves per Cyl
Drivetrain
Transmission
Number of Forward
Speeds
CurbWt(lbs)
ETW (Ibs)
GVWR (Ibs)
GCWR (Ibs)
Front Track Width
(in.)
Wheelbase (in.)
Displacement /
Weight Ratio
(L/ton)
Power /Weight
Ratio (HP/ton)
Standard Car
Toyota Camry
327
DOHC 14
2.4
154
160
VVT (DCP)
4
FWD
Auto
5
3108
3500
-
-
62
109.3
1.54
99.1
Full Size Car
Chrysler 300
409
SOHC V6
3.5
250
250
Fixed
4
RWD
Auto
5
3721
4000
-
-
63
120
1.88
134.4
Small MPV
Saturn VUE
415
DOHC 14
2.4
169
161
VVT (DCP)
4
FWD
Auto
4
3825
4000
4300
-
61.4
106.6
1.25
88.4
Large MPV
Dodge Grand
Caravan
435
OHVV6
3.8
205
240
Fixed
2
FWD
Auto
4
4279
4500
5700
-
63
119.3
1.78
95.8
Large Truck
Ford F-150
575
SOHCV8
5.4
300
365
VVT (CCP)
3
4WD
Auto
4
5004
6000
6800
14000
67
144.5
2.16
119.9
*-Estimated CC>2 equivalent, taken from EPA adjusted combined fuel economy ratings.
3.2.1.3 Technologies Considered by EPA and Ricardo in the Vehicle Simulation
A number of advanced gasoline and diesel technologies were considered in the Ricardo
study, comprising the majority of the technologies used in the Volpe model, with the exception
of the hybrid electric vehicle technologies. In developing a comprehensive list of technologies to
be modeled, EPA surveyed numerous powertrain and vehicle technologies and technology
trends, in order to assess their potential feasibility in the next one to ten years. The list of
technologies considered therefore includes those that are available today (e.g., variable valve
timing, six-speed automatic transmissions) as well as some that may not be ready for five to ten
39
-------
years (e.g., camless valve actuation and HCCI engines). Table 3.2-4 below lists the technologies
that Ricardo included in the vehicle simulation models.
Tal
)le 3.2-4. Technologies Included in the Ricardo Vehicle Simulation
Engine Technologies
Abbrev.
DOHC
SOHC
OHV
CCP
DCP
DVVL
CVVL
Deac
CVA
Turbo
GDI
Diesel
HCCI
LUB
EFR
Description
Dual Overhead Camshaft
Single Overhead Camshaft
Overhead Valve (pushrod)
Coordinated cam phasing
Dual (independent) cam phasing
Discrete (two-step) Variable Valve Lift
Continuous Variable Valve Lift
Cylinder Deactivation
Camless Valve Actuation (full)
Turbocharging with engine downsizing
Gasoline Direct Injection
Diesel with advanced aftertreatment
Homogeneous Charge Compression Ignition (gasoline)
Low-friction engine lubricants
Engine friction reduction
Transmission Technologies
Abbrev.
L4
L5
L6
DCT6
CVT
ASL
TORQ
Description
Lockup 4-speed automatic transmission
Lockup 5-speed automatic transmission
Lockup 6-speed automatic transmission
6-speed dual clutch automated manual transmission
Continuously variable transmission
Aggressive shift logic
Early torque converter lockup
Accessory Technologies
Abbrev.
ISG (42V)
EPS
EACC
HEA
Description
42V Integrated Starter-Generator
Electric Power Steering
Electric Accessories (water pump, oil pump, fans)
High-Efficiency Alternator
Vehicle Technologies
Abbrev.
AERO
ROLL
Description
Aerodynamic drag reduction (10%-20%)
Tire rolling resistance reduction (10%)
40
-------
3.2.1.4 Choice of Technology Packages
EPA chose a number of technology packages representing a range of options that
manufacturers might pursue. In determining these technology combinations, EPA considered
available cost and effectiveness numbers from the literature, and applied engineering judgment
to match technologies that were compatible with each other and with each vehicle platform.
Also, where appropriate, we applied the same technologies to multiple vehicle classes, to
determine where specific vehicle attributes might affect their benefits and synergies. These
technologies represent most of those listed in Section 2 of this report. Table 3.2-5 below
describes in detail the technology content in each technology package simulated by Ricardo.
Table 3.2-5 Description of the Vehicle Technology Packages Modeled by Ricardo
Vehicle
E
« *
re °
^
Q.
^
"re
E
if\
re
o>
_N
i/5
LL
Q> •>,
Pa.
™ ^
x.
o
£
d>
O)
ll
re
_i
Technology
Package
baseline
Z
1
2
baseline
Z
1
2
15
15a
15b
5
baseline
4
5
Y1
Y2
6a
16
baseline
4
6b
16
baseline
9
10
11
12
17
X1
X2
Engine
2.4-Liter 14
2.4LI4, PFI
2.4LI4, GDI
2.4LI4, GDI
2.4-Liter 14
2.4LI4, PFI
2.4LI4, GDI
2.4LI4, GDI
1.5LI4, GDI, Turbo
2.4LI4, GDI
2.4LI4, GDI, HCCI
1.9LI4, Diesel
3.5-Liter V6
2.2LI4, GDI, Turbo
2.8L 14, Diesel
3.5LV6, GDI
3.5LV6, GDI, HCCI
3.0LV6, GDI
3.5LV6, GDI
3.8-Liter, V6
2.1LI4, GDI, Turbo
3.0LV6, GDI
3.8LV6, GDI
5.4-Liter V8
5.4LV8, GDI
3.6LV6, GDI, Turbo
4.8LV8, Diesel
5.4LV8, GDI
5.4LV8, GDI
5.4LV8, GDI
5.4LV8, GDI, HCCI
Valvetrain
DOHC, DCP
CCP, DVVL
DCP, DVVL
DCP
DOHC, DCP
CCP, DVVL
DCP, DVVL
DCP
DCP
CVA
DCP, CVVL
DOHC
SOHC
DCP
DOHC
CVA
DCP, CVVL
DCP, CVVL
CCP, Deac
OHV
DCP
CCP, Deac
CCP, Deac
SOHC, CCP
CCP, Deac
DCP
DOHC
CCP, Deac
DCP, DVVL
CVA
DCP, CVVL
Transmission
L5
DCT6
CVT
L6
L4
DCT6
CVT
L6
DCT6
DCT6
DCT6
DCT6
L5
L6
DCT6
DCT6
DCT6
DCT6
L6
L4
L6
DCT6
L6
L4
DCT6
DCT6
DCT6
L6
L6
DCT6
DCT6
Accessories
ISG (42V), EPS, EACC
EPS, EACC, HEA
ISG (42V), EPS, EACC
EPS
ISG (42V), EPS, EACC
EPS, EACC, HEA
ISG (42V), EPS, EACC
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
ISG (42V), EPS, EACC
EPS, EACC, HEA
EPS, EACC, HEA
ISG (42V), EPS, EACC
ISG (42V), EPS, EACC
EPS, EACC, HEA
EPS, EACC, HEA
ISG (42V), EPS, EACC
EPS, EACC, HEA
EPS, EACC, HEA
EPS, EACC, HEA
Other:
20% Aerodynamic drag reduction, 10% tire rolling resistance reduction assumed for all vehicles, except Large Truck
10% Aerodynamic drag reduction assumed for Large Truck
Low-friction lubricants and moderate engine friction reductions are assumed for all vehicles
Aggressive shift logic and early torque converter lockup strategies are assumed for all vehicles, where applicable.
41
-------
3.2.1.5 Simulation Results
The CC>2 emissions results from the vehicle simulation are summarized below in Table 3.2-6
(for cars) and Table 3.2.7 (for light-duty trucks). The CC>2 estimates are given for the combined
city and highway test cycles, according to the EPA Federal Test Procedure (FTP), with the
technology package results compared with the baseline vehicle as shown.
It is important to reiterate that each of the technology package results were obtained with
performance determined to be equivalent to the baseline vehicle. No attempt was made to
project trends in performance during the proposed regulatory period, nor did we downgrade
performance to give improved fuel efficiency. A full comparison of vehicle acceleration
performance is given in the Ricardo final report.
Table 3.2-6. CO2 Emissions Estimates Obtained from Vehicle Simulation (Cars)
Vehicle
Standard
car
>_
(0
o
o>
N
(/>
"5
LJ_
Technology
Package
baseline
Z
1
2
baseline
4
5
Y1
Y2
6a
16
Major Features*
2.4L 14, DCP, L5
CCP, DWL, DCT, ISG
GDI, DCP, DWL, CVT
GDI, DCP, L6, ISG
3.5L V6, L5
2.2LI4, GDI, Turbo, DCP, L6
2.8L 14 Diesel, DCT
GDI, CVA, DCT
GDI, HCCI, DCT
GDI, DCP, CWL, DCT
GDI, CCP, Deac, L6, ISG
CO2
City
g/mi
338
250
294
277
420
346
315
278
290
331
301
CO2
Hwy
g/mi
217
170
198
180
279
236
221
199
197
235
205
CO2
Combined
g/mi
284
214
251
233
356
296
273
242
248
288
257
CO2
Reduction
%
X
24.7%
1 1 .5%
17.8%
X
16.9%
23.5%
32.0%
30.4%
19.2%
27.7%
*-Please refer to Table 3.2-4 for a full description of the vehicle technologies
42
-------
Table 3.2-7. COi Emissions Estimates Obtained from Vehicle Simulation (Light-Duty
Trucks)
Vehicle
0.
"re
E
0) ^,
^
o
1-
0>
0)
(0
Technology
Package
baseline
Z
1
2
15
15a
15b
5
baseline
4
6b
16
baseline
9
10
11
12
17
X1
X2
Major Features*
2.4L 14, DCP, EPS
CCP, DWL, DCT, ISG
GDI, DCP, DWL, CVT
GDI, DCP, L6, ISG
1.5LI4, GDI, Turbo, DCP, DCT
GDI, CVA, DCT
GDI, HCCI, DCT
1.9LI4 Diesel, DCT
3.8L V6
2. 1LI4, GDI, Turbo, DCP, L6
GDI, CCP, Deac, DCT
GDI, CCP, Deac, L6, ISG
5.4L V8, CCP
GDI, CCP, Deac, DCT, ISG
3.6L V6, GDI, Turbo, DCP, DCT
4.8LV8 Diesel, DCT
GDI, CCP, Deac, L6, ISG
GDI, DCP, DWL, L6
GDI, CVA, DCT
GDI, HCCI, DCT
CO2
City
g/mi
367
272
310
291
272
262
270
282
458
357
333
325
612
432
404
444
459
492
422
425
CO2
Hwy
g/mi
253
208
227
211
212
193
197
205
313
256
248
225
402
315
319
326
328
333
314
311
CO2
Combined
g/mi
316
243
272
255
245
231
237
247
393
312
295
280
517
379
366
391
400
420
374
374
CO2
Reduction
%
X
23.0%
13.7%
19.3%
22.5%
26.8%
24.8%
21 .8%
X
20.6%
24.9%
28.7%
X
26.7%
29.3%
24.4%
22.6%
18.8%
27.8%
27.7%
*-Please refer to Table 3.2-4 for a full description of the vehicle technologies
3.3 Comparison of Lumped-Parameter Results to Modeling Results
Considering the following:
1) EPA's lumped-parameter package estimates are comparable with those obtained from the
detailed Ricardo simulations. This is illustrated in Figure 3.3-1 below.
2) EPA is confident in the plausibility of the individual technology effectiveness estimates
in Table 2.2-1 through Table 2.2-5, based on the sources from which that information
was assimilated, as detailed in Section 2 of this report.
3) Additionally, EPA expresses confidence in the overall Ricardo package results due to our
knowledge of the robust methodology used in building the models and generating the
results.
43
-------
Figure 3.3-1. Comparison of Ricardo package results to equivalent lumped parameter
package results
30%
25%
5%
o
w
Z
i
_
O
w
1
-, _
"1
ri R-i r
r
066666660000
lggggggg||||
2 Z 1 2 15 15a 15b 5 4 5 6a 16
r
I"!
-1
1 n
00&&&ggggggg
£? £P u u u
i-J i-J i-J
Yl Y2 4 6b 16 9 10 11 12 17 XI X2
Package
Based on this, EPA concludes that the synergies derived from the lumped parameter approach
are generally plausible (with a few packages that garner additional investigation). EPA will
continue to analyze this data, focusing on those packages where the differences between the two
approaches are large.
The simulation results may present opportunities to improve the fidelity of the lumped-parameter
approach by identifying differences between different platforms or important vehicle traits (such
as displacement-to-weight ratio, e.g.). There might also be opportunity to infer (through detailed
analysis) the individual effectiveness values for some technologies by comparing and isolating
Ricardo package results across different vehicle platforms.
3.4 Using the Lumped-Parameter Technique to Determine Synergies in a Technology
Application Flowpath (Identifying "Technology Pairs" to account for synergies)
In order to account for the real world synergies of combining of two or more technologies,
the product of their individual effectiveness values must be adjusted based on known
interactions, as noted above. When using an approach in which technologies are added
sequentially in a pre-determined application path to each individual vehicle model, as used in
NHTSA's 2006 fuel economy rule for light trucks10, these interactions may be accounted for by
44
-------
considering a series of interacting technology pairs. EPA believes that a lumped parameter
approach can be used as a means to estimate and account for synergies for such a technology
application method. When using a sequential technology application approach which applies
more than one technology, it is necessary to separately account for the interaction of each unique
technology pair. Moreover, if the sequential technology application approach applies a
technology that supersedes another, for example, where a VVLT system is substituted in place of
a cylinder deactivation system, its incremental effectiveness must be reduced by the sum of the
synergies of that technology with each individual technology that was previously applied,
regardless of whether any of them have also been superseded. Figure 3.4-1 below provides an
example of how technology pairs are identified for a specific technology application path similar
to one used by NHTSA. In this example, an interaction is identified between each of the engine
technologies (except GDI) with each of the transmission technologies. So, in this example, were
the model to couple a turbocharged and downsized GDI engine with a 6-speed transmission, it
would apply a series of many synergy pairs to the combined individual effectiveness values to
arrive at the overall effectiveness.
Figure 3.4-1 Illustration of technology pairings for a specific technology application path
Engine Technology Trans Technology
VVT (ICP)
1
1 VVT(
CCP) |
DISP
|VVLT(DWL)
L
T(D\
GDI
L
TURB
(Lines indicate potential synergies)
45
-------
4 Costs for Technologies
4.1 Methodology for Estimating Variable Piece Costs
This section describes the costs associated with the new vehicle technologies described in
Section 2. The costs described here represent the piece costs for an individual piece of hardware
or system, e.g., an intake cam phaser to provide variable valve timing. To estimate piece costs,
we relied upon a number of sources for cost related information. Our objective was to use those
sources of information that we considered to be most credible for projecting the costs of
individual vehicle technologies. These sources included: the 2002 NAS report on the
effectiveness and impact of CAFE standards;11 the 2004 study done by NESCCAF;12 the recent
California Air Resources Board (CARB) Initial Statement of Reasons in support of their carbon
rulemaking;13 a 2006 study done by Energy and Environmental Analysis (EEA) for the
Department of Energy;14 and our own vehicle fuel economy certification data. We also
considered confidential data submitted by vehicle manufacturers in response to NHTSA's
request for product plan information,15 and confidential information shared by automotive
industry component suppliers in meetings with EPA and NHTSA staff held during the second
half of the 2007 calendar year. These sources of data do not present their values in terms of 2006
dollars as was desired for this analysis. To adjust to 2006 dollars, we have used the appropriate
Producer Price Index as determined by the Department of Labor's Bureau of Labor Statistics,
and we present our methodology for these adjustments in Appendix 4. A to this report. Where
estimates differ between sources, we have used engineering judgment to arrive at what we
believe to be the best cost estimate available today, and explained the basis for that exercise of
judgment. The following discussion summarizes our piece cost estimates and how we used these
data sources to arrive at our best estimate of piece costs.
4.2 Piece Costs Assigned to COi Reduction technologies
Table 4.2-1 presents our estimated costs associated with the technologies we believe will be
used to reduce carbon dioxide emissions from passenger cars and light trucks. Following the
table is a detailed description of how each of these costs was developed. The costs are meant to
represent the incremental compliance costs for each technology. As such, these costs account for
both the direct manufacturing costs and the indirect costs. These indirect costs include
production-related costs (research, development, and other engineering), business-related costs
(salaries, pensions), or retail-sales-related costs (dealer support, marketing), and profits.16 For
this analysis, we first developed piece cost estimates for each technology or system at the auto
manufacturer level, i.e., the price paid by the manufacturer to a Tier 1 component supplier.E To
these costs, we then added an indirect cost markup factor of 50 percent to generate the
compliance costs presented in the table.F We believe that this indirect cost markup overstates the
incremental indirect costs because it is based on studies that include cost elements—such as
E A Tier 1 supplier is one that sells its products directly to the automobile manufacturer, or any original equipment
manufacturer (OEM) that sells its products at the consumer retail level. A Tier 2 supplier would be one that sells its
products to a Tier 1 supplier, and so on.
F We have a more detailed discussion of markup factors and what cost elements they capture in section 4.3 of this
report.
46
-------
funding of pensions—which we believe are unlikely to change as a result of the introduction of
new technology. Consequently, the incremental compliance costs we have developed should not
be understood as estimated price increases for vehicles, but rather an estimate of the incremental
cost of the technology at the retail level which accounts (conservatively) for all direct and
indirect OEM costs.0 We have a detailed discussion of markup factors in Section 4.3 of this
report.
Note that throughout this discussion we compare our estimated compliance costs to
manufacturer costs (i.e., costs without any indirect cost markup) and retail price equivalents (i.e.,
costs with indirect cost markups) from other studies. These comparisons are sometimes
necessarily imprecise given different markups and/or lack of markups among the estimates. In
addition, please note that the 2002 NAS study and the NESCCAF study used a markup of 40
percent to arrive at their retail price equivalent (RPE) estimates. Also please be aware that the
EEA study referred to throughout this discussion applied no markups, presenting only the
manufacturer's direct costs. Lastly, the CBI submittals reported RPEs but contained no
information as to how those RPEs were calculated (i.e., no information showing the direct cost
versus indirect cost portions).
We differentiate between cost and price where cost is meant to capture the concept of what it costs entity A to
produce and make available a product that can be purchased by entity B, while price is meant to capture the concept
of what entity B actually pays to entity A for the product. The price is generally higher than the cost but can also be
lower since so many factors impact the price. The incremental compliance costs we have developed are meant only
to capture all of the incremental business expenses that entity A - the original equipment auto manufacturer - would
incur.
47
-------
Table 4.2-1 Incremental Compliance Costs for Technologies
(2006 Dollars per Vehicle)
Technology
Incremental to
Vehicle Class
Small
Car
Large
Car
Minivan
Small
Truck
Large
Truck
Engine Technologies
Low friction lubricants
Engine friction reduction
Overhead Cam Engines
VVT - intake cam phasing
VVT - coupled cam phasing
VVT - dual cam phasing
Cylinder deactivation
Discrete VVLT
Continuous VVLT
Overhead Valve Engines
Cylinder deactivation
VVT - coupled cam phasing
Discrete VVLT
Continuous VVLT (includes
conversion to Overhead Cam)
Camless valvetrain (electromagnetic)
GDI - stoichiometric
GDI - lean burn
Gasoline HCCI dual-mode
Turbocharge+downsize
Diesel - Lean NOx trap
Diesel - urea SCR
Optimized E20-E30
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine
Base engine w/ VVT-coupled
Base engine
Base engine
GDI - stoich
GDI - stoich
Base engine
Base gasoline engine
Base gasoline engine
Base gasoline engine
3
0-84
59
59
89
n.a.
169
254
n.a.
59
169
599
336-673
122-420
750
263
690
2790
713
3
0-126
119
119
209
203
246
466
203
59
246
1262
336-673
204-525
750
390
120
3045
143
3
0-126
119
119
209
203
246
466
203
59
246
1262
336-673
204-525
750
390
120
3120
143
3
0-126
119
119
209
203
246
466
203
59
246
1262
336-673
204-525
750
390
120
3405
143
3
0-168
119
119
209
229
322
508
229
59
322
1380
336-673
228-525
750
685
810
4065
833
Transmission Technologies
Aggressive shift logic
Early torque converter lockup
5-speed automatic
6-speed automatic
6-speed AMT
6-speed manual
CVT
Base trans
Base trans
4-speed auto
4-speed auto
6-speed auto
5-speed man
4-speed auto
38
30
76-167
76-167
141
107
231
38
30
76-167
76-167
141
107
270
38
30
76-167
76-167
141
107
270
38
30
76-167
76-167
141
107
n.a.
38
30
76-167
76-167
141
107
n.a.
Hybrid Technologies
Stop-Start with 42 volt system
IMA/ISA/BSG (includes engine
downsize)
2-Mode hybrid electric vehicle
Power-split hybrid electric vehicle (P-
SHEV)
Full-Series hydraulic hybrid
Plug-in hybrid electric vehicle
(PHEV)
Full electric vehicle (EV)
Base engine w/ upgraded 42V
accessories & base trans
Base engine & trans
Base engine & trans
Base engine & trans
Base engine & trans
Base engine & trans
Base engine & trans
563
2477
3754
750
4500
12000
600
3153
4655
825
6750
15000
600
n.a
4655
825
6750
600
n.a
4655
900
6750
600
n.a
6006
1200
10200
Accessory Technologies
Improved high efficiency alternator &
electrification of accessories (12 volt)
Electric power steering (12 or 42
volt)
Improved high efficiency alternator &
electrification of accessories (42 volt)
Base accessories
Base accessories
Improved high efficiency alternator &
electrification of accessories (12 volt)
89-119
118-197
89-119
89-119
118-197
89-119
89-119
118-197
89-119
89-119
118-197
89-119
89-119
118-197
89-119
Vehicle Technologies
Aero drag reduction (20% on cars,
10% on trucks)
Low rolling resistance tires (10%)
Low drag brakes (ladder frame only)
Secondary axle disconnect (unibody
only)
Front axle disconnect (ladder frame
only)
Base vehicle
Base vehicle
Base vehicle
Base vehicle
Base vehicle
0-75
6
676
0-75
6
676
0-75
6
676
0-75
6
87
676
114
0-75
87
114
48
-------
For some of the technologies presented in Table 4.2-1, we believe that learning effects would
reduce future costs from the levels shown.H 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 EPA has often done in past
regulatory analyses—to apply at the industry-wide level particularly in industries that utilize
many common technologies and component supply sources.17 We believe there are factors that
cause hardware 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 which allows them to lower the per-unit
cost of production (i.e., the manufacturing learning curve).18
The learning curve is a well documented phenomenon. The general concept is that unit costs
decrease as cumulative production increases. Learning curves are often characterized in terms of
a progress ratio, where each doubling of cumulative production leads to a reduction in unit cost
to a percentage "p" of its former value (referred to as a "p cycle"). Organizational learning,
which brings about a reduction in total cost, is caused by improvements in several areas. Areas
involving direct labor and material are usually the source of the greatest savings. Examples
include, but are not limited to, a reduction in the number or complexity of component parts,
improved component production, improved assembly speed and processes, reduced error rates,
and improved manufacturing process. These all result in higher overall production, less
scrappage of materials and products, and better overall quality. As each successive p cycle takes
longer to complete, production proficiency generally reaches a relatively stable plateau, beyond
which increased production does not necessarily lead to markedly decreased costs.
Companies and industry sectors learn differently. In a 1984 publication, Button and Thomas
reviewed the progress ratios for 108 manufactured items from 22 separate field studies
representing a variety of products and services.19 The distribution of these progress ratios is
shown in Figure 4.2-1. Except for one company that saw increasing costs as production
continued, every study showed cost savings of at least five percent for every doubling of
production volume. The average progress ratio for the whole data set falls between 81 and 82
percent. Other studies (Alchian 1963, Argote and Epple 1990, Benkard 1999) appear to support
the commonly used p value of 80 percent, i.e., each doubling of cumulative production reduces
the former cost level by 20 percent.
H During the development of our cost estimates, EPA technical staff had several discussions and shared drafts of
technical write-ups with colleagues at the Department of Transportation's Volpe Center regarding the learning
curve. While the final write-up presented here on the theory and application of the learning curve was not reviewed
by DOT, it benefited greatly from the input and many suggested additions from technical staff at the Volpe Center.
49
-------
Figure 4.2-1 Distribution of Progress Ratios (Button and Thomas 1984)
Distribution of Progress Ratios
From 22 field studies (n = 108).
The learning curve is not the same in all industries. For example, the effect of the learning
curve seems to be less in the chemical industry and the nuclear power industry where a doubling
of cumulative output is associated with 11 percent decrease in cost (Lieberman 1984,
Zimmerman 1982). The effect of learning is more difficult to decipher in the computer chip
industry (Gruber 1992).
A typical experience curve can be described by three parameters: (1) the initial production
volume that must be reached before cost reductions begin to be realized (referred to as the
"threshold volume"); (2) the percentage at which costs are reduced with increases in cumulative
production beyond this initial volume (usually referred to as the "learning rate"); and (3) the
production volume after which costs reach a "floor," and further cost reductions no longer occur.
As such, a typical cost curve can be expressed by the following set of equations where Costt is
the current cost and Cost0 is the original cost.
Costt = Cost0 * (1 - decay)1™'
lVolt = max(o, log 2 (m Volt fseedV}}
m Volt = Ts\m(cVolt, a * kD * seedV}
i=t
cVol =
where
a
= the number of stages of learning-related cost reductions. Setting a=2 results in
two full learning stages.
decay = the learning rate.
IVol = zero until the threshold volume, seedV, is reached.
50
-------
mVol = the cumulative volume, cVol, until the volume floor, a*kD*seedV is reached.
The volume floor, a*kD*seedV represents the volume at which learning effects
cease.
kD = the volume factor which defines the volume floor. Setting kD=2 and a=2 would
result in a volume floor of four times the threshold volume, seedV.
seedV= the threshold volume at which learning effects begin to occur.
Figure 4.2-2 illustrates an experience curve for a vehicle technology with an initial average
unit cost, Costo, of $100 and a learning rate, decay, of 20 percent. In this hypothetical
example—illustrated by the curve, Costh named "Typical Learning Curve"—the initial
production volume, or threshold volume, before cost reductions begin to be realized is set at
12,500 units and kD is set at 2 (i.e., the volume floor is set at 50,000 units). As shown in the
figure, costs remain constant until the threshold volume, seedV, of 12,500 is reached at which
point learning begins to decrease the part cost. Upon a doubling of the threshold volume, the
learning curve effect has resulted in a 20 percent reduction in part costs (i.e., mFo/=25,000 so
that Wol=\ and then Costt = $100*(1-0.2)). Another doubling of volume at 50,000 units results
in another 20 percent reduction in costs. Since a cumulative volume of 50,000 units represents
the volume floor, costs then stabilize and no further learning occurs.
Figure 4.2-2 also shows a "Traditional EPA Curve." As discussed more below, EPA has
traditionally used a simple approach to applying the learning curve by ignoring the threshold
volume and assuming that learning occurs in a step-wise fashion. We have traditionally applied
two learning steps to our initial costs which would result in the same final cost as the approach
described above while simplifying the analysis. Further, while the more detailed approach
described above more closely approximates reality, our traditional approach has slightly
underestimated the learning impacts (i.e., our approach has not accounted for the cost reductions
represented by the area between the "typical" curve and the "EPA" curve).
Figure 4.2-2 Typical Experience Curve
$120 -|
$100 -
o $80 -
o
$60 -
$40
10 20 30 40
Cumulative Volume
50
60
-Typical Learning Curve
-Traditional EPA Curve
51
-------
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 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 no longer occur, instead depending on the asymptotic behavior of the
effect for learning rates below 100 percent to establish a floor on costs. Table Table 4.2-2
summarizes estimates of learning rates derived from studies of production costs for various
products.
Table 4.
2-2 Estimated Learning Rates and Associated Volumes for Various Products
Product(s)
Photovoltaic cells
Wind turbines
Gas turbines
Semiconductors
Automobile assembly
Truck manufacturing
Battery -electric LDV
Fuel cell hybrid LDV
Fuel cell LDV powertrain
Costs Affected
Total costs
Total costs
Total costs
Total costs
Assembly labor
Total costs
Total costs
Total costs
Total costs
Threshold Volume
Not reported
100 MW
100 MW
Not reported
Not reported
Not reported
10,000 units
10,000 units
10,000 units
Learning Rate
20%
20%
10%
13-24%
16%
10%
10%
16%
19%
In past rulemaking analyses, as noted above, EPA has used a learning curve factor of 20
percent for each doubling of production volume. In those analyses, we simplified our approach
by using a time based learning progression rather than a pure production volume progression
(i.e., after two years of production we have assumed that production volumes have doubled and,
therefore, costs are reduced by 20 percent). This approach has served the Agency well,
especially considering that those rulemaking analyses have reflected programs in which every
new engine or vehicle beginning in year one of implementation would be equipped with the
newly required piece of technology.
We believe that a 20 percent learning factor has been appropriate for all newly applied
technologies in past EPA rules, and we believe it is appropriate in the future for most carbon
dioxide reducing technologies. One exception is learning applied to diesel technologies where
we believe that a 10 percent factor is more applicable going forward because those costs build on
past EPA estimates and consider already at least two learning steps (resulting from the Tier 2
highway and 2007 heavy-duty highway rules). That said, we believe there is still room for more
learning on diesel technologies since so few light-duty diesels with aftertreatment devices exist
in the United States and because General Motors has made announcements recently about
potential cost savings associated with their new 4.5 liter Duramax diesel V8 engine.
For each of the technologies presented in Table 4.2-1, we have considered whether we could
project future cost reductions due to manufacturer learning. In making this determination, we
considered whether or not the technology was in wide-spread use today or expected to be by the
model year 2011-2012 time frame, in which case estimating future learning may not be
52
-------
appropriate because the technology is already in wide-spread production by the automotive
industry today, e.g., on the order of multi-millions of units per year. Savings from learning are
thus already reflected in our estimates. (Examples of these include 5-speed automatic
transmissions and intake-cam phasing variable valve timing. These technologies have been in
production for light-duty vehicles for more than 10 years.) In addition, we carefully considered
the underlying source data for our cost estimate. If the source data specifically stated that
manufacturer cost reduction from future learning would occur, we took that information into
account in determining whether we would apply manufacturer learning in our cost projections.
Thus, for many of the technologies, we do not believe it would be appropriate to consider any
learning curve cost reductions during the timeframe of consideration.
However, there are a number of technologies that are not yet in mass production for which
we believe the initial cost would be reduced in the time frame of consideration due to
manufacturer production learning. As indicated in Table 4.2-3, we believe that application of
learning effects for some technologies would be appropriate beginning today while for others the
learning effects should not be considered for another four to six years. The distinction between
the application of learning in the near term versus its application in the longer term is due to the
source data for our cost estimates. For those technologies where the source of our cost estimate
did not take into account manufacturer learning, we believe that learning effects are applicable in
the near term.
Table 4.2-3 Technologies Expected to See Cost Reductions due to Learning Effects
Cost Reductions Measured Relative to Costs Shown in Table 4.2-1
Technology
Cylinder deactivation - overhead cam
Continuous VVLT - overhead cam
Camless valvetrain (electromagnetic)
GDI - lean burn
Gasoline HCCI dual-mode
Turbo+downsize
Diesel - Lean NOx trap
Diesel - urea SCR
6-speed AMT
Stop-Start with 42 volt system
IMA/ISA/BSG (includes engine downsize)
2-Mode hybrid electric vehicle
Power-split hybrid electric vehicle (P-S HEV)
Plug-in hybrid electric vehicle (PHEV)
Improved high efficiency alternator & electrification of
accessories (42 volt)
Secondary axle disconnect (unibody only)
Learning curve cost reductions upon
doubling of production starting in
the given time frame a
Year
longer term
longer term
near term
near term
near term
longer term
near term
near term
near term
longer term
longer term
longer term
longer term
near term
near term
near term
Learning
Factor
20%
20%
20%
20%
20%
20%
10%
10%
20%
20%
20%
20%
20%
20%
20%
20%
Cost reductions would occur at the first doubling of production for production beginning in the time frame
shown (near term or longer term). Cost reductions would occur again at the second doubling of production.
Note that the time frame designation—near term or longer term—is not meant to be an absolute measure, but
rather a relative measure tied only to this analysis. Please refer to the text for detail on the meaning of these
terms within the context of this report. The learning factor represents the level of cost reduction that would
occur at each step. Technologies not shown may experience cost reductions from our estimated levels, but
we believe those reductions would not occur during the timeframe of consideration.
Certain other technologies are based on a source that we understand to have taken into
account manufacturer learning and, therefore, we believe that the cost estimates we present for
those technologies should not have any learning applied to them in the near term. The
53
-------
technologies for which we believe that longer term learning is more appropriate, we have used as
our primary source the 2004 NESCCAF study, for which the sub-contractor was The Martec
Group. In the work done for the 2004 NESCCAF report, Martec relied upon actual price quotes
from Tier 1 automotive suppliers to develop automotive manufacturer cost estimates. During the
process of developing the cost estimates for this proposal, EPA staff met directly with
representatives from Martec to better understand how their cost estimation methodology was
developed. Based on this information, we understand that the Martec cost estimates already
incorporate some element of manufacturer learning. Martec informed us that the Tier 1 suppliers
were specifically requested to provide price quotes which would be valid for three years (2009-
2011), and that for some components the Tier 1 supplier included cost reductions in years two
and three which the supplier anticipated could occur, and which they anticipated would be
necessary in order for their quote to be competitive with other suppliers. Therefore, for this
analysis, we believe that some learning effects are already reflected in the Martec-sourced costs
and additional learning effects should not be applied to those costs for several years, at least until
after 2013. However, the theory of manufacturer learning is that it is a continuous process,
though the rate of improvement decreases as the number of units produced increases. While we
were not able to gain access to the detailed submissions from Tier 1 suppliers upon which Martec
relied for their estimates, we do believe that additional cost reductions will occur in the future for
a number of the technologies for which we relied upon the Martec cost estimates. Those
technologies are noted in Table 4.2-3 with learning curve effects being applicable in the longer
term.
4.2.1 Piece Costs Associated with Engine Technologies
The technologies listed here are discussed in detail in Section 2 of this report.
4.2.1.1 Low-Friction Lubricants
A change in lubricant, whether engine oil or transmission fluid, usually requires some
durability testing to ensure that durability is not compromised. It may also require some bearing
changes, but we suspect these to be minimal. The 2002 NAS study estimated the low friction
lubricant RPE at $8 to $11 using a 40 percent markup, the NESCCAF study showed an RPE of
$5 to $15 with a 40 percent markup, and the EEA report to DOE showed manufacturer costs of
$10 to $20. By contrast, many of the manufacturer CBI submittals had lower or even no costs
associated with low friction lubricants. We believe these manufacturer estimates are more
accurate (among other things, it is in manufacturers' interests to use higher cost estimates), but
also believe that a change in any lubricant would involve some level of verification and
durability testing and have estimated the incremental compliance cost at $3. We believe that this
estimate is independent of vehicle class since the engineering work required should apply to any
engine size.
4.2.1.2 Engine Friction Reduction
Several friction reduction opportunities (piston surfaces and rings, crankshaft design,
improved material coatings, etc.) have been identified that are still available to a significant
number of engine designs. Additionally, as computer-aided modeling software continues to
improve, more opportunities for evolutionary friction reduction might become apparent. The
2002 NAS study estimated the engine friction reduction RPE at $35 to $140 (40 percent
markup); NESCCAF showed an RPE of $5 to $15 (40 percent markup); the EEA report to DOE
54
-------
showed manufacturer costs of $10 to $55. The CBI submittals suggest that these ranges are
reasonable, although they contained values ranging from $0 to $140. For this analysis, we have
estimated the incremental compliance cost to range from $0 to $84 for small cars, $0 to $168 for
large trucks, and $0 to $126 for applications in between. We have estimated such a wide range
here because there are so many friction reduction opportunities - piston surfaces and rings,
crankshaft design, improved material coatings, low-tension piston rings, roller cam followers,
material substitution, more optimal thermal management, piston surface treatments, as well as
lubricant friction reduction - and manufacturers may do anywhere from none to many or all of
them.
4.2.1.3 Variable Valve Timing Systems
Manufacturers are currently using many different types of variable valve timing, which have
a variety of different names and methods. The major types of VVT are listed below.
a. Intake Camshaft Phasing (ICP)
Valvetrains with VVT-ICP, which is the simplest of the cam phasing technologies,
can modify the timing of the intake valve while the exhaust valve timing remains fixed.
This requires the addition of a cam phaser for each bank of intake valves on the engine.
An in-line 4-cylinder engine has one bank of intake valves, while V-configured engines
would have two banks of intake valves. In our efforts to understand the estimates
presented in the 2002 NAS study, we believed that they estimated the cam phaser RPE at
$35 (40% markup),1 while the EEA report showed a manufacturer cost (rather than an
RPE) of $35. The NESCCAF study showed a manufacturer cost of $35 which would
mean an RPE of $49 (40% markup). Consistent with the EEA report and NESCCAF
study, we have used this $35 manufacturer cost to arrive at our PPI-adjusted incremental
compliance cost of $59 per cam phaser or $59 for an in-line 4 cylinder and $119 for a V-
type engine. We have developed an estimate for VVT-ICP associated with overhead cam
engines only. For overhead valve engines we do not expect use of VVT-ICP since they
typically use a form of VVT called coupled cam phasing, described below.
b. Coupled Camshaft Phasing (CCP)
Coupled (or coordinated) cam phasing is a design in which both the intake and
exhaust valve timing are varied using the same cam phaser. For an overhead cam engine,
the same phaser added for VVT-ICP would be used for VVT-CCP control. As a result,
its costs are identical to those for VVT-ICP. For an overhead valve engine, only one
phaser would be required for both 4-cylinder and V-configured engines since only one
camshaft exits. Therefore, for overhead valve engines, the incremental compliance cost
is estimated at $59 regardless of engine configuration, for the reasons given in the
previous sub-section.
1 The 2002 NAS study estimated the lower end of the RPE range for variable valve timing at $35 and for variable
valve lift and timing at $70. VVT requires one cam phaser while WLT requires two.
55
-------
c. Dual (Independent) Camshaft Phasing (DCP)
The most flexible VVT design is dual cam phasing, where the intake and exhaust
valve opening and closing events are controlled independently. This design allows the
option of controlling valve overlap, which can be used as an internal EGR strategy. Our
estimated incremental compliance cost for this technology is built upon that for VVT-ICP
where an additional cam phaser is added to control each bank of exhaust valves less the
cost to the manufacturer of the removed EGR valve. For example, the incremental
compliance cost for a V6 engine would be $59 for each bank of intake valves (i.e., 2
banks times $59/bank = $119), $59 for each bank of exhaust valves (i.e., another $119),
less $29 incremental compliance cost for the removed EGR valve; the total incremental
compliance cost being $209.J Note that we do not anticipate VVT-DCP being used on
overhead valve engines and, hence, do not have a cost associated with VVT-DCP on
overhead valve engines.
4.2.1.4 Engine Cylinder Deactivation
Cylinder deactivation allows for some (usually half) of the cylinders to be "shut down"
during light load operation. Noise and vibration issues reduce the operating range to which
cylinder deactivation is allowed, although manufacturers are exploring the possibility of
increasing the amount of time that cylinder deactivation might be suitable. The 2002 NAS study
estimated the RPE to range from $112 to $252 (40% markup) while the NESCCAF study
estimated the RPE at $161 to $210 (40% markup). The EEA report showed manufacturer costs
of $105 to $135 (no markup), depending on vehicle class. In reviewing all our sources, we have
attempted to determine the cost associated with individual components needed to employ
cylinder deactivation. This way, we can use a "bottom-up" approach to estimate costs. Doing
this, we have estimated the cost for individual components of these systems at $15 per cylinder
being deactivated, $15 per engine for controls, and $60 per engine for engine mounts to address
noise and vibration. Adjusting to 2006 dollars this results in incremental compliance costs of
$203 for a V6 engine and $229 for a V8 engine, as shown in Table 4.2-1. These incremental
compliance costs are consistent with the NAS Report and the NESCCAF study as well as with
the CBI submissions from manufacturers. Note that 4-cylinder engines are not expected to add
this technology because noise and vibration problems become very difficult to control.
4.2.1.5 Variable Valve Lift Systems
Controlling the lift height of the valves provides additional flexibility and potential for
further reduction in pumping losses. There are two major classifications of variable valve lift,
described below.
a. Discrete Variable Valve Lift
1 Note that rounding may impact the totals presented throughout this discussion. For example, $59.30 is stated in the
text as $59 for clarity, while 2 x $59.30 is stated as $119.
56
-------
The 2002 NAS study shows DVVL system RPEs to range from $70 to $210 (40%
markup) depending on engine size. The NESCCAF study shows a RPE range of $105 to
$210 (40% markup) for electro-hydraulic DVVL systems depending on engine size and
overhead cam versus overhead valve engines. We have used the NESCCAF values and
added to those a $25 cost to the manufacturer for controls and associated oil supply needs
(costs not reflected in the NESCCAF study). We have also estimated that a single valve
lifter could control valve pairs, so engines with dual intake and/or dual exhaust valves
would require one lifter per pair of valves being controlled. As a result, our estimates for
overhead cam and overhead valve engines are the same. The end result, including PPI
adjustments, is an incremental compliance cost of $169, $246, and $322 depending on
vehicle class.
b. Continuous Variable Valve Lift
Continuous variable valve lift (CVVL) typically employs a mechanism that varies the
pivot point in the rocker arm. The NESCCAF study showed estimated RPEs from $210
to $420 (40% markup), depending on vehicle class. The EEA report showed
manufacturer costs of $180 to $350 (no markup), depending on vehicle class and
assuming presence of overhead cams. Consistent with NESCCAF, we estimated the PPI-
adjusted incremental compliance cost for these systems on overhead cam engines at
$254, $466, and $508 for a 4-, 6-, and 8-cylinder engine, respectively.
We consider this technology to be limited to overhead cam engines. As a result, for
an overhead valve engine to add CVVL, it would first have to be converted to an
overhead cam engine. (The NESCCAF value of $420 for a V8 overhead valve engine did
not include costs associated with conversion to overhead cam(s).) Such a conversion is
not inexpensive, as it entails the addition of one to three camshafts, additional valves, and
then addition of the CVVL costs just discussed. As shown in Table 4.2-1, we have
estimated this PPI-adjusted incremental compliance cost at $599 to $1,380 depending on
vehicle class.
4.2.1.6 Camless Valve Actuation Systems
Camless valve actuation relies on electromechanical actuators instead of camshafts to open
and close the cylinder valves. Camless valvetrains have been under research for decades because
it would allow for considerable fuel economy improvement potential and tremendous flexibility.
In reviewing our sources for costs, we have determined that the values presented in the 2002
NAS study - $280 to $560 (40% markup), depending on vehicle class - represent the best
available estimates. These are shown in Table 4.2-1, after applying our markup and PPI
adjustments, as ranging from $336 to $673, independent of vehicle class. For comparison, the
NESCCAF study showed RPE estimates ranging from $805 to $1,820 (40% markup), and the
EEA report to DOE showed manufacturer costs ranging from $210 to $600 (no markup). The
NESCCAF study shows considerably higher values than our estimates. Importantly, the
NESCCAF study estimated the costs for camless valve actuation on both the intake and the
exhaust valves. We believe that a more likely scenario would be using camless valve actuation
on only the intake valves. Therefore, we believe that our lower estimates represent the more
likely technology application.
57
-------
4.2.1.7 Stoichiometric Gasoline Direct Injection Technology
Gasoline direct injection (GDI), also known as 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). The Stoichiometric GDI engine operates at Stoichiometric conditions and
uses a spark to initiate ignition, unlike a compression ignition engine. This requires injector
design advances, new fuel pumps to deliver higher injection pressures, and new fuel rails to
handle the higher fuel pressures. The NESCCAF study estimated the RPE for these systems at
$189 to $294 (40% markup), depending on vehicle class. The EEA report to DOE shows a
manufacturer cost range of $77 to $135. The CBI submittals from manufacturers suggest these
ranges are low. For our analysis, we have estimated the costs of individual components of a GDI
system and used a "bottom up" approach looking at incremental costs for injectors, fuel pumps,
etc., to arrive at system incremental compliance costs ranging from $122 to $420 for small cars
and up to $228 to $525 for large trucks. The lower end of the ranges represent our best estimate
using a bottom up approach while the upper end of the ranges represent levels more consistent
with the manufacturer CBI submittals.
4.2.1.8 Lean-Burn Gasoline Direct Injection Technology
One way to dramatically improve an engine's thermodynamic efficiency is by operating at a
lean air-fuel mixture (excess air). Such operation presents difficult challenges for a gasoline
fueled engine. To achieve such operation while meeting emissions standards requires everything
mentioned above for Stoichiometric GDI engines. In addition, some incremental costs would be
likely for the aftertreatment system since NOx aftertreatment for lean-burn engines is generally
more costly than for Stoichiometric engines. The NESCCAF study estimated the RPE for lean-
burn GDI aftertreatment to range from $539 to $1,260 (40% markup) depending on vehicle class.
We do not believe that lean-burn GDI engines can be expected to penetrate the light-duty market
anytime soon due to gasoline sulfur levels being slightly too high.K Nonetheless, we have
estimated the incremental compliance cost for these systems at $750, independent of vehicle
class, and incremental to a Stoichiometric GDI engine.
4.2.1.9 Homogeneous Charge Compression Ignition
Homogeneous charge compression ignition (HCCI), also referred to as controlled
autoignition (CAI), is an alternative engine operating mode that does not rely on a spark event to
initiate combustion. As the combustion is more closely aligned with diesel compression ignition
combustion, the engine operates at the higher compression ratios and efficiencies typical of
diesel engines. However, proper control of the combustion process is difficult to achieve and
requires in-cylinder pressure sensors and very fast engine control logic to optimize combustion
timing, especially considering the variable nature of operating conditions seen in a vehicle. The
NESCCAF study estimated the RPE to range from $560 to $840 (40% markup), depending on
vehicle class, including the costs for a Stoichiometric GDI system and VVLT-DVVL. We have
based our estimated incremental compliance cost on the NESCCAF estimates and, after applying
our markup and adjusting to 2006 dollars, have arrived at $263 to $685, depending on vehicle
K Note that gasoline lean-burn aftertreatment is essentially the same technology as diesel aftertreatment. We explain
in detail in our 2007 Heavy-duty Highway and Nonroad Tier 4 rules that lean-burn aftertreatment works only if fuel
sulfur is below 15 ppm. Current gasoline sulfur levels in the U.S. are, on average, 30 ppm with a maximum of 80
ppm.
58
-------
class. Note that our estimated incremental compliance cost is incremental to a stoichiometric
GDI engine.
4.2.1.10 Gasoline Turbocharging and Engine Downsizing
Turbocharging and supercharging (grouped together here as boosting) are two methods to
increase the intake manifold and cylinder pressures above typical levels. Boosting increases the
airflow into the engine, thus increasing the specific power level, and with it the ability to reduce
engine size while maintaining performance. The technology considered here involves addition
of a boost system, removal of two cylinders in most cases (from an 8-cylinder to a 6, or a 6 to a
4) and associated valves, and the addition of some form of cold start control system (e.g., air
injection) to address possible cold start emission control. Consistent with NESCCAF, we have
estimated the boost system incremental compliance cost at $600. The incremental compliance
cost for material associated with a cylinder we estimated at $75, valves at $15, and camshafts at
$150. An air injection pump incremental compliance cost was estimated at $90. Using these
values, we have estimated the incremental compliance cost for a boosted/downsized engine
system at $690 for small cars, $810 for large trucks, and $120 for other vehicle classes. The
small car value is higher than the mid-range classes because it does not eliminate any cylinders
so that there are no cost savings associated with that elimination. For the large trucks, the costs
are higher than the mid-range classes because we have assumed an overhead valve engine as the
baseline and an overhead cam engine once downsized. That results in the addition of camshafts,
rather than removal of camshafts and associated costs, which outweighs the removal of cylinders
and associated costs.
4.2.1.11 Diesel Systems
Diesel engines have several characteristics that give them superior fuel efficiency to
conventional gasoline, spark-ignited engines. The diesel combustion cycle operates with fewer
pumping losses, at a higher compression ratio, with a very lean air/fuel mixture, and typically at
much higher torque levels than an equivalent-displacement gasoline engine. These features carry
with them higher costs relative to gasoline engines. These higher costs result from:
• Improved fuel systems (higher pressures and more responsive injectors)
• Advanced controls and sensors to optimize combustion and emissions performance
• Higher compression ratios which require a more robust engine
• A turbocharger
• More costly aftertreatment systems
We have considered the costs for two types of diesel systems: one using a lean-NOx trap
(LNT) along with a particulate filter; and one using a selective catalytic reduction (SCR) system
along with a particulate filter. In our discussions with industry over the past couple of years -
both auto manufacturers and aftertreatment device manufacturers - we have been repeatedly told
that LNT systems would probably be used on smaller vehicles while the SCR systems would be
used on larger vehicles and trucks. The primary reason given for this choice is the trade off
between the rhodium needed for the LNT and the urea injection system needed for SCR. The
breakeven point between these two cost factors appears, at present, to occur at roughly the 3.0
liter engine size - below that, LNT is less costly while above that SCR is less costly. Other
59
-------
factors impact a manufacturer's decision on which system to use, but we have used that rule-of-
thumb for our analysis.
We have estimated the incremental compliance cost for diesel systems in a manner
consistent with our recent clean diesel rulemakings (2007 Heavy-duty Highway and Nonroad
Tier 4), building upon that approach by also considering the incrementally higher costs
associated with the diesel engine relative to the gasoline engine. Our estimated incremental
compliance costs range from $2,790 for the small car to $4,065 for the large truck. For
comparison, the NESCCAF study showed RPEs of $2,100 to $2,730 (40% markup), although
that analysis did not specify LNT versus SCR so direct comparisons are difficult to make.
4.2.1.12 E20-E30 Optimized Ethanol Engines
None of the other cost sources have directly addressed or considered engines optimized for
ethanol use. For these systems, we believe that the only hardware cost required for optimization
for ethanol use is to substitute some materials in the fuel system to accommodate the ethanol
because ethanol reacts differently with materials than does gasoline. The cost to substitute
affected materials should be low and we estimate them to be on the order of $15. However, for
true optimization, the engine would also have to be boosted either using a turbo charger or
supercharger. Further, we would expect such optimization to include engine downsizing to
minimize fuel consumption. Neither of these costs is captured in the $15 material substitution
cost. As a rough estimate of the incremental compliance costs, we have added the $15 material
substitution cost, or $23 with markup, to our engine downsize and turbocharging costs presented
in Section 4.2.1.10, to arrive at incremental compliance costs of $713 for small cars, $833 for
large trucks, and $143 for other vehicle classes. These compliance costs would be incremental to
a base gasoline engine (i.e., an engine that has not been downsized or turbocharged).
4.2.2 Piece Costs Associated with Transmission and Hybrid Technologies
4.2.2.1 Automatic 5-speed Transmissions
As automatic transmissions have been developed over the years, more forward speeds have
been added to improve fuel efficiency, and performance. Increasing the number of available
ratios provides the opportunity to optimize engine operation under a wider variety of vehicle
speeds and load conditions. We have relied on the 2002 NAS study for our estimated
incremental compliance costs associated with migrating from a 4-speed automatic transmission
to a 5-speed automatic. Those RPEs range from $70 to $154 (40% markup) which becomes $76
to $167 for using our indirect cost markup and adjusting to 2006 dollars, independent of vehicle
class. This range is consistent with the NESCCAF report which showed RPEs of $140 (40%
markup), independent of vehicle class, while the EEA report showed a manufacturer cost
estimate of $130.
4.2.2.2 Aggressive Shift Logic
In operation, an automatic transmission's controller decides when to upshift or downshift
based on a variety of inputs such as vehicle speed, and throttle position according to programmed
logic. This logic can be biased towards maximizing fuel efficiency by upshifting earlier and
inhibiting downshifts under some conditions. Additional adaptive algorithms can be employed
to maintain performance feel while improving fuel economy under most driving conditions. This
technology consists of calibration of computer software as no hardware is required. We have
estimated the incremental compliance cost for this calibration effort at $38 based on the 2002
60
-------
NAS study which estimated the RPE of aggressive shift logic to range from $0 to $70 (40%
markup).
4.2.2.3 Early Torque Converter Lockup
As with aggressive shift logic, early torque converter lockup requires no new hardware and is
accomplished through calibration to force lockup earlier than done traditionally. The 2002 NAS
study did not provide a RPE estimate for this technology. The NESCCAF study estimated the
RPE to range from $0 to $10 (40% markup) and, in its 2006 report for the Department of Energy,
EEA estimated the manufacturer cost at $5. Here, we have estimated the incremental
compliance cost of this technology (i.e., the calibration effort) at $30 based in part on NESCCAF
and the CBI submissions which were slightly higher than NESCCAF. We have used a higher
value here than NESCCAF and EEA because we have tried to account for the engineering effort
in addition to the hardware which we believe NESCCAF and EEA did not do.
4.2.2.4 Automatic 6-, 7- and 8-speed Transmissions
In addition to 5-speed automatic transmissions, manufacturers can also choose to utilize 6-,
7-, or 8-speed automatic transmissions. Additional gears allows 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. According to EEA in its 2006 report for DOE, a Lepelletier gear
set design provides for 6-speeds at the same cost as a 5-speed automatic. Based on that analysis,
we have estimated the incremental compliance cost of a 6-speed automatic to be equivalent to
that for a 5-speed automatic. We have not developed any estimate costs for 7- or 8-speed
transmissions because of the diminishing returns in efficiency versus the costs for transmissions
beyond 6-speeds.
4.2.2.5 Automated (shift) Manual Transmissions
An Automated Manual Transmission (AMT) is mechanically similar to a conventional
manual transmission, but shifting and launch functions are controlled by the vehicle rather than
the driver. A switch from a conventional automatic transmission with torque converter to an
AMT incurs some costs but also allows for some cost savings. Savings can be realized through
elimination of the torque converter which is a very costly part of a traditional automatic
transmission, and through reduced need for high pressure hydraulic circuits to hold clutches (to
maintain gear ratios in automatic transmissions) or hold pulleys (to maintain gear ratios in
Continuously Variable Transmissions). Cost increases would be incurred in the form of
calibration efforts since transmission calibrations would have to be redone, and the addition of a
clutch assembly for launce and gear changes.
While AMTs are becoming more common in Europe, a primary concern with respect to this
technology is the production capacity in the United States. Transmission manufacturing
facilities in the United States are primarily for automatic transmissions. They are very costly and
cannot be easily converted from producing traditional automatic transmissions to AMTs which
are more like manual transmissions in design. However, as facilities are upgraded and
transmission manufacturing equipment is replaced, existing facilities can be converted to AMT
production at little to no additional cost relative to retooling for continued production of
automatic transmissions. General Motors has made investments recently in manufacturing
facilities geared toward production of traditional automatic transmissions so a widespread
migration to AMTs would be very costly and unlikely for GM. In 2007, Getrag and Chrysler
61
-------
announced plans to build a new transmission plant in the U.S. to supply Chrysler with AMTs
beginning as soon as 2009.
We believe that, overall, the hardware associated with an AMI, whether single clutch or dual
clutch, is no more costly than that for a traditional automatic transmission given the savings
associated with removal of the torque converter and high pressure hydraulic circuits.
Nonetheless, given the need for engineering effort (e.g., calibration and vehicle integration work)
when transitioning from a traditional automatic to an AMI, we have estimated the incremental
compliance cost at $141, independent of vehicle class.
4.2.2.6 Continuously Variable Transmissions
A Continuously Variable Transmission (CVT) is unique in that it does not use gears to
provide ratios for operation, but instead uses two V-shaped pulleys and a belt. The pullies are
split in half and a hydraulic actuator moves the pulley halves together or apart which causes the
belt to ride on either a larger or smaller diameter section of the pulley. This changes the
effective ratio of the input to the output shafts providing for a very wide range of ratios. The
2002 NAS study estimated the RPE for this technology at $140 to $350 (40% markup) with the
higher end of the range apparently meant for larger vehicles and trucks. The NESCCAF study
estimated the RPE at $210 to $245 (40% markup) noting that the technology was not applicable
to large trucks (i.e., V8 engines). For this analysis, we are consistent with the NESCCAF study
and have estimated the incremental compliance cost at $231 to $270 in 2006 dollars, depending
on vehicle class.
4.2.2.7 Manual (clutch shifted) 6-, 7-, and 8-speed Transmissions
As with automatic transmissions, 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 at a given speed. Typically, this is achieved through adding additional
overdrive ratios to reduce engine speed (which saves fuel through reduced pumping losses). Six-
speed manual transmissions have already achieved significant market penetration, so
manufacturers have considerable experience with them and the associated costs. Based on CBI
submissions, we have estimated the incremental compliance cost of a 6-speed manual relative to
a 5-speed manual at $107, regardless of vehicle class.
4.2.2.8 Hybrid Systems
A hybrid is a vehicle that combines two or more sources of propulsion energy, where one
uses a consumable fuel (like gasoline), and the other is rechargeable (during operation or by an
external energy source). There are three primary ways to make use of hybrid technology to
reduce fuel consumption as described in more detail in Section 2.5 of this report. The major
hybrid concepts we have considered are listed below.
a. Integrated Starter-Generator with Idle-Off
Integrated Starter-Generator (ISG) systems are the most basic of hybrid systems and
offer mainly idle-stop capability. The most common ISG systems replace the
conventional belt-driven alternator with a belt-driven, higher power starter-alternator. In
addition, when idle-off is used (i.e., the petroleum fuelled engine is shut off during idle
operation), an electric power steering and auxiliary transmission pump are added to
62
-------
provide for functioning of these systems which, in a traditional vehicle, were powered by
the petroleum engine. The 2002 NAS study estimated the RPE of these systems at $210
to $350 (40% markup) with a 12 volt electrical system and independent of vehicle class,
while the NESCCAF study estimated the RPE for these systems at $280 (40% markup)
with a 12 volt electrical system for a small car. We have estimated the incremental
compliance cost of these systems, including the costs associated with upgrading to a 42
volt electrical system (discussed below in Section 4.2.3.3) and expressed in 2006 dollars,
at $563 to $600, depending on vehicle class.
b. Integrated Motor Assist (IMA)/Integrated Starter-Alternator-Dampener (ISAD)
Honda's IMA system uses an electric motor bolted to the engine's crankshaft and
connected to the transmission through a torque converter or clutch. This electric motor
acts as both a motor for helping to launch the vehicle and a generator for recovering
energy while slowing down. It also acts as the starter for the engine and the electrical
system's main generator. The Continental ISAD system is similar to Honda's IMA and
allows for idle-stop capability. The 2002 NAS study did not consider this technology
while the NESCCAF study estimated the RPE for these systems at $2310 to $2940 (40%
markup) for a small car and large car, respectively. We have used these estimates as the
basis for our incremental compliance costs of $2477 for the small car and $3153 for the
large car, expressed in 2006 dollars. We have not estimated incremental compliance
costs for the other vehicle classes because we do not believe those classes would use this
technology and would, instead, use the hybrid technologies discussed below.
c. 2-Mode Hybrids
This technology uses an adaptation of a conventional stepped-ratio automatic
transmission by replacing some of the transmission clutches with two electric motors,
which makes the transmission act like a CVT. The 2002 NAS study did not consider this
technology, while the NESCCAF study estimated the RPE at $4340 to $5600 (40%
markup), depending on vehicle class. We have used these estimates as the basis for our
incremental compliance costs of $4655 to $6006 in 2006 dollars, depending on vehicle
class. We have not estimated incremental compliance costs for small cars because we do
not believe that this technology is well suited to small cars which would, instead use ISO
or IMA/ISAD for a mild hybrid approach or power split for a more aggressive approach.
d. Power Split Hybrids
Power Split HEV systems are used currently by Toyota, Ford and Nissan. The Power
Split system replaces the transmission with a single planetary gear and a motor/generator.
A second, more powerful motor/generator is permanently connected to the vehicle's final
drive and always turns with the wheels. This allows for removal of the conventional
transmission, replacing it with a much simpler single planetary and motor/generator.
Because load capacity is limited by the first motor/generator's capacity to resist the
reaction torque of the drive train, this technology is best suited to low load applications
63
-------
(i.e., smaller vehicles), although recent advances have expanded Power Split system
applicability. The 2002 NAS study did not consider this technology, while the
NESCCAF study estimated the RPE at $3500 (40% markup) for a small car. Based on
the NESCCAF study, we have estimated the incremental compliance cost at $3754 for a
small car, expressed in 2006 dollars.
e. Full-Series Hydraulic Hybrids
A Full Series Hydraulic Hybrid Vehicle (HHV) is somewhat similar in concept to a
full-series electric hybrid vehicle, except that the energy is stored in the form of
compressed nitrogen gas and the power is transmitted in the form of hydraulic fluid.
Series HHV technology is under development by EPA. We have estimated the
incremental compliance cost for this technology at $771 to $1233 expressed in 2006
dollars, depending on vehicle class.20 Because this technology is still under development
and has not yet been commercialized, we believe that it would not be available until the
second half of the next decade.
f Plug-in Hybrids
Plug-In Hybrid Electric Vehicles (PHEVs) are very similar to Hybrid Electric
Vehicles, but would have a larger battery pack with more energy storage and a greater
capability to be discharged. A PHEV would also have a control system that allows the
battery pack to be significantly depleted during normal operation. These changes mean
that PHEVs are expected to be more costly than conventional vehicles and some other
advanced technologies. Neither the 2002 NAS study nor the NESCCAF study
considered this technology as manufacturers have only recently made statements about it
being a technology for serious consideration. Based on our own work, we have estimated
the incremental compliance cost at $4500 to $10200, depending on vehicle class. This
incremental compliance cost assumes a 20 mile "all electric" range.
g. Full Electric Vehicles
The recent intense interest in hybrid vehicles and the development of hybrid vehicle
battery and motor technology has helped make Electric Vehicle (EV) technology a more
viable candidate for consideration. Electric vehicles require much larger batteries than
either HEVs or PHEVs and the batteries must be of a high-energy and lower-power
design to deliver an appropriate amount of power over the useful charge of the battery.
While electric motor and power electronics designs are very similar to HEV and PHEV
designs, they must be larger, more powerful, and more robust since they provide the only
motive power for the vehicle. However, cost savings can be realized by removing the
internal combustion engine, fuel system, and possibly the transmission. Neither the 2002
NAS nor the NESCCAF study considered this technology and, based on our own work,
we have estimated the incremental compliance cost at $12000 for the small car and
$15000 for the large car. We have not made estimates for other vehicle classes because
we do not consider this to be a viable technology for vehicles larger than a large car.
64
-------
4.2.3 Piece Costs Associated with Accessory Technologies
4.2.3.1 High Efficiency Alternators, electric water pumps and electrification of other
accessories for 12 Volt systems
Replacing the traditionally belt-driven accessories - the alternator, coolant and oil pumps -
with electrically controlled accessories, or simply improving the efficiency of the belt driven
accessories would provide an opportunity to reduce the accessory loads on the engine. Some
large trucks also employ mechanical fans, some of which could be improved or electrified as
well. Additionally, there are now higher efficiency alternators which require less of an accessory
load to achieve the same power flow to the battery. The 2002 NAS study estimated the RPE for
this at $84 to $112 (40% markup), independent of vehicle class. The NESCCAF study estimated
an RPE of $56, but that estimate included only a high efficiency generator and did not include
electrification of other accessories. We have used the NAS estimates to arrive at our incremental
compliance costs of $89 to 119 expressed in 2006 dollars, independent of vehicle class. Note
that air conditioning and power steering are other candidates for accessory load reduction. We
discuss those below and have not included them in this incremental compliance cost estimate.
4.2.3.2 Electric Power Steering for 12 Volt and 42 Volt systems
Electric power steering (EPS) is advantageous over hydraulic steering in that it only draws
power as needed when the vehicle is cornering, which is only a small percentage of a vehicle's
operating time. EPS may be implemented on many vehicles with a standard 12V system, but
heavier vehicles may require a 42V system adding cost and complexity. The 2002 NAS study
estimated the RPE for a 12V system to range from $105 to $150 (40% markup), independent of
vehicle class. The NESCCAF study estimated the RPE to range from $28 to $56 (40% markup)
for a 12V system, independent of vehicle class. We have estimated the incremental compliance
cost to range from $118 to $197 for a 12V system expressed in 2006 dollars, independent of
vehicle class.
4.2.3.3 Upgrade Electrical Systems to 42V
Most vehicles today (aside from hybrids) operate on 12 V electrical systems. At higher
voltages, the power density of motors, solenoids, and other electrical components increases to the
point that new and more efficient systems, such as electric A/C compressors and electric power
steering (for heavier trucks) may be feasible. A 42V system also allows for smaller-gauge
wiring and smaller, lighter electric motors and actuators. A 42V system also acts as an enabler
for an integrated starter generator. The 2002 NAS study estimated the RPE for upgrading to 42V
at $70 to $280 (40% markup), independent of vehicle class. We have estimated the incremental
compliance cost at $89 to $119 in 2006 dollars, independent of vehicle class and exclusive of
improvements to the efficiencies or electrification of 12V accessories.
4.2.4 Piece Costs Associated with Other Vehicle Technologies
4.2.4.1 Aerodynamic Drag Reduction through reduced drag coefficient and reduced
frontal area
A vehicle's size and shape determine the amount of power needed to push the vehicle
through the air at different speeds. Changes in vehicle shape or frontal area can therefore reduce
CC>2 emissions. Areas for potential aerodynamic drag improvements include skirts, air dams,
65
-------
underbody covers, and more aerodynamic side view mirrors. The CBI submittals generally
showed the RPE associated with these changes at less than $100. We have estimated the
incremental compliance cost to range from $0 to $75, independent of vehicle class.
4.2.4.2 Low Rolling Resistance Tires
Tire characteristics (e.g., materials, construction, and tread design) influence durability,
traction control, vehicle handling, and comfort. They also influence rolling resistance. This
technology is applicable to all vehicles, with the exception of body-on-frame light trucks. We
have based our estimates on a 2006 NAS/NRC report which showed a $1 per tire cost for low
rolling resistance tires.21 For four tires, our incremental compliance cost estimate is $6 per
vehicle, independent of vehicle class, although not applicable to large trucks.
4.2.4.3 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 shoes are pulled away from the rotating disc. While most
passenger cars have already adopted this technology, there are indications that this technology is
still available for body-on-frame trucks. Based on the recent NHTSA light-duty truck CAFE
rule, we have estimated the incremental compliance cost for low drag brakes at $87 per truck.22
As noted, this technology is already on most passenger cars so we have estimated the
incremental compliance cost for trucks only.
4.2.4.4 Secondary Axle Disconnect (front axle for ladder frame and rear axle for unibody
frame)
To provide shift-on-the-fly capabilities, many part-time four-wheel drive systems use some
type of axle disconnect which, whether front or rear axle, reduces parasitic losses and fuel
consumption. For front-axle disconnect systems, we have estimated the incremental compliance
cost at $114 based on CBI submittals. This technology is considered applicable for light-duty
ladder-on-frame trucks. For unibody vehicles and trucks, expected to use a secondary-axle
disconnect system, the incremental compliance cost is estimated at $676, based again on CBI
submittals.
4.3 Estimates of Indirect Costs and the Use of Markup Factors
Regulatory agencies, including EPA, have frequently relied upon multiplicative adjustment
factors to account for the indirect costs associated with changes in direct manufacturing costs.
(Indirect costs include research and development, salaries, pensions, marketing, and other
expenses.) The resulting cost after applying the factor is often called the "Retail Price
Equivalent" (RPE): thus these factors are frequently called "RPE factors" or "RPE multipliers."
Clearly 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, this is not always feasible within the constraints of an agency's time or
budget, or the necessary information to carry out such an analysis is simply unavailable. Given
this, EPA has continued to rely on the use of an indirect cost multiplier for some of our
regulatory cost analyses.
There is a history of what multiplier to use. In past mobile source regulatory actions, EPA
has sometimes used an RPE factor of 1.26 to account for the indirect costs associated with the
variable cost impacts of a regulation. This factor was originally derived in the late 1970's and
66
-------
updated in a 1985 report by Jack Faucett Associates under contract to EPA.23 In 2000, Argonne
National Laboratory (ANL) published a technical memorandum comparing three different
estimates of RPE multipliers for vehicle manufacturing.24 In this memorandum, ANL compared
their own estimates with those from Chrysler and Energy and Environmental Analysis, and
found that the three estimates of indirect cost multipliers, when put on a comparable basis, were
very similar. The ANL analysis made a distinction between components made by suppliers and
components developed and manufactured by the vehicle manufacturer, reasoning that for
outsourced components the outside vendor would incur some of the costs that would otherwise
be borne by the vehicle manufacturer (i.e., indirect cost items such as warranty, research and
development, and depreciation and amortization). The ANL analysis estimated retail price
equivalent factors of 1.5 for outsourced components and 2.0 for products developed and
manufactured internally. A more recent analysis commissioned by the automobile manufacturing
industry and conducted by Sierra Research, Inc., suggested that the 1985 Faucett report contains
"basic methodological errors that make it unreliable for use in a regulatory analysis."25 The
Sierra analysis concluded that a retail price equivalent factor for components manufactured
internally should be about 2.0 to accurately account for the per-vehicle indirect costs.
In Section 4.2 above we applied a multiplicative adjustment factor of 1.5 to our estimates of
the direct manufacturing costs to account for the incremental indirect costs associated with each
CO2-reducing technology. The derivation of this factor, which is consistent with the ANL
analysis, is described below. As explained in Section 4.3.1 below, EPA believes that this factor
is reasonable, if one utilizes the RPE methodology. However, EPA also believes that the RPE
multiplier, as it has been traditionally derived, may not be entirely appropriate for the purpose of
estimating the impacts on a manufacturer's indirect costs of a new EPA regulation, and may, in
fact, overestimate the overall costs. For the reasons described below, EPA is initiating a new
study that will evaluate the methodology's appropriateness for use in EPA cost analyses.
EPA has a fundamental concern with using an RPE factor, as they are currently derived, to
estimate incremental indirect costs that result from a regulatory action. The RPE approach
assumes that all costs rise in direct proportion to the incremental cost change of an individual
cost element. We believe that many of the indirect cost elements that get marked up by the RPE
factor do not in fact behave in this way. Including such elements in the development of an
indirect cost markup factor would be clearly inappropriate, and could result in significantly
overestimating the costs. Consider an illustrative example where the only direct cost is energy
and all other costs are indirect. In this example, assume a company's energy costs are 10 percent
of the manufacturer suggested retail price (MSRP), and an RPE factor is created relating MSRP
to energy costs. The very nature of the RPE multiplier approach assumes that the 10:1
relationship of MSRP to energy costs remains constant, meaning that a doubling of energy costs
would, using the RPE multiplier methodology, effectively result in a doubling of the MSRP. In
the case of energy costs, we believe that a more likely and more realistic response would be that
the indirect (i.e., non-energy) cost components of the MSRP would remain unchanged. In this
case, the correct response to a doubling of an element that comprises 10 percent of the total
MSRP would be a 10 percent increase in MSRP, not the 100 percent increase that results from
uncritical application of the RPE multiplier.
A key question when using an indirect cost multiplier is which of the individual indirect cost
elements increase in proportion to the direct cost elements (and if so, by what factor) and which
remain relatively fixed. Unfortunately, the literature at our disposal does not address this issue
67
-------
specifically. EPA's study will consider each manufacturer cost element for the automotive
industry and carefully evaluate whether or not each cost element should be included in a total
indirect cost markup due to an EPA regulatory action. In a recent study NHTSA identified eight
components of a manufacturer's indirect costs:
Maintenance and repairs;
Research and development;
Taxes other than income;
Selling, general, and administrative;
Depreciation of property, plant, and equipment;
Amortization of special tooling;
Pension expense; and
Other retirement benefits expense.
The new EPA study will analyze each of these, breaking them down in to sub-elements when
possible, to determine the extent to which each cost element responds to an increase in direct
costs resulting from an EPA regulation of the automotive industry. There may also be costs on
the dealer side that do not scale in proportion to increases in variable costs, and EPA intends to
evaluate these as well.
EPA also intends to evaluate the appropriateness of using an RPE multiplier to estimate
price. EPA believes that using such a multiplier achieves results that are inconsistent market
realities, and that a market model based on supply and demand elasticities would be a more
appropriate method for estimating price. Standard microeconomic theory tells us that setting the
MSRP based simply on an RPE multiplier will result in establishing a price that fails to
maximize corporate profits, and we do not believe that automotive manufacturers behave in this
way with respect to their pricing strategies.
4.3.1 Current Methodol ogy
For this analysis, we have estimated direct costs for individual technologies based on input
from existing published reports and information from automotive component suppliers and auto
manufacturers. We provide estimates of the direct costs incurred by the auto manufacturer, or, in
other words, what the auto manufacturer pays the component supplier for the component or
system. To estimate the indirect costs to the auto manufacturer and auto dealer, we are relying
on an indirect cost adjustment factor applied to the direct costs. This indirect cost factor is an
empirically derived multiplier intended to account for all indirect costs, including auto
manufacturer R&D, other engineering costs (primarily product integration), corporate overhead,
pensions, marketing, dealer support and profits. For this analysis, we are using an adjustment
factor of 1.5. As such, an indirect cost adjustment of 1.5 implies that direct manufacturing costs
represent two-thirds of the total costs while indirect costs represent 1/3 of the total costs. The 1.5
ANL factor is consistent with our approach in that we estimated that most of the technology
elements that we have considered are likely to be purchased from outside suppliers, rather than
developed and produced internally by the auto manufacturer.
68
-------
NHTSA developed the adjustment factor we are using in a report that analyzed the financial
reports of the domestic three auto manufacturers for the years 1989 through 1997.L We have
studied the NHTSA report and have considered the elements contained in the factor and the
amounts associated with those elements. We believe that an indirect cost markup factor of 1.5 is
a reasonable reflection of the cost structure typical in the auto industry. The relevant information
from the report and its analysis are reproduced here in Table 4.3-1.
Table 4.3-1 Domestic Three Auto Manufacturers Weighted Average Marginal Analysis of
Operating Results, 1989-1997 (Reproduced from Reference 6)
Chrysler, Ford, & GM
Net sales
Variable Costs - Manufacturing
Contribution Margin
Fixed & Discretionary Costs
Maintenance & Repairs
Research & Development
Selling, General, & Administrative
Taxes Other than Income
Pension Expense & Other Pension Related Benefits
Depreciation
Amortization, Tooling
Provision for Plant Closing
Amortization, Intangibles
Subtotal
Operating Margin
Net Profit Margin
Average for
1989-1997
100%
73.3%
26.7%
3.4%
4.7%
6.6%
2.6%
1.9%
2.8%
2.2%
0.3%
0.1%
24.6%
2.1%
0.6%
Adjustment
Factor
1.00
0.36
0.046
0.064
0.090
0.035
0.026
0.038
0.030
0.004
0.001
0.34
0.028
0.008
For this analysis, we have first estimated the cost to the automobile manufacturer for a
supplied part, or the price charged by the supplier to the manufacturer. That value would
represent the direct cost of manufacturing in the table above (i.e., "Variable Costs -
Manufacturing"). To estimate the markup that the manufacturer would apply to that value we
use the contribution margin above which is meant to reflect the firm's ability to cover
discretionary costs, fixed costs, interest expenses, taxes, and still leave a residual net profit. That
markup factor can be calculated as:
Contribution Margin/Variable Costs = Markup on Variable Costs
or,
26.7%/73.3% = 0.36
L Bruce Spinney, Barbara Fagin, Noble Bowie, and Stephen Kratzke, "Advanced Air Bag Systems: Cost, Weight,
and Lead Time Analysis: Summary Report," Contract No. DTNH22-96-0-12003, Task Orders 001, 003, and 005,
National Highway Traffic Safety Administration.
69
-------
In other words, we can estimate wholesale price by multiplying supplier price by a 1.36
factor. We can then estimate dealer price or retail price by accounting for dealer related costs.
The NHTSA study does this by using an 11% markup factor which was determined by
calculating a sales weighted comparison of suggested retail prices to wholesale prices for the
domestic manufacturers. In the end, the markup from manufacturer cost to retail price equivalent
is:
Retail price equivalent = Variable Cost x Markup on Variable Cost x Dealer Margin
or,
RPE = Variable Cost x 1.36 x 1.11 = Variable Cost x 1.5
4.3.2 Limitations and Uncertainties with Current Methodologies
As noted above, EPA is concerned that there are inherent limitations in an RPE approach
which may lead to overly conservative cost estimates. EPA also believes that even if one uses an
unqualified RPE approach, all of the work to date suffers from a number of limitations and
uncertainties in addition to the broad concerns with the use of an RPE approach noted above.
Consequently, in addition to the fundamental concerns with an RPE approach, EPA will attempt
to address the following limitations and concerns in our upcoming study.
4.3.2.1 Outdated Research
While the Jack Faucett Associates paper served as a source for a multiplier for a number of
EPA's regulatory actions, it is undeniable that the automotive industry has evolved in many ways
since the mid-1980s. Although the various critiques have suggested improvements on the
Faucett paper, EPA believes that it has been overtaken by time. EPA intends to use the most
recent financial and engineering information available in our upcoming analytical work.
4.3.2.2 Studies Limited to U.S. Domestic Manufacturers
All of the research to date has been limited to analyses of the financial information from the
U.S. domestic manufacturers. If, possible, it would be desirable to include financial information
from other large manufacturers, such as Honda and Toyota, in the analysis. However, due to
differences in accounting principles in different countries, past work has found it difficult to
incorporate foreign manufacturers due to the difficulty of reconciling their financial statements
with U.S. accounting methods. Nevertheless, these manufacturers represent a significant market
share in the U.S., and EPA intends to further evaluate the possibility of incorporating other
manufacturers in the analysis.
4.3.2.3 Treatment of Outsourced vs. Internally-Developed Technologies
As noted earlier, some of the more recent research has made a distinction between
outsourced and in-house technology sources. Given the large quantity of vehicle components
that are outsourced, EPA agrees that this distinction is one which should be drawn. It seems
reasonable to assume that a given vehicle component costs the same amount to develop and
manufacture, whether it is outsourced or produced in-house by the automobile manufacturer.
For an outsourced part, however, many of the indirect costs associated with making the part are
70
-------
assumed by the supplier and are included in the price charged to the vehicle manufacturer for the
component. If the part is developed and manufactured by the automobile manufacturer, then the
manufacturer bears all of the indirect costs associated with producing the part. We believe that
multiplying the cost charged by a supplier by an indirect cost multiplier should logically produce
the same total cost as multiplying the direct costs to a manufacturer by an indirect cost
multiplier. Given the assumption of some overhead costs by the supplier in the case of a
supplier-sourced part, the indirect cost multiplier for supplier-sourced parts should be lower than
the multiplier applied to a manufacturer's direct manufacturing costs if the same total RPE for
the part is the expected result. This concept is consistent with the approach that we have used in
our analysis described above and with the results of the ANL analysis. EPA's upcoming work
will evaluate markup factors for both outsourced components and those developed in-house.
4.3.2.4 Short-term and Long-term Adjustment Factors
EPA believes that a two-stage indirect cost multiplier may more accurately reflect the cost
impacts of a new regulation on auto manufacturers. The markup factor as it has traditionally
been used results in incremental indirect costs that continue in perpetuity. In reality, however,
we believe that a portion of the indirect costs included in the multiplier will not occur beyond the
first few years of implementation of a new regulatory program. For example, we believe that
manufacturers would continue to conduct research and development and would incur indirect
costs, but after the new regulation is fully implemented those costs would no longer be attributed
by a manufacturer to the regulation. Instead, indirect research and development costs would be
redirected to performance improvements, drivability, entertainment features, new product
development, etc. EPA expects our upcoming study to evaluate this issue and determine an
appropriate long-term indirect cost multiplier and when it should apply relative to the
implementation of a new regulation.
71
-------
Appendix 4.A Producer Price Index Adjustments
Throughout this analysis, we have presented our incremental cost estimates in terms of 2006
dollars. However, some of the data sources upon which we have based many of our estimates
presented costs in terms of dollars for other years. To convert the sourced estimates to 2006
dollars, we used the Producer Price Index (PPI) for the data series that most closely represented
the technology of interest. Table 4.A-1 shows the PPI series we have used for this analysis.
Table 4.A-1 Producer Price Indexes Used in this Analysis
Series Name
Lubricating oil and greases
Passenger car pneumatic tires
Carburetors, new and rebuilt (All types) [see note a\
Valves (engine intake and exhaust) [see note a]
Gasoline engines and engine parts for motor vehicles, new
Battery charging alternators, generators, and regulators
Motor vehicle steering and suspension components, new
Motor vehicle brake parts and assemblies, new
Motor vehicle drive train components except brakes and wheels,
new
Original equipment automotive stampings
Parts for manual and automatic transmissions, new [see note b]
Code
3241107
3262111
3363111
3363115
3363121
3363223
3363301
3363401
3363501
3363701
33635013
1997
80.9
92.1
137.8
130.0
99.4
129.8
NA
107.5
105.7
109.5
100.9
1998
59.3
89.5
139.2
126.9
97.7
128.7
NA
109.2
106.3
109.1
101.0
1999
70.2
88.1
140.3
126.4
96.6
127.4
NA
109.0
105.7
108.2
100.8
2000
113.6
88.1
144.1
126.9
95.4
126.6
99.7
108.6
105.0
108.3
99.7
2001
98.8
89.3
148.5
126.2
95.6
128.5
99.8
107.3
105.1
107.9
98.9
2002
102.6
90.6
153.6
124.8
96.8
128.8
100.0
106.7
104.7
107.1
98.8
2003
137.8
93.1
157.5
124.9
94.9
126.8
99.8
106.5
103.2
108.3
97.6
2004
136.4
97.7
163.3
143.3
96.6
126.1
101.3
105.4
102.4
109.7
96.6
2005
202.8
103.7
163.3
151.2
97.7
126.2
103.8
106.4
102.9
112.0
98.5
2006
219.4
111.9
181.7
151.2
107.2
126.9
104.9
107.0
104.4
112.5
100.3
Source: Bureau of Labor Statistics, U.S. Department of Labor, http://data.bls.gov/PDO/servlet/SurvevOutputServlet
a Data for years 2005 and 2006 are estimated due to incomplete monthly data for the year.
b Data for year 2006 is estimated based on data for December of 2005.
Table 4. A-2 shows the sources of data used in this analysis and our understanding of what
year in which the costs are expressed for each. With that, we were able to calculate a PPI
adjustment factor that could be applied to source estimates to express those source estimates in
2006 dollars. The resultant PPI adjustments are shown in Table 4.A-3, along with the PPI series
we considered to be the most applicable.
Table 4.A-2 Data Sources and Base Year Dollars
Source
NAS Study (Reference 11)
NESCCAF Report (Reference 12)
2006 EEA report to DOE (Reference 14)
EPA Interim Technical Report (Reference 20)
Transportation Research Board Special Report 286 (Reference 21)
NHTSA Light-duty Truck Rule (Reference 22)
Base Year of Costs
2001
2003
2006
2003
2006
2003
72
-------
Table 4.A-3 PPI Adjustments Used to Express Source Costs in 2006 Dollars
Technology
Low friction lubricants
Engine friction reduction
Overhead Cam Engines
VVT - intake cam phasing
VVT - coupled cam phasing
VVT - dual cam phasing
Cylinder deactivation
Discrete VVLT
Continuous VVLT
Overhead Valve Engines
Cylinder deactivation
VVT - coupled cam phasing
Discrete VVLT
Continuous VVLT (includes conversion to Overhead Cam)
Camless valvetrain (electromagnetic)
GDI - stoichiometric
GDI - lean burn
Gasoline HCCI dual-mode
Turbocharge + engine downsize
Diesel - Lean NOx trap
Diesel - urea SCR
Optimized E20-E30
Aggressive shift logic
Early torque converter lockup
5-speed automatic
6-speed automatic
6-speed AMT
6-speed manual
CVT
Stop-Start with 42 volt system
IMA/IS A/BSG (includes engine downsize)
2-Mode hybrid electric vehicle
Power-split hybrid electric vehicle (P-S HEV)
Full-Series hydraulic hybrid
Plug-in hybrid electric vehicle (PHEV)
Full electric vehicle (EV)
Improved high efficiency alternator & electrification of accessories (12
volt)
Electric power steering (12 or 42 volt)
Improved high efficiency alternator & electrification of accessories (42
volt)
Aero drag reduction (20% on cars, 10% on trucks)
Low rolling resistance tires (10%)
Low drag brakes (ladder frame only)
Secondary axle disconnect (unibody only)
Front axle disconnect (ladder frame only)
PPI Series
3241107
3363121
3363121
3363121
3363121
3363121
3363121
3363121
3363121
3363121
3363121
3363115
3363121
3363121
3363121
3363121
3363121
3363121
3363121
3363121
33635013
33635013
33635013
33635013
33635013
33635013
33635013
3363223
3363223
3363223
3363223
33635013
3363223
3363223
3363223
3363301
3363223
3363701
3262111
3363401
3363501
3363501
PPI Adjustment
1.000
1.121
1.130
1.130
1.130
1.130
1.130
1.130
1.130
1.130
1.130
1.211
1.121
1.000
1.000
1.130
1.000
1.000
1.000
1.000
1.014
1.000
1.014
1.000
1.000
1.000
1.028
1.001
1.001
1.001
1.001
1.028
1.000
1.000
0.988
1.051
0.988
1.000
1.000
1.005
1.012
1.012
73
-------
References
1 "Light-Duty Automotive Technology and Fuel Economy Trends: 1975 Through 2007", EPA420-S-07-001,
September 2007
2 "Reducing Greenhouse Gas Emissions from Light-Duty Motor Vehicles," Northeast States Center for a Clean Air
Future, September 2004, Table B-13.
3 "Reducing Greenhouse Gas Emissions from Light-Duty Motor Vehicles," Northeast States Center for a Clean Air
Future, September 2004.
4 This estimate is primarily based on the 2002 NAS report findings which projected a 5-10% COa reduction over
WLT systems. Information from EEA (2006) suggests a 12-13% CO2 reduction over fixed valvetrains.
5 "An HCCI Engine Power Plant for a Hybrid Vehicle" (SAE 2004-01-0933). Sun, R., R. Thomas and C. Gray, Jr.,
EPA, 2004.
6 Dominique Portmann, DaimlerChrysler presentation to 2006 EDTA http://www.vito.be/ieahev/annexl2/2006-ll-
30%20PHEV%20Prototype%20and%20Field%20Test%20Evaluation%20Program_Portmann.pdf
7 "Reforming them Automobile Fuel Economy Standards Program," Docket NHTSA-2003-16128-1120
8 National Energy Modeling System, Energy Information Administration, U. S. Dept of Energy.
9 "A Study of Potential Effectiveness of Carbon Dioxide Reducing Vehicle Technologies," EPA Report No.
EPA420-R-08-004, available on the Internet at http://www.epa.gov/OMS/technology/420r08004.pdf.
10 Preliminary Regulatory Impact Analysis: Corporate Average Fuel Economy and CAFE Reform for MY 2008-
2011 Light Trucks, NHTSA, August 2005.
11 "Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards," National Research Council,
National Academy of Sciences, 2002.
12 "Reducing Greenhouse Gas Emissions from Light-Duty Motor Vehicles," Northeast States Center for a Clean
Air Future, September 2004.
13 "Staff Report: Initial Statement of Reasons for Proposed Rulemaking," California Environmental Protection
Agency, Air Resources Board, Regulations to Control Greenhouse Gas Emissions from Motor Vehicles, August 6,
2004.
14 "Technology to Improve the Fuel Economy of Light Duty Trucks to 2015," Energy and Environmental Analysis,
Inc., May 2006.
15 National Highway Traffic Safety Administration, Request for Product Plan Information, 70 FR 51466, August
30, 2005.
16 "Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards," National Research Council,
National Academy of Sciences, 2002; Vyas, A., Santini, D., Cuenca, R., "Comparison of Indirect Cost Multipliers
for Vehicle Manufacturing," Argonne National Laboratory, April 2000; "Update of EPA's Motor Vehicle Emission
Control Equipment Retail Price Equivalent (RPE) Calculation Formula," Work Assignment 3, Contract No. 68-03-
3244, September 4, 1985; "Final Peer Review Document: NHTSA's Cost, Weight, and Lead Time Estimating
Methodology for New Safety Initiatives," DTNH22-02-D-02104, Task Order 09, October 23, 2006.
17 "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur
Control Requirements," EPA420-R-00-026, December 2000; "Final Regulatory Analysis: Control of Emissions
from Nonroad Diesel Engines," EPA420-R-04-007, May 2004; "Draft Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Locomotive Engines and Marine Compression-Ignition Engines Less than 30 Liters
per Cylinder," EPA420-D-07-001, March 2007.
74
-------
18 "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February 23, 1990, Vol. 247, pp.
920-924.
19 "Treating Progress Functions As Managerial Opportunity", J.M Button and A. Thomas, Academy of
Management Review, Rev. 9, 235, 1984, Public Docket A-2001-28, Docket Item II-A-73.
20 "Progress Report on Clean and Efficient Automotive Technologies Under Development at EPA, Interim
Technical Report," EPA420-R-04-002, January 2004.
21 "Tires and Passenger Vehicle Fuel Economy," Transportation Research Board Special Report 286, National
Research Council of the National Academies, 2006.
22 "Final Regulatory Impact Analysis: Corporate Average Fuel Economy and CAFE Reform for MY 2008-2011
Light Trucks," Office of Regulatory Analysis and Evaluation, National Center for Statistics and Analysis, U.S.
Department of Transportation, March 2006.
23 "Update of EPA's Motor Vehicle Emission Control Equipment Retail Price Equivalent (RPE) Calculation
Formula," Jack Faucett Associates, September 4, 1985.
24 Anant Vyas, Dan Santini, and Roy Cuenca, "Comparison of Indirect Cost Multipliers for Vehicle Manufacturing,"
Center for Transportation Research, Energy Systems Division, Argonne National Laboratory, April 2000.
25 "Study of Industry-Average Mark-up Factors Used to Estimate Changes in Retail Price Equivalent (RPE) for
Automotive Fuel Economy and Emissions Control Systems," Sierra Research, Inc., November 21, 2007.
75
-------
Attachment 1
-------
THE NATIONAL ACADEMIES
Advisers to the Nation on Science, Engineering, and Medicine
Division on Engineering and Physical Sciences 500 Rflh Street, NW
Board on Energy and Environmental Systems Washington, DC 20001
Phone: 202 334 3344
Fax; 202 334 2019
www.nationalacademiBS.org
March 4,2008
The Honorable Stephen L. Johnson
Administrator
United States Environmental Protection Agency
1200 Pennsylvania Avenue, N.W.
Mail Code 1101A
Washington, DC 20460
Dear Administrator Johnson:
I am serving as the chair of the National Research Council's Committee on the Assessment of
Technologies for Improving Light-Duty Vehicle Fuel Economy. The committee is tasked with
providing updated estimates of the cost and potential efficiency improvements of light-duty
vehicle technologies that might be employed to improve fuel economy. Our group has received
a peat deal of cooperation from your staff, including presentations related to the agency's
activities on control of vehicle greenhouse gas emissions. I am writing today to request that the
Environmental Protection Agency provide to the committee its report, Greenhouse Gas Vehicle
Proposal, which would become part of the committee's public record on the informatlbn-used in
its deliberations. It is my understanding that this report details much of the agency's analysis of
the control of greenhouse gas emissions from light-duty vehicles. I believe the information
would greatly aid the committee in its work. I look forward to your reply.
Sincerely.,
Trevor 0. Jones, Chair
Committee on the Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy
cc: K. John Holmes, National Research Council
NATIONAL ACADEMY OF SCIENCES • NATIONAL ACADEMY OF ENGINEERING * INSHTUTE OF MEDICINE « NATIONAL RESEARCH COUNCIL
ii
------- |