Summary and Analysis of Comments
on THE
Notice of Proposed Rulemaking
for
Revised Gaseous Emission Regulations
for 1984 and Later Model Year Light-Duty
Trucks and Heavy-Duty Engines
July 1983
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Sources
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency
-------
Summary and Analysis of Comments
on the
Notice of Proposed Rulemaking
for
Revised Gaseous Emission Regulations
for 1984 and Later Model Year Light-Duty
Trucks and Heavy-Duty Engines
July 1983
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Sources
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
-------
Table of Contents
Page
I. Introduction . iii
II. List of Commenters iv
III. Analysis of Issues
A. Primary Issues 1
1. Technology/Standards 1
2. Useful Life 43
3. Alternative Cycles and Standards 72
4. Environmental Impact 118
B. Secondary Issues . 130
1. Deterioration Factors 130
2. Idle CO Test and Standards 134
3. Fuel Economy 140
4. Allowable Maintenance . 149
5. Minor Amendments to HDE/LDT SEA 159
6. Split Standards - Gasoline-Fueled vs.
Diesel Engines 181
7. Cold-Start Test Requirements . 188
8. Diesel Engine CO Measurement 196
9. Parameter Adjustment 197
10. Potential Impacts on Specific
Manufacturers 200
11. Transient Test Procedures 206
12. Possible "Migration" from Class IIB
to Class III 225
13. Diesel Engine Closed Crankcase
Requirements 230
-------
-ii-
Table of Contents (cont'd.)
Page
IV. Appendices
A. Draft Technological Feasibility from NPRM ... A1
B. Transient Test Study B1
-------
I. Introduction
On January 13, 1982, the Environmental Protection Agency
(the Agency) published a Notice of Proposed Rulemaking (NPRM)
in which the Agency considered revised gaseous emission
regulations for 1984 and later model year light-duty trucks and
heavy-duty engines. Although the major thrust of this action
was to propose non-catalyst emission standards for heavy-duty
engines, the Agency also requested and received comment on a
large number of other issues related to the 1984 emission
control requirements for light-duty trucks and heavy-duty
engines.
To seek further clarification and comment on issues raised
by the initial NPRM, several opportunities were offered for
comment, including a further request for comments published in
the Federal Register on March 12, 1982. A final rule was
published in January 1983. Also, to achieve final resolution
on the useful-life requirement, a further NPRM on the 1985
light-duty truck and heavy-duty engine useful-life requirements
was published in January 1983.
This document presents a Summary and Analysis of Comments
received in response to the NPRM and the subsequent requests
for comment mentioned above. The useful-life discussion
presented as Primary Issue 2 serves as the study of the
useful-life requirements discussed in the Federal Register
Notice of April L3, 1981. The transient test study undertaken
as a result of the same Federal Register notice is included as
Appendix B.
-------
-iv-
II. List of Commenters
1. American Motors Corporation (AMC)
2. Caterpillar Tractor Company
3. Chrysler Corporation
4. Cummins Engine Company
5. Engine Manufacturers Association (EMA)
6. Ford Motor Company
7. Freightliner Corporation
8. General Motors Corporation OGM)
9. U.S. Senator Gary Hart
10. Hino Motors, Limited
11. International Harvester Company (JHC}
12. League of Women Voters of Carson City, Nevada
13. League of Women Voters of the Doyleston Area
14. League of Women Voters of the United States
15. Mack Trucks, Inc.
16. Manufacturers of Emission Controls Association (MECA)
17. Mercedes-Benz of North America (MB)
18. Mrs. V, H. Worse
19. Motor Vehicle Manufacturers Association (MVMA)
20. National Association of Van Pool Operators (HAVPO)
21. National Automobile Dealers Association (NADA)
22. Natural Resources Defense Council (NRDC)
23. New York City League of Women Voters
24. Regional Air Pollution Control Agency, Dayton, Ohio
25. Frances Scherer
26. Toyota Motor Company
27. Western New York Allergy and Ecology Association
28. Volkswagen of America (VWoA)
-------
A. Primary Issues
1. Issue: Technological Feasibility
Summary of the Issue
This analysis addresses the technological feasibility of
emission standards for heavy-duty engines (HDEs) for 1985 and
later model years. Two separate analyses are contained
herein: an analysis and derivation of hydrocarbons (HC) and
carbon monoxide (CO) emission standards for 1985 which are
achievable without catalysts, and an analysis of the
feasibility of catalyst-based standards for 1987 and later
model years.
A. NON-CATALYST STANDARDS FOR 1985
Summary of the Comments/Synopsis of Events
There have been several iterations of EPA action and
public reaction as this issue has developed over time. For
purposes of clarity, a brief synopsis of significant events is
appropriate; public comments to each iteration will be
summarized as they chronologically occurred.
On January 21, 1980, EPA promulgated final regulations for
the control of gaseous emissions from HDEs applicable to the
1984 and later model years.[1] The regulations included the
new EPA transient test cycles, the full useful-life concept,
and statutory emission standards of 1.3 grams per brake
horsepower-hour (g/BHP-hr) HC, 15.5 g/BHP-hr CO, and 10.7
g/BHP-hr oxides of nitrogen (NOx).* Compliance with these
emission standards on the transient test almost certainly
requires the use of oxidation catalysts on heavy-duty gasoline
engines {HDGEs).
On April 6, 1981, the Vice President's Task Force on
Regulatory Relief announced that EPA would propose emission
standards for 1984 which would not require the use of
catalysts. It was intended that this action defer the capital
investments required for catalyst development, and thus provide
economic relief to an industry beset by recession and decreased
sales. On January 13, 1982, EPA officially proposed
non-catalyst emission standards of 1.3 HC/35.0 CO/lO.7 NOx for
1984 HDEs.[2]
The associated Draft Regulatory Analysis[3] tentatively
concluded that emission standards of 1.3 HC/35.0 CO were
feasible without cataiysts. The analysis discussed the
transient test in great detail, and presented modal emissions
All standards are based upon the EPA HDGE transient cycle.
-------
from 12 1979 model year HDGEs. Emission levels for current
technol'ogy were discussed, as were the emissions impact of
specific operational modes of the transient test. By comparing
low. to high emitting engines and by identifying specific
technologies and calibrations, EPA made the judgment that
1.3/35.0 appeared feasible for 1984 non-catalyst HDGEs.
In public comments to the Notice of proposed Rulemaking
(NPRM) received by April 1982,[4,5] only Ford Motor Company
(Ford) and General Motors (GM) submitted transient test
emissions data. Chrysler and international Harvester Company
(IHC) did not comment on technological feasibility, and in
fact, indicated that they were leaving the HDGE market for
reasons unrelated to these regulations. (Both Chrysler and IHC
have since indicated that they may reverse these decisions.)
In its comments, Ford stated its position as follows:
"Ford believes that its recommended 3.3 HC and 42 CO
g/BHP-hr standards for the 1985 model year HDGEs represent
the lowest levels achievable without unreasonable
sacrifices in performance, fuel economy, or driveability."
General Motor's comments stated that:
"lieview of available 1984 prototype HDGE development data
indicated that most GM HDGEs could achieve low-mileage
emission levels of approximately 2.0 HC and 32 CO
g/BHP-hr."
General Motors subsequently recommended emission standards
of 2.9 g/BHP-hr HC and 43.0 g/BHP-hr CO, on the basis of
increased certification deterioration factors and assumed
production variability. GM also argued that standards of 3.5
g/BHP-hr HC and 70 g/BHP-hr CO were justified on the basis of
air quality needs, fuel economy, and cost. in comments and
later discussions, GM also raised the point that the emission
control strategies required to reduce HC and CO emissions to
EPA's proposed standards could severely degrade engine
durability. GM claimed that the need for full-power mixture
e'nleanment and increased oxidation of pollutants in the exhaust
system will raise in-cylinder and exhaust system temperatures
to excessive levels. GM said that this will not necessarily be
seen on EPA's transient test procedure, but more than likely
will be seen in severe in-use applications for engines
calibrated to meet EPA's proposed requirements.
Emissions data for 1984 prototype HDEs, as submitted by GM
and Ford in April of 1982, are listed in Table 1-1.
-2-
-------
Table 1-1
Manufacturers' 1984 Prototype Heavy-
Duty Engine Data (submitted by April 1982)
Engine
Emission
q/BHP-hr*
BSFC
Manufacturer
Displacement
Control System
HC
CO
NOx
lb/BHP-hr
Ford
4.9L
AIR/EGR/EFE
1.66
23.2
7.68
0.560
Ford
6.1L
AIR/EGR
2.33
28.8
7.25
0.654
Ford
7.0/7.5L
AIR/EGR
2.21
24.3
4.82
0.633
GM
292 in3
AIR/EGR
1.65
17.42
6.52
--
GM
350 in3
AIR/EGR
1.76
25.07
5.25
GM
366 in3
AIR/EGR
1.33
20.19
6.91
--
GM
454 in3
AIR/EGR
0.90
20.93
7.42
—
EPA cycle based.
-3-
-------
In reviewing these comments, EPA staff felt that
additional engineering data were required to determine the
lowest emission standards achievable without catalysts.
Specific requests for more detailed information were made to
HDGE manufacturers on June 17, 1982;[6] Ford provided
additional engineering data, GM provided a more detailed
qualitative discussion and emission data {see Table 1-2) but
declined ;to submit detailed engineering data, and Chrysler and
IHC wer£ unable to provide any additional data or information.
At a meeting with EPA staff on January 28, 1983,
representatives of GM again made the claim that 1.3/35.0
non-catalyst emission standards would adversely affect engine
durability and fuel economy.
In a Federal Register notice of January 12, 1983, EPA
officially delayed the 1984 model year emission requirements
until 1985. This revision of the 1985 standards was justified
on the basis of Leadtime,* economics,* and the number of other
issues yet to be resolved (i.e., alternative test cycles,
useful life, etc.].
Reviewing all comments and data available at the time, and
taking into account the additional year of development
leadtime, EPA then, analyzed the level of non-catalyst emission
standards achievable for 1985. This analysis [8] went
hand-in-hand with an EPA staff paper[7] in which both short and
long-term strategies for the control of HDGE emissions were
discussed. The staff paper, which was released for public
comment on March 16, 1983, developed a control scenario whereby
lighter heavy-duty gasoline trucks (HDGTs) would be equipped
with catalysts in the 1987-88 time frame, and heavier gasoline
truck engine standards would remain at non-catalyst levels.**
At the same time, the staff paper summarized EPA's most recent
analysis of 1985 standard feasibility, which had recommended
non-catalyst standards of 2.5/35.0 g/BHP-hr.
This feasibility analysis,[8] the summarized results of
which were discussed at an April 6, 1983 Public Workshop, was
also distributed for public comment on April 12, 1983. The
analysis recommended that non-catalyst emission standards of
2".5/35.0 g/BHP-hr be promulgated for 1985. This recommendation
revised EPA's earlier conclusion that a 1.3 g/BHP-hr HC was
feasible for 1984 without catalysts.
* See appendix, Chapter 3 of the Transient Test Study.
** See the POST-1985 EMISSION STANDARDS section of this issue.
-4-
-------
Table 1-2
Additional Emission Data
Provided by GM in August of 1982[5]
Low-Mileage Emissions*
Enqine
HC
CO
NOx
BSFC
GM 292
2.17
24.9
6.80
.639
GM 350-2V
1.57
28.2
6.11
.604
GM 350-4V
1.99
27.2
5.14
.649
GM 366
.75
17.9
3.57
.582
GM 454
1.01
22.2
4.28
.666
EPA cycle-based g/BHP-hr. GM claimed that these data are
representative of heavy-duty gasoline engine emission
control systems and calibrations which, in August of 1982,
were believed to be "at least plausible for production."
None of these engine configurations had been durability
tested, but all. had been driven in a small sample of
vehicles and had been determined to provide commercially
acceptable performance and driveability.[5]
-5-
-------
EPA's more recent conclusion that .the non-catalyst HC
standard must be relaxed was based upon several
considerations. A review of the actual development data
submitted by Ford and GM in April 1982 (see Table 1-1)
indicated that substantial progress had been made in reducing
emissions. However, all but one engine family were still well
above the low-mileage target emission levels needed to assure
compliance with a 1.3 g/BHP-hr HC standard. (All engine
families were very close to the low-mileage target level for CO
more than two years before required compliance, hence no
relaxation of the CO standard was recommended.) EPA's analysis
then discussed the remaining technology which could be applied
to reduce HC emissions further. Since only Ford supplied
detailed engineering data in response to EPA's June 17, 1982
request, only an analysis of Ford's product line was possible.
(Since GM's engines in Table 1~1 all exhibited HC emission
rates less than most Ford engines, it was judged that GM would
have no problem complying with an HC emission standard based
upon Ford's higher emitting engines.) Using data provided by
Ford, EPA concluded that further reductions in HC were
certainly possible, and that compliance with a 2.5 g/BHP-hr HC
standard in 1985 would be possible even for Ford's highest
emitting engine. HC standards less than 2.5 g/BHP-hr were
considered, but were rejected on the basis of reasonable risk
of non-compliance and fuel economy penalties for higher
emitting engines. In summary, 2.5/35.0 were recommended as
reasonable interim emission standards.
In comments[5] received by May 6, 1983, the conclusions
and methodology of EPA's latest feasibility analysis[8] were
again disputed. These comments are summarized below for each
commenter.
Chrysler
Chrysler again commented that it was in no position to
recommend specific interim standards, primarily because its
transient test facility was not yet operational. Based upon
testing performed for it under contract, however, Chrysler did
not believe that the 35.0 g/BHP-hr CO standard was feasible
even with a catalyst. Chrysler recommended continued provision
of the 9-mode steady-state option until 1986.
Engine Manufacturer's Association (EMA)
The EMA and its member companies have not disputed the
feasibility of the 1.3/35.0 g/BHP-hr standards for heavy-duty
diesel engines (HDDEs).
-6-
-------
Ford
Ford provided a comprehensive review of the emissions
status of its HDGE product line. Ford>s data showed that
significant progress has been made in reducing emissions.
However, Ford disputed EPA's feasibility analysis,
characterizing it as overly optimistic. "EPA's suggestion that
manufacturers not only can meet 2.5/35.0 but can also achieve
substantial further reductions is overstated." Ford also
argued that "EPA has overestimated the capabilities of some
heavy-duty engines," notably Ford's 6.1L-4V (Ford's largest
seller and occupant of the heaviest gasoline vehicle weight
classes). The major problem associated with feasibility,
according to Ford, is not so much the effectiveness of
technology but rather the relatively low target levels which
are forced upon a manufacturer by the full useful-life and
Selective Enforcement Audit (SEA) requirements. Ford
recommended half-life standards of 2.19/42.6 based upon the
MVMA cycle; according to Ford, these are equivalent to
full-life EPA cycle standards of 3.07/47.8.
General Motors
General Motors vigorously disputed the conclusions of
EPA's feasibility analysis, characterizing the analysis as
"entirely inadequate," and mostly "guesswork." GM insisted
that EPA's engineering judgment was based upon limited and
outdated emission data, very few research studies, and limited,
incomplete data supplied by manufacturers. GM also criticised
EPA for "engineering on paper," for failing to construct and
evaluate through testing, any engine conforming to EPA's design
recommendations, and for failing to generate any current data.
General Motors qualitatively discussed several engineering
aspects of achieving low levels of HDGE emissions without
catalysts. GM also discussed durability, driveability, and
fuel economy problems associated with "unreasonably stringent
standards." in GM's opinion, forced compliance with the 35.0
g/BHP-hr CO standard would preclude the production of
reasonably durable engines. indeed, GM argued that the poor
performance of engines produced under these emission
constraints would invite tampering in the field.
General Motors questioned EPA's apparent policy of
establishing stringent interim standards, especially given the
major changes occurring in the 1987-8B timeframe, the "risk to
the heavy-duty industry," and "the lack of demonstrated
feasibility." GM recommended half-life non-catalyst emission
standards of 2.9/43.0 (EPA cycle) for 1985.
-7-
-------
With respect to data, GM submitted a large confidential
discussion of various aspects of its development work.
Included were qualitative discussions of GM's calibration
strategies and hardware for complying with 1.3/35.0 g/BHP-hr
standards, qualitative discussions of GM's engine durability
experience, a description of GM's in-house durability test
procedures, comparisons of 1983 versus 1985 prototype timing
and air/fuel (A/F) calibrations, actual test reports from
characterizations of wide-open throttle (WOT) timing versus
detonation requirements, on-road fuel economy data, and actual
test reports from exhaust system temperature and durability
studies.
On the other hand, no new emission data were submitted by
GM; the latest data indicating the position of GM's product
line with respect to compliance was that submitted by April and
August of 1982 {see Tables 1-1 and 1-2) . GM went on to
characterize the April 1982 data as being unrepresentative of
its true compliance capability, having been acquired long
before subsequent testing discovered durability problems.
Furthermore, GM stated that its test experience and comments
only address the feasibility of the 1.3/35.0 standards. GM
claimed that it had only just begun to evaluate the
implications of the 2.5/35.0 standards. Nevertheless, GM
recommended that EPA promulgate half-life standards of 2.9/43.0
(EPA cycle) g/BHP-hr for 1985.
Analysis of Comments
Overview
This analysis will develop and recommend non-catalyst
standards for 1985 and later model year HDGEs. Aside from the
specific hardware and applicable emission control techniques to
be addressed, it is equally important to address the effect of
other factors on the stringency of the interim emission
standards. The most important of these other factors are the
full useful-life concept, the SEA requirements, and the
correlation between the EPA and MVMA test cycles. These issues
will be discussed first, because of their inherent impact on
standard stringency.*
The relationship of two of these factors to low-mileage
emission targets and emission standards is typically
expressed as:
[Emission Standard - DF] = Low-mileage target.
-8-
-------
The analysis will then review the status of the HDGE
manufacturers with respect to current emission levels of their
"best effort" engines; where possible. As in earlier analyses,
judgments will be made whether further reductions can be made
for the 1985 model year.
Deterioration Factors
Manufacturers are required to correct "low-mileage"
emission levels from certification engines for expected in-use
deterioration. Current requirements, and those applicable for
1985, require that deterioration be assessed in an additive
fashion. Current deterioration factors (DFs) have largely been
derived from durability testing performed on engine
dynamometers. Very little, if any, data exists on the degree
of deterioration which actually occurs in use. Dynamometer
durability testing results have never been validated, and there
is substantial uncertainty as to the magnitude of true in-use
dfs for all hdes.
On the other hand, the process of compliance in 1985 will
be based upon dfs derived and supplied by the manufacturer in
whatever manner they deem appropriate. Techniques of DF
derivation can range from simple engineering judgment to
continued use of dynamometer testing to actual in-use tests.
Given the lack of an officially imposed method, one would
expect manufacturers to base their DF determinations upon past
practice and experience.
Certification dfs for HDGEs have typically been quite
small. Table 1-3 presents a summary of official certification
DFs for Ford's and GM's HDGEs for the 1983 model year. in
almost all cases, emissions decreased after completion of
durability test runs on the engine dynamometers. Substantial
changes, however, are being made to engine hardware for the
1985 model year. This new hardware will also be required to
maintain compliance for a full useful life (110,000 miles), as
opposed to the previous half-life (50,000 miles) requirement.
Therefore, the dfs in Table 1-3 may be somewhat less than dfs
derived and used for 1985.
in past analyses, EPA has converted from half- to
full-life DFs by assuming linear deterioration (i.e, the
full-life DF is equal to the half-life DF multiplied by
110,000/50,000, or 2.2). This methodology is straightforward,
and fits the general trend of deterioration observed in
dynamometer testing of non-catalyst engines. While EPA has
confidence in this adjustment, assessing the deterioration
rates of new engine hardware not yet in production is more
problematic.
Based upon current prototypes, 1985 HDGEs will likely be
equipped with the following hardware: large dual air pumps,
-9-
-------
Table 1-3
1983 Model Year Certification
Deterioration Factors (DFs) for Ford and GM HDGEs
Certification DFs*
Manufacturer/Engine Family HC CO
Ford 4.9L "Q" 0.00 0.00
Ford 6.1L "E" - 2V 0.00 1.91
Ford 6.1L "E" - 4V 0.00 1.91
Ford 7.0L "E" 0.00 0.00
Ford 7.5L "E" 0.00 0.00
Ford 5.8L (W) "E" 0.00 0.48
GM DGM07.0ABB4:
- L86 (366 CID) 0.00 0.00
- L43 (427 CID) 0.00 0.00
GM DGM07.4ABB9:
- LF8 (454 CID) 0.00 0.00
GM DGM04.8ABA6:
- L25 (292 CID) 0.00 0.00
GM DGM05.7ABB9:
- LF5 (350 CID - 2V) 0.00 0.00
- LS9 (350 CID - 4V) 0.00 0.00
Additive g/BHP-hr, half-life basis.
-10-
-------
EGR, early fuel evaporation systems, heated air intake systems,
and automatic chokes. Carburetor and ignition timing
calibrations will be different from current models, as will
manifold designs and air injection systems. in-cylinder and
exhaust system temperatures will be hotter than those of
current engines because of leaner mixtures and increased
thermal reaction. as a total package, these modifications are
uncharacterized in heavy-duty engine applications with respect
to deterioration and long-term performance. Given the
significant changes from 1984 to 1985, it is reasonable to
expect that manufacturers will run at least some dynamometer
durability tests out to the full useful-life equivalent of
3,300 hours.
Quantification of expected deterioration is by necessity
somewhat speculative, but there are a number of reasons why
1985 DFs should not be exceptionally high:
1. No inherent increase in deterioration rates should
be expected from recalibrations of ignition timing or
carburetors. Deterioration rates of this hardware have been
previously established, and simple changes to timing settings
or fuel flow rates should not alter the functional durability
of the hardware.
2. Catastrophic or significant causes of deterioration
to minor components will be identified during accelerated
durability testing, at which time corrective redesign can take
place.
3. If problems arise with component-related durability,
especially during dynamometer testing corresponding to the
second half of the useful life, new maintenance provisions can
be specified to alleviate the problem.
4. Prototype air injection systems are merely larger
versions of existing systems whose durability performance have
already been characterized. Other changes simply represent
changes to static piping and manifolds; these hardware
experience minimal emission-related deterioration.
5. Finally, most of the hardware new to HDGEs have
already been successfully used on production LDVs and LDTs for
several years. EPA expects the manufacturers to have acquired
considerable experience with the design, maintenance, and
in-use durability of such hardware. This experience is
directly relatable to HDGEs.
Given the above, and given the dfs presented in Table 1-3,
EPA does not expect large dfs to be used or needed for 1985.
-11-
-------
Referring to Table 1-3, only two engine families exhibited
a non-zero CO DF for 1983 (the Ford 6.1L and 5.8L). EPA does
not expect CO deterioration rates to be significantly different
from 1983; CO emission control is primarily a function of
leaner carburetor calibration and improved air injection,
neither of which should affect durability to any great extent.
For purposes of this analysis, EPA will use the worst case DF
from 1983 (1.91 for the Ford 6.1L), corrected from half- to
full-life (i.e., 1.91 multiplied by 110,000/50,000 or 2.2 to
equal 4.20).
Quantification of HC deterioration is more speculative.
All 1983 Ford and GM engine families exhibited HC DFs of 0.00
or less, but there is some reason to believe that HC DFs may
increase in 1985. Cold start emission control apparatus is
new, as would be more elaborate ignition timing controls (if
used). These systems will primarily affect HC emissions. On
the other hand, systems of this type have already been used for
several years on production LDVs and LDTs, and EPA presumes
that the manufacturers have well characterized. their
performance. For purposes of this analysis, EPA will use the
same additive DF used in the earlier analysis, [8] a DF of .25.
This is likely to be a representative DF, given the performance
of current engines, the existing experience with such equipment
on LDTs and LDVs, and EPA's assumption of a moderate increase
in DFs for 1985.
SEA Requirements
SEA testing requirements are scheduled to take effect in
the 1986 model year. Therefore, a manufacturer cannot be
subjected to the jeopardy of failing a production line audit
until 1986. For this reason, it is entirely reasonable to
ignore SEA requirements in establishing emission target levels
for 1985. On the other hand, it is also reasonable to expect
that a manufacturer would wish to conclude development work
prior to 1985, and rely upon carryover for the next year to
avoid continued recertification expenses. This feasibility
analysis will include the effect of SEA requirements in
establishing feasible emission standards for 1985, since
recertification in 1986 would not be desirable from the
manufacturers' standpoint. However, in the event that one or
two engines may appear to be having difficulty in achieving
SEA-based low-mileage target levels for 1985, EPA cannot ignore
the additional flexibility provided manufacturers by the
effective relaxation of low-mileage target levels afforded by
EPA's deferral of SEA requirements.
Production line emission variability is fairly well
characterized. EPA's earlier analysis[8] and Ford's May 6,
1983 comments[5] used numerical values of 1.136 and 1.200,
-12-
-------
respectively, for HC, and 1.266 and 1.300, respectively, for
CO. (GM has previously used a 40 percent AQL factor of 1.10
for all gases.[5]) EPA's and Ford's values are essentially in
agreement; for purposes of this analysis, arithimetic averages
of Ford's and EPA's numbers will be used (i.e., 1.168 for HC
and 1.283 for CO), and represent conservative values in EPA's
judgment. For worst case engines, however, a value of 1.000
would be available for the 1985 model year.
Alternative Test Cycles
EPA's earlier analyses[3,7,8,9 ] were all based upon EPA
cycle test results, and the emission standards discussed were
also based upon the EPA cycle. All of the latest."best effort"
emission data, however, is MVMA cycle based. For purposes of
this analysis, MVMA cycle-based standards will be developed
from this "best effort" data. For purposes of comparability
with previous analyses, equivalent EPA cycle-based standards
will also be presented.*
Only Ford gave EPA specific information on the current
emissions status of their product line. As in its earlier
analysis,[8] EPA will evaluate the feasibility of emission
standards for HDGEs based largely upon Ford's data. In the
absence of any specific emissions data to the contrary, and by
reviewing the latest GM emission data made available to. EPA in
August of 1982, EPA will assume that the emissions capabilities
of GM's engines are not substantially different from Ford's.
The emissions capabilities of Chrysler's and IHC's engines are
unknown, however, the necessary technologies are widely
available and well understood. EPA does not expect the
emissions capabilities of Chrysler's or IHC's engines to be
fundamentally different from those of Ford or GM.
Current Status of HDGE Emission Levels
Tables 1-4, 1-5, and 1-6 present EPA's evaluation of
Ford's "best effort" data; Table 1-5 also includes GM's most
recent data. Clearly, significant improvements have been made
since 1979 and earlier model years (see Figures 1-1 and 1-2).
Using Ford's recent MVMA cycle-based low-mileage results, these
levels have been converted to equivalent emission standards,
both in terms of the MVMA cycle and the EPA cycle (see Tables
* Equivalent EPA cycle-based emissions will be based upon
the following equations (derived in Issue A.3. of this
Summary and Analysis of Comments):
HC: MVMA = .886 (EPA) - 0.318
CO: MVMA = 1.0 3 (EPA) - 4.04
-13-
-------
Table 1-4
Review of Ford's "Best Effort" Data,[5] Estimated
Current Attainable MVMA Cycle-Based Emission Standards*
May 1983 Data
Ford
Engine Model
4.9L-1V
6.1L-2V
6.1L-4V
7.0L-4V
7.5L-4V
Low-Mileage
Emission Results*
HC
CO
1.21(1.72) 25.0(28.2)
1.08(1.58) 23.4(26.6)
1.70(2.28) 28.5(31.6)
1.50(2.05) 17.7(21.1)
1.36(1.89) 23.6(26.8)
Equivalent
Deteriorated
HC Emission
Standard*
1.66(2.26)
1.51(2.09)
2.24(2.91)
2.00 (2.65)
1.84(2.46)
Equivalent
Deteriorated
CO Emission
Standard*
36.3(40.4)
34.2(38.4)
40.8(44.7)
26.9 (31.3)
34.5(38.6)
MVMA cycle based, g/BHP-hr; the numbers in parenthesis are
EPA cycle based, g/BHP-hr.
-14-
-------
Table 1-5
Currently Attainable
EPA Equivalent Emissions Standards
Engine Family
Equivalent, Deteriorated
EPA Cycle-Based
HC Emission Standard*
Equivalent, Deteriorated
EPA Cycle-Based
CG Emission Standard*
May 1983 Data (derived from Table 1-4)**
Ford 4.9L-1V
Ford 6.1L-2V
Ford 6.1L-4V
Ford 7.0L-4V
Ford 7.5L-4V
2.26
2.09
2.91
2.65
2.46
40.4
38.4
44.7
31.3
38.6
August 1982 Data (derived from Table 1-2)
GM 292
GM 350-2V
GM 350-4V
GM 366
GM 4 54
2.78
2.08
2.57
1.13
1.43
36,
40,
39,
27,
32.7
April 1982 Data (derived from Table 1-1)
Ford 4.9L
Ford 6.1L
Ford 7.0/7.5L
GM 292
GM 350
GM 366
GM 454
2.19
2.97
2.83
2.18***
2.31***
1.80***
1.30***
34.0
41.2
35.4
26.5***
36.4***
30.1***
31.1***
* EPA cycle-based, g/BHP-hr, assumes deterioration and
includes SEA requirements.
** Calculated by first correcting MVMA to EPA low mileage
emissions, then adjusting for SEA and deterioration.
*** These emission levels were claimed by GM in May of 1983 to
have promoted unacceptable engine durability and
performance, and were reported to EPA before discovery of
such problems.
-15-
-------
Table 1-6
Compliance Ability of Ford's
Product Line at Various Levels of EPA
Cycle-Based Emission Standards (May 1983 data)
EPA Cycle-Based
Emission Standards Number of Engine Models
HC CO in Compliance by May 1983*
1.3 35.0 0 out of 5
2.5 35.0 1 out of 5
2.5 40.0 3 out of 5
2.6 40.0 4 out of 5
2.9 45.0 5 out of 5
Assumes deterioration; includes 1986 SEA requirements.
-16-
-------
Figure 1-1
CO Emissions (EPA Cycle-Based) from Ford HDG Engines
(Note: Engines of similar displacement over time are
considered to be in the same "familv.")
i
Measured
BSCO
Emissions
(g/BHP-hr)
250
1969 Baseline Average
Y (Uncontrolled]^
Ford Engine "Fami"
~ = 4 . 9 L
A = 6 . 1 L
= 6 . 6 L
X = 7.0/7.5L
~ = Mean of al
"Families"
I 363
I 373
1973 Baseline Average
1979 Baseline Average
_ 1 9 85 Non-Catalyst Standard
Statutory Standard
j
I 373 I3BE 13B3
Model Year
-------
Figure 1-2
i
i—
CO
I
Measured
BSiiC
Erai ssions
(g/BUP-hr)
IIC Emissions (EPA Cycle-Based) from Ford HDG Engines
(Note: Engines of simialr displacement over time are
considered to be in the sa^oe "family.")
Ford Engine "Family"
Model Year
-------
1-4 and 1-5). Table 1-6 summarizes the ability of Ford's
engines, in May of 1983, to comply with a variety of potential
emission standards.
Tables 1-5 and 1-6 indicate that only two of Ford's five
engine models would today exceed an HC standard (EPA cycie
based) of 2.5 g/BHP-hr (and one marginally so), whereas four
out of five would exceed the 35.0 g/BHP-hr CO standard. An
important observation to make is the fact that very little has
changed in Ford's compliance ability between April 1982 and May
1983 (see Table 1-5). Ford made the statement in their
comments of May 6, 1983 that significant progress had been made
relative to Ford's reported status of April 1982. However,
EPA's present analysis indicates that much of the reported
progress was illusory, arising primarily from a change in test
cycles. (The April 1982 data were EPA cycle-based; the May
1983 data were MVMA cycle-based.)
Based solely upon Ford's May 1983 data, the current
critical range of feasibility apparently lies between 2.5-2.6
g/BHP-hr HC and 35.0-40.0 g/BHP-hr CO (EPA cycle based).
Assuming for the moment that there' is little practical
difference between a 2.5 and 2.6 HC standard, then relaxing the
proposed CO standard of 35.0 to 40.0 would allow all but one of
Ford's engines to comply with 1985 requirements in May of 1983,
taking into account deterioration and 1986 SEA requirements for
1985 . Reviewing GM's latest data (presented in Table 1-5),
this relaxation would also allow all but one of GM's engines to
comply, even if no improvement in emission levels have been
made since August of 1982.
1985 Standards Derivation
EPA's draft feasibility analysis[8] attempted to evaluate
the detailed calibrations, hardware, and associated emission
levels of Ford's April 1982 engines. using these facts as
starting points, EPA surmised how additional emission
reductions could be made for the few engines which the April
1982 data indicated actually required further work to meet 2.5
g/BHP-hr HC and 35.0 g/BHP-hr CO (see Tables 1-1 and 1-5).
Emission reductions were predicted based upon established
principles of emissions engineering, whereby given changes of
calibrations produce predictable trends in emissions.
Both Ford and GM took issue with EPA's analysis.
Criticism of EPA abounded, but no approaches for further
emission control were recommended as having promise. GM
harshly criticized EPA for drawing conclusions from a limited
data base, despite its refusal to provide EPA with specific
-19-
-------
calibration information. Ford was less critical than GM,
discussing several of EPA's evaluations and indicating where
Ford thought they were incorrect or why they would prove
ineffective. Despite Ford's presentation on the emissions
status of their product line, EPA still has only generalized
information as to the specific emission control techniques
attempted, which were discarded, and which remain available
(with and without trade-offs). Furthermore, without specific
engine calibration information it is difficult for the Agency
to identify which levels of emission standards represent the
most stringent standards possible without unreasonable impacts
on cost or fuel economy, as EPA is required by law to
promulgate. Commenters are correct in maintaining that EPA is
not close enough to engine development efforts to anticipate
engine specific problems which arise as each control technique
is applied to each engine. For this reason alone, the
manufacturers are responsible for providing EPA with the
detailed, unbiased information it needs to make reasoned
decisions.
Without such information, EPA can only review the best
available data, and make a judgment as to what represents
reasonable interim standards, given the state of current engine
development and given the remaining leadtime until 1985.
As shown in Table 1-6, only one of Ford's engines would
significantly exceed an HC standard of 2.5 g/BHP-hr. The
remaining engine, the 6.1L-4V, would require a 16 percent
reduction in low-mileage emissions to meet the 2.5 standard
(see Table 1-7). If Ford takes advantage of the certification
flexibility provided by EPA for 1985 (SEA requirements do not
apply), the 6.1L-4V would already meet a 2.5 standard if only
deterioration is included with the low-mileage emissions to
determine compliance, in short, one extra year of leadtime is
available, if necessary, for attaining what appears to be a
modest reduction in HC emissions. EPA will not speculate as to
which technologies will be used to achieve the reduction,
although in the worst case ignition timing retard is
available. More importantly, EPA cannot allow the
technological laggard to set the pace for standard setting; to
do so surrenders the gains already achieved with the majority
of engines, and does little to motivate a manufacturer to lower
emissions from its engines.
To some extent, the same argument holds true in
determining a feasible CO standard. However, as shown in Table
1-6, the majority of Ford's product line will require
additional work to achieve the 35.0 g/BHP-hr CO standard. Some
reduction in low-mileage emissions will be necessary for four
out of five engines, including a substantial reduction (26
percent) for the 6.1L-4V family (see Table 1-7). Given the
-20-
-------
Table 1-7
Percentage Reductions in low Mileage Target (LMT)
Emissions Required to Comply with Emission standards of
2.5 g/BHP-hr HC and 35.0 g/BHP-hr CO (EPA cycle based)
Required HC Required CO
Engine Family LMT Reductions (%) LMT Reductions (%)
Ford 4.9L-IV*
0
16
Ford 6.1L-2V*
0
10
Ford 6.1-4V*
16
26
Ford 7.0L-4V*
5
0
Ford 7.5L-4V*
0
11
GM 292**
9
2
GM 350-2V**
0
13
GM 35C-4V**
1
10
GM 366*+
0
0
GM 454**
0
0
* May, 1983 data.
** August, 1982 data.
-21-
-------
industry's claims of decreased engine durability with further
enleaned fuel mixtures at WOT, and given the remaining
leadtime, there appears to be some risk in a 35.0 g/BHP-hr CO
standard. if, on the other hand, the emissions from Ford's
worst emitter (again the 6.1L-4V) were to be reduced to the
level of the remainder of Ford's fleet, an emission standard of
40.0 g/BHP-hr would be required. Again, EPA rejects the idea
that the technological laggard set the pace of emissions
reduction; therefore, a standard greater than 40.0 g/BHP-hr
would be unjustified.
Selecting a CO standard between 35.0 and 40.0 g/BHP-hr
then becomes an exercise in evaluating trade-offs.
Promulgation of 35.0 g/BHP-hr, or any standard which requires
the majority of the product line to achieve further reductions,
will increase the risk of durability problems, and at the same
time direct development efforts away from the 1987 standards.
Requiring the highest emitters to achieve further reductions,
however, is both appropriate and necessary to retain reductions
already achieved. From Table 1-5, EPA notes that the majority
of Ford's engines (according to the latest data) lie at the
high end of the 35.0-40.0 g/BHP-hr range.
EPA does not believe that compliance with a 35.0 g/BHP-hr
CO standard is infeasible. However, some additional
development work would be necessary for four of Ford's five
families, and significant work for one family. Given the fact
that some development work is still required to meet both the
2.5 g/BHP-hr HC standard and a 40.0 g/BHP-hr CO standard, given
EPA's desire not to preempt significant development efforts
from the 1987 model year, given the fact that many of the
engines for which data is currently available exhibit CO
emissions closer to 40.0 g/BHP-hr than 35.0, and given the risk
to engine durability entailed in meeting a 35.0 g/BHP-hr CO
standard within short leadtimes, EPA believes that 40.0
g/BHP-hr would be a reasonable non-catalyst CO standard for
1985.
EPA's evaluation of the latest GM data leads it to the
same conclusions. as can be seen in Table 1-5, GM's August
1982 data indicates that only a single engine would
significantly exceed standards of 2.5/40.0, and it would only
exceed the 2.5 HC standard. GM has repeatedly expressed
concern about the durability implications of stringent
non-catalyst CO standards, as noted in Table 1-7, some of GM's
engines still require reductions in low-mileage CO emissions to
meet a 35.0 g/BHP-hr standard. However, the lack of specific
calibration information for GM's engine has made EPA's review
of the reasonableness of GM's claims difficult, at best. (For
example, EPA would not consider durability data taken on
engines with WOT A/F calibrations leaner than stoichiometry to
be at all representative; such calibrations would be
-22-
-------
unnecessary for compliance and understandably severe on
durability.) EPA notes that GM's criticism of EPA's earlier
feasibility analysis only addressed GM's concern with complying
with 1.3/35.0 standards. The latest GM and Ford data indicate
that low-mileage compliance with standards of 2.5/40.0 g/BHP-hr
represents no problem whatsoever for almost all engine
families; the feasibility issue essentially breaks down to the
level of emission standards which would not degrade engine
durability or performance. EPA believes that relaxation of the
proposed 1.3 HC standard to 2.5 will preclude the need for
substantial ignition timing retard, both preserving fuel
economy and precluding increased exhaust temperatures. EPA
also believes that relaxation of the proposed 35.0 CO standard
to 40.0 will also preclude the need for a/F calibrations lean
enough to promote excessively high temperatures and durability
problems. EPA bases these judgments on the current performance
of Ford's product line, upon Ford's claims that these emission
levels will not impair engine durability, upon GM's own test
data, and upon the lack of GM's comments and data to the
contrary for engines designed to meet emission standards at
these levels.
Conclusion
Revised gaseous emission standards of 2.5 g/BHP-hr HC and
40.0 g/BHP-hr CO (or 1.9 g/BHP-hr HC and 37.1 g/BHP-hr CO based
upon the MVMA cycle) are feasible without catalysts, will not
degrade engine performance or durability, and therefore should
be promulgated for the 1985 model year.
B. POST-1985 EMISSION STANDARDS
Summary of Comments/Synopsis of Events
Soon after the decision was made to propose non-catalyst
standards for the 1985 model year, EPA began evaluating when
further progress towards the statutory standards would be
appropriate for gasoline engines. it is generally accepted
that compliance with the statutory 1.3 HC/15.5 CO standards
will require oxidation catalysts. (Diesel engines easily
comply with the statutory HC and CO standards.)
EPA has never altered its conclusions of January 21,
1980[1,9] that catalysts are ultimately feasible for use on
HDGTs. The i justification for deferring catalyst-based
standards beyond 1984 was based principally upon economic
grounds and leadtime concerns, not technical feasibility.
on March 16, 1983, EPA distributed a staff paper[7] for
public comment, and subsequently held a Public Workshop on
-23-
-------
April 6, 1983. The staff paper presented options for the
long-term control of HC and CO emissions from heavy-duty
trucks. The major provision of the recommended option was that
HDGTs would be split along traditional class lines. vehicles
up to 14,000 lbs. gross vehicle weight (GVW) would be required
to meet statutory standards (and thus have catalysts); all
heavier gasoline vehicle engines would continue to meet
non-catalyst standards. This approach attempted to capitalize
on the transferability of light-duty truck (LDT) catalyst
technology to the largest fraction of HDGTs (the lighter
classes), while acknowledging the decreasing number of heavier
HDGTs on which catalyst application would be most expensive (on
account of the need to design increased survivability into
catalyst systems used in the more extreme heavier truck
environment). in short, emission reductions were hoped to be
achieved in the most cost-effective fashion. The suggested
implementation date for this strategy was the 1987-88
timeframe. Public comments on the staff paper were solicited
and accepted up until May 6, 1983.
prior to the May 6 close of comments, GM advanced an
alternative approach at an April 13, 1983 meeting with EPA
staff.[5] GM proposed that most* HDGTs under 10,000 lbs. GVW
("light heavy-duty vehicles") be required to meet emission
standards similar to those required for LDTs, and be certified
on the light-duty chassis dynamometer test procedure. Vehicles
above 10,000 lbs. would continue to have their engines
certified on EPA's heavy-duty engine test at non-catalyst
emission levels. GM proposed that the scenario take effect in
1987 .
Public comments received by May 6, 1983[5] addressed both
the EPA and GM scenarios and are summarized by commenter below.
Chrysler
Chrysler cannot support the GM proposal, because of the
proposed more stringent standards for LDTs below 6,000 lbs. GVW
and proposed relaxation for LDTs between 8,500 and 10,000 lbs.
GVW. Chrysler also opposes the creation of the light
heavy-duty class, arguing it would require an additional test
fleet for durability testing, thereby increasing costs.
Chrysler also claimed that EPA's engine dynamometer test
is not representative of vehicles less than 10,000 lbs. GVW.
Chrysler implied that another test would be better, but did not
specify any particular test.
Some exemptions would be allowed on the basis of larger
frontal area, etc.
-24-
-------
Ford
Ford argued that catalysts are not feasible for all
trucks, but may be feasible on trucks in. the 8,500-14,000 lb
category. Ford argued that catalyst standards should not be
implemented before 1988, because of production leadtimes and
the required 4-year leadtitae provisions of the Clean Air Act.
Ford suggested that the heavy-duty class be split at
10,000 lbs., primarily because there are not many class III
trucks,. Ford did not disagree, however, with EPA's concern
about potential migration, should HDTs be split at 10,000 lbs.
indeed, if EPA does split HDTs at 14,000 lbs., Ford recommended
specific vehicle types for exemption. These vehicles are those
which see the most severe operation, and thus, would be those
vehicles most difficult to equip with durable catalysts. Ford
also agreed with GM that the LDT chassis test procedure would
be appropriate for trucks under 10,000 lbs. GVW. Ford urged
EPA to consider this testing alternative seriously.
With respect to catalyst feasibility and the feasibility
of the 15.5 g/BHP-hr CO standard, Ford argued that temperatures
above 1,600°F will cause thermal degradation of the catalyst.
Catalyst protection systems are possible, but an
overtemperature protection system of air injection cutoff at
full load also cuts off CO control at its most significant
mode. This trade-off between catalyst durability and CO
control has not been characterized. Focd did claim, however,
that their experience with LDT truck catalyst technology will
be applicable to the 8,500-14,000 lb vehicle classes.
General Motors
General Motors argued that EPA's split-class approach was
flawed. Specifically, EPA's approach does not make compliance
any different for lighter HDTs because they would still be
certified on the HDE test. GM argued that the test procedure
itself will determine which technology is applied for emission
control. in fact, much more than minor modifications to LDT
systems would be required for usage on the heavy-duty test. GM
argued that catalyst-equipped HDGEs will exhibit unacceptable
durability and performance if certified on the transient engine
test procedure. GM claimed that they were unable, based upon
the lack of data, to define regulatory requirements based upon
the engine dynamometer test procedure.
General Motors also took issue with EPA's rationale for
splitting the classes. GM disagreed with EPA's conclusion that
LDTs and lighter HDTs were not significantly different;. GM
argued that EPA has not proved that they are sufficiently
similar to permit "easy" transfer of LDT control technology.
-25-
-------
General Motors did agree that the heavier the total
vehicle weight, the higher the catalyst temperatures were
likely.to be over the road. GM did not address the feasibility
of catalyst protection systems.
Manufacturers of Emission Controls Association
The Manufacturers of Emission Controls Association (MECA)
stated that EPA's split-class approach better balanced the
needs and costs of controlling emissions from HDGTs. MECA
further stated that if the operating environments of Classes
IIB and III trucks are "...not significantly different both in
'terms of emission levels and thermal exposure from that
experienced with vehicles currently equipped with catalysts,
then it is expected that conventional light-duty truck catalyst
technology could be applied with relatively minor modifications
to trucks in those classes."
MECA also stated that several of its member companies are
already working to develop catalysts for the Classes IIB and
III trucks, and also to develop catalyst components that will
withstand higher temperatures.
With respect to leadtimes, if LDT catalyst technology is
readily transferable, MECA claims that adequate quantities of
catalysts "...could be produced well within the timeframe
needed to supply 1987 model year trucks." "If more heat
resistant systems are needed for certain Class IIB and III
vehicles, some additional development time will be necessary."
Natural Resources Defense Council
The Natural Resources Defense Council (NRDC) took strong
exception to EPA's performance on the regulation of HC and CO
emissions from HDTs. NRDC stated that EPA's split-class
approach should mandate the entire 90 percent reduction in HC
and CO emissions for the lighter class by 1985, instead of
1987-88 as EPA's staff paper suggested. NRDC supported the
provision of a 1-year "safety valve" exemption for vehicles
subjected to more severe operating conditions, _if a need for
such could be publicly demonstrated. NRDC also recommended
that EPA seriously consider extending the lighter class upper
weight limit from 14,000 to 20,000 lbs. GVW to prevent vehicle
migration to higher weight classes.
NRDC also argued that the heavier classes should not be
given a permanent exemption from the 90 percent reduction
standards, even if such an exemption were technically justified
for 1985 or 1986. NRDC claimed that a permanent exemption is
not only detrimental to air quality, but also beyond EPA's
legal authority.
-26-
-------
Analysis of Comments
There are four basic questions concerning the issue of
catalyst feasibility for HDEs: 1) can the catalyst-based
standards be met at low mileage, 2) what type of catalyst
system and hardware are needed to allow compliance, 3) what
type of overtemperature protection is necessary for a catalyst
operating at HDE conditions, and 4) how much leadtime is
required for the development and production of such systems?
These questions will be addressed in the following analysis,
along with public comments to EPA's March 16, 1983 staff paper
wherein EPA originally proposed the "split-class" approach.
Low-Mileage Feasibility of Catalyst-Based Standards
EPA's decision to defer catalyst-based standards beyond
1984 was not a technical one, but based primarily upon economic
grounds and leadtime concerns. EPA concluded on January 21,
1980[1] that catalysts are feasible for use on HDGEs, and this
analysis will not reiterate the detailed findings of that
rulemaking. The associated Summary and Analysis of Comments
document, published in December 1979,[93, discussed a limited
test program which had been conducted by EPA during which the
statutory standards had been achieved at low mileage on two
test engines using catalysts. The conclusion of feasibility
was, therefore, supported by actual testing conducted by EPA.
Since that time, EPA has collected data from three
additional catalyst-equipped heavy-duty gasoline engines. (All
five catalyst-equipped heavy-duty gasoline engines and their
weighted cold/hot start transient test emissions are listed in
Table 1-8.) In this more recent testing, EPA retrofited an IHC
404 CID engine with two three-way catalysts and two oxidation
catalysts. A Ford 1985 prototype 7.5L HDE equipped with
oxidation catalysts was also tested at the EPA facility.
Finally, a GM 350-CID engine, with both a three-way and an
oxidation catalyst, was tested at Southwest Research
Institute. In addition, EPA notes that GM has tested a 1985
prototype 350-CID engine equipped with oxidation catalysts, and
submitted that data to the Agency as part of the cooperative
effort to determine the correlation between EPA and MVMA test
cycles. All engines yielded emissions well below the 15.5
g/BHP-hr CO and 1.3 g/BHP-hr HC standards (see Table 1-8).
Thus, laboratory testing of heavy-duty gasoline engines
equipped with catalyst systems has established that these
engines can comply with the statutory standards at low mileage.
Likely Emission Control Stategies
EPA believes that LDTs and most lighter HDTs are not
subjected to significantly different operational environments,
-27-
-------
and that existing LDT catalyst technologies and strategies can
be modified for use by HDEs (see below) . The only significant
difference affecting compliance technology between LDTs and the
HDEs for which catalyst standards will apply is, as properly
noted by GM, the larger engine exhaust mass flow induced by the
heavy-duty transient engine test. This difference should only
be manifest in the CO emissions. Necessary modifications to
LDT control technology to permit compliance with the CO
standard on the heavy-duty test include both changes to the air
injection system and to the catalyst system itself.
Adding air to the catalyst ensures that there is
sufficient oxygen to allow the oxidation of CO emissions. Air
injection is most important, and potentially problematic, at
full-power modes when the engine is operating under relatively
richer mixtures. Most of the CO emissions generated on the
transient test arise during these modes, and therefore
high-power CO emission control is critical. Given the already
high exhaust temperature at full power, substantial oxidation
of the relatively abundant concentrations of CO could
potentially raise catalyst temperatures to unacceptable
levels. It has been argued by manufacturers that this fact may
be the most difficult development problem to solve: any
emission control system with sufficient air injection to permit
CO compliance on the heavy-duty test, if that calibration is
carried through to the in-use vehicle operating for sustained
periods at full power, will create catalyst overtemperature
problems in-use. In turn, catalyst durability could be
severely impaired..
EPA's testing of the Ford 7.5L (see Table 1-8) examined
the relationship between CO emissions and the injection of air
to the catalyst. (Evaluations of catalyst temperatures were
also made, and are discussed below with respect to catalyst
protection systems.) Solenoid valves were installed in the
engine's air injection system so that complete control of when
air was being injected into the catalyst was achieved. Testing
was conducted such that different amounts of air were added to
the catalyst at wide open throttle (WOT) . (WOT was defined as
the condition when the manifold vacuum was equal to or less
than 2 inches Hg, the point at which power enrichment was
observed to substantially begin.) Figure 1-3 shows the
observed trade-off between hot start CO emissions and the
diverted air; Table 1-9 lists the hot start emission data.
Even though WOT represents only a small amount of the total
test time (4.5 percent), the CO emissions attributable to this
fraction of operation are relatively high. By allowing more
air to reach the catalyst, (i.e., air was injected a greater
percentage of the time the engine was at WOT) , there was a
dramatic reduction in CO emissions. In fact, with full-time
air injection, CO emissions were virtually eliminated.. EPA's
-2 8-
-------
Table 1-8
Catalyst Feasibility Testing
on the EPA Transient Test Cycle
Engine
1979 GM 292
1978 Production
IHC 404
Test
Facility
EPA
EPA
EPA
Weighted Emissions
(q/BHP-hr)*
HC
.58
.28
.32
CO
12.25
8.98
3.74
Comments
Dual 50 g/ft3 catalysts,
2:1 ratio of platinum
palladium
Dual air pumps, 4-113 in3
oxidation catalysts.
Extrapolated emissions
with fourfold increase in
air injection, four
oxidation catalysts.
EPA .68
1975 GM 350 SwRI .39
1985 Prototype EPA .72
Ford 7.5L
Prototype GM 350 GM .53
3.6 Dual-bed system and EGR,
closed loop feedback
carburetor, 2-151 in3
TWC and 2-173 in3
oxidation catalysts.
5.6 COC/TWC pelletized
catalysts, closed-loop
feedback carburetor.
7.22 4-150 in3 COC LDT
catalysts in parallel,
dual air pumps.
5.62 2-260 in3 COC new
pelletized catalysts.
EPA cycle based.
-29-
-------
Table 1-9
Hot Start CO Emissions As A Function Of
Air Diversion at WOT
CO Hot Start
Air Dumping Emissions
% Time at WOT % Time of Complete Transient Test (g/BHP-hr)
0 0 .79
30.4 1.4 3.40
47.4 2.2 5.88
100.0 4.5 11.40
Note: Maximum catalyst bed temperature did not vary
significantly between any of these tests, and never exceeded
1,600°F.
-30-
-------
Figure 1-3
Hot-Start CO Emissions As a Function of Diverted Air From the Catalyst
I2.0r
10.0-
i
(jJ
B . 0
Hot-Start
BSCO
g/BHP-hr
E . 0
H . 0
2 . 0
0 . 0
0
20 H0 E0 00
% of Time Air is Diverted at WOT
00
-------
testing also indicated that as much as 60 percent of the air
can be diverted from the catalyst at WOT while still attaining
a target CO emission level of 7.1 g/BHP-hr (EPA cycle based).*
This ability to "by-pass" air, while still attaining required
emission levels, has important implications for catalyst
protection systems, as discussed later in this analysis. In
summary, however, EPA sees no obstacle which would prevent
modification of existing LDT or HDE air pump systems for usage
with HD cf-atalysts.
With respect to catalyst design, the two most important
factors with respect to catalyst application to HDGEs are the
noble metal loading and catalyst size. Location and geometry
of the, catalyst also affect its efficiency, as does substrate
and noble metal material and density. Due to the higher mass
flow of exhaust observed at full power on the HD test cycle,
some changes may need to be made to existing LDT catalyst
systems to maintain adequate CO oxidation efficiencies at these
modes. In the worst case, larger, more heavily loaded
catalysts may be needed. In other cases, changes to the
exhaust and catalyst system geometry to increase gas residence
time and eliminate "break through" at maximum exhaust flow will
be necessary.
To evaluate how changing the geometry of the catalyst
system affects its efficiency, EPA recently tested a Ford 7.5L
engine with two catalyst configurations: 1) four catalysts in
parallel, and 2) two sets of two catalysts in series. The
brake specific carbon monoxide emissions for the parallel
version were 59 percent lower than the emissions for the series
version. By splitting the exhaust four ways instead of two,
the exhaust flow velocity decreased, thereby increasing the
residence time of the gas in the catalyst. This presumably
allowed more time for the oxidation reaction to occur
(eliminating "breakthrough"), and thus yielded lower overall CO
emissions.
In summary, industry has several design options to
maintain the required catalyst efficiency at full-power modes
and to ensure that catalyst-equipped HDGEs meet the required CO
target level. (EPA's earlier analysis[5] concluded that HC
emissions will be reduced as a matter of course, and will be
achieved primarily by assuring sufficiently prompt catalyst
light-off on the cold start; EPA's recent data substantiated
these earlier conclusions.) EPA believes that these
Target CO Emission Level = 1/DF X 1/AQL X Emission
Standard, where AQL = 1.283 (from above), and the
multiplicative DP = 1.7 (from Reference 9).
-32-
-------
modifications to air injection and catalyst systems are
possible. Moreover, EPA believes that they represent the
transfer of known emission control technology to different
applications, and do not represent fundamental technological
unknowns.
Catalyst Survivability
High operating temperatures create problems for
maintaining catalyst efficiency over time:
"The primary material used currently to support the noble
metal catalyst in automotive converters is gamma alumina,
in the form of either pellets or a washcoat on cordierite
monoliths. At elevated temperatures, a phase change to
alpha alumina begins which is accompanied by a reduction
in the structural strength and surface area of the
material. Active catalyst sites tend to diffuse and
agglomerate as well as become inaccessible due to the loss
of porosity; this process effectively reduces the number
of sites available for catalysis and hence lowers the
efficiency of conversion. Finally, the magnitude of the
physical changes which occur in the alumina above the safe
operating temperatures is a function of temperature, time
of exposure, and the presence of certain ions which
stabilize the gamma lattice."[9]
Due to the time and condition dependency of catalysts,
there is ho exact temperature above which a catalyst will
suddenly fail. It is generally accepted that above 1,800°F, a
catalyst will suffer serious damage. Operation between 1,600°F
and 1,800°F is possible, but thermal degradation increases with
time spent within that temperature range.
While none of the manufacturers in their comments disputed
that the emissions from a heavy-duty engine can be reduced
below the standards, they argued that catalysts are not
feasible for all trucks. They argued that heavier trucks cause
special problems for catalysts, such as the continually higher
temperatures and greater mass flow of emissions from vehicles
which spend a large percentage of operational time at full or
very high power. They contended that these conditions
seriously threaten the durability of currently available
catalysts. EPA has since recognized the manufacturers'
concerns of increased difficulty and cost of protecting
catalysts under these circumstances, and thus EPA proposed the
split-class approach as a solution.[7] In essence, the
split-class approach allows more time for application of
catalyst technology to worst case operational applications.
The heaviest HDGVs (above 14,000 lbs. GVW), and also a limited
-33-
-------
number of lighter vehicles intended for the most severe
applications, would continue for now to meet non-catalyst
standards of 2.5/40.0. For these reasons, EPA believes that
catalysts would not see extended service in severe operating
temperatures if the "split-class" approach were adopted. This
"split-class" approach is therefore the most important factor
in assuring catalyst durability, presuming the use of existing
catalyst materials and substrates.
Recent testing at EPA also included examination of the
catalyst temperatures under various types of operation. The
catalyst bed temperature of the Ford 7.5L heavy-duty engine
never rose above 1,600°F during the transient test cycle for
the parallel catalyst version; the version with series catalyst
had maximum bed temperatures approximately 50°F higher in the
catalysts closest to the exhaust manifold. A Ford 302 LDT
engine that was tested on the hde transient cycle by EPA had a
maximum catalyst temperature of 1,640°F. Catalyst temperatures
were also observed under conditions more severe than the
transient test. Table 1-10 lists the maximum catalyst
temperatures during WOT engine maps for two engines tested by
EPA. Mapping conditions are extreme, and as expected, the
catalyst temperatures are higher. Indeed, catalyst
temperatures typically reach a maximum after the engine
operates for sustained periods of time at WOT. Protection of
the catalyst from too much oxidation of CO would be necessary
at these conditions, if such conditions were expected to
routinely occur in-use. (Note that these conditions are not
seen on the transient test.) Again, however, EPA believes that
the "split-class" approach would virtually eliminate sustained
full-load operation from the vehicles required to use catalysts.
Aside from the elimination of the applications most
detrimental to catalyst survivability, there are other
strategies available to protect catalysts on HDGEs.
With increased air injection, the catalyst bed temperature
increases as more oxidation occurs. One obvious means of
protecting the catalyst at WOT is to divert the injected air
from the catalyst mechanically, thus precluding increased
oxidation. However, air injection cutoff at full load also may
cut off CO control at the most significant moment. This
creates an inherent trade-off between catalyst temperature and
CO emissions.
This trade-off, however, is not significant enough to
preclude compliance with the statutory CO standards. EPA bases
this judgment on the test data discussed above, with which it
was demonstrated that, for at least one engine, air injection
could be completely diverted for up to 60 percent of the time
spent at WOT on the transient test, and sufficiently low CO
emission levels could still be maintained (see Figure 1-3 and
-34-
-------
Table 1-10
Maximum Catalyst Temperature During WOT Engine Map
Engine
1980 GM 305-CID LDVE
equipped with air/
oxidation catalyst/
EGR
»
II
1985 Ford 7.5L HDE,
equipped with air/
oxidation catalyst
Catalyst Configuration Temperature, °F
One catalyst, stock 1,734
location
4.5 feet downstream 1,660
6.5 feet downstream 1,402
behind muffler
4 catalysts in parallel 1,650
6 feet downstream
-35-
-------
Table 1-9) . This indicates to EPA that some form of the WOT
air injection cutoff presently found on LDTs could be applied
to HDGEs. In short, additional catalyst protection can be
provided while still maintaining acceptable emissions levels.
Over-temperature protection systems, particularly for
non-catalyst engines, have been discussed at some lenth by GM
in earlier submissions. GM discussed several concepts,
including one which would completely cut off full-power air
injection after a certain amount of time (e.g., one minute) at
sustained full power. GM noted that maximum temperatures
require a certain amount of time to build up, and that such a
system would protect the engine, and at the same time would be
required very little in typical urban driving. GM's apparent
concern, however, is that EPA may rule such a system to be a
"defeat device" and forbid its use. EPA at this time cannot
specifically approve or disapprove any system described to the
Agency in a cursory or qualitative fashion; indeed, EPA's
"split-class" approach should eliminate the need for such a
system for now. However, past EPA policy with respect to the
determination of defeat devices does not necessarily preclude
the use of such a system. In general, EPA policy has been not
to classify a technology as a defeat device if it can be
demonstrated that such a device is essential for protecting the
integrity of the engine, the integrity of the emission control
system (e.g., catalysts), or the safety of the vehicle. In
short, provided that such demonstrations can be made (for
either catalyst or non-catalyst engines), additional
flexibility could be available for protection from excessive
temperatures, despite EPA's present belief that such protection
is not currently necessary.
An additional means of providing temperature protection
for the catalyst system has been discussed in earlier EPA
analyses[9]--the ability to relocate that catalysts further
downstream in the exhaust system. There are limitations to the
degree of relocation protection available, primarily because HC
emissions tend to increase dramatically as catalyst light-off
time is sufficiently increased. However, a limited amount of
temperature protection should certainly be available through
relocation of the catalyst.
Finally, one additional measure providing major
flexibility for certification will be available to the
industry; EPA intends to retain the option of allowing a
manufacturer to certify any vehicle of 10,000 lbs. GVW or less
on the LDT chassis test procedure to LDT emission standards.
Whether or not this option is exercised will be based upon the
manufacturer's judgment of relative compliance costs; it is an
option, however, which remains available.
-36-
-------
in summary, EPA does not expect HDGE catalyst durability
to represent a major problem. The "split-class" approach
eliminates the most problematic applications. Several other
techniques of catalyst and engine temperature protection are
available. Air injection diversion at full load would be a
technical derivative of systems already in production for LDVs
and LDTs, and can be calibrated in such a way to both provide
protection and achieve required emission levels. Catalyst
relocation can also provide additional minor protection. Other
protection systems could be used, if EPA were convinced that
they were truly necessary for engine or catalyst survival.
Leadtime
This discussion focuses on the technical leadtime
necessary to allow compliance with the "split-class" approach;
legal issues regarding leadtime are specifically addressed in
the Preamble of this rulemaking.
An outline of the technical ability of the manufacturers
to comply with the statutory standards for HDGEs in Classes iib
and III applications is presented below. This general schedule
(Figure 1-4) assumes a significant, although certainly
attainable, compliance effort by the manufacturers, of course,
the specifics of the situation facing each manufacturer will
determine exactly how much time is necessary for each phase of
the effort and what sequence will be followed. The schedule in
Figure 1-4 and the discussion below are intended to illustrate
what needs to be known and what needs to be done; by allowing a
reasonable amount of time for each phase, the feasibility of
compliance is demonstrated.
The work can be viewed as phases of development,
dynamometer and vehicle assurance testing, dynamometer-based DF
determination, and certification. These phases are not
necessarily sequential; in fact, there is certain to be a
considerable amount of overlap. Assumptions that have been
made in developing this schedule are noted where appropriate.
There are a few decisions that will be made, at least on a
tentative basis, early in the development process. under the
split-class approach, where all worst case HDGEs (in terms of
difficulty of catalyst application) are certified to
non-catalyst standards, manufacturers are expected to divide
their HDGEs into families on the bases of displacement and
catalyst use. The families will be divided in such a way as to
minimize disruption to the manufacturer's product line and
inconvenience to the consumer. in addition, manufacturers will
avoid situations that could result in competitive disparities;
for example, a manufacturer would not want to use
catalyst-equipped engines in vehicles that are in direct
competition with similar vehicles without catalysts from
-37-
-------
Figure 1-4
General Leadtime Schedule - MY87 HDGE Standards
m
ro sr
00
00 oo
CO
r—i-
r—
a\
r—
r-.
i
r~
1
I
CO
H
H
in
00
1
in
i
m
oo
I
I
u>
CO
I
(Pre-FRM
JeyeiapjnentjL _
Accelerated dynamometer
(catalyst design tests)
(test to
failure)
Initial
Development
Production tooling
commitments —
A
V
<£>
30
oo
NO
CO
1
H
|
r-i
1
lO
r--.
1
>£>
00
I
' -1
'e>
(tooling)
Dynamocnete r-b a sed
DF- determinations
Production
Certifica :ion
Final design modifications (if necessarj
Vehicle assurance testing
iynamometer-based
sspgSTiipnrs
-------
a different manufacturer. In short, the manufacturers will
substantially determine the structure of their 1987 product
lines. The manufacturers will also make tentative judgments of
likely DFs and the impact of the SEA requirement (effective in
the 1986 model year) , and thereby estimate target emission
levels and the likely hardware and engine calibrations needed
to begin development.
The most important phase, which EPA expects to last until
October 1984, is for development work. It is clear from the
history of this action and from manufacturer comments that
preliminary development work has already been underway for some
time, that a significant portion of the necessary work (i.e.,
reduction of engine-out emission levels) will have been
performed in complying with the 1985 model year interim
standards, and that early catalyst testing has been in progress
since January 21, 1980. (For example, in comments submitted to
EPA in April of 1982, GM provided a lengthy submission covering
their heavy-duty catalyst development work to date.) EPA
believes that the most significant problem to be solved during
development is determining the catalyst configurations and
engine calibrations that will be needed in order to demonstrate
compliance on the HDGE transient test. Technically, this is a
relatively straightforward engineering problem of applying
known technology to new applications. As noted earlier in this
chapter, the same generic technology will be used, with
problems and engineering parameters similar to those
encountered in applying catalysts to LDTs.
It is assumed that accelerated dynamometer testing, a
fundamental part of engine and catalyst development, will occur
during the development phase. Limited dynamometer-based DF
assessments can also be conducted as part of the development
phase in order to provide preliminary DFs. Following this,
worst case durability assessments will be run to check for
catastrophic failures. Such failures will become apparent in
accelerated dynamometer testing, the last round of which is
estimated by EPA to extend three months beyond actual
development. (For mileage/service accumulation purposes, this
testing may proceed 24 hours a day under automatic control.
Thus the 1,500-hour half-life equivalent could be reached in as
little as 2.4 months, assuming operation for six days per
week. Note that GM's standard corporate durability test
generally runs about 200 hours.) Approximate DF determinations
based on dynamometer operation for the equivalent of the
full-life (3,300 hours) could therefore be completed
conservatively within eight months after the development phase
is concluded.
After the worst case durability assessments are completed,
EPA estimates that basic vehicle assurance testing could be
-39-
-------
done in about four months. This vehicle assurance testing will
allow comparison of in-vehicle and dynamometer-based DFs, as
well as assessment of the effectiveness of vehicle
modifications (heat shields, catalyst location and mounting),
system reaction to on-road phenomena (vibration), and overall
effect on performance characteristics (driveability). Four
months is adequate time to accumulate at least 25,000 miles of
in-vehicle use, assuming 10 hours per day on the road for 5
days per week at an average speed of 30 mph. Vehicles in this
program could be left in service; only the catalysts need to be
switched periodically for inspection and oxidation efficiency
testing on well-characterized engines. Assuming that the
in-vehicle testing conditions are appropriately planned, four
months should be more than enough time for identification of
any vehicle-related flaws.
EPA assumed three months beyond the work described above
for final design modifications to be implemented, if any are
found to be necessary. All of this could be completed by
September 1985, at which time production tooling commitments
could be made. At this point a full year would remain before
model year 1987 "Job 1" production must begin.
Although EPA considers the possibility to be remote, any
fundamental problems that may arise should be evident after the
completion of vehicle assurance testing and dynamometer-based
preliminary DF assessments. Existing regulatory provisions
would allow a manufacturer to petition EPA for relief in the
event a serious risk of non-compliance appeared likely at this
time. EPA does not believe, however, that such relief will be
necessary on the basis of all information available at this
writing.
The remaining 12 months before "Job 1" would be used for
final dynamometer-based full-life DF determinations, which
should take eight months or less, and certification. Under
procedures applicable for 1985 and later model years,
durability testing is not required to be a part of the formal
certification process. If further changes to calibrations or
hardware appear necessary, manufacturers would have the option
of foregoing the eight month durability assessment, and merely
use engineering judgment or use predetermined DFs for
certification. Certification should then begin no later than
May 1, 1986, and should take no longer than four months.
In summary, EPA estimates that compliance with the
"split-class" approch is feasible for the 1987 model year.
Having outlined a general schedule that demonstrates the
feasibility of compliance by model year 1987, EPA takes issue
with the technical leadtime estimates supplied by Ford in its
-40-
-------
comments. Ford considerably overstated the amount of time
necessary, and seemingly disregarded the considerable progress
that has already been made. Where Ford has included three
complete iterations of designing, building, and testing, EPA
believes that at most two iterations will be required,
particularly because of existing experience with ldt technology
and because all worst case applications are excluded under the
split-class approach. Ford also estimated the certification
process to last for three years, which EPA finds unreasonable
and unlikely.
on the other hand, the EPA leadtime estimates allow for
little slack time. Despite EPA's judgment that legal authority
exists for requiring compliance by model year 1986, the
elimination of 12 months from the time estimates discussed
above would preclude orderly development and make the risk of
non-compliance for 1986 unacceptably high. with respect to
1987, EPA again stresses that all truly worst case HDGE
applications, in terms of catalyst use, are excluded from the
statutory standards by the split-class approach. in addition,
catalyst-forcing emission standards for HDGEs were first
promulgated in 1979. The interim standards for 1985-86 were
never intended to defer catalyst standards permanently, but
merely to provide short-term economic relief. Thus the. Agency
is confident that implementation of this approach by 1987 poses
no insurmountable difficulties for the industry.
Conclusions
Statutory emission standards (1.1 g/BHP-hr HC and 14.4
g/BHP-hr CO based upon the MVMA cycle) for Classes IIB and III
HDGEs should be promulgated for the 1987 model year. All
heavier HDGEs should continue to meet non-catalyst standards,
as would the small number (5 percent of total Classes IIB and
III sales; see the "migration" issue, Section B.12, for further
information) of lighter vehicles allowed to certify to
non-catalyst standards on the basis of application.
-41-
-------
References
1. "Gaseous Emission Regulations for 1984 and Later
Model Year Heavy-Duty Engines," Federal Register 4136, Vol. 45,
No. 14, Monday, January 21, 1983.
2. "Control of Air Pollution from New Motor Vehicles
and New Motor Vehicle Engines; Revised Gaseous Emission
Regulations for 1984 and Later Model Year Light-Duty Trucks and
Heavy-Duty Engines," Federal Register 1642, Vol. 47, No. 8,
Wednesday, January 13, 1982.
3. "Revised Gaseous Emission Regulations for 1984 and
Later Model Year Heavy-Duty Engines and Light-Duty Trucks,
Draft Regulatory Analysis," U.S. EPA, OANR, OMSAPC, ECTD, SDSB,
Chapter II, September 1981.
4. Derived from comments submitted to EPA Public Docket
No. A-81-20.
5. Derived from comments submitted to EPA Public Docket
No. A-81-11.
6. Letter from Charles L. Gray, Jr. to Ford Motor Co.,
General Motors, Chrysler, and International Harvester, June 17,
1982, EPA Public Docket No. A-81-11.
7. "Issue Analysis - Final Heavy-Duty Engine HC and CO
Standards," EPA Staff Report, March 1983, EPA Public Docket No.
A-81-11.
8. Letter to commenters from Charles L. Gray, Jr., plus
attachment, EPA Public Docket No. A-81-11, April 12, 1983.
9. Summary and Analysis of Comments to the NPRM,
"Gaseous Emission Regulations for 1984 and Later Model Year
Heavy-Duty Engines and Light-Duty Trucks," EPA, OANR, OMS,
ECTD, SDSB, December 1979.
-42-
-------
2. Issue: Useful Life
Summary of the Issue
This issue addresses the useful-life provisions that will
apply to 1985 and later model year light-duty trucks (LDTs) and
heavy-duty engines (HDEs). On January 12, 1983, EPA proposed a
revised full-life useful-life approach and an alternative for
comment (48 FR 1472). This section of the Summary and Analysis
of Comments deals with the responses to the proposal and the
selection of the appropriate useful-life approach in response
to those comments.
Summary of the Comments
Introduction and Synopsis of Events
Useful life is the period, expressed in terms of time or
vehicle miles, over which in-use vehicles/engines are required
to demonstrate compliance with the applicable emission
standards and the period for which they are required to warrant
the emissions performance of their products. In 1979 and 1980
EPA promulgated regulations effective for 1984 and later model
years that contained revised useful-life periods for LDTs and
HDEs. Useful-life periods were changed from fixed intervals,
representing periods representing somewhat less than half the
service life of these vehicles/engines, to
manufacturer-determined periods representing the full average
period to engine retirement or rebuild.
EPA adopted these reguations over concern about the in-use
performance of HDEs. Half-life regulations provided no
incentive for manufacturers to be concerned about the long term
emissions durability of their engines, since they had no
liability for their performance past the half-life
certification period. This problem could be only partly dealt
with by establishment of lower emission standards. Lower
standards would lower overall average emissions, but would not
control departures from standards during the second half of a
vehicle's life due to what are often know as "gross emitters".
Gross emitters are those vehicles whose emissions are increased
severalfold above normal due to the failure of emission control
hardware. The in-use failure of emission control components
could completely eliminate the improvements gained by lower
standards. Indeed, indications are that as more advanced
technology comes into use for control of emissions, the effects
of in use failure becomes much more pronounced.
The goal, then, was to focus manufacturer efforts more
toward in-use performance and durability of their engines and
to insure that emission control systems were fully capable of
lasting as long as the average engine. Full-life useful life
provided that incentive, and gave EPA enforcement authority to
-43-
-------
deal with problems in the second half of a vehicle's life. It
helped insure that durability would not be sacrificed in an
effort to minimize costs. EPAs analysis of both the costs and
air quality impacts of full-life useful life indicated that it
was a beneficial and cost effective program.
Subsequent to that rulemaking, LDT and HDE manufacturers
raised a number of issues relating to the practical problems of
useful-life determinations and possible high costs of
implementing a full-life useful-life approach. As a result, in
April of 1981 EPA agreed to undertake a further study of the
useful-life issue as a part of the President's program to
provide regulatory relief to the automotive industry.
Comments received during the several comment periods and
public hearings held during the course of this study led to the
January 13 proposal. EPA offered two useful-life options in
the NPRM: 1) a modified full-life requirement designed to
address previously expressed concerns regarding full-life
implementation, and 2) an extended half-life proposal with
slightly more stringent emission standards to compensate as
well as possible for the reduced stringency of half life. A
formal durability testing program accompanied the half-life
proposal, whereas the full-life allowed manufacturers to design
their own programs. EPA's stated preference was for the
modified full-life option; the half-life plan was provided only
in the event of unforeseen problems with resolving full-life
implementation issues.
The majority of the manufacturers favored a half-life
useful-life definition; however, none found the EPA half-life
proposal with the adjusted standards and extended durability
testing to be acceptable. Although some manufacturers were
willing to accept a longer useful-life period than presently
exists none was willing to also accept the downward adjustment
in the emission standards. Rather, they advocated a half-life
useful life with no adjustment of the standards or durability
testing requirement. However, acceptability of the half-life
plan to EPA was fully contingent upon the adjusted standards to
account for the decrease in the compliance period and upon
extended durability testing requirements to increase the focus
on emission control performance at higher mileages. The Agency
felt that without those compensating qualifications, all of the
environmental benefits of full life would be lost and such a
change would effectively reduce the stringency of the
standards. Since no commenters supported the provisions of the
extended half-life approach as proposed by EPA, and since three
commenters expressed a preference for modified full life, it is
EPA's intention to retain the modified full-life approach.
Therefore, the half-life plan will not. be analyzed further, and
the remainder of the analysis will concern itself only with
comments pertinent to modified full life.
-44-
-------
Before turning to analysis of specific comments it is
important to reaffirm EPAs belief in the value of full-life
useful life. EPA continues to hold to its original
justification for this program as outlined above. Commenters
have argued that EPA has not conclusively demonstrated that the
in-use need exists to an extent which justifies taking action.
While it is true that there is not a large body of data to
demonstrate the need, EPA believes that the logic of the
situation, along with what data is available argues strongly
for the establishment of full-life useful life. Indeed, the
very vigor of much of the opposition to extending
manufacturers' responsibilities into the second half of the
useful life argues in favor of the need for this action. If,
as argued by manufacturers, the current durability of emission
control components is adequate, then there is little risk
involved in extending the useful life period. EPA believes
that full-life useful life is needed to insure durable
components and to provide an enforcement mechanism for in-use
problems.
The comments have been divided into major and minor issues
for convenience of analysis. Within the group of major issues,
five significant areas have been identified. These include:
1) legal objections to EPA's modified full-life approach, 2)
concerns related to the recall provisions, 3) the heavy-duty
diesel engine subclasses, 4) the assigned useful-life periods,
and 5) the air quality benefits associated with full life. In
response to the last issue, an update of the environmental
impact and cost effectiveness of useful life was undertaken
which forms a part of the Regulatory Support Document.
Briefly, since it will not be considered further here, this
analysis shows that the adoption of EPA's modified full-life
approach will produce up to a 1 percent improvement in air
quality for ozone and CO in the mid-late 1990s. The analysis
further shows that full life is very cost effective in
comparison with other emission control strategies, projecting
costs-effectiveness values of $206-484 per ton for HC and
&12-24 per ton for CO. Interested parties are referred to
Chapter 3 of the Support Document for further analysis in this
area. Discussions of the other four main issues and several
minor issues are presented below.
Legal Issues
Summary of the Comments
A large number of comments were received concerning EPA's
legal authority for implementing the modified full-life
approach. Comments fell in four major areas: 1) statutory
authority for the full-life concept, 2) authority to establish
different useful-life periods for purposes of certification,
warranty and recall, 3) authority to group LDTs under 6,000
-45-
-------
lbs. GVW with heavier LDTs and HDEs foe the purposes of useful
life# and 4) the appropriate period of recall liability for
in-use vehicles and engines. These are discussed below.
Many comments were received that reiterated (either
directly or by reference) claims made during the original
rulemaking that EPA does not have the statutory authority to
implement a full-life useful-life definition. A substantial
number of commenters addressed the issue of Congressional
intent with respect to a half-life versus a full-life
definition, and cited portions of the legislative history of
the Clean Air Act which they believed demonstrated that
Congress intended that the half-life concept be retained for
LDTs and HDEs, regardless of the actual language in the Act.
Second, comments were received on the modified full-life
proposal which argued that the Act limits EPA to a single
useful-life period for both certification under Section 202(a),
and the in-use programs contained in Section 207 (warranty and
recall). The commenters therefore concluded that EPA was
precluded from establishing a useful-life period for warranty
that was different from the useful-life period for
certification and recall liability.
Third, Volkswagen of America (VW) stated that EPA had no
statutory authority to create a separate LDT class for
useful-life purposes for LDTs under 6,000 lbs. GVW. VW argued
that the court decision which initially led to the creation of
the LDT class by EPA (International Harvester vs. Ruckelshaus,
478 F. 2d 615, D.C. Circuit, 1973) applies only to the level of
the emission standards, and does not extend to other regulatory
requirements. Based on this premise, VW took the position that
LDTs of less than 6,000 lbs. GVW may not be required to conform
to a period longer than the statutory period for light-duty
vehicles (LDVs) (i.e. 5 years/50,000 miles).
Fourth, several commenters expressed concern about the
s,cope of their liability during a recall action. They believed
t,hat the full-life recall provisions would force them to "fix"
a;ll the vehicles/engines in the recalled group regardless of
their age, mileage, or condition. The comments took the
position that their recall liability should end with the
assigned useful life and that they should not be responsible
for any vehicle engine which has been rebuilt, regardless of
mileage.
Analysis of Comments
The question of Congressional intent and EPA's statutory
authority to adopt a full-life useful-life definition for LDTs
and HDEs was also raised when the full-life concept was first
proposed for these vehicle/engine classes. During those
-46-
-------
rulemakings, EPA prepared two separate Summary and Analysis of
Comments documents on the full-life useful-life proposals, one
each for LDTs and HDEs, and these are herein incorporated by
reference.[1,2] These analyses concluded, as EPA still
concludes, that the Clean Air Act as amended in 1977 provides
the Administrator full authority to set the LDT and HDE useful
life at any period of time and/or mileage longer than 5
years/50,000 miles if it was determined to be appropriate.
Manufacturers based their findings of Congressional intent
for half life on the differences between the Senate version of
the Clean Air Act Amendments of 1977 (S.252) and the final
version that emerged from the conference committee and was
later enacted. In the Senate version, the useful life for a
"motorcycle or any other motor vehicle or motor vehicle engine
would be a period of use the Administrator shall determine."[3]
In the Amendments as they were enacted, however, "motor vehicle
and motor vehicle engine" were removed from this clause and
placed in a new clause which read that useful life for "any
other" (than light duty) "motor vehicle or motor vehicle engine
(other than motorcycles or motorcycle engines)" was to be a
period of 5 years/50,000 miles "unless the Administrator
determines that a period of use of greater duration or mileage
is appropriate."[4] From this change in language, and the past
use of the half-life concept for LDVs, LDTs, and HDEs, the
commenters inferred that Congress intended EPA to retain the
half-life concept for LDTs and HDEs.
First, it should be noted that EPA's authority to
establish longer useful-life periods for LDTs and HDEs was
established in 1970. The 1977 amendments did not address LDTs
and HDEs directly, but were concerned with the problems of
existing law created with respect to motorcycles. Thus, the
1977 amendments are not directly relevant to EPA's authority to
set useful-life periods for LDTs and HDEs.
Moreover, in EPA's view, setting "any other motor vehicle
or motor vehicle engine" apart from motorcycles and light-duty
vehicles/engines simply retained a minimum 5 year/50,000 mile
useful life for LDTs and HDEs and did not alter the
Administrator's specific authority to set a period longer than
5 years/50,000 miles if it was determined to be appropriate.
Congress was aware of the ongoing litigation between Harley
Davidson and EPA on the issue of motorcycle useful life, and
specifically provided statutory language to permit EPA to
establish a useful life other than 5 years/50,000 miles for
motorcycles in Section 202 (d) (3). [5] Had this change not been
made, and the Senate version retained, the 5 year/50,000 mile
minimum would have been lost for LDTs and HDEs. Congress
desired to keep that minimum, which led to the creation of
Section 202(d)(2), which also contains EPA's authority to set a
useful-life period longer than half-life. Therefore,
-47-
-------
reducing vehicle emission," but are emission-related, have to
be warranted for only 2 years/24,000 miles under Subsection
207(a) (2) . [6] Nevertheless, Section 202 (a)(1) refers to "a
period of use" for light-duty vehicle useful life; Congress
evidently believed varying useful-life periods could exist
notwithstanding the apparent reference to a single period.
Since Congress was clearly aware of the possibilities involved
and yet did not specifically prohibit the Administrator from
making similar determinations for LDTs and HDEs, EPA concludes
that authority exists under the general and specific authority
mentioned above to allow the establishment of reduced warranty
periods, and that the Administrator is not restricted to only
one useful-life period for certification, warranty, and recall
purposes.
It should be kept in mind that the reduced useful-life
period for warranty is an attempt to be responsive to
manufacturers' valid concerns with having to warrant LDTs and
HDEs for their full useful lives, while not sacrificing the air
quality and durability benefits of the earlier full-life
useful-life requirement. EPA could have promulgated more
stringent half-life standards with increased durability
requirements, but opted instead for an approach that at least
was favored by some manufacturers.
Finally, although EPA believes, for the reasons set forth
above, that it would have authority to establish a different
useful-life period for purposes of recall, that is not what the
Agency has done. As discussed below, manufacturers in a recall
will be required to repair non-conforming LDTs and HDEs
regardless of age or mileage at the time of repair. EPA, as
part of this rulemaking, has simply established a policy that
LDTs and HDEs will not be tested for purposes of recall if they
exceed 75 percent of their useful life. Indeed, even in the
established LDV recall program EPA typically tests cars that
are only two to three years old, notwithstanding a 5-year
useful-life requirement. The recall policy established today
for LDTs and HDEs is an attempt to be responsive to
manufacturers' concerns that wornout or otherwise
unrepresentative engines may inadvertently be selected for
recall testing.
Turning to the issue raised by VW, EPA cannot accept VW's
contention that the decision of the Court of Appeals in
International Harvester v. Ruckelshaus was applicable only to
compliance with emission standards. The decision of the Court
led EPA to initiate a rulemaking which ultimately established
the definition of a new LDT class and an entire set of emission
regulations for new 1975 and later model year light-duty
trucks. (85 CFR - Subpart C) (See 38 FR 21362, August 7,
1973) . Since that time LDVs and LDTs have shared common
requirements only when it was found to be technologically
-49-
-------
appropriate (e.g., test cycle). Beginning in 1979 the LDT
class wa!s expanded from 0-6,000 lbs. GVW to 0-8,500 lbs. GVW,
and in. subsequent regulatory actions EPA has used the general
authority of Section 202(a)(1) to group the lighter weight LDTs
with the heavier LDTs for purposes of complying with mandates
of other portions of Section 202. In 1977, when Congress added
language to the Act authorizing EPA to establish classes and
categories for setting standards for HDEs, it specifically
ratified EPA's approach for LDT regulations.[7]
Given that the Court of Appeals ordered EPA to remove
light-duty trucks from the light-duty vehicle class in 1973,
and that EPA has operated with a distinct set of LDT emission
regulation and standards since 1975, EPA sees no merit in VW's
argument. EPA believes that setting LDT useful life under
Section 202(d) (2) is consistent with the past practice of
establishing separate LDT provisions, and is a correct usage of
Section 202(d)(2) since LDTs are neither LDVs nor motorcycles.
Finally, EPA recognizes the manufacturers' comments on
recall liability. Current EPA policy is that all
non-conforming vehicles/engines in a recalled family must be
"repaired" regardless of their mileage, age, or condition at
the time of repair. Recall evaluation testing will not be
conducted past 75 percent of the assigned useful life; however,
if a defect is discovered during such testing, it must be
remedied for all non-conforming vehicles/engines in that family.
The Agency is now involved in litigation over this
requirement (General Motors v. EPA, No. 80-1868, D.C. Circuit,
1980) , so it is subject to possible revision based on the
outcome. Final EPA response to these concerns is therefore not
possible at this time.
Conclusion
The Act contains the necessary authority to establish the
certification, recall, and warranty provisions embodied in the
modified full-life useful-life approach. EPA has significantly
revised these provisions in a way that should alleviate the
manufacturers' most pressing concerns, while preserving the
benefits of a full-life useful life-approach.
Recall Provisions
Summary of the Comments
A number of manufacturers have anticipated problems with
the three-quarter-life recall provisions proposed by EPA as
part of the modified full-life approach. The Engine
Manufacturers Association (EMA) and several industry commenters
stated that limiting recall evaluation testing to 75 percent of
the assigned useful life would not fully address the problem of
-50-
-------
including rebuilt and wornout engines in a sample selected for
recall testing. The commenters claimed that due to the
variability found in actual engine service lives, it was quite
possible that a substantial percentage of engines would be
rebuilt before reaching the 75 percent of assigned useful-life
cut-off point. For example, EMA estimated that 22-36 percent
of heavy-duty diesel engines would have been rebuilt and an
additional unspecified percentage would be in need of rebuild.
Commenters believed that the difficulties in screening such
engines would add to the cost of recall testing and might lead
EPA to "cut corners" by basing a recall on too small a sample
or by including marginal engines in the test program.
The commenters also expressed several other concerns
related to the recall program. Mack Trucks, Inc. expressed
concern over the potential impact of the 40 percent Acceptable
Quality level (AQL) of EPA's Selective Enforcement Audit
program on the recall evaluation program, stating that as a
result of the 40 percent AQL, there is a near 40 percent chance
that an engine taken randomly for recall evaluation may have
been above the standard when it left the production line.
Mack Trucks also stated that laboratory-to-laboratory
variability must be considered in any recall evaluations, since
results of the EPA/EMA round-robin test program indicated that
up to a 25 percent variation existed between certain test
facilities.
Mack also requested that EPA provide a three model year
"grace period" from recall liability for newly introduced
engine lines. Mack was concerned that even the best
engineering practices may not allow them to predict their
in-use emissions deterioration accurately for these new engine
lines, and that their in-use engines may exceed the emission
standards as a result.
Some manufacturers wanted EPA to limit recall liability to
a select list of emission control components only, although the
American Trucking Association (ATA) doubted that EPA had the
authority to do so.
Analysis of the Comments
The problem of screening vehicles/engines for improper
maintenance, abuse, rebuild, wearout, etc., prior to inclusion
in a recall sample is not new. Such screening is now
successfully conducted in the LDV and LDT recall programs, and
EPA expects to use a similar approach under full life for LDTs
and in the recall program currently being developed for HDEs.
The manufacturers are given several opportunities for
participation in the recall program. Under the current program
manufacturers are given the opportunity to comment or otherwise
-51-
-------
respond to the Maintenance/Use Criteria questionaire which
serves as the first level of screening for prospective
vehicles. The second level of screening is a physical
inspection of those prospective vehicles/engines which pass the
first level of screening. At this point, the manufacturers are
invited to be present at the inspection and to provide input to
EPA as to why a given vehicle may or may not be representative
for recall evaluation testing. Should any disagreement arise,
the current recall program allows manufacturers a full
opportunity to challenge vehicle selection. And, of course,
the manufacturers are. involved in the recall provisions as
discussed in Subpart S of 40 CFR Part 85. Finally, in the
unlikely event that disagreements with the recall sample
remain, manufacturers are given an opportunity, in an
adjudicatory hearing, to contest EPA's determination that the
class is in non-conformity. The results of that hearing are,
of course, judicially appealable.
The above procedures involving recall screening ensure
that the manufacturers are indeed involved in the current
recall screening process, and EPA fully expects that such
involvement will continue in the full-life LDT program and the
developing HDE recall program. EPA expects considerable
dialogue with the industry on the implementation details of
these new programs, and in fact some preliminary discussions
have been held. EPA presented a brief synopsis of the current
LDV/LDT recall program at the Useful-Life Workshop on February
18, 1983, and a subsequent meeting was held between EPA and EMA
representatives on June 2, 1983.[8]
In any event, EPA and the industry are in agreement on the
need to develop procedures and implementation approaches for
minimizing the possibility that a rebuilt or wornout engine
might be included in a recall evaluation sample. Since few
LDTs are rebuilt and no HDE recall program currently exists,
EPA believes that the full-life useful-life requirement can be
implemented now, and the details for implementing the
provisions of the recall program for LDTs and HDEs can be
refined in the future, through discussions between EPA and the
industry.
Although EPA can understand how Mack might make a
connection between the 40 percent AQL and its impact on the
recall program, there simply is none. SEA and recall are two
distinct EPA programs, addressing compliance on the assembly
line and in use, respectively. The AQL in SEA testing was not
established at 40 percent to condone nonconformance, but was
set at that level in recognition of manufacturing
practicalities and economic and other negative impacts of an
SEA failure (i.e., lost production, lost wages while a fix for
the problem is implemented, etc.). In fact, any vehicle/engine
which fails during an SEA must be fixed before it can be sold.
-52-
-------
It is still the Agency's desire and should also be the
manufacturers' goal, that every vehicle/engine produced meets
the emission standards when produced. Therefore, EPA does not
feel constrained by the 40 percent figure for the recall
program, since the goals of the two programs are different.
EPA also notes that the LDV SEA program includes a 40 percent
AQL and the LDV recall program has not suffered as a result.
In response to Mack's comments concerning
laboratory-to-laboratory variability and also as a partial
response to Mack's concern over the impact of the 40 percent
AQL, it should be noted that the lack of rigidly defined
procedures for recall evaluation and for determining that a
substantial number of vehicles/engines are in nonconformity,
provides EPA some flexibility for accounting for the impact of
such factors. EPA expects to continue judicious use of this
flexibility in the future to account for factors such as these.
EPA cannot agree to Mack's request for a 3-year grace
period from recall liability for new engine lines, while the
manufacurer gathers in-use data on the performance of its new
engines. Manufacturers do not introduce new engine lines to
the marketplace without extensive durability and assurance
testing both on engine dynamometers and in actual vehicles
before production begins. Given this practice, and EPA's
provisions which allow manufacturers to determine their
deterioration factors by any means they deem appropriate, the
manufacturers should be able to utilize the results of such
durability and assurance testing to determine a reasonably
accurate deterioration factor. To account for unforeseen
problems in use, the manufacturer can always build a cushion
into the certification deterioration factor or decrease
low-mileage targets and thereby minimize in-use noncompliance
risk. Manufacturers cannot be spared the liability of not
complying with the emission standards in use. This is a
central and important part of the mobile source control
program; it ensures that manufacturers build engine/vehicles
that perform well in use.
Regarding the idea of limiting recall liability to a
specific list of emission control components, EPA cannot accept
the manufacturers' position. The industry has argued that
emission-related components (i.e., those that affect emissions,
but are not specifically designed for emission control—fuel
injection systems, for example) should be excluded from recall
liability becuase they will be kept in good repair to avoid
degradation in performance and fuel economy. This may be true,
and, if so, defects uncovered in emission-related components
would be rare and should pose no problem to the manufacturers.
Conversely, recall evaluation testing may find that there are
significant problems with emission-related components, and a
recall program would assure correction of these problems.
-53-
-------
Publishing a list of select emission control components would
preclude the possibility of correcting such problems should
they be uncovered. Finally, in the recall provision of the
Act, Congress did not limit recalls to non-conformities caused
by certain components, but rather required remedial action when
a class of vehicles fails for any reason to conform to the
applicable emission standards.
Conclusions
EPA will work closely with the LDT and HDE manufacturers
to ensure that the new recall programs are implemented in an
equitable and reasonable manner, and that manufacturers'
concerns over wornout and rebuilt engines are properly
addressed. These implementation provisions will be developed
with public involvement, and will be modified in the future as
experience dictates. There is every reason to believe that
these new recall programs can work as smoothly as the current
LDV and LDT programs. EPA concludes that no additional recall
provisions are required at this time to implement the modified
full-life useful-life approach for 1985.
Heavy-Duty Diesel Engine Subclasses
Summary of the Comments
EMA and several manufacturers did not agree with EPA's
approach of subdividing the heavy-duty diesel engine (HDDE)
class on the basis of gross vehicle weight (GVW). In the
proposal, EPA subdivided the HDDE class into three distinct
subclasses based on a range of GVWs and then assigned
useful-life periods to each subclass. Under the EPA approach,
an engine's assigned useful-life period would then be derived
from the GVW of the vehicle in which the engine was installed!*
EMA commented that this approach was flawed because a given
engine line might be sold for use in applications which
encompassed more than one HDDE subclass. EMA also did not like
the nomenclature which EPA used to identify its three HDDE
subclasses (i.e., medium, light heavy, and heavy heavy).
As an alternative approach, EMA suggested splitting the
HDDE class into three subclasses based on the primary intended
service application for which the engine was designed and
sold. These three subclasses would be called light heavy-duty
diesel engines (LHDDEs), medium heavy-duty diesel engines
(MHDDEs), and heavy heavy-duty diesel engines (HHDDEs). The
LHDDE subclass would cover applications such as motor homes,
multi-stop vans, large utility vehicles, pickup trucks, and
delivery vans. The MHDDE subclass would cover engines that
were designed for short haul or intracity operation such as van
trucks, stake trucks, single axle tractor/trailers
combinations, and school buses. The HHDDE subclass would
-5 4-
-------
primarily entail engines designed for full-load, long haul
intercity operation, such as those used in over-the-road
tractor/trailer trucks and intercity commercial buses. EMA
further recommended that EPA review each manufacturer's primary
service category designation to ensure that engines were
properly classified for regulatory purposes.
Virtually all of the HDDE manufacturers concurred with
EMA's comments. However, Daimler-Benz and ATA suggested engine
horsepower as another plausible approach since it would also
allow the manufacturers to characterize the engines in the
manner in which they were normally used.
Analysis of Comments
The HDDE classification approach suggested by EMA has
considerable merit. Basing the HDDE subclasses on primary
intended service applications is preferable to the GVW-based
approach by EPA, because it avoids two potential problems of
the GVW-based approach. First, it avoids the potential design,
certification, and recall complications which arise if an
engine model would be used in more than one of the GVW-based
subclasses proposed by EPA. This problem is avoided simply
because GVW is removed as a useful-life determinant. Second,
it avoids the potential problems associated with atypical
applications within the GVW-based subclasses proposed by EPA.
For example, even though garbage trucks fall in GVW Classes VII
and VIII (HHDDE under the EPA GVW-based approach) their engine
requirements and vehicle usage patterns are not typical of most
Class VII and VIII vehicles. Under EPA's proposed approach
these engines would have been assigned the same useful-life
period as over-the-road trucks, which probably would not be
appropriate. The primary intended service application approach
avoids this GVW-based complication, and recognizes that a
typical MHDDE may be efficient in this application, and would
have a useful life typical of MHDDEs, not HHDDEs.
The HDDE classification approach suggested by EMA is
preferable to that proposed by EPA. For those engines which do
not readily fall into either the light, medium, or heavy
heavy-duty subclass, EPA is retaining the provisions which
allow the manufacturer to petition the Administrator for a
different useful-life period.
At this time, EPA foresees no need to review the
manufacturers' primary intended service determinations as
suggested by EMA, and does not desire to establish the need for
additional approvals during certification. EPA believes that a
labeling requirement could be used to assure that engines are
not misclassified. Under this approach, manufacturers will be
required to label HDDEs as to the subclass for which they are
certified. The label will also include alternative assigned
-55-
-------
useful-life periods, if applicable, as described above. Market
forces should then help ensure proper engine classification by
the manufacturer, and selection of the appropriate engine by
the purchaser. Although this approach should guard against
abuses, EPA retains the right to challenge any manufacturer's
practice in determination of subclasses should
misclassifications occur.
The horsepower-based approach proposed by two commenters
may also be plausible, because there is generally a
relationship between engine horsepower and other parameters
such as the load factor which could in turn adequately
delineate an engine's application. However, this approach is
not preferable to that proposed by EMA because no body of data
is readily available which could be used to develop an
appropriate relationship between engine horsepower and average
useful life.
Conclusions
The HDDE class will be split into three subclasses on the
basis of primary intended service application, as suggested by
EMA. Each engine will be labeled with the subclass for which
it is certified. The provision allowing a manufacturer to
request a different useful-life value under special conditions
will also be retained. However, these values will have to be
printed on the label.
Assigned Useful-Life Periods
Summary of the Comments
All of the manufacturers claimed that one or more of EPA's
proposed assigned useful-life periods (period to engine
retirement or rebuild) as too long. Since the comments
pertaining to the various assigned useful-life periods are
fairly specific and detailed, they will be grouped by
vehicle/engine class and each will be prefaced by EPA's
rationale for establishing the useful-life value which was
originally proposed. The development of the assigned
useful-life periods is more fully documented in an EPA
memorandum which was released concurrent with the proposal.[13]
a. Light-Duty Trucks
EPA's proposed assigned useful life of 12 years/130,000
miles was based on an average of the following data:
Engine rebuild surveys
Survey Data Research (SDR) 171,000 miles
"maximum likelihood"
-56-
-------
SDR "median ranks"
141,000 miles
Scrappage data
DOE 124,000 miles
Michigan Technological
University (MTU) 120,000 miles
Engineering Estimate
Myers, SAE 750128 100,000 miles
Average 131,200 miles
The two Survey Data Research (SDR) survey numbers were based on
the same set of survey data, aggregated and analyzed in two
different ways. Ford Motor Company stated that this was
inappropriate because if the data set forming the basis for the
two analyses were biased in any way, the effect of the error
would be doubled (since it would constitute 40 percent of the
average, rather than 20 percent). Ford believed that the data
were biased, because the two projected engine rebuild mileages
were higher than the mileage at which the average vehicle would
be scrappped by roughly 20,000 and 50,000 miles, respectively.
Ford claimed that this was contrary to the common sense
conclusion that the miles to rebuild should be less than the
mileage at the vehicle scrappage point.
EMA and GM stated that the use of scrappage data in
developing useful-life mileage values for LDTs and HDEs was
inappropriate because the data included engines that had been
rebuilt, therefore raising the average scrappage point
mileage. They also asserted that use of scrappage-rate data
represented a departure from the Agency's original regulatory
intent of basing useful life on mileage to engine rebuild. GM
carried this argument one step further, saying that useful-life
periods should be based on the need for rebuild rather than on
"owner action," (i.e., actually having the engine rebuilt). GM
did not offer any suggestions, however, as to exactly how this
determination was to be made, other than to say that they felt
EPA's previous effort to provide objective. end-of-life
indicators for screening wornout engines out of recall samples
(the rebuild criteria in 40 CFR §86.084-21) was "unworkable" in
terms of accomplishing the stated objective.
VW argued that the data used by EPA to develop the LDT
assigned useful life did not include the smaller 4-cylinder
pickups that have been introduced in the last few years and
which they felt are not designed for a useful life of 130,000
miles.
-57-
-------
b. Heavy-Duty Gasoline Engines
EPA averaged the following data in proposing an assigned
useful life of 120,000 miles for HDGEs:
Rebuild surverys
SDR "maximum likelihood" 134,000 miles
SDR "median ranks" 124,000 miles
Fleet Maintenance and
Specifying magazine[9] 100,000 miles
Scrappage rate data
DOE 129,000 miles
MTU 114,000 miles
Average 120,200 miles
Ford, GM, and EMA all felt that the useful life for HDGEs
should be 100,000 miles or less. This contention is based on
the arguments mentioned above for LDTs regarding scrappage data
versus rebuild data and also on their belief in the possibility
that the SDR survey data overstated mileage to rebuild. EMA
suggested the inclusion of data from a rebuild survey conducted
by the ATA and also engineering estimates from a draft study
done under EPA contract by Arthur D. Little, Inc.[10,11]
c. Heavy-Duty Diesel Engines
In the proposal, EPA split the HDDE class into three
subclasses based on GVW, and proposed useful-life periods based
on the general design and usage characteristics of each. The
"medium-duty diesel" subclass, all HDDEs in vehicles less than
19,500 lbs. GVW (Classes IIB-V), represented a relatively new
diesel application in a. field heretofore dominated by gasoline
engines and there were few data available regarding average
service life. However, EPA reasoned that since these engines
introduced as a replacement for HDGEs, they should last as long
as the gasoline engines they were designed to replace in order
to be competitive, and so a similar useful-life period of
120,000 miles was proposed.
The second subclass proposed, "light heavy diesel,"
(19,501-26,000 lbs. GVW - Class VI) had a useful-life period of
200,000 miles, which was determined by averaging the following
data:
-5 8-
-------
Rebuild Surveys
SDR (Class VI engines)
Fleet Maintenance and
203,000 miles
Specifying Magazine
229,000 miles*
Engineering Estimate
Myers
16 2,500 miles**
Average
198,167 miles
* Average of values for sleeved and non-sleeved bus engines.
** Midpoint of range.
The third subclass was called heavy heavy-duty diesel (GVW
above 26,000 lbs. - Classes VII-VIII). Since the data
available indicated that virtually all heavy heavy-duty diesel
engines were rebuilt, the proposed useful-life value was based
on an average of two rebuild surveys:
Since most manufacturers supported the EMA alternative
HDDE classification scheme discussed earlier, their comments
concerned both the methodology used by EPA to develop the
proposed useful-life values and EPA's methodology and its
results as they applied to the HDDE subclasses suggested by
EMA. Since EPA has accepted the EMA classification system, the
summary and analysis of comments for HDDEs will focus on those
comments pertaining to EPA's methodology for estimating
useful-life periods and the relationship between the assigned
useful-life periods proposed by EPA and the EMA HDDE subclasses.
For the sake of clarity, further references to the HDDE
subclasses will use the EMA terminology (LHDDE, MHDDE, and
HHDDE). When the subclasses proposed by EPA are mentioned,
their full names with be used (medium-duty diesel, light
heavy-duty diesel, heavy heavy-duty diesel).
Having presented the necessary preliminary information, we
turn now to the comments. First, no significant comments were
received on EPA's proposal that the medium-duty diesel assigned
useful-life period should be the same as that used for HDGEs.
Ford agreed with this approach.
SDR
267,000 miles
Fleet Maintenance and
Specifying Magazine
281,000 miles
Average
274,000 miles
-59-
-------
For the next subclass, EMA disagreed with EPA's use of the
vocational groupings in the Fleet survey as being
representative of LHDDE and MHDDE usage for establishing
useful-Life periods. The Fleet survey aggregated the data on
the basis of vocational application (e.g., bus fleet, utility)
as well as on the basis of some significant engine design
characteristics (sleeved versus non-sleeved engines). EMA
suggested that since most MHDDEs are non-sleeved, the MHDDE
assigned useful-life period should be based on the Fleet
rebuild mileage for non-sleeved engines (175,000) rather than
on vocational categories.
EMA and Ford suggested that the SDR survey rebuild mileage
for Class VI engines used in the EPA calculation of the
useful-life period proposed for light heavy-duty diesels was
inappropriate for use in calculating the useful-life for MHDDEs
under the EMA classification system. Even though the
light-heavy-duty subclass proposed by EPA and the MHDDE
subclass proposed by EMA are quite similar, it was thought that
the SDR sample for Class VI engines likely included a number of
premium HHDDEs which would raise the average rebuild mileage.
Thus, EMA and Ford believed that the MHDDE useful life should
be less than that determined for EPA's GVW-based light
heavy-duty diesel subclass.
It was also suggested that data from the ATA maintenance
survey and from the draft study by Arthur D. Little be added to
the data used for calculating the average useful-life periods.
Given this information, EMA, Ford and GM felt that the MHDDE
average full-life value should be 170,000 miles, based on the
change in methodology. Caterpillar felt the MHDDE figure
should be 150,000 miles, based on an average value of different
applications of their 3208 model.
There was not significant disagreement on the assigned
useful-life period of 275,000 miles EPA proposed for heavy
heavy-duty diesel engines, although several commenters noted
that operation of Class VII trucks is not typical of that of
Class VIII trucks. In their view, the 275,000 miles was far
more representative of Class VIII operation than Class VII
operation.
EMA also suggested that the EPA assigned useful-life year
values were not equivalent to the mileage values for the
various classes/subclasses. They argued that equivalency
should be maintained.
Analysis of the Comments
a. Light-Duty Trucks
EPA accepts the Ford comment regarding the use of two
different numbers resulting from alternative analyses of the
-60-
-------
same set of data in the SDR survey. EPA intended to use the
SDR data as a rebuild mileage survey, and to let the scrappage
rate data serve for high-mileage non-rebuilds. Since the SDR
"maximum likelihood" figure in question includes non-rebuilt
engines in determining average lifetime mileages, it would be
inappropriate to include it as a rebuild figure. Therefore,
EPA has dropped the SDR "maximum likelihood" value from the
averaging total and has used only the "median ranks" value,
which is limited to rebuilt engines.
Regarding the comments on including scrappage data in the
useful-life calculation, EPA believes the manufacturers have
misinterpreted the Agency's intent regarding what constitutes
useful life. In the original full-life useful-life
regulations, useful life was defined as "the average period of
use up to engine retirement or rebuild, whichever occurs first"
(emphasis added).[12] Under the modified full-life definition
it is the Agency's intent to retain this concept. Thus, it is
not EPA's intent that useful life should be only mileage to
rebuild when establishing the assigned useful-life values in
the modified full-life plan but rather that it should also
consider vehicle scrappage. Moreover, both "rebuild" and
"retirement" (scrappage) can be described as "owner actions"
and there is no mention of "need" for a rebuild in the above
definition. Available data indicate that the average LDT is
far more likely to be scrapped than to be rebuilt. An analysis
of the SDR survey data indicate that only about 12 percent of
all LDTs are ever rebuilt. [13] Thus, for 88 percent of the
vehicles in question, useful life is the mileage to
"retirement" rather than the mileage to rebuild, and exclusion
of scrappage data would overlook a significant body of data in
the calculation of average useful life.
While EPA acknowledges the point made by EMA and GM that
there may be some bias introduced into the scrappage rate data
by the presence of high mileage rebuilt engines, the percentage
of rebuilds (about 12 percent) is not large enough to have a
significant effect. Also, it should be recognized that there
are also biases in the other direction. Scrappage totals
include many low-mileage wrecks, for example, which tend to
lower the average. A major driveline failure may also result
in scrappage of a vehicle with additional miles remaining in
the engine because retirement and replacement would be more
cost-effective than repair. None of the available data on
average useful-life periods are without some drawbacks. With
the exception of HHDDEs where virtually every engine is rebuilt
at least once, neither rebuild data nor scrappage data is
adequate in and of itself to unequivocally establish
useful-life periods. Therefore, in light of these unavoidable
uncertainties, EPA has averaged data from a wide variety of
sources to minimize the effect of the deficiencies in the data
bases. These deficiencies were judged to be minor, and in some
-61-
-------
cases offsetting, thus allowing EPA to derive a representative
useful-life value.
While VW is correct that the data used in deriving the LDT
useful-life value do not include the newer, smaller light-duty
trucks, EPA sees no reason why the service lives of these
latter vehicles should differ significantly from those of the
standard size LDTs. While the lighter LDTs are powered by
smaller, less powerful engines, usually of 4 cylinders, and
operate at somewhat higher engine revolutions than conventional
LDTs, the trucks themselves are also lighter in weight and have
less payload and frontal area than their standard-size
counterparts. Few if any of these small pickups are likely to
be loaded to maximum capacity with any degree of regularity
and, as VW's comments indicated, the vast majority will in fact
be used for personal transportation, as are many standard
LDTs. EPA, therefore, concludes that there is no need for a
shorter assigned useful life for these vehicles. Since neither
VW nor any of the other commenters submitted any data to
substantiate the need for a shorter useful life for the smaller
LDTs, EPA will continue a common useful-life period for all
LDTs. LDT manufacturers also have the option of requesting an
alternative useful-life value in cases where the assigned
useful-life value is significantly unrepresentative of the
useful life for a particular engine family.
Therefore, the only change necessary to the LDT assigned
useful-life period calculation is to drop the SDR maximum
likelihood rebuild number, and reaverage the remaining four
sources. An average of the four sources remaining yields a
figure of 121,000 miles, so EPA will reduce the assigned useful
life for LDTs from 130,000 miles to 120,000 miles.
b. Heavy-Duty Gasoline Engines (HDGEs)
EPA rejects the arguments advanced by EMA and others
regarding the exclusion of scrappage-rate data in the HDE
useful-life calculation for the same basic reasons outlined
above in the LDT discussion. The SDR survey data indicate that
only 28 percent of the HDGEs are rebuilt or replaced, so again,
for the vast majority of HDGEs, useful life is the mileage to
retirement.
Although the above-mentioned Ford comment regarding SDR
survey data was made in reference to LDTs, the same general
considerations hold for HDGEs as well. The two HDGE rebuild
mileages from the SDR survey are not as disparate as the LDT
figures. However, if the SDR data are to be representative of
engine rebuild data in the HDGE average useful-life
calculation, the "maximum likelihood" value should be dropped,
since it includes non-rebuilt engines.
-62-
-------
EPA will also include rebuild data from the American
Trucking Association survey and estimates from the Arthur D.
Little draft truck usage study as suggested by EMA. Although
the contract under which the latter study was done was
terminated prior to completion, the useful-life estimates in
the report are reasonably consistent with other engineering
estimates, and EPA has no objection to inclusion of the A. D.
Little figures in the useful-life calculation. As shown below,
an average of the two scrappage rate values, the three rebuild
survey mileages, and the two engineering estimates results in a
useful-life period of about 108,000 miles. Therefore, EPA will
assign a value of 110,000 miles for HDGEs, rather than the
proposed value of 120,000 miles, based on the following
calculation:
Scrappage Rate Surveys:
Michigan Technological University 114,000
DOE 12 9,000
Rebuild Surveys:
SDR 124,000
Fleet 100,000
ATA 89,000*
Engineering Estimates:
* Sales-weighted average of trucks under 20,000 lbs. GVW (73
percent of sales) and of trucks over 20,000 lbs. GVW (27
percent of sales). These projected sales percentages were
multiplied by ATA mean survey mileages of 91,447 and
82,450 miles, respectively. If the modal or median values
are used, the sales-weighted rebuild mileages are 94,600
and 90,950 miles, respectively.
c. Heavy-Duty Diesel Engines
As in the Summary of the Comments, the EMA subclass
designations will be used throughout this section of the
analysis to avoid the confusion that would result from use of
both EPA and EMA terminology. The analysis will also be
oriented toward the EMA subclasses, since EPA is adopting them
over the subclasses as defined in the proposal. EPA agrees
with EMA that the Fleet vocational based rebuild figures used
in the MHDDE and LHDDE useful-life calculation may have
Little
Meyers
100,000
100,000
Average
108,000
-63-
-------
deficiencies. In any event, as stated in the above-mentioned
support memorandum concerning useful-life derivation, the LHDDE
subclass was assigned the same useful-life period as HDGEs, on
the judgment that LHDDEs should last about as long as the HDGEs
they were designed to replace. [14] The Ford comments and a
request by EMA to certify these engines to the HDGE useful life
provide additional support for this position.[15,16] Since the
HDGE useful-life period is based in part on the use of
scrappage-rate data, the EMA objection to its use also carries
oyer to the LHDDE subclass value. There are relatively few
data on the subject for LHDDEs. However, the Fleet survey
fbund that 43 percent of the non-sleeved engines in the survey
(the vast majority of which would be classified as MHDDEs) are
never rebuilt.[17] Most of the new LHDDEs are also non-sleeved
and are less expensive than MHDDEs, being designed to compete
with HDGEs. Since they are less costly to replace, and in some
cases are not designed to be rebuilt, it is likely that even
fewer LHDDEs than MHDDEs would be rebuilt. EPA will therefore
continue the linkage between LHDDEs and HDGEs, and establish
the assigned useful-life period for LHDDEs at 110,000 miles.
Turning now to MHDDEs, EMA felt that the figure of 203,000
miles quoted in the SDR survey for Class VI vehicles was too
large for the MHDDE subclass, because the Class VI vehicles
probably used some engines which would be considered as premium
HHDDEs. EPA concurs with EMA's assessment. Second, although
data in the Fleet article indicated that buses are typically
powered by MHDDEs, this application may not be representative
of MHDDE usage. Therefore, EPA accepts EMA's suggestion that
MHDDE useful life should be based in part on the average
non-sleeved engine mileage to overhaul reported in the Fleet
survey (175,000). The ATA survey rebuild mileage of 176,000
for diesel straight trucks also lends support to this figure.
An average of the sources suggested by EMA (including the SDR,
ATA, and Fleet, rebuild surveys and engineering estimates by
Little and Myers) yields an average rebuild mileage of 173,300
miles.
However, while EPA accepts the average of these data as
valid for the MHDDEs that are rebuilt, the Fleet survey also
indicates that an average of 43 percent of the non-sleeved
engines do not get rebuilt. This is a significant percentage
and must also be factored in to the determination of the useful-
* The straight truck data in the ATA survey included both
gasoline and diesel trucks. The median was 150,000 miles
and the mean was 170,470 miles. The relative values of
the median and the mode depict a disjointed data set.
Based on the HDGE analysis above, it was concluded that
the median value was relatively low due to the HDGEs. So
the modal value probably represented the diesel straight
trucks. Therefore, the 200,000-mile value was used.
-64-
-------
life period for MHDDEs. EPA expects that non-rebuilt engines
would be operated somewhat past the point where the average
rebuild would normally occur, as owners attempted to extract
the maximum service. The ATA survey provides a modal value of
200,000 miles for trade-in of straight trucks.[18]* Since few
owners are likely to go to the expense of a rebuild at 173,000
miles, and then trade in the truck 27,000 miles later, 200,000
miles is a reasonable estimate of useful life for non-rebuilt
engines.
Therefore, addition of 57 percent of 173,300 miles
(98,781) to 43 percent of 200,000 miles (86,000) yields a
weighted average of 184,781 miles, so EPA will establish a
period of 185,000 miles as the MHHDE assigned useful life.
Given the change in the HDDE classification approach from
gross vehicle weight to what is essentially an
application-based approach, there also is a need for a
reassessment of the assigned useful-life period for HHDDEs,
just as was done for MHDDEs. The heavy heavy-duty diesel
engine subclass originally proposed by EPA covered GVW Classes
VII and VIII trucks and buses, and these vehicles included some
engines that would now be classified as MHDDEs or even LHDDEs
under the EMA approach. Caterpillar's comments indicated, for
example, that its 3208 engine, which the manufacturer
considered an MHDDE, was sold "almost solely" in Class VII or
VIII GVW vehicles. [19] International Harvester Corporation,
which also manufactures LHDDE and MHDDEs, stated that "every
diesel engine" the company offers for sale could be found in
Class VII or VIII GVW vehicles.[20] The assigned useful-life
period for heavy heavy-duty diesel engines in the proposal
reflected this vehicle mix and is therefore understated for
HHDDEs under the EMA approach. With the adoption of EMA's
subclasses based on application rather than GVW, the HHDDE
subclass will now be predominantly premium-engines designed for
long-haul, high-mileage service applications, necessitating an
adjustment in the assigned useful-life period. The SDR Classes
VII and VIII data are not adequately representative of premium
HHDDEs to serve as the basis for the analysis since these
vehicles would use some MHDDEs, just as EMA asserted that the
Class VI SDR figure was overstated because it included some
HHDDEs. However the SDR survey determined a rebuild mileage of
303,000 for "long haul" usage engines, which would clearly
reflect HHDDEs. Also, the Fleet average rebuild mileage for
sleeved engines {281,000 miles) and the ATA mean rebuild
mileage for "tractors" (296,862 miles) are clearly
representative of the type of engine and operation in
question. An average of these three sources plus the A.D.
Little engineering estimate (290,000 miles) yields a mileage of
292,716 miles. Based on this average, EPA will assign a
useful-life value of 290,000 miles for this subclass.
-65-
-------
If the median or- modal values in the ATA survey (300,000
miles) were substituted for the mean, the resulting average of
the four sources would be 293,500 miles. Using the EMA
approach to the HDDE subclasses effectively addresses the
comment concerning the grouping together of Class VII and Class
VIII trucks in the heavy heavy-duty diesel subclass proposed by
EPA.
The final area to be considered is the number of years, as
opposed to miles, in the assigned useful-life values. EMA
s"tated that EPA's position appeared to be that a truly
representative years-to-rebuild value was not necessary "if an
accurate mileage value is prescribed."[21] EMA did not agree
with this position, saying that an accurate figure for
equivalent years to rebuild was necessary due to the fact that
many HDDEs accumulate a great many hours of running time
w:ithout accumulating many miles. Actually, EPA has never
maintained the position claimed by EMA. In most cases, the
period of years was roughly equivalent to miles of use as
described above. The Michigan Technological University (MTU)
vehicle mileage tables used in the EPA Emission Factors
Program, the National Highway Traffic Safety Administration
(NHTSA) mileage tables, and the Department of Energy "Highway
Fuel Consumption Model" (DOE) annual vehicle mileage tables
were consulted in determining years to the end of useful life
for all categories except Classes VII and VIII.[22] In the
latter case EPA found that while most applications were very
high mileage (i.e., the useful-life mileage would be
accumulated in 3-5 years), enough relatively low-mileage
applications would be included so that a considerably longer
period of years was necessary to be representative of their
full useful lives. Examples of these low-mileage applications
include concrete mixers, fire trucks, and garbage packers.
Although some of these applications will now likely be included
in the MHDD service class, EPA believes that some HHDDEs will
continue to be sold for this kind of use. An extended period
of years should not affect long-haul intercity vehicles for
which the useful-life mileage total will become the limiting
factor, but will allow a more representative useful-life period
for the lower mileage applications. Therefore, EPA will adjust
the useful-life year values for approximate equivalency with
the revised useful-life mileages, except for HHDDEs, which will
be assigned the same useful-life years as MHDDEs (i.e., 8
years).
Conclusions
Based on the above analysis, and the HDDE classification
system suggested by EMA, EPA finds it appropriate to revise the
useful-life values as follows:
-6 6-
-------
Mileage
Years
Proposed Final
Proposed Final
LDT
HDGE/LHDDE
MHDDE
HHDDE
130,000 mi. 120,000 mi.
120,000 mi. 110,000 mi.
200,000 mi. 185,000 mi.
275,000 mi. 290,000 mi.
12 years 11 years
10 years 8 years
10 years 8 years
10 years 8 years
EPA concludes that scrappage data should not be excluded
from the assigned useful-life calculations, as EMA suggested,
but has no objection to inclusion of other data as desired by
EMA. The assigned useful-life period will be the same for all
LDTs, rather than setting a shorter period for the lighter
4-cylinder vehicles as desired by VW and others. Finally, it
should be repeated that a manufacturer has the option to
request an alternative useful-life period for an individual
engine family if there is reason to believe that the assigned
useful-life value is unrepresentative.
Minor Issues
Summary of Comments
Manufacturers have raised a number of minor issues under
the modified full-life provisions. GM asked whether a
manufacturer would be expected to test engines after they were
worn out to determine a deterioration factor (DF) for the full
assigned useful life and also how a manufacturer would
determine the DF if the test engine failed before reaching the
assigned useful-life value. Mack Trucks suggested that bench
testing would be sufficient for checking the durability of HDDE
emission control components and that there would be no need for
full-life useful life. Ford presented its opinion that the
manufacturer's certification statement, to the effect that a
properly maintained engine will conform to the applicable
standards for its full useful life, must be qualified to take
into consideration engine wearout before the end of the
assigned useful-life period. The ATA wanted EPA to publish
useful-life values for all HDEs. Lastly, AMC stated that full
life would break the link allowing shared technology between
light-duty vehicles (LDVs) and LDTs, since LDTs would now
require more durable components. AMC predicted increased costs
to both manufacturers and consumers as a result of breaking
this LDV/LDT link.
Analysis of Comments
In response to GM's concerns, there will be no specific
durability testing requirement under modified full life. As
long as the manufacturers are satisfied that emissions will not
exceed the standard for the useful life of their
vehicles/engines, they are free to determine deterioration
-67-
-------
factors using any method desired, as long as good engineering
practices are followed. The bench testing option advanced by-
Mack Trucks might be a more cost-effective way to assess
component durability, for example, if the manufacturer felt
confident of the accuracy of that approach. However,. EPA
disagrees that such bench testing is adequate to eliminate the
need for the full-life useful-life requirement. EPA believes
that the full-life requirement is still necessary to provide
increased assurance of durable component design by holding the
manufacturer accountable for lifetime emissions compliance.
Bench testing represents one potential approach to durability
assessment which the manufacturers may choose.
With regard to Ford's point concerning the certification
statement, elimination of the useful-life labeling requirement
also removes the current compliance statement required for LDTs
and HDEs under 40 CFR 86.084-35. However, for vehicle/engine
classes where a single assigned useful-life value is specified,
that label is replaced by a general compliance label, as is
currently specified for LDTs and HDEs. This label states that
the vehicle/engine conforms to the applicable model year EPA
regulations. Since HDDEs are not all assigned the same
useful-life period, they will also be labeled to indicate the
subclass for which they are certified. Any LDT/HDE for which
an alternative useful-life value is approved by the
Administrator will also be labeled with the alternative
useful-life value. To address the Ford concern, EPA will
retain the current qualifying statement that "This engine's
actual life may vary, depending on its service applications."
Since engine-specific useful-life values . are replaced by
assigned useful-life periods, there should be no need to
publish individual useful-life data as requested by the ATA.
The assigned useful-life values for classes/subclasses are
published in this rulemaking. In addition, as outlined above,
HDDE manufacturers will be required to label their engines with
the service classes for which they are certified. As requested
by ATA, any vehicle/engine for which an alternative useful-life
value was approved would also be labeled with the alternative
value. Since ATA!s interests seem to be primarily in the area
of HDDEs which will all be labeled with the subclass as a
matter of course, and since alternative useful-life values will
be indicated if applicable, EPA feels ATA's needs will be
addressed by the above measures.
Finally, EPA does not believe the impact of full-life
useful life will be so great as to inhibit the sharing of
technology between LDVs and LDTs, as AMC suggests. It is
hardly cost-effective to redesign a component or design a
replacement for it and then continue to produce the old one in
parallel with production of the new or redesigned component.
-6 8-
-------
The Agency finds it difficult to believe that any manufacturer
would choose this course of action, particularly since it
provides a much narrower base for amortization of development
and tooling costs than would application of the component to
the entire product line. Since the applications and technology
are basically similar for both LDVs and LDTs, EPA believes that
commonality of components will continue to be standard
practice. If the durability of some of these components
improves as a result of the full-life requirement for LDTs, it
will result in an additional benefit to the LDV buyer and
improve the emissions of the LDV. There is no reason, however,
why separate components need be produced for LDVs and LDTs.
Conclusions
Most of the minor certification issues require no action
on EPA's part. The qualifying statement on the label that
average useful life will vary according to service application
will address Ford's concerns. ATA's request for publication of
the assigned useful-life values is addressed by the values
published in the rulemaking and by the labeling requirement for
HDDE subclasses and for alternative useful-life periods. AMC's
fear that full-life useful life will break the traditional
technology link between LDVs and LDTs appear overstated and EPA
rejects the idea, that any duplication of effort or waste of
resources will result.
-69-
-------
References
1. Summary and Analysis of Comment to the NPRM, 1983
and Later HDEs, pp. 105-110, December 1975.
2. Summary and Analysis of Comment to the NPRM, 1983
and Later LDTs, p. 1-3, May 1980.
3. "Report of the Committee on Public Works, 94-717,"
Legislative History of the CAA Amendments of 1977, Vol. 3, p.
671, May 10, 1977.
4. "The Clean Air Act as Amended, August 1977," Ser.
95-11, p. 11, November 1977.
5. Harley-Davidson Motor Company, Inc. vs. EPA, U.S.
Court of Appeals, D.C. Circuit, No. 77-1104, March 9, 1979.
6. See Reference 4, p. 122.
7. "House of Representatives Report 95-564," 75th
Congress, 1st Session, p. 164, 1977.
8. EPA Memorandum, "Summary of the February 18, 1983
HDE/LDT Useful-Life Workshop," Terry P. Newell, March 1, 1983.
9. Fleet Maintenance and Speci fying, Vol. 7, No. 3,
March 1981, pp. 39-43.
10. "Report on Commerical Trucking Diagnostic Needs
Survey," American Trucking Associations, Inc., c.1981.
11. Draft Report "HD Vehicle Engine Service Accumulation
Cycle, A. D. Little, Inc., EPA Contract 68-03-2712, April 1977.
12. 40 CFR 86.084-4(b)(1).
13. "Calculation of Total Rebuild Percentages for LDTs
and HDGEs," EPA Memorandum from Robert J. Johnson, July 25,
1983.
14. "Determination of Useful-Life Values for Light-Duty
Trucks and Heavy-Duty Engines," EPA Memorandum from Robert J.
Johnson, December 13, 1982.
15. Response to January 12, 1983 NPRM, Public Docket
A-81-11, IV-D-86, p. 16, April 18, 1983.
16. Letter from Thomas C. Young to Kathleen M. Bennett,
Office of Air, Noise and Radiation, May 3/ 1982.
-70-
-------
References (cont'd)
17. Fleet Maintenance and Specifying, Vol. 7, No. 7, May
1981.
18. See Reference 9, p. 6.
19. Caterpillar Response to January 12, 1983 NPRM,
Public Docket A-81-11, IV-D-77, p. 4, April 13, 1983.
20. International Harvester Response to January 12, 1983
NPRM, Public Docket A-81-11, IV-D-81, p. 23, April 18, 1983.
21. EMA Response to January 12, 1983 NPRM, Public Docket
A-81-11, IV-D-83, p. 53.
22. "Conversion of Useful-Life Mileages to Periods of
Years," EPA Memorandum, Robert J. Johnson, July 25, 1983.
-71-
-------
3. Issue: Alternative Test Cycles
A. The Real Time Cycle for Heavy-Duty Diesel Engines
Summary of the Issue
An alternative heavy-duty diesel test cycle, the Real Time
Cycle (RTC), has been developed by the Caterpillar Tractor
Company (Caterpillar). It was developed in response to
industry-wide concern over the methodology used to generate the
EPA cycle and its resulting representativeness.
Summary of Comments
Most manufacturers and the Engine Manufacturers
Association (EMA) have made specific recommendations concerning
the use of the RTC for certification testing. Some of the
recommendations, however, have changed over time as additional
data were gathered on the RTC. A brief review of the
chronology of events is appropriate for this discussion.
The EMA and member companies recommended in April 1982
that EPA adopt as a test option the use of the RTC. This
recommendation was based upon the industry's concern about the
representativeness of the EPA cycle.
Shortly thereafter, EPA reviewed the technical basis for
the creation of the RTC. (Part of EPA's earlier analysis is
reproduced below.) EPA also reviewed the available data base
wherein emission results from both cycles were compared. At
the time, about 30 comparative data points were available.
EPA's draft analysis noted that a net difference in emissions
existed between the test cycles, and recommended that the
applicable emission standards be adjusted to account for the
offset. This was recommended so as to preclude an effective
relaxation of the emission requirements promulgated on January
21, 1980. EPA's original analysis (see Appendix, Chapter 5 of
the Transient Test Study) was distributed for public comment in
early summer 1982.
The EMA and member companies reviewed EPA's draft
analysis, and over time, in both informal and formal
communications, took issue with two of EPA's conclusions.
First, EMA disputed the need for an emission standard
adjustment. The argument was made that the heavy-duty diesel
cycle was never used in the standards development process, and
the use of a specific diesel cycle is decoupled from the level
of the standards. (The heavy-duty gasoline engine cycle was
used to derive the statutory emission standards.) Secondly, if
an adjustment were to be made, EMA disagreed with the
methodology EPA used to derive the equivalently stringent
-72-
-------
standard. EMA proposed a methodology which yielded an
RTC-based hydrocarbon (HC) emission standard of 1.20 grams per
brake horsepower-hour (g/BHP-hr), using the latest available
data. EMA and its member companies formalized their position
in a May 13, 1983 submission to EPA. They also recommended,
contrary to earlier recommendations, that only a single test
cycle bp used for certification. If the RTC cycle was not made
available with a HC standard of 1.2 g/BHP-hr, the industry
preferred the use of the EPA cycle at the 1.3 g/BHP-hr HC
standard.
Analysis of Comments
in this analysis, we address the construction,
representativeness, and relative stringency of the RTC.
Methodologies for emission standard adjustments are also
evaluated, as is the justification for such an adjustment.
Finally, the selection of a certification test cycle is made.
Cycie Development!3]
Caterpillar developed the RTC because of concern over the
accuracy of simulation of in-use truck operation represented by
the EPA cycles. This concern stemmed from alleged
instrumentation problems in the CAPE-21 project which they
argued resulted in a significant amount of questionable data
being accepted into the data base, and from the methodology EPA
used to generate the cycle. Caterpillar's objectives in
developing the RTC were to generate a cycle from the portion of
the CAPE-21 data base which it considered valid, and to
construct- the cycle so it better represented its judgment of
real-world truck operation.
The entire data base was first edited to remove what
Caterpillar believed to be questionable data. This editing
left 23 truck-days of data, or about 25 percent of the original
data base. statistical parameters were then chosen to
characterize the edited data base. These were mean values and
cumulative distributions of percentage rpm, percentage power,
and positive percentage rpm. The percentage idle time and
distribution in length of idle were also used. These
statistical parameters then became the target values for the
construction of the new test cycle.
To construct the new test cycle, the data were broken down
into the smallest elements which did not interrupt the normal
driving sequence. These elements were defined as the vehicle
operational events which occurred between vehicle stops. The
elements were then assembled into trial test segments which
matched, as closely as possible, the desired statistics of the
categories they represented. The categories were: Hew York
-73-
-------
Freeway (NYP), New York Non-Freeway (NYNF), Los Angeles Freeway
(LAF), and Los Angeles Non-Freeway (LANF). The idle time and
category weighting were adjusted to match the original CAPE-21
data base, since it was judged unlikely that instrument error
would change these parameters. The trial test segments were
then tested against the data base for maximum deviation of
cumulative distributions and then were compared visually. The
best cycles were selected and assembled into an entire driving
cycle.
The result was a heavy-duty diesel engine (HDDE) driving
cycle, the "Real Time Cycle," which matched very closely the
statistics of the edited CAPE-21 data base, and which the
diesel engine manufacturers believed was more representative of
in-use truck operation than the EPA cycle.
Statistical Analysis
A comparison of the target statistics from the edited data
base, the RTC statistics, and the EPA cycle statistics is shown
in Table 3-1. Additional statistics from the RTC, EPA cycle,
and the original CAPE-21 data base are listed in Table 3-2.
The most important statistical differences between the RTC and
the EPA cycle are:
1. The RTC includes a NYF segment, while the EPA cycle
does not.* (The NYF segment is higher in mean percentage rpm
and higher in mean percentage power than the NYNF segment.)
2. The RTC is 5.2 percent higher in mean percentage
power, overall, than the EPA cycle.
3. The RTC is 4.8 percent lower in percentage idle
time, overall, than the EPA cycle.
4. The sequential ordering of the cycle segments on the
RTC is LANF, LAF, NYF, and NYNF. The ordering on the EPA cycle
is NYNF, LANF, LAF, NYNF.
The statistical differences cited here may or may not
affect engine emission levels. A potentially significant
factor is that the observed engine work done over the test
cycle (BHP-hr) is higher by 16-18 percent on the RTC.
Furthermore, the reordering of the segments in the RTC permits
the engine to operate in the high power LAF mode earlier in the
cycle, which may lead to an earlier engine warm-up (although
EPA omitted the NYF segment because invalid data had been
included in the data base and the weighting for this
segment was small compared to the other segments. [4]
-74-
-------
Table 3-1
Target, RTC, and EPA Cycle Statistics [3]
Los Angele
s Non-Freeway
Los Angeles Freeway
Target Real Time EPA
Target
Real Time
EPA
Average rpm (%)
40.7
41.8 43
80.0
83.5
83
Average Power
(%)
24.1
25.9 26
58.9
56.4
56
Average Positive
Acceleration
4.6
5.7 6.1
1.9
1.2
2.4
Idle Time (%)
35.0
32.7 34
2.0
1.4
2.3
Category
Weighting
23.7
27.3 25.0
26,3
25.1
25.0
Los Anqeles Non-Freeway
Los Anqeles Freeway
Target Real Time EPA
Target
Real Time
EPA
Average rpm (%)
41.5
47.1
17.7
19.8
20
Average Power
(%)
Average Positive
Acceleration
41.0
2.8
54.4
4.6
19.4
3.8
22.3
3.6
16
5.6
Idle Time (%)
19
21
51.0
51.0
55
Category
Weighting
9.0
5.9 0
41.0
Overall
41.7
50.0
Target
Real Time
EPA
Average
rpm <%)
41.7
43.4
41.5
Average
Power (%)
32.8
33.7
28.5
Average Positive
Acceleration
3.9
4.2
4.6
Idle Time (%)
31.4
31.8
36.6
-75-
-------
RTC, EPA
Table 3-2
Cycle, CAPE-21 Data
Base Statistics
Parameter
RTC
EPA
CAPE-21
Torque
Mean (%)
30.57
28.32
27.00
Percent of
Cycle Time
Acceleration
(%)
18.21
15.68
15.10
Deceleration
(%)
18.37
16.85
15.25
Cruise (%)
22.48
20.43
18.75
Motor (%)
7.98
11.43
15.00
Idle (%)
32.96
35.61
35.00
RPM
Mean (%)
42.78
41.52
41.75
Percent of
Cycle Time
Acceleration
(%)
23.45
21.77
21. 50
Deceleration
<%)
22.48
21.93
19.50
Cruise (%)
19.74
16.10
19.50
Idle (%)
34.33
40.20
39.00
-76-
-------
operation in the LAF segment of the RTC is initially cooler
than on the EPA cycle) . Inclusion of the NYF segment in lieu
of another NYNF segment is one obvious reason why the RTC
BHP-hr is higher than that of the EPA cycle.
Test Cycle Correlation
Heavy-duty diesel engine manufacturers and EPA have now
tested many engines on both the RTC and the EPA cycle for the
purpose of comparing emissions results. All of the available
data have been collected and are summarized in Table 3-3.
Results are also plotted in Figures 3-1, 3-2, and 3-3. Since
typical diesel carbon monoxide (CO) emissions are much lower
than statutory levels, this pollutant comparison was not
included. Immediately obvious is the fact that emission levels
are different between the candidate test cycles.
The HC emissions difference between the test cycles is
explainable and expected. A decrease in brake specific HC
emission rates at higher engine loads is typically observed on
diesel engines. Consider the following mechanism for such an
observation: in diesel engines, HC emissions are in large part
attributable to residual fuel in the injector sac. The sac
volume is constant regardless of the amount of fuel injected;
as the load is increased (i.e., more fuel is injected), the
mass rate of HC emissions from the sac remains constant.
However, the brake specific rate of HC emissions (g/BHP-hr)
decreases at higher loads since the denominator (power-hour)
increases while the numerator (mass HC) remains the same. This
could explain most of the difference in HC emissions, given
that the RTC is a higher power test cycle and that both cycles
exercise the engine in fundamentally the same way.
The constant residual sac volume is likely not the only
mechanism by which emissions from each cycle are different.
However, further discussion of exact mechanisms at this point
is not important. What is important is the fact that the RTC
cycle correlates very well with the EPA cycle (and vice versa)
for many different engines. This indicates that emissions from
one cycle can be accurately predicted from those of the other.
The excellent correlation also indicates that both cycles
should be comparable in the ability to predict in-use emission
reductions, and that there is no inherent advantage in using
one cycle over the other. Given the difference in cycle
generation methodologies and the correlatable emission results,
and the reasonable presumption that the HC emissions offset is
primarily attributable to the difference in load factor between
the cycles, EPA concludes that both cycles are comparably
representative. Each by itself would be technically acceptable
for certification testing.
-7 7-
-------
Table 3-3
Summary of Bnissions Data EPA Cycle vs. Real Time Cycle
Number and Contained Cold/Hot Start Hot Start Only
Engine/
Test Number[a]
Test Lab
Cycle
Type
(CS)
of Tests
(HS)
Regressions
BC
(q/BHP-hr)
NOx
(q/BHP-hr)
Part.
(q/BHP-hr)
Regressions [b]
IC
(q/BHP-hr)
NOx
(g/BHP-hr)
Part
(q/BHP-1
Mack ETSX-676[c]
EPA
EPA
RTC
4
4
—
~
1,8,10
.65
.55
8.40
8.56
.752
.777
Cummins
VTB-903 [c]
EPA
EPA
KPC
4
4
—
—
•—
1,8,10
1.60
1.27
5.07
5.01
.544
.504
IHC DT-210
IHC[e]
EPA
RTC
1-3
1-3
2-7
2-7
1,2,3,4,
8,9,10,11
.89
.78
7.20
6.80
—
—
—
—
IHC DTI-210
IHC
EPA
RTC
1-3
1-3
.2-7
2-7
1,2,3,4,
8,9,10,11
1.07
• 95 [i]
4.15
4.16
—
—
—
—
IHC OTI-180
IHC
EPA
fox:
1-3
1-3
2-7
2-7
1,2,3,
4,5,6,7,
8,9,10,11
1.18
1.06
4.94
4.73
5,6,7
1.14
1.05
—
—
IK 9.0L
IHC
EPA
PTC
1-3
1-3
2-7
2-7
1,2,3,4,5,
6,7,8,9,
10,11
2.03
1.90
7.18
7.52
5,6,7
2.04
1.90
;;
"
Cumnins
#l[f]
Cumnins
EPA
RTC
2
2
—
—
—
1,8,9,10
.55
.48
7.50
7.46
.46
.43
Cummins
#2
Cummins
EPA
RTC
2
2
—
—
--
1,8,9,10
1.19
.91
8.10
7.92
.66
.66
Cummins
#3
Cummins
EPA
RTC
2
2
—
—
--
1,8,9,10
.87
.63
7.37
7.29
.70
.56
Cumnins
#4
Cummins
EPA
RTC
2
2
—
—
~
1,8,9,10
.94
.67
4.63
5.42
.94
.94
Cat 3208
IHC
EPA
PTC
5
4
12
6
—
--
—
1
1.30
.84
7.68
8.57
.70
.60
-------
Table 3-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Number and
Combined Cold/Hot Start
I
-j
I
Engine/
Test Number[a]
Test Lab
Cycle
Type of
(CS)
Tests
(HS)
Regressions
HC
(q/BHP-hr)
NOx
(q/BHP-hr)
Part.
(q/BHP-hr)
Reqressions[b]
fC
(q/BHP-hr)
NOx
(q/BHP-hr)
Part.
(q/BHP-hr)
Mack #1[g]
Mack
EPA
RTC
lf2[d],8,
9,10,11
.46
•41
5.6
5.9
.51
.46
__
Mack #2
Mack
EPA
KIC
1/2[d],8,
9,10,11
.55
.46
7.8
8.4
.79
.69
—
—
—
Mack #3
Mack
EPA
RIC
If 2 [d] ,8,
9,10
1.10
.87
10.3
9.0
.85
.69
—
—
—
Cat 3208 [c)
IHC
EPA
KIC
5
4
11
6
1,2,5,
8,10,11
1.30
.85
7.59
8.59
.70
.60
5
1.24
.81
—
—
Mack ETSX-
676 [c]
Cat
EPA
FTC
2
1
12
7
1,2,3,
4,5,6,8,10
.73
.65
6.82
7.62
.53
.63
5,6
.74
.64
--
—
IHC DCT-
466B [c]
Cat
EPA
RIC
2
1
15
6
1,2,3,
4,5,8,10
1.00
.90
4.44
4.30
.69
.70
5
.95
.90
—
—
Cat 3208
Cat
EPA
RTC
5
1
11
7
1 [h] ,2,3,
4 [h] ,5,6,7,
8,9,10,11
.97
.88
8.40
8.79
.86
.88
5,6,7
.92
.85
—
—
Cat 3406
Cat
EPA
RTC
3
1
14
6
1,2,3,
4,5,6,7,
8,9,10,11
.49
.40
4.82
5.00
.83
.73
5,6,7
.48
.39
—
—
Cat 3208
Cat
EPA
RIC
2
3
6
8
1,2,3,
4,5,6,7,8,
9,10,11
1.07
.98
9.11
9.24
.854
.712
5,6,7
1.07
.97
—
—
Cat 3208
Model 1
Cat
EPA
RIC
2
5
3
5
1,2,3,4,5,
6,7,8,9,10,11
1.04
.88
14.13
13.96
1.04
.820
5,6,7
1.02
.88
—
--
-------
Table 3-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Engine/
Test Number[a]
Cat 3208
Model 2
Cat 3406
IHC DT-466
(210)
Cat 3406,
Model 1
Cat 3406,
Model 2
Cat 3406,
Model 3
Cuirmins
VTB-903 [c]
DDA 8V-92,
Model 1
Test fr^h Cycle
Number and
Type of Tests
(CS) (HSF
Combined Cold/Hot Start
Cat
Cat
IHC
Cat
Cat
Cat
DDA
DDA
EPA
RFC
EPA
KPC
EPA
KPC
EPA
KIC
EPA
RTC
EPA
RTC
EPA
FTC
EPA
RTC
Regressions
1,2,3,4,5,
6,7,8,9,10,
11
1,2,3,4,
5,6,7,8,9,
10,11
2.5.6.7,
8,9,10,11
1,2,3,4,
5,6,7,8,9
10,11
1,2,3,4,
5,6,7,
8,9,10,11
1,2,3,4,
5.6.7.8,
9,10,11
HC
(g/BHP-hr)
2.70
2.40
.48
.37
1.02
.95
.60
.47
.57
.50
.89
.77
NOx
(q/BHP-hr)
6.38
6.42
7.62
7.26
11.82
11.56
4.03
3.78
4.12
3.64
Part.
(q/BHP-hr)
1.24
.974
.782
.653
Hot Start Only
.726
.601
1.79
1.33
2.20
2.27
Regressions[b]
5,6,7
5,6,7
5,6,7
5,6,7
5,6,7
5,6,7
1,8,10
1,8,9,10
HC
(q/BHP-hr)
2.73
2.38
.45
.35
NOx
(q/BHP-hr)
Part.
(q/BHP-hr)
.98
.94
.52
.43
.53
.49
.82
.74
1.98
1.73
.81
.72
5.07
4.80
4.60
4.26
-------
Table 3-3 (cont'd)
Summary of Bnissions Data EPA Cycle vs. Real Time Cycle
Number and Combined Cold/Hot Start Hot Start Only
Engine/
Test Number[a)
Test Lab
Cycle
Type of Tests
(CS) (HS>
Regressions
HC
(q/BHP-hr)
NCbc
(q/BHP-hr)
Part.
(q/BHP-hr)
Reqress ions[b]
HC
(q/BHP-hr)
NOx
(g/BHP-hr)
Part.
(q/BHR-hr)
DDA 8V-92,
Model 2
DDA
EPA
RFC
3
3
—
—
—
1,8,9,10
.68
.73
8.38
7.69
—
OCA 8V-71TA[C]
Cat
EPA
RFC
2,3,4,5,6,
8,10
.63
.61
—
—
5,6
.63
.62
—
--
IHC 466B[c]
Cummins
EPA
'FTC
—
—
--
8,10
.66
.62
4.01
4.09
.76
.77
DDA 8.2L
DDft
EPA
KIC
3
3
—
—
—
1,8,9,10
1.14
.92
5.78
5.44
—
Currinins
VTB-903
Cummins
EPA
KIC
2,8,9,10,11
1.66
1.37
4.97
5.14
.91
.76
—
—
—
IK DT-466 [c]
EPA
EPA
PTC
—
—
—
8,10
.64
.62
3.53
3.53
.65
.66
DDA 8V-71TA
DDA
EPA
PTC
2,8,9,10,11
.55
.59
6.75
6.96
.35
.33
—
—
--
Mack EISX-676
Mack
EPA
ETC
—
—
—
8,9,10
.78
.71
7.86
7.4
.56
.53
IHC OT-466
IHC
EPA
FTC
2,5,6,7,
8,9,10,11
.75
.63
—
—
5,6,7
.68
.62
—
—
3241
3242
Cunmins
EPA
FTC
1
1
—
—
8,9,10
.88
.69
5.45
5.57
—
3261
3263
Cumnins
EPA
FTC
1
1
—
8,9,10
1.3
2.51[j]
5.90
5.80
—
-------
Engii
it NUI
3301
3302
3321
3322
3341
3342
3391
3393
3413
3414
3461
3463
3501
3502
3531
3532
3612
3613
3641
3642
Test Lab Cycle
Cummins
Cummins
Cummins
Cummins
Currmins
Cummins
Cummins
Cumnins
Cuirmins
Cummins
Number and
Type of Tests
(CS) (HS)
Table 3-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Combined Cold/Hot Start
HC NOx Part.
Regressions (q/BHP-hr) (q/BHP-hr) (q/BHP-hr)
Hot Start Only
Regressions[b]
HC NOx
(q/BHP-hr) (q/BHP-hr)
EPA 1
KDC 1
8,9,10
.89
.68
5.80
5.84
EPA 1
KDC 1
8,9,10
.84
.64
6.56
6.53
EPA 1
RTC 1
8,9,10
.91
.73
6.53
6.42
EPA 1
FTC 1
8,9,10
1.08
.60
6.96
6.63
EPA 1
KIC 1
8,9,10
.84
.64
7.16
6.98
EPA 1
ETC 1
8,9,10
.82
.58
6.68
6.77
EPA 1
RTC 1
8,9,10
.80
.65
5.37
5.66
EPA 1
RTC 1
8,9,10
.83
.72
4.42
4.53
EPA 1
FTC 1
8,9,10
.77
.61
7.40
7.38
EPA 1
FTC 1
8,9,10
.72
.57
7.32
7.28
-------
Engii
it Nu]
3761
3762
3742
3743
3761
3762
3781
3782
3791
3792
3841
3842
3852
3853
3861
3862
3871
3872
3881
3882
3961
3962
Table 3-3 (cont'd)
Sunmary of amissions Data EPA Cycle vs. Fteal Time, Cycle
Test Lab Cycle
Cunmins
Cummins
Cumnins
Cumnins
Number and
Type of Tests
(CS) (HS)
Combined Cold/Hot Start
BC NOx Part.
Regressions (g/BHP-hr) (q/BHP-hr) (g/BHP-hr) Regress ions{jbl
Hot Start Only
HC NOx
(g/BHP-hr) (g/BHP-hr)
Cuirmins
Cuirmins
Cummins
Cuirmins
Cunmins
Cunmins
Cunmins
EPA 1
RFC 1
8,9,10
2.07
1.58
6.13
6.27
EPA 1
FTC 1
8,9,10
.93
.72
6.77
7.01
EPA 1
KIC 1
8,9,10
.89
.65
6.24
6.11
EPA 1
RFC 1
8,9,10
1.58
.98
6.72
6.80
EPA 1
PTC 1
8,9,10
.26
.17
6.05
6.10
EPA 1
RFC 1
8,9,10
.88
.54
7.86
8.11
EPA 1
FTC 1
8,9,10
.85
.67
7.19
7.52
EPA 1
KIC 1
8,9,10
.60
.47
7.22
7.54
EPA 1
KIC 1
8,9,10
.56
.43
7.49
7.54
EPA 1
KIC 1
8,9,10
.65
.53
7.35
7.27
EPA 1
KIC 1
8,9,10
.74
.59
6.55
6.58
-------
Table 3-3 (cont'd)
Summary of Bnissions Data EPA Cycle vs. Real Time Cycle
Number and Combined Cold/Hot Start Hot Start Only
Engine/ Type of Tests fC NQx Part. HC NOx Part.
Tfest Number[a] Test Lab Cycle (CS) (HS) Regressions (g/BHP-hr) (g/BHP-hr) (g/BHP-hr), Regressions[b] (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
4011
4012
Cumnins
EPA 1
KTC 1
8,9,10
.69
.50
6.36
6.38
4031
4032
Cumnins
EPA 1
RIC 1
8,9,10
.65
.53
7.67
7.92
4042
4043
Cummins
EPA 1
RTC 1
8,9,10
1.48
1.18
6.82
6.91
4051
4052
Cunmins
EPA 1
KTC -1
8,9,10
.67
.53
7.43
7.55
4081
4082
Cummins.
EPA 1
KTC 1
8,9,10
.86
.72
6.33
6.44
4245
4246
Cumnins
EPA 1
KIC 1
8,9,10
.83
.66
3.83
4.74
4261
4262
Cummins
EPA 1
KIC 1
8,9,10
.75
.63
6.57
6.90
4331
4332
Cumnins
EPA 1
RTC 1
8,9,10
.77
.68
6.05
6.26
4351
4352
Cunmins
EPA 1
KIC 1
8,9,10
.88
.74
5.91
5.62
4381
4382
Cummins
EPA 1
KIC 1
8,9,10
1.02
.77
6.43
6.59
4401
4102
Cumnins
EPA 1
KDC 1
8,9,10
.76
.53
6.39
6.29
-------
Table 3-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Number and Combined Cold/Hot Start Hot Start Only
Engine/ Type of Tests HC NOx Part. fC NOx Part.
Test Number fa) Test Lab Cycle (CS) (HS) Regressions (g/BHP-hr) (q/BHP-hr) (g/BUP-hr) Regressions [b] (g/BHP-hr) (g/BHP-hr) (q/BHP-hr)
4451
4452
Cummins
EPA 1
KIC 1
8,9,10
.61
.58
6.30
6.23
4521
4522
Cummins
EPA 1
KIC 1
8,9,10
.75
.56
6.47
6.48
4561
4562
Cummins
EPA 1
RTC 1
8,9,10
.72
.63
4.09
4.14
4581
4582
Cummins
EPA 1
rac 1
8,9,10
.66
.56
4.36
4.35
4611
4612
Cumnins
EPA 1
FTC 1
8,9,10
.80
.65
3.93
3.99
3661
3663
Cujtmins
EPA 1
KIC 1
8,9,10
.87
.65
6.13
6.32
4661
4662
Cummins
EPA 1
FTC 1
8,9,10
.79
.56
6.94
6.01
4801
4802
Cummins
Cummins
EPA 1
FTC 1
8,9,10
.86
.69
8.88
9.06
4721
4722
Cummins
Cummins
EPA 1
FTC 1
8,9,10
.91
.68
6.19
6.24
4773
4774
Cummins
Cummins
EPA 1
FTC 1
8,9,10
.72
.59
7.91
7.70
4713
4714
Cummins
Cummins
EPA 1
FTC 1
8,9,10
.98
.76
6.20
5.73
-------
Table 3-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Number and Combined Cold/Hot Start Hot Start Only
Engine/ Type of Tests HC NQx Part. HC NOx Part.
Test Number [a] Test Lab Cycle (CS)
-------
Table 3-3 (cont'd)
Summary of amissions Data EPA Cycle vs. Real Time Cycle
Number and Combined Cold/Hot Start Hot Start Only
Engine/ Type of Tests HC NOx Part. HC NOx Part.
Test Number [a] Test Lab Cycle (CS) (HS) Regressions {g/HHP-hr] (g/EHP-hr) (q/BttP-hr) Regressions[b] (g/BHP-hr) (q/BHP-hr) {g/BHP--hr)
4461
4685
Curtmins
Cummins
EPA
KDC
1
1
— —
8,9,10
.52
.37
4.99
5.08
4694
4695
Cummins
Cunmins
EPA
ETC
1
1
—
8,9,10
.42
.37
4.40
4.42
4697
4696
Cumnins
Cummins
EPA
KDC
1
1
—
8,9,10
.46
.35
4.73
4.82
4861
4864
Cummins
Cummins
EPA
RTC
1
1
—
8,9,10
.53
.52
4.70
5.06
DDR-"A"
DDft
DDA
EPA
RIC
6,9,11
.57
.52
—
.57
.52
—
DDR-"B"
DDA
DDA
EPA
RPC
8,9,11
.49
.43
—
.48
.42
—
Mercedes
CM362IA
MB
MB
EPA
RTC
8,9,11
1.16
1.11
—
1.12
1.09
~
[al Engines are listed per original EPA analysis; more recent data are included at the end of the list.
[b] An explanation of the regressions appears in Table 3-4.
[c] Duplicate engine.
[dj Data changed from hot only to combined per EMA docket submittal, May 13, 1983.
[e] Particulate data not included.
[f] Engine models not specified.
[g] Omissions data derived from plots. Engine models and number and type of tests not specified.
[h] Regression analysis calculated with wrong data (EPA = 1.08, RTC = 1.18).
[ij Changed from .85 to .95 per telephone conversation with IHC on May 16, 1983. The value of .95 was used in Methodologies 4, 5, 8, and 9.
IjJ Spurious point; not used in HC regressions.
-------
FIGURE 3-1
EPA Cycle vs. RTC BSHC Emissions
EPA Cycle BSHC
(g/BHP-hr)
-88-
-------
FIGURE 3-2
RTC vs. EPA Cycle BSNOx Emissions
RTC
BSNOx
(g/BHP-hr}
r = .965
y = . 968x + .223
At EPA = 10.7, RTC = 10.581
6 7 8 9
EPA Cycle BSNOx
(g/BHP-hr)
10 11 12 13 14 15
-89-
-------
FIGURE 3-3
RTC vs. EPA Cycle Particulate Emissions
S.EBr
2.50 ¦
a.aa ¦
RTC
Particulate
(g/BHP-hr) 1 'sa '
i ,«|.
0.S0 -
0.00
r = .830
y = .966x + .0]
0.00 0.50
1.00 1.50 2.00
EPA Cycle Particulate
(g/BHP-hr)
2.50
-90-
-------
Justification for a Standard Adjustment
Givfen the consistent HC offset between the cycles, EPA
believes that there are compelling reasons to adjust the
emission standard. EPA cannot agree with EMA's argument that
the diesel test cycle should not be linked to the level of
emission standards.
First of all, emission standards promulgated on January
21, 1980 were derived using the EPA test cycle. That final
rulemaking established the "baseline" against which all
subsequent actions must be judged. Test procedures and
emission standards are fundamentally related; any significant
change to one without the appropriate change to the other
represents a net change in the stringency of compliance
requirements.
Secondly, the construction of the diesel cycle is not
entirely independent from the standard setting process. The
gasoline test cycle was used to establish the uncontrolled
emission baseline, from which 90 percent reductions were taken
to derive the statutory standards. The absolute emission level
of the baseline was fundamentally determined by the
construction of the gasoline test cycle. Both the EPA gasoline
and diesel cycles were composed of the same subcycles in the
same sequence (NYNF-LANF-LAF-NYNF) . They were both intended to
represent characteristic operation of gasoline and diesel
trucks over comparable road conditions. This comparability
gave EPA confidence that the level of emissions representing
the full 90 percent reductions would be achieved by both
classes of engines. As noted above, however, the RTC cycle
incorporates a NYF segment in lieu of the second NYNF, and some
of the operational comparability between the RTC cycle and the
gasoline engine baseline is lost. Indeed, had such an
operational change been made to the gasoline cycle, EPA is
convinced that both the HC baseline and the statutory HC
standard would be lower. (Gasoline engine "brake specific HC
emission rates are substantially lower on the LAF segment than
on the LANF segment.) For this reason, the construction of the
diesel cycle is not independent of the standard-setting
process; comparability in represented road type between
gasoline and diesel cycles assures that subsequent emission
test results are also comparable.
For the two reasons cited above, EPA does not believe that
the specific diesel engine test cycle is independent of either
the standard-setting process or the level of the standards. A
change in test cycle, therefore, requires an adjustment in
emission standards to maintain equivalent stringency.
-91-
-------
Standard Adjustment Methodology
To determine equivalently stringent standards for the rtC
cycle, EPA evaluated several different methodologies, the
results of which are shown in Table 3-4. EPA's first
evaluation, distributed for public comment in the summer of
1982, used data from 30 engines/configurations. Twenty-one
pairs of the data were combined (cold/hot) results; the
remainder were hot-only results. Nine of the
engines/configurations are "duplicates," (i.e., they represent
the same engines included elsewhere in the data base, but the
additional data come from tests performed at different
laboratories). This analysis yielded an equivalently stringent
HC standard of 1.1 g/BHP-hr (see Table 3-4, Methodology 1).
In early March 1983, Caterpillar recommended another
methodology based upon its evaluation of the original data
base. Caterpillar concluded that only 16 of the 30 data points
used in the EPA evaluation were valid. Caterpillar's
evaluation excluded: 1) the hot-only data, and 2) the
duplicate engines which were not tested in the laboratory of
the engines' manufacturer. Caterpillar excluded the hot-only
data claiming that they were "incomplete" tests. (The Federal
Test Procedure requires the use of combined cold and hot
data.) Caterpillar also excluded duplicate engines from their
analysis, claiming that lab-to-lab sensitivity as well as
cycle-to-cycle sensitivity would be reflected. After omitting
these data points, Caterpillar recommended an RTC equivalent HC
standard of 1.2 g/BHP-hr (see Methodology 3, Table 3-4).
EPA then reviewed Caterpillar's analysis to determine if
the inclusion of hot-only data and duplicate engines had indeed
biased EPA's analysis. (Caterpillar's methodology was accepted
and recommended by EMA on May 13, 1983.) EPA staff first
contacted the manufacturers and requested all additional data
which had been generated since the initial analysis. These
data were incorporated into the data base and Caterpillar's
(and EMA's) two main concerns were evaluated.
The assertion that the inclusion of hot-only data unduly
influenced EPA's analysis was evaluated by directly comparing
hot-only data and combined (cold/hot) data in three linear
regression analyses (see Methodologies 5, 6, and 7, Table
3-4). The comparisons were done only on engines which had both
combined (cold/hot) and hot-only data available. Duplicates
were both included and excluded in separate methodologies.
Use of either methodology on identical engines produced
insignificant differences in the adjusted emission standards
(see Table 3-4). Far more error is induced in the adjusted
standard by excluding the hot-only data than including them,
primarily because their exclusion reduces the size of the data
-92-
-------
Table 3-4
Comparative Methodologies and Results
i
10
OJ
I
Methodology
1. EPA's original
analysis using
data available in
March 1982:
Sample
Size
30
2. All available
combined data
as of March 1982
(duplicates
included): 21
3. Caterpillar's
methodology of
early 1983, using
EPA's original data
(excluding one
erroneous point): 16
4. Caterpillar's exact
methodology of early
1983 (EPA's original
data with one errone-
ous point) : 16
Correlation
Coefficient
(r2)
.943
,970
,993
.983
Regression Regression At EPA = 1.3,
Slope, m Intercept, b EMA =
.873
-.0302
1.105
.884
-.013
1.134
.920
-.039
1.156
.922
-.029
1.170
-------
Table 3-4 (cont'd)
Methodology
Cold/Hot Versus Hot-Only Comparison
Sample
Size
Direct comparison
of combined cold/
hot data vs. hot-
only data (dupli-
cates included):
Correlation
Coefficient
(r2)
Regression
Slope, m
Regression
Intercept, b
At EPA
EMA =
= 1-3,
l
kO
a. Combined data
b. Hot-only data
Same as 5, but all
duplicate engines
excluded if "home"
lab has both com-
bined and hot-only
data:
17
17
,974
,974
,905
,883
-.035
+ .000
1.142
1.148
a. Combined data
b. Hot-only data
Same as 5, but only
"home" lab data used
(all duplicates ex-
cluded) :
a.
b.
Combined data
Hot-only data
15
15
13
13
.995
,994
.996
.995
,922
,897
,928
.899
-.034
+ .002
-.047
-.003
1.164
1.167
1.160
1.166
-------
Table 3-4 (cont'd)
Comparative Methodologies and Results
i
vo
Ul
I
Methodology
Sample
Size
8. All combined data,
plus hot-only data
for engines where
combined data is not
available (duplica-
tes included) . For
all data available
by June 1, 1983.
(EPA1s r ecommended
methodology):
9. Same as Methodology
8 (EPA's recommended
methodology), but dup-
licates excluded
(home lab data
only):
10. EPA's recommended
methodology, but
"sales-weighted r"
using each manufac-
turer's percentage
of total sales, as
shown in Table 3-5:
Correlation
Coefficient
U2)
Regression Regression
Slope, m Intercept, b
Afc^EPA = '"1-.3.V-
EMA =
99
.925
.817
.007
1.069
90
.923
,826
-.0018
1.072
162
.920
.810
,037
1.09
11. EMA's proposed meth-
odology (May, 1983) ,
excluding hot-only data
and duplicates: 23 .988 .901 -.018 1.153
-------
base by almost 70 percent, including the exclusion of all but
one engine of the major manufacturer (Cummins Engine Company).
The importance of the additional data can be seen in Figure
3-4, in which RTC HC equivalent emissions are plotted as a
function of sample size. As the data base increases, the
adjusted standard converges oh 1.1 g/BHP-hr. Again, this may
not so much be an effect of sample size, but more an effect of
the inclusion of Cummins's engines in a more representative
number. in short, the most accurate representation of the
difference between the test cycles is derived from the larger
data base; the starting condition of the engine has been
demonstrated to be unimportant.
The resulting equivalently stringent HC standard using all
available data (Methodology 8) is 1.1 g/BHP-hr. Note that
Methodology 9 excluded duplicate engines but included hot-only
data; the impact of the duplicate engines on the magnitude of
the adjustment is insignificant once hot-only data is
included.
As a final evaluation of the sensitivity of the standard
adjustment to methodology, and to ensure that one
manufacturer's data didn't bias the adjustment, EPA also
performed a "sales-weighted" analysis (see Methodology 10).
Table 3-5 shows each manufacturer's percentage of total sales
and the "weight," (i.e., the number of times added to the
regression) of each manufacturer's engines used in the
analysis. Again, the equivalent HC emission standard was found
to be 1.1 g/BHP-hr.
Based upon the insensitivity of the standard adjustment to
engine starting condition, the best and most representative
data base is that which includes all of the available data.
The emission standards for the RTC are derived by substituting
the EPA standards in linear regression equations derived from
the most representative data base. using the regression
equations from Figures 3-1 and 3-2 (derived using Methodology
8), and EPA standards of 1.3 g/BHP-hr HC and 10.7 g/BHP-hr
nitrogen oxides (NOx), the respective standards for the RTC
would be:
HC: 1.1 g/BHP-hr nOx: 10.6 g/BHP-hr
Cycle selection
in their final comments, the EMA recommended that either
the RTC cycle be adopted with a HC standard of 1.2 g/BHP-hr, or
the EPA cycle be retained with the existing 1.3 g/BHP-hr
standard. in any case, EMA argued, only a single cycle should
be set in place.
-96-
-------
Table 3-5
"Sales Weighted" Regression Analysis
Existing Data:
Number of Engines Factor New Data Base Percent of Total
10 Caterpillar X 2 20 12.4
66 Cummins XI 66 40.7
5 DDA X 8 40 24.7
9 IHC X 2 18 11.1
6 Mack X 3 18 11.1
Total: 162 100%
These percentages correspond roughly to 1979 market
shares, based upon actual production volumes.
-97-
-------
Figure 3-4
Sample Size vs. RTC HC Equivalent Emissions
Sample Size
-98-
-------
With respect to the number of test cycles, EPA concurs
with EMfr's rationale for a single cycle. The use of more than
one cycle should be avoided when possible, since it can create
unnecessary testing and can add unwarranted complexity to the
certification process. With respect to the representativeness
of the^TC cycle, ePA considers the development work done by
caterpillar to be technically sound and to have produced a
thoroughly valid and representative test cycle. On the other
hand, the strong correlation between both cycles increases
EPA's already strong confidence in the ability of its own test
cycle to predict in-use emission reductions {see Appendix,
Chapter 4 of the Transient Test Study).
The issue then boils down to the adjustment of emission
standards. Given the observed difference in emissions between
test cycles, and given that statutory standards with the
existing cycle have already been promulgated, there is no
alternative but to adjust the standards for a change in test
procedure. If the new test cycle represented an increase in
stringency at the sane numerical standard, instead of the
decrease in stringency seen with the RTC, a standard adjustment
would likewise be appropriate. With respect to the magnitude
of the adjustment, EPA has been consistently open in presenting
its methodologies and results, and has been open to industry's
comments. EPA's analysis yields a greater HC adjustment than
EMA's recommended methodology, but in EPA's judgment represents
a more accurate characterization of the average cycle-to-cycle
relationship for the average engine.
EPA has attempted, over time, to reach a consensus with
the EMA on the technical issue of the test cycle. EMA's final
recommendation to EPA is to promulgate a single cycle, either
the EPA cycle at 1,3 g/BHP-hr HC or the RTC cycle at 1.20
g/BHP-hr. For the reasons cited above, EPA can promulgate the
RTC cycle with an HC standard of 1.1 g/BHP-hr.
Conclusion
The EPA cycle will be retained as the single driving cycle
for the certification of 19&4 and later model year HDDEst in
conjunction with the 1.3 g/BHP-hr EC and the 10.7 g/BHP-hr NOx
standards.
B. The WVMA Cycle for Beavy-puty Gasoline Engines
Summary of the issue
The Motor Vehicle Manufacturers Association (MVMA) has
developed an alternative heavy-duty gasoline engine (HDGE)
-99-
-------
driving cycle. The MVMA cycle was developed because of its
concern about the representativeness of the EPA cycle.
Summary of Comments
MVMA and member manufacturers have on several occasions
submitted specific recommendations for the MVMA cycle. A brief
synopsis of events is again appropriate.
In earlier submissions to EPA, Ford Motor Company (Ford),
General Motors Corporation (GM), and the MVMA recommended that
EPA replace its own test cycle with the MVMA cycle. This
position was reiterated in comments made to the Agency in April
of 1982.
EPA's original evaluation (see Appendix, Chapter 6 of the
Transient Test Study) of the MVMA test cycle was distributed
for public comment in the early summer of 1982. That analysis
drew several conclusions about the MVMA cycle. First of all,
the MVMA cycle was shown to correlate well with the EPA cycle.
Secondly, both HC and CO emissions measured over the MVMA cycle
were less than those measured on the EPA cycle, and an
adjustment of emissions standards was recommended. Finally,
the available data base comparing both cycles was small, and
given the undocumented nature of the MVMA cycle's generation,
EPA was cautious in its recommendations. More comparative
testing between cycles was recommended; EPA judged on the basis
of available evidence that the MVMA cycle might perhaps be
acceptable as a test option.
Industry's reaction to EPA's analysis initially disputed
the need for an adjustment of emission standards, but also
agreed with the need for more testing between cycles. The need
for more testing was especially clear at the level of the
statutory HC and CO standards. No comparative data existed at
these low emission levels, creating substantial uncertainty as
to the proper adjustments to the statutory standards.
Since then, EPA and the manufacturers have cooperated in
generating more test data. The original data base of 14
engines/engine configurations has been expanded to 35. The new
data base includes engines of all technologies, ranging from
uncontrolled 1969 baseline engines to catalyst-equipped 1985
prototypes. (The analysis of this data base is presented
below.)
Both EPA and the industry evaluated the new emission
data. In letters to EPA dated on June 10 and June 16, 1983,
the MVMA recommended provision of the MVMA cycle as a test
option (contrary to earlier recommendations). The industry
agreed that a standard adjustment was appropriate, and a
specific standard adjustment methodology was recommended,
-100-
-------
whereby the data base would be split into catalyst and
non-catalyst groupings, and emission standards would be
adjusted from analysis of the appropriate data base.
Analysis of Comments
In this analysis, we address the construction,
representativeness, and relative stringency of the MVMA cycle.
Methodologies for emission standards adjustment are discussed,
as is the selection of a test cycle for certification testing.
Cycle Development
The MVMA HDGE driving cycle was developed because of
industry concerns that the EPA cycle was inadequate in the
following two areas:
1. It was not representative of real world truck
operation.
2. The irregular nature of the cycle could create
inter laboratory correlation problems.
In an attempt to alleviate some of these concerns, MVMA
modified the EPA cycle to obtain a driving cycle which they
felt was more representative and more acceptable. MVMA
established four basic objectives for constructing the modified
test cycle. The modified cycle had to:
1. Maintain the general character of the EPA cycle.
2. Improve the relationship between simultaneous speed,
power, and acceleration.
3. Reduce momentary speed excursions.
4. Reduce excessive throttle manipulations.
To accomplish these objectives, the cycle was simply examined
on a second-by-second basis; using engineering judgment, the
speed and torque specifications were revised where deemed
appropriate. The resulting driving cycle was a smoothed
version of the EPA cycle with a revised synchronization between
speed and torque commands. Technical justification for
specific cycle changes were not submitted or documented by MVMA.
Statistical Analysis
A comparison of overall statistical parameters from the
MVMA cycle, EPA cycle, and the CAPE-21 data base is listed in
Table 3-6. The CAPE-21 statistics are included for comparison
purposes, although the MVMA cycle was not directly derived from
the CAPE-21 data base.
-101-
-------
Table 3-6
Cycle Statistics: MVMA Cycle,
EPA Cycle, CAPE-21 Data Base
Parameter
Torque
Mean (%)
Percent of Cycle Time
Acceleration (%)
Deceleration (%)
Cruise (%)
Motor (%)
Idle (%)
RPM
Mean (%)
Percent of Cycle Time
Acceleration (%)
Deceleration (%)
Cruise (%)
Idle (%)
MVMA EPA PE-21
37 36 34
15 17 15
19 20 16
28 26 28
9 10 13
28 27 28
31 30 31
20 24 20
26 21 26
26 23 26
28 31 28
-102-
-------
As can be seen from the table , the EPA cycle and MVMA
cycle are very similar statistically. There are no major
discrepancies, which is to be expected since the MVMA driving
cycle is directly derived from the EPA cycle. However, data
from engine tests indicate total engine work (BHP-hr) over the
MVMA cycle is about 10 percent higher than on the EPA cycle.
This increase in cycle work is attributable to the
resynchronization of the speed and torque commands. The MVMA
cycle is also less transient than the EPA cycle. The speed and
torque sequences are smoother, and numbers of torque
accelerations have been completely eliminated, thereby reducing
the number of throttle position changes. (This reduces
accelerator pump operation and transient fuel enrichment.)
The MVMA cycle is statistically similar to the EPA cycle,
but not operationally identical.
Test Cycle Correlation Analysis
EPA, Ford, and GM have now tested 3 5 gasoline engine
configurations to compare the MVMA and the EPA cycle. Both
catalyst and non-catalyst configurations have been tested, as
have engines at all levels of emission control. (The emission
data from these tests are summarized in Table 3-7,)
Excellent statistical correlations were observed between
the MVMA cycle and the EPA cycle. The data were split into
non-catalyst and catalyst sets, on which linear regression
analyses were performed. For non-catalyst emissions of HC and
CO, coefficients of determination (r^) values were found to
be .972 and .987, respectively. The r^ values for catalyst
emissions of HC and CO were .915 and .975, respectively. For
the entire data base, the r^ value for HOx emissions was
.974. The above data indicates that in all cases, the
correlation between the test cycles is strong.
In both sets of data, however, MVMA cycle emissions ace
consistently less than those measured on the EPA cycle. These
differences are explainable by the operational differences
between the cycles, (i.e., the MVMA cycle is smoother, and that
the speed and torque commands follow each other more closely on
the MVMA cycle resulting in an increase in . integrated
power-hour). These changes are illustrated graphically in
Figure 3-5 where the same characteristic sections from both
test cycles have been overlaid. The decrease in the transience
of the MVMA cycle results In less movement of the engine
accelerator pump, which would be expected to result in lower HC
and CO emissions. The rephasing of the speed and torque
commands results in different modes of engine operation on the
two test cycles, with fewer events at both lower speed and load
on the MVMA cycle. The observed increase in power-hour over
the MVMA cycle may also explain the decrease in the brake
-103-
-------
V >
1
2
3
4
5
6
7
8
9
10
Table 3-7
Test
Facility Tests
EPA 3C/5H
EPA 2C/4H
EPA 2C/5H
Ford 2C/2H
Ford 1C/1H
Ford 2C/4H
Ford 1C/3H
Ford 1C/1H
GM
GM
1C/1H
1C/1H
MVMA Transient Test Cycle Adjustment Analysis
Emissions Data; q/BHP-hr
EPA Cycle
HC
CO NOx Tests
MVMA Cycle
HC
CO NOx
Comments
6.12 118.4 6.54 2C/5H 4.72 109.4 6.38 1969 GM 4.8L (292 CID) -
original data base
7.64 126.6 7.74 2C/4H 6.49 125.0 7.50 1969 Ford 4.9L (300 CID)
- original data base
8.14 135,5 4.43 2C/5H 7.71 143.2 4.22 1969 GM 5.8L (350 CID) -
orignial data base
2.86 28.4 8.04 1C/1H 2.40 21.7 8.75 1985 prototype Ford 4,9L
(300 CID) - original
data base
2.36 28.9 7.42 1C/1H 1.50 27.8 6.67 1985 prototype Ford 6.1L
(370 CID) - original
data base
2.46 30.5 8.29 2C/6H 1.59 25.7 8.01 1985 prototype Ford 6.1L
(370 CID) - original
data base
3.28 31.3 8.55 1C/3H 1.81 27.7 8.77 1985 portotype Ford 6.1L
(370 CID) - original
data base
2.34 30.6 8.17 1C/1H 1.48 25.5 8.04 1985 prototype Ford 6.lL
(380 CID) - original
data base
1.28 47.9 5.06 1C/1H 1.44 52.4 4.91 1981 GM 7.5L (454 CID) -
original data base
3.12 98.7 5.48 1C/1H 2.89 100.4 4.25 Ibid: less controls -
original data base
-------
Table 3-7 (cont'd)
0
cn
1
11
14
15
16
17
18
19
MVMA Transient Test Cycle Adjustment Analysis
. Emissions Data: g/BHP-hr "
Test
EPA Cycle
No. Facility Tests
1C/1H
MVMA Cycle
HC
CO
NOx Tests
HC
CO
22
GM
12 GM
13 GM
GM
Ford
EPA
Ford
GM
7.45 63.4 6.22 1C/1H
1C/1H 10.06 129.6 5.67 1C/1H
1C/1H 3.33 26.7 8.13 1C/1H 2.70 26.3
1H 1.66 12.3 8.93 1H 1.08 10.8
4C/4H 3.21 34.6 8.09 4C/4H 2.02 27.1
4C/4H
2C/2H
GM 4C/4H
Ford 2C/2H
20 Ford 2C/2H
21 Ford 4C/4H
2H
4.05 30.8 7.04 4C/4H
1.75 3 6.9 7.32 2C/2H
2.23 33.3 7.90 3C/3H
1.84 35.3 5.56 3C/3H
1.70 19.7 5.44 1C/1H
0.47 18.2 5.43 3C/3H
3.30
1.46
1.79
1.28
1.20
0.37
NOx
Comments
52.9 6.47 1981 GM 7.0L (427
-------
Table 3-7 (cont'd)
0
cr>
1
26
27
28
Test
No¦ Facility Tests
23 EPA 3C/3H
24 SwRI 2C/2H
25 EPA
EPA
EPA
EPA
29 GM
30 EPA
2C/2H
1C/1H
2C/2H
5H
1C/1H
2C/2H
31 EPA 1C/2H
32 EPA 2C/2H
MVMA Transient Test Cycle Adjustment Analysis
Emissions Data: g/BHP-hr
EPA Cycle
HC
CO
NOx Tests
0.52 40.8 2.97 3C/3H
0.39
0.78
72.6
.19 2C/2H
4.15 105.5 3.84 1C/1H
2.42 94.5 1.80 2C/2H
4.04 153.6 5.51
2H
.53
5.6 4.44 1C/1H
MVMA Cycle
.89 20.0 9.54 2C/2H
2.49 3 3.3 8.88 1C/2H
1.68 59.4 2.62 2C/2H
HC
CO
NOx
Comments
0.48
5.6 2.50 2C/2H 0.34
.79
45.0 2.63
7.3 2.30
75.3 1.07
19 82 LDV-S/W GM 5.0L (30 5
CID): 260 in3 COC
pelletized catalyst
1975 5.7L (350 CID):
COC/TWC pelletized
catalyst
1982 LDT Ford 5.0L (302
CID): 128 in3 COC/TWC
system
2.59 106.0 3.90 Ibid: without catalysts
1.88 83.8 1.90 Ibid: catalyst system
moved to location behind
muffler
3.57 164.7 4.60 1981 LDV GM 5.8L (350 CID)
TWC system: tested
without catalyst
.41
.61
1.95
1.43
4.7 4.54
18.7 9.18
30.9
55.7
9.21
2 .72
Chevy 350 HD prototype:
2 260 in3 COC pellet
catalysts
1985 prototype Ford 7.5L
(454 CID): 2 150 in3
COC LDT new catalysts
Ibid: without catalysts
1982 LDT Ford 5.0L (454
CID): 128 in3 COC
catalyst
33 EPA
1C/2H 4.74 89.7 4.18 1C/2H 3.84 87.7 3.97 Ibid: without catalyst
-------
Figure 3-5
MVMA Cycle and EPA Cycle Comparison
Time (seconds)
Key
MVMA
l:pa
-------
specific emissions (i.e., more emissions divided by increased
output work). Changing the engine speed at which motoring
(defined as -10 percent maximum engine torque) occurs would
certainly create HC emission differences between the cycles.
Smoothing of the MVMA's cold start cycle may also yield lower
HC emissions, especially for catalyst-equipped engines.
In summary, the MVMA cycle does not yield emissions
equivalent to the EPA driving cycle; it is not equivalently
stringent at the same numerical emission standards for HC and
CO. The MVMA cycle does, however, correlate well with the EPA
cycle for a wide variety of engines. This strong correlation
implies that there is no advantage in using one cycle over the
other to predict in-use emission reductions.
Standard Adjustment Methodology
EPA's review of the available data indicates that
different correlations exist between the test cycles, depending
upon the technology applied to the engine. specifically, the
relationship between the test cycles is affected by the
presence of a catalyst, especially for HC. Given this fact, we
also note the fact that the standards to be adjusted, 1.3/15.5
and 2.5/40.0, represent 100 percent catalyst and 100 percent
non-catalyst technologies, respectively. The most rigorous
technical approach for adjusting the emission standards would
therefore be to split the data base into catalyst and
non-catalyst groupings. The MVMA cycle-based non-catalyst
standards would be obtained from an analysis of only
non-catalyst data. Similarly, the MVMA cycle-based catalyst
standards would be obtained from only the catalyst data.
The non-catalyst data and the resulting linear regression
equations for HC and CO are presented in Figures 3-6 and 3-7.
The non-catalyst analysis is straightforward, and uses all
available non-catalyst data. However, the derivation of
appropriate linear regressions for the catalyst data base
cannot be made without first exercising some engineering
judgment. Emissions observed on these catalyst-equipped
engines lay over a very wide range. (Several of the engines
were light-duty truck engines, with catalysts and air injection
systems ill-designed to control CO emissions over the
heavy-duty test.) Some data lay far enough outside of the
range expected for HDGEs that they should be judged
unrepresentative and excluded from analysis. in addition,
excluding all unrepresentatively high CO data for
catalyst-equipped engines leaves only six representative data
pairs. This is a data base whose small size may raise concern
as to the accuracy of the derived MVMA cycle-based standard.
All in all, five data pairs should be discarded from the
CO analysis as unrepresentative. (Each of the five lies above
-108-
-------
Figure: 3-6
EPA Cycle vs. MVMA Cycle
BSHC Non-Catalyst Emissions
10.01-
B . 0 -
MVMA Cycle
BSHC
(g/BHP-hr)
E . 0 -
H . 0 -
2 . 0
0 . 0
0 . 0
2 . 0
H . 0
r = . 9716
At EPA = 2.5, MVMA = 1.897 (1.9]
E . 0
B . 0
0 . 0
EPA Cycle BSHC
(g/BHP-hr)
-109-
-------
Figure 3-7
EPA Cycle vs. MVMA Cycle
BSCO Non-Catalyst Emissions
X
0 30 S0 30 I 20 ISO
EPA Cycle BSCO
(g/BHP-hr)
-11.0-
-------
40.0 g/BHP-hr, well beyond the range of the 15.5 g/BHP-hr
standard.) similarly, one data pair should be discarded from
the HC analysis. (This data was taken by EPA with the catalyst
relocated behind the muffler in an attempt to characterize the
effects of catalyst location. The HC emissions were well above
the 1.3 g/BHP-hr level, and the engine is not representative of
a typical catalyst-equipped engine.) using the remaining data,
the catalyst regression equations for HC and CO are presented
in Figures 3-8 and 3-9.
Finally, for the adjustment of the NOx standard of 10.7
g/BHP-hr, all 35 data pairs were used. NOx emissions are not
significantly affected by the presence of a catalyst, and it is
not necessary to segregate the data base. This analysis and
its accompanying regression equation are presented in Figure
3-10.
Based upon the 1985 non-catalyst EPA cycle-based standards
of 2.5 g/BHP-hr HC, 4 0.0 g/BHP-hr CO, 10.7 g/BHP-hr NOx, the
1987 EPA cycle-based standards of 1.3/15". 5/10.7, and the
derived regression equations, equivalently stringent standards
for the MVMA cycle are as follows:
HC CO NOX
(g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
198 5 MVMA Standards 1.9
(non-catalyst)
1987 MVMA Standards 1.1
(catalyst)
37.1
14.4
10.6
10.6
EPA is confident in the accuracy of the derived
adjustments for both HC standards (catalyst and non-catalyst),
both NOx standards, and the non-catalyst CO standard. EPA was
initially concerned, however, about the accuracy of the
adjustment for the 1987 CO standard because of the small
sample. Upon reviewing all data, however, EPA is reasonably
confident in its accuracy. For six data pairs included in this
analysis, the offset between cycles is fairly consistent; the
ratios of MVMA cycle CO to EPA cycle CO exhibit a coefficient
of variation of 17.3 percent, but only 4.4 percent if the
single outlier is excluded. in other words, the offset is
repeatable. More significantly, the ratio of the adjusted MVMA
cycle-based standard to the EPA cycle-based standard for
catalyst engines is virtually identical to that observed in the
adjustment of the non-catalyst standard (14.4/15.5 equals .929,
whereas 37.1/40.0 equals .928.) Assuming substantially similar
test cycles, and assuming that the catalyst operates at a
constant oxidation efficiency over the test cycles, this
observation is to be expected. Note that the same observation
-111-
-------
Figure 3-8
EPA Cycle vs. MVMA Cycle
BSHC Catalyst Emissions
2 . 00 r
I . SB ¦
MVMA Cycle
BSHC
(g/BHP-hr)
I . B0 -
0 . SB j-
0 . 00
0 . 00
y = .8374x - .0.195
r2 = .915
At EPA = 1.3, MVMA = 1.06 92 (1.1)
0.SB I.00
EPA Cycle BSHC
(g/BHP-hr)
I . SB
2 . 00
-112-
-------
Figure 3-9
EPA Cycle vs. MVMA Cycle
BSCO Catalyst Emissions
EPA Cycle BSCO
(g/BHP-hr)
-113-
-------
Figure 3-10
EPA Cycle vs. MVMA Cycle
BSNOx (Catalyst and Non-Catalyst)
EPA Cycle BSNOx
(g/BHP-hr)
-114-
-------
is not true for HC: 1.1/1.3 equals .846, whereas 1.9/2.5
equals .760. This is also to be expected, however, because the
catalyst does not maintain a constant efficiency over the test
cycle for HC oxidation. The cold start produces the majority
of HC emissions on catalyst-equipped engines; the HC offset
between the MVMA and EPA cycles on catalyst-equipped engines is
primarily attributable to the offset in HC emissions before
catalyst light-off. (HC emissions on either cycle after
light-off are virtually eliminated by the catalyst.) Primarily
for this reason, the degree of HC adjustment between cycles
differs between catalyst and non-catalyst engines. Most of the
CO emissions, however, for both catalyst and non-catalyst
engines come from high-power, warmed-up operating modes. For
this reason, EPA believes that its assumption of constant
catalyst efficiency in evaluating the CO adjustment is valid,
and that the derived MVMA cycle-based statutory CO standard is
correct. Any error in the adjustment of the statutory CO
standard is likely to be small.
Test Cycle Selection
For the reasons discussed in the RTC analysis, EPA prefers
the use of a single cycle for certification. MVMA, however,
recommended that its cycle be adopted as an option for 1985,
primarily because different member manufacturers have conducted
development work on different cycles. Selection of a single
cycle for 1985 may penalize a manufacturer who has used the
rejected cycle for all development work. In a letter to EPA
dated June 16, 1983, MVMA was unable to identify conditions
under which it would accept a single cycle (uniike EMA's final
recommendation on the EPA/RTC cycle selection). MVMA also did
not specifically recommend which cycle should eventually be
chosen as the single certification cycle beyond 1985, although
it agreed with EPA that a single cycle shculd eventually be
selected.
EPA can appreciate the position a manufacturer would find
itself in if the test cycle on which all its development work
was based was suddenly eliminated. For this reason, EPA can
accept the use of an optional test cycle in 1985, and the
Agency will conduct all its confirmatory testing, SEA testing,
etc., with the specific cycle on which a manufacturer
certifies, provided that the manufacturer certifies its entire
product line on the same cycle. (Required use of a single
cycle by a manufacturer for all its engines will eliminate the
potential for gamesmanship by selecting the "best" cycle for a
specific engine family.) Under these conditions, EPA finds the
use of the MVMA cycle as an optional procedure for 1985 to be
acceptable.
On the other hand, EPA cannot accept the indefinite
provision of two test cycles. Our analyses indicate that
-115-
-------
either cycle would be acceptable, provided that the emission
standards were appropriately adjusted. For this reason, and in
the interest of reaching a technical accommodation with the
industry, EPA has no objection to adopting the MVMA cycle as
the official EPA certification test cycle. It is EPA's
judgment that the MVMA cycle is preferred by the industry.
This adoption is most reasonably made in the 19.87 model year,
when the next major recertification of hdges occurs.
Conclusions
1. Both the EPA and MVMA test cycles will be permitted
for certification in 1985 and, 1986; optional use of either
cycle will be permitted, provided that any single manufacturer
certifies all its engines on the same test. Similarly, all
confirmatory and other regulatory testing will be conducted on
the same cycle on which the manufacturer originally certified.
2. After 1986, all certification and running change
testing (except carryover for non-catalyst engines previously
certified on the EPA cycle) will be conducted on the MVMA
cycle.
3. The following emission standards, as derived by this
analysis, will be used:
1985: EPA Cycle
MVMA Cycle
1987: MVMA Cycle
BSHC
(g/BHP-hr)
2.5
1.9
BSCO
(g/BHP-hr)
40.0
37.1
BSNOX
(g/BHP-hr)
1077
10.6
1.1
14.4
10.6
-116-
-------
References
1. Derived from Comments Submitted to EPA Public Docket
NO. A-81-20.
2. Derived from Comments Submitted to EPA Public Docket
No. A-81-11.
3. "Evaluation of the Federal Test Procedure for Heavy-
Duty Diesel Engines for 1984 and the Development of the Real
Time Test Cycle," W. L. Brown, Jr., Research Report 88-29, File
18967, Caterpillar Tractor Company, June 22, 1981.
4. "Transient Cycle Arrangement for Heavy-Duty Engine
and Chassis Emission Testing," Chester J. France, EPA Report
HDV 78-04, August 1978.
5. MVMA-Modified Heavy-Duty Gasoline Engine Transient
Emission Test Cycle, Attachment, Letter to EPA Administrator,
Motor Vehicle Manufacturers Association, February 15, 1982 (see
EPA Public Docket No. A-81-11, IV-D-2 and IV-D-2a).
-117-
-------
4. Issue: Environmental Impact
Summary of the Issue
The impact of this rulemaking on the nation's air quality-
was a topic of substantial comment. Many commenters were
critical of the revised rule on the grounds that it would hot
lead to the maximum possible air quality improvements, while
others criticized it as being more stringent than is necessary.
Summary of the Comments
Comments arguing that this rule fails to force the maximum
achievable air quality benefits were received from the
following individuals and organizations: Senator Gary Hart of
Colorado; Frances J. Scherer of New York, a private citizen;
the National League of Women Voters (LWV); the Manufacturers of
Emission Controls Association (MECA), an industry trade group;
the Natural Resources Defense Council (NRDC); the Regional Air
Pollution Control Association (RAPCA) of Dayton, Ohio; and the
Western New York Allergy and Ecology Association (WNYAEA).
These state and local LWV affiliates alsb submitted comments:
Michigan; Carson City, Nevada; New York City; and Doylestown,
Pennsylvania.
Those maintaining that the revised rule is still
unnecessarily stringent from the standpoint of achieving the
desired improvements in air quality were all manufacturers.
The comments of Ford Motor Company (Ford), General Motors
Corporation (GM), International Harvester (IH), and Mack Truck
(Mack) are summarized after those of the commenters listed
above.
All of the commenters in the former group (opposing
relaxation) maintained that these revisions to the light-duty
truck (LDT) and heavy-duty engine (HDE) emission rules pose a
threat to the public health and welfare. Citing figures from
the December 1979 EPA Regulatory Analysis projecting average
improvement of 7 percent in 1995 for carbon monoxide (CO),
Senator Hart noted that for cities with very high CO levels
such as Denver and Los Angeles, this difference could
"...determine whether or when the ambient air quality standards
will be achieved." The LWV, MECA, NRDC, and RAPCA all cited
this figure, and the 2 percent average improvement for ozone in
1995 projected in the same document, to argue that air quality
improvements of that magnitude are necessary if areas currently
in nonattainment status for either pollutant are to be brought
into compliance.
-118-
-------
The NRDC, MECA, and LWV all noted that control of
hydrocarbon (HC) and CO emissions from HDEs has previously been
found to be cost effective, and that the control technology
necessary to meet the statutory standards is available.
Emissions from HDEs have not been reduced to nearly the extent
that light-duty vehicle (LDV) emissions have, NRDC and LWV
stated; thus, HDEs have not borne their proportionate share of
mobile source emissions reductions and associated costs to date.
MECA listed some of the air quality problems foreseen by
the Association of State and Territorial Air Pollution
Administrators (ASTAPA) "...if auto and truck emission
standards are relaxed." Citing To Breathe Ciean Air, the 1981
final report of the National Commission on Air Quality (NCAQ),
and EPA-supplied data, MECA noted that violations of the
National Ambient Air Quality Standard (NAAQS) for ozone are
projected to occur through at least 1995 and stated that "...it
is generally agreed that if the ozone air quality standard is
ever to be achieved all feasible and reasonable hydrocarbon
controls will be needed."
MECA indicated that even if it were concluded, contrary to
"clear and compelling evidence," that adequate control of HDE
emissions of HC and CO could be achieved without the use of
catalysts, catalyst technology should still be implemented.
Catalysts offer "an attractive answer to [future] gasoline-
truck NOx control," MECA stated, and rejection of catalysts for
HC and CO control at this time will make it more difficult to
implement such technology in the future.
RAPCA was critical of the lack of detailed air quality
analysis data included in the Federal Register publication of
this rulemaking, stating that the information provided "...is
so sparse as to make it virtually impossible to determine the
impact of the anticipated emission increases on the Dayton
Region." Since the Dayton area is currently operating under a
nonattainment State Implementation Plan (SIP> for ozone, RAPCA
finds it difficult to accept EPA's ".. .conclusory assertion of
no impact." RAPCA also called it unseemly for EPA to propose
"a large increase in truck emissions" and assert that the
impact on air quality will be small, while simultaneously
"restricting access" to the detailed information (air quality
modelings) necessary for independent evaluation of EPA's
conclusions.
In addition to their concerns over the ozone and CO air
quality impacts of this rule, LWV expressed reservations about
anticipated increases in lead emissions from HDEs as compared
to the original 1984 rulemaking. They contended that the
relaxation of the HDE emission standards to non-catalyst levels
will increase lead emissions both directly, through continued
-119-
-------
HDE lead emissions which would have been eliminated under the
originally promulgated standards, and indirectly, through
extension of a legitimate source of demand for leaded gasoline
at the retail level, thereby extending the opportunity for
misfueling of catalyst-equipped LDVs and LDTs.
NRDC documented its opposition to this rule using many
quotes and figures taken from EPA's December 1979 Regulatory
Analysis. In addition, pertinent quotes were taken from House
Report No. 95-294 (95th Congress, ist Session, 1977), the NCAQ
final report, other reports by the National Academy of Science,
the Library of Congress, and the New York City Department of
Air Resources, a study conducted by the Jet Propulsion
Laboratory, and former Senator Edmund Muskie, floor manager for
the 1977 Clean Air Act (the Act) amendments* All of these
stressed the need for further control of HDE emissions.
Finally, NRDC also claimed that the air quality impacts
calculated by EPA and included in the September 1981 Draft
Regulatory Support Document are significantly understated when
the deterioration factors (DFs) contained in EPA's January 15,
1982 response to questions from Senator Robert Stafford of
Vermont on motor vehicle emission standards are used. NRDC
claimed that EPA used different (and lower) DFs in the
Regulatory Support Document. On this basis, NRDC urged
"...that EPA reanalyze the air quality impactst impacts on
nonattainment status, and impacts on the number of exceedances
using the more recent deterioration factors submitted to
Congress."
The remaining comments concerning the air quality impacts
of this rule are those of the manufacturers. All felt that the
rule, even as revised, is unnecessarily stringent for
attainment of the air quality benefits sought. Several
different bases for this position were advanced.
Ford and Mack both questioned the need to control HC
emissions from HDEs to the extent required in the rule, on the
grounds that all areas exceeding the NAAQS for ozone are urban,
while much of the HC from heavy-duty trucks (HDTs) is emitted
in rural areas. Ford stated that approximately half of all HDT
vehicle miles travelled (VMT)'are in rural areas, and that the
air quality impacts for ozone and CO projected by EPA are
therefore approximately twice the magnitude of the actual
impacts. Mack quoted the Department of Commerce 1977 Truck
Inventory and Use Survey, which showed that only 22 percent of
the VMT of Class VIII heavy-duty diesels (HDDs) are accumulated
in urban areas. On this basis Mack, which manufactures only
heavy-duty diesel engines (HDDEs) for Class VIII applications,
-120-
-------
stated that EPA must consider the fraction of all HDT VMT that
are accumulated in urban areas when performing air quality
analyses.
Mack also accused EPA of failing to use the findings of
the August 1980 pollutant-specific study (?SS) for HC when
setting the standards for HDDEs. Its argument can be
summarized as follows: The 57 areas in violation of the NAAQS
for ozone are all urban areas. According to the PSS, in 1999
HDDEs will be contributing only 4.7 percent of total HC
emissions in those 57 areas. (Mack added that a report by
Southwest Research Institute (SwRI) shows that this HDDE
fraction of total HC will be only 3.6 percent in 1999.) The
SwRI report also said that i>377 million would be spent during
the 1990s on control of HC from HDEs, in order to bring "only
one or two AQCRs" into compliance. Since further control of HC
from HDDEs is "obviously" not cost effective based on this
information, Mack concluded, EPA did not use the findings of
the PSS in setting the standards. In failing to do so, Mack
claimed that EPA has "...overlooked a very important and
significant issue."
IH recommended that EPA perform a complete reanalysis of
all air quality and cost/benefit questions, taking into account
two factors that they maintained were not considered. The
first of these dealt with the multiplier that EPA used to
convert 1979 certification HC emission rates to equivalent 1984
transient cycle HC emission rates for HDDEs. In the December
1979 Regulatory Analysis, EPA used a multiplying factor of 2.4
to make this conversion. IH states that their testing and that
of other manufacturers indicates that the value of this
multiplier should have been 1.3; therefore, EPA overestimated
pre-1984 HDDE HC emissions by a factor of 1.8 (2.4/1.3).
According to IH, EPA also used "unrealistic estimates of
the trend to diesels in the heavy-duty market" in the
Regulatory Support Document. By underestimating the magnitude
of the shift to diesels in the 1980s and overestimating the
level of HC emissions from 1979-83 HDDEs, IH argued, EPA has
based its ambient air quality arguments for the transient test
and emission standards on faulty assumptions. IH maintains
that the air quality benefits intended to result from this rule
will "by and large" be accomplished through continued diesel
penetration of the HDE market in the 1980s.
In arguing for their proposed HDE emission standards, Ford
also made reference to the latest projections for diesel
penetration of the heavy-duty market in the 1980s. Ford then
described the results of their own air quality analyses in
which the impact of the standards being set at 3.3 HC/42 CO,
-121-
-------
rather than 2.5 HC/35 CO as specified in the proposal, is seen
to be quite small. These impacts are given as "considerably
less than one percent" foregone improvement for ozone as of the
year 2000 and "one percent or less" foregone improvement for CO
as of 1995.
General Motors criticized the estimates of HDE fuel
economy (FE) that EPA used, arguing that they may have been too
low by as much as a factor of two and, that as a result, HDE
emissions and their contribution to overall air quality may be
overstated by a factor of two. General Motors noted that EPA
used FE estimates of 5.0 miles per gallon (mpg) for heavy-duty
gasoline engines (HDGEs) and 5.8 mpg for HDDEs, derived from
the EPA transient HDE test cycles, which in turn were based on
CAPE-21 survey data. They argued that the survey did not
include any heavy-duty gasoline vehicles (HDGVs) from Class IIB
(8,501-10,000 lbs. gross vehicle weight (GVW)), which are the
largest subset of all HDGVs and have average fuel economy of
considerably more than 5.0 mpg. In addition, EPA assumed that
these FE values would be constant throughout the projection
period; actually these values are expected to rise
significantly during the 1980s, GM said, partly due to
increasing diesel penetration of the lower-GVW heavy-duty
classes.
Most of the comments made by GM sought to minimize the
significance of this rule to national air quality. GM argued,
for example, that "...clearly, any HC standard more stringent
than the 1979-83 HDE standard would be adequate to avoid
significant effect on urban air quality." On the basis of the
"negligible" and "insignificant" air quality improvements
projected, GM maintained: 1) that a Selective Enforcement
Audit program for HDEs cannot be justified; 2) that extended
useful-life requirements are unnecessary; and 3) that the 1984
LDT requirements will have no significant impact on air quality
violations. GM concluded that HDE standards of 3.5 HC/70 CO
will "allow early attainment of the NAAQS in even the worst
areas of the country," while having minimal cost impact,
eliminating the need for overtemperature protection controls,
and not imposing any fuel penalty.
Analysis of the Comments
The subissues raised by the commenters are discussed in
this section in roughly the same order as they were presented
in the summary section.
At the outset, it is important to, keep in mind the
statutory authority and Congressional guidelines for this
rulemaking. The emission standards are being revised
-122-
-------
principally under the authority of Sections 202(a)(3)(B) and
(C) . Although EPA has evaluated the air quality effects of
this rulemaking, the revised standards are based on findings
concerning cost, technology and leadtime, as explained in the
preamble and elsewhere in this document. Congress has
specified that revised standards provide for "the maximum
degree of emission reduction which can be achieved by means
reasonably expected to be available" for the duration of the
revised standards, set against an ultimate Congressional goal
of 90 percent emission reductions, also established in these
rules for lighter HDEs.
Thus, the comments on both sidies of the issue of the
appropriate degree of air quality protection are somewhat
misplaced. Although the air quality effects of these rules are
important, air quality considerations are not the driving force
behind the amendments.*
In the December 1979 Regulatory Analysis, EPA projected
average air quality improvements in 1995 of 7 percent (CO) and
2 percent (ozone) . Commenters noted that improvements, of this
magnitude are very important for areas that exceed or just meet
the NAAQS for either pollutant, and could be the deciding
factor in whether and when cities with vfery high CO levels
reach attainment of the standard. EPA concurs with the
importance of HDE emission reductions to such areas, and notes
that air quality improvements of 5 percent (CO) and 1 percent
(dzone) are still projected in 2000 as a result of this rule.
EPA considers the HDE emission standards being promulgated in
this action to be the most stringent reasonably available at
this time, taking into consideration such issues as leadtime,
cost effectiveness, technological feasibility, and fuel economy
effects. These factors are dealt with in more detail in other
sections of this document.
Commenters indicated that further emission controls for
HDEs have been shown to be cost effective and technologically
feasible, and that the statutory standards mandated in the 1977
amendments to the Act can be achieved. It was stated that the
1979-83 HDE emission standards are too lenient; the sharp
* As described elsewhere, air quality effects do bear on
related portions of these rules. For example, EPA selected a
modified full-life useful-life requirement over a half-life
approach based in part on the Agency's determination that the
former approach helps assure that the full air quality benefits
of these rules will be realized. So, too, EPA in allowing the
use of manufacturers' test cycles has adjusted standards to
assure that the air quality benefits of the previously
promulgated standards will not be compromised.
-123-
-------
contrast in the degrees of emission control required of
light-duty vehicles (LDVs) and LDTs, and of HDEs in the same
time period, was cited. EPA agrees that the current
discrepancy in HDE and LDV/LDT emission control requirements is
inequitable in the long run, and that significant reductions in
HDE emissions may be achieved at reasonable cost. This rule
substantially reduces the inequality in the stringency of
light-duty and heavy-duty emission control requirements, and
results in lifetime per-vehicle emission reductions for HDGEs
of 0.25 tons HC and 16.42 tons CO (representing reductions of
39.5 and 73.6 percent, respectively, from model years 1979-83
lifetime emission levels).
According to MECA the use of catalytic converters on HDGEs
should be required, even if the HC and CO emission standards
are set at levels that would not require the use of catalysts,
because substantial fuel economy gains could be realized and
catalysts provide an attractive method for future NOx control
from HDGEs. The LWV said that requiring catalysts on HDGEs
would decrease future misfueling, while MECA said it would
allow more stringent future NOx control and increase fuel
economy. These points are acknowledged, but it should be noted
that EPA does not specify what emission control technology
should be used to meet any emission standards. In addition,
the revisions in the HC and CO standards for HDGEs contained in
this rulemaking for 1985-86 are only temporary. Catalysts will
be used on the majority of HDGEs beginning in 1987 to meet the
statutory standards.
The concerns of ASTAPA as outlined in the MECA comments
are understandable; especially if, as stated in the comments,
both auto and truck emission standards were being relaxed.
This rule has no bearing on LDV emission standards or test
procedures; and while the HDE standards are being temporarily
revised, the standards contained in this rule are still
considerably more stringent than those in effect for model
years 1979-83. The air quality analyses, which are discussed
in detail in Chapter 2 of the Regulatory Support Document, show
that the "worst-case" fears of ASTAPA are unfounded.
After receiving the RAPCA comments, which criticized the
lack of detailed air quality data included in the Federal
Register Notice of this rulemaking, EPA immediately provided
them with copies of all air quality analyses. Since the period
for public comment on this rule was subsequently extended for
21 days, RAPCA had the opportunity to comment further after
receipt of those analyses.
The National Resources Defense Council's claim that EPA
used two different sets of deterioration factors, one shown in
-124-
-------
the September 1981 Draft Regulatory Support Document and the
other submitted in response to Senator Stafford's questions, is
erroneous. NRDC apparently took deterioration factor (DF) and
deterioration rate (DR) to be synonymous. They are not.
Senator Stafford asked for HDG vehicle DRs, while the
calculation of lifetime per-vehicle emissions in the Draft
Regulatory Support Document used DFs.
Deterioration rates are based on testing of vehicles in
the field, while DFs are derived from manufacturers'
certification data. Both quantities attempt to describe the
deterioration in the emissions of a vehicle or engine.
However, the DR accounts for many in-use causes of
deterioration not accounted for in the DF, including causes
that might not be directly within a manufacturer's control.
In-use deterioration (i.e., the DR) includes the effects of
climatic extremes, inadequate maintenance, and tampering and
abuse, as well as the normal wear and tear which the DF is
supposed to represent. Thus, as the comment by NRDC reflected,
the DR and the DF can be and usually are very different numbers.
The Agency's air quality model uses DRs, not DFs, to model
the deterioration of emission levels with increasing mileage.
The DFs are used in the calculation of the zero-mile emission
levels (ZMs) , and thus are only used indirectly by the model.
The air quality analyses used in the September 1981 Draft
Regulatory Support Document, and in the final Regulatory
Support Document which accompanies this Final Rule, used the
DRs submitted to Senator Stafford. The model's outputs
(including the number of urban areas in violation, the number
of exceedances, the average percent reduction, and the
inventory in tons of pollutant) therefore result from the use
of DRs, not DFs. The only instance where DFs were used was in
the calculation of the per-vehicle lifetime emissions for the
Draft Regulatory Support Document. DRs can also be used to
calculate the per-vehicle lifetime emissions, and are more
accurate if absolute numbers are desired. However, the
calculation in the Draft Regulatory Support Document used DFs
because the focus was on relative numbers, that is, on the
differences between the scenarios rather than the absolute
number of tons under each scenario. These relative numbers
using either a DR or a DF calculation are about the same. The
final Regulatory Support Document uses the DR calculation,
since the focus is on the absolute, as well as the relative
numbers.
The remainder of this section is devoted to discussion of
the comments submitted by the manufacturers. Several of these
comments raised issues that, while valid points of concern, are
beyond the immediate issue (the air quality impact of this
-125-
-------
rulemaking) and are not subject to quick or simple resolution.
Neither are they considered by EPA to be of sufficient
magnitude as to affect the rulemaking decision process. Such
issues include the urban/rural VMT split cited by Mack and
Ford, and the criticism by GM of the heavy-duty fuel economy
estimates and the representativeness of the CAPE-21 data base.
EPA allows that improvements in the accuracy of the air quality
projections are possible. However, dealing thoroughly and
appropriately with questions such as those mentioned above is
not a trivial exercise. EPA is concerned with improving the
accuracy of the air quality model and the assumptions that go
into it, and efforts to do so will continue in the future.
However, at this time EPA notes that there is no reason to
consider rural HC emissions to be unimportant. Ozone is a
regional pollutant, and in the time that HC emissions are
reacting to form ozorie, they could travel a considerable
distance from their original emission points.
Mack cited the August 1980 pollutant-specific HC study to
argue that further HDE HC emission controls are not cost
effective. While cost effectiveness may to some extent be
relative, EPA feels that the cost effectiveness of further HDE
emission control has been demonstrated to be good, as described
in a recent EPA staff paper.[1] The staff paper analysis
considers EPA's best estimates of. both costs of this action and
its associated air quality benefits in arriving at that
conclusion. The point raised by Mack about the relatively
small contribution from HDDEs could be equally applied to many
other HC emission source categories and lead to the erroneous
conclusion that none of these sources need be controlled
according to Mack's logic. Since HC emissions include a large
number of relatively small sources, it is important if progress
is to be made to control HC emissions wherever that can be done
in a cost-effective manner.
EPA rejects iH's contention that the multiplying factor
used to convert 1979 certification HDDE HC emission data to
equivalent 1984 transient test emission data should have been
1.3, and not 2.4 as used by EPA. The value of 2.4 used by EPA
was based on the results of tests of HDDEs manufactured by
Caterpillar, Cummins, and Detroit. Diesel Allison (GM). While
EPA acknowledges that these tests were conducted several years
ago and that considerable additional testing has since been
conducted, it is also noted that IH did not submit any new data
to support their claim. The value of 2.4 is intended to be
representative of the heavy-duty industry as a whole; thus it
is entirely possible that the 1.3 value may be more accurate
for IH engines alone, for example. Basically, this comment is
unrelated to the air quality impact of this rule.
-126-
-------
In response to claims by the manufacturers that EPA used
unrealistically low projections of diesel penetration of the
heavy-duty market in the 1980's, thereby overestimating the
contribution of the heavy-duty fleet to overall air quality,
EPA notes several things. First, a major premise of this
argument is invalid. Stating that HDDEs will always have lower
lifetime HC emissions than HDGEs, given the same standard and
useful life for both engine types, is simply not true. The
zero-mile emission rate is higher for HDDEs than for HDGEs
under the same standard. Our calculations show that, under the
same standard, HDDEs will emit more HC than will HDGEs over the
lifetime of the engine. It is only the fact that the HDDE HC
standard will be lower than the HDGE HC standard for 1985-87
that makes HDDEs "cleaner." Second, EPA acknowledges that the
nature of the heavy-duty market has changed somewhat since the
original Regulatory Analysis was published. However, as was
noted by IH in their discussion of this point, there are
factors (such as sudden changes in fuel costs) that can cause
the rate of diesel penetration of the heavy-duty market to
change dramatically in a short time; thus, any projections,
whether by EPA or the manufacturers, are at best educated
guesses and subject to quickly being overtaken by events.
Finally, in terms of air quality the crucial estimate is
the relative change in HDDE and HDGE VMT, not the changes in
vehicle registrations or HDE market shares. EPA assumed that
HDDE VMT would increase by 5 percent annually, while HDGE VMT
would decrease by 2 percent annually during the same time
period. These estimates still appear reasonable.
Ford's air quality projections, showing almost no decrease
in air quality if the standards of 3.3 HC/42 CO advocated by
Ford are implemented rather than the 2.5 HC/35 CO specified in
this rule, appear to be valid. EPA simply notes that small,
incremental relaxations in emission standards will, by
definition, result in relatively small air quality impacts.
Extending Ford's line of reasoning, any and all emission
standards could be discarded incrementally, since each
incremental air quality impact would be minimal. Given both
the current and projected future need for improvements in ozone
ambient air quality, EPA must reject Ford's approach. All
reasonably attainable HC control is important in terms of air
quality. The proposal by GM that interim standards be set at
3.5 HC/70 CO must also be dismissed on the same grounds, since
this proposal represents virtually no reduction from the
1979-83 emission levels. Finally, EPA also notes that
§202(a)(3)(B) of the Act requires that when interim emission
standards less stringent than those mandated are implemented,
those interim standards represent "...the maximum degree of
emission reduction which can be achieved by means reasonably
-127-
-------
expected to be available for production." Both the Ford and GM
proposals are inconsistent with this requirement of the Act.
Conclusions
Although EPA recognizes the concerns of those opposed to
any revisions to the HDE gaseous emission standards, arid also
recognizes that some of the points raised by the manufacturers
merit further study, EPA concludes that the Agency is acting
within its legislative authority in promulgating both the
interim non-catalyst emission standards and the long-term
reductions in these rules. Further changes to this rule are
not justified on the basis of these comments.
-128-
-------
References
1. "Issue Analysis - Final. Heavy-Duty Engine HC and CO
Standards," Staff Paper, U.S. EPA, OMS, OANR, ECTD, SDSB, March
1983.
-129-
-------
B. Secondary Issues
1• Issue: Deterioration Factors
Summary of the Issue
In the original FRM, EPA finalized provisions for the
application of multiplicative deterioration factors to HDE
exhaust emissions. Commenters opposed this change from the
previously used additive deterioration factors. Comments were
also received indicating that negative deterioration factors
should be accepted by EPA. One comment concerned the methods
used to determine deterioration factors.
Summary of the Comments
Most of the comments criticized the application of
multiplicative deterioration factors to HDEs. These comments
argued that no justification exists for the use of
multiplicative deterioration factors with non-catalyst emission
control systems. It was also noted that, in the opinion of the
commenters, multiplicative deterioration factors effectively
increase the stringency of the applicable emission standards,
particularly those at low numerical values, thereby increasing
the control system development costs to the manufacturers.
Comments requesting that EPA recognize the validity of
negative deterioration factors were received from EMA, with
supporting data on HDEs being provided by several
manufacturers. These data show that some HDE exhaust
emissions, particularly NOx, may actually decrease over the
useful life of the engine. One engine manufacturer also
submitted data which it claimed demonstrated that properly
maintained HDEs have no significant deterioration in emissions
during useful life, and that emissions have been observed to
decrease in some cases.
One comment was made concerning the method used to
determine deterioration factors. The commenter maintained that
deterioration factors should be determined through 1,000-hour
durability runs per §86.082-28(c)(4) , and not by fleet tests
with uncontrolled parameters.
Each of the comments concerning the application of
multiplicative deterioration factors to HDEs cited the lack of
justification for extending the use of multiplicative
deterioration factors to vehicles and engines using
non-catalyst emission control systems. A few also noted that
the use of multiplicative deterioration factors cannot be
justified now on the basis of possible regulations implementing
trap-oxidizer technology for HDDEs in the future.
-130-
-------
Analysis ^Qf the Comments
EPAs analysis in support of the original FRM[1] did not
provide conclusive evidence that one type of deterioration
factor was more appropriate than the other for engines without
aftertreatment devices. That analysis did conclude that
multiplicative deterioration factors are more representative of
actual emission deterioration when aftertreatment technology is
used, however, since such technology reduces emissions on a
proportional basis. Therefore the use of multiplicative
deterioration factors should still be required for vehicles or
engines utilizing aftertreatment technology.
The possibility that durability testing of a vehicle or
engine may result in an additive deterioration factor less than
zero, or a multiplicative deterioration factor less than one,
is recognized. However, at least for HC and CO, EPA views such
results as anomalous and clearly not indicative of actual
in-use deterioration. At best, a well-maintained engine could
be expected to exhibit stable emission levels; there is no
mechanical reason for in-use HC or CO emissions to decrease
with accumulated time or mileage. In addition, accepting such
deterioration factors would allow relaxation of low-mileage
target levels to values above those otherwise required for
compliance at low mileages. This would be incompatible with
the purpose behind the use of deterioration factors in
certification, which is to estimate the highest emission level
a vehicle is expected to exhibit over its life so that
compliance is assured on that basis. If emissions were
expected to decline with mileage or time, -hen the level of
concern for certification purposes would be the unadjusted
new-vehicle level. Thus, EPA feels that -he current rule,
under which an additive deterioration factor of less than zero
is considered to be zero and a multiplicative deterioration
factor of less than one is considered to be or.e, is justified.
The comment concerning methods of determining
deterioration factors can be addressed quite briefly. The
determination of deterioration factors is entirely the
responsibility of the manufacturer; within certain constraints,
so are the methods and procedures used in the determination.
Section 86.082-28(c)(4) does not specify that 1,000-hr
durability runs or fleet tests by used to determine
deterioration factors, but refers only to "...deterioration
factors, determined from tests of engines, subsystems, or
components conducted by the manufacturer."
Conclusions
While studies[l] have been inconclusive regarding the
appropriateness of multiplicative deterioration factors for
-131-
-------
non-aftertreatment vehicles and engines, they are clearly more
accurate in describing deterioration in the performance of
proportional-reduction devices. Thus, EPA has decided to delay
the required use of multiplicative deterioration factors for
HDEs until such time as more stringent emission standards
requiring the use of catalysts (for HDGEs) or particulate traps
(for HDDEs) are established and implemented. The first use of
multiplicative deterioration factors will then be for lighter
HDGEs certifying to the statutory standards in 1987. For
reasons cited in the analysis, EPA also has decided that
additive deterioration factors of less than zero and
multiplicative deterioration factors of less than one will
continue to be taken as equal to zero and one, respectively.
-132-
-------
References
1. "Summary and Analysis of Comments to. the NPRM: 1983
and Later Model Year Heavy-Duty Engines, Proposed Gaseous
Emission Regulations," U.S. EPA, OANR, OMS, ECTD, SDSB,
December 1979.
-133-
-------
2. Issue; Idle CO Test and Standards
Summary of the Issue
In the original FRM, EPA finalized a separate standard and
test procedure for idle CO emissions from gasoline-powered LDTs
and HDEs. The commenters on this issue were unanimous in their
opposition to these requirements. Much of the criticism
stressed the allegedly redundant nature of the test and the
planned use of the test by EPA to detect failed catalysts and
set lower I/M cutpoints. One manufacturer claimed that the
idle test requirements will force it to include additional
hardware on its LDTs. Several procedural and technical
questions were also raised.
Summary of the Comments
The manufacturers commenting on this issue all criticized
the idle CO test as redundant, unnecessary, and unjustified.
Several claimed that the 26.8 percent of the transient
certification test spent idling guarantees that idle CO
emissions must be closely controlled in order to pass the
entire test. The comments indicated that the added cost and
complexity of certification including the idle test would thus
be an unnecessary burden on the manufacturers.
EPA was also criticized for planning to use data from idle
CO tests in the detection of failed catalysts and the
establishment of lower I/M cutpoints. One comment specifically
cautioned EPA to "avoid the belief that idle CO measurements
would be a viable method of in-service compliance checking."
Several commenters indicated that EPA cannot promulgate
the idle CO test without demonstrating that a reasonable
correlation exists between the idle test and the other required
CO measurements (transient cycle and performance-warranty short
test). According to the commenters, this correlation is
required under Sections 206 and 207(b) of the Clean Air Act and
has not been demonstrated.
The numerical level of the standard was criticized in
several of the comments. One manufacturer criticized the
standard as infeasible and said that it should be revised
upward to reflect non-catalyst technology, while another
indicated that the dry volumetric measurements used make the
same numerical standard more stringent for smaller-engine
vehicles. Volkswagen questioned the authority of EPA under
Section 202(a)(1) of the Clean Air Act to implement the same
numerical standard across the entire LDT class, noting that it
would be forced to install "new systems" on its LDTs less than
6,000 lb. GVWR simply because of the idle CO standard.
-134-
-------
Volkswagen submitted data on idle emission characteristics
relative to FTP results from two of its light-duty pick-up
trucks. While these vehicles met the current FTP standards for
HC, CO, and NOx emissions, the volumetric tailpipe idle CO
measurements were between 1.3 and 1.4 percent. VW claimed that
these data show that it would be forced to install new systems
on these trucks solely because of the idle requirements. VW
also noted that these vehicles would otherwise be able to meet
the emission standards promulgated for 1984 and later LDTs with
minor calibration changes.
In a follow-up conversation between EPA and VW staff, the
possibility of adjusting the idle A/F mixture to a leaner
setting in order to reduce idle CO was discussed. VW expressed
concern that leaning the idle A/F ratio, combined with the
possibility of in-use drift of this setting, could result in
engine stalling problems. Should leaning of the idle A/F mix
either fail to bring idle CO under the standard, or result in
unacceptable driveability problems, VW stated it would be
forced either to install a "new system" (air pump) in its LDTs
or to go to a closed-loop system. VW indicated that it would
prefer the closed-loop solution *
Finally, several minor issues were addressed in the
comments: the inclusion of the idle CO test in SEA testing,
the applicability of DFs to idle emission data, and the
scarcity of data on idle CO deterioration throughout the useful
life of a vehicle or engine.
Analysis of the Comments
Each of the comments criticizing the HDGE idle CO standard
and test as redundant pointed to the 26.8 percent of total time
spent idling in the transient test cycle. It was argued that
since the transient cycle is deemed representative of in-use
operation, idle mode emissions are adequately represented.
(One manufacturer made the same argument for LDTs, noting that
18 percent of the time in the FTP cycle is spent idling.)
Strict control of idle emissions was claimed to be necessary in
order to certify under the transient cycle test procedure.
EPA rejects the manufacturers' contention that strict
control of idle emissions is prerequisite for certification
under the transient cycle test procedure. This contention is
based on the large portion of the time in the transient cycle
(26.8 percent) that is spent at Idle. In calculating CO
emissions for certification, the total mass CO emissions
generated during the test are divided by the total work
performed by the engine during the test, yielding a result in
g/BHP-hr that is measured against the applicable standard.
-135-
-------
Since the volumetric exhaust flow is much lower at idle than at
higher engine speeds, the mass contribution of CO during the
idle portions of the transient cycle is not proportional to the
time spent at idle. As a result, the 26,8 percent of the cycle
time spent idling contributes much less than 26.8 percent of
the total mass CO emissions. Thus, the statement that strict
idle mode emission control is required in order to be certified
using the transient cycle test procedure is not true.
In fact, one manufacturer1s comments supported the EPA
position on this issue. Data submitted on two HDGEs (4.9L and
6.1L) showed that of the total CO emissions during the
transient test, only 14 percent and 3 percent respectively were
contributed by the idle mode segments of the transient cycle.
The manufacturer states that "...the idle test in no way
reflects the ability of an engine to comply with the transient
test.™ EPA notes that the converse of this statement, that the
transient test does not reflect the ability of an engine to
comply with the idle CO test, logically follows; this undercuts
the assertion that the idle test is redundant.
The cost-per-vehicle of the idle test requirements is
minimal. Since compliance with the standard is virtually
automatic with the use of catalysts, there are no associated
development or hardware costs. Even in the case of
non-catalyst systems, only small development and calibration
costs are likely. The only other cost is that of conducting
the idle tests during certification and SEA, which is very
small on a per-vehicle or per-engine basis. With the benefits
discussed herein, EPA cannot agree that these requirements
constitute an unnecessary or unreasonable burden on the
manufacturers.
The detection of failed in-use catalytic emission control
systems will have a positive impact on air quality. These
benefits will be achieved through reduction of the number of
gross-emitting in-use vehicles. While several commenters
stated that the idle standard cannot be used as a practical I/M
cutpoint, no data were provided supporting this assertion. The
only substantive comment received in this respect noted that
idle CO levels are largely a function of previous operating
conditions, including pre-test idle time, evaporative content,
over-temperature conditions, and fuel volatility. EPA remains
convinced that the idle CO requirements are appropriate for
catalyst-equipped vehicles and engines, and will be a useful
tool in the detection of failed catalysts. Since this is the
most important application of the idle test requirements,
however, EPA agrees that these requirements should be deferred
for HDGEs until more stringent HC/CO emission standards
requiring the use of catalytic control technology take effect
in 1987.
-136-
-------
MVMA and several of the manufacturers challenged the idle
standard and test procedure on the basis that EPA has not yet
established a "reasonable correlation" between the idle test
and the other standards and procedures applicable to the
control of CO emissions, as required by Sections 206 and 207(b)
of the amended Clean Air Act. The issue is premature. EPA has
not proposed to use the idle CO test as a "short test" for
enforcing the performance warranty under Section 207(b). If
EPA takes that step, the issue of "reasonable correlation" will
then be ripe.
Cdmments regarding the numerical level of this standard
contained no information to justify a relaxation. One
manufacturer suggested that the proposed standard be revised
upward to reflect non-catalyst technology. Since the EPA
recommendation (above) is to limit the applicability of the
idle CO test to vehicles and engines utilizing aftertreatment
technology, and since all HDGEs will be capable of meeting the
revised HC/CO emission standards without utilizing such
technology, this comment need not be addressed further.
The question of the appropriateness of the dry volumetric
method of measurement used in the idle test and whether the
standard is thereby effectively made more stringent for
vehicles using smaller engines was raised. The method of
measurement to be used in the idle test procedure was taken
into account in the setting of the standard, and so the
stringency of the standard is not greater than was intended.
Smaller engines must have a slightly richer A/F mixture at idle
to avoid problems with stalling, which implies that the idle CO
emissions of a smaller engine could be somewhat greater than
those of similar but larger engines. However, the use of
catalysts should make compliance with the idle standard easily
attainable by engines of all sizes that are affected by these
requirements. In addition, EPA notes that data submitted by
one manufacturer, on idle CO emissions from 15 LDTs with engine
displacements ranging from 1.9L to 5.7L, showed that the
average idle CO emissions of well-maintained vehicles with
properly functioning catalytic systems were markedly below the
standard. These data do not support the contention by the
manufacturers that vehicles using relatively smaller engines
will have an effectively more stringent idle CO standard to
meet.
Several commenters discussed the applicability of the idle
CO standard and test procedure to LDTs, although this issue was
not officially open for comment. The issues and analyses
surrounding LDTs are the same as those discussed above for
HDEs. In particular, EPA notes these relevant facts: The
inclusion of the idle test requirements in the certification
-137-
-------
procedure serves a valid purpose. Section 206(a)(1) of the Act
allows such test requirements to be implemented. The earlier
discussion, indicating that the mass CO contribution of the
idle portions of the transient test is proportionally much
lower than the percent of cycle time spent at idle, is equally
applicable to the idle portions of the FTP.
In reference to vw's assertion that these regulations will
require the use of additional emission control hardware on its
LDTs, EPA notes that data submitted by other manufacturers
showed idle CO levels for LDTs to be well within the standard.
The idle CO standard went through an extensive proposal and
comment period as part of the original LDT rulemaking, and the
record indicates that neither VW nor any other LDT manufacturer
raised any issue over the feasibility of the standard. Since
that time, VW has certainly had an adequate period of leadtime
to meet the new requirements. VW should investigate the
possibility of meeting the idle standard through adjustment of
idle A/F settings. If this approach results in driveability
problems unacceptable to VW or fails to bring idle CO levels
under the standard, then one of the other two options
(closed-loop system or air pumps) should be exercised.
Turning now to the lesser issues raised, the first
concerns the use of the idle test in future SEAs. The idle
test procedure, as an integral part of the certification
procedure for vehicles and engines utilizing aftertreatment
technology, will be included in SEA testing.
The use of DFs with idle emission data was questioned by
one manufacturer, who noted the lack of data on idle CO
deterioration during useful life and the fact that negative DFs
are not allowed. In response, EPA notes that the application
of DFs is required for all emission standards, and therefore
will be required for this standard. Although negative DFs are
not allowed, manufacturers having data showing that no
deterioration occurs for a given regulated emission during the
useful life, can use a multiplicative DF of 1.0 or an additive
DF of zero, thereby demonstrating useful-life compliance with
that standard at the time of certification.
The lack of idle CO deterioration data for non-catalyst
vehicles/engines, which was addressed by another manufacturer,
is not an issue. As noted earlier in this section, vehicles
and engines not utilizing aftertreatment control technology
will not be subject to the idle CO standard and test
requirements.
-138-
-------
Conclusions
The primary benefit of the idle CO standard and associated
test procedure will be in the detection of failed in-use
catalytic emission control systems. With this in mind, EPA has
decided to delete the idle test requirement for all vehicles
and engines that do not utilize aftertreatment control
technology, but to retain it for catalyst-equipped vehicles and
engines. Hence for HDGEs, the idle test requirements will be
delayed until more stringent HC/CO standards requiring the use
of catalysts take effect. To make the standard and the test
more practically useful, in terms of the degree of accuracy
needed for both certification and in-use testing, the original
standard of 0.47 percent will be rounded to 0.50 percent.
The comments submitted on this issue contained no
information justifying additional changes in these requirements.
-139-
-------
3. Issue: Fuel Economy
Summary of the Issue
This analysis addresses the fuel economy impact of
emission standards for heavy-duty gasoline engines (HDGEs) for
1985 and later model years. Two separate issues are included:
1) the fuel economy effect of 1985 HC and CO emission standards
that are achievable without catalysts, and 2) the fuel economy
effect of catalyst-based HC and CO standards for 1987 and later
model years.
Summary of the Comments
The consensus of the gasoline engine industry is that a
substantial fuel economy penalty would result from the use of
stringent non-catalyst standards, such as those originally
proposed.
General Motors (GM) asserted in July 1981 that with full
life and 40 percent AQL requirements, the fuel economy penalty
will be around 2 percent for gasoline engines meeting standards
of 3.7 g/BHP-hr HC and 45 g/BHP-hr CO, when compared to the
1979 baseline mpg. Its reasoning for the penalty was that the
larger air pumps required to meet the standards will require
more energy than that gained by having a leaner full power
calibration.[1]
On March 16, 1983, EPA released for public comment a staff
paper[2] which, among other things, discussed the expected fuel
economy impact of the non-catalyst standards. For standards of
2.5 g/BHP-hr HC and 35 g/BHP-hr CO, EPA expected HDGEs to
experience as much as a 10 percent improvement in fuel economy
relative to 1979 engines. This estimate was based upon a
review of data submitted by Ford in April 1982.
In its most recent comments, [3] GM criticized EPA for
basing its estimate of a 10 percent fuel economy benefit on
only two prototype Ford engines. GM presented confidential
data to show that wide open throttle (WOT) power and fuel
economy losses would occur on engines calibrated to meet
standards of 1.3/35. (GM's WOT calibration was leaner than
stoichiometry, and required substantial timing retard to
preclude knock.) General Motors did not comment on the fuel
economy impact of catalyst standards, nor has it commented on
the fuei economy impact of non-catalyst standards of 2.5/35.
Ford's most recent comments of May 1983 also disputed the
conclusions of EPA's staff paper. Based upon the current
position of its product line, as submitted in its "best effort"
-140-
-------
data of May 1983, Ford claimed that those emission levels
result in no significant change in power, fuel economy, or
durability relative to 1979 requirements.[4]
Ford did not comment on the fuel economy impact of
catalyst standards.
Analysis of the Comments
Non-Catalyst HDGEs
This section will review available fuel economy data,
identify the likely emission control techniques to be used, and
discuss the fuel economy effects as these techniques are
applied to allow compliance with emission standards of 2.5
g/BHP-hr HC and 40 g/BHP-hr CO.
The data that are available for non-catalyst HDGEs are
presented in Tables 3-1 and 3-2. Table 3-1 addresses GM's
concern that only Ford data were used to assess the fuel
economy impact of the standards. The GM data in Table 3-1[5]
show that decreases in fuel consumption for their development
engines range from approximately zero to 19 percent. None of
the prototype engines had increased fuel consumption relative
to their 1979 counterparts. The engine with the lowest
emissions in both HC and CO had both the lowest fuel
consumption, and the largest decrease in fuel consumption (19
percent) relative to 1979.
Examination of Ford's 1984 prototype engine data in Table
3-2 [6] also shows that fuel consumption has decreased relative
to 1979 HDGEs. (More recent data submitted by Ford[4] did not
include BSFC.) Fuel consumption decreased by more than 7
percent when the average of all the April 1982 prototype tests
are compared to all of the corresponding 1979 baseline engine
tests. (Ford's concern that lab-to-lab correlation problems
could lead to EPA drawing erroneous conclusions from available
data is unfounded. Tentative results from the EPA/MVMA
correlation project show superb agreement between laboratories
for CO2 emissions, the emission with the most direct bearing
on fuel consumption calculations, and between BSFC results
themselves.)
Aside from the actual data, there are theoretical reasons
why fuel economy should improve as technology is applied to
engines to meet non-catalyst standards of 2.5 HC and 40 CO.
These theoretical reasons are based upon the combined fuel
economy effects of the technologies which will likely be
-141-
-------
Table 3-1
GM Development
Data from
Auqust 1982[5]***
BSFC*
% Decrease in
Prototype
Emissions**
Engine
1979 Baseline
Prototype
Fuel Consumption
HC
eo
292-L6
.655
.655
.640
.639
-2.29
-2.44
2.41
2.17
21.82
24.93
350-2V8
.717
.604
-15.76
1.57
28.20
350-4V8
.727
.727
.656
.649
-9.77
-10.73
2.08
1.99
29.02
27.22
366-V8
.719
.582
-19.05
.75
17.88
454-V8
.668
.6 66
-.30
1.01
22.18
Average (by engine family): -9.6 percent
* lbs/BHP-hr, EPA cycle based.
** g/BHP-hr, EPA cycle based.
*** GM stated in August 1982 that these data "are representative
of HDGE emission control systems and calibrations which are
currently believed to be at least plausible for production...
[although]...[n]one of these arrangements have been
durability tested..."[5]
-142-
-------
Table 3-2
Ford Development Data from April 1982[6]
Prototype
BSFCri]
% Decrease in
Emissions[2]
Engine Family
1979 Baseline
Prototype
Fuel Consumption
HC eo
4.9L
.696
.560
-19.54
1.66 23.2
6.1L
.681
.654
-3.96
2.33 28.8
7.5L
.633
.633
0.00
2.21 24.3
Average (by engine family): -7.8 percent
[T] lbs/BHP^hr, EPA cycle based.
[2] g/BHP^hr, EPA cycle based.
-143-
-------
applied to comply with the non-catalyst emission standards:
primarily leaner A/F ratios, retarded spark timing, and
increased air injection.
A/F ratios for current technology (1979 requirements)
HDGEs are generally quite rich, and it is expected that the new
emission standards will require leaner carburetor
calibrations. For example, GM stated that it has leaned out
its A/F ratios on its 1985 prototypes at wide-open throttle,
but that its calibrations are still on the rich side of
stoichiometric. As a case in point, Table 3-3 shows the
relationship between the A/F ratio, fuel consumption, and power
for a Chevrolet 350-CID V-8 HDGE operating at wide-open
throttle (WOT). As the A/F ratio was changed from 12:1 to
14.6:1, the fuel consumption dropped by 14 percent while the
power declined by 6 percent.[7] The HDGE engine data presented
in Table 3-3 represent performance only at WOT, and EPA
concedes that WOT constitutes a small percentage of total cycle
operating time. Logic suggests, however, that leaning A/F
ratios to reduce HC and CO emissions over all combinations of
operating modes on the transient test will also significantly
improve HDGE fuel economy. Generally, leaner A/F mixtures
decrease fuel consumption (BSFC) and therefore improve fuel
economy (mpg).
Retarding spark timing also reduces HC emissions by
raising post-combustion cylinder gas and exhaust gas
temperatures, thus promoting oxidation of the HC emissions.
This technique was widely used in pre-catalyst light-duty
vehicles to control HC. However, retarding spark timing
typically causes an increase in fuel consumption. For example,
Ford data showed that by retarding initial spark timing by 4°
(from 12° to 8° BTDC) on a 4.9L development engine, there was a
15 percent decrease in HC emissions, but a 5 percent increase
in fuel consumption.[6,8] A similar analysis of a GM 350-V8
engine in a light-duty vehicle showed that retarding timing 20°
from MBT resulted in a 10 percent fuel consumption increase at
a 14:1 A/F ratio.[9] However, because of the relaxation of the
non-catalyst HC standard from 1.3 to 2.5 g/BHP-hr, very little
timing retard should be necessary to allow compliance, as
suggested by the actual fuel economy data presented in Tables
3-1 and 3-2.
Increased air injection will also be used to reduce HC and
CO emissions. For example, for its development engines, Ford
replaced the standard two 19 in^ pumps with two 23 in^
pumps and added multiple injection points. These pumps had a
37 percent higher flow capacity.[6] There is, however, a
practical limit to the amount of air injection; too much air
can actually quench the oxidation reactions and preclude
further emission reductions. (Ford experimented with a 50
-144-
-------
Table 3-3
Chevrolet 350-CID V-8 Engine Data[8]
Air/Fuel
Fuel
% Change in
% Change
Ratio
Consumption[1]
Fuel Consumption[2]
in Power[2]
12.0:1
.575
_ —
_ _
12.8:1
.542
-5.7
-2.1
13.2:1
.525
-8.7
-2.9
13.8:1
.493
-14.3
-3.5
14.6:1
.493
-14.3
-5.9
15.0:1
.493
-14.3
-9.4
Til lbs/HP-hr.
[2] Relative to 12:1 A/F ratio.
-145-
-------
in^ air pump, and observed no significant incremental
emission reductions.) As with retarded spark timing, increased
air injection reduces fuel economy. (Air pumps require energy
to be driven.) EPA's own test datafioj indicate that there may
be a 2.5 to 4 percent increase in fuel consumption if air
injection rates are increased to the extent necessary for
catalysts. (EPA expects the air injection rates for 1985
non-catalyst and 1987 catalyst engines to be similar.).
Other emission control techniques EPA expects to be used
in 1985 should affect overall fuel economy very little. These
techniques include early fuel evaporation systems, heated air
intake, temperature-actuated timing retard, and automatic
chokes. Early fuel evaporation systems use exhaust gases to
heat the A/F mixture, resulting in reduced emissions and
shorter warm-up periods. Shorter warm-up periods would promote
better efficiency and therefore better fuel economy. Heated
air intake also reduces engine warm-up time and allows leaner
carburetor calibrations, thus better fuel economy. Cold
temperature-actuated timing retard reduces cold start emissions
at the expense of a slight increase in fuel consumption. These
technologies are not anticipated to have any noticeable effect
on overall fuel economy, however, because of the small
percentage of operating time that engines in the field spend
cold.
The theoretical picture painted for fuel economy is one of
trade-offs. Fuel economy would be predicted to improve
significantly with leaner A/F mixtures but would be predicted
to decrease marginally with larger air injection systems. (EPA
does not expect significant timing retard to be required to
meet the 2.5 g/BHP-hr HC standard.) All of the actual emission
data available to EPA show that there is a greater probability
for an overall fuel economy benefit rather than a fuel economy
penalty, and that the gains from leaner A/F calibrations will
more than offset the losses attributable to increased air
injection. The fuel economy data for Ford's and GM's prototype
engines (Tables 3-1 and 3-2) show that these engines are
actually more fuel efficient - with increased emission control
- than they were in 1979. On the average basis, these engines
are running 7-10 percent more efficient than their 1979
counterparts. Based upon this prototype fuel economy data, a
modest fuel economy increase for 1985 HDGEs is anticipated
relative to the 1979 baseline engines. Certainly, no aggregate
fuel economy penalty is likely.
Catalyst-Equipped HDGEs
Little has changed with respect to the availability of
information on the fuel economy effect of catalyst standards
since the December 1979 Final Rulemaking. The fuel economy
-146-
-------
analysis associated with that rulemaking[10] concluded that a 4
to 9 percent improvement in fuel economy would be achieved,
relative to 1979 engines, when catalysts were applied to
HDGEs." This conclusion was based to a large extent on th'e
performance of light-duty vehicles when catalysts were applied
for increased emission control.
Much of the emission control required for
catalyst-equipped 1987 HDGEs is being accomplished for 1985
{i.e.,1 the first step in applying catalysts to heavy-duty
engines was to reduce engine-out emission levels). As
discussed above, the techniques used to reduce HDGE engine-out
emissibns have also yielded a fuel economy benefit. This is
much of the same benefit which would have been observed had
catalysts been immediately applied to HDGEs in 1985. The
remaining question is how much of an incremental change in fuel
economy is to be expected relative to 1985 when catalysts are
applied in 1987?
Application of catalysts has traditionally removed much of
the need for engine calibrations which tended to reduce fuel
economy (e.g., spark retard). However, EPA does not expect the
significant use of timing retard, or other engine-out emission
control calibration strategies which would degrade fuel
economy, to be used for 1985. Therefore, EPA does not expect
the addition of catalysts in 1987 to provide much additional
flexibility relative to 1985. Catalysts also create modest
increases in exhaust backpressure which may somewhat decrease
fuel economy; these backpressure increases, however, can be
easily offset by larger diameter exhaust systems. Finally, EPA
anticipates no increase in air injection rates relative to 1985
significant enough to affect fuel economy. Given the absence
of major potential calibration optimizations, and given modest
but correctable increases in backpressure as catalysts are
applied, EPA judges that little change in vehicle fuel economy
will be seen between 1985 and 1987 on account of the change in
emission standards. Much of the fuel economy benefit predicted
in 1979 as attributable to catalysts will already have been
achieved in 1985.
Conclusions
1985 HDGEs are expected to incur a fuel economy benefit as
a result of the non-catalyst standards. Prototype engine data
indicates that this benefit, on average, could be as large as
7-10 percent. No net change in fuel economy relative to 1985
is expected, however, when catalysts are applied in 1987.
-147-
-------
References
1. "Cost Effectiveness Analysis of Alternatives to 1984
Heavy-Duty Engine Emissions Regulations," General Motors
Corporation, July 31, 1981.
2. "Issue Analysis - Final Heavy-Duty Engine HC and CO
Standards," EPA Staff Report, EPA Public Docket No. A-81-11,
March 1983.
3. "General Motors Comments on the March, 1983 EPA
Staff Report Issue Analysis - Final Heavy-Duty Engine HC and
CO Standards," May 6,1983.
4* "Ford Motor Company Response to the Environmental
Protection Agency on Gaseous Emission Regulations for 1985 and
Later Model Year Heavy-Duty Engines," May 6> 1983.
5. Letter from T. M. Fisher of General Motors, to
Charles L. Gray, Jr., U.S. EPA, dated August 9, 1982.
6. "Response to Revised Gaseous Emission Regulations
for 1984 and Later Model Year Light-Duty Trucks and Heavy-Duty
Engines," Ford Motor Company, April 1982.
7. "Heavy-Duty Fuel Economy Program: Evaluation of
Emissions Control Technology Approaches," EPA Paper No.
460/3-77-010, July 1977.
8. Letter from R. E. Bisaro of Ford Motor Company, to
W. M. Pidgeon, U.S. EPA, June 25, 1982.
9. "Optimizing Engine Parameters with Exhaust Gas
Recirculation," SAE Paper No. 740104, 1974.
10. "Summary and Analysis of Comments to the NPRM: 1983
and Later Model Year Heavy-Duty Engines Proposed Gaseous
Emission Regulations," U.S. EPA, December 1979.
-148-
-------
4. Issue: Allowable Maintenance
Summary of the Issue
In 1980, EPA published revised allowable-maintenance
intervals for LDTs and HDEs. The primary purpose for these
intervals was to encourage the design of long-life
emission-related components and to limit maintenance to that
which was considered technologically necessary. In the NPRM,
EPA proposed to add an HDGE spark plug maintenance interval for
leaded fuel, but no other specific changes were proposed. Even
though the general area of allowable maintenance was not
formally reopened, both LDT and HDE manufacturers submitted
comments criticizing the intervals and the HDE manufacturers
requested relaxation of several specific intervals.
Summary of the Comments
Light-Duty Trucks
General Motors commented that the LDT requirements were
not cost effective, had no air quality benefit, and were
inappropriate. It recommended that EPA drop its current
requirements and adopt the LDV requirements, thus allowing GM
to recommend the maintenance it believes is appropriate.
Heavy-Duty Gasoline-Fueled Engines
Heavy-duty gasoline engine manufacturers generally
accepted EPA's leaded-fuel spark plug maintenance interval of
12,000 miles. However, Chrysler asked that the unleaded-fuel
spark plug maintenance interval be revised from 25,000 miles to
18,000 miles, primarily because it had no data beyond that
point. Ford also requested that the intervals for the EGR
valve, PCV valve, heat-control valve, and checking the choke
system be revised because they are also subject to lead
fouling. As before with LDTs, GM stated that. the
allowable-maintenance intervals for HDGEs were inappropriate
and should be dropped.
Heavy-Duty Diesel Engines
Several commenters stated that the allowable-maintenance
intervals for HDDEs were too long. Further, the commenters
contended that setting allowable-maintenance intervals for
HDDEs was not necessary because heavy-duty diesel truck owners
maintain their vehicles due to business reasons, and the very
competitive nature of the HDDE business drives the development
of more durable components. Specifically, several commenters
stated that the current intervals were too long for the newly
-149-
-------
emerging medium-duty diesel engines, and in some cases were
longer than the anticipated engine lifetime.
General
One general comment was received which stated that the
allowable-maintenance intervals do not allow the manufacturers
to recommend more frequent maintenance than that specified by
the interval.
Analysis of the Comments
Light-Duty Trucks
Even though these requirements were not formally reopened
for comment, EPA has carefully reviewed the comments received
on the LDT allowable-maintenance requirements. Although there
is clearly some disagreement between the manufacturers' and
EPA's assessments of the cost effectiveness and air quality
impact of these provisions, EPA finds no compelling evidence
for revising these requirements.
EPA believes it is important to encourage the design and
use of more durable, low-maintenance emission-related
components, and believes the 1984 LDT allowable-maintenance
intervals effectively accomplish this task. Adopting the
current LDV requirements would be a step backwards and would do
nothing toward meeting that objective. Unfortunately, there
are no strong market forces acting to encourage the
manufacturers to develop and use more durable, low-maintenance
components.
No real data was submitted to question the technological
feasibility of these requirements, and EPA continues to
believe, based on its original analysis,[1] that these
requirements are technologically feasible and are an
appropriate and cost-effective means of improving air quality.
It is also important to note that while the new
allowable-maintenance requirements are more restrictive than
existing provisions in some areas, they at the same time
reclassify a great deal of maintenance items as non-emission
related. For these items, the manufacturers are free to
recommend whatever maintenance provisions they believe are
reasonable and necessary, without other regulatory requirements.
Heavy-Duty Gasoline-Fueled Engines
As with LDTs, EPA believes that allowable-maintenance
intervals are necessary to encourage the use of more durable,
low-maintenance emission-related components. It does not
-150-
-------
appear that market forces and business competition can be
relied upon to meet the stated goal. For example, for the four
HDGE manufacturers, there is a range of 6,000 miles in the
manufacturers' current recommended maintenance intervals for
spark plugs. This discrepancy has existed for several years,
and yet there appears to be no effort on the part of the
manufacturers at the lower end of the range to lengthen these
intervals. It was to deal with this type of situation that the
allowable-maintenance provisions were first adopted.
Turning first to Chrysler's request for a relaxation in
the unleaded-fuel spark plug maintenance interval (25,000 miles
to 18,000 miles), EPA notes that the sole basis for Chrysler's
request is that it does not have data beyond its present
interval of 18,000 miles, and thus Chrysler is uncertain about
the feasibility of the 25,000-mile interval. (Two of
Chrysler's three HDGE families are currently certified using
unleaded fuel.)
Chrysler's request for a relaxation appears to be based
primarily on a desire not to conduct any further testing,
which, given EPA's goals in establishing these provisions, is
insufficient reason to delete the requirement. If Chrysler
decides to remain in the HDGE market after 1984, new testing
will be required for development and certification. At this
time, Chrysler will then have the opportunity to demonstrate
compliance with the longer interval, assuming that Chrysler
continues to choose emission-control technology which requires
unleaded fuel.
EPA believes that the 25,000-mile spark plug maintenance
interval is achievable with Chrysler's present technology.
Chrysler's present LDT recommended maintenance interval is
30,000 miles. Chrysler's present heavy-duty gasoline
vehicles/engines are so similar to their light-duty
trucks/truck engines that compliance could be projected based
almost purely on extrapolation. Chrysler's comments even
indicate that it has tested some of its heavy-duty gasoline
vehicles/engines on the LDT chassis-roll procedure. EPA is
confident that Chrysler can meet the 25,000-mile unleaded-fuel
spark plug maintenance interval with minimal effort.
EPA concurs with Ford's request that the EGR maintenance
interval for leaded fuel be revised to allow one scheduled
maintenance prior to 50,000 miles. Past performance of EGR
systems on engines using leaded fuel leaves some doubt about
the feasibility of the 50,000-mile interval before 1985 with
the current level of lead used in leaded fuel. It is the
judgment of both EPA and the manufacturers that the. proper
function of the EGR valve/system could be affected by lead
deposition.
-151-
-------
However, EPA does not agree with Ford's request for a
relaxation of the 50,000-mile PCV valve interval, EPA believes
"f ¦ ¦
that any plugging or hang-up which may occur in the valve is
caused xJy contaminated blowby oil and not lead deposition.
Ford submitted no data to support its request or its position
that lead deposition is a major contributor to problems with
the PCV valve. EPA believes that the 50,000-mile interval is
technologically feasible and will encourage the use of durable,
low-maintenance PCV valves.
Ford also requested that HDGE manufacturers be allowed to
service (lubricate) their heat-control valve system once during
the first 50,000 miles. (Similar systems used by other
manufacturers are called early fuel evaporation (EFE).) This
request was based on the tight clearances within such systems
and the concern that lead buildup might hinder the free
operation of the valves.
The current allowable-maintenance provisions (§86.084-25)
do not specify maintenance of this type to be emission
related. Therefore, the manufacturer is free to perform the
maintenance as deemed necessary, provided that such maintenance
is recommended to the consumer.
In a follow-up conversation on this issue Ford withdrew
its request for additional choke-system maintenance.
Heavy-Duty Diesel Engines
Even though the HDDE allowable-maintenance intervals were
not formally opened for comment, the EMA submitted the results
of a substantial survey of fleet and owner/operator maintenance
practices. EPA is always open to substantive input and data on
past regulatory decisions and is considering the EMA submittal
accordingly.
On its face, it appears that there is some validity to the
manufacturers* contention that the ¦ business nature of the
heavy-duty truck and bus industry leads to more routine
maintenance than might otherwise occur, and drives the HDDE
manufacturers toward continually lengthening the recommended
maintenance intervals. However, the EMA report on maintenance
practices tends to cast some doubt on the manufacturers'
assertions that routine maintenance is the norm for HDDVs.
Tables 4-1 and 4-2 summarize the maintenance practices for
the components which are currently covered by EPA's allowable-
maintenance intervals.* In only a few cases was routine
No data was submitted on diesel EGR or PCV system
maintenance, presumably because neither is in widespread
use on current HDDEs.
-152-
-------
Table 4-1
Maintenance Practices -- Total Fleets
Injector Nozzles -- Function Incidence Rates*
Total GVW Fleet Size Usage Owner/
Item Fleets** 8 7-6 3-49 50+ Long Haul Other Operator
Clean/Recali-
brate/Check:
Perform
Routine Main-
tenance (%)
Routine
Maintenace
Interval
(miles)
(x 1,000)
Perform
Maintenance
Upon
Failure (%)
Replace:
Perform
Routine Main-
tenance (%)
Routine
Maintenance
Interval
(miles)
(x 1,000) 143 153 93 143 145 178 111 155
Perform
Maintenance
Upon
Failure {%) 78 77 83 80 69 76 80 79
Routine and failure maintenance do not always sum to 100 percent due
to responses which fell in neither category.
Does not include owner/operator^ which is considered as a separate
group in this study.
46 48 35 46 44 55 41 34
88 B9 81 85 100 87 89 98
53 52 62 53 54 44 58 66
21 23 15 19 29 24 19 21
-153-
-------
Table 4-2
Maintenance Practices -- Total Fleets
Turborcharger - Function Incidence Rates*
Total GVW Fleet Size Usage Owner/
Item Fleets** 8 7-6 3-49 50+ Long Haul Other Operator
Rebuild:
Perform
Routine Main-
tenance (%) 27 30 16 27 29 31 24 44
Routine
Maintenance
Interval
(miles)
(x 1,000) 201 207 149 211 169 190 219 124
Perform
Maintenance
Upon
Failure (%) 64 61 75 63 66 56 70 56
Replace:
Perform
Routine Main-
tenance (%) 11 il 14 11 14 8 13 26
Routine
Maintenance
Interval
(miles)
(x 1,000) 154 179 81 121 182 61 183 135
Perform
Maintenance
Upon
Failure (%) 79 79 79 78 82 76 81 74
Routine and failure maintenance do not always sum to 100 percent due
to other responses which fell into neither category.
Does not include owner/operator, which is considered as a separate
group in this study.
-154-
-------
maintenance conducted half the time or more, and in virtually
no cases could routine maintenance be considered dominant. In
short, the data submitted by EMA tend to refute its contention
that allowable-maintenance intervals for HDDEs are not
necessary due to good maintenance practices by heavy-duty truck
owners.
However, the data submitted by EMA provide some useful
information on the average-mileage intervals followed by those
users who do perform routine maintenance. Even though there
are some substantial disparities among the intervals followed
by the various HDDE users, the data are useful for comparing
the length of the EPA allowable-maintenance intervals against
current field practices as represented by the EMA data.
The EMA data show that EPA's allowable-maintenance
interval for cleaning of injector tips is generous. EPA's
interval is 50,000 miles, and the EMA data indicate a fleet-
average value of 88,000 miles, and an owner/operator average
value of 98,000 miles.
EPA currently has an allowable-maintenance interval of
200,000 miles for replacement of injectors. Data submitted by
EMA indicate a wide range of values in current practices.
Intervals tend to be lower for Gross Vehicle Weight Rating
(GVWR) Classes VI and VII trucks or non-long haul applications
(93,000-111,000 miles), and higher for GVWR Class VIII trucks
or fleets involved in long-haul applications (153,000-178,000
miles). For owner/operators the average-mileage interval is
155,000 miles. These data indicate that EPA's interval is too
stringent, especially for the HDDE class as a whole. An
interval of 200,000 miles might be reasonable for engines
designed for long-haul/Class VIII trucks, but is probably too
stringent and not as cost effective for less durable engines.
A revision of the current EPA interval appears appropriate if
one interval is to serve for the entire HDDE class. In this
case, setting a revised interval of 150,000 miles seems
appropriate based on the EMA data.* This would tend to extend
the intervals in the cases where routine maintenance appears
least prevalent (and the intervals are shortest) , and would
extend the intervals on average for the total fleet. EPA
believes an interval extended to 150,000 miles is feasible for
HDDEs, including those used in Classes VI and VII trucks.
Also, in a December 1980 study prepared by a task force of
the American Trucking Association, forty respondents to
its survey indicated a mean injector replacement interval
of 170,125 miles. The range in values was 50,000-375,000
miles, the median was 150,000 miles, and the mode was
100,000 miles.
-155-
-------
The EMA data also address the area of turbocharger
rebuild/replacement. The EMA data show that rebuilds are far
more prevalent than replacement, and that replacements tend to
occur as a result of catastrophic failure instead of routine
maintenance. As before with injectors, EPA's current interval
of 200,000 miles for rebuild of turbochargers appears
reasonable for GVWR Class VIII trucks used in fleets in most
applications (169,000-219,000 miles), but might be too
stringent for GVWR Classes VI and VII trucks and
owner/operators (124,000-149,000 miles). EPA's interval of
200,000 miles also covers the replacement of turbochargers.
The EMA data are not as useful here because of the heavy
dominance of non-scheduled maintenance practices. Even so, it
is evident that when replacement does occur, it is at shorter
intervals than rebuilds. Considering the EMA data, EPA
believes that a revision of the turbocharger
rebuild/replacement interval is appropriate. Setting the
interval at 150,000 miles would accomplish the goals of the
allowable-maintenance program, while at the same time serving
as a reasonable compromise value for the rebuild of GVWR
Classes VI and VII truck turbochargers and the average fleet
interval for replacement. EPA also believes that there is a
greater likelihood that the turbocharger maintenance will be
performed because of the likely negative performance and fuel
economy impacts.
In summary, EPA continues to believe that the
allowable-maintenance intervals are necessary for HDDEs,
because routine maintenance of emission-related items is not as
prevalent as claimed by the HDDE manufacturers. EPA's goal is
to certify HDDEs under maintenance intervals that reflect the
actual in-use maintenance schedule as closely as possible.
However, EPA does see some validity to the manufacturers'
contention that the competitive nature of the HDDE business
will tend to provide an impetus to lengthen recommended
maintenance intervals and to improve general component
durability. Relaxing the allowable-maintenance intervals for
the two components discussed above is appropriate, because the
data submitted tend to indicate that the technologically
necessary intervals set by EPA in 1980 are too long for the
HDDE class as a whole. The intervals set in 1980 are
reasonable for engines used in Classes VII-VIII long-haul
trucks, but appear too stringent for less durable medium-duty
diesel engines designed for trucks in GVWR Class VI and below.
If one interval is to serve for the entire HDDE class then it
may by necessity have to be shorter than is technologically
necessary for Classes VII-VIII trucks/engines.
-156-
-------
General
One commenter claimed that the allowable-maintenance
intervals preclude the manufacturers from recommending more
frequent maintenance than permitted by EPA. This is not the
case. Manufacturers may recommend maintenance at more frequent
intervals if they desire, but any maintenance beyond that
prescribed by the allowable-maintenance intervals cannot be
tied to emission warranty eligibility.
It is also important in this context to remember that, as
stated earlier, EPA's allowable-maintenance requirements apply
only to emission-related maintenance. Manufacturers are
allowed to recommend any maintenance intervals that are
reasonable and necessary for non-emission-related maintenance.
Conclusions
1. No changes will be made to the LDT allowable-
maintenance provisions.
2. EPA has decided to include an HDGE leaded-fuel spark
plug maintenance interval of 12,000 miles, but not to revise
the present unleaded-fuel interval.
3. A leaded-fuel EGR valve/system maintenance interval
for HDGEs which allows servicing at 24,000 mile intervals will
be included.
4. The PCV maintenance interval for HDGEs will not be
revised.
5. The injector replacement and turbocharger rebuild/
replacement intervals for HDDEs will be reduced, from 200,000
to 150,000 miles.
-157-
-------
References
1. "Summary and Analysis of Comments in the Proposed
Rulemaking for Gaseous Emission Regulations for 1983 and Later
Model Year Light-Duty Trucks," U.S. EPA, OANR, OMS, ECTD, SDSB,
May 1980.
-158-
-------
5. Issue: Minor Amendments to HPK/LDT SKA
Summary of the Issue
On January 13, 1982, EPA proposed several technical
and procedural amendments to the requlations governing
Selective Enforcement Auditinq (SKA) of HDEs and LDTs
contained in Subparts A, K, and N. These reaulations were
originally promulgated for HDEs at 45 FR 4167 and 4170
(January 21, 1980), and were updated on September 25, 1980
to include LDTs at 45 FR 63767 and 63772.
These amendments were intended to clarify specific
aspects of the existing requlations, to improve the efficiency
with which the HDE/LDT SEA proqram will be conducted in the
future, and to reduce the compliance burden on the affected
manufacturers where practical. Throuqh these amendments,
EPA expects the RDE/LDT manufacturers to accrue substantial
cash expenditure and cash flow savings.
Summary and Analysis of the Comments
The HDE/LDT manufacturers did not have many major
concerns with the amendments to the HEE/LDT SEA procedures.
The manufacturers did however, raise numerous minor issues
pertaining to various technical points and details of the
amended, as well as the original FTDE/LDT SEA procedures.
The majoritv of these comments came from General Motors
who stated that its proposal was a resubmittal of
its earlier comments (submitted on the HDE NPRM which
-159-
-------
was promulgated as a final rule on January 21f 1980) on the
subject of Subpart K. Therefore, this entire summary and
analvsis is dedicated to the new comments on the technical
and procedural SEA amendments as well as GM's resubmittal
of its original comments regarding the FTDE/LD? SEA procedures.
The comments received fall into a number of subissues.
Each of these subissues will be treated separately.
a. Applicability ($86.1001-84)..
suggested that this section include a provision to
allow a phase-in period for trial test orders for heavy-duty
engine (FTOE) SEAs. "A minimum period of one year, after the
first Heavy-Duty engines are certified on a new test cycle,
is recommended."
On January 13, 1982, EPA proposed several regulatory
relief initiatives related to the TOE/ID^ industrv. One of
these initiatives was a two-year delay in the start of the
BDE SEA proqram until 1986. The two-year delay in the
HDE SEA program already satisfies GM's concern of a phase-in
period of one year after the first RDEs are certified on a
new test cycle (gas and diesel HPEs are scheduled to be
certified on a new transient test cycle in the 1985 model
year, with optional transient test standards for the 1984
-160-
-------
model year).
In addition, the Aqency will make its SEA personnel
available, to the extent possible, to monitor trial SEAs
prior to the 1^86 model year (anytime in the 1985 model
year). Any HDE manufacturer that is interested in conductincr
a trial audit, pursuant to the provisions of Subpart K, may
contact EPA in writing to make the appropriate arrangements.
Also, the Aaency prefers that any manufacturer requesting a
trial audit invite other HDF manufacturer representatives to
observe the audit in order to maximize its usefulness.
These trial audits are designed to provide both the manufac-
turers and EPA with logistical and procedural experience
in running the new SEA program and will be performed on a
voluntary basis.
b. Definition of "Configuration" (§86.1002-84(b)).
The present regulations state that a HDE/IDT
configuration will be "...described on the basis of...other
parameters which may be designated by the Administrator."
CM contested this definition as being unreasonably broad and
vague and wanted protection aoainst arbitrary selection of
parameters by EPA.
This provision about "other parameters" is similar to
a provision contained in the present LDV SEA definition of
"configuration" (when the present LDV regulations or pro-
-161-
-------
gram are mentioned henceforth, they also include LDTs
until the 1984 model year). A LDV configuration has never
been defined beyond the specific parameters contained in
that definition.
Present HDE/LDT configurations can be described using the
specific parameters in the HDE/LDT definition. However, EPA
needs some flexibility in specifying configurations, because
new emission control technologies developed in response to
1984 and later HDE standards may result in emission control
parameters not presently identified. EPA does not intend to
use this flexibility in an unreasonable manner but has retained
the proposed definition in the final rule.
c. Test Orders - Instructions in test order (S86.1003-84(b)j.
GM stated that the phrase "...instructions in the test
order.", in the last sentence of paragraph (t), be eliminated
as redundant and unnecessary in that the Clean Air Act (the
Act) mandates compliance with test orders issued under the
regulations. EPA prefers not to delete the phrase because
it alerts the manufacturer of their obligations directly in
the regulations under Subpart K.
d. Test engine or vehicle selection procedures in the test
order (§86.100 3-84(c)).
Present reaulations state that "The test order will
specify... the procedure by which engines or vehicles of
the specified configuration must be selected." General
Motors believes that this provision is too vague and ambiauous;
-------
it stated that the test sample selection process should be standard-
ized and placed in the regulations. In addition, General Motors
recommended revising this provision to ensure enaines or vehicles
are selected in a quantity not to exceed that required to meet test-
ino schedules while not disrupting normal production activities.
It is not oossible to standardize the test sample selection
procedure because of the varying production practices and assembly
plant operations of the different PDE/IDT manufacturers. This
conclusion is based on visits by EPA personnel to domestic HDE/LDT
manufacturers and EPA's experience with the LDV SEA program.
Also, the sequential sampling plans were designed to prevent
severe disruption of a manufacturer's production and customer
delivery schedules. The impact on these schedules should be
minimized because these sampling plans allow configurations to
be tested as expeditiously as possible and the test engines or
vehicles may even be selected over several days. It should be
emphasized that paragraph §86,1007-84(a) allows for manufacturer
input into the determination of the appropriate test sample selec-
tion procedure. Therefore, EPA has made no changes in its proposed
statement for the final rule.
e. Other standardized test order instructions (§86.1003-
84(c)).
The current regulations state that "In addition, the
test order may include other directions or information
essential to the administration of the required testing."
General Motors stated that the latitude allowed EPA bv this
provision is too broad, and that any instructions which can
-163-
-------
be standardized should be placed in the reflations and any
information which is deemed "essential" should also be
included. FPA determined that some of the specific instruc-
tions presently incorporated in LDV SKA test orders are
applicable to HDE/LDT SEA testing and included them in the
January 2lf 1980 final rule as new paragraph §86.1003-84(c)(2).
However, the provision to include "other directions or infor-
mation" essential to administer SEA testing has been retained
to allow some flexibility in SEA operating procedures. This
flexibility can be in both the interest of the manufacturers
and EPA, as it will allow audits to be conducted in the most
expeditious manner practical, given circumstances unique to
a particular manufacturer.
The latitude built into the test order and sample
selection sections of the SEA regulations is intended to
accommodate procedural variations, especially in the area of
test engine selection. Specific instructions may be made
to minimize the impact on each manufacturer's normal produc-
tion activities while still assuring the generation of
accurate, reDresentative test results.
f. Selection at non-preferred plants (§86.1003-84(d) and
§86.1007-84(a)).
The current regulations assert that, even thouqh a
manufacturer has submitted a list of assembly plants preferred
for engine or vehicle selectionthe Administrator mav
order selection at other than a preferred location."
stated that this paragraph should be revised to ensure that
-164-
-------
selection is performer? at non-preferred locations onlv if it
will not disrupt normal oroduction activities and onlv upon
makina the determination that evidence exists indicatinq a
noncompliance at other than the manufacturer's preferred
plant, •"'he seauential samplinci plans contained in this
regulation were desianed to allow flexibilitv. in sample
selection to orevent, to the greatest extent possible,
disruptinq a manufacturer's normal oroduction and delivery
schedules. FPA intends to select test ennines and vehicles
at preferred locations, but reauires the flexibility of
selectinq at non-preferred plants when that would allow
the audit to be performed expeditiously or permit the
auditino, based upon available evidence, of specific cases
of noncompliance. For example, in the LDV SEA proaram,
audits have had to be canceled or significantly delayed due
to the preferred plant beinq down for a couple of weeks,
closed indefinitelv, or otherwise unavailable for selection.
Tn such cases, the Aoencv needs to be able to select its
test enaines or vehicles at non-oreferred plants. "^o retain
this flexibility, EPA made no chanqe to the final rule.
q. Additional test orders for noncompliance
f Sfl6. 10 0 3 - 8 4 ( f ) ( 3) ) .
FPA provided that after the annual limit has been
met, the Administrator mav issue additional test orders for
which evidence of noncompliance exists. funeral "otors
arnues that test orders should not be indiscriminantlv and
-165-
-------
unreasonably issued on the basis of "any evidence." Further
GM says, test orders when issued, should count against the
annual limit the same as any other test order. Otherwise,
the "annual limit" would be open ended and the manufacturer
would be subject to an indefinite number of test orders.
EPA has a responsibility to investigate those engine
configurations for which it has evidence of noncompliance,
and, therefore, has not incorporated this provision into the
regulations. Also, this provision is consistent with the
present LDV regulations. The Agency is however, sensitive
to GM•s concern that manufacturers may be subjected to an
indefinite number of test orders. Based on evidence of
noncompliance/ a test order issued within the annual limit
will count toward the annual limit, if the configuration
passes the audit. If the limit has been reached, additional
test orders may be issued only on the basis of evidence of
noncompliance. In addition, the provision requiring a
statement of the reason for issuance of a test order beyond
the annual limit will be retained.
h. Discrepancies between EPA test results and manufacturer
test results (§86.1004-84(b) and (c)).
The present reaulations state that EPA's test re-
sults comprise the official data for a test engine or
vehicle when there is a disagreement with a manufacturer's
results. GM disagrees with the assumption that the manufac-
turer's test facility is deficient and that it bears the
-166-
-------
burden of proving that its own data are correct. It argues
that the certificate of conformity should not be suspended
with respect to the vehicle or engine configuration in
auestion until the reasons for the lack of correlation are
determined. However, the regulations provide two mechanisms
for resolving differences between data: (1) paragraph
§86 .1004-84(c)(2) allows a manufacturer to demonstrate that
EPA's data were erroneous and its own data were correct; and
(2) if EPA invokes a suspension of the certificate of conformity
based on the Administrator's test data, the manufacturer can
request a hearing under paragraph §86.1012-84(1) to determine
whether the tests were conducted properly. Therefore, this
provision is unchanged.
i. Retaining names of involved personnel
(§86.1005-84(a)(2)(iii) and (a)(2)(iv)).
Paragraph §86.1005-84(a)(2)(iii) requires the manufacturer
to retain the names of all personnel involved in the conduct
of an audit and paragraph §86.1005-84(a)(2)(iv) requires the
manufacturer to retain the names of all personnel involved
in the supervision and performance of a repair. GM proposed
that these provisions be deleted because this information
is unnecessary and irrelevant for EPA's needs and the
information goes beyond that required by the current LDV/LDT
regulations.
EPA does agree that these provisions- should be consistent
with the requirements of the present LDV/LDT regulations which
-167-
-------
only require the names of supervisory personnel he retained
hv the manufacturer. Further, EPA believes that the names
of manufacturer personnel involved in repairing vehicles or
engines and conducting audits can be obtained from supervisory
personnel if an investigation of an audit is ever necessary,
therefore, EPA revised paragraphs §86.1005-84{a)(2 J(iii)
and
-------
to help determine compliance of HDEs or LDTs with applicable
emission standards. In addition, Subpart K does not require
the manufacturer to submit emission test results on ADP
equipment. What the provision says is i_f emission test
results are available on ADP equipment and the manufacturer's
storage device is compatible with EPA's ADP equipment, then
the manufacturer would submit the information in a form
available for automatic processing. EPA will even furnish
the necessary ADP storage devices upon a manufacturer's
request.
This submission of test data requirement has been
proven workable in the current LDV SEA proqram and does not
appear to be unreasonably burdensome to manufacturers. EPA
believes that the reporting period (quarterly) and require-
ments it promulgated for the RDE/LDT manufacturers are
reasonable and are similar in scope to those currently being
met by LDV/LDT manufacturers. A semiannual reporting
period, .with closing dates of January 31 and July 31 (as
suggested by GM), would not adequately meet EPA's needs.
Emission's data received so late in the model year (the
first reporting period's data would not be received until
late February or early March and the last reporting period's
data would not arrive until the end of the model year) would
provide the Agencv with little help in determining compliance
of HDEs and LDTs with applicable emission standards. The
manufacturer is required to describe the emission test used
-169-
-------
to obtain the data submitted (see §86.1005-84(c)(1)) to help
EPA evaluate the value of the data when compared to the FTP.
EPA, therefore, proposes no changes to §8fi.1005-84(c).
k. Additional information which the Administrator may
require (§86.10O5-84(e)).
CM recommended deleting this requirement because it
is ambiguous, provides unlimited discretion to the Adminis-
trator, and goes beyond the scope of Section 208(a) of the
Clean Air Act (GM also recommended that §86.1009-84(d)(5)(vi)
be deleted for similar reasons). EPA however, has made no
changes to these paragraphs for the final rule because the
Agency needs some flexibility in requiring information on a
case-by-case basis. Paragraph §86.1005-84(e) states that
the Administrator may request information not specifically
provided under the other sections of 586.1005-84. However,
the Administrator is still bound by Section 208(a) to require
only the information that will enable a determination to be
made of whether a manufacturer has acted or is acting in
compliance with Title II, Part A of the Clean Air Act and
the regulations promulgated thereunder.
1. Entry and access (§86.1006-84(b)(4)).
In this paragraph, GM recommended that only "emission
relatedn parts or aspects of an engine or vehicle be inves-
tigated, but it did not give a reason for this comment.
586.1006-84(a) states that matters related only to this
-170-
-------
subpart (Subpart K) will be investigated. Therefore, EPA
has not cevised §86.1006-84(b)(4) in response to this
comment.
m. Entry and access (§86.1006-84(h)(3)) .
GM recommended that this paragraph be revised in order
to more accurately reflect the current practice of the
LDV regulations. Paragraph (h)(3) of $86.1006 deals
with the definition of "operating hours" at facilities
or areas other than those where engine or vehicle storage is
concerned. EPA concurs with GM's comment and has amended
§86.1006-84{h)(3) to be consistent with the LDV regulations,
for purposes of uniformity and clarity.
n. Authorization for personnel appearance and entry without
24 hours notice (§86.1006-84(h)(4) and (5)).
GM recommended that paragraph §86.1006-84(h)(4) be
amended and a new paragraph S86.1006-84(h)(5) be added to
require the Assistant Administrator for Air, Noise, and
Radiation to approve these authorizations. EPA believes it
is unnecessary to require the Assistant Administrator to
authorize either appearances of personnel or entry without
24 hours prior notice because these authorizations can be
performed bv other responsible Agency officials. If a
manufacturer refuses to consent to personnel appearance or
entry without 24 hours notice, EPA is required to seek a
search warrant before attempting to conduct these activities.
-171-
-------
Therefore, no changes relating to these issues have been
made in the final rule.
o. Selection of incomplete test engines or vehicles
(§86.1007-84(b)).
The present regulations state that a test order will
specify the manner in which assembly of incomplete test
engines or vehicles will be completed. GM recommended that
this provision be revised to allow the assembly to be
completed according to applicable production and assembly
quality control methods and procedures. GM first proposed
this revision during the initial comment period on the HDE
gaseous emission regulation notice of proposed rulmaking
{NPRM). EPA agreed with the request and amended paragraph
§86.1007-84(b), in the January 21, 1980 HDE final rule, to
allow the use of these methods. However, EPA qualified GM's
suggestion by adding that the procedures must be "documented
by the manufacturer" and eliminated GM's sugqested phrase,
"assembled to normal certification dress." These qualifica-
tions were necessary to ensure that engines are assembled
using only standard assembly line procedures and quality
control checks and that these test enqines duplicate, as
closely as possible, the configuration of the manufacturer's
engines being distributed into commerce. EPA continues to
believe such documentation is important and will retain
paragraph §86.1007-84(b) as is.
-172-
-------
p. Exception to sample selection (586.1007-84(c)).
GM proposed that the last portion of this paragraph
be clarified and rewritten to include an exception that the
Administrator may approve a modification in the normal
assembly procedures. Although GM did not describe a situa-
tion in which such an exemption would be needed, the comment
has been adopted.
g. Allowance for "dealer preparation" procedures
(§86 .1008-8 4(b)(1)).
GM recommended that an additional paragraph be added to
the end of this section to reflect the current practice of
the LDV regulations. The recommended new paragraph states that
a manufacturer may perform "dealer preparation" procedures on
the new vehicles or engines, provided that these procedures are
documented in written instructions or are approved by the
Administrator in advance of their performance. EPA believes
that SEA vehicles or test engines that have undergone dealer
preparation procedures will represent "real world" conditions
to the extent that these procedures are actually and correctly
performed by dealers, EPA's experience with LDVs indicates
that in several cases, dealer preparation procedures are not
performed, or are not performed correctly by the dealers.
However, the current regulations do permit dealer preparation
procedures to be performed if they are approved in advance
by the Administrator. EPA approval will be facilitated if
-173-
-------
the manufacturer orovides sufficient dealer survey data or
other information to allow FPA to conclude that the pro-
cedures are actually beina correctlv performed bv dealers.
Therefore, EPA did not alter this provision.
r. Se'rvice accumulation requirements ($86.1008-84(c}(1)).
The current FPA reoulations require that service accumu-
lation prior to engine testing be performed at a minimum rate
of 16 hours per 24 hour period, unless otherwise approved by
the Administrator. GM proposed an 8 hour minimum rate to make
this requirement more consistent with the LDV requlations, which
does not require the manufacturer to maintain a two shift oper-
ation at a test facility. GM did not justify on a technical
basis why test engines could not be run for a minimum of 16
hours per day. The Aqency would like to conduct the audits
in the most expeditious manner possible. We believe that
this is still a reasonable requirement because HDE service
accumulation does not require a full-time "driver" and can
be monitored automatically for emergency shut-down. In
addition, there is an existing provision in this paraqraph
of the reaulations that allows the Administrator to approve
an alternate service accumulation rate based on a justifiable
manufacturer request. Therefore, FPA has not made a revision
to this paraqraph in the final rule.
-174-
-------
s. Test per day requirement (§86.10 08-84(g)(1)).
GM recommended that EPA's requirement of a minimum of
two SEA tests per 24 hour period be revised to one test per
24 hour period on the averaqe. GM proposed this change to
make this regulation consistent with current IDV requirements.
GM stated that the LDV regulations require only the use of
a single test cell for the expeditious completion of an
audit, whereas, for HDEs the manufacturer would have to
dedicate two test cells for the purpose of an audit.
In its HDE SEA program, EPA desires to conduct the
audits in as expeditious and non-disruptive a manner as
possible while still obtaining accurate test results. Based
on the time required to perform the transient test procedure,
taking into account the ''forced cool-down" allowed in the
final rule (see Subpart N), EPA has determined that two
tests can be performed in a 24 hour period, given two test
cells (especially with double-ended dynomometers}. EPA used
this test cell requirement in its analysis of the cost of
these regulations. To require only one test per day
would make the SEA last about twice as long, with resultant
demands on both the manufacturer's and the Agency's resources.
If a manufacturer has a justifiable reason for being unable
to perform the minimum number of tests, the regulations
allow them (under §86.1008—84(g)(4)) to ask EPA for a
reduction. EPA has therefore made no change in the test
-175-
-------
per day requirement in this paragraph for the final rule.
t. Option to retest (§86.1008-84(i)) .
wanted EPA's present regulation revised to allow
retesting at any time during an audit (as opposed to only
after a fail decision has been reached) and to delete the
requirement for testing each engine or vehicle the same
number of times. It justified these changes on the basis of
possible logistic, storage, and economic impacts on manufac-
turer operations and comparability with the LDV SEA regulations.
To permit a manufacturer to retest, before an actual failure
has occurred, may unnecessarily delay the audit and may even
cause the negative impact on operations that GM wished to
avoid. There is nevertheless, an existing provision in this
paragraph that allows the Administrator to approve retesting,
before a fail decision has been reached, based on a manufacturer's
request accompanied by a satisfactory justification. However,
the engines or vehicles must still be tested the same number of
times. To permit retesting of only failed engines or
vehicles or to allow some enqines or vehicles to be tested
more times than others will bias the test results from a
statistical viewpoint because of inherent test-to-test
variability. EPA has therefore made no changes to this
paragraph for the final rule.
-175-
-------
u. Failed engine or vehicle report (§86.1012-84(i)(2)).
CM proposed that this paragraph be changed to delete
the requirement that a written report to the Administrator
be submitted within five working days after successful
completion of testing on a failed vehicle or engine. GM
stated that this change will more accurately reflect the
current requirement of the LDV regulations.
In the original HDE/IOT NPRMs, EPA proposed regulations
that required the written report on corrective testing of
engines or vehicles that failed emission testing during an
SEA be submitted to EPA within five working days after
completion of that testing. While EPA needs to receive
reports on the repair of noncomplying engines or vehricles in
a timely manner, the Agency acknowledges that corrective
action need not be taken immediately after an engine or
vehicle failure. To clarify its intent, RPA revised this
paragraph (in the January 21, 193 0 and September 25, 198 0
final rules) so as not to limit the time a manufacturer may
take to complete, testing of failed engines or vehicles. The
Agency also concurs with GM's statement that this paragraph
be chanqed to reflect the current requirements of the LDV
regulations, which do not require a five working day time
limit for submission of failed vehicle reports. EPA will
revise paragraph §86*1012—84(i)(2) to relect GM's comment
and the LDV regulations.
-177-
-------
v. Applicability (586.1001-84).
American Motors (AM) commented that the applicability
in 586.1001-84 was not changed to reflect the deletion of
HPEs fdr the 1984 model year, and that LDTs are still
included in §86.601 applicability (Subpart (1). AM goes on
to say that FPA could avoid much confusion and simplify the
regulation by continuing the reauirements of Subpart G for
LdTs until such time as EPA requires HDE SEAs (1986 model
year). AM opposes grouping LDTs with HDEs for any emission
certification or compliance related matters. Therefore,
it recommends postponing the applicability of Subpart K
until 1986.
On January 21, 1980, EPA promulgated gaseous emission
regulations for 1984 and later model year HDEs and a similar
rulemaking affecting 1984 and later model year LDTs was
promulgated on September 25, 1980. The primary function of
these rulemakings was to promulgate the statutory HC and CO
standards called for in the 1977 Clean Air Act Amendments
(202(a)(3)(A)(ii)). In addition to the statutory standards
these rulemakings implemented a number of other provisions
to be effective for the 1984 model year, such as: sequential
sampling plans for SEA, revised certification requirements, a
revised useful life definition, and an idle test and idle
emission standard for gasoline-powered LDTs and HDEs.
These new requirements for LDTs and HDEs were promulgated
simultaneously to avoid the procedural disruption and waste
-178-
-------
associated with freoruent changes in emission regulations.
The Agency chose this comprehensive approach to controlling
IDT and HDE emissions because it was the most efficient
approach in that it allows the manufacturers to deal with the
effects of several regulations at once. This will avoid
repeated financial outlays for research, development,
recertification, and retooling.
The applicability of $86.1001-84 (Subpart K) was not
changed to reflect the deletion of HDE SEAs for 1984 because
the regulations still apply for 1984 and later model year
HDEs and LDTs ( LDTs are still subject to SEAs under Subpart
K starting in the 1984 model year - LDTs are currently
subject to SEAs under the provisions of Subpart G). The
Agency, as part of their regulatory relief initiative, has
made a commitment to the HDE manufacturers not to begin the
HDE SEA program until the 1986 model year. In addition, Sub-
part K will be used to implement the nonconformance penalty
(NCP) provisions of Section 206(g) of the Act, which, where
applicable, may apply to LDTs (greater than 6000 pounds GVW) as
well as HDEs. As far as §86.601 (the applicabilitv provision in
Subpart G) is concerned, LDTs are still included because this
provision for LDTs is effective through the 1983 model year.
In another EPA rulemaking, we intend to propose several changes
-179-
-------
to Subpart G to provide greater consistency with Subpart K.
That proposal would make the sampling plans, test procedures,
etc. coincide for LDTs and LJDVs. Therefore, EPA has not
revised paragraphs §86.601 or §86.1001-84 in response to
these comments.
-180-
-------
6. Issue: Split Standards - Gasoline-Fueled vs. Diesel
Engines
Summary of the Issue
Virtually all HDVs are powered by either gasoline- or
diesel-fueled engines. In the past, the gaseous exhaust
emission standards for each pollutant (i.e., HC, CO, NOx) have
been the same for all HDVs regardless of the type of powerplant
used in the vehicle. This has been true even though the two
types of powerplants have different operating characteristics,
which lead to significantly different levels of the various
pollutants in the uncontrolled case. Diesel engines are
inherently low in HC and CO, while being relatively high in
particulates. Gasoline-fueled engines are relatively high in
HC and CO but inherently low in particulate. Both types of
engines produce similar levels of NOx in the uncontrolled case,
but gasoline engine NOx is more easily controlled.
In the past, emission standards were of a level of
stringency such that both types of engines could meet them with
relative ease. However, as tighter emission standards are
considered for the future, the difference in inherent emission
levels between the two types of engines may need to be
considered. An emission standard which is a practical lower
limit for one type of engine may be quite a high level of
emissions for the other type of engine. By setting separate
standards for the two types of powerplants, the maximum degree
of emission reduction might be obtained for all pollutants in a
cost-effective manner.
At the February 18, 1982 public hearing, for this
rulemaking as well as in its March 23, 1982 notification
extending the comment period, the Agency requested comments on
this issue of setting separate standards for gasoline- and
diesel-fueled engines. The comments received are summarized
and analyzed below.
Summary of the Comments
EPA received comments from seven HDV manufacturers and one
trade association on this issue. The seven manufacturers
were: Caterpillar Tractor Company, International Harvester
Company (IHC), Chrysler Corporation (Chrysler), Mack Trucks
(Mack), Daimler-Benz A.G., General Motors Corporation (GM), and
Ford Motor Company (Ford). The trade association was the
Engine Manufacturers Association (EMA). Only one major
domestic engine manufacturer, Cummins Engine Company, provided
no position on this issue.
Caterpillar recommended "...the complete separation of HD
diesel regulations from HD gasoline regulations." Caterpillar
-181-
-------
stated that it would not be opposed to a less stringent HC
standard for gasoline-fueled HDVs as compared to the diesel
standard/ but it requested that similar consideration be given
to future NOx standards.
International Harvester Corporation stated that the
standards mandated by the 1977 amendments to the Clean Air Act
{CAA) are not feasible when the need to avoid excessive cost
and fuel economy losses is considered. Therefore, IEC
recommended that the Agency revise all of these standards to
utilize , the best control technology while minimizing costs.
IHC recommended that "...in the future EPA should set standards
based upon best and most cost-effective emission control
technology taking into consideration the type of basic engine
(diesel or gasoline)." IHC did not oppose the concept of split
standards.
Chrysler stated that it saw no compelling reason for EPA
to consider establishing separate standards that would be
applied to comparable vehicles performing similar operations.
Chrysler claimed that the CAA made no provisions for separate
standards for different types of engines. It quoted from
§202 (a)(3)(A)., which authorizes the Administrator to establish
classes based on "gross vehicle weight, horsepower, or other
factors as may be appropriate," but Chrysler claimed the
legislative history is clear that the use of diesel fuel is not
a proper determinant for establishing a separate class of
eng ines.
Mack stated at the public hearing that since gasoline
engines are not in competition with its diesels, it has "...no
problem with a different standard for gasoline." Mack also
expressed a hope that "similar concessions might be in the
cards for the diesel should they be needed," with regard to
future NOx or particulate standards.
Daimler-Benz agreed with EPA that "...the Clean Air Act
permits the establishment of separate standards for diesel and
gasoline engines based on the technical capability of each
engine class." Furthermore, Daimler-Benz claimed that it is
appropriate to establish separate standards for all regulated
pollutants.
General Motors stated that it did not believe separate
standards should be established for gasoline-fueled vs.
diesel-fueled engines, and gave three main reasons to support
its position. First, GM commented that the cost of control
increases sharply as the lowest achievable levels are
approached. As these low level of emissions are approached,
setting different standards of approximately equal stringency
would become very difficult. Therefore, a competitive
-182-
-------
advantage would be artificially induced for one or the other
engine type.
Second, GM stated that there would be little incentive to
develop a less costly, low-emission powerplant of new
technology which would be capable of complying with given HDE
emission standards if it is likely that after introduction that
engine type would be subjected to more stringent standards
based on the best available control technology for that
engine. The new engine's original cost advantage over other
available powerplants would be eroded.
Third, GM claimed that Congress intended one set of
standards to cover both engine types. GM quoted the House
report which accompanied the 1977 CAA amendments as saying, "In
permitting the Administrator to specify separate classes or
categories of vehicles or engines, the Committee did not intend
to authorize the Administrator to prescribe separate standards
for gasoline-powered and diesel-powered engines." Furthermore,
GM claimed that the standards which are specified in CAA
§202(a) (3) (A) (ii) are to apply to all heavy-duty engines.
General Motors was also concerned that if separate
standards were promulgated someone might erroneously conclude
that the diesel HC and CO standards should be 90 percent
reductions from uncontrolled diesel levels. GM claimed such
levels would in some cases be infeasible, thus resulting in
unnecessary and expensive regulatory activity to periodically
revise the standards as required by the CAA.
Finally, GM noted that EPA discussed the issue in the
Advance Notice of Proposed Rulemaking (ANPRM) for LDT/HDE NOx
(46 FR 5838) . The ANPRM took the position that the same NOx
standard should apply to both types of engines within a given
class. GM also pointed out EPA's reasoning that it would be
inequitable to establish different requirements for competing
engines within the same class and that to do so could have the
appearance of favoring one powerplant over another. GM stated
that EPA's reasoning expressed in the NOx ANPRM is equally
applicable to the HDE HC and CO standards.
Ford claimed that "...the public interest is best served
by the establishment of uniform standards for competing classes
of vehicles." Since Ford did not elaborate on the above
statement, we will assume that Ford's comment was concerned
with the competitive effects similar to those expressed in GM's
comments.
Ford also claimed that EPA does not have statutory
authority to set separate standards for gasoline- and
diesel-fueled HDEs. Ford quoted from the 1977 report by the
-183-
-------
Senate Committee on Environment and Public Works which
accompanied Senate Bill 252. The Committee stated, "Diesel
vehicles, which inherently emit less hydrocarbons and carbon
monoxide, must meet the standards set for gasoline-powered
vehicles." Also, Ford submitted the same quote from the House
Committee on Interstate and Foreign Commerce that GM submitted.
F6rd. claimed that Congress' intent to establish uniform
standards is further evidenced by the fact that Congress did
allow some separate standards, but it did so in a very limited
and specific fashion. For example, unique NOx standards for
1981-84 diesel-powered light-duty vehicles (LDVs) that qualify
for waivers, and waiver provisions for "small manufacturers"
and "innovative technology," are specifically authorized.
Finally, Ford stated that EPA itself has recognized the
need for uniformity of standards. Ford, as did GM, pointed out
EPA's intention to propose a NOx standard for all HDVs that
represents the level that can be achieved by diesel engines.
Ford claimed that this same uniformity must apply to all
standards.
The Engine Manufacturers Association recommended that the
Agency propose a rule in response to which interested parties
could comment. EPA should consider how separate standards
might correct the problems created by the statutory NOx
standard which the Agency "...has already indicated it believes
is not technologically feasible." EMA suggested that EPA
consider the establishment of future standards based on more
representative baselines or control technologies. EMA stated
that it would submit additional comments at such time as the
Agency articulates a policy which addresses the issues.
Analysis of the Comments
The four commenters who produce only HDDEs stated that EPA
could set separate standards for the two engine types. In
fact, most of these commenters urged EPA to do so. These
commenters felt that split standards were consistent with that
requirement in the CAA for EPA to consider the impact of
available technology in setting standards.
General Motors, Ford, and Chrysler all produce HDGEs, and
all were opposed to split standards. GM also produces HDDEs.
Three reasons for opposition were common to both GM and Ford:
1) competitive effects, 2) statutory intent, and 3) EPA
precedent. Chrysler's main reason for opposition was statutory
intent. Each of these will be discussed in the order given,
followed by discussion of the other two concerns of GM (i.e.;
infeasibility of 90 percent reduction and incentives for new
technology).
-184-
-------
EPA agrees with the commenters that the possibility of a
competitive advantage being established does exist. This
situation might occur if the Agency attempted to set standards
at the very lowest possible emission limits of the two engine
types. Since the cost of control often increases rapidly as
such limits are approached, great care would need to be taken
to assure that such separate standards would not lead to an
unreasonable cost differential for the two engine types. (This
clearly would not apply in cases such as evaporative HC or
particulate/smoke standards, where only one engine type is
regulated.)
However, split standards would not necessarily lead to a
competitive advantage for one or the other engine type. For
example, the public record for this rulemaking clearly
indicates that HDDEs are already achieving the statutory CO
standard. HDGEs, on the other hand, could have substantial
difficulty meeting that standard by 1985, and in fact would
need to utilize an oxidation catalyst-based control system in
most cases. Thus, having the same standard for both engine
types results in a large initial cost disadvantage for HDGEs.
Even when the gasoline standards are relaxed to levels where
catalysts would not be necessary, but the diesel standards
remain at the statutory level, the average incremental diesel
engine cost would be less than the average incremental gasoline
engine cost. In this case, the promulgation of split standards
would clearly promote equity for the two engine types, while
retaining the same standards results in a definite cost
advantage for diesels. Therefore, EPA has determined that
while the possibility of creating a competitive advantage due
to split standards does exist, each individual instance must be
carefully analyzed on its own merits to determine if such an
advantage would be created.
EPA disagrees with the claim by GM, Ford, and Chrysler
that the CAA disallows the setting of separate standards for
HDGEs and HDDEs. While GM and Ford raise valid points of
legislative history, it is important to realize that no action
was ever taken to write these Committees' intents into the CAA,
nor is there any indication that such intents were endorsed by
the conference committee or the Congress as a whole in
establishing the final 1977 amendments. Moreover, although the
legislative history cited by both GM and Ford may indicate the
House Committee's intent as to how EPA should exercise its
discretion, the quotation is more suggestive than mandatory.
The actual wording of the CAA, on the other hand, confers broad
authority on EPA in this area. According to §202(a)(3)(A)(iv) ,
the Administrator "...may base such classes or categories on
gross vehicle weight, horsepower or such other factors as may
be appropriate" (emphasis added).
-185-
-------
The CAA obviously established the same statutory standard
for both engine classes in §202(a)(3)(A)(ii), as GM pointed
out. EPA continues to move toward that goal. However, in the
provisions for temporary revised standards the CAA calls for
technology-based interim standards. The technology, cost,
leadtime, and fuel economy considerations involved in
establishing such interim standards are all fundamentally
engine-type dependent. Thus, it would seem that consideration
of basic engine type is clearly an "appropriate factor" under
the §202 (a) (3) (A) (iv) definition, and that the CAA allows the
Agency to establish split standards for HDDEs and HDGEs.
It should also be remembered that distinctive,
technology-related standards for HDEs are not foreign to EPA's
application of the CAA. Smoke emission standards currently
apply to HDDEs, but not HDGEs. EPA has also proposed
particulate emission standards to be applicable to diesel
engines only, and has recently established evaporative
emissions standards for HDGEs only. In none of these cases has
any question been raised as to the appropriateness of split
standards.
Both GM and Ford claimed that statements made by EPA in
the LDT/HDE NOx ANPRM (46 FR 5845), which indicated the
Agency's intent to propose a single revised NOx standard for
both gasoline and diesel engines, must apply to this rulemaking
as well. EPA disagrees with the commenters' claim. Those
statements did not reflect a final Agency policy .statement, but
indicated a preliminary EPA position on the single vs. separate
standards issue for HDE NOx, published for.public comment in an
ANPRM. That position was clearly subject to change as is
Agency policy in general. This is especially true when
circumstances and conditions change or when new regulatory
situations arise. It is in this light that EPA has raised the
issue of split standards for HDE HC and CO.
The possible application of split standards for HC and CO
likewise should not be taken as precedent setting for HDE NOx.
EPA analysis here indicates the Agency's authority to set such
standards, and will momentarily discuss further considerations
in any decision to use this authority. All of this analysis
should make it clear that the approach for NOx standards could
be different than the approach for HC and CO standards. For
example, in the range of standards now being considered for HDE
NOx, gasoline and diesel engine control costs are similar for
the same standards. However, if a single set of HC/CO
standards were adopted, such as the statutory standards, then
inequitable costs due to widely differing technology
requirements could result. Therefore, a single set of HC/CO
standards might be inappropriate because of significantly
different costs between two engine types, while a single NOx
-186-
-------
standard might be appropriate because emission control costs
are nearly the same for gasoline and diesel engines.
General Motors was concerned that the incentive for
developing innovative technology could be diminished if the
standards applicable to such new technology are overly
stringent. When and if the Agency considers standards for new
technology engines, it will need to consider this valid
concern. However, this rulemaking does not involve new
technology engines, and is unlikely to have an impact on the
development of new technologies. Therefore, while the Agency
may need to evaluate this concern in future rulemakings, the
problem does not arise in this final rule.
General Motors' final concern was that if split standards
were developed, then the Agency might consider as standards 90
percent reductions from uncontrolled levels for both diesel and
gasoline engines. GM stated that a 90 percent reduction in HC
and CO for diesels is technologically infeasible. We conclude
that GM's concern is ill-founded. As already stated above, the
CAA obviously established the same statutory standards for both
engine classes. EPA recognizes that diesel engines are
inherently low emitters of HC and CO. EPA always has and will
continue to analyze the technical feasibility of standards it
promulgates.
The Engine Manufacturers Association suggested that the
Agency propose a rulemaking on split standards. EPA concludes
that since the Agency has requested comment on this issue both
at the public hearing and in a published notice (47 PR 1236.6) ,
and has received substantial comment, the requirements for
establishing the Agency's position have been met. Thus, a
separate rulemaking on split standards is not necessary.
Conclusion
EPA concludes that the CAA gives the Administrator
authority under §202(a) (3) (A) (iv) and §202(a) (3)(C)(i) to set
split standards. Furthermore, while there may be in some cases
potential problems concerning competitive advantage and
innovative technology incentives, each situation must be
analyzed on its own merits. Therefore, split standards will be
employed where necessary and appropriate, as is the case for
this rulemaking,
-187-
-------
7. Issue: Colt! Start Requirements
Summary of the Issue
The heavy-duty engine (HDE) emission test procedures
applicable to 1984 and later model year HDEs require that the
test begin with a cold engine, and that a 1/7 weighting be
applied to the emissions measured from the cold start segment
of the test. Comraentecs questioned the need for the cold start
requirement and disagreed with the cold start weighting.
Summary of the Comments
Commenters stated that the cold start requirement for
diesel engines is not necessary because the weighted emission
results (1/7 of cold start and 6/7 of hot start) are almost
identical to the emission results obtained from the hot start
portion of the test. Comparisons of the results obtained
either by individual manufacturers or from EM A/EPA round-robin
tests were expressed in several forms to support the position
that the cold start requirement is not necessary.
For diesel engines, the comparisons were expressed in
terms of a correlation coefficient. Daimler-Benz expressed
their results as a range of differences between the ratio of
cold start results to hot start results. The Engine
Manufacturers Association (EMA) analyzed the data to determine
the predicted "error" at the 95 percent confidence level in the
emission results if based upon only hot start test data. The
projected errors were: 0.1 grams per brake horsepower-hour
(g/BHP-hr) NOx, 0.08 g/BHP-hr . HC, and 0.05 g/BHP-hr
particulate. These errors were also noted to be less than the
variations seen from one test to the next.
For gasoline engines, commenters made the comparison in
terms of the ratio of hot start emissions to weighted
emissions, and expressed the result as a percentage. All of
the comments indicated that the cold start test had very
little, if any, effect on total test results.
The reasons given by the commenters for the good agreement
between hot test results and the weighted results (for both
diesel ana gasoline) were: 1) engine warm-up requires about
five minutes, and this warm-up period occurs in the very first
segment of the cold start portion of the test, 2) the exhaust
mass flow rates are low during the first segments of both the
cold start and hot start portions of the test, while being high
and essentially equal during the third segment of both portions
of the test, 3) the high exhaust gas mass flow rate of segment
three tends to overpower the effects of the other segments, and
4) application of a -weighting factor to the cold start portion
-188-
-------
of the test further reduces the overall effect of these
emissions on the final weighted test result. Commenters agreed
that HDEs are started from cold, but argued for the elimination
of the cold start requirement because of its minimal impact on
total test emissions.
Commenters stated that significant cost savings would be
realized by manufacturers through removal of the cold start
requirement. The cost savings would accrue from: 1) better
utilization of test facilities, 2) fewer test cells would have
to be built because of better test cell utilization, 3)
reduction in the number of lost tests (instrumentation and
hook-up of equipment cannot be checked prior to the start of a
cold test), and 4) reduction in development and certification
leadtimes. Forced cooldown does not solve the facility problem
because it still requires four to five hours to perform and
still results in only one test per day. Cummins provided an
estimate of the costs attributable to lost tests associated
with the cold start requirement. The estimate was between
$160 ,000 and $200 ,000 per year.
Commenters also disagreed with the weighting applied by
EPA to the cold start portion of the test. Commenters stated
that the CAPE-21 data showed that 1.6 percent of total vehicle
operation was with a cold engine. Commenters stated that if
EPA believes that test engines must be cold started, it was
recommended that the weighting for the cold start portion of
the test be changed. Ford recommended that the weighting for
the cold start portion of the test be 1/16 instead of 1/7. The
1/16 weighting was developed by making each of the four
segments of the cold start portion of the test equal to 1.6
percent of the total with the resulting cold start portion
equal to 6.4 percent or 1/16 of the total. As part of this
issue, commenters also disagreed with the EPA methodology foe
determining the number of truck trips per day. Commenters
stated that, on the basis of the CAPE-21 data, the average
truck was used for nine trips per day (based upon mean values)
and that it was an error to use the median number of truck
trips per day, as EPA had done, as the basis for the cold start
we ighting.
In final comments submitted by May 6 , 1983, EMA also
presented a detailed reanalysis of the cold start weighting
factors. EMA used mean values for calculating total operating
time per day per CAPE-21 truck (EPA used median values) , and
alternately, used median values for operating time, but
increased all median values to the extent necessary for median
total accountable time to equal an 8-hour day. Using either
method, EMA derived and recommended a cold cycle weighting
factor of .03. (By deriving total operating time per aay per
truck, by assuming a single cold start per day, and by knowing
-189-
-------
reasonably well the warm-up time for a typical engine, the
percentage of time an engine spends "cold" per day is easily
calculated.) EMA recommended that EPA adopt the .03
weighting. Furthermore, considering the relative stringency of
the HC standard, EMA argued that the entire cold start cycle be
abandoned as unnecessary. "However, if EPA wishes to retain a
check on new technical developments as they affect cold
starting, then EPA should permit a much reduced cold start
measurement effort. For example, EPA could adopt a method
similar to the CO emission measurement waiver...."
Analysis of the Comments
Cold Start Requirement for Diesel Engines
Diesel engines designed to meet existing emission
standards show good agreement between the hot start and the
composite test results, as measured by the ratio of hot start
results to composite test results.
For most current technology diesel engines, EPA agrees
that there is little difference between the hot start segment
result and the composite result. Comments submitted by EMA and
Cummins substantiate this fact. For many diesel engines, EPA
cannot find fault with the argument that the cold start cycle
has a marginal effect on total test results. EPA also
recognizes the economic implications of 100 percent cold start
testing. A significant percentage of dynamometer space is
idled while engines are cooling (thereby increasing the number
of dynamometers necessary for a given program, and thus
increasing the facility expenditures). Additional cost is
involved both in running the extra cold cycle and in procuring
equipment necessary for forced engine cooling. EPA concurs
that it makes no sense to impose a costly cold start testing
burden if no benefits are to be achieved.
On the other hand, EPA is reluctant to remove the cold
start requirement entirely for heavy-duty diesel engines
(HDDEs}. Some engines do show a difference between cold and
hot emissions. Furthermore, HDDEs have yet to experience the
most technologically difficult emission reductions (i.e., NOx
and particulate) . These will probably require the use of new
and elaborate emission control techniques. It is EPA's
experience with other internal combustion engines that, as
emission standards become more stringent, unique operational
modes such as cold starts take on greater significance and
contribute more to the total test result. This may also be
observed in future HDDEs. If so, a cold start test will become
increasingly necessary. However, if the cold start
requirements are abandoned today, they will be administratively
difficult to reimpose in the future when they may be most
needed.
-190-
-------
An alternative to the "all-or-nothing" approach has been
suggested by EMA. EMA recommends that an approach be taken
similar to that taken with the measurement of CO emissions from
HDDEs. HDDEs emit CO well below the level of the applicable CO
standards. In recognition of both this and the expenses
incurred in measuring CO, EPA waived the requirement for HDDE
manufacturers to report CO emission levels. This waiver was
made with the explicit condition that CO emission standards
Still apply, and that the risk of non-compliance still rests
with the manufacturer.
Such an approach is appropriate for heavy-duty diesel cold
start test results. Under this approach, EPA could allow
submission of only hot start data in certification
applications. The official test procedure, however, will
remain a cold/hot test which will still be run for all
confirmatory tests. The manufacturer would then accept any
jeopardy arising from potential differences in test results.
As always, a manufacturer may run whatever tests deemed
necessary for in-house development testing. In this way, the
cold start testing burden is minimized if a manufacturer is
confident that a cold start is actually insignificant. In
fact, marginally greater cold start emissions may be adequately
simulated much the same way that expected in-use deterioration
is: by downwardly adjusting hot start emission target levels.
On the other hand, if a cold start is indeed significant for a
given engine, these are the very engines on which cold start
testing should be performed. Since the jeopardy of
non-compliance still would rest with the manufacturer, EPA has
sufficient assurance that necessary testing will take place.
Cold Start Requirement for Gasoline Engines
EPA has reviewed the emission data collected during its
baseline testing programs. For uncontrolled engines, about 11
percent of total hydrocarbons measured over the transient test
were attributable to the cold start segment. In later testing
(the 1979 current technology baseline), the cold start
contribution to composite test results ranged from 4.5 to 37.7
percent for HC, and from 1.5 to 10.2 percent for CO. As total
emissions decreased, the percentage contribution of the cold
start was observed to increase. Finally, for emission tests on
engines equipped with catalysts, the cold start test dominates
total HC emissions, and becomes a greater percentage of total
CO emissions. This finding is nothing new or surprising: all
testing on catalyst-equipped vehicles substantiate the
importance of the cold start on the emissions of
gasoline-fueled vehicles.
The implications of this data are clear. For current and
future technology heavy-duty gasoline-fueled engines (HDGEs),
-191-
-------
the cold start is not only significant for HC and CO emissions,
in the future it will become the dominant source of HC
emissions. This conclusion was challenged to some extent by
the gasoline engine industry. Emissions data from prototype
1985 engines were submitted which indicated that the cold start
had very little effect on total emissions. However, these data
were collected on engines with cold start emission control;
without a cold start test, such control would not be necessary
and cold engine emissions would again be significant.
EPA continues to believe that the cold start test is
critical for accurate characterization of HDGE emissions, and
should be retained.
Cold Start Weighting
In the "Summary and Analysis of Comments to the NPRM: 1983
and Later Model Year Heavy-Duty Engines, Proposed Gaseous
Emission Regulations" (December 1979), EPA showed that the
average percentage cold operation observed in the CAPE-21 study
for gasoline trucks was 5.5 percent and that for diesel trucks
it was 4.3 percent. These values were developed from the
median number of trips per day per truck (4.43 for diesels and
9.06 for gasoline) , and the median time of each trip (26
minutes for diesels and 10 minutes for gasoline). EPA also
assigned a cold operating period of five minutes only to the
first trip of the day, thereby treating all other trips as hot
start trips. (In practice, some of these other trips will be
started from temperatures colder than fully warmed-up because
of engine cooling between trips.) The Summary and Analysis of
Comments document went on to determine the percentage of cold
operation during the cold start portion of the test and
compared these results to the CAPE-21 data. During the cold
start segment of the test, cold operation was calculated to be
3.7 percent for gasoline engines and 3.6 percent for diesel
engines. Based upon the comparison of the test cycle's
percentage of time in cold operation to that of the CAPE-21
data, EPA concluded that the test slightly understated the
on-road condition.
In recent comments, both EMA and the gasoline engine
manufacturers have disputed the derivation and values of EPA's
weighting factors. Ford recommended a cold start weighting of
1/16 (.0625); EMA recommended a weighting of .03. The
differences between EPA's and the industry's weighting factors
are based upon two differences in assumptions:
1. EPA used median CAPE-21 values as the necessary
parameters to calculate a cold start weighting, while the
industry used mean values; and
-192-
-------
2. The industry assumed that the entire cold start test
cycle was "cold", (i.e., the engine did not warm-up) and
continued to produce cold emissions during the entire 20-minute
cycle.
EPA believes that its assumption in 1. above is more
reasonable than the industry's; EPA also believes that the
industry's assumption in 2. above is incorrect.
The truck population sampled in CAPE-21 was highly
diverse. Any given parameter, especially those used to
determine the cold start weighting, was decidedly non-normal
(see Figure 1.) In non-normal distributions, medians are far
better indicators of central tendency (i.e., the "typical"
truck) . Means tend to be skewed by a small number of very
different parameters. For this reason EPA's use of medians
represents a more reasonable method of determining "typical"
values.
Also, to hold that the engine remains cold during the
entire cold start cycle is incorrect. Oil temperature data
gathered by EPA, and EMA's own test data indicate that the
engine reaches a warmed-up state somewhere between 5 and 10
minutes into the test. In other words, emissions during the
remaining portion of the cold test cycle are no different than
those of the warmed-up hot cycle, and for this period of time
the weighting factor value is irrelevant. If we assume that
the first five minutes, or 300 seconds, of the 1,199 second
diesel test cycle are actually cold, then 300/1,199 or 25.0
percent of the cycle is cold. Since the entire cold cycle is
then weighted by 1/7, the cold engine emissions are actually
weighted by 1/7 x .250, or 3.6 percent of the total test
result. If we continue to make the assumption that the engine
warms-up in the first five minutes of operation, as did EPA
when it derived its original weighting factors, we find that
the percentage of cold operation in the test cycles are 3.6
percent for diesels and 3.7 percent for gasoline-fueled
engines. If a 5-minute warm-up is similarly assumed for the
first trip of the day from the CAPE-21 data, the actual
percentages of on-road time spent with a "cold" engine are 5.5
percent for gasoline-fueled engines and 4.3 percent for diesels.
In summary, EPA cannot agree that the present cold start
weightings are unrepresentatively high; if anything, they might
understate those observed in CAPE-21.
Conclusions
1. The cold start requirement for both gasoline and
diesel-fueled HDEs will be retained.
-193-
-------
Fiaure 1
tjhe frn -rnjp oriisJTr fuuctjom ron wr en sou n't!
D.
I 11 i i—i i i
i 11—i r—i
10 30 50 70 00 I ID 130
20 HO 00 00 100 120 1H0
i hie-rrn-trip tiiihutrsi
-194-
-------
2. The present 1/7 cold start weighting will be
retained
3. Diesel engine manufacturers may report only hot
start, data when making application for certification. For
confirmatory, SEA and recall testing, however, EPA will retain
the option of using either the hot start or cold start tests;
the cold start test will be retained as the official test.
-195-
-------
8. Issue: Diesel Engine CO Measurement
Summary of the Issue
Measurement of CO emissions is currently required for all
regulated vehicles and engines, including HDDEs, during both
certification and SEA testing. All of the comments received on
this issue were in agreement: the CO emissions of HDDEs are so
far below all applicable standards, current and proposed, that
any requirements for HDDE CO measurements are unnecessary and
should be deleted.
Summary of the Comments
All of the commenters referred to the very low CO
emissions of HDDEs, exemplified by the mean CO level of 3.27
g/BHP-hr for all 1981 model year certified HDDEs, less than
one-sixth of the 25 g/BHP-hr standard applicable under the
13-mode test. The cost savings to the manufacturers resulting
from deletion of the diesel CO test requirements are estimated
at $20,000 annually per manufacturer by Mack and EMA. Cummins
estimated that deletion of these requirements would result in a
25 percent annual reduction in their equipment, maintenance,
and storage costs. Several manufacturers noted that valuable
laboratory test time would also be made available if the CO
test requirements are deleted for HDDEs.
The manufacturers also indicated that they feel that the
deletion of diesel CO test requirements should be completed as
soon as possible, rather than awaiting the final implementation
of the transient test procedure as was originally proposed.
All of the manufacturers commenting noted that since HDDE
CO emissions are inherently so low, they had no substantive
comment on the proposed revisions to the level of the standard.
Analysis of the Comments
As was noted in several of the comments, EPA has already
agreed to delete all certification and SEA CO measurement
requirements for HDDEs. This action has been implemented
through technical amendments to the regulations. Revision of
the diesel CO measurement requirements took effect as soon as
those amendments were issued (47 FR 49802, November 2, 1982).
This action is appropriate as a regulatory relief measure, and
is not expected to have any negative air quality impact.
Conclusion
No further action is necessary, since the requested
changes have been made.
-196-
-------
9. Issue: Parameter Adjustment
Summary of the Issue
The original FRM (45 FR 4136, January 21, 1980) contained
regulations describing the parameters subject to adjustment
during certification and SEA testing. Few comments were
received on this issue. The only significant concern of the
manufacturers appeared to be the possibility of manual choke
settings being subject to the parameter-adjustment regulations.
Summary of the Comments
Two of the three manufacturers commenting indicated
concern over the possibility that, under a strict
interpretation of the current rules, the Administrator could
require manual chokes to be adjusted over the full range of
their authority during certification and SEA testing. The
operation of a manual choke should not be confused with
tampering, one manufacturer noted. A slight revision to the
wording of §86.084-22(e) (1) (i) was suggested by the second
manufacturer as a means of removing any ambiguity concerning
the applicability of the parameter-adjustment regulations to
manual chokes.
One manufacturer noted that EPA had previously determined
that the parameter-adjustment regulations do not apply to idle
speed or to ignition spark timing. The comment suggested that
references to these two parameters be deleted from
§86.084-22(e)(1)(i) in accordance with this determination.
Analysis of the Comments
EPA agrees that the manual choke should not be considered
a parameter in the context of the parameter-adjustment
provisions. While a manual choke is clearly an adjustable
parameter, its operation is governed by Subpart N test
procedure provisions and therefore manual chokes were not
included in the list of adjustable parameters of
§86.084-22 (e) (1) (i) . In fact, to add manual chokes to that
list would require public notice and comment plus a minimum of
two years of leadtime. Inclusion of choke operation under
Subpart N is based upon viewing a manual choke as an operating
control (as are, for example, shift points on a manual
transmission) rather than a parameter subject to the
parameter-adjustment provisions. EPA believes this is the
appropriate approach.
EPA also believes that to subject manual chokes to the
parameter adjustment provisions would in effect prohibit
-197-
-------
their use. At this time, EPA has no evidence of systematic
improper use of chokes on HDEs to iustify such an action. In
addition, it is likely that as future emission reductions are
required, manual chokes will gradually be phased out of use by
manufacturers (similar to what has already happened for LDVs).
Concerning the request to delete idle speed and ignition
spark timing from the regulations, the Agency has already taken
the necessary steps to implement its findings that these two
parameters need not be adjusted. Manufacturers were notified
by two letters that these parameters would not be subject to
adjustment under the parameter-adjustment requirements.[1,2]
Since the list of parameters in §86.084-22(e) (1) (i) is
discretionary rather than mandatory, EPA sees no need to make
changes to the regulations.
Conclusions
In order to eliminate any possible ambiguity over the
adjustment of manual choke settings during certification and
SEA testing, EPA has decided to add a new paragraph
586.085-22 (e) (1) (iv) to read, "Manual chokes will not be
considered an adjustable parameter for HDEs subject to
adjustment under this paragraph." In addition, EPA will revise
paragraph §86.085-22 (e) (.1) (i) to read, "Except as noted in
§86.085-22(e) (1) (iv), the Administrator may determine...." EPA
has also decided not to change the references to idle speed and
ignition spark timing in §86.085-22(e)(1)(i).
-198-
-------
References
1. EPA Memorandum From Michael P. Walsh, OMSAPC, To
Light-Duty Vehicle and Heavy-Duty Engine Manufacturers, August
22, 1980.
2. Deletion of Spark Timing Parameter Adjustment
Requirement, EPA Memorandum From Michael P. Walsh, OMSAPC, To
Light-Duty Vehicle, Light-Duty Truck and Heavy-Duty Engine
Manufacturers, October 28, 1980.
-199-
-------
iO. Issue: Potential Impacts on Specific Manufacturers
Summary of the Issue
Several commenters claimed that even though EPA has
proposed revisions to many of the 1984 HDE requirements, these
provisions may still cause substantial harm to the industry.
The impact of these revised regulations on the HDGE
manufacturers' future product offerings and financial
situations, based on the comments received, are discussed in
this section. The impact of the rule on Chrysler is considered
separately from, the impact on the other manufacturers.
Summary of the Comments
Chrysler Corporation
In its initial submission to the docket, Chrysler
indicated that it was planning to withdraw from the HDGE market
in the near future. At that time, Chrysler stated that the
potential profitability of the HDGE market in the 1980's was
thought to be insufficient to justify directing scarce capital
resources into the development of the necessary transient test
facilities and the development of HDGE emission control systems
capable of meeting the revised standards. More recent comments
received from Chrysler indicate that, based on the improving
financial condition of the corporation and a reassessment of
the profit potential of manufacturing HDGEs, it is now planning
to remain in the market.
However, these recent comments also include several
reservations that Chrysler continues to have concerning the
regulatory requirements. Primary among these is the claimed
inability of Chrysler to develop transient test capabilities
for at least three years. For this reason, Chrysler suggests
the creation of a "small-volume manufacturer" category, defined
as any HDGE manufacturer building only engines that are derived
from passenger car engines? and that such "small-volume
manufacturers" be allowed to certify under the steady-state
procedure for up to three more years.
In its comments, Chrysler preliminarily rejected two
additional options (beside the extension of the steady-state
test option) that are available for compliance with these
regulations. First, vehicles up to 10,000 lbs. GVWR and
equipped with HDGEs may now be certified, at the manufacturer's
discretion, to LDT emission standards under the light-duty
chassis test procedure (FTP). During the public hearings, EPA
asked Chrysler representatives whether an increase in the
10,000 lbs. maximum, for example to 11,000 lbs., would make it
easier for Chrysler to take advantage of this certification
-200-
-------
option.. Its answer was> no; Chrysler does not currently plan to
certify iany of its 1984 vehicles in the 8,500-10,000 lbs. GVWR
range as LDTs, and in its opinion, increasing the upper bound
of this range offers no meaningful relief. The second option,
to use outside engineering services to conduct transient
testing .and certification, was rejected by Chrysler on the
basis of excessive cost and insufficient leadtime remaining
before scheduled compliance.
In response to EPA's inquiry as to what would constitute
an appropriate level for steady-state test emission standards,
in the event that this procedure is allowed as an option to the
transient test, Chrysler maintained that the present standards
cannot be made more stringent if engines without catalysts are
to meet them. It went on to state that with additional
development it may be possible for Chrysler to meet the current
steady-state standards without catalysts, but that the present
standards appear to be at the limit of non-catalyst emission
control technology. Chrysler does not feel that the 1984
California emission standards (0.5 HC, 25 CO, 4.5 HC+NOx) can
be met on a steady-state test without catalysts.
Chrysler also noted that it is the only manufacturer now
using catalysts on its HDGEs (5.2L and 5.9L engine families), a
decision that was made on the grounds that development costs
and manufacturing complexity would be minimized. Chrysler
maintains that it is confident that "real-world" emissions from
these engines are low,* thus, it claims it is ironic that the
proposed revisions to these rules, intended to make it possible
to certify HDGEs without catalysts, may result in Chrysler
being forced to withdraw from the market.
Other Commenters
In addition to Chrysler, comments on the potential impact
of these rules were also submitted by IHC, American Motors
Corporation (AMC), and the National Association of Van Pool
Operators (NAVPO). These comments are summarized below.
International Harvester Corporation has already made
public its intention to abandon the HDGE market when the
revised HDGE regulations take effect, after which it will only
manufacture HDDEs. This decision was based primarily on the
rapid and continuing decline in the demand for HDGEs, although
IHC noted that the implementation of a transient test procedure
was a contributing factor. If these regulations were to take
effect for the 1984 model year, as originally planned, IHC
would leave the HDGE market at the end of the 1983 model year.
Therefore, IHC has requested that the effective date of these
regulations be delayed until the 1985 model year, thereby
-201-
-------
allowing it to plan a more orderly withdrawal from the HDGE
market.
American Motors Corporation noted that it has not
certified any HDGEs in recent years; however, Renault is
planning the introduction of both HDGEs and HDDEs to the
medium-duty truck market over the next few years. AMC does not
have, or plan to acquire, transient test capabilities, but it
will be responsible for the certification of the Renault HDGEs
when they are introduced. Therefore it intends to contract for
this work, although it expressed some concern over the
availability of and competition for independent laboratory
time. Given this background information, AMC requested that
the implementation of the 1984 HD standards arid test procedures
be delayed until 1985, and that EPA "...consider waivers for
low-volume (less than 10,000 units) domestic manufacturers."
The National Association of Van Pool Operators' comments
were concerned entirely with the potential impact of these
rules on the manufacturers' product offerings. They expressed
concern over the possibility that the 12- to 15-passenger vans
that are most economical for van-pooling programs may no longer
be available if the emissions regulations applicable to them
are strengthened. A later conversation between EPA and a
representative of NAVPO revealed that their concern is focused
on the larger passenger vans manufactured by the Chrysler
Corporation.
Analysis of the Comments
Chrysler Corporation
As noted in the summary of Chrysler's comments, Chrysler
has decided to remain in the HDGE market. This decision must
have been based on the improving financial condition of the
company, as well as the belief that the profit potential of
HDGE manufacturing in the 1980's will be sufficient to justify
the necessary capital expenditures. The significant
stabilization of gasoline prices late in 1981 and in 1982, and
the 1-year delay in the effective date of these regulations,
may also have contributed to Chrysler's reevaluation of its
decision.
EPA has estimated that the capital costs to Chrysler for
transient test facilities, plus additional engineering costs
for facilities checkout and engine development, would total
approximately $2.9 million in 1982 dollars. Considering that
Chrysler, in its first and second quarterly reports, showed
profits of $256.8 million for the first six months of 1982, it
appears that it currently has liquid assets adequate to
-202-
-------
underwrite this investment. While recognizing that the firm
currently has major debt servicing obligations, and that
therefore not all of these recent profits are available for
capital expenditures, EPA can only conclude that Chrysler is
capable of making these investments, given the decision by
management to do so.
Chrysler maintains that it will be unable to develop
in-house transient test capability for three years. It has
preliminarily rejected the options of - contracting with
independent laboratories for the development and testing
required during the next three years, and of certifying its
HDGE vehicles under 10,000 lbs. GVWR to LDT emission
standards. Claiming that both of these options are
unacceptable, it requests that a "small-volume" category of
HDGE manufacturers, defined so as to include Chrysler, be
allowed to certify HDGEs under the steady-state test for the
next three years. The implication is that Chrysler may not
remain in the HDGE market unless such an exemption is granted.
Chrysler's position on whether to continue to compete in
the HDGE market is primarily a business decision. The 1984
HDGE emission regulations have been discussed in the public
forum for more than four years, and should have been taken into
account in any earlier decisions by Chrysler regarding its HDGE
manufacturing operations. Other affected manufacturers have
made the capital investments necessitated by these rules (Ford,
GM), or have determined that it is more economical for the work
to be performed under contract {PMC), or have decided, that the
profit potential of the HDGE market is insufficient to justify
further capital expenditures in this area (IHC). These
decisions have been based on business considerations, as
Chrysler's eventual decision should be. EPA notes that it
appears that Chrysler could now afford to pursue either of the
options discussed above, which it has preliminarily rejected,
and thus that it has three approaches to meeting the
requirements of these rules available to it.
In its comments, Chrysler also noted it has elected to
equip two of its three current HDGE families (5.2L and 5.9L)
with catalysts "...in order to minimize development costs and
manufacturing complexity." While it might be ironic for the
only HDGEs currently equipped with catalytic emission controls
to be forced from the market by rules designed to negate the
necessity of catalysts, EPA does not feel that this will be the
case. Since production and sales of these catalyst-equipped
engines has continued while other HDGE manufacturers did not
use catalysts, the cost disadvantage resulting from catalyst
use must be relatively small. Additional development work,
aimed at improving the emission characteristics of Chrysler's
-203-
-------
HDGEs, will have to be conducted if Chrysler really wishes to
remain in the market.
Since Chrysler has decided that its position in the HDGE
market warrants the decision to stay in that market, it must be
willing to commit the necessary resources to the development of
transient test facilities and improvement of its HDGE line.
EPA cannot justify granting what would amount to a 3-year delay
in the effective date of these regulations to some, but not
all, HDGE manufacturers. This is particularly true since other
firms, as noted above, have undertaken the investments required
by these regulations, and since the financial condition of
Chrysler has now improved. In addition, the 1-year delay in
the effective date of these rules provides Chrysler (and the
other manufacturers) additional time for compliance.
Other Comments
The major interest of IHC is that the effective date of
these HDGE regulations be delayed for an additional year, so
that its planned withdrawal from the market may proceed in an
orderly fashion. Due primarily to leadtime considerations,
this is being done. As was indicated by IHC, its decision to
withdraw from the HDGE market was made more on the basis of
financial considerations than on the effect of these
regulations.
The position of AMC with respect to the certification of
HDGEs manufactured by Renault is recognized as the basis for
its decision to contract with independent laboratories for this
work. As noted previously, the effective date of these
standards is being delayed until the 1985 model year, as
desired by IHC and AMC. However, AMC's request that EPA
consider waivers for "low-volume domestic manufacturers" is
unclear. Historically, EPA has rejected requests for waivers
from the use of applicable test procedures, and EPA sees no
other suitable way to determine compliance. On the other hand,
waivers from durability testing requirements and certain other
certification procedures have been granted in the past and
would be available in this context. These waivers are
available to manufacturers whose combined U.S. sales of LDVs,
LDTs, and HDEs are under 10,000 units.
The National Association of Van Pool Operators' concerns
about the availability of 12- to 15-passenger vans under the
new regulations appear to be groundless. The only manufacturer
of such vehicles that was considering dropping out of the HDGE
market as a result of these regulations (Chrysler) has decided
to remain, as noted above. Aside from that decision, none of
Chrysler's passenger vans are currently certified as HDGEs--all
-204-
-------
are less than 8,500 lbs. GVWR--and so these rules should have
no effect on the continued production of these vehicles.
Conclusions
For the reasons discussed in the preceding analysis, EPA
does not feel that allowing Chrysler (or any of the other
manufacturers) to use the steady-state test rather than the
transient test for the next three years can be justified.
Therefore, EPA rejects the "small-volume manufacturer"
exemptions proposed by Chrysler and AMC.
The concerns of IHC about being able to plan its
withdrawal fron. the HDGE market in an orderly manner, and of
AMC about having adequate time to plan and contract with
independent laboratories, are addressed by the delay in the
effective date of these rules until the 1985 model year.
The concerns expressed by NAVPO are unfounded. EPA has
decided not to make further changes in these rules based on the
comments received from NAVPO.
-205-
-------
11. Issue; Transient Test Procedure - Technical Details
Summary of the Issue
On June 17, 1981, EPA solicited manufacturers for
information regarding operational aspects of running the
transient test (46 PR 31677). On January 13, 1982, EPA
reopened for comment all aspects of the transient test as part
of the proposed revisions to the 1984 requirements.
In their comments, the heavy-duty industry recommended
that large numbers of technical amendments be made to the
transient test procedure. These amendments were justified as
necessary on the basis of technical merit and cost reduction.
EPA has also recognized the need to modify specific
sections of the transient test. This has become apparent as
more actual testing exerience was gained by both EPA and the
industry.
Summary and Analysis of the Comments
Each comment and technical amendment is not significant
enough to justify devoting an individual section to its
discussion. Collectively, the amendments represent a
clarified, streamlined, and technically improved test procedure.
The format for this discussion will be a section-by-
section breakdown of the transient test procedure (Subpart N) .
Specific modifications will be noted, as will the rationale for
the changes. Note that some technical amendments were
requested by industry, while others are being made by EPA's
initiative.
Also note that some technical amendments were necessary in
the heavy-duty diesel engine smoke test procedure (Subpart I)
and the heavy-duty gasoline engine and light-duty gasoline
truck idle test procedure (Subpart P) . A list of these changes
will follow those of the transient test procedure.
A. Subpart N - Transient Test Procedures for 1984 and Later
Model Year Heavy-Duty Gasoline and Diesel Engines
Overview of Technical Amendments
Large numbers of technical amendments are being made to
the transient test procedure.
In general, amendments have been made to correct errors
and omissions, to clarify requirements, to minimize prior
-206-
-------
approvals by the Administrator for inconsequential deviations
from t.fte existing procedures, and to reduce costs associated
with running the test.
All changes except corrections of typographical errors are
listed below.
Specific Technical Amendments
The following sections from 40 CFR Part 86 are being
amended:
§86.1308-84 (a) Torque and speed accuracies
rereferenced. Eliminated need for
Administrator's approval for using
dynamometer currents for torque
measurement.
Most accuracies within the test procedure were respecified
to provide greater traceability to NBS standards. Also,
several manufacturers have developed methods for using
dynamometer currents as surrogates for direct torque
measurements; EPA is reasonably convinced that the
techniques are technically acceptable and need not have
advance EPA approval.
§86.1308-84 (b) Torque cycle verification equipment
accuracy changed from +3 percent to _+2
percent (to equal speed cycle accuracy).
Torque cycle accuracy was changed to be comparable to that
required for speed, to correct an earlier oversight.
§85.1308^84 (e) Clarification of dynamometer calibration
procedures, and rereferencing of
accuracies.
Existing dynamometer calibration procedures were unclear,
and required procedural clarifications. No substantive
technical changes have been made. Again, accuracies were
rereferenced to provide greater traceability to NBS
standards.
§86.1308-84(f) Added specification for mass fuel flow
measurement device for diesel engines.
The option for direct measurements of mass fuel flow for
diesel engines was added; this addition required inclusion
of an accuracy specification for the flow measurement
equipment.
-207-
-------
§86.1309-84(a)(5)
Clarified required degree of compliance
with analytical system schematic.
This change represents a simple clarification of minor
deviations which are allowable under the existing
equipment specifications. Previously, there was some
uncertainty within the industry as to what deviations
would be acceptable to EPA.
§86.1309-84 (b) {1) Clarified rationale and means of
and (c)(1) verifying that CVS-induced pressure
variations on the exhaust system are not
excessive.
Both the rationale for and the means of verifying this
specification were questioned by the industry; this
procedural change clarifies both the intent and the
procedure itself.
§86.1309-84 (b) (2) Rereferenced CVS gas mixture temperature
accuracy from the temperature at the
start of the test to the average
operating temperature during the test.
This is a minor change, simply changing the reference
temperature against which the temperature excursions of
the dilute exhaust are measured. Because the temperatures
during a test never go below the temperature at the start
of the test, the previous specification was actually twice
as stringent as needed be.
§86.1309-84 (c) Clarified sensor accuracy requirements
to include the signal transmission and
readout equipment.
This amendment corrects a previous omission, and more
correctly includes all sources of equipment error within
required accuracy specifications.
§86.1309-84 (c) (2) Relaxed temperature measurement system
response time from 0.100 to 1.50
seconds; eliminated response time
requirement for CVS equipment with heat
exchange.
This response time relaxation reflects the uncertain
commercial availability of fast-responding temperature
sensors for CFV-CVSs. CVSs with heat exchangers do not
require temperature sensors with fast response
characteristics, and are thus exempted from the response
time specification.
-208-
-------
§86.13iQ-=-84 (a) Permitted measurement of mass fuel
consumption in lieu of CO2 exhaust
measurement. Clarified general sampling
system requirements.
This option, requested by EMA, permits a manufacturer to
measure mass fuel consumption in lieu of CO2 exhaust
concentration. Either of the two measurements is
acceptable for calculation of exhaust emissions, however,
the equipment for measuring mass fuel flow is much less
expensive to procure and maintain. In addition, general
sampling system requirements were clarified where
ambiguous or misinterpreted by the industry.
§86.1310-84 (a) (5) Clarified required degree of compliance
with analytical system schematics.
The industry requested clarification of the degree of
compliance which EPA requires for components of the
exhaust analytical system. Specifically, minor deviations
in equipment components are permitted; many of these
deviations are indicative of the different equipment a
manufacturer may use.
§86.1310-84 (b) (2) (iii) Rereferenced CVS gas mixture temperature
accuracy from the temperature at the
start of the test to the average
operating temperature during the test.
This is a minor change, simply changing the reference
temperature against which the temperature excursions of
the. dilute exhaust are measured. Because the temperatures
during a test never go below the temperature at the start
of the test, the previous specification was actually twice
as stringent as needed be.
§86.1310-84(a)(3) Removed requirements for Administrator
approval for use of continuous sampling
systems.
Many diesel engine manufacturers are already using
continuous sampling systems, the viability of which have
been demonstrated in EPA/EMA correlation programs. The
test procedure already contains generalized specifications
for continuous sampling systems; EPA feels that these are
sufficient to guarantee correlatable test results, without
the unnecessary step of requiring advance EPA approval.
§86.1310-84(b)(3)(i) Revised HC "overflow" technique to be an
optional calibration, but mandatory zero
and span check of the sampling system.
-209-
-------
EPA's earlier requirement that the HC emissions analyzer
be calibrated through the overflow system has been
changed; calibration will now take place at the analyzer
ports, with zero and span checks still being made through
the overflow system. EPA believes that the revision is
more technically correct, consistent with light-duty
practice, and still adequately permits the identification
of potential hang-up problems.
§86.1310-84 (b) (3) (ii) Included provision for use of a single
sample line.
The industry requested this change, and suggested wording
which allowed only the use of a single sample pump. EPA's
earlier requirement that different analyzers use different
sample lines was based upon concern about potential errors
arising from pressure fluctuations induced by more than
one sample pump. The revised wording as suggested by the
industry satisfies EPA's concern, and has been
incorporated into the test procedure.
§86-1310-84 (b) (3) (iii) Reduced HC "overflow" gas flow rate to
at least 105 percent.
The earlier version of the test procedure required
excessive overflow gas flow rates; the industry argued
that this was wasteful of calibration gases, and that any
quantity of gas greater than 100 percent total flow was
sufficient. EPA concurs with this observation.
§86.1310-84 (b) (3) (v) Eliminated requirement that gaseous HC
(A) probe point only upstream.
The industry requested this modification because probes
pointing upstream are susceptible to contamination by
large particles (for example, collected particulate matter
intermittently shaken off the walls of the exhaust system,
engine parts, etc.). Exhaust gas flow is sufficiently
isokinetic to allow the probe to point in any direction
without impacting the accuracy of the measurement of gas
concentration.
§86.1310-84(b)(3)(v) Eliminated specific insulation
(C) requirement as the only means of
maintaining sample probe wall
temperature.
Measurement system integrity requires that the wall of the
HC sample probe be maintained at a sufficiently high
temperature. EPA's earlier requirement specifically
dictated how that temperature was to be maintained; the
-210-
-------
revised test procedure simply requires that the
temperature be maintained, and leaves the method of
temperature maintenance to the discretion of the
manufacturer.
§86.1310S84 (b) (3) (vi) Clarified sensor accuracy requirements
(<&) and (B) to include signal transmission and
readout equipment.
This amendment corrects a previous omission, and more
correctly includes all sources of equipment error within
required accuracy specifications.
§86.1310-84(b)(3)(vi) Eliminated.
(C)
This paragraph was redundant, served no purpose to the
test procedure, and was eliminated.
§86.1310-84(b)(3) Increased analyzer response time from no
(vii) (B) greater than 5.5 to no greater than 20.0
seconds.
The industry recommended this change, providing data that
sampling system response times up to 20 seconds yielded
equivalent emission results. EPA's original response time
requirement reflected primarily a concern that the
integrity of longer sample lines is more difficult to
maintain, especially if heated. EPA believes, however,
that sufficient requirements already exist within the test
procedure for sample line heating, leak checks, and zero
and span checks, in addition to the verification provided
by the industry data, that increasing the sample system
response time will not adversely affect test accuracy. In
addition, the allowance of longer sample lines (by
allowing greater system response times) gives the
manufacturer much greater flexibility in modifying
existing dynamometer cells for running the transient test.
§86.1310-84 (b) (4) Eliminated requirement that gaseous HC
(ii)(F) probe point only upstream.
The industry requested this modification because probes
pointing upstream are susceptible to contamination by
large particles (for example, collected particulate matter
intermittently shaken off the walls of the exhaust system,
engine parts, etc.). Exhaust gas flow is sufficiently
isokinetic to allow the probe to point in any direction
without impacting the accuracy of the measurement of gas
concentration.
-211-
-------
§86.1310-84 (b) (5)(ii) Increased analyzer response time from
(B) from 5.5 to 20.0 seconds.
The industry recommended this change, providing data that
sampling system response times up to 20 seconds yielded
equivalent emission results. EPA's original response time
requirement reflected primarily a concern that the
integrity of longer sample lines is more difficult to
maintain, especially if heated. EPA believes, however,
that sufficient requirements already exist within the test
procedure for sample line heating, leak checks, and zero
and span checks, in addition to the verification provided
by the industry data, that increasing the sample system
response time will not adversely affect test accuracy. In
addition, the allowance of longer sample lines (by
allowing greater system response times) gives the
manufacturer much greater flexibility in modifying
existing dynamometer cells for running the transient test.
§86.1311-84 (a) Clarified required degree of conformance
with analytical system schematic.
The industry requested clarification of the degree of
compliance which EPA requires for components of the
exhaust analytical system. Specifically, minor deviations
in equipment components are permitted; many of these
deviations are indicative of the different equipment a
manufacturer may use.
§86.1314-84(g) Allowed use of gas dividers, subject to
accuracy requirements of ±1.5 percent of
NBS gas standards.
Gas dividers were permitted under the old test procedure;
however, accuracy specifications for their use were never
provided, creating uncertainty within the industry as to
what EPA actually required. This technical amendment
corrects that omission by providing gas blending accuracy
specifications.
§86.1316-84 (c) (3) Added weekly check (not mandatory
calibration) of torque feedback signals
at steady-state operating conditions.
EPA believes that this procedural modification is easily
performed, and reflects good engineering practice; this
amendment is therefore made part of the test procedure.
§86.1318-84(b) Added required electronic check and
adjustment of torque feedback signal
before each test.
-212-
-------
EPA believes that this procedural modification is easily
performed, and reflects good engineering practice; this
amendment is therefore made part of the test procedure.
§86.1319-84 (a) Defined flowmeter traceable to NBS as a
reference standard for CVS calibration;
removed need for Administrator's
approval.
EPA believes that any flowmeter traceable to NBS standards
which conforms to EPA's accuracy specifications is
technically acceptable, and does not require advance
approval by EPA for its use.
§86.1319-84 (c) (2) (i) Eliminated pump pressure tap
specifications.
This part of the original test procedure was drafted
verbatim from light-duty vehicle test procedures; in fact,
this procedure is outdated, and is removed from the test
procedure by EPA inititive.
§86.1319-84(c)(4) Changed accuracy tolerances for
and (d) (3) measurements of barometric pressure
(from +.01 inches Hg to ^+.10 inches Hg) ,
pressure head at CVS pump outlet and
inlet depression at CVS pump inlet (from
+.05 inches fluid to +.13 inches fluid),
and elapsed time for test (from +^.05
seconds to +^.5 seconds) . Changed air
temperature measurement tolerances from
+0 .5°F to +2 .00F for PDP-CVS, and from
0.5°F to 4.0°F for CFV-CVS.
EMA submitted data and calculations which argued that
relaxed calibration accuracies would not impair overall
test accuracy. The requirements that these measurement
accuracies be very stringent necessitated the use of very
accurate but very expensive calibration equipment. EPA
has reviewed EMA's calculations, and agrees that no net
impact on test accuracy would be incurred. In fact, the
requirement for CVS system verification using propane will
still serve as an overall system check. For these
reasons, EPA accepts EMA's recommendations and relaxes the
tolerances.
§86.1319-84 (d) Added missing sections from light-duty
CVS calibration procedure, but deleted
correlation function between pump RPM
and pressure differential.
-213-
-------
During drafting of the original test procedure, several
paragraphs were inadvertantly deleted; these paragraphs
were substantially similar to CVS calibration procedures
applicable to light-duty vehicles. These paragraphs have
now been restored, with the exception of a single but
unnecessary correlation function.
§86.1319-84(e)(1) Eliminated carbon monoxide as a
recommended CVS verification gas.
Both EPA and industry use propane as a CVS verification
gas; propane is adequate for all verification purposes.
Given the adequacy of propane, and the risk to safety
associated with the use of carbon monoxide, EPA no longer
recommends its use.
§86.1319-84 (e) (4) Corrected density of propane to 17.30
g/ft3.
This is a minor numerical correction that makes the
heavy-duty test procedure consistent with light duty.
§86.1321-84 (b) Clarified requirements for HFID analyzer
calibration.
This technical amendment eliminates the requirement for
overflow calibration of the analyzer; the exact analyzer
calibration procedure is reworded to reflect this change.
§86.1324-84 (c) Permitted use of span gases for CO2
analyzer calibration.
Span gases are "named" to a lesser degree of accuracy than
calibration gases, and for this reason, calibration gases
have always been used to maximize accuracy of analyzer
calibrations. EMA has provided evidence that slightly
less accurate calibration of the CO2 analyzer will not
affect overall test results. (CO2 emissions
measurements are used only to calculate overall dilution
factor and fuel consumption.) EPA concurs with EMA's
analysis, and specifically allows use of span gas for the
CO2 analyzer calibration.
§86.1327-84 (d) (4) Eliminated prior approval of
Administrator for inclusion of engine
accessor ies.
Since the heavy-duty transient test is based upon
normalized engine parameters, EPA is no longer concerned
about parasitic effects of engine accessories. EPA
-214-
-------
therefore allows their inclusion on certification engines
if the manufacturer so desires, without the requirement
for advance EPA approval.
§86.1327-84 (d) (5) Eliminated mandatory use of production
starter.
The earlier test procedure required the use of a
production starter motor at the beginning of the transient
test sequence. The industry has argued for some time that
this represents an unnecessary test burden. EPA no longer
believes that use of a dynamometer to start the engine
will significantly impact overall test results, especially
since the dynamometer will be required to simulate the
characteristics of a production starter.
§86.1327-84 (f) Clarified and modified exhaust system
requirements.
Significant clarifications to exhaust system requirements
for diesel engines have been made. Specifically, use of a
facility exhaust system in lieu of a chassis-type exhaust
system has been required. This change has been made to
provide uniformity with future exhaust system requirements
which will be necessary for the measurement of
particulates.
§86.1330-84(a)(1) Permitted dilution air temperatures
above 86°F.
This change has been made to accommodate problems several
manufacturers were having in maintaining a CVS dilution
air temperatures below 86°F, especially in the summer
months. Rather than force the installation of expensive
air cooling equipment, EPA is eliminating the upper
temperature limit of the CVS dilution air temperature.
EPA does not believe that this will have any impact on
test results. (Note that the dilution air temperature can
readily exceed 200°F when mixed with engine exhaust.)
§86.1330-84(a)(3) Permitted test cell and engine intake
air to exceed 86°F if no temperature
dependent auxiliary emission control
devices are used.
This modification applies almost certainly to diesel
engines only, and will preclude the installation of
expensive air handling and temperature conditioning
equipment where such equipment is not necessary.
-215-
-------
§86.1330-84(b)
Eliminated need to control test cell,
engine intake, and CVS dilution air
humidity.
EPA is specifically providing the use of a humidity
correction factor for both gasoline and diesel engines;
for this reason, control of humidity during the test
sequence is ho longer necessary or required.
§86.1330-84(e) Specified inlet and exhaust restrictions
for diesel engines, both naturally
aspirated and turbocharged.
This technical amendment represents a clarification of
earlier requirements, and was recommended by EMA. This
amendment constitutes no net change in the test procedure.
§86.1330-84 (f) Clarified pre-test procedures.
This amendment specifically addresses when certain
operational checks of the engine and other procedural
steps may be performed during the test sequence.
§86.1332-84 (b) Minimum mapping speed redefined as curb
idle speed.
This technical amendment eliminates the need to map the
engine below idle speed. This eliminates engine and
equipment stresses associated with running the engine at
full load at very low speeds. No compromise in test
accuracy is incurred, because very few of the engine
speeds required during transient testing actually lie
below idle.
§86.1332-84 (d) (2) (vii) Added +20 rpm tolerance to 100 rpm
mapping steps.
This accuracy tolerance was requested by Ford because EPA
had provided no tolerance in the earlier test procedure.
§86.1332-84(d)(2)(x) Added allowance for avoiding lengthy
and (d) (3) (viii) engine warm-up before mapping if the
engine is already warm.
This technical amendment was requested by MVMA as a means
of avoiding unnecessary warm-up required under the earlier
test procedure. Since EPA's intent is merely that certain
portions of the test be conducted with a warm engine, EPA
is allowing that these portions of the test be conducted
without warm-up, provided that certain engine temperature
specifications are met.
-216-
-------
§86.1332-84 (d) (3)(iv) Eliminated mandatory 10-minute minimum
time for temperature stabilization.
EPA is eliminating this unnecessary requirement on its own
initiative. That the engine temperature be stabilized is
the only necessary criterion; if this criterion is
achieved in less than 10 minutes, there is no need to
maintain warm-up for the full 10 minutes.
§86.1332-84(e)(1) Added goodness of fit criteria for cubic
spline technique.
MVMA requested this amendment, so that EPA would provide
an accuracy specification where the original test
procedure had failed to do so.
§86.1332-84 (f) Removed requirement for Administrator
approval for alternate mapping
techniques based upon safety or
representativeness criteria.
EPA is removing the requirement that alternate mapping
techniques be approved in advance by EPA, if such
techniques are in the manufacturer's judgment required to
maintain test safety or representativeness. General
guidelines for alternate mapping techniques are provided,
along with the requirement that the specific mapping
technique used be reported to EPA in the manufacturer's
application for certification.
§86.1332-84(g) Added conditions under which remapping
need not occur.
EPA has added this clarification because several
manufacturers had misinterpreted the earlier test
procedure to require that an engine be mapped before each
and every test. This was never EPA's intent, nor EPA's
test practice.
§86.1333-84 (d) (3) Clarified point deletion allowances.
This technical amendment represents a clarification of
earlier requirements.
§86.1333-84 (f) Added clutch allowance.
EPA has specifically added to the test procedure the
allowance to use a clutch during engine testing. The
earlier procedure never specifically precluded the use of
a clutch; indeed, EPA recommends its use in certain
circumstances. Several manufacturers, however, had
-217-
-------
misinterpreted EPA's earlier procedure, and requested that
EPA specifically address the use of a clutch to alleviate
any uncertainty.
§86.1333-84 (g) Added required method of calculating
measured rated rpm, or usage of
manufacturers' specified rated rpm,
whichever is greater.
EPA is initiating this technical amendment. In testing
practice, EPA has found this revised methodology to be
less susceptible to errors induced by unusual engine
mapping curves.
§86.1335-84
-------
§86.1336-84 (b) (2) Eliminated need for approval by
Administrator of longer cranking times.
EPA is eliminating the need for prior Administrator
approval for engine cranking times which are longer than
nominal/ but nevertheless typical of the engine.
§86.1336-84 (b) (3) Eliminated need to report malfunctions
and (4) during engine start to the Administrator.
EPA considers this to be an unnecessary requirement and
eliminates it.
§86.1336-84(c) Clarified action to be taken during
engine stalling.
This amendment represents a clarification of the earlier
test procedure.
§86.1337-84 (a) {10) Added requirement that sampling systems
and (21) continue sampling until system response
times have elapsed.
This amendment goes hand in hand with EPA's allowance for
longer sampling system response times. This amendment
assures that emissions generated by the engine are not
lost at the very end of the test, as they would be if
sampling systems with longer response times were shut down
simultaneously with the engine.
§86.1337-84 (b) Eliminated mandatory time increments for
emission tests using more than one bag
or mode.
EPA sees no need to require manufacturers to conform with
specific time increments for modal analysis.
§86.1337-84 (c) Added clarification cf conditions under
which an engine on which a void test was
run may be recooled and retested.
This amendment represents a clarification of the earlier
procedure.
§86.1338-84(a)(2) Added procedure for calibration below 15
percent of analyzer's full scale.
This procedure was requested by the industry to provide
clarification of the specific conditions and applicable
procedures for calibrating analyzers below 15 percent of
full scale. This amendment represents no net change in
procedure accuracy.
-219-
-------
§86.1338-84(b)(1) Clarified permissible deviations from
requirement that analyzer response
remain between 15 and 100 percent of
full scale.
This amendment represents a clarification of the earlier
procedure, as requested by the industry.
§86.1340-84 (a) (1) Clarified stability requirement for
background sample response.
This amendment corrects an inadvertantly stringent
specification contained within the original test
procedure; a more reasonable stability requirement is
promulgated.
§86.1340-84(a)(2) Eliminated need to store all ADC input;
only an average integrated value need be
stored.
This amendment corrects an overly burdensome requirement
contained within the original test procedure. EPA now
requires only that a manufacturer record a single emission
value for a given test cycle, and not the second-by-second
ADC output. (This is conceptually identical to the
requirements imposed for bag sampling.)
§86.1340-84(d) and (e) Reorganized the procedures for clarity,
and modified continuous HC sampling and
hang-up check procedures.
This amendment represents a clarification of the original
test procedure.
§86.1340-84 (f) Changed hang-up check to include entire
sample probe.
This technical amendment makes the hang-up check more
technically correct, and better able to verify the
integrity of the entire sample probe.
§86 .1341-84 (h) Added to address the handling of closed
rack torque reference points in cycle
validation. Clarified method of
validation foe HHP points when torque
reference calls for motoring.
This amendment addresses the treatment of certain feedback
points in the cycle performance regression analyses; these
specific points and their treatment in the regressions
were inadvertantly ignored in the original test procedure.
-220-
-------
Clarified regression analysis point
deletions for diesels at closed rack;
specifically allowed use of clutch.
Original Figure N84-11 deleted, and
original Figure N84-12 substituted in
its place. Regression line tolerances
clarified to represent a percentage of
power-map values. An additional torque
and power deletion added if closed
throttle and torque feedback greater
than torque reference
These amendments reflect clarifications, elimination of an
unnecessary figure, and the inclusion of an additional
point deletion allowance which EPA has determined to be
appropriate.
§86.1342-84 (c) Corrected omission of humidity
correction factor from flow compensated
NOx measurement calculations.
§86.1341-84
Figure N84-11
This corrects an error in the earlier test procedure, and
reflects EPA's provision of a humidity correction factor
for diesel engines.
§86.1342-84 (d) (3) Added calculation for mass fuel flow to
be used in approximating dilute exhaust
C02.
This option, requested by EMA, permits a manufacturer to
measure mass fuel consumption in lieu of CO2 exhaust
concentration. Either of the two measurements is
acceptable for calculation of exhaust emissions, however,
the equipment for measuring mass fuel flow is much less
expensive to procure and maintain.
§86.1342-84 (d) (5) Added dilution factor calculation based
and (6) upon approximated dilute exhaust CO2.
Specified humidity correction factors
for diesel engines.
This additional calculation was necessitated by the
allowance that mass fuel flow measurement be substituted
for exhaust CO2 measurement. In addition, the newly
provided humidity correction factor for diesel engines is
specifically included here.
§86.1342-84 (i) Added calculations for dry to wet
exhaust concentration conversion,
accounting for both dilution air
humidity and approximate exhaust H2O
concentration.
-221-
-------
This correction calculation was suggested by the EMA as an
improvement. EPA concurs with their recommendation, and
believes that the omission of this calculation from the
earlier procedure was an error.
§86.1344-84 (e) (6) Added requirement for description of
mapping technique.
This requirement has been added by EPA to ensure that the
manufacturers inform EPA in their application for
certification if an alternate mapping technique has been
used.
Appendix 1(f)(1) An optional driving cycle for heavy-duty
gasoline engines has been added.
(See Chapter 3.A.3 of this Summary and Analysis of
Comments.)
B. Subpart I - Heavy-Duty Diesel Engine Smoke Test Procedure
Overview of Technical Amendments
The following sections from 40 CFR Part 86 (as printed
July 1, 1982) are being superseded, and are hereby deleted:
Sections
86.877-1
86.877-13
86.877-2
86.877-14
86.877-3
86.879-5
86.877-4
86.879-6
86.877-5
86.879-7
86.877-6
86.879-8
86.877-7
86.879-9
86.877-8
86.879-10
86.877-9
86.879-11
86.877-10
86.879-12
86.877-11
86.879-13
86.877-12
86.879-14
The following sections are being added to 40 CFR Part 86
(as printed July 1, 1982). Aside from changes in references,
specific allowances for the use of automated data collection
equipment and electric dynamometers, and changes to permit
consistency with Subpart N and other 1984 rules, no significant
change distinguishes this procedure from earlier versions:
-222-
-------
Sections
86.884irl General Applicability.
86.884-2 Definitions.
86.884t3 Abbreviations.
86.884^4 Section numbering.
86.8$!-; 5 Test procedure.
86.884^6 Diesel fuel specifications.
86.884-7 Dynamometer operation cycle for smoke emission tests,
86.884-8 Dynamometer and engine equipment.
86.884-9 Smoke measurement system.
86.884-10 Information.
86.884-11 Instrument checks.
86.884-12 Test run.
86.884-13 Data analysis.
86.884-14 Calculations.
C. Subpart P - Heavy-Duty Gasoline Engine and Light-Duty
Gasoline Truck Idle Test Procedure
Overview of Technical Amendments
In general, the following changes were made throughout the
entire subpart:
1. All references to diesels were deleted; references
to and procedures for light-duty trucks were added.
2. Miscellaneous clarifications were made.
3. References to Subparts N, B, and D were clarified.
4. Requirements were made consistent with Subparts N
and B where possible.
Specific Amendments
The following sections from 40 CFR Part 86 Subpart P were
modified enough to merit specific mention. These modifications
represent no substantive change to the fundamental test
procedure:
86.1514-84 Analyzer gas specifications made consistent
with Subpart N and B.
EPA considers it to be unnecessary to have differential
requirements for calibration and span gases for all
emission test procedures applicable to any given vehicle
-223-
-------
or engine. This amendment makes analyzer gas requirements
consistent between applicable test procedures.
86.1516-84(b) Minimum calibration frequency changed from
weekly to monthly as in Subpart N.
This amendment makes calibration procedures consistent
between subparts, such that the same equipment can be used
for either test.
86.1516-84 (c) Check interval changed from daily to before
each test.
This amendment makes calibration procedures consistent
between subparts, such that the same equipment can be used
for either test.
86.1527-84(a) Clarified test run sequence, especially for
light-duty trucks.
The earlier test procedure addressed light-duty trucks
only by reference, and left much of the test sequence
unspecified. This amendment corrects that omission, and
provides a specific test sequence for light-duty trucks.
86.1527-84 (b) Ambient test cell requirements made consistent
with those of Subpart N and B.
This amendment allows the use of the same equipment in the
same test cells for testing conducted under either subpart.
86.1542-84 Information requirements made consistent with
Subparts N and B.
This amendment allows the manufacturer to more easily
combine test results in a single application for
certification.
-224-
-------
12. Issue: Possible "Migration" from Class IIB to Class III
Summary of the Issue
In the March 1983 staff paper, EPA proposed a split-class
approach to HDGE HC/CO control. In that proposal, HDGEs
intended for use in Class IIB and III applications (up to
14,000 lbs. GVW) would be required to meet the statutory
standards of 1.3 HC/15.5 CO in 1987, while HDGEs intended for
use in heavier applications (over 14,000 lbs. GVW) would
continue to meet non-catalyst standards (assumed to be 2.5 HC/
35 CO). Commenters disagreed with the choice of "break point,"
maintaining that Class III (10,001-14,000 lbs. GVW) HDGEs
should be included with the heavier HDGE applications, and
therefore allowed to meet the non-catalyst standards rather
than the statutory standards.
The critical problem with lowering the "break point," as
advocated in the comments, is the possibility that LDTs in the
upper portion of the Class IIB weight range (8,501-10,000 lbs.
GVW) could be slightly "redesigned" so as to be heavy enough to
be included in Class III. This "migration" of Class IIB
vehicles into Class III, thereby avoiding the catalyst-forcing
statutory HC/CO standards, has been discussed in some detail
elsewhere.[1] The discussion below is limited to where the
classes should be split, and how the potential problem of
"migration" should be addressed.
Summary of the Comments
Of the comments submitted in response to the staff paper,
only Ford and GM specifically addressed the issue of where the
"break point" should be set under EPA's proposed split-class
approach.
Ford suggested that a more logical "break point" would be
at 10,000 lbs. GVW; in other words, Class IIB HDGEs would meet
the catalyst-forcing statutory standards while all other HDGEs
(Classes III-VIII) would meet the proposed non-catalyst
standards. Ford acknowledged the validity of EPA's concern
over "migration" under this HDGE split, but maintained that
HDGEs in dual rear-wheel and fifth-wheel ("pop-truck") Class
III applications are more appropriately grouped with the
heavier HDGEs. This is due to the in-use service environment
of such vehicles, a significant portion of which is spent
operating at full-load, WOT conditions. Ford indicated that
the high temperatures characteristic of these conditions render
catalyst use infeasible.
-225-
-------
General Motors responded to the staff paper with an
entirely new proposal, maintaining the general notion of
splitting HDGEs into two groups but having little else in
common with the split-class approach described in the staff
paper. Other aspects of the GM proposal are dealt with in
detail .elsewhere; [1] only the choice of "break point" is
discussed here.
General Motors paralled Ford in proposing that the "break
point" be set at 10,000 lbs. GVW, and claimed that its proposal
is responsive to two of the concerns expressed by EPA. As
identified by GM from the staff paper, these concerns are:
First, that the majority of HDGEs as currently defined be
subject to the catalyst-based statutory standards; and second,
that the air quality benefits resulting from implementation of
the statutory standards not be significantly reduced by
"migration" of HDGEs below, but close to, the "break point" to
just above that point.
General Motors noted that HDGEs in Class IIB represent
about 65 percent of all HDGEs; thus, the IIB/III-VIII "break
point" would continue to require a majority of HDGEs to be
catalyst-equipped. Citing the small HDGE sales in Class III,
GM stated that "insignificant air quality improvement" would
occur, relative to its proposal, if the "break point" is set at
14,000 lbs. GVW. GM also indicated, although less
specifically, a concern similar to that expressed by Ford:
that the use patterns and operating conditions characteristic
of Class III applications are likely to result in higher
temperatures than catalyst technology can endure.
General Motors also claimed that not only would there not
be a "migration" problem, but that there would actually be
strong incentives not to move vehicles to higher GVW classes.
As justification for this assertion, GM states that the cost
increase for catalyst technology on HDGEs meeting the statutory
standards would be comparable to the cost increase for added
non-catalyst emission control technology on the heavier HDGEs.
In addition, raising the GVW would involve cost increases for
the upgrading of other vehicle components (e.g., springs,
axles), thus further reducing any possible motivation for
vehicles to "migrate" from Class IIB to Class III.
Analysis of the Comments
EPA acknowledges that the air quality impact of control of
HC and CO to catalyst-fcircing levels for HDGEs in Class III
applications will be small compared to the control from Class
IIB vehicles (because of sales). The selection of 14,000 lbs.
GVW as the "break point" was based on EPA analysis of where
-226-
-------
HDGE types change, in terms of use and application. EPA was
also concerned over possible HDGE "migration" if the dividing
line between catalyst-based and non-catalyst based standards
were set at the lower level (10,000 lbs. GVW).
The comment by Ford about the similarity of use patterns
and operating environments for some Class III applications and
the Classes IV-VIII applications is a valid concern. However,
EPA does not believe that lowering the "break point" to 10,000
lbs. GVW is the best approach to dealing with this concern,
since it fails to address how "migration" might be avoided. A
method of accounting for both EPA's concern over "migration,"
and Ford's concern over the inappropriateness of requiring
catalyst technology on some HDGEs in Class III applications, is
to maintain the "break point" as EPA proposed (14,000 lbs. GVW)
while providing for reclassification of a limited number of
HDGE configurations. This is explained in more detail below.
The disincentives to "migration" cited in the GM comments
are based on assumptions contained in the GM counter-proposal,
not on the EPA split-class proposal. While the arguments may
sound reasonable on first examination, EPA's concerns over
"migration" are not alleviated. Previous "migration" of Class
IIA vehicles up to Class IIB, to avoid more stringent emission
standards and fuel economy regulations, demonstrates the
validity of EPA's concern. This earlier trend of "migration"
is evidenced by the relatively large concentration of vehicles
with GVWs in the 8,501-8,600 lbs. range,[1] which can logically
be assumed to have "migrated" above 8,500 lbs. GVW for the
reasons cited above.
In EPA's analysis of the GM counter-proposal,[1] an
attempt to estimate the potential magnitude of the "migration"
of Class IIB into Class III is made. Although these estimates
must be considered "soft," due to the unavailability of sales
data for Class IIB alone, they do provide an estimate for
consideration. The analysis showed that roughly 70 percent of
Class IIB vehicles have GVWs of 9,000 lbs. or more, which means
that they could conceivably be redesigned so as to enter Class
III. How much actual "migration" would occur if the "break
point" were set at 10,000 lbs. GVW is difficult to predict with
any certainty. However, EPA believes that the potential for
migration is strong because of the number of vehicles sold near
the Class IIB upper GVW limit, coupled with a desire by
manufacturers to apply catalyst systems to as few vehicles as
possible (if for no other reason than because of an anticipated
strong buyer preference for non-catalyst vehicles). Further,
EPA finds the risk of migration avoidable.
-227-
-------
EPA believes that the concerns of both the Agency and the
industry over the inclusion of Class III vehicles in the
proposed new Classes IIB-III subcategory can be adequately
addressed by a fairly simple modification to the approach
developed in the staff paper. The "break point" between
catalyst and non-catalyst HDGES should remain at 14,000 lbs.
GVW, as proposed by the staff paper. However, whereas that
proposal effectively contained no exemption provisions, EPA
recommends modifying the proposal so that manufacturers would
be permitted to reclassify a limited portion of their Classes
IIB and III configurations to Class IV. The choice of
configurations to be reclassified would be left to the
discretion of the manufacturers, providing them maximum
flexibility in choosing the configurations where catalyst
application would be. the most difficult. However, the size of
the reclassified group would have to be limited by EPA to
insure that no significant environmental losses would occur.
The limit on reclassification would be expressed as a
percentage of all sales in Classes IIB and III. Based on the
actual 1980 and projected 1990 sales data used in the staff
paper, this limit would be in the range of 2 to 7 percent,
approximating Class III sales as a fraction of combined sales
in Classes IIB and III. There is a tendency in the sales
projections for this ratio to increase slowly over time;
however, as was noted by Ford in its comments, it is not
necessary for all Class III HDGEs to be exempted from the
statutory standards. Balancing these considerations, EPA has
decided to limit to 5 percent of combined Classes IIB and III
sales the reclassification of Classes IIB and III HDGEs to
Class IV.
Under the split-class approach, modified as detailed
above, there should be little change of air quality benefits
from the staff paper proposal, while the legitimate concerns of
the manufacturers over a limited number of Class III
applications would be addressed. In fact, the manufacturers
will gain an added degree of flexibility in compliance with the
new regulations. They will be able to minimize their costs by
reclassifying the more severe applications.
Conclusions
EPA will maintain the LHDGE/HDGE "break point" at 14,000
lbs. GVW, as was proposed in the staff paper. EPA will include
provisions for up to 5 percent of combined sales of HDGEs in
Classes IIB and III to be reclassified and certified to
non-catalyst levels, on a configuration-specific basis.
-228-
-------
References
1. Evaluation of General Motors' Heavy-Duty Engine
Proposal, EPA Memo from Chester J. France, Standards
Development and Support Branch, to Richard D. Wilson, Office of
Mobile Sources, May 16, 1983.
-229-
-------
13. Issue; Diesel Engine Closed Crankcase Requirements
Summary of the Issue
The regulations promulgated on January 21, 1980 presently
require that all naturally aspirated heavy-duty diesel engines
have closed crankcases (i.e., zero crankcase emissions are to
be discharged into the ambient atmosphere).
Summary of the Comments
General Motors claimed that no technology was available to
safely allow closing of the crankcase for 2-stroke heavy-duty
diesel engines. General Motors' primary concern is that the
internal fuel system used in these engines may leak. This
would create a safety problem if fuel overflows into the engine
intake through the crankcase ventilation system and causes an
uncontrolled engine runaway.
General Motors also noted that 2-stroke engines require a
blower to induct intake air into the cylinders. To route
crankcase emissions into the intake air would require either an
expensive pumping system to force the crankcase vapors into the
higher pressure air downstream of the intake blower, or, if
crankcase vapors were ventilated into the intake air upstream
of the blower, fouling and deterioration of the blower may
occur. These problems led EPA to decide not to finalize closed
crankcase requirements for turbocharged diesel engines in
December of 1979.
General Motors recommended that the closed crankcase
requirement for these engines be rescinded.
Analysis of the Comments
There are two aspects to the closed crankcase issue for
2-stroke HDDEs: feasibility and cost effectiveness.
EPA notes that GM engine families, other than 2-stroke
engines, utilize internal fuel systems. GM has stated to EPA
that a safe closed crankcase system for its internally fueled
8.2L engine, while presenting an initial challenge to
designers, will likely be available for the 1985 model year.
Given GM's claim.that a feasible closed crankcase system can be
applied to its 8.2L engine in 1985, it is difficult to accept
GM's assertion that the application of such systems to 2-stroke
engines will be permanently infeasible. On the strict basis of
feasibility, EPA finds no merit in GM's request that 2-stroke
engines be permanently excluded from closed crankcase
requirements.
-230-
-------
On the other hand, the parallel drawn by GM between
2-stroke and turbocharged engines is valid. Both engines would
require similar closed crankcase systems in the sense that a
more expensive pumping system is needed to overcome the high
pressure intake air. Otherwise, turbocharger/blower fouling
may occur if crankcase effluents are added to the intake air
upstream of the turbocharger/blower. EPA. recognized this
problem in the January 21, 1980 rulemaking: such a system for
turbocharged engines would be roughly ten times the cost of a
closed crankcase system for naturally aspirated engines. For
this reason, EPA did not finalize closed crankcase requirements
for turbocharged engines at that time. It was not a question
of feasibility, but rather an acknowledgement of the poor cost
effectiveness of the requirement.
Failure to include all engines which rely upon forced
induction of intake air with this deferral of closed crankcase
requirements occurred mainly because the manufacturers did not
raise it as a significant issue. (GM and other manufacturers
never raised such an issue during the earlier rulemaking.)
However, it would now be technically appropriate to make this
change to the regulations. Furthermore, the number of engines
affected by this (i.e., the number of naturally aspirated
2-stroke engines) is quite small, and getting smaller as
turbochargers become more universally adopted. (Only 3.3
percent of GM's 1983 sales were naturally aspirated 2-stroke
engines; no other manufacturer makes 2-stroke engines.) Given
this small impact, and given the technological similarity
between the 2-stroke and turbocharged engines with respect to
closing the crankcase, EPA concurs that closed crankcase
requirements should not apply to 2-stroke engines until a
similar requirement for turbocharged engines is promulgated.
This conclusion is based entirely on the relative cost
effectiveness of closing the crankcase on engines which require
turbochargers, blowers, etc., to induct intake air.
Conclusion
The closed crankcase requirements should not apply to 1985
and later model year heavy-duty diesel engines which require
forced induction of intake air (e.g., by turbochargers,
blowers, etc.).
-23.1-
-------
Appendix A
Draft Technological Feasibility Analysis
from the NPRM "Revised Gaseous Emissions
Regulations for 1984 and Later Model Year
Light-Duty Trucks and Heavy-Duty Engines"
-------
CHAPTER II
technological feasibility/
ATTAINABLE HOH-CATAL^ST STANDARDS
A. Introduction
In this chapter, EPA analyzes available technologies and pro-
jects what levels of HC and CO emissions for heavy-duty gasoline
(HDG) engines are attainable for 1984, assuming that oxidation
catalysts are not employed.
B. Current HC and CO Emission Rates
To properly evaluate potential non-catalyst emission reduc-
tions from HDG engines, current emission rates must be reviewed.
Because absolute emission levels are inherently affected by the
test procedure over which they are measured, a review of the tran-
sient emission, test is appropriate.
1. Overview; The Transient Test
The transient test is performed on a computer-controlled en-
gine dynamometer. During the test, the engine is driven through
continuously-varying speeds and loads according to prescribed cy-
cles* These speed and load cycles were developed from in vehicle
performance data taken from 57 urban HDG trucks: 30 in the joint
industry/EPA CAPE-21 study in New York. City, and 27 in the EPA-
conducted Los Angeles CAPE-21 study. These trucks were actual
commercial vehicles operated by their own drivers; the performance
data was taken in the course of their daily business. These data
were then used to generate driving cycles representative of the
input data.
There are several key aspects of the transient test;
a. It is engine specific,
b. It is composed of subcycles, each of which retains the
characteristic driving patterns of specific, urban localities, and,
c. It is performed on a "cold" engine, and then "repeated
with the engine in a warmed-up state.
Each of the above characteristics is critical in evaluating cur-
rent and future emission trends.
Engine specific means that the cycles are defined in terms of
percent speed and percent load, i.e., any two engines are required
to deliver identical percent powers throughout the cycle even
though their absolute power levels may be different. This, and
the fact that emissions are expressed as mass per output work
-------
(work is simply power multiplied by the time at that power), make
emission results between engines comparable, regardless of their
specific rated power and varying performance characteristics.
Secondly, the cycle is actually four subcycles joined end to
end, each one characteristic of a particular geographic area and
type of driving;
Subcycle
1. New York Non-Freeway (NYNF)
Duration
(sec)
272
Characteristics
low power; stop-and-
go; 45% idle;
avg. spd. 7.8 mph
2. Los Angeles Non-Freeway (LANF)
309 moderate power, tran-
sient; 26% idle;
avg. spd. 15.1 mph
3. Los Angeles Freeway (LAF)
316 high-speed, high-
power cruising;
avg. spd. 45<54 mph
New York Non-Freeway (NYNF)
272
repeat of 1,
Each subcycle demands different performance from the engine, and
produces different absolute emission levels. These performance
demands can be isolated and their emissions impact reasonably
estimated.
Thirdly, the heavy-duty engine dynamometer test is similar to
the light-duty vehicle test in that the total emission results are
derived from a weighted average of a "cold" engine cycle and a hot
engine cycle. For the heavy-duty test, the cold start emission
cycle consists of the above four subcycles (NYNF, LANF, LAF,
NYNF), and is weighted 1/7 of the total; the hot start cycle is
identical to the cold, begins 20 minutes after shut down of the
engine from the cold start, and is weighted 6/7 of the total.
These weighting factors were derived from the observed in-use
ratio of cold starts to hot starts in the CAPE-21 survey. Since a
cold engine characteristically emits higher amounts of HC and CO,
the cold start cycle is significant when discussing current and
future emission levels.
2. Current Technology Engines
Table II-l presents a list of 1979 MY HDG engines tested by
EPA on the transient cycle. Table II-2 presents subcycle by sub-
cycle HC emission breakdowns for each engine, along with a percent
contribution of each subcycle to the total emission results.
Table II-3 presents the same data for CO.
A-2
-------
Immediately noticeable in Table II-l are the high levels of
HC and CO emissions. Note that the engines were certified for
1979 at 1.5 g/BHP-hr HC and 25 g/BHP-hr CO, but on the 9-mode
steady-state test procedure. In complying with any motor vehicle
emission standard, the design approach is to match the engine cal-
ibration and emission control system to the test procedure it-
self. This is the case in light-duty (see Reference 2), and in-
deed in heavy-duty. Table II-4 presents comparative HC and CO
emission data for both transient and 9-mode test procedures for
the current technology (1979) engine baseline. The large dif-
ferences in measured emissions are explainable by the readily
identifiable differences in required engine performance under each
test.
3. The 9-Mode Test
The 9-mode test procedure consists of nine steady state en-
gine operating modes which are weighted into a composite emission
number:
Mode
Speed (RPM)
% Power
Weighting
1
Idle
0
.232
2
2000
25
.007
3
2000
55
.147
4
2000
25
.077
5
2000
10
.057
6
2000
25
.077
7
2000
90
.113
8
2000
25
.007
9
2000
Closed Throttle
.143
The 9-mode is performed with the engine in a warmed-up state, at
only one engine speed (except idle). To date, it can be firmly
stated that on all current production engines all efforts at emis-
sion control on HDG engines have been directed primarily at these
modes.
There are three major areas of engine operation which the
transient test contains, but not the 9-mode:
a. Full power operation;
b. Transient operation, at all speeds and loads;
c. Cold engine operation.
These areas give rise to the measurable emission differences, and
reflect where control technology will need to be directed for
1984. In this analysis we will show that full power (power en-
riched) LA Freeway modes are the major source of CO emissions in
current technology engines, and also a significant source of HC on
the higher emitting engines. Secondly, the major source of HC on
the lower HC emitting engines will be shown to be the cold engine
A-3
-------
operation. Finally, ori the lower-emitting engines, it will be
shown that non-cold start HC and the remaining CO emissions are
not as attributable to any one mode or source, and are primarily
relatable to inadequately controlled mixture calibration as the
engine undergoes transients at all speeds and loads throughout the
entire test cycle.
4. Full Power Operation
Under wide open throttle (WOT) conditions, additional fuel is
added to the combustion mixture. This power enrichment causes
richer than stoichiometric mixtures, thereby promoting power and
driveability, but drastically increasing unburned fuel (HC) and
partially oxidized fuel (CO) emissions due to lack of oxygen.
Present day engines certified to the 9-mode were emission con-
trolled primarily up to 90 percent power (at only a single speed);
note that current technology engine power valves are calibrated to
cause power enrichment above 90 percent power. Thus, full power
emissions on current technology engines are uncontrolled.
This observation is demonstrated by the data presented in
Tables II-2 and II-3. In both tables, data from all twelve cur-
rent technology engines tested at EPA are presented. In addition,
the engines are also grouped into three categories: high, medium,
and low emitters of a given pollutant. Note mode 7, the LA Free-
way (LAF) in the hot-start portion of the test: 29.6 to 65.7 per-
cent of brake specific CO (BSCO) emissions are attributable to
this high-power segment. More interesting are the trends observed
in segment percentage contributions from the highto the low-
emitting engines. As the average composite BSCO emissions go from
105.5 g/BHP-hr (higher emitters) to 46.1 (lower emitters), i.e., a
2.3 fold decrease, all other subcycle model percentages increase
by approximately two-fold except for the LAF mode, which decreases
in contributing percentage from 56.3 to 36.7 percent (i.e., a
lower percentage of a lower composite number). Had all modes de-
creased proportionally, the model percentages should remain con-
stant. Clearly the major difference between high and lower CO en-
gines is the amount of CO generated during the LAF segment. This
is primarily a result of power enrichment in the carburetor during
the LAF's characteristic high speed, high'power operation. (Per-
haps most indicative is the actual mass of CO generated during the
LAF segment. Note in Table II-3 that total grams of CO generated
in the LAF segment are 50-650 percent higher than those of the
next highest hot start segment.)
The data for HC (Table II-2) is less dramatic with regards to
LAF dominance, but the trends are nevertheless the same. Every
high CO engine, (i.e, those with LAF dominance of CO emissions)
also has dominant LAF HC emissions (ranging from 23.7 to 36.0 per-
cent total contribution). This is logical since in this opera-
tional mode both emissions arise primarily from inadequate oxygen
for total combustion in the fuel-enriched mixture. Again, the
A-4
-------
lower the total HC emissions are, the lesser the percent contribu-
tion of the LAF segment to that total.
In summary, power enrichment occurs at the high power points
throughout the entire transient test cycle, but the majority of
this high power operation is found in the LA Freeway segment.
Emissions performance over this segment is the major differenti-
ating factor between lower and higher emitting engines. Control
or power enrichment is the first and most effective step in re-
ducing CO emissions with or without a catalyst. This will be dis-
cussed further below.
5. Transient Operation/All Speeds and Loads
As the LAF emissions contribution drops when going from the
higher to lower emitting engines, the contribution from other seg-
ments tend to increase until no single segment is dominant. (The
obvious exception to this is cold start HC, which is discussed be-
low.) Aside from certain physical factors,* these emissions arise
from less than accurate fuel metering and mixing as the engine
drives over the entire test cycle. If the fuel flow does not pre-
cisely match the engine inlet air flow at any instant in time,
then too lean or too rich mixture conditions prevail, along with
ensuing lean misfire (high EC) or incomplete combustion (high HC
and CO). This matching is complicated by the inevitable need to
closely match the fuel and air flows at continually varying speeds
and loads while also maintaining power and driveability. All cur-
rent technology engines were emission optimized at idle, and at
eight different steady-state power modes at 2000 RPM. This repre-
sented a reasonably simple design/calibration problem, as evi-
denced by the engines' emission performance over the 9-mode test.
Once outside that limited regime of emission-optimized modes, how-
ever, such as on the road or on the transient test, emissions re-
main virtually uncontrolled. Little design attention with respect
to emissions has been given to the majority of the engines' opera-
ting ranges.
Precise matching of fuel and air flows under varying condi-
tions, including transient enrichment by the accelerator pump for
driveability, is a major emission-related problem of mixture con-
trol. Another is the problem of achieving as homogeneous (per-
fectly mixed) a fuel/air (F/A) mixture as possible. Incomplete
mixing (including liquid fuel deposition on the manifold or com-
bustion chamber walls) produces- localized pockets of rich and lean
mixtures, resulting in an overall increase in HC and CO emis-
sions. Complete mixing is also critical to achieving uniform A/F
ratios from cylinder to cylinder, again to optimize overall emis-
sion performance.
* Combustion chamber design affects wall quenching. Inlet
manifold design affects mixture distribution between cylinders,
fuel deposition in the manifold, and heat exchange charac-
teristics. All of these in turn affect HC and CO emissions.
A-5
-------
The above problems are not new, are well recognized, and have
already been addressed in the light-duty passenger car fleet. Ex-
perience with the light-duty fleet has indicated, however, that
there exists a definite limit to the amount of HC and CO emission
reductions achieveable through recalibration before power, drive-
ability, and/or fuel economy become unacceptable. For this rea-
son, catalysts become inevitable at lower emission standards, both
for their effectivenss and the flexibility in engine calibration
their effectiveness permits.
6. Cold Engine Operation
Cold start emissions are substantially higher than those of a
fully warmed-up engine, and usually require separate attention
during control system design. Again referring to Tables II-2 and
II-3, we note that cold start HC contributions are high, and be-
come dominant at lower overall levels of HC emissions. Cold start
CO on the other hand has a relatively minor effect on an overall
basis. This phenomenom is typical, though perhaps exaggerated by
the lack of design control in the past, and is attributable to the
fact that a very rich mixture is needed for starting and drive-
ability in a cold engine, to compensate for deposition of a large
part of the fuel on cold manifold walls. This rich mixture is
provided by the choke mechanism, either manual or automatic.
Emissions arise both from this overall rich mixture, misfire, and
from the eventual evaporation of the condensed fuel. Emissions
have not been a design constraint in the past for cold starting,
only startability, driveability, and power. The transient test
procedure itself is demanding, requiring both emission control and
high power driveability early in the cold start cycle.
C. Available Control Techniques
1. Overview
Widespread introduction of new non-catalyst technologies is
assumed to be an unrealistic scenario for the 1984 model year.
This is a function of the remaining leadtime, and cost - the in-
tent of this rulemaking is to ease the capital expenditure burden
on the industry. Technologies which EPA expects to be implemented
for 1984 will not be new, but rather will represent refinements,
recalibrations, and optimizations of current technologies.
2. Improvements to Fuel Metering
By and large, fuel metering improvements will be the single
most effective strategy for reducing overall HC and CO emissions
in 1984 engines, especially when optimized for the transient
test. These improvements include modifications to carburetors to
achieve more precise F/A ratio control, and recalibration to lean-
er F/A ratios on an overall basis, and especially under transient
conditions and WOT.
A-6
-------
Figure II-l presents the CO emission distribution of the 1979
baseline engines. Note that two mutually exclusive sets of car-
buretors are found above and below 70 grams/BHP-hr, representing
higher and lower emitting engines. Some carburetors (those below
70 g/BHP-hr) meter fuel more accurately under transient conditions
even though also optimized for the 9-mode. Power enrichment,
sometimes observed at 4-6 percent CO (40,000 to 60,000 ppm in the
raw exhaust) contributes substantially to these CO levels, as
shown above in Table II-3. At any rate, we infer from Figure II-l
that since two groups of carburetors produce two radically dif-
ferent emission rates on a test procedure for which neither was
optimized, the higher emitting group is unrepresentative of cur-
rent technology and should not be considered a realistic starting
point when extrapolating achieveable emission reductions. They
represent excessive power enrichment/inaccurate fuel metering pro-
ducing twice the CO emissions of other engines of equivalent power
and displacement. The realistic current technology CO baseline
is, therefore, presumed to be in the range of 40-60 g/BHP-hr. It
is from this range downward in which development work will be con-
centrated.
The prime result of recalibration will need to be leaner mix-
ture calibration, and leaner WOT and transient enrichment, thereby
reducing both HC and CO emissions.
3. Improved Mixture Distribution
As overall calibrations get leaner, it becomes more important
from a power, driveability, and emissions standpoint that the F/A
mixture be as homogeneously mixed as possible and the mixture dis-
tribution to each cylinder is uniform. Localized rich or lean
"pockets" in the mixture should be eliminated by the time it en-
ters the cylinder. Assuring uniform F/A mixture distribution to
each cylinder is also important. Too lean a mixture in one or
more cylinders will force recalibration to a richer operating
point to accommodate the needs in that cylinder, which will in
turn cause too rich a mixture in other cylinders.
This is essentially a problem of improving the mixing of air
and fuel in the manifold prior to cylinder induction. The liquid
fuel must be vaporized and then mixed, requiring heat energy and
substantial turbulence. Deferring the problem of cold starting
until later, heat energy arises from the air itself and from the
warm manifold. Improvements would come from redesign of the mani-
fold to increase turbulent mixing, and to increase heat transfer
(perhaps by heating intake air by drawing it across the exhaust
manifold) to the intake air or air/fuel mixture.
4. Other Physical Modifications
Other physical changes to the engine have been proven to re-
duce unburned fuel emissions, such as decreasing surface-to-volume
ratio of the combustion chamber to minimize wall quenching, reduc-
tions in cylinder "dead" volume, etc. Although these may be per
A-7
-------
formed on some engine families, we do not consider fleetwide phy-
sical redesign of engine combustion chambers for all families to
be realistic or necessary for 1984.
5. Other Calibration Optimizations
As mixture calibration optimization reaches its limit with
respect to attainable reductions, other calibrations - notably
spark timing - can be utilized to further reduce HC and CO. Iron-
ically, these reductions are made possible by the other 1984 MY
emission standard for heavy-duty engines: the NOx standard of
10.7 - grams/BHP-hr. NOx emissions at this level are relatively
uncontrolled, and will allow ignition timing calibration to be set
near MBT* - the most efficient calibrations. The higher NOx
standard permits both lean mixtures and optimum timing advance -
both of which increase NOx but decrease HC and CO emissions and
fuel consumption.
Furthermore, spark timing can also be optimized for the cold
start portion transient test procedure. The light duty fleet cur-
rently uses electronically-controlled spark timing to optimize ig-
nition under all engine operating conditions in the Federal Test
Cycle to minimize emissions and maximize fuel economy. The
methodology and technology is entirely applicable, if necessary,
to HDG engines on the transient test.
6. Improved Warm-up Characteristics
As emission levels decrease with mixture and Ignition timing
optimizations, the limiting factor for HC reductions is clearly
the engine's performance on the cold start portion of the tran-
sient test. As Table II-2 above indicated, cold start HC emis-
sions are the dominant fraction of engine-out HC.
Two strategies exist for reducing cold start emissions: re-
strict the amount of cold mixture enrichment, and Increase the
warm-up rate of the engine. The former is straightforward, and
limited by the amount of leaning a cold engine can withstand and
still maintain the high driveability and performance both the road
and the transient test require. This is done by choke recali-
bration. Increasing the warm-up rate of the engine can be ac-
complished in primarily two ways: decrease the efficiency of the
overall combustion cycle, and use exhaust gas heat to rapidly warm
the intake manifold and/or intake air. Cycle efficiency reduc-
tions are best achieved by changing spark timing as a function of
engine temperature: less efficient spark timing calibrations re-
duce engine efficiency, and Increase the amount of waste heat re-
jected to the combustion products and thereby conducted to the en-
gine itself. The result is a faster warm-up; less time spent in a
cold state reduces cold emissions.
* "MBT" denotes the minimum timing retard (i.e. maximum timing
advance) at which maximum power is obtained without inducing knock
reactions.
A-8
-------
Cold start HC emissions, as elaborated above, are presently
uncontrolled, and generally dominate at lower overall HC emission
levels. Table II-5 lists current technology engines, and the per-
cent increase in composite total transient test QC and CO emis-
sions attributable to the cold start cycle. (The cold start cycle
is identical in every way to the hot cycle, with the sole excep-
tion of engine temperature.) From this we can infer the amount of
emissions generated by the "cold"* engine temperature. Figure
II-2 graphically portrays the percentage attributable to cold en-
gine temperature versus the total composite test result, and il-
lustrates the general trend of increasing impact of cold HC emis-
sions with lower overall HC emission rates. (Note that there are
exceptions to the trend). All of the 1979 baseline engines tested
by EPA were equipped with automatic chokes; the high degree of
scatter in the Table II-5 data indicates that varying choke cali-
brations are possible. Since the varying engine calibrations were
not optimized for either a transient test or a cold start, the
available data does not lend itself to determining the exact con-
tribution of the cold start to overall test results at any given
emission level. The data do indicate, however, that it can be
significant ( probably 10-40 percent). The real question is to
what degree cold start HC emissions can be reduced by choke recal-
ibration/improved warm-up. Experience tells us that significant
reductions are achievable from uncontrolled engines.
7. Summary of Possible Control Techniques
Based on the discussion above EPA has identified a number of
potential means of reducing HC and CO- emissions from HDG engines.
These are summarized below.
a. Carburetion - modifications and improvements to the
power enrichment, accelerator pump, and general fuel metering sys-
tems .
b. Calibrations - spark timing, A/F ratio, and EGR flow
ratfi calibrations.
c. Manifold/Combustion Chamber Redesign - intake manifolds
could be redesigned to improve the homogeneity of the F/A ratio.
Combustion chamber surface-to-volume ratio could be decreased and
cylinder dead volume minimized to lessen fuel quenching on
cylinder walls.
d. Air Injection System - Increased air injection to the
exhaust manifolds will increase the HC and CO oxidation. This
system could be further improved by an air modulation system and
possible recalibrations of the pressure relief and diverter
valves. Some exhaust manifold modifications may also aid the
efficiency of the air injection system.
* "Cold," for laboratory test procedure purposes, is a tempera-
ture between 68° and 86°F.
A-9
-------
e. Automatic Choke - the use of a properly calibrated
automatic choke would decrease cold start HC and CO emissions and
improve warm-up time.
f. Early Fuel Evaporation (EFE) - this system involves the
use of exhaust gases to warm the air-fuel mixture by directing
some of the exhaust gases through a passage below the carburetor.
A warmer A/F mixture improves the fuel distribution to the cyl-
inders and results in lower emission levels and shorter warm-up
periods.
g. Heated Air Intake - heated air intake or a modulated
air cleaner system uses exhaust gases to warm the intake air to
the carburetor. This improves engine warm-up time and reduces
emissions by allowing leaner carburetor calibrations.
h. Exhaust Gas Recirculation (EGR) - EGR primarily used
for NOx control, can also be beneficial with regards to HC con-
trol. Besides its overall leaning effect on the mixture, it also
permits recombustion of a percentage of the exhaust gases. Sim-
ilarily, increased valve overlap works as a form of "internal
EGR."
The effectiveness of modifications and hardware of this type
has been demonstrated in the light-duty vehicle and light-duty
truck fleets for several years. These control strategies should
be available for the 1984 model year HDG engines and should pro-
vide substantial HC and CO reductions over current levels.
8. Tradeoffs
The emission control strategies discussed above have trade-
offs with respect to fuel economy, power, and driveability. Lean-
er mixtures, less power enrichment, and quicker engine warm-up all
improve fuel economy, but when carried to excessive degree could
impair power and driveability. An increase in air injection would
also cause a small fuel economy loss. EPA now believes that the
fuel economy impacts of these regulations will be basically neu-
tral. The limits to emission reduction will be determined equally
by power requirements and driveability needs in addition to any
fuel economy concerns.
D. Attainable Reductions/Proposed Emission Standards
As described above, several relatively simple and effective
means of emission control are available. At this time, EPA has
limited data as to the absolute effectiveness of a given technique
on heavy-duty gasoline engines. For example, no testing has been
performed to date on a current engine where mixtures were leaned
out, spark timing curves optimized, power enrichment limited, and
fast warm-ups or fast opening chokes were initiated. It is dif-
ficult to quantitatively predict attainable emission reductions
without results of such testing.
A-10
-------
One approach to deriving achievable standards would be to use
an engineering estimate of the efficiencies of the previously des-
cribed reduction techniques * These efficiency estimates could
then be applied to the current baseline emissions data to calcu-
late what emission levels could be reached. Lacking any other
substantive data or technique at this time this methodology will
be used.
The emission reduction efficiencies used in this analysis are
those expected from the lower emitting engines in the current
technology baseline (see Tables II-2 and II-3) so the average HC
and CO emission rates from the low emitting engines will serve as
the baseline levels. One might question the use of the lower
emission levels as not being representative of the average emis-
sion levels. However for the higher emitting engines the effic-
iency estimates would in turn be substantially larger. We have
chosen to use the lower emitting engines because they already re-
flect what could easily be achieved on other current technology
engines with even minor calibration changes.
Tables II-2 and II-3 clearly indicate that the HC and CO
emission levels in certain modes are so large that they require
specific attention in this analysis. HC emissions could be di-
vided into "cold/warm start" and "other." CO emissions could be
divided into "LAF" and "other." Table II-6 lists the emission re-
duction techniques together with the modes in which they will be
effective in gaining emission reductions. This information will
serve as a background for the discussion which follows.
1. Hydrocarbons [3]
As shown in Table II-2 cold/warm start emissions account for
49 percent of the HC emissions. Thus the remaining 51 percent
comes from the "other" six portions of the test. In terms of the
i^erage of the low emitting engines from Table 11-3 the "cold/warm
start" portions account for 0.92 g/BHP-hr and the "other" portions
account for 0.96 g/BHP-hr.
With the emission control strategies shown in Table II-6 we
believe that substantial reductions in HC emission levels are
easily achievable. Our current belief is that reductions of 50-60
percent are possible in the "cold/warm start" portions of the test
through the means shown in Table II-6. For all practical purposes
"start" emissions are uncontrolled on the current test procedure.
EPA also believes that reductions of 30-40 percent are also avail-
able on the other portions of the test procedure. Assuming the
ranges of engineering estimates of reduction efficiencies given
above, achievable emission levels can be calculated.
a. Cold/Warm Start Reductions
High Estimate; (0.92 g/BHP-hr)(60%) = 0.55 g/BHP-hr
A-ll
-------
Low Estimate: (0.92 g/BHP-hr)(50%) = 0.46 g/BHP-hr
New Range: 0.37 - 0.46 g/BHP-hr
b. Reductions in Other Portions
High Estimate: (0.96 g/BHP-hr)(40%) = 0.38 g/BHP-hr
Low Estimate: (0.96 g/BHP-hr)(30%) = 0.29 g/BHP-hr.
New Range: 0.58 - 0.67 g/BHP-hr
c. Achievable Emission Levels
Using "High Estimate": 1.88 - 0.55 - 0.38 =0.95 g/BHP-hr
Using "Low Estimate": 1.88 - 0.46 - 0.29 = 1.13 g/BHP-hr
Emission levels in the 0.95 - 1.13 g/BHP-hr range would support an
HC emission standard of 1.3 g/BHP-hr.
Using a full life multiplicative deterioration factor of 1.2
and an HC variability of 10 percent, the expected target HC levels
are 1.1 g/BHP-hr for 1984 (no SEA) and 1.0 g/BHP-hr when SEA be-
gins in 1986. The range of achievable emission levels shown above
supports the feasibility of these targets and thus the 1.3 g/BHP-
hr standard.
2. Carbon Monoxide
As shown in Table II-3 the "LAF" (LA Freeway) CO emissions
account for 43.2 percent of the total. Thus the remaining 56.8
percent arises from the "other" portions of the test. When these
percentages are applied to the average low CO engines of Table
11-3, the "LAF" accounts for 19.9 g/BHP-hr and the other portion
accounts for 26.2 g/BHP-hr.
With the emission control strategies shown in Table II-6 sub-
stantial reductions in CO emission levels are easily achievable.
Reductions of 40-50 percent are possible in the "LAF" portion of
the test through the means in Table II-6. Emissions under the
high-speed, high-power operation characteristic of the LAF portion
are relatively uncontrolled because of the limited power demands
of the 9-mode test procedure. Reductions of 30-40 percent are al-
so possible from the "other" portions of the test procedure.
Given the engineering estimates of reduction efficiencies shown
above, achievable emission levels can be calculated.
a. LAF Reductions
High Estimate: (19.9 g/BHP-hr)(50%) = 10 g/BHP-hr
A-12
-------
Low Estimate: (19.9 g/BHP-hr)(40%) = 8 g/BHP-hr
New Range: 9.9-11.9 g/BHP-hr
b. Reductions in Other Portions
High Estimate: (26.2 g/BHP-hr)(40%) = 10.5 g/BHP-hr
Low Estimate: (26.2 g/BHP-hr)(30%) =7.9 g/BHP-hr
New Range: 15.7 - 18.3 g/BHP-hr
c. Achievable Emission Levels
Using "High Estimate": 46.1 - 10 - 10.5 = 25.6 g/BHP-hr
Using "Low Estimate": 46.1 - 8 - 7.9 =30.2 g/BHP-hr
Emission levels in the 25.6 - 30.2 g/BHP-hr range would support a
CO emission standard of about 35 g/BHP-hr. Using a full life mul-
tiplicative deterioration factor of 1.1 and a CO variability of 20
percent, the expected target CO levels are 31.8 g/BHP-hr for 1984
(no SEA) and 25.5 g/BHP-hr when SEA begins in 1986. The range of
achievable emission levels shown above supports the feasibility of
these targets and thus the 35 g/BHP-hr standard proposed here.
Considering all of the factors bearing on this analysis
(cost, fuel economy, leadtime, power, and driveability), EPA be-
lieves that the standards herein discussed are achievable for all
HDG engines for the 1984 model year. However if during the com-
ment period further data and information would prove the standards
to be infeasible the option for further relaxation for final rule-
making exists.
E- Idle Emission Standard
For heavy-duty gasoline engines, the 1984 idle CO standard is
0.47 percent (raw exhaust composition). Table 11-7 presents the
current technology idle CO baseline. Note that five of twelve en-
gines already comply. Given the fact that substantial leaning of
mixtures will be performed to meet the transient standards, there
is no reason to believe the idle circuits of the remaining engines
cannot be improved. EPA judges compliance with the idle standard
to be relatively straightforward and will pose no problems to
manufacturers even considering any small deterioration factor
which may need to be included.
A-13
-------
Table II-l
1979 HDG Current Technology Baseline
Engine
Family
HC*
(g/BHP-hr)
CO*
(g/BHP-
Ford 400
6.6L "E"
4.89
(H)**
112.4
Chrysler
440
RBM
3.83
(K)
112.4
Ford 370
6.1L "E"
3.51
(H)
47.8
IHC 446
MV8
3.27
(E)
90.4
GM 350
113
3.14
(M)
118.1
Chrysler
360
LAI
2.67
(M)
96.1
GM 350
113
2.48
(M)
64.8
IHC 345
V345
2.44
(M)
34.4
GM 454
114
2.30
W
51.6
GM 366
114
2.16
(L)
43.4
GM 292
112
2.12
a)
55.0
GM 454
115
1.31
(L)
78.5
(H)
(H)
(H)
(H)
* Average of several tests.
** Engines are classified as high (H), moderate (M), or lower (L)
emitters of a given pollutant. Note that a high HC engine is also
usually a high CO engine, but not in every case.
A-14
-------
Table II-2
1HC 446
IHC 345
CM 366
CM 350
F 400
F 370
Engine by iingLre Transient HC Emission Breakdown
lc]
Cold
Sta rt
20
Hot
Start
Compos il
1
2
3
4
Minute
5
6
7
8
Test
NYNF
lamf
LAF
NYHF
j'iuse
NYNF
LAMF
LAF
HtNF
Re s u 11
24,73
17.11
15.10
5.55
9.66
11.14
12.53
5.71
_
23.26
6.97
1.61
5.23
10.65
4.61
1.50
5.36
3.32
.269
.18B
,166
.-06-1
.677
.736
.830
.373
3.32
6. n
5.62
5,0%
1.8%
20.4%
22.2%
25.0%
11.2*
100%
25.50
9.09
4.93
3.80
4.B2
6.29
4.b4
3.40
-
64,64
5.51
0.78
5.20
7.14
3.40
0.75
4.38
2.35
.40
.13
.07
.06
.44
.54
.42
.29
2.35
17.0%
5.5%
3.0%
2.6%
18.7%
23.0%
17.9%
12.3%
100%
47. 86
12.73
5.95
2.61
4.96
4.89
5.07
2.62
-
94.7
6.2 4
0.81
2.69
5.69
2.28
0.69
2.74
2.22
.64
.16
.08
.03
.39
.36
.37
.19
2.22
28.8%
7.2%
3.6%
1.4%
17.6%
16.2%
16.7%
8.6%
100%
61.49
12.57
6.42
2.81
4.71
6.50
4.5b
2.74
-
95,0
6.30
.92
3.06
5.56
3.16
.65
2.95
2.57
.66
.17
.08
.04
.37
.49
.35
.21
2.57
33.531
6.6%
3.1%
1.6%
14.4%
19. IX
13.6%
8.2%
100%
32.9L
16.16
14.67
6.68
8.62
10.11
13.17
5.69
_
46.77
9.66
2.60
. 9.38
12.10
5.77
2.33
8.02
4.80
.56
.26
.24
.11
.87
.96
1.25
.55
4.80
11.7%
5.4%
'5.0%
2.3%
18.1%
20.0%
26.0%
11.5%
-
20.U
8.13
7.39
1.71
8.05
7.65
6.96
3.4b
_
52.25
5,29
1,35
2.47
13.37
4.78
1.26
5.02
3.31
.36
,14
.13
-03
.85
.76
.69
.35
3.31
10.93;
4,2%
3.9%
.9%.
25.7%
23.0%
20. b%
10.6%
100%
High
Medium, or
Low Emitterle]
-------
Tabid H-2 (cont'd)
Engine by Engine Tr-anaient BC Emission Sre-akdourj
Cold Start
20
Hot
Start
Composite
liifih
1
2 3
4
Minute
5-
6
7
8
Test
Medium, or
NVNF
LANF LAF
NYHF
Pause
NYHF
LANF
LAF
MYNF
He s u 11
Low tin i c t e r
1.
8.56
7.18 3.22
3.63
10.23
6.41
7.87
3.52
___
2.
13.11
3.42 i.oa
3.96
13.54
2.99
1.04
3.77
2.45
M
3.
.11
.09 .10
.05
.80
.47
.56
.26
2.45
4.
4.5*
3.62 4.0%
2.0%
32-73!
19.2%
23. IX
10.21
100%
I.
17.38
10.57
24.67
7.76
C 440
2.
20.12
4.10
2.78
7.41
3.
.19
.11
.26
.08
4.
5.02
2.9%
6.8%
2.1%
1.
16.38
3.88
5.34
1.68
CM 454
2.
19.06
1.74
.63
1.83
3.
.20
.05
.06
.02
4.
15.5%
3.9%
4.7%
1.6%
1.
47.31
4.33
2.08
1.64
CM 292
2.
65.65
2.62
0.39
2.08
3.
.60
.01
.03
.03
4.
37. 7%
3.3S
1.4%
1.4%
1.
44.54
15.43
6.30
6.43
CM 454
2.
62.3S
5.75
0.68
5.83
3.
.44
.15
.06
.06
4.
17.9%
6.1%
2.4%
2.4%
10.25
9.32
22.22
9.10
-
11.32
3.65
2.40
8.69
3.81
H
.67
.57
1 - 3 J
.56
3.81
17.6%
15.0%
36.0%
14.7%
100%
4.94
2.39
4.95
1.57
_
5.82
1.07
0.60
1.72
1.29
L
.35
.17
.33
.11
1.29
27.1%
13.2%
25.6%
d. 51
100%
4-12
6.1?
.43
20. iX
11.&5
11.24
.70
26.5%
3.71
2.20
.37
17.5%
6.97
2.52
.39
15.91
1.95
0.37
.2ii
9.4%
5.65
0.57
.33
13.4%
1. 83
2.33
.19
9.0J
5 . bO
5.25
.33
13. 4S
2.12
2.22
1005!
2.46
2.46
100%
-------
Table 11-2 (cont'd)
CH 350
Cold
Start
*nsinr
. -En.i>ine
20-
Transient
HC Emission Breakdown
Ho t S L£ rt
Tota 1
liiph
1
2
3
A
. suite
5
6
7
B
Teat
H^diusa,. or
NYNF
LAWF
LftP
NYNF
•aur.fi
NYNF
LANF
LAF
NYHF
Compos ice
Low tmitter
1.
21.04
6.13
10.39
3.69
3.66
5.01
9.51
3.71.
-
2.
31.48
3.53
1. 71
5.34
5.13
2.93
1.57
4.51
2.66
M
3.
.34
.09
.16
.06
.34
.46
. 87
.34
2.66
A.
12.82
3.4%
6.02
2.32
12.82
17.32
32. 7X
12.82
ioux
Ave ra^e HC,
tmission Level
Averaga:
All
Engines
17.OX
4.B2
4.U 1.92
21. IS IB.43! 21.7% 11.OX 1D0X (12 engines) 2.7b
Average:
High HC
Engines
>
' Average:
~jMert. HC
Engines
Average:
Lou HC
Engines
9.IX
4. 17.IX
4. 27.32
A.5Z 5.2% 1.82
5.02 3.7* 2.22
4.62 3.22 1.52
20.52 20.12 27.02 12.02 1002 H (A engines) 3.bl
21.AX IB.92 20.3X 11.42 10UX h (5 engines) 2.50
21.72 15.62 17.22 8.7X 1002 L (3 engines) l.bb
[a J Total grams per subcycle.
(b| Grams per brake-horscpower-hour pec subcycle.
|cl Subcycle contribution, in e£fectively-weighted grams per brakeHiorsepower-hour, to the composite test result. (When
added together, all subcycle contributions arid up to the composite test result). For methodology, see Keterence 1,
pp. 4-5.
[dj Relative percentage of subcycle contribution (3) to the total composite test result.
(el In graitB per brake-horsepo-uer-hour: High (H)^ 3.3
3.3 medium 2.3
Low (L)^2.2
-------
Cold
Start
1
2
3
4
HVHF
LAKF
LAP
NYNF
1 [a 1
236.4
245.2
774.8
127.2
IHC 446
211> 1
222.4
99.9
93.0
119.9
3[cJ
2.79
2.73
8.63
1.42
4 Id J
3.0%
2.9%
9.3%
1.5%
1.
90.3
84.7
153.2
60.0
IHC 345
2.
229.6
51.4
24.2
79.2
3.
1.40
1. 24
2.24
.88
4.
4.3%
3.8%
6.8%
2.7%
1.
143.7
140.7
187.7
86.1
CM 366
2.
284.3
69.1
25.5
88.8
3.
1.9
1.8
2.4
1.1
4.
4.5%
4.3%
5.7%
2.6%
1.
171.2
155.2
404.5
102.6
CM 350
2.
264.5
77.7
57.fi
116.6
3.
2.4
2.1
5.3
1.4
4.
3.5%
3.1%
7.8%
2.1%
1.
222.9
162.4
620.6
130.6
F 400
2.
316.7
97.0
109.9
183.5
3.
3.8
2.6
10.0
2.1
4.
3.4%
2.3%
8.8%
1.9%
1.
85.2
106.6
230.7
21.4
F 370
2.
221.5
69.4
42.1
30.9
3.
1.5
1.8
3.9
.4
4.
3.3%
4.0%
8.7%
.9%
Emission Breakdown
Hot Start Composite Hitjh
6 7 8 Tost Medium, or
LAHF LAF hYNF hesu 11 Low Emitter [ft |,
2U0.0 70B.1 122.7
82.7 84.7 115.1 92.t>« M
13.3 47.1 B.20 92.&8
14.32 50.7X b.OX lUOi
60.Z 150.6 56.1
32.5 23.2 72.3
5.29 13.11 4.94
16.1% 40.0% 15.1%
32. 8 L
32. b
100%
113.3 167.0 B8.2
52.9 22.8 92.1 41.9
8.7 12.5 6.6 41.9
20.8% 29.8% 15.b% 100%
130.2 376.6 95.3
63.4 53.9 102.5 67.8
10.1 29.) 7.4 67.6
14.92 43. HX 10.9* lOOSt
161.6
5 JJ2.3
127.3
-
92.2
103.0
179.6
113.2
H
15.6
56.2
12.3
113.2
13.$%
49.6%
10.9%
100%
80.1
206.7
40.3
-
50.0
37.4
58.4
45.0
L
8.1
21.0
4.1
45.l>
18.Q%
46.7%
9.12
100%
-------
Table 11-3 (cont''d)
Erftine-ay-Enfeine Transient CO Emission Breakdown
Cold Stare
2
LANF
3
LA.F
4
r. >-
Pause
hoc Start
6
LMif
7
LhlF
K
trtliF
Composite
TliSt
He s u 11
HlUh
Medium, or
Lou fcmitter
1.
107.5
144.7
863.6
61. 2
56.6
127.9
783.0
56.5
-
C 360
2.
164.7
68.8
113.fi
66.7
76.3
59.5
103.3
62.6
92.0
H
3.
L. 4
1.8
10. 8
.a
4.6
9.6
58.6
4 .4
92.0
4.
1.5%
2.02
11.1%
¦9%
5. OX
10. AX
63. VI
4.62
100*
1.
228.3
203.6
1262.0
100.6
75.2
161. 1
1217.2
94.1
-
C 440
2.
264.3
78.9
142.1
96.0
83.0
63.0
131. 7
89.9
115.6
tl
3.
2.5
2.1
13.1
1.0
5.0
10.0
75.9
5.9
115.6
4.
2.2%
1.83!
11. 3%
.9%
4.3%
8.7%
65. n
5.1*
-
1.
250.3
86.2
769.b
65.2
86.9
102.8
714.1
69.7
-
GH 454
2.
291.3
38.7
91.3
71.1
L02.4
45.8
87.2
76.2
61.9
M
(Short
3.
3.1
1.0
9.0
.8
6.4
7.2
4V.5
4.9
81.9
Block)
4.
3.8%
1.2%
11.OX
1 - CIS.
7.62
b.bx
60.431.
6.OX
lim
1.
315.0
115.7
159.4
64.7
89.4
111.0
161.5
70.0
-
GH 292
2.
437.1
69.9
30.2
81.9
133.7
65.9
30.5
89.1
55.0
L
3.
5.6
2.0
2.7
1.1
9.6
11.2
16.3
7.1
55.0
4.
10.22:
3.62
4.9 X
2.OX
17.5%
20.4X
29.6%
12.92
1004
1.
204.6
175.6
3?6. L
144.6
153.9
157.1
366.2
13&.2
-
CM 454
2.
286.9
65.5
37.9
131.1
146.1
56.7
36.9
124.9
55.9
L
(Tall
3.
2.1
1.7
3.6
1.4
9.3
9.0
20.9
7.9
55.9
Block)
4.
3.8%
3.OX
6.42
2.5%
16.7%
16.IX
37.45i
14. IX
10UX
-------
Table H-3 (cont'd)
GM 350
Cold
Start
Engine-by-Engine Transient
20-
CO Emission Breakdown
Hot Start
To t a 1
High
1
2
3
4 Minute
5
6
7
8
Test
Medium, or
NYNF
LANF
LAF
NYNF Pause
NYNF
LANF
LAI'
NYNf
Compos ite
Low bmitter
1.
196.1
108.9
805.5
68.1
92.1
104.8
640.8
64.8
-
2.
293.3
62.6
132.21
98.6
130.1
61.4
106.0
78.8
101.5
H
3.
3.2
1.7
12.1
1.1
B-5
9.7
59.2
6.0
101.5
4.
3.2%
1.7%
11.8%
1.1%
8.4%
9.6%
58.3%
5.9%
1002
Ave rage
All 4.
Engines
Average
High CO 4.
> r ¦
I Engines
O
Average
Mod CO 4.
Engines
Average
Low CO 4.
Engines
3.931 2.82 9.5% 1.72
2.6% 2.1% 12.6% 1.2%
3.7% 2.22 9.4% 1.6%
5.2% 3.7% 6.52 2.1%
10.7% 14.2% 47.92 9.92 1002 (12 engines)
7.12 11.1% 56.32 7.02 100% (5 engines)
10.8% 11.9% 52.12 8.52 1002 (2 engines)
14.3% 18.3% 36.7% 13.42 100% (5 engines)
Avera6e llC
tmission Level
75. 7
105.5
74.9
4b. 1
[a]Total grams per subcycle.
[b] Grams per brake-horsepower-hour per subcycle.
[c] Subcycle contribution, in effectively-weighted grams per brake-horsepower-hour, to the composite test result. IWhen
added together, all subcycle contributions add up to the composite test result). For methodology, see Reference 1,
pp. 4-5.
[d] Relative percentage of subcycle contribution (3) to the total composite test result.
[e] In grains per brake-horsepower-hour: High (H)^>90
90S medium (M^> 60
Low (L)<^60
-------
Table II-4
9-Mode Versus Transient Emissions
Current Technology Engines[1][2]
BSHC BSCO
Engine
9-Mode
Transient
9-Mode
Transient
1979 GM 292
0.42
2.12
26.86
54.98
1979 GM 454
0.39
2.30
17.33
51.55
1979 GM 350
0.79
3.14
14.62
118.07
1979 IHC 446
0.42
3.27
24.28
90.40
1979 GM 366
0.50
2.16
17.40
43.43
1979 IHC 345
2.73
2.44
17.68
34.44
1979 GM 350
0.59
2.48
20.40
64.76
1979 Ford 400
2.15
4.89
53.16
112.43
1979 Ford 370
1.20
3.51
37.12
47.75
1979 Chrysler 360
1.18
2.67
21.38
98.14
1979 Chrysler 440
0.83
3.83
10.47
112.38
1979 GM 454
0.47
1.31
20.11
78.49
[1] Engines were tested as received from the manufacturers.
[2] All levels are undeteriorated.
A-21
-------
Table II-5
Cold Start Contribution to Composite Emission Results
HC
CO
Composite
Composite
% Due
Composite
Composite
% Due
Engine
HS
Total Test
To CS
HS
Total Test
To CS
Ford 400
4.26
4.80(H)
11.3%
110.4
113.2
2.5%
Chrysler
440
3.70
3.81(H)
2.9%
112.5
115.6
2.7%
Ford 370
3.10
3.31(H)
6.3%
43.5
45.0
3.3%
IHC 446
3.06
3.32(M)
7.8%
90.5
92.9
2.6?.
GM 350
1.71
2.57(M)
33.5%
66.0
67.8
2.7%
Chrysler
360
2.46
2.45(H)
neg.
90.0
92.0
2.2%
GM 350
2.36
2.66(M)
11.3%
97.2
101.5
4.2%
IHC 345
1.98
2.35(M)
15.7%
31.3
32.8
4.6%
GM 454
1.14
1.29(L)
11.6%
79.8
81.9
2.6%
GM 366
1.55
2.22(L)
30.2%
40.4
41.9
3.6%
GM 292
1.38
2.12(L)
34.9%
51.2
55.0
6.9%
GM 454
2.04
2.46(K)
17.1%
54.8
55.9
2.0%
HC
Averages t
High (H);
6.3%
Med. (M);
14.2%
Low (L):
25.6%
Grams/BHP-hr, results of individual tests, unweighted.
A-22
-------
Table II-6
Test Portions/Emission. Reduction Technologies
HC CO _
Other[4]
X
X
X
X
X
X
X
[1]
Sample Bags
1 & 5
[2]
Sample Bags
2, 3, 4, 6, 7, 8
[3]
Sample Bags
3 & 4
[A]
Sample Bags
1, 2, 4, 5> 6, 8
A-23
Cold/Warm Start I'll Qther[2] LAF[3]
Carburetion X XX
Calibrations X XX
Manifold/Combustion X XX
Chamber
Air Injection X X
Automatic Choke X
EFE X
Heated Air Intake X
EGR X
-------
Table II-7
Idle CO Current Technology Baseline Emissions
Complies with
Engine Idle CO (%) 1984 standard?
IHC 446 .299 yes
IHC 345 .402 yes
GM 366 .913 no
GM 350 1.158 no
Ford 400 1.853 no
Ford 370 .515 no
Chrysler 360 .226 yes
Chysler 440 1.279 no
GM 454 .596 no
GM 292 .308 yes
GM 454 .888 no
GM 350 .242 yes
A-24
-------
Figure II-l
CO Distribution of 1979 Baseline Engines
Ford 400
(Kotorcraft
2 bbl)
IHC
V-345
(Hoiley
2bbl)
GM 366
GM 292
(Holley
(Roch.
4 bbl)'
1 bbl)
Ford 370
GM 454
(Holley
(Holley
4 bbl)
4 bbl)
GM 350
(Roch.
2 bbl)
GM 454
(Roch.
4 bbl)
IHC 446
(Carter
4 bbl)
Chrysler 360
(Carter
4 bbl)
30-40
Chrysler 44C
(Carter
4 bbl)
G2M 350
(Roch.
4 bbl)
40-50
50-60 60-70 70-80
CO (grams/BHP-hr)
80-90
90-100
100-120
-------
Figure. II-2
Cold Start Contribution to Total Test
HC Emissions as a Function of Total Test Emissions
H : Average, All higher emitting engines
M : Average, All moderate emitting engin<
L : Average, All lower emitting engines.
40-
35 .
30-
Due to
aid Start 25"
20"
15
10-
—T
5.0
T
©
®*
T
4.0 3.0 2.0
Total Test HC (grams/BHP-hr)
—I—
1.0
0.0
A-25
-------
References
1. Cox, Timothy P., "Heavy-Duty Gasoline Engine Emission
Sensitivity to Variations in the 1984 Federal Test Cycle," SAE No.
801370.
2. Auiler, J., et. al., "Optimization of Automotive Engine
Calibration for Better Fuel Economy-Methods and Applications,"
SAE Paper No. 770076.
3. Here we are addressing total hydrocarbon emissions and
a total hydrocarbon emission standard. EPA intends to propose an
optional non-methane hydrocarbon standard for HDEs in a future
rulemaking.
4. The terms "High Estimate" and "Low Estimate" refer to
the range of reduction efficiencies. The percent figures shown
are the actual efficiencies.
A-27
-------
The Transient Test Study:
A Review of the 1984 Heavy-Duty
Engine Testing Requirements
An ECTD Staff Report
June 1982
-------
Appendix B
The Transient Test Study
-------
-i-
Table of Contents
Page
1. Executive Summary 1-1
2. introduction 2-1
3. Economic impacts, Leadtime, and Facility Status . . . 3-1
4. A Review of EPA's Transient Test Cycle . 4-1
5. The Real Time Cycle 5-1
6. The MVMA Cycle 6-1
7. The Diesel Transient Test/13-Mode Option 7-1
8. Summary and Recommendations 8-1
Appendix 1: 46 FR 31677 (June 17, 1961) APl-1
Appendix 2: 47 FR 1642 (January 13, 1982) . AP2-1
Appendix 3: List of Pertinent References AP3-1
Appendix 4: ECTD's Analysis of the A. D. Little Report . AP4-1
B-2
-------
CHAPTER 1
EXECUTIVE SUMMARY
As part of EPA's efforts at regulatory relief, this study
reexamines transient testing requirements for 1984 and later
model year heavy-duty engines, and recommends changes where
necessary. This study addresses economic issues (e.g., testing
costs, available leadtime, test implementation dates) and
technical issues (e.g., alternative test cycles). This
document will serve as a reference for the upcoming revised
1984 gaseous emission regulations for heavy-duty engines, in
which the specific changes recommended by this study will be
promulgated.
Chapter 2 presents an overview of the study, a brief
history of the transient test procedure and associated issues,
and enumerates the specific issues addressed by the study. In
particular, the issues addressed within are: 1) economic
impact of the transient test, and the possibilities for relief,
2) the technical adequacy of EPA's test cycles, 3) the
acceptability and implementation options for the industry's
proposed alternative test cycles, and 4) the steady-state
option for heavy-duty aiesel engines. Specific technical
modifications to the test procedure will be addressed later in
the revised 1984 gaseous emission regulations for heavy-duty
engines.
Chapter 3 reviews the economic impact, facility status,
and leadtime requirements associated with running the transient
test procedure, as of November 1981, the majority of facility
expenditures had been made; these investments are unrecoverable
if the transient test is delayed or withdrawn. As a result,
most manufacturers support the retention of a transient test.
Leadtime requirements, however, indicate that some relief is
necessary in 1984, i.e., steady-state testing or carryover on
the steady-state test in conjunction with optional
certification on the transient test. international Harvester
(gasoline) currently plans to leave the gasoline engine market
the year the transient test becomes effective; a single year
deferral will ease its withdrawal. Chrysler has refused to
make investments in both transient facilities and engine
development work to meet stricter standards, thereby allowing
itself to be driven from the market when the new requirements
become effective. There is no fair administrative way,
however, by which EPA can provide relief to Chrysler alone on
what is essentially its business decis-' n
Chapter 4 reviews the more recent criticisms of EPA's
specific gasoline and diesel test cycles. The criticism
applies to the representativeness of the test cycles and
operational problems associated with actually running the
test. A.D. Little inc.'s Report to the MVMA, harshly critical
B-3
-------
1-2
of EPA's methodology, is reviewed. (EPA's comprehensive
analysis of the A.D. Little Report is attached as Appendix
IV.) EPA finds no reason to conclude that the EPA cycles or
test are unrepresentative. The industry criticism is almost
entirely speculative with no emission evidence for
substantiation. All recent emission data {including emission
correlation of EPA's cycles with the alternative cycles
proposed by the industry) continue to confirm the acceptability
of EPA's test, both in construction and in operation.
Chapters 5 and 6 review the alternative test cycles
developed and proposed by the EMA and MVMA. In general, the
cycles are statistically similar despite the fact that they
were developed using different methodologies. With respect to
emissions, they correlate well with EPA's cycles; absolute
emission levels measured over the alternate cycles are less
(especially for HC) than those measured on EPA's cycles. In
fact, at a given numerical standard the alternate cycles are
less stringent; adoption of the alternate cycles without
adjusting the EPA cycle-based standards represents a relaxation
of emission standards by test procedure change. More
comparative testing between the EPA and MVMA cycles may be
necessary, due to the small existing data base and unresolved
technical concerns about the MVMA cycle.
Chapter 7 addresses the issues sue rounding the diesel
engine transient test and the 13-mode option. Based upon
leadtime considerations, a single model year deferral for 1984
appears necessary. The latest 13-mode versus transient data
indicate that the steady-state test is indeed a poor predictor
of transient emissions. With respect to emission control,
there is little difference between optional steady-state HC
standards of 0.5 g/BHP-hr and 1.3 g/BHP-hr; a single year
deferral is recommended simply to provide maximum flexibility
to the industry as they spread transient certification over two
model years. Also reviewed is the option of allowing continued
carryover of steady-state data . for several years until
recertification is required. This option would likely defer
almost all transient testing. It would also preclude HC
emission reductions made possible by the transient test, in
spite of the substantial investments already made by the
industry for transient test facilities. This degree of relief
is unnecessary and unjustified.
Chapter 8 summarizes the conclusions of the study and
makes the following recommendations:
1. Finalize the transient test for 1985; allow its use
as an option in 1984. Allow gasoline and diesel engine
manufacturers to carryover 1983 steady-state data (1979
standards and regulations) for 1984. After 1984, all
certification testing will be based upon the transient test.
B-4
-------
1-3
2. Allow the use of the EMA and MVMA alternative test
cycles as options, certifiable to optional emission standards
which are equivalently stringent to EPA cycle-based standards.
{Note that these alternative cycle standards may change,
especially for the MVMA cycle, as the comparative data base is
increased.) Eventually select single cycles and standards.
3. Finalize necessary technical amendments to the test
procedure in the upcoming revised 1984 gaseous emission
regulations, to minimize both cost and complexity.
4. continue cooperation with the heavy-duty engine
industry to establish correlation between laboratories and to
refine the transient test procedure as necessary.
These recommendations provide the needed relief to the
industry; the environmental impacts of these recommendations
are expected to be minimal.
B-5
-------
CHAPTER 2
INTRODUCTION
I. Overview
As part of its commitment to regulatory relief, EPA
initiated this study to review the mandatory 1984 heavy-duty
engine certification test procedure. Specifically, the study-
is intended to determine if there is a need to revise transient
testing requirements. Information used in the study was
solicited directly from the manufacturers in a separate Federal
Register notice (Appendix I) and as part of a Notice of
Proposed Rulemaking for Revised Gaseous Emission Regulations
for 1984 and Later Model Year Heavy-Duty Engines (Appendix
II). This study will serve as a reference document for that
upcoming final rulemaking. In particular/ it will serve as a
large part of the Summary and Analysis of Comments pertaining
to the test procedure. The recommendations for revisions to
testing requirements contained within this report will be
incorporated into that final rulemaking.
II. The History of the Transient Test Procedure
The transient test procedure for heavy-duty engines is not
a recent phenomenon* its origins lie at the beginning of the
effort to control air pollution from mobile sources.
As early as 1967, studies indicated that steady-state
emission tests were unrepresentative of on-road truck
emissions.* Based upon this and light-duty experience,** EPA
began the development of a more representative heavy-duty
engine test procedure in 1972. (This was the same year in
which the Federal transient emission test became effective for
light-duty vehicles.) Following studies which developed a
sampling plan for New York City and Los Angeles trucks, an
on-road truck operational study (the CAPE-21 Project) was
initiated in late 1973 under joint management by EPA and the
industry-staffed Coordinating Research Council. The New York
and Los Angeles phases of the project were completed by May of
1975. (Note that the data collection methodologies were
derived by both the industry and EPA.) Editing of this on-road
data and computer generation of the test cycles***
* See Appendix III for a complete listing of pertinent
technical reports and references.
** The original LDV emission test was the California 7-mode,
a steady-state test, and was used through 1971. The modal
test was found to be inadequate, especially when emission
control technology was applied to meet it, and it was
replaced in 1972 by the transient CVS-72 procedure.
*** See Chapter 4 of this study.
B-5
-------
2-2
began immediately, followed by selection of the final test
cycles in November of 1977. During this time EPA also
evaluated testing methods (engine vs. chassis dynamometer
testing) ana emissions sampling methodologies to develop the
accompanying test procedure. The Draft Recommended Practice
for the new test procedure was published in August of 1978.
The Clear. Air Act Amendments of. 1S77 directed EPA to
conduct^ uncontrolled EC and CO baseline studies to derive
emission standards for the 1983 model year. EPA began baseline
testing of 1969 model year engines using the new test procedure
in February of 1978} EPA's contractor {the Southwest Research
institute in San Aiitonio, Texas) also began gasoline engine
transient testing in June of 1978. Following this testing, the
resulting emission standards and the new transient test
procedure were proposed in February cf 1979 and promulgated the
following December. During and after this time period, EPA
conducted extensive transient testing projects, including 1969/
1972/73i current technology, and prototype gasoline engine
baselines, and current technology diesel engine baselines.
*The manufacturers raised several issues in response to the
proposed rulemaking of February 1979; 1; EPA's technical
justification for a transient test, 2) the representativeness
of EPA's test cycles, 3) the lack of validation of the test
procedure, 4) the availability of test procedure alternatives,
5) the industry's lack of transient experience and its
inability to comment meaningfully, 6) the technical adequacy of
the overall test procedure, 7) alternative t^st cycles for eddi
current dynamometers, and 8) technical details of the
procedure. ihe original summary and Analysis of Comments to
the HFRH dealt at length with all of these issues; the analysis
strongly defended EPA's original proposal, although several
changes to the test procedure were made end a single year
steady-state option was given to the diesel engine industry.
Since promulgation of the test procedure, the industry has
continued to raise concerns about the transient test. As part
of EPA's recent efforts to provide regulatory relief, the
Agency solicited manufacturers for information on June 17,
1981, {46 FR 31677);* this information pertained to leadtime,
economic impact, transient versus steady-state emissions, ana
technical problems associated with the new test procedure. EPA
then reopened the entire transient test issue for comment on
January 13, 1982, (47 FR 1642), as part of the NPRM for Revised
Gaseous Emission Regulations for 1984 and Later Model Year
Light-Duty Trucks ana tieav^-Diity Engines.** The last period
for submitting information and comment on -he transient test
procedure closed April 12, 1982.
See Appendix I, and Public Docket No. A-81-2G.
See Appendix II, and Public Docket No. a-81-11.
B-7
-------
2-3
III. Outstanding Issues
The major test procedure issues raised by the
manufacturers since the transient test's original promulgation
are easily grouped into five separate categories: 1) economic
impact, 2) the validity of EPA's test (in concept ana in
operation), 3) proposed alternative test cycles, 4) technical
amendments to the test procedure, and 5} issues, surrounding a
transient test for heavy-duty aiesel engines/optional 13-mode
standard.
A. Economic Impact (Chapter 3)
Almost all comments received in this area described the
capital expenditures which have been invested in transient
facilities, and the leadtime required to complete these
facilities. The comments were not in great detail or length,
and all were provided in response to the June 17, 1981, Federal
Register Notice. Aside from facility expenditures, no
manufacturer addressed incremental development costs (to design
engines to meet transient standards), or aggregate costs
associated with actually running the test. (Some manufacturers
did, however, identify specific technical amendments to the
test procedure which would yield cost savings. See below.).
B. The Validity of EPA's Test (Chapter 4}
In most cases, the industry reiterated its concerns with
the representativeness of EPA's test cycles, and problems
associated with actually running the test. Doubts about the
validity of EPA's cycle were directed mainly at the cycle
development methodology and the CAPE-21 data base. The
criticism was similar to that made over the last several
years. Kith respect to running a transient test, the industry
is concerned with potential repeatability and correlation
problems, and its inherent jeopardy if such problems arise
during confirmatory or SEA testing.
B. Alternative Test Cycles (Chapters 5 and fa)
The Engine Manufacturers' Association (EMA) and the Motor
Vehicle Manufacturers' Association (MVMA) both proposed that
EPA adopt their specific alternative test cycles, as an option
or in lieu of EPA's test cycles. The major issues associated
with these test cycles are their tech" acceptability, their
relative stringency, and how - if at axl - they should be
implemented.
D. Technical Amendments to the Test Procedure
The industry, in particular the EMA, has suggested
numerous technical amendments to the procedure to reduce cost
and complexity, and enhance technical aspects of the test.
B-8
-------
2-4
These suggestions range from recommended el imitation of the
cold start requirements to modification of more routine aspects
of the test (e.g. sampling system calibration procedures).
These recommended changes are perhaps as important in
reducing the economic impacts of the transient test as any
other possible action. They do not, however, materially affect
the overall implementation decisions associated with the
transient test. For this reason, they will not be specifically
addressed in this study, but in the separate Summary and
Analysis of Comments document to be published with the revised
HD 1S84 gaseous emission regulations. We intend to make as
many technical and test efficiency improvements as possible,
using the comments ar.d accumulated experience of both the
industry and EPA.
E. Diesel Transient Test/13-Mode Option (Chapter 7)
Some diesel manufacturers have questioned the need for a
transient test, especially for the small increments of HC
emission reductions required for diesel engines for 1984-86.
They have argueo that the 13-mode would be adequate to ensure
that mandated HC emission reductions would in fact take place.
Use of the transient test would, in their judgment, be an
expensive overkill of the problem. All diesel manufacturers
have invested in transient testing facilities, but all support
a more extensive and less stringent 13-mode option than the one
EPA has already provided.
IV. Issues Not Addressed
A few issues are not addressed in this study. The
technical justification for a transient test for gasoline
engines was not challenged by the manufacturers; EPA's original
analysis stands firm.* in addition, steady-state tests are
poor predictors of transient emissions for gasoline engines:
there is no means by which a steady-state test can adequately
represent actual gasoline engine operating characteristics and
be equivalently stringent to the transient. This removes any
possibility of providing a technically acceptable steady-state
option for gasoline engines. Finally, no diesel engine
manufacturers have pursued the use of eddy current dynamometers
for transient testing, and their acceptability need not be
addressed.
See EPA's earlier analysis of this issue in chapter 1 of
the Summary arid Analysis of Comments to the NPRM: "198 3
and Later Model Year Heavy-Duty Engines - Proposed
Emission Regulations,* December 1979.
B-9
-------
CHAPTER 3
ECONOMIC IMPACTS, LEADTIME, AND FACILITY STATUS
I. Overview
Manufacturers of heavy-duty gasoline ano diesel engines
have been concerned about the cost and leadtime for transient
test facility development since the transient test was first
proposed in February of 1979. Most manufacturers have
indicated that economic factors and leadtimes have not been
sufficient for engine development and certification on the
transient test for 1984. Regulatory relief is intended to
decrease the economic burden of implementing regulations on the
manufacturers. In this chapter the economic impacts, facility
status, and leadtimes associated with the manufacturer's
capability to perform transient engine tests and their effects
on regulatory relief efforts will be discussed.
II. Summary of Comments[1,2 ]
Comments have been received from nine manufacturers
regarding the status and the economic impact of developing
transient test capabilities. The facility requirements,
estimated costs, capital committed, and possible deferred
expenditures are summarized on a manufacturer-by-manufacturer
basis in Table 3-1. The manufacturers also provided
projections of available leadtimes, and their estimates of when
certification on the transient test will be feasible. These
projections are summarized in Table 3-2. Comments concerning
manufacturer's individual situations and recommendations are
summarized below.
A. Caterpillar
As of 11/1/81 Caterpillar had spent $10.1M of an estimated
$12.9M necessary for transient test facilities. Caterpillar
stated that 10 transient test cells would be required for
development and certification of 13 engine families for model
years 1984-85. Caterpillar is opposed to any delay in the
implementation of a properly structured transient test for
Federal certification in 1984 for two reasons:
1. Caterpillar has already invested significant
resources to develop transient test capabilities, and this
investment cannot be retrieved if the transient testing is
delayed; and,
2. The California Air Resources Board (CARB) has
adopted optional transient standards for 1984 which Caterpillar
feels are more reasonable than CARB's steady-state standards.
Caterpillar opposes any EPA action which may force CARB to
withdraw its transient standards.
B-10
-------
Table 3-1
Manufacturers' Economic Impact of Developing Transient Test Capabilities
Manufacturer
Test
Completed
Cells
Required
Estimated
Facility
Costs[1]
Already
CommitteUl11
Possible
Defer red
E x penditures 11J
Caterpillar
4
10
$12.9M
$10.1M
(78%)
$2. 8M
(22%)
Chrysler
0
2
$1.9M
$0.0 M
(0%)
$1.9M
(100%)
Cummins
2
12
$8 . 1M
$6.3M
(78%)
$1.8M
(22%)
Daimler-Benz
1
3
$5 .8M
$3.8M
(66%)
$2 .0M
(34%)
Ford
4
4
$6 .0M
$6. 0M
(100%)
-
(0%)
General Motors
(gas)
(diesel)
4
1
4
12
$3 .0M
$18 .0M
$3 .0M
$13.0M
(100%)
(72%)
$5 .0M
(0%)
(28%)
Hino
0
1
$2 .2M
$2 .2M
(100%
-
(0%)
International
(gas)
(0%)[2]
(diesel)
Harvester
0
1
4
2
$2 .0M[2 J
$5. 7M
$0 .0M
$2 . 1M
( 0 % } [ 2 ]
(37%)
$3 .6 M
$0 .014
(63%)
Mack
2
6
$5. 0M
$1.0M
(20%)
$4. 0M
(80%)
TOTAL COST
$6 8 .6M
$47 .5M
(69%)
$21.1M
(31%)
[1] 1981 dollars.
[2] IHC plans to leave the gasoline engine market before the transient test is
mandatory. IHC's estimated costs for gasoiine facilities therefore are not
included in the Total Cost or Possible Deferred Expenditures.
-------
Table 3-2
Model Year for Availability of Transient
Test Facilities for Certification of All Engine Families
Manufacturer
Caterpillar
Chrysler
Cummins
Daimler-Benz
Ford
General Motors
Hino
International
Harvester
Mack
Model Year - Gas
NA
1986*
NA
NA
1984
1984
NA
NA**
NA
Model Year - Diesel
1985
NA
1985
1984
NA
1985
1984
19S4
1984
* Dependent on Chrysler decision to acquire its own test
facilities.
** IH will abandon the gasoline engine market when transient
tests are required for certification.
B-12
-------
3-4
Although Caterpillar stated it was opposed to any delay in
the implementation of the transient test for 1984
certification, it also indicated that the leadtime for facility
development is not sufficient for certification of all of their
1984 engines on the transient test, and some families will
require carryover on the optional steady-state standards.
Caterpillar recommends the optional steady-state standard be
revised to 1.0 g/bhp-hr. This would allow more of
Caterpillar's engines to qualify for carryover and would result
in a more orderly phase-in of the transient test.
B. Chrysler
Chrysler has not invested any money to develop its
transient test capability and estimates it would require $1.9M
and two years to do so. Because of the small projected sales
volume of heavy-duty gasoline engines and its need to use all
available capital for passenger car development, Chrysler
decided to abandon this market segment when the transient cycle
regulation was promulgated. Unless an alternative to the
transient cycle is offered, Chrysler feels it must either buy-
certified engines from other manufacturers or pay an outside
engineering service to develop Chrysler engines to comply with
the transient test standards. Chrysler fears either of these
approaches would make the prices of its heavy-duty vehicles
uncompetitive.
The possibility also exists that Chrysler Corporation will
offer diesel engines for heavy-duty trucks for the 1984 model
year. These engines will be purchased and responsibility for
all emission testing will rest with the engine supplier.
Chrysler requests a steady-state test be allowed as an
option, at least for smaller manufacturers, for model years
1984 through 1987. This would allow Chrysler to continue
production of heavy-duty gasoline engines.
C. Cummins
Cummins has committed $6.3M of an estimated $8.1M
necessary for transient test facility development. Cummins
stated that it would have two transient test cells operational
by January 1982. Cummins also indicated that it would need 10
transient test cells in the beginning of 1983 and would need 12
transient test cells by the end of 1983. Their present
planning indicates that only 8 cells will be available at the
beginning of 1983, and 4 additional test cells, for a total of
12, will become available for use at the end of 1983. Cummins
said that the lack of test facilities would prevent it from
certifying all of its engine families to a 1.3 HC standard for
the 1984 model year. cummins recommends the option of
certification on either the steady-state test or the transient
test until the new NOx and particulate standards take effect.
After this, it recommends only the transient test be used.
B-13
-------
3-5
D. Daimler-Benz
Daimler-Benz has committed $3.84M and requires an
additional $1.8M-$2.0M for the development of three transient
test cells. Daimler-Benz facility development is behind
schedule, but it did not claim that it would be unable to
certify its four 1984 model year engine families on the
transient test. Daimler-Benz recommends allowing the 13-mode
test with a 1.1 g/BHP-hr HC standard as an option through the
1985 model year.
E. Ford
Ford has already committed all of its required $6.0M for
transient testing facilities (including four double-ended
dynamometers) for certification of three 1984 heavy-duty
gasoline engine families. Ford indicates that none of this
investment is recoverable if the transient test is delayed or
cancelled. However, Ford recommends delay of implementation of
the transient test because:
1. There is insufficient leadtime to conduct an orderly
development and certification of engines for 1984,
2. A delay would be beneficial in terms of gaining
additional experience with the new test cycle and engine
calibrations, and
3. A delay would enable a laboratory-to-laboratory
correlation program to be conducted.
E. General Motors
General Motors has spent $16.OM of an estimated $21.OM
toward the construction and development of transient test
facilities. GM stated that four gasoline engine transient test
cells have been completed and another is under construction.
These five cells will be adequate for certification testing of
all of its gasoline engine families for the 1984 model year if
the standards will not require development testing. However,
additional leadtime is necessary to comply with the proposed
standards. GM, also indicated the conversion of test cells
from 9-mode to transient test capability will not allow it to
certify on the 9-mode test in 1984.
One test cell for diesel engine testing is complete, but
GM requires twelve diesel test cells. GM will not have
sufficient facilities for transient test certification of
diesel engines until the 1985 model year. Also, due to the
test cell conversions, 13-mode facilities will not be adequate
to allow certification testing of all of its 1984 model year
diesel engine families. Therefore, GM is requesting that EPA
B-14
-------
3-6
allow carryover for the 1984 model year and allow the transient
test to be used as an option.
GM'-s recommendations for Loth gasoline and diesel engine
certification are:
1. Allow carryover of current certification procedures
and emission standards through 1984, but also allow the option
of the transient test for 1984. This will allow GM to certify
new engines,, first introduced in 1984, on the transient test.
Previously certified engines, which are carried over in 1984,
would be tested on the transient procedure in 1985. since the
new engines tested on the transient procedure in 1984 would
qualify for carryover in 1985, this plan would spread the
burden on GM's test facilities over a two-year period.
2. Allow the option of either the transient test or the
steady-state test with emission standards of equivalent
stringency as the transient test for 1985 and 1986. This
recommendation is entirely for manufacturers who may still be
unable to certify on the transient test for these years. GM
has explicitly stated that it does not require this action
itself.
3. Beginning in 1987, use only the transient test.
F. Hino
Hino is planning investment of $2.2M for transient
certification test facilities. Hino plans to certify two
engine families and only need one transient test cell if
implementation of SEA is delayed. Hino anticipated completing
this transient test cell at the end of 1983. Hino recommends
"adoption of the steady-state test as an optional method of
certification for 1984 and after."
G. International Harvester
International Harvester has invested $2.1M of an estimated
$5.7m required for heavy-duty diesel engine transient testing
facilities and has cancelled investment of S2.QM for gasoline
engine transient testing facilities
IHC indicated that the dramatic shift . in the market from
gasoline engines to diesel engines will force them out of the
gasoline engine market. IHC indicates that the major factor in
leaving the market is simply the market change, but the
implementation date of the transient test will influence when
this takes place. IHC has indicated that a single year
deferral (until 1985) of transient testing requirements will
allow it to withdraw from the market in a more orderly fashion.
B-15
-------
3-7
IHC recommends delaying the implementation of the
transient test for diesel engines until the time the new NOx
and particulate standards take effect. IHC's recommendation is
mainly based on a desire to develop control technology to meet
all four regulated pollutant levels at the same time. IHC did
not provide EPA with any information to indicate that 1984
model year diesel engine certification would be delayed due to
test facility problems.
H. Mack
Mack has invested $1.0M of an estimated $5.0M required to
develop transient testing capability. Mack has completed two
transient test cells out of a planned total of six. Because of
the depressed and unanticipated economic condition of the
industry, Mack has decided to defer a decision on the
development of the rest of its facilities until the end of
1982. Mack stated that the two transient test cells now
completed are sufficient to certify all of its 1984 engine
familes to a 1.3 HC standard.
Mack recommends delaying transient test requirements until
the new particulate standards become effective, and using the
13-moae steady-state test with emission standards of equivalent
stringency as the transient test for certification until that
time.
Ill. Analysis of Comments
Manufacturers of heavy-duty gasoline and diesel engines
have collectively committed $47.5M, or 69 percent of a total
$68.6M required for transient testing facilities. These costs
are higher than those projected earlier by EPA,[3,4] but we do
not challenge the manufacturers' estimates for money which has
already been spent. It is universally argued by the industry
that most of these sunk costs cannot be recovered if the
transient test is withdrawn or delayed. If the transient test
implementation is delayed, investment of $21.1M industry-wide,
or 31 percent of the total could be deferred for approximately
the same period of time as the delay. For the most part, the
outstanding transient test issues involve the technical and
timing aspects rather than the need for additional
expenditures. The leadtime for engine development is also an
important issue.
The transient test is, however, more of an economic issue
for two manufacturers which have been more severely impacted by
the recent economic conditions and which have experienced the
greatest difficulty in amortizing investments because of
smaller production volumes. These companies are Chrysler and
International Harvester. These two manufacturers claim that
they have delayed investment in transient test facilities out
of financial necessity, although no economic data were
B-16
-------
3-8
submitted to substantiate these difficulties. (This was true
for all manufacturers.) A delay in transient test
implementation would allow Chrysler to continue production of
gasoline engines. (1HC will phase out gasoline engines by
1985. After 1984, IHC will be unaffected by any test
procedural change).
Table 3-2 summarizes the manufacturers' comments regarding
the model year that sufficient transient test facilities would
be available for certification of all engine families. Except
for Chrysler's and international Harvester's gasoline engines,
all manufacturers will have transient test facilities to
certify some, if not all, of their engine families in 1984.
Chrysler's cancellation of investment in transient test
facilities and intention to abandon the gasoline engine market
when the transient test becomes mandatory have been business
decisions based on what it perceives to be low profitability of
this portion of its product line. Chrysler has the options of
certifying its engines on the light-duty truck procedure or of
contracting out the testing but has apparently decided against
these choices (see Chapter 8). Based on Chrysler's business
decisions, the only possible courses of action EPA could take
which would allow them to continue production of heavy-duty-
gasoline engines would be either a special provision for
Chrysler alone or a relaxation of the requirements for all of
the industry by eliminating the transient test. Neither option
is fair to the rest of the industry because all have made
substantial financial commitments. Chrysler cannot be
classified as a small manufacturer, therefore, a special
provision for Chrysler would result in a fundamental business
disadvantage for other manufacturers. Elimination of the
transient test for .whatever reason would result in nearly 50
million dollars industry-wide in sunk costs which could not be
recovered, and would yield no environmental benefits.
IV. Conclusions
1. Manufacturers of heavy-duty gasoline and diesel
engines have collectively committed $47.5M or 67 percent of a
total $70.6M required for the development of transient testing
facilities. Most of these sunk costs cannot be recovered if
the implementation of the transient test is cancelled or
delayed. The outstanding transient test issues are, therefore,
primarily technical or related to timing of implementation
rather than economic.
2. A delay in the implementation of the transient test
would allow the deferral of a collective $21.1M in investment
capital for approximately the same period of time as the delay.
3. Most gasoline and diesel manufacturers have
indicated that a steady-state test with equivalently stringent
B-17
-------
3-9
standards as the transient test, and/or carryover of emissions
data from 1983 certification, will be required as an option to
the transient test for 1984. Most diesel manufacturers
recommend the steady-state test be retained as an option until
the new NOX and particulate standards become effective.
4. Chrysler has indicated it may leave the gasoline
engine market at the time the transient test becomes
mandatory. There is no fair way to provide regulatory relief
for Chrysler alone.
B-18
-------
3-10
References
1. Derived from Comments submitted to EPA Public Docket
No. A-81-20.
2. Derived from Comments Submitted to EPA Public Docket
No. A-81-11.
3. "Summary and Analysis of Comments to the NPRM, 1983
and Later Model Year Heavy-Duty Engines Proposed Gaseous
Emission Regulations," December 1979.
4. "Regulatory Analysis and Environmental Impact of
Final Emission Regulations for 1984 and Later Model Year
Heavy-Duty Engines," U.S. EPA, OMSAPC, December 1979.
B-19
-------
CHAPTER 4
A REVIEW OF
EPA'S TRANSIENT TEST CYCLES
I. Introduction
Since their original proposal in February of 1979/ EPA's
transient test cycles have been the subject of a vigorous
technical controversy. The heavy-duty gasoline and diesel
engine manufacturers have continously criticized the cycles,
characterizing them as unrepresentative of real truck
operation, the victims of flawed data and incorrect development
methodologies. These arguments were reiterated in response to
the June 17, 1981, Federal Register questions, and in comments
received in response to the "Revised 1984 Heavy-Duty Engine
Emission Regulations."
A great deal has already been written on this subject.
The original Summary and Analysis of Comments of the 1984
FRM[1] argued strongly against most of the industry's technical
criticism. Later, continued industry criticism arose in A.D.
Little Inc.'s report to the MVMA[2] and Caterpillar Tractor
Co.'s Real Time Cycle Report. [3] This study will briefly
review recent criticisms and review the available emission data
to determine if measurable irregularities are indeed present in
the EPA test cycles.
II. Summary of Comments
A. Flawed Data Base
The raw data used in generating EPA's test cycles were
accumulated in the early 1970's in the CAPE-21 Project. Over
100 trucks and buses were instrumented, and a variety of
operational parameters (speed, RPM, load factors, etc) were
recorded.
The EMA characterized the CAPE-21 project as "plagued by
equipment malfunctions and overwhelmed with false data."[5]
EPA in fact classified approximately one-third of the total
data base as erroneous, and removed it. EMA, however, argued
that all data taken on the same instrumentation on the same day
should have been discarded, not merely that which was clearly
incorrect. Using this and other criteria,* Caterpillar[3]
discarded nearly three-fourths of the CAPE-21 data. Following
this logic, a large percentage of the upon which the EPA
cycle is based was incorrect. EPA's cycle is thereby
questionable, if not completely unrepresentative, in the
judgment of the EMA.
* See Chapter 5 of this study, "The Real Time Cycle: An
Alternative Heavy-Duty Diesel Engine Driving Cycle."
B-20
-------
4-2
Another criticism of the CAPE-21 data base, expressed most
recently by A.D. Little,[2]" was that the data base sample size
(i.e. number of trucks) was too small. Too small a sample size
limits the accuracy with which the "average" truck can be
characterized, ana, hence, the representativeness of the test
cycles.
B. Inappropriate Generation Methodology
EPA used the Monte Carlo statistical technique to
construct candidate cycles from the second-by-second CAPE-21
data. In practice, one second of operational data (engine
speed, engine power) was randomly selected by a computer; the
next second was probabilistically selected, based upon the
observed frequency of occurrence of that, transition in the
CAPE-21 data base. This probabilistic sampling continued until
a candidate cycle was created. These candidate cyles (of which
there were thousands) were then compared to and screened
against statistical summaries of CAPE-21 parameters. The
cycles closest to the CAPE-21 data base were then selected to
make up the final test cycle.
The most exhaustive criticism of this methodology was made
by A.D. Little[2], although over time almost every manufacturer
criticized EPA on this matter. Among A.D. Little's conclusions
were the following:
1. The methodology failed to consider all important
operational parameters (e.g., engine temperature, time
sequencing, and history of operation).
2. The Monte Carlo methodology cannot adequately
represent actual operation, since it neglected sequential
engine operation and produced erratic second-by-second test
cycles - not the characteristic speed/load shift patterns seen
in actual trucks.
3. The cycle segments were too short, requiring
unrepresentative manipulations to give the Monte Carlo
technique the appearance of sampling the entire data base.
4. The statistical filter used to screen the candidate
cycles was incomplete and inconsistently applied, permitting
unrepresentative parameters to enter the test cycle.
In short, A.D. Little concluded that the EPA cycles were
unrepresentative of both the data base and re§l world trucks.
The industry has used A.D. Little's analysis to substantiate
its earlier criticisms, and to argue for adoption of its own
engine test cycles. (See Chapters 5 and 6).
C. Operational Problems
The industry has also argued that the EPA cycles' erratic
construction will lead to repeatability and correlation
B-21
-------
4-3
problems when used in actual emission testing. The gasoline
engine manufacturers cite a study by EPA staff[6] which showed
that gasoline engine emissions, as measured over EPA's test
cycle, are sensitive to the dynamometer control system
calibration {and perhaps the control system itself). In
addition, the study indicated that other test validation
criteria may need to be tightened to minimize the chances of
repeatability and correlation problems. The industry has used
this study to argue that a test cycle which is too transient
(high frequency of changing speed and torque commands) may
yield unrepresentatively high emissions results, and when
combined with the variety of potential control systems and
control system calibrations, may yield unrepeatable emissions
from lab-to-lab. Some manufacturers have also argued that the
EPA cycle is too transient to provide repeatable results from
test to test, especially on cold starts.
The diesel engine industry also raised similar concerns
about the repeatability and lab-to-lab correlation potential of
the EPA cycle. Based upon these concerns, the EPA/EMA Round
Robin correlation program was initiated, in which several
diesel engines were circulated among the manufacturers and
EPA. The gasoline manufacturers, through MVMA, have also
requested that a similar (though smaller scale) project be
initiated with EPA for their engines.
II. Analysis of Comments
A. Test Cyle Representativeness (overview)
To begin this discussion, it is best to review how the
representativeness of an emissions test cycle can be evaluated.
To date, both EPA and the industry have argued questions
of representativeness on the basis of statistical comparisons:
if the test cycle is operationally and statistically
representative of the "average" on-road truck, then the
emissions measured over that test cycle will be representative
of on-road emissions. This is a logical, but hypothetical
argument; its utility in discriminating between test cycles is
valid only up to a point. Its fundamental weakness is that
statistical differences represent unquantified, perhaps
undetectable, emission differences. Secondly, the correlation
between test cycles and the real world, i.e., the predictive
utility of the test cycles in insu. > ; real world emission
reductions, may not be as sensitive to minor statistical and
operational variations as some have claimed. Comparative
emission data represent the only means for resolving how
statistically and operationally representative a test cycle has
to be.
The preferred method is to review laboratory emission
data. In as fine a breakdown as possible, compare the emission
effects of changes to test cycle parameters. These results can
B-22
-------
4-4
be used to extrapolate the emissions effect of statistical and
operational differences between cycles, or a particular t^pe of
operation , in a given cycle. This has already been done in
several cases by EPA,[4,6] and will likely be continued during
analysis of the data generated from the EPA/EMA Round-Robin
Test Program[7,8] and the suggested EPA/MVMA correlation
project. EMA's and MVMA's alternate test cycles provide a
further basis for comparison, i.e., between different cycles
developed in different ways as a means of evaluating the cycle
generation methodologies. From this analysis, one can judge if
ahi given test cycle yields grossly different and, hence,
unrepresentative emissions.
B. The Validity of Specific Criticisms
The diesel engine industry concentrated it's criticism on
the CAPE-21 data base. EPA had argued earlier[1] that the data
collection and editing were performed with sufficient care, to
preclude the use of void data for test cycle development. in
generating their own Real Time Cycle, engineers at Caterpillar
were stricter in their criteria for editing data (see Chapter
5). Rather than debate the merits of either approach, however,
we need only review the similarity of outcomes, as discussed
in detail in Chapter 5, the EPA and EMA (Real Time) cycles are
statistically similar, both being close to their respective
CAPE-21 data bases (which themselves are quite similar). Both
cycles yield NOx emissions which are comparable; particulate
and HC emissions are somewhat less on the EKA cycle, but to a
much lesser degree than differences observed between subcycles
(NYNF vs. LAF), i.e., between cycles of significantly different
operational characteristics. Furthermore, emissions measured
over both cycles correlate extremely well with each other. In
other words, if applied technology affects a change in
emissions on one cycle, a proportional change will be
measurable on the other cycle (and presumably in-use if both
cycles are representative). If applicable emission baselines
and standards are adjusted to reflect the differences in
absolute emission levels between cycles, there remain no
practical differences between the test cycles. The editing of
the CAPE-21 data base becomes irrelevant, as do other claims of
EPA cycle unrepresentativeness, from a practical testing
standpoint.
EPA's analysis of the A.D. Little report[4] addressed the
issue of truck sample size. EPA concluded that A.D. Little
held EPA to unreasonable and impractical standards of
accuracy. For example, A.D. Little concluded that over 1000
trucks were necessary to adequately characterize certain truck
parameters. At more reasonable accuracy criteria, the truck
sample size was large enough to confidently establish 41 of 42
parameters. Furthermore, CAPE-21 was the largest on-road truck
study attempted up to that time. Such criticism neglects real
world practicalities and the diminishing returns of accuracy
with continued testing.* Despite the fact that they rejected
B-23
-------
4-5
75 percent of the CAPE-21 data (as opposed to EPA's 33
percent), Caterpillar's technical personnel concluded that a
sufficiently large data base remained to generate a
representative test cycle - the Real Time Cycle.[3]
The gasoline engine industry's criticism focused mainly on
the manipulation of the CAPE-21 data base, and the generation
of EPA's cycles. Much has also been written about the cycle
generation methodology, in particular, the Monte Carlo
technique. A.D. Little's report to the -MVMA (2) is the best
example of the industry's position. To address this criticism,
we repeat the conclusions of EPA's analysis of the A.D. Little
report (4):
1. "A.D. Little's criticisms of EPA's cycle development
methodology are based upon statistics and
conjecture. No emission data is presented or
referenced to substantiate any claims of engine or
chassis cycle unrepresentativeness.n
2. "A.D. Little's statements concerning combustion
engine emissions are based upon a brief literature
review, and not upon actual "hands-on" experience
with emission testing or emissions control. A.D.
Little therefore attributes many engine
parameters...with larger than life emission
significance, whereas only a few...will in actuality-
dominate the results of laboratory emissions test."
3. "The majority of A.D. Little's analysis of test
cycle validation is incorrect, being based upon
incorrectly-applied statistical tests and
incorrectly-chosen comparative parameters."
4. "A.D. Little's harsh criticism of the weakness in
EPA's methodology neglects to mention that the
weakness was not imposed by EPA error, but rather by
fundamental limitations which the input data placed
upon the applicable statistical techniques, and by
the compromises required to compress the vast data
base into a workable test procedure." (Note that in
their cover letter to the MVMA, A.D. Little stated
that they knew of no entirely correct statistical
way that test cycles could be derived from the
CAPE-21 data.)
5. "A.D. Little's arguments were not only narrow in
focus (neglecting practicalities and ignoring
emissions), they were entirely negative. No single
alternative to EPA's methodology was advanced. No
* The complete in-depth analysis of sample-size, and EPA's
comprehensive review of the A.D. Little Report can be found in
[4].
B-24
-------
4-6
evidence was presented that EPA neglected a more
representative approach in lieu of the methodology
chosen. No specific suggestions were made about how
the cycles could be improved."
6. "ECTD sees nothing in the A.D. Little Report to
warrant changing the conclusion that the engine
cycles are sufficiently representative of the data
base. The implications from all emission data
collected to date on the engine cycles are that
emissions are not exaggerated by the Monte Carlo
technique. Changes to the engine cycles to improve
the agreement of their operational parameter summary
percentages with those of the CAPE-21 data cause
negligible or minor (less than 5 percent) changes in
measured emissions. This is true even for the most
operationally-sensitive modes and emissions
evaluated...As substantiated by spa's engine cycle
emission data, A.D. Little overestimates the
emissions sensitivity of the test cycles to the
cycle generation methodology. . .it was a highly-
conjectural overestimation, being based upon
absolutely no emission data whatsoever."
Emission data is available to support these conclusions.
To evaluate the effect of the cycle generation methodology, EPA
reviewed data on transient throttle response,[6] and transient
test emissions vs. steady-state emissions for uncontrolled
gasoline engines.[4] The data[6] indicated that below the
range of dynamic throttle activity normally used to control a
gasoline engine during a transient cycle, emissions were
unaffected, i.e., decreasing transient operation did not affect
emissions. This implies that the EPA cycle is not excessively
transient, since a dampened throttle controller mechanically
smoothed the cycle and emissions remained unaffected.
Furthermore, average emission levels of fifteen 1969 gasoline
engines on both the EPA transient test and the old 9-moae
steady-state procedures were compared. If unrepresentative
transient operation were present, one would expect uncontrolled
transient emissions to significantly exceed those measured on
the steady-state tests. in fact, the emission levels of both
tests were not dramatically different;* EPA transient HC
emissions were actually 9 percent less.
In the same report,[4] EPA evaluated statistical
differences between cycles. Small differences in operational
percentages (which A.D. Little argued rendered the EPA cycle
unrepresentative) produced minimal emission changes (less than
5 percent) for the most operationally-sensitive modes in the
* Note that this comparability between steady-state and
transient test procedures has only held for uncontrolled
engines, i.e., engines to which no technology has been applied
for emission control.
B-25
-------
4-7
test cycle. in short, emissions measured over the entire test
cycle are not sensitive to small changes in summary
statistics. Finally, the close, correlatable relationship
between the emissions measured on EPA cycles and the MVMA cycle
(a smoothed, minimally-restructured version of the EPA cycle),
and the EMA cycle (a cycle developed using an entirely
different methodology), leads one to the conclusion that the
EPA cycles are technically sound. We are unaware of any
contradictory emission data.
C. Operational problems
Correlation and repeatability are important facets of
emission testing. This is especially so when a manufacturer's
certificate of conformity depends upon the outcome of
confirmatory testing. For light-duty vehicles, correlation ana
repeatability concerns led to the standardization of test
equipment (e.g. chassis dynamometers). Calibration and testing
protocols have also been standardized over time. All represent
the accumulated work and experience of technical personnel
throughout the industry and EPA.
The old heavy-duty test procedures are no different, nor
is the new transient test. Resolution of repeatability and
correlation problems is an inevitable task when a new test
procedure is adopted. This is especially true when new
equipment for engine control and emission sampling and
measurement are required. EPA anticipates that a number of
technical amendments to the transient test will need to be
made; many have indeed already been maae.[l] Aside from the
repeatability and correlation aspects, changes recommended by
industry will also be made to reduce complexity and cost. (All
of these forthcoming technical changes will not be discussed
here; they will be addressed in the Summary and Analysis of
Comments to the "Revised 1984 Heavy-Duty Engine Emission
Regulations.")
The industry has speculated that EPA's transient test will
be more conducive to repeatability and correlation
difficulties. A report by EPA staff[6] indicated that there is
potential for gasoline engine emission variability within the
range of the transient test validation criteria. This by no
means guarantees that repeatability is a problem. Other test
procedures also have validation criteria, usually specified as
ranges of operation which the engine vehicle may not exceed
during the test.* Good engineering practice, however, requires
testing not at the full range of possible calibrations, but at
that closest to nominal to minimize repeatability problems.
* The 9-mode steady-state test is run at 2000 +100 RPM, the
13-mode at the specified speed +50 RPM; torque values for both
tests are run at +2 percent of the maximum engine torque. The
light-duty test procedure requires a driver to remain within a
specified mph band about the driving cycle.
B-26
-------
4-8
The same is true when running a transient test cycle - any
transient test cycle - on an engine dynamometer. However, a
technical problem arises when using an electro-mechanical
control system to maneuver a high inertia engine/dynamometer
combination through a continuously varying cycle. The control
system 'design may in fact determine what the optimal validation
statistics are.* This is a hardware/control system problem -
not a cycle problem, i.e., the system will respond similarly
for any reasonably similar cycles. (For example, there are no
substantive differences between cycle validation statistics for
engines tested by EPA on both the EPA and MVMA cycles.) There
is no evidence that EPA's cycle is any mote susceptible to
control system calibration than any other transient cycle.
* For example, EPA uses a proportional-feedback control
system.[9,10] The difference in electrical signal between the
command voltage and the feedback voltage "drives" the throttle
controller to correct the discrepancy, i.e., until the
differential voltage is zero. in this way the engine is made
to follow the cycle as the feedback signal continually "chases"
the command signal. However, the inertia of the
electro-mechanical system (including the engine) is high; if
too strong an initial "push" (too high an electrical gain) is
given to effect a torque change, the control system will
overshoot the desired torque. An even higher gain will cause
oscillatory motion (an "underdampea" response) about the
desired torque, pumping the accelerator pump in an
unrepresentative manner and drastically increasing
emissions!6]. This dynamic characteristic is entirely a
function of the analog control system's design. Another type
of analog control system (e.g., proportional plus derivative
feedback - PDF) or a real-time microprocessor system will yield
entirely different transient characteristics.
Transient cycle validation criteria are least squares
regression statistics; mathematically perfect regression
statistics are 1.000 (regression line slope) and 1.000
(correlation coefficient). EPA's prototype control system, in
order to respond quickly enough to approach the mathematically
perfect regression statistics, must be given so much gain that
oscillatory, dynamic instability occurs. EPATs control system
is quite capable of following a transient cycle very closely
without resonant instability . This does not occur, however,
at the mathematically perfect regression values, but at values
somewhat less. A different type control system will no doubt
perform differently, and have its own optimum validation
statistics based upon its own characteristic equations of
motion.
B-27
-------
4-9
The report by EPA staff[6] recommends further testing. in
this way, the speculation of the past gives way to reasoned
engineering analyses of technical issues such as these. A
logical result of this testing may be standardized control
system design, and perhaps revised validation criteria.
Assessments of emission sampling equipment would also be a
natural extension of further testing, especially for the
continuous sampling systems required to test diesel engines.
Such testing is already taking place in the EPA/EKA Round
Robin Correlation Program. Preliminary results[7,8] indicate
reasonable correlation between laboratories, and repeatable
measurements within labs. Widespread fears of dramatic
correlation problems have not been realised. Steady-state
13-mode particulate results were most variable (with a
coefficient of variation of +16.7 percent relative to +8.7
percent for the EPA transient test); this is primarily due to
the lack of a standardized steady-state particulate test, which
itself indicates how important standardization is in achieving
repeatability between labs. Transient HC data were more
variable than 13-mode HC (+12.6 percent vs. +8.3 percent), but
not to an extent indicative of fatal problems, ana likely
attributable to differences in emission sampling equipment (to
which continuously sampled diluted exhaust HC is sensitive).
The initial results of this test project are encouraging, and
give no indication to any reasonable reviewer that fundamental,
unresolvable problems exist. (A more comprehensive review of
the results can be found in [7,8].)
Repeatability of emissions with baseline gasoline engines
on the transient procedure at EPA's laboratory have been
reported earlier.[9,10] Average coefficients of variation for
all engines tested were +5.9 percent (HC), +5.8 percent (CO),
and +6.3 percent (NOx). These are reasonable but not
exceptional, and the degree of repeatability was engine
dependent. Again, no fundamental problems were observed. An
in-depth correlation effort between EPA and the gasoline engine
manufacturers has yet to begin, although MVMA has requested
such an effort. Since gasoline engines may indeed be most
sensitive to control system parameters, it is important for the
iaentification of test procedure refinements for EPA to follow
through with such a program. (Emission sampling for heavy-duty
gasoline engines is relatively straightforward, being a direct
extension of light-duty sampling technology; only minor
problems, if any, are anticipated in t!~' ~ Cacet of correlation.)
Finally, with mandatory certification on the transient
test procedure likely beginning for 1985, over two years remain
for procedure refinement. We are aware of no fundamental
problems with the test (either cycle related or operationally
related) which would preclude or potentially invalidate ongoing
development testing. (The gasoline engine manufacturers are
well along in characterizing their product lines on the
B-28
-------
4-10
transient test, and have begun initial development work; the
diesel engine industry is routinely running transient tests,
and their engines need far less development work to meet
applicable standards.) We also note that the industry's
transient facilities are all new and incorporate the latest
technology, whereas EPA still uses its prototype facility.
Should test equipment standardization be necessary to assure
correlation, upgrading EPA's test sites to match the
manufacturers' would impact the industry least, would ease
industry concerns about correlation problems at the time of
confirmatory testing, and would give EPA a better, more
productive test site. At any rate, we see no time constraint
preventing procedural refinements, should they be found
necessary.
IV. conclusions
1. No test data have yet been advanced to find EPA's
transient cycles to be unrepresentative. All emission data to
date continue to confirm their acceptability. In fact, the EPA
cycles correlate well with the EMA ancl MVMA alternative cycles,
from which the industry removed all operation with which it
objected.
2. No fundamental problems of repeatability and
inter-lab correlation have been observed for the EPA test. We
anticipate a succession of technical amendments to the
procedure as experience is gained throughout the industry, and
in response to joint correlation projects. All work necessary
to refine the test procedure should proceed as required. A
joint correlation program between EPA and the gasoline engine
manufacturers should be initiated as soon as practical.
B-29
-------
4-11
References
1. "Summary and Analysis of Comments to the KPRK: 198 3
and Later Model Year Heavy-Duty Engines - Proposed Emission
Regulations," December 1979.
2. "A Review of the Heavj-Duty Gasoline Engine
Certification Test Cycles," Final Report to the KVMA, Arthur D.
Little, Inc., Cambridge, April 1981.
3. "Evaluation of the Federal Test Procedure for
Heavy-Duty Diesel Engines for 1984 and the Development of the
Real Time Test Cycle," w.L. Brown, Caterpillar Tractor Co.,
Research Report 88-29, File 18967, June 22, 1981.
4. "Analysis of Arthur D. Little inc.'s Report to the
MVMA: A Review of the Heavy-Duty Gasoline Engine Certification
Test Cycles,'" ECTD Staff Report, January 1982.
5. Derived from comments submitted to EPA Public Docket
No. A-81-11.
6. "Heavy-Duty Gasoline Engine Emission Sensitivity to
Variations in the 1984 Federal Test Cycle," T. Cox, SAE Paper
No. 801370, October 1980.
7. "Status of EPA/EMA Cooperative Test program," A.
Azary, EPA status Report, March 1982.
B. "Preliminary Report on Statistical Analysis of
Heavy-Duty Diesel Engine Emissions Data," N.J. Barsic, prepared
for the Engine Manufacturers Association/Technology and Methods
Subcommittee.
9. "1969 Heavy-Duty Engine Baseline Program ana 1983
Emission standards Development," T. Cox, et al., EPA Technical
Report No. SDSB-79-23, May 1979.
10. "1972-73 Heavy-Duty Engine Baseline Program and NCtx
Emission Standard Development," T. Cox, et al., EPA Technical
Report Mo. AA-SDSB-81-01, March 1981.
B-3D
-------
CHAPTER 5
THE REAL TIME CYCLE: AN ALTERNATIVE HEAVY-DUTY
DIESEL ENGINE DRIVING CYCLE
An alternative heavy-duty diesel driving cycle, the Real
Time Cycle (RTC), has been developed by the Caterpillar Tractor
Company. It was developed in response to industry-wide concern
over the methodology used to generate the EPA driving cycle and
the resulting representativeness of EPA's cycle. The Engine
Manufacturers Association has proposed the RTC be allowed as an
option to the EPA driving cycle for the certification of
heavy-duty diesel engines. The specific recommendations of the
manufacturers, and factors under consideration by EPA
concerning the RTC as a test option, will be discussed in this
chapter.
A. Summary of Comments[1,2]
Most manufacturers and the EMA have made specific
recommendations concerning the use of the Real Time Cycle for
certification testing. These comments are summarized below.
Engine Manufacturers Association - EMA recommends EPA
allow manufacturers to certify engines on either the Federal
Test Procedure or on a modified cycle such as the RTC.
Caterpillar - Caterpillar recommends EPA provide the
option of using either the RTC or the EPA cycle for
certification of heavy-duty diesel engines.
Cummins .- Cummins has no reason to support or object to
the RTC as an option as long as EPA and industry work toward
the eventual use on only one test cycle.
Daimler-Benz - DB's April 12 submission of comments
indicated that it had not had time to comment, but will do so
later.
General Motors - GM recommends EPA initiate a cooperative
test program with industry to improve the Federal Test
Procedure (EPA driving cycle) , However GM is not opposed to
the use of the RTC as an option.
International Harvester - IH believes the RTC should be
used instead of the EPA cycle, however IH does not oppose the
RTC as an option.
Mack - Mack would consider the use of the RTC only after a
thorough study of the cycle, including round robin testing at
manufacturers' and EPA laboratories.
B-31
-------
5-2
B. Analysis of Comments
In this analysis, we address the technological basis,
representativeness, and relative stringency of the Real Time
Cycle.
1. Cycle Development[3]
Caterpillar developed the RTC because of concern over the
accuracy of simulation of in-use truck operation represented by
the EPA cycles. This concern stemmed from alleged
instrumentation problems in the CAPE-21 project which resulted
in a significant amount of questionable data being accepted
into the data base, and from the methodology EPA used to
generate the cycle.* Caterpillar's objectives in developing
the RTC were to generate a cycle from the portion of the
CAPE-21 data base which it considered valid, and to construct
the cycle so it better represented its judgment of real-world
truck operation.
Caterpillar constructed the cycle in the following
manner: First, the entire data base was edited to remove the
questionable data using the following criteria for acceptance
or rejection:
a. Entire truck-days of operation were either accepted
or rejected. The reasoning is that repairs were probably, not
made on instrumentation during the same days in which
malfunctions occurred.
b. Truck-days which exhibited more than 10 percent
spurious data for both engine rpm and power were rejected.
c. A positive correlation between engine torque and
change in rpm for at least 50 percent of the time of operation
was required for acceptance.
d. Discrepancies between time at idle rpm and idle
power could not occur more than 15 percent of the total
operating time for acceptance.
e. Truck days which had over 10 percent of the events
at less than negative 30 percent power were rejected.
This editing left 23 truck-days of data, or about 25 percent of
the original data base.
See Chapter 4 of this study.
B-32
-------
5-3
Statistical parameters were then chosen to characterize
the edited data base. These were mean values and cumulative
distributions of %rpm, %power, and positive %rpm. The % idle
time and distribution in length of idle were also used. These
statistical parameters, which characterized the edited data
base, then became the target values for the construction of the
new test cycle.
To construct the new test cycle, the data were broken down
into the smallest elements which did not interrupt the normal
driving sequence. These elements were defined as the vehicle
operational events which occurred between vehicle stops. The
elements were then assembled into trial test segments which
matched, as closely as possible, the desired statistics of the
categories they represented. The categories were: New York
Freeway (NYF), New York Non-Freeway (NYNF), Los Angeles Freeway
(LAF) , and Los Angeles Non-Freeway (LANF) . The idle time and
category weighting were adjusted to match the original CAPE-21
data base, since it was unlikely that instrument error would
change these parameters. The trial test segments were then
tested against the data base for maximum deviation of
cumulative distributions and then were compared visually. The
best cycles were selected and assembled into an entire driving
cycle.
The result was a heavy-truck engine driving cycle, the
"Real Time Cycle," which matched very closely the statistics of
the edited CAPE-21 data base, and which the diesel engine
manufacturers determined was more representative of in-use
truck operation than the EPA cycle.
2. Statistical Analysis
A comparison of the target statistics from the edited data
base, the RTC statistics, and the EPA cycle statistics is shown
in Table 5-1. Additional statistics from the RTC, EPA cycle,
and the original CAPE-21 data base are listed in Table 5-2.
The most important statistical differences between the RTC and
the EPA cycle are:
a. The RTC includes a NY Freeway segment; the EPA cycle
does not include the NY Freeway segment.* (The NY Freeway
segment is higher in mean %rpm and higher in mean %power than
the NY non-Freeway segment.)
EPA omitted the NY Freeway segment because invalid data
had been included in the data base and the weighting for
this segment was small compared to the other segments.[4]
B-33
-------
5-4
Table 5-1
Target/ RTC, and EPA Cycle Statistics [3]
Average % RPM
Average % Power
Average Positive
Acceleration
% Idle Time
Category
Weighting
Average % RPM
Average % Power
Average Positive
Acceleration
% Idle Time
Category
Weighting
Los Angeles Non-Freeway
Target Real Time EPA
Los Angeles Freeway
Target Real Time EPA
40.7
24.1
4.6
35.0
23.7
41.8
25.9
5.7
32.7
27.3
43
26
6.1
34
25.0
80.0
58.9
1.9
2.0
26.3
83.5
56.4
1.2
1-4
25.1
83
56
2.4
2.3
25.0
New York Non-Freeway New York Freeway
Target Real Time EPA Target Real Time EPA
41.5
41.0
2.8
19
9.0
Average % RPM
Average % Power
Average Positive
Acceleration
47.1
54.4
4.6
21
5.9
17.7
19.4
3.8
51. 0
41.0
Overall
19. B
22.3
3.6
51. 0
41.7
20
16
5.6
55
50.0
Target
41.7
32.8
3.9
Real Time
4 3.4
53.7
4.2
EPA
41.5
28.5
4.6
% Idle Time
31. 4
31. 8
36.6
B-34
-------
Table 5-2
RTC, EPA Cycle, CAPE-21 Data Base Statistics
Parameter
RTC
EPA
CAPE-21
Torque
Mean {%)
30.57
28.32
27.00
Percent of
Cycle Time
Accel. (%)
Decel. (%)
Cruise (%)
Motor (%}
Idle {%)
18.21
18.37
22.48
7.98
32.96
15.68
16.85
20.43
11.43
35.61
15.10
15.25
18.75
15.00
35.00
RPM
Mean {%)
42.78
41.52
41.75
Percent of
Cycle Time
Acceleration
Deceleration
Cruise (%)
Idle (%)
(%)
{%)
23.45
22.48
19.74
34.33
21.77
21.93
16.10
40.20
21.50
19.50
19.50
39.00
B-35
-------
5-6
b. The RTC is 5.2 percent higher in mean %power,
overall than the EPA cycle.
c. The RTC is 4.8 percent lower in %idle time, overall
than the EPA cycle.
d. The sequential ordering of the cycle segments on the
RTC is LANF, LAF, NYF, and NYNF. The ordering on the EPA cycle
is NYNF, LANF, LAF, NYNF.
The statistical differences cited here may or may not
affect engine emission levels. A potentially significant
factor is that the observed engine work done over the test
cycle (BHP-hr) is higher by 16-18 percent on the RTC.
Furthermore, the reordering of the segments in the RTC permits
the engine to operate in the high power LAF mode earlier in the
cycle, which may lead to an earlier engine warm-up (although
operation in the LAF segment of the RTC is initially cooler
than on the EPA cycle) . Inclusion of the NYF segment is one
obvious reason why the RTC BHP-hr is higher, possibly affecting
emissions.
3. Emissions Analysis
Heavy-duty diesel engine manufacturers and EPA have tested
several engines on both the Real Time Cycle and the EPA cycle
for the purpose of comparing emissions results. All of the
available data has been collected and is summarized in Table
5-3. Results are also plotted in Figures 5-1, 5-2, and 5-3.
Since typical diesel CO emissions are much lower than statutory
levels, this pollutant comparison was not included.
Data from 30 engines/configurations showed excellent
statistical correlation between the two test cycles for all
three pollutants; r^ values were .94, .98, and .90 for HC,
NOx and particulates, respectively. However, definite
emissions differences between the two test cycles were
observed. There was an average RTC HC emissions offset* of
-16.0 percent from the EPA cycle with a range of -35.4 percent
to +9.4 percent and a coefficient of variation of .62. The
average NOx emissions offset was +1.0 percent from the EPA
cycle with a range of -12.6 percent to +17.1 percent and a
coefficient of variation of .72. The average particulate
emissions offset on the RTC was -9.4 percent from the EPA cycle
with a range of -25.7 percent to +18.9 percent and a
coefficient of variation of 1.17.
% Offset = RTCgpaEPA x 100%
EPA
B-36
-------
Table 5-3
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Engine
Tested
By
Cycle
Number and
Type of Tests
HC
(q/BHP-hr >
NOx
(g/BHP-hr)
Particul,
(g/BHP-h
Mack 676
EPA
RTC
4HS
.55
8.56
.777
Mack 676
EPA
EPA
4HS
.65
8.40
.752
% Difference*
-15.4%
1.9%
3.3%
Cummins VTB903
EPA
RTC
4HS
1.27
5.01
.504
Cummins VTB903
EPA
EPA
4HS
1.60
5.07
.544
% Difference
-22.6%
-1.2%
-6.7%
IHC DT-210
IHC[a]
RTC
1-3CS, 2-7HS
.78
6.80
_
IHC DT-210
IHC
EPA
1-3CS, 2-7HS
.89
7.20
-
% Difference
-12.4%
-5.6%
-
IHC DTI-210
IHC
RTC
1-3CS, 2-7HS
.85
4.16
-
IHC DTI-210
IHC
EPA
1-3CS, 2-7HS
1.07
4 .15
-
% Dif feren- e
-20 .6%
-0.2%
-
IHC DT1-180
IHC
RTC
1-3CS, 2-7HS
1.06
4.73
_
IHC DTI-180
IHC
EPA
1-3CS, 2-7HS
1.19
4.94
-
% Difference
-10.9%
-4.3%
-
IHC 9.0L
IHC
RTC
1-3CS, 2-7HS
1.90
7.52
—
IHC 9.0L
IHC
EPA
1-3CS, 2-7HS
2.03
7.18
-
% Difference
-6.4%
-4.7%
„ - • -- RTC - EPA
* % Difference =
[a] Particulate data not included.
-------
Table 5-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Tested Number and HC NOx Particulate
Engine By Cycle Type of Tests (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
Cummins #l[b]
Cummins
RTC
2HS
.48
7.4 6
.43
Cummins #1
Cummins
EPA
2HS
.55
7.50
.46
% Difference
-12.7%
- .53%
-6.5%
Cummins #2
Cummins
RTC
2HS
.91
7.92
.66
Cummins #2
Cummins
EPA
2HS
1.19
8.10
.66
% Difference
-23.5%
-2.2%
0.0%
Cummins #3
Cummins
RTC
2HS
.63
7.29
.56
Cummins #3
Cummins
EPA
2HS
.87
7 .37
.70
% Difference
-27.6%
-1.18
-20.0%
Cummins #4
Cummins
RTC
2HS
.67
5.42
.94
Cummins #4
Cummins
EPA
2HS
.94
4.63
.94
% Difference
-28.7%
17.1%
0.0%
Cat. 3208
IHC
RTC
4CS
, 6HS
.84
8.57
.600
Cat. 3208
ICH
EPA
5CS
, 12HS
1.30
7.68
.704
% Difference
-35.4%
11.6%
-14.8%
Mack # 1;Ic J
Mack
RTC
.41
5.9
.46
Mack #1
Mack
EPA
.46
5.6
.51
% Difference
-10.9%
5.4%
-9.8%
Tb] Engine models not specified.
[c] Emissions data derived from plots. Engine models and number and type of
tests not specified.
-------
Table 5-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Enqine
Tested
By
Cycle
Number
Type of
and
Tests
HC
(q/BHP-hr)
NOx
(q/BHP-hr)
Particu
(q/BHP-:
Mack #2
Mack
RTC
.46
8.4
.69
Mack #2
Mack
EPA
.55
7.8
.79
% Difference
-16.4%
7.7%
-12.7%
Mack #3
Mack
RTC
_
.87
9.0
.69
Mack #3
Mack
EPA
1.10
10.3
.85
% Difference
-20.9%
-12.6%
-18.8%
CAT 3208
IHC
RTC
4CS,
6HS
.85
8.59
.60
CAT 3208
IHC
EPA
5CS,
11HS
1.31
7.59
.70
% Difference
-35.1%
13.2%
-14.3%
Mack ETSZ-676
Cat
RTC
ICS,
7HS
.65
7.62
.63
Mack ETSZ-676
Cat
EPA
2CS,
12HS
.73
6.82
.53
% Difference
-11.0%
11.7%
18.9%
IHC DTI-4CjB
Cat
RTC
ICS,
6HS
.90
4.30
.70
IHC DTI-4' 5B
Cat
EPA
2CS,
15HS
1.00
4.44
.69
% Difference
-10.0%
-3.2%
1.5%
Cat 3208
Cat
RTC
ICS,
7HS
1.18
8.79
.88
Cat 3208
Cat
EPA
5CS,
11HS
1.08
8.40
.86
% Difference
9.3%
4.6%
2.3%
Cat 3406
Cat
RTC
ICS,
6HS
.40
5.00
.73
Cat 3406
Cat
EPA
3CS,
14HS
.49
4.82
.83
% Difference
-18.4%
3.7%
-13.1%
-------
Table 5-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Tested Number and HC NOx Particulate
Engine
By
Cycle
Type of
: Tests
(q/BHP-hr)
(g/BHP-hr)
(g/BHP-hr)
Cat 3208
Cat
RTC
3CS,
8HS
.98
9.24
.712
Cat 3208
Cat
EPA
2CS,
6HS
1.07
9.11
.854
% Difference
-8.4%
1.4%
-16.6%
Cat 3208 Model
1
Cat
RTC
5CS,
5HS
.88
13.96
.820
Cat 3208 Model
1
Cat
EPA
2CS,
3HS
1.04
14.13
1.04
% Difference
-15.4%
-1.2%
-21.2%
Cat 3208 Model
2
Cat
RTC
3CS,
6HS
2.40
6.42
.974
Cat 3208 Model
2
Cat
EPA
2CS,
8HS
2.70
6.38
1.24
% Difference
-11.1%
.63%
-21.4%
Cat 3406
Cat
RTC
3CS,
4HS
.37
7.26
.653
Cat 3406
Cat
EPA
3CS,
4HS
.48
7.62
.782
% Difference
-24.2%
-4.7%
-16.5%
CAT 3406 Model
1
Cat
RTC
3CS,
8HS
.47
11.56
.601
CAT 3406 Model
1
Cat
EPA
2CS,
5HS
.60
11.82
.726
% Difference
-22.0%
-2.2%
-17.2%
Cat 3406 Model
2
Cat
RTC
4CS,
4HS
.50
3.78
1.33
Cat 3406 Model
2
Cat
EPA
5CS,
5HS
.57
4.03
1.79
% Difference
-11.8%
-6.2%
-25.7%
Cat 3406 Model
3
Cat
RTC
2CS,
6HS
.77
3.64
2.27
Cat 3406 Model
3
Cat
EPA
3CS,
4HS
.89
4.12
2.20
% Difference
-12.6%
-11.7%
3.2%
-------
03
I
>£-
Table 5-3 (cont'd)
Summary of Emissions Data EPA Cycle vs. Real Time Cycle
Tested Number and HC NOx Particulate
Engine By Cycle Type of Tests (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
Cummins VTB-903 DDA[a] RTC ICS, HS 1.73 4.80
Cummins VTB-903 DDA EPA ICS, HS 1.98 5.07
% Difference -12.6% 5.3%
Det. Diesel 8V- DDA[a] RTC 2HS .72 4.26
92 Model 1
Det. Diesel 8V- DDA EPA 2HS .81 4.60
92 Model 1
% Difference -10.9% -7.4% Y"
Det. Diesel 8V- DDA[a] RTC 3HS .73 7.69 - M
92 Model 2
Det. Diesel 8V- DDA EPA 3HS .68 8.38
92 Model 2
% Difference -7.3% -8.2%
Det. Diesel 8.2L DDA[a] RTC 3HS .92 5.44
Det. Diesel 8.2L DDA EPA 3HS 1.14 5.78
% Difference -19.3% -5.9%
[a] Particulate data not included.
-------
5-12
These emissions offsets are significant, especially
considering the excellent statistical correlations between the
two test cycles. The relatively high coefficients of variation
indicate there is a wide range of emissions differences among
the different diesel engines tested.
In diesel engines, HC emissions are in large part
attributable to residual fuel in the injector sac. The sac
volume is constant regardless of the amount of fuel injected;
as the load is increased {increased fuel injection), the mass
rate of HC emissions from the sac remains constant. However,
the brake specific rate of HC emissions (g/BHP-hr.) decreases
at higher loads, since, the denominator (power) is increasing
while the numerator (mass HC) remains the same. This may
explain in large part the difference in HC emissions, given
that the test cycles both exercise the engine in fundamentally
the same way. Mathematically, an increase of 17 percent in
BHP-.hr. would result in a 15 percent reduction in the
measurement of brake-specific HC:
_! epa _
1.17 RTC ~ '°3
This is nearly the entire observed HC offset.
The observed emissions offsets between the RTC and the EPA
cycle indicate that the RTC is not as stringent as the EPA
cycle at the same numerical level of emission standards. The
RTC does, however, correlate very well with the EPA cycle for
many different engines indicating that emissions from one cycle
can be accurately predicted from those from the other. The
excellent correlation also indicates both cycles are comparable
in the ability to predict in-use emissions, and that there is
no inherent advantage in favoring one cycle over the other - as
long as the respective emission standards reflect equivalent
stringency. Given the difference in cycle generation
methodologies and the correlatable emission results, ana the
reasonable presumption that the HC emissions offset is
primarily attributable to the difference in load factor between
the cycles, we conclude that both cycles are comparably
representative.
The emission standards for the RTC can be adjusted to
levels of equivalent stringency as the EPA cycle by
substituting the EPA standards in the linear regressions from
Figures 5-1 and 5-2. Using this method and EPA standards of
1.3 g/bhp-hr HC and 10.7 g/bhp-hr NOx, the respective standards
for the RTC would be:
HC 1.10 g/bhp-hr NOx 10.6 g/bhp-hr
The advantage to using this approach is that it is direct and
accurately reflects the emission difference at the level of the
B-42
-------
FIGURE 5-1
EPA Cycle vs. RTC BSHC Emissions
EPA Cycle BSHC
(gm/BHP-hr)
B-43
-------
5-14
FIGURE 5-2
RTC vs. EPA Cycle BSNOx Emissions
-1 . Q
2 . 0
10.0
RTC
BSNGx
B . 0
(gm/BHP-hr)
E . ~
H . 0
2 . ~
Yy
X / X
X X
X
y = .97 8x + .156
r = .9567
2 . C1 H.0 B.0 B.0
0.0 I 2 . 0 I H .0
EPA Cycle BSNOx
(gm/BHP-hr)
B-44
-------
5-15
FIGURE 5-3
RTC vs. EPA Cycle Particulate Emissions
EPA Cycle Particulate
(gm/BKP-hr)
B-45
-------
5-16
statutory standard and level typical of current technology
diesels.
4. Economic Impact
It should be noted that the choice of test cycle is a
technical issue, not an economic issue. The cost of operating
a transient test is essentially unaffected by the driving cycle
itself. Also, the emission control strategy for a given engine
should not vary with the test cycle if the level of standards
for given cycles reflect equivalent stringency and the cycles
exercise the engine in fundamentally the same way.
C. Conclusions
1. Most diesel engine manufacturers support, or at
least would not oppose, the RTC as a certification test
option. Inclusion of the NYF segment in the RTC cycle makes
the RTC somewhat more representative of the entire CAPE-21 data
base than the EPA cycle.
2. The RTC is statistically comparable to the EPA
cycle. The most important differences are that the work over
the RTC is higher than the EPA cycle, and with the RTC the
engine is operated over the highest power LAE segment of the
cycle sooner into the test.
3. Data from 30 engine configurations indicate that at
the level of the standards, HC emissions are 15.3 percent lower
on the RTC, and NOx emissions are 1.0 percent lower on the
RTC. These differences in measured emissions arise, most
likely, from the differences in engine work between the
cycles. The emissions results from the RTC correlate extremely
well with EPA cycle results, and both cycles can be expected to
accurately predict in-use emission reductions. Emission
standards for the RTC, adjusted to levels of equivalent
stringency as the emission standards for the EPA cycle, have
been calculated to be:
HC NOx
Cycle (g/BHP-hr) (g/BEP-hr)
EPA 1.30 1C.7
RTC 1.10 1C.6
4. The RTC appears to be an acceptable option for
certification testing as long as the standards are adjusted to
reflect equivalent stringency. This precludes the arbitrary
relaxation of in place emission standards by test procedure
change. Retention of the EPA test as an option would also
B-46
-------
5-17
guarantee that no increase in statutory standard stringency is
forced by inaccurately derived offsets, both on an average, or
on an engine-specific basis. Finally, use of both tests as
options would allow further comparative testing between
cycles. In the future, the use of only one test for
certification would be the simplest and the preferable approach.
5. The choice of test cycles is a technical, not an
economic issue. Relief to the manufacturers at this point in
time with respect to alternate test cycles will have no effect
economically. Accepting alternate cycles as options fully
addresses industry's technical concerns.
B-47
-------
5-18
References
1. Derived from Comments Submitted to EPA Public Docket
No. A-81-20.
2. Derived from Comments Submitted to EPA Public Docket
No. A-81-11.
3. "Evaluation of the Federal Test Procedure for Heavy-
Duty Diesel Engines for 1984 and the Development of the Real
Time Test Cycle," W. L. Brown, Jr., Research Report 88-29, File
18967, Caterpillar Tractor Company, June 22, 1981.
4. "Transient Cycle Arrangement for Heavy-Duty Engine
and Chassis Emission Testing," Chester J. France, EPA Report
HDV 78-04, August 1978.
B-48
-------
CHAPTER 6
THE MVMA CYCLE: AN ALTERNATE HEAVY-DUTY
GASOLINE ENGINE DRIVING CYCLE
The Motor Vehicle Manufacturers Association (MVMA) has
developed an alternative heavy-duty gasoline engine driving
cycle and has proposed it be used as a complete replacement for
the EPA cycle for certification testing. The MVMA cycle was
developed because of concern about the representativeness of
the EPA cycle. in this chapter the industry recommendations,
cycle development, statistical and emissions analyses, and
economic impact concerning the use of the MVMA cycle as a test
alternative are discussed.
I. Summary of Comments[1,2]
MVMA and member mar.uf acturers have submitted specific
recommendations with regards to • the MVMA cycle as a test
alternative to the EPA cycle. These recommendations are
summar i2ed below.
MVMA - MVMA recommends EPA adopt the MVMA cycle in lieu of
the EPA cycle for certification testing. MVMA also recommends
an inter laboratory correlation program be initiated to
determine and remedy any deficiencies.
Chrysler - since Chrysler does not have transient testing
facilities, it has no experience with the MVMA cycle.
Chrysler, therefore, has not made specific recommendations on
the use of MVMA cycle.
Ford - Fcrd recommends the MVMA cycle be adopted instead
of the EPA cycle for certification testing. Ford also believes
an interlaboratory correlation program should be conducted.
General Motors - GM recommends that EPA adopt the MVMA
cycle as a replacement for the EPA cycle. GM also recommends
that an EPA/Industry cooperative study be initiated to
determine if interlaboratory correlation exists and to address
other test procedural concerns.
International Harvester - Since IHC is deemphasizing the
gasoline engine segment of its market, it has not been
following the MVMA cycle developments and does not have
specific comments.
II. Analysis of comments
In this analysis, we address the technical adequacy,
representativeness, and relative stringency of the MVMA cycle.
B-49
-------
6-2
A. Cycle Development[3]
The MVMA heavy-duty gasoline engine driving cycle was
developed because of industry concerns that the EPA cycle was
inadequate in the following two areas:
1. It was not representative of real world truck
operation.
2. The irregular nature of the cycle could create
interlaboratory correlation problems.
In an attempt to alleviate some of these concerns, MVMA
modified the EPA cycle to obtain a driving cycle which they
felt was more representative and more acceptable. MVMA
established four basic objectives for constructing the modified
test cycle. The modified cycle bad to:
1. Maintain the general character of the EPA cycle„
2. Improve the relationship between simultaneous speed,
power, and acceleration.
3. Reduce momentary speed excursions.
4. Reduce excessive throttle manipulations.
To accomplish these objectives, the cycle was simply examined
on a second-by-second basis; using engineering judgment, the
speed and torque specifications were revised where it was
deemed appropriate. The resulting driving cycle was a smoothed
version of the EPA cycle with a revised synchronization between
speed and torque commands. Technical justifications for
specific cycle changes were not submitted by MVMA. EPA was
merely presented with the cycle and the claim that it
represents MVMA's best engineering judgment. unlike the Real
Time Cycle alternative, in-depth documentation of the MVMA
cycle development methodology was not provided to EPA.
B. Statistical Analysis
A comparison of overall statistical parameters from the
MVMA cycle, EPA cycle, and the CAPE-21 data base is listed in
Table 6-1. The CAPE-21 statistics are included for comparison
purposes, although the MVMA cycle was not directly derived from
the CAPE-21 data base.
As can be seen from the table, the EPA cycle and MVKA
cycle are very similar statistically, with respect to speed and
torque parameters. There are no major discrepancies, which is
to be expected since the MVMA driving cycle is directly derived
from the EPA cycle. However, data from engine tests indicate
B-50
-------
6-3
Table 6-1
Cycle Statistics:MVMA Cycle,
EPA Cycle, CAPE-21 Data Base
Parameter MVMA EPA CAPE-21
Torque
Mean {%) 3 7 36 34
Percent of Cycle Time
Accel (%) 15 17 15
Decel (%) 19 20 16
Cruise (%) 28 26 28
Motor (%) 9 10 13
Idle (%) 28 27 28
RPM
Mean (%) 31 30 31
Percent of Cycle Time
Accel (%) 20 24 20
Decel (%) 26 21 26
Cruise (%) 26 23 26
Idle (%) 28 31 28
B-51
-------
6-4
total engine work (BHP-HR) over the MVMA cycle is about 10
percent higher than on the EPA cycle. This increase in cycle
work is5 ^attributed to the resynchronization of the speed ana
torque commands so there are more events where speed and torque
are increasing or decreasing at the same time, resulting in
increased integrated power-hr. Another important difference is
that the MVMA cycle is less transient than the EPA cycle. The
speed and torque sequences are smoother and numbers of torque
accelerations have been completely eliminated/ reducing the
frequency of throttle position changes for engine operation
(reducing accelerator pump operation and transient fuel
enrichment). In summary, the MVMA cycle is statistically
similar to the EPA cycle, but as explained in greater detail
below, not operationally identical.
3. Emissions Analysis
EPA, Ford, and GM have to date tested a small number of
gasoline engines and configurations to compare emissions
results on the MVMA cycle and the EPA cycle. The emissions
data from these tests are summarized in Table 6-2 and plotted
in Figures 6-1, 6-2, and 6-3 with their accompanying regression
lines, a total of 12 engines/configurations are included.
Excellent statistical correlations were observed between
the MVMA cycle and the EPA cycle; for emissions of HC, CO and
NOx, values of r2 were .97, .98, and .93 respectively.
Definite emissions offsets were also observed. The MVMA cycle
HC emissions offset averaged -21.32 percent from the EPA cycle
with a range of -44 percent to +13 percent and coefficient of
variation of .724; the CO emissions offset averaged -7.45
percent with a range of -24 percent to +9 percent and
coefficient variation of 1.27; and the NOx emissions offset
averaged .98 percent with a range of -4.2 percent to +10.1
percent and coefficient of variation of 4.64. As indicated by
the regression lines, the percentage emissions offsets between
cycles increase as absolute emission levels decrease.
The HC and CO emissions differences are explainable by the
operational differences between the cycles. The most
significant differences between the MVMA cycle and the EPA
cycle are that the MVMA cycle is smoother, and that the speed
and torque commands follow each other more closely on the MVMA
cycle (resulting in an increase in integrated power-hour).
These changes are illustrated graphically in Figure 6-4 where
the same characteristic sections from both test cycles have
been overlaid. The decrease in the transience of the MVMA
cycle will result in less movement of the engine accelerator
pump, which could be expected to result in lower HC and CO
emissions.
The degree to which the cycle smoothing explains the
observed emission differences, however, is unknown. The
B-52
-------
Table 6-2
Summary
of Emissions Data EPA Cycle vs. MVMA
Cycle
Enyine
Tested
by
Cycle
Number and
Type of Tests
HC
(q/BHP-hr)
CO
(q/BHP-hr)
NOx
{y/BHP-h
1969 GM 292
1969 GM 292
EPA
EPA
EPA
MVMA
3CS,5HS
2CS,5HS
6 .12
4.72
118.4
1U9.4
6 .54
6.38
% Difference
-22 .88%
-7.6%
-2.45%
1969 Ford F300
1969 Ford F300
EPA
EPA
EPA
MVMA
2CS,4HS
2CS,4HS
7 .64
6.49
126.64
124.99
7.74
7 .50
% Difference
-15.05%
-1.32%
-3.10%
1969 GM 350
1969 GM 350
EPA
EPA
EPA
MVMA
2CS,5HS
2CS,5HS
8.14
7.71
135.53
143.15
4 .43
4.22
% Difference
-5.60%
5.62%
-4.74%
Ford 4.9L
Ford 4.9L
Ford
Ford
EPA
MVMA
2CS,2HS
ICS/1HS
2.86
2 .40
28.4
21.7
8.04
8.75
% Differ >nce
-16 .08%
-23.59%
8.83%
Ford 6.1L [a]
Ford 6.1Lj I
Ford
Ford
EPA
MVMA
ICS,1HS
ICS,1HS
2 .36
1.50
28.9
27.8
7 .42
6.67
% Difference
-36.44%
-3.81%
L0.11%
Ford 6.1L II[a]
Ford 6.1L II
Ford
Ford
EPA
MVMA
2CS,4HS
2CSi6HS
2.46
1.59
30.5
25.7
8.29
8.01
% Difference
-35.37%
-15.74%
3 .38%
Ford 6.1L III[a]
Ford 6.1L III
I Ford
Ford
EPA
MVMA
ICS, 3HS
ICS, 3HS
3 .28
1.81
31.3
27 .7
8.55
8.77
% Difference
-44 .82%
-11.50%
2.57%
-------
Table 6-2 (cont'd)
Summary of Emissions Data EPA Cycle vs. MVMA Cycle
Engine
Ford 6-..1L IV[a]
Ford 6.1L IV
% Difference
1981 454
1981 454
% Difference
1981 454
less controls
1981 454
less controls
% Difference
1981 427
1981 427
% Difference
1981 427
less controls
1981 427
less controls
% Difference
1981 292
1981 292
Tested
by
Ford
Ford
GM
GM
GM
GM
GM
GM
GM
GM
GI«1
GM
Cycle
EPA
MVMA
EPA
MVMA
EPA
MVMA
EPA
MVMA
EPA
MVMA
EPA
MVMA
Number and HC
Type of Tests (g/BHP-hr)
ICS/1HS
ICS,1HS
ICS / 1HS
ICS/1HS
ICS/1HS
ICS/1HS
ICS,HS
ICS,HS
ICS,1HS
ICS,1HS
ICSi1HS
ICSi1HS
2.34
1.48
-36.75%
1.28
1.44
12 .50%
3 .12
2 .89
-7 .37%
7 .45
5.85
-21.48%
10.06
8.51
-15.41%
3 .33
2 .70
CO
(g/BHP-hr)
30.6
25.5
-16.6 7%
47.9
52 .4
9.39%
98.7
100 .4
1.72%
63 .4
52 .9
-16.56%
129.6
116 .0
-10.49%
26.7
26.3
NOx
(q/BHP-hr)
8.17
8.04
-1.5 9%
5.06
4 .91
-2.96%
5.48
4.25
-4 .20%
6.22
6 .47
4.02%
5.67
5.93
4 .59%
8 .13
8.11
-------
Table 6-2 (cont'd)
Summary of Emissions Data EPA Cycle vs. MVMA Cycle
Engine
% Difference
1984 292
prototype
1984 292
prototype
% Difference
Tested
fay
GM
GM
Number and HC
Cycle Type of Tests (g/BHP-hr)
-18.92%
EPA
MVMA
1HS
1HS
1.6 6
1.08
-34.94%
CO
(g/BHP-hr)
-1.50%
12 .3
10.8
-12.2%
NOx
(y/BHP-hr )
-.25%
8.93
8.89
-.45%
DO
I
U1
LP
% Differences =
MVMA -EPA
EPA
[a] Unique hardware and/or calibration.
a>
I
-------
6-b
FIGURE 6-1
EPA Cycle vs. MVMA Cycle BSHC Emissions
2 . 0
10.0-
MVMA
Cycle
BSHC
(gm/BHP-hr)
a . 0
e . 0
h . a
2 . £3
a . a
a . a 2.a h.h e.b b.b ih.q 12.0
EPA Cycle BSHC
(gm/BHP-hr)
B-56
-------
6-5
FIGURE 6-2
EPA Cycle vs. MVMA Cycle BSCO Emissions
iWMA
Cycle
BSCO
(gm/BHP-hr)
I '10
I 22]
I Hfcl
Utl
BH
Itl
!E1
El
CI
20
-|£1 EO BfJ I Hiil I 20 I MB
EPA Cycle BSCO
(gm/BHP-hr)
B-57
-------
6-10
FIGURE 6-3
EPA Cycle vs. MVMA Cycle BSNOx Emissions
i ta . ej
ei . a
MVMA
Cycle
BSNOx
(gm/BHP-hr)
E . a
-I . trl
0/
¦I.; El
a Q
y = 1.083x - .7142
r = .9257
E . 0
a. a
i a. a
EPA Cycle BSNOx
(gm/BHP-hr)
B-58
-------
FIGURE 6-4
MVMA Cycle and EPA Cycle Comparison
O20U ObM . 441 OA 14100 44000
462 00 iU 00 460 01 40*00 408 fit 02.00 476 00 4M00 404 00 <68 00 492 00
It? an 4 'Mi 0£) ' 110 CM) 44400
tie 00 4&2 00 4SOOO
4/0 00 400 00 404 00 400 00 492.00
Time (seconds)
Key
MVMA
EPA
-------
6-12
difference in transience of the MVMA cycle and EPA cycle has
not been quantified, and previous data (see Chapter 4)
indicated that a decrease in transient operation resulting from
dampening of the throttle controller had no effect on
emissions.* The rephasing of the speed and torque commands
results in different modes (combination of speed and torque) of
engine operation on the two test cycles, with fewer events at
both lower speed and load on the MVMA cycle. The observed
increase in power-hour over the MVMA cycle may also explain the
decrease in the specific emissions (i.e., more emissions
divided by increased output work). This may contribute to the
lower BSHC and BSCO emissions on the MVMA cycle, although the
effect has not been quantified. Changing the engine speed at
which motoring (defined as -10 percent maximum engine torque)
occurs may also explain emission differences between the
cycles. Ford has mentioned the possibility of modal and
subcycle testing[2] to help understand emissions offsets
between the two cycles. This type of information would be
beneficial and its acquisition is encouraged.
The MVMA cycle does not yield emissions equivalent to the
EPA driving cycle, i.e., it is not equivalently stringent at
the same numerical emission standards for HC and CO. The MVMA
cycle does, however, correlate well with the EPA cycle for the
different engines tested to date. A strong correlation between
an alternate test cycle and the EPA cycle implies that there is
no advantage in using one c^cle over the other as long as the
emission standards reflect equivalent stringency.
The MVMA cycle HC and CO standards have been adjusted to
levels of equivalent stringency as the EPA cycle by
substituting EPA cycle standards in the least squares
regression lines from Figures 6-1 to 6-3. The standards are
summarized below:
1984 statutory emission standards based upon the MVMA cycle would,
therefore, be .8 g/BHP-hr HC, 12.5 g/BHP-hr CO, and 10.9 g/BHP-hr
NOx. Interim revised standards would be adjusted in an identical
manner. (For example, the proposed revised EPA standards of
1.3/35.0 would be equivalent to MVMA cycle standards of .8/31.9.)
There is perhaps a significant difference between
mechanically smoothing accelerations (as was done by EPA's
sensitivity testing) and completely eliminating accelerations
(as was done by MVMA).
HC(g/bhp-hr) CO(g/bhp-hr) NOx(g/bhp-hr)
1984 EPA Cycle
1984 MVMA Cycle
1.3
0.8
15.5
12 .5
10.7
10.9
B-60
-------
6-13
This method of standard adjustment results in large
percentage differences (38 percent HC, 19 percent CO) between the
MVMA cycle and the statutory EPA cycle standards. A large
difference in cycle standards is appropriate if the MVMA cycle is
indeed significantly more lenient than the EPA cycle. Futhermore,
it is not known how the operational differences between the MVMA
and EPA cycles would affect compliance strategies. For example,
the CO emissions difference between test cycles for the Ford 6.1
liter engine (see Table 6-2) varied from 1.1 g/bhp-hr (4 percent
difference) to 5.1 g/bhp-hr (17 percent difference) with
calibration and/or hardware changes, indicating that the MVMA
cycle may be more amenable to a given emission control strategy.
However, the data on which this analysis is based are limited to
14 configurations of 8 different engines, including three 1969
baseline engines. More testing will be required to more
confidently determine the emissions differences between cycles at
the level of the emission standards, and more adequately
characterize the MVMA test cycle.
D. Economic Impact
As was discussed in Chapter 5, the choice of test cycles will
not have an effect economically. The cost of operating a
transient test is unaffected by the driving cycle itself, assuming
the standards are adjusted to levels of equivalent stringency to
the EPA cycle, and the cycles exercise the engine in fundamentally
the same way. The choice of test cycles is a technical, not an
economic issue.
Ill. Conclusions
A. MVMA. and its members who are most significantly-
affected, Ford and GM, recommend the MVMA cycle be used for
certification testing instead of the EPA cycle. They also
recommend an EPA/Industry cooperative test program be initiated..
B. The MVMA cycle is statistically comparable to the EPA
cycle. The most significant differences are that the speed and
the torque sequence is smoother, total engine work (BHP-hr) is
higher on the MVMA cycle, and the speed and torque cycles have
been rephased with each other.
C. Data from 14 engines/configurations indicate that, at
the level of the standards, HC emissi^s are 38.5 percent lower on
the MVMA cycle, CO emissions are 19.3 percent lower on the MVMA
cycle, and NOx emissions are 1.9 percent higher on the MVMA
cycle. However, the . sample size may be too small to accurately
assess the emissions differences for all gasoline engines.
Conclusions based upon this data alone should be regarded as
tentative until more engine testing, such as an EPA/Industry test
program, can be conducted.
B-61
-------
6-14
D. The MVMA cycle may be an acceptable option for
certification testing. However/ the undocumented and perhaps
arbitrary method of MVMA cycle generation gives us less assurance
of its overall adequacy relative to EPA's cycle. This caution is
warranted both by the very small data base at present/ and also by
the data presented above, in which engine recalibrations can yield
significantly greater emission reductions on the MVMA cycle than
the EPA cycle. These facts argue for more comparative testing
before the MVMA cycle can be judged to be an equivalent option to
the EPA cycle.
B-62
-------
6-15
References
1. Derived from Comments Submitted to EPA Public Docket
No. A-81-20.
2. Derived from Comments Submitted to EPA Public Docket
No. A-81-11.
3. MVMA-Modified Heavy-Duty Gasoline Engine Transient
Emission Test Cycle, Attachment, Letter to EPA Administrator,
Motor Vehicle Manufacturers Association, Feb. 15, 1982. (See
EPA Public Docket No. A-81-11, IV-D-2 and IV-D-2a.)
B-63
-------
CHAPTER 7
THE DIESEL TRANSIENT TEST/13-MODE OPTION
I. Summary of Comments
Diesel engine manufacturers again questioned the need for
an expensive transient test for diesel engines. Several
manufacturers and the EMA submitted emission data to argue that
a sufficient correlation exists between the 13-mode
steady-state and EPA's transient test for regulated
pollutants. In addition, the manufacturers argued for a more
extensive steady-state option for 1984 and later years than the
single-year option represented by the 0.5 g/BHP-hr optional
13-mode HC standard which EPA originally provided.
II. Analysis of Comments
Figures 7-1, 7-2, and 7-3 present available comparative
emission data for the 13-mode and EPA transient tests. In no
cases were the regression lines "forced" through zero—as were
many of the regression analyses performed by the industry. Not
only is the zerg-zero point a physically meaningless point of
singularity, it is incorrect statistical practice to
arbitrarily force data through any non-data point. This causes
the correlation to appear much better than it actually is.
Such is the case with the industry's analyses.
Compiling all data submitted by the industry, [1] all data
taken by EPA and Southwest Research, [2] and all available data
from the EPA/EMA Round Robin Program,[3] EPA evaluated the
correlation between tests. For BSHC, the least squares
regression line had an value of .51, with an intercept of
.45 g/BHP-hr. For BSNOx, and BS particulate emissions, the
calculated values were .72/2.1 and .73/.34 respectively.
At current levels of BSHC (both above and below the
statutory transient standard of 1.3 g/BHP-hr), the correlation
between transient and steady-state test procedures is poor.
Only 51 percent of the variation in a transient result is
explainable by a variation in a 13-mode result, i.e., the
13-mode is a poor predictor of transient BSHC. This is
indicated most clearly by the dramatic scatter of data
presented in Figure 7-1. Note that 13-mode results as low as
.5 g/BHP-hr are found on several engines whose corresponding
transient results exceed the 1.3 g/BHP-hr transient standard.
At current levels of NOx emissions, L ~'"arf the 13-mode is a
better but not good predictor of transient emissions. (See
Figure 1-2.) We note that this "correlation" exists at levels
of emissions which represent little applied emission control.
EPA's earlier analysis of the justification for a transient
test[4] indicated that steady-state tests traditionally become
ineffective as technology is applied to achieve substantial
B-64
-------
7-2
FIGURE 7-1
Comparative Transient vs. 13-Mode HC Emissions
2. 5_
EPA Cycle
BSHC
(gm/BHP-hr)
2.0
1.5
1.0
0.5
Least
Squares
Regressio
Line
of
Perfect
Agreement
o.o
+ Coterplllar
* Cununltis
i UUA
Bint erna clonal
A Hock
2.0
13-Mode Cycle BSHC
(gm/BHP-hr)
Regression Line Equation: y = . 96x + .449
2
r = .514
B-65
-------
7-3
FIGURE 7-2
Comparative Transient vs. 13-Mode Emissions
for NOx
10
EPA
Cycle 6
USNOx
(gm/BHP-hr) 5
Least
Squares
Regression
Line
KEY
4- Cocetpillar
jr Cununlus
* UDA
a lnternacional
A Hack
10
11
13-Mode Cycle BSNOx
(gm/BHP-hr)
Regression Line Equation: y = .685x + 2.085
r2 = .7 22
B-66
-------
7-4
FIGURE 7-3
Comparative Transient vs. 13-Mode Particulate Emissions
EPA Cycle
(gm/BHP-hr)
.8
Least
'Squares
Regression
Line
~ Caterpillar
x Ciiiiuni'ns
A 00A
8 Iiiternationa 1
A Hock
13-Moae Cycle
(gm/BHP-hr)
Regression Line Equation: y = .758x + .342
r = .728
B-67
-------
7-5
emission reductions. This is also true of the "correlation"
between transient and steady-state particulates (see Figure
7-3) for which future standards will also be promulgated. The
promulgation of more stringent standards (NOx and particulate)
makes the use of the transient test all the more necessary to
ensure that laboratory emission reductions also occur in the
field. Using all available emission data, we find that the
transient test is technically justifiable today, and even
moreso when more stringent NOx and particulate standards become
applicable.
This justification did not prevent EPA from originally
providing a steady-state option for 1984 for heavy-duty diesel
engines. EPA's rationale was to allow time for the industry to
investigate the use. of existing eddy-current dynamometers, and
perhaps save money on facility conversion. The optional 0.5
g/BHP-hr HC standard was set to minimize the possibility that
no compromise in engine emissions would result (relative to the
1.3 transient standard), although EPA has previously stated
that the standard was intended to be approximately as stringent
as the transient standard and to not require extensive
development work. It now appears that a single year option
would effectively spread transient recertification requirements
out over two model years, as our leadtime analysis in Chapter
III indicates is necessary for the diesel manufacturers.
The industry has argued that a 0.5 g/BHP-hr optional
13-mode standard is too stringent, and is essentially a
non-option.[1,5] As indicated by Figure 7-4 (representing the
HC emission distribution of all domestic and foreign HD diesel
engines certified in 1982 on the 13-mode test), as the optional
13-mode standard becomes more stringent, fewer engines can be
carried over without development work and recertification.
Figure 7-4 indicates that with a 0.5 standard and a full useful
life, 21 percent (13/67) of 1982 engine families would be
eligible for carryover. Revising the optional standard to the
industy's recommended 1.0 g/BHP-hr permits 73 percent (46/67)
of 1982 engine families to carryover. This would certainly
provide increased flexibility for allocation of testing and
development resources for both model years 1984 and 1985. (We
note that these percentages do not include engines which can
still be recertified on the 13-mode test in 1983, and then be
eligible for carryover.) In short, EPA's acceptance of the
industry's recommended optional standw' n would alleviate much
of the recertification testing burden for 1984.
Based upon the correlation presented in Figure 7-1,
however, we have little confidence that any 13-mode standard
between 0.5 and 1.3 can adequately predict compliance with a
1.3 transient standard; hence, its level should be set based
upon th,e industry's requirements for relief, not upon emission
control requirements. A steady-state HC standard of 1.0
B-68
-------
Number of
Engine
Families
CO
i
cn
vc
0.0
0.5 1.0
HC Emissions (gm/BHP-hr)
1 1 r
1.5
FIGURE 7-4: 1982 Diesel Certification Data (13-Mode BSHC)
-------
7-7
g/BHP-hr would be appropriate to provide some flexibility to
the industry as they spread out recertification over an
additional year. A 0.5 g/BHP-hr standard would likely yield
some additional emission reductions, but not to any predictable
extent. Practically, however, the relative difficulty in
meeting a 0.5 g/BHP-hr standard may in fact force transient
certification, which is not consistent with the relief
requirements identified. A 0.5 g/BHP-hr steady-state option
may actually misdirect technology from transient certification.
As for the provision of a relaxed steady-state option
(e.g. to 1.0 g/BHP-hr), we see no justification for such action
at this time. The poor correlation between the 13-mode and the
transient test gives us little confidence that a 1.0 g/BHP-hr
steady-state standard would yield true reductions from the
current 13-mode HC standard of 1.5 g/BHP-hr. Since the
provision of the steady-state option is intended to allow
recertification to be spread over an additional year (i.e.,
allow carryover of nearly half of the engine families), those
engines carried over on the option will receive no additional
development work anyway. For these engines, decreasing the
steady-state standard from 1.5 to 1.0 g/BHP-hr is irrelevant.
As noted above, optional 13-mode standards more stringent than
1.0 g/BHP-hr cannot predict further emission reductions, will
likely either force transient certification or misdirect
technology towards steady-state certification, and will
decrease industry's flexibility in selecting the order in which
they recertify.
In summary, a one year deferral in transient testing
requirements, (i.e., maintaining the present steady-state HC
standard of 1.5 g/BHP-hr, along with the option to certify on
the transient test for 1984) is appropriate for relief
purposes, and is not practically different in effect from a 1.0
g/BHP-hr standard.
The environmental impact of this deferral is difficult to
quantify, since the transient data presented in Figure 7-1 do
not represent a complete characterization of all diesel engine
families. (In most cases, the manufacturers declined to
identify specific engine families when they submitted their
comparative data to EPA.) We note, however, that the majority
of engines for which data are available already comply with the
transient 1.3 standard, and we think A. unlikely that their
emissions would increase. This will mitigate the overall
impact of the 1984 deferral. For those engines which now
exceed the transient standard and from which we expect all the
environmental gains, a single year deferral is expected to
cause an unmeasurable air quality impact.
When advocating a less stringent steady-state option, the
industry also argued for an extension of the 13-mode option
until recertification (i.e., until the engine family is
B-70
-------
7-8
redesigned, a new family is introduced, or new emission
standards apply). With an optional steady-state HC standard of
1.0 g/BHP-hr, this would absolve 73 percent of all engines
certified in 1982 from certifying on the transient test. (This
does not include those engines which could be certified on the
13-mode in 1983 and 1984, which would then be eligible for
carryover). With this option, it is quite likely that few
engines - other than new introductions - would be required to
be tested on the transient test. This would certainly defer
incremental testing and development expenses, but it would also
defer for several years most of the emission reductions the
transient test was intended to achieve. The analysis in
Chapter III indicates that some sort of steady-state option or
deferral is required for 1984 to accommodate facility
changeover. Beyond that year, however, diesel engine
manufacturers will be able to comply with transient testing
requirements and do not need further relief.
In summary, to carryover engines on the steady-state test
until recertification is required would defer transient testing
requirements for the large majority of diesel engines. This
would allow continued production for several years of the very
engines the transient test was intended to bring into
compliance. This is fundamentally different from a single year
option or deferral, in which a manufacturer can space out
development and certification work over an additional year, but
still works to have all engines in compliance by 1985. Since
the majority of test facilities have been paid for with
investments which are unrecoverable, such a carryover would
assure that little environmental returns would be realized from
the large facility investment.
III. Conclusions
A. The technical justification for a transient test for
heavy-duty diesel engines remains strong. The evidence
substantiates EPA's earlier ana.lyses[4] that a transient test
is justified in 1984 and even moreso when more stringent NOx
and particulate standards apply.
B. A single year deferral in mandatory transient
testing requirements would accommodate the facility changeover
which the diesel industry indicated was necessary. This
deferral represents no practical difference from a relaxed
steady-state option, and does not present the difficulties
associated with a more stringent steady-state option. A
carryover option beyond this (e.g., until recertification was
required) would defer most transient testing requirements, and
also most of the benefits the transient test was intended to
achieve, for no demonstrated reason.
B-71
-------
7-9
References
1. Derived from Industry Comments Submitted to Public
Docket No. A-81-20.
2. "Emissions from Heavy-Duty Engines Using the 1984
Transient Test Procedure, Volume II - Diesel," S. Martin,
prepared by the Southwest Research Institute foe the US EPA,
EPA-460/3-81-031, July 1981.
3. "status of EPA/EMA Cooperative Test Program," A.
Azary, EPA Technical Report, March 1982.
4. Summary and Analysis of Comments to the NPRM: 1983
and Later Model Year Heavy-Duty Engines," Proposed Gaseous
Emission Regulations," December 1979.
5. Derived from Industry Comments Submitted to Public
Docket No. A-81-11.
B-7 2
-------
CHAPTER 8
SUMMARY AND RECOMMENDATIONS
I. Discussion
A. implementation Issues
The driving force behind implementation of a transient
engine test is its technical justification, i.e., its increased
ability over steady-state tests to predict in-use emissions.
For gasoline engines, this justification is unquestionably
solid.* As discussed in Chapter 7, the justification for
diesel engines has become stronger as more data are collected.
Given that a transient test remains technically necessary
for heavy-duty engines, we reviewed in Chapter 4 whether the
specific test procedure - EPA's cycles and procedures - was
technically adequate. The development and procedural details
of EPA's test have been the subject of vigorous industry
criticism for some time. We reviewed criticism which has
arisen since promulgation of the test in January, 1981, and all
new emission data acquired since then to address this issue.
With respect to the validity of EPA's test cycle, industry's
criticisms were speculative. The manufacturers have continually
failed to provide real evidence (i.e., factual emission data)
that EPA's cycles iield unrepresentative results. in fact,
EPA's cycles correlate very well with the alternative cycles
proposed b^ the industry (discussed below), cycles from which
the industry removed all objectionable operation. Other
available emission evidence also substantiates the adequacy of
EPA's cycles. Operationally, claims of repeatability and
correlation problems have not been proven. The EPA/EMA Round
Robin correlation project yielded reasonably good results, and
gave no indication that fundamental problems exist. Such a
program has not been run with the gasoline engine
manufacturers, although MVMA has asked that one begin. EPA's
experience with light-duty vehicle testing indicates that such
a program is an important milestone in identifying and
correcting problems. That minor problems will exist is almost
inevitable; maintaining correlation between emission testing
laboratories is an ongoing project for even the most
established tests (e.g., LDV test procedure). We anticipate
that a number of future technical amendments will be necessary
as industry experience with the procedure accumulates. Again,
no fundamental problems have been identified and there are no
conceptual reasons why ones should exJ v In short, EPA's test
procedure remains acceptable (although certainly capable of
* See EPA's earlier analysis of this issue in Chapter 1 of
the summary and Analysis of comments to the NPRM: "1983
and Later Model Year Heavy-Duty Engines - Proposed
Emission Regulations," December 1979.
B-73
-------
8-2
refinements) ; the large baseline and current technology data
bases acquired on the EPA test have not been compromised by any
test procedural flaws.
Given a technically necessary and adequate test procedure,
we reviewed the economics and leadtime issues associated with
the transient test in Chapter 3. All gasoline manufacturers,
except Chrysler and international Harvester (gasoline), have
made the investment in transient facilities; this investment is
not recoverable if the transient test is delayed or withdrawn.
The gasoline manufacturers claim that they will not have
sufficient leadtime to certify all their engine families on a
transient test in 1984; GM will not be able to certify on the
9-mode test in 1984 because of facility changeover. This is
not a problem. Due to practical leadtime requirements, the
earliest that revised transient emission standards will be
implementable is 1985. Since both Ford and GM (gasoline) have
spent 100 percent of the required investments in facilities,
this effective delay does not provide economic relief. It is
advantageous, however, since it allows more leadtime for engine
development* and test procedure refinements. For 1984,
carryover of 1983 9-mode certification data will be necessary,
as will an option to certify on the transient test, to better
accommodate the changeover of facilities.
This effective deferral also benefits Chrysler and IHC.
After 1984, however, International Harvester currently plans to
leave the gasoline engine market and will be unaffected by
future test procedural changes. Chrysler's situation remains
uncertain. Chrysler has apparently made the business decision
that investments in transient facilities will not be made. it
is also doubtful whether Chrysler would invest in development
work to meet an equivalently stringent steaa.y-state test, were
one feasible. (Chrysler has had the option of certifying its
relatively lightweight product line on the light-duty truck
procedure, but has neglected to do so on account of the
procedure's stringency. Chrysler was the only gasoline engine
manufacturer to opt for the "quick fix" use of catalytic
converter technology in response to the far less stringent 1980
steady-state California emission standards.) Chrysler has also
argued that it may not purchase the technology nor contract out
the testing, on the grounds that their product may then become
cost uncompetitive. There is no fair administrative means by
which EPA can exempt only Chrysler from transient
requirements. Chrysler's 1982 production will exceed 20,000
engines, so it is difficult to classify it as a small-volume
manufacturer for relief purposes. Chrysler's products are not
sufficiently different from those of the other gasoline engine
manufacturers to provide a basis for any other form of separate
classification.
Note that EPA will not finalize revised HDE emission
standards until the second half of 1982.
B-74
-------
8-3
The diesel engine manufacturers have also committed a
large percentage of their required investments, which are also
unrecoverable. All diesel manufacturers will be able to
conduct some degree of transient testing in 1984. All
facilities will be operational by 1985. This argues for the
provision of a steady-state option or deferral for 1984. In
Chapter 7, we reviewed all available data comparing transient
to 13-mode emissions. For HC, the correlation between test
procedures is poor (r2 = .51); 13-mode HC emissions between
0.5 g/BHP-hr and 1.3 g/BHP-hr cannot predict transient
emissions in excess of the transient 1.3 g/BHP-hr standard.
For emission control purposes, the 13-mode is inadequate. A
deferral in transient testing requirements for a single year is
necessary to accomodate the changeover of testing facilities.
The industry's recommended level of the optional steady-state
standard (1.0 g/BHP-hr) would afford the industry a great deal
of flexibility. (Indeed, over 70 percent of all domestic and
foreign engine families would be eligible for carryover without
recertification.) This option, however, represents no
practical difference from a one-year deferral. A more
stringent optional steady-state standard would not result in
reductions comparable to those achievable using the transient
test, and would either misdirect development work needed to
achieve true emission reductions or force transient
certification anyway.
We have also reviewed in Chapter 7 a less burdensome
approach, i.e., allowing indefinite carryover until
recertification is required. We concluded that the industry
did not need this much relief, based upon our economic and
leadtime analysis, and that such a carryover option would
sacrifice a large part of the HC reductions attainable on the
transient test and already paid for by the facility
expenditures. (Such an option effectively defers transient
testing requirements for the large majority of diesel engine
families.)
In summary, given the single year deferral and the need
for relatively little development work to meet applicable
standards, the diesel industry will have little problem
complying with our recommended transient requirements.
B. Technical Issues
Both the EMA and MVMA submitted ai..tnative transient test
cycles (described at length in Chapters 5 and 6).
Statistically, these alternate cycles are quite similar to the
EPA cycles, and the CAPE-21 data base. They are not completely
similar, however; identifiable operational differences exist
between the cycles and their EPA counterparts. These
differences in operation have led to differences in measured
emissions. In fact, both alternative test cycles were less
stringent than the respective EPA cycle.
B-75
-------
8-4
On the other hand, emissions from both cycles correlated
very well with results from the EPA cycles, and vice versa.
(Coefficients of determination (r2) for all emissions were
.90 or greater.) In other words, one cycle should predict
emission reductions as well as the other. Given that the
industry cycles were developed using different methodologies,
the observed correlation tends to confirm all cycles as equally
representative, although we note that the number of engines
tested to date on the MVMA cycle is smsll and technical
reservations about the MVMA cycle still exist.
To maintain comparable stringency with the alternate test
cycles, i.e., achieve the same degree of on-road emission
reductions as the EPA cycle, the EPA cycle standards must be
corrected. To not do so, i.e., to maintain EPA-based standards
with the alternate test cycles, represents a de facto
relaxation of standard stringency through test procedure
change. This would increase on-road emissions (notably HC) by
the percentage of effective relaxation.
The alternate standards for both test cycles were
determined by substituting the EPA standards in linear
regression equations derived from all available comparative
data. Presently there are 30 comparative data points for the
RTC and 14 points for the MVMA cycle. More testing may be
necessary to more confidently determine the relationship
between EPA and MVMA cycle emissions, and to more conclusively
establish the viability of the MVMA cycle as a long-term option
or replacement.
Equivalent Emission Standards (q/BHP-hr)
BSHC BSCO BSNOx BSPart
EPA Cycle (gas) 1.3* ** 15.5+/35.0** 10.7 HA
EPA Cycle (diesel) 1.3 NA 10.7 .25(P)
MVMA Cycle 0.8* ** 12.5V31.9** 10.9 NA
EMA Cycle 1.1 NA 10.6 .25(P)
HA: Ho applicable standard.
(P): Proposed particulate standard
* Statutory standards.
** Proposed revised emission standards.
MVMA recommended that EPA adopt its cycle as a complete
replacement for the EPA cycle, primarily because of its
judgment that EPA's cycle was unrepresentative. EMA requested
that its cycle be adopted as an option.
Complete replacement of EPA's cycle at this time is
unwise. Such action would be justified if the MVMA cycle were
B-76
-------
8-5
technically superior to the EPA cycle; in our judgment that is
not the case. Such action could open to question the
accumulated data base which EPA has collected using EPA's
cycle, including several baseline studies on which both
finalized and proposed emission standards are based.
Furthermore, the data base for the MVMA cycle is relatively
small, ana the calculated EPA/MVMA emission standard offsets
may not agree with those derived from a larger data base. (The
alternate emission standards are only as accurate as the
calculated regression lines.) Furthermore, technical
reservations still exist for the MVMA cycle, notably its
undocumented derivation and the apparent ease of attaining
emission reductions relative to EPA's cycle. (See Chapter 6.)
Finally, complete replacement precludes any future flexibility
which EPA would retain if the alternate cycles were accepted as
options, i.e., the flexibility to analyze and review data
collected over several years and during actual certification.
Adoption of both alternate cycles as at least short term
options, however, addresses all of the industry's technical
concerns while leaving EPA the flexibility of selecting a
single cycle in the future. Optional use of either procedure
at equivalently stringent standards is the wisest course of
action.
Finally, as mentioned earlier, a number of technical
amendments to the test are being reviewed to cut costs and
enhance technical aspects of the test. These will be
expeditiously pursued, since they will likely have as great an
economic impact as any other test procedural action.
II. Recommendations
1. For gasoline engines, finalize a mandatory transient
test for 1985. In 1984, allow both a carryover of 1983
steady-state data and optional certification on the transient
test at 1985 standards. This approach gives no special relief
or consideration to Chrysler beyond 1984.
2. For diesel engines, finalize the transient test for
1985. For 1984, allow certification on the 13-mode at existing
standards to accommodate facility changeover. Also, allow
optional certification on the transient test for 1984. After
1984, all engine families would be certified on the transient
procedure.
3. Allow the optional use of the EMA and MVMA cycles,
certifiable to standards which are equivalently stringent as
the EPA cycle-based standards. Based upon experience gathered
over time, eventually select a single cycle for gasoline
engines and one for diesel engines.
4. As necessary, make refinements and technical
amendments to the test procedure itself, with as many changes
B-77
-------
8-6
as possible being finalised in the upcoming revised 1984
heavy-duty engine emission regulations.
5. Continue cooperation with the industry on joint
correlation projects. Pursue standardization of procedures,
protocols, and equipment between laboratories as much as is
required.
Ill. Summary of Relief Provided/Anticipated Environmental
Impacts
The above recommendations provide the following specific
relief to the industry, at the following environmental costs:
1. A one-year deferral in transient testing
requirements and applicable emission standards has been
provided to the gasoline engine industry. This is due entirely
to leadtime requirements and the revised 1984 emission
regulations. The extra year will provide greater time for
engine development and facility/test procedure improvement} it
will also result in another year's production of relatively
uncontrolled gasoline engines.
2. A one-year deferral in transient testing
requirements will afford the diesel engine manufacturers much
flexibility, and will allow them to spread out recertification
of their product lines over two model years. some engine
families will continue to exceed the transient HC standard of
1.3 g/BHP-hr for this single year, although the environmental
impact will be minimized by the development work which will
bring them into compliance by 1985.
3. Specific technical amendments to the test procedure
(not addressed here) have the potential to significantly reduce
yearly operating costs. Most can be made with no adverse
impact on engine emissions. Others (e.g., cold start
requirements) will need to be carefully reexamined since
environmental impacts are possible.
4. The technical issue of test cycle selection
addresses the manufacturers' technical concerns, but is
anticipated to have no economic effects either way. if
equivalently stringent standards accompany the optional test
cycles, no adverse environmental impacts should result from
their acceptance.
B-78
-------
APPENDICES
B-79
-------
APPENDIX 1
REQUEST FOR INDUSTRY COMMENTS ON THE
TRANSIENT TESTING REQUIREMENTS (46 FR 31677, June 17, 1981)
"Study of the 1984 Heavy-puty Truck Requirements
Regulations for 1984 ana later model year heavy-duty
engine emissions of HC and CO were promulgated on January
21, 1981 (45 FR 4136). These regulations implemented a
broad range of new provisions for heavy-duti engines. One
of the key provisions of the regulations was EPA's
adoption of a transient engine test procedure to replace
the current steady-state test as being more representative
of actual truck use.
Since heavy-duty engine manufacturers currently are
conducting transient test programs, this study will survey
manufacturers' progress to date in developing the
transient testing capability needed to meet the 1984
requirements for implementation of the new test
procedure. The results of this survey will be used to
evaluate whether there is any need to revise those
requirements....
...The particular areas in which EPA requests
information are as follows:
a. Please identify your needs for transient
testing facilities for 1984-85. These needs should be
based upon EPA's announced intention to delay
implementation of selective Enforcement Auditing for two
years. Include an identification of the number of engine
families you plan to certify.
b. Please describe the status of your transient
facilities. Identify how many cells you now have and when
they became operational plus how many are currently under
construction and expected completion dates for those.
Describe the equipment complement of your test cells.
c. Please describe any difficulties you are
experiencing in developing your testing capabilities. Are
you having problems locating vendors to supply necessary
equipment? What are delivery times associated with key
equipment items? Have economic conditions led you to
either delay or cancel purci..."\s of the necessary
facilities and equipment? What is your assessment of
remaining leadtime for the 1984-85 model years?
d. Please describe the economic impact you are
experiencing in developing transient test capability.
Itemize the cost of items in your test facilities. Have
B-80
-------
APl-2
all outstanding equipment items been purchased, and what
would be the impact of a one or two-year delay, in the
required implementation date? How are you financing the
required investments and what do you consider to be your
overall cost of capital? What effect, if any, have the
transient test costs had on your future product plans for
the heavy-duty market? Please provide a detailed
calculation of cost impact on a per-engine basis.
e. Please provide any information you have on
transient versus steady-state emissions for regulated
pollutants. Have you attempted to establish a
relationship on either a family-by-family basis or a
product line basis? Please provide any data you have
developed.
f. For aiesel engine manufacturers, have you
done, or are you doing, work to develop transient test
capability on eddy-current dynamometers? Please describe
your effort and any results.
g. Have you identified any problems with the
transient test itself, such as problems with repeatability
of test results? please submit supporting data.
h. Are there specific modifications that could be
made to the test to make it more representative of real
world conditions and manner of use?"
B-81
-------
APPENDIX 2
SOLICITATION OF INDUSTRY COMMENTS AS
PART OF THE RULEMAKING FOR REVISED HDE GASEOUS
EMISSION STANDARDS (47 FR 1642, JANUARY 13, 1982)
"As mentioned previously, a new HDE test procedure
and revised useful life requirements for LDTs and HDEs are
both effective beginning in the 1984 model year. As a
part of the Administration's regulatory relief program,
EPA has committed to a study of the issues pertinent to
these provisions and has formally solicited public comment
(46 FR 31677, June 17, 1981) (public docket A-81-20). The
period to submit information and comment on these issues
closed November 1, 1981, for the HDE test procedure ana
closed December 1, 1981, for the useful life provisions.
EPA recognizes that the transient test and useful
life provisions play a key role in the overall emission
control program. They directly affect both the
manufacturers' compliance strategies and the feasibility
of the emission standards. These issues are especially
important in this rulemaking and cannot be completely
separated from the proposed revised emission standards for
HDEs. Given the importance of these provisions in this
rulemaking, docket A-81-20 is herein incorporated by-
reference.
EPA is preserving the option to modify the HDE
transient test and useful life provisions as an integral
part of the final rulemaking process for this proposed
rule. Commenters asserting that changes are needed to
make the transient test more accurate or precise are urged
to suggest specific amendments where possible. to the
extent that issues raised in this rulemaking might elicit
additional comment for either the transient test or useful
life studies, EPA will accept that comment in conjunction
with this rulemaking....
Concerning the transient test, this procedure has
been the subject of extensive discussion between EPA and
the regulated industry since its promulgation in December
of 1979. EPA has visited several transient test
facilities and met with manufacturers to discuss their
concerns about the transient test procedure. EPA has also
received a great deal of correspondence on this topic
primarily from the engine manu^*nrers and their trade
associations. EPA remains open to additional substantive
comment on the transient test procedure and will fully
consider any new information which identifies significant
defects in either the method used to develop the transient
test or the transient test procedure itself.
B-82
-------
AP2-2
EPA has maintained that tighter standards under the
steadj-state test would provide little assurance that
appreciable further emission reductions would be achieved
by engines in actual use. Those who assert that the
transient test should be deferred should also address
whether an^ significant emission reductions beyond current
levels can be attained under the existing steady-state
test procedure, and if so, what standards would be
appropriate during any period of deferral. if the need
for deferral of the 1984 implementation date for the
transient test is established during this rulemaking, EPA
will consider all comments on the appropriateness of
interim standards; EPA will then make a final
determination of whether it would be rrtore appropriate to
adopt interim standards or to retain the pre-1984
standards with the steady-state test procedure during any
period of deferral."
B-83
-------
Appendix 3
List of Pertinent References
Report
Number Report/Summary
1 "Truck Driving Pattern and Use Survey, Phase II,
Implementation Plan," by Wilbur Smith and
Associates, May 7, 1973.
This report outlines a sampling and
instrumentation plan by which on^road heavy-duty
engine operational parameters can be recorded.
2 "Heavy-Duty Vehicle Driving Pattern and Use
Survey, Final Report, Part I, New York City,"
Report No. APT.D-1523, by Wilbur Smith and
Associates, May, 1973.
This report characterizes usage patterns and
population data for heavy-duty trucks in New York
City.
3 "Heavy-Duty Driving Pattern and Use Survey: Part
II - Los Angeles Basin Pinal Report," Report No.
EPA-460/ 3-75-005, by Wilbur Smith and
Associates, February, 1974.
This report characterizes usage patterns and
population data for heavy-duty trucks in Los
Angeles.
4 "Engine Horsepower Modeling for Diesel Engines,"
EPA Technical Report No. HDV 76-03. by C. France,
October, 1976.
This report summarizes the methodology used in
deriving horsepower models for gasoline engines
used in the CAPE-21 study.
6 "Truck Driving Pattern and Use Survey, Phase II,
Final Report, Part I," Report No.
EPA-460/3-77-009, by Wilbur Smith and Associates,
June, 1977.
This report summarize0 the sampling plan,
instrumentation, and data collected in the New
York phase of CAPE-21.
B-84
-------
AP3-2
Report/summary
"Truck Driving Pattern and Use Survey, Phase II,
Final Report Part II, Los Angeles," EPA Technical
Report NO. HDV 78-03, by L. Higdon, May, 1978.
This report summarizes the sampling plan,
instrumentation/ and the data collected in the
Los Angeles phase of CAPE-21.
"Analysis of CAPE-21 Horsepower Models," by
Systems Control, Inc., July, 1978.
In this report to the MVMA, the horsepower models
used in CAPE-21 were investigated and checked for
their validity.
"Heavy-Duty Vehicle Cycle Development," Technical
Report No. EPA 460/3-78-008, by Malcolm Smith,
July, 1978„
This report summarizes the data editing, data
manipulation, engine parameter models used, and
the overall statistical methodology used in
generating heavy-duty engine and chassis
dynamometer test cycles.
"Category selection for Transient Heavy-Duty
Chassis and Engine Cycles," EPA Technical Report
No. HDV-78-01, by C. France, May, 1978.
This report summarizes the methodology and
statistical comparative procedures used to
meaningfully combine truck categories to simplify
the CAPE-21 data base.
"Analysis of Hot/Cold Cycle Requirements for
Heavy-Duty Vehicles," EPA Technical Report No.
HDV-78-05, by C. France, July, 1978.
This report analyzes the need for separate cold
cycles for heavy-duty emission testing; it
extrapolates the amount of cold operation present
in the CAPE-21 data base.
"Selection of Transient Cycles for Heavy-Duty
Engines," EPA Technical Report No. HDV-78-04, by
C. France, August, 1978.
This report summarizes the statistical
methodology used in selecting the final test
cycles, from the several cycles generated in
Report No. 9.
B-85
-------
Report
Number
AP3-3
Report/Summary
13 "Transient Cycle Arrangement for Heavy-Duty
Engine and Chassis Emission Testing," EPA
Technical Report No. HDV-78-04, by C. France,
August, 1978.
This report summarizes the final analysis used in
arranging the cycles selected in Report No. 12
for the transient certification test cycles, and
also selects the final cold/hot weighting factors.
14 "1969 Heavy-Duty Engine Baseline Program and 1983
Emission Standards Development," EPA Technical
Report, by T. Cox, G. Passavant, and L. Ragsdale,
May, 1979.
This report summarizes the baseline test program
from which the transient standards were derived,
and summarizes experience and technical
discoveries gained in an actual transient test
program.
15 Summary and Analysis of Comments to the NPRM:
"1983 and Later Model Year Heavy-Duty Engines,
Proposed Gaseous Emission Regulations," December,
1979.
This document discusses the manufacturers'
concerns about the transient test, and places the
transient test in its technical and historical
perspective.
16 "Heavy-Duty Gasoline Engine Emission Sensitivity
to Variations in the 1984 Federal Test Cycle,"
Cox, T., SAE Paper No. 801370, October, 1980.
This report summarizes EPA testing on the
variability of transient test emissions
associated with dynamometer controller
calibration, and gives a detailed description of
the transient test procedure with respect to
emissions.
17 "Emissions from Trucks by Chassis Version of 1.983
Transient Procedure," Dietzman, H., et al., SAE
Paper No. 801371, October, 1980.
This report summarizes truck emissions generated
on a chassis dynamometer when exercised over
EPA's transient chassis cycle.
B-86
-------
Report
Number
AP3-4
Report/Summary
18 "Emissions From Heavy-Duty Engines Using the 1984
Transient Test procedures," volumes I and II,
Urban, C. and Martin, S., prepared for EPA under
Contract No. 68-03-2603, SwRl Report No.
EPA-460/3-81, July 1981.
This report by EPA's contractor summarizes their
transient testing baseline work and experience
with the transient test procedure in three
separate baseline studies..
19 "The Application of a Three-Way Conversion
Catalyst System to a Heavy-Duty Gasoline Engine,"
Hansel, J., et al., February, 1981.
This report summarizes work performed on a
prototype heavy duty gasoline engine, in which
statutory emissions standards for HC, CO, and NOx
were achieved by application of a closed-loop
three-way catalyst system, over the EPA transient
test.
20 "1972-73 Heavy-Duty Engine Baseline Program and
NOx Emission Standard Development," EPA Technical
Report No. EPA-AA-SDSB-81-01, Cox T., et al.,
March, 1981.
This report summarizes EPA's in-house transient
1972/73 baseline testing program and the
generation of statutory NOx emission standards.
21 "A Review of the Heavy-Duty Gasoline Engine
Certification Test Cycles," Final Report to the
MVMA, by Arthur D. Little, Inc., Cambridge,
April, 1981.
In this report to the MVMA, A.D. Little inc.
harshly criticized EPA's test cycles and
characterized them as unrepresentative.
22 "Evaluation of the Federal Test Procedure for
Heavy-Duty Diesel Engines for 1984 and the
Development of the Real Time Test cycle," by W.L.
Brown, Caterpillar Tractor Co., Research Report
88-29, File 18967, June 22, 1981.
In this report, Caterpillar's technical personnel
review the CAPE-21 data base and, using their own
methodology, develop the Real Time Test Cycle for
diesel engines.
B-87
-------
Report
Number
AP3-5
Report/summary
23 Draft Regulatory Support Document "Revised
Gaseous Emission Regulations For 1984 and Later
Model Year Light-Duty Trucks and Heavy-Duty
Engines," Chapter II, Section B., prepared by
OMSAPC, September, 1981.
This chapter of the regulatory support document
gives a detailed description of current
technology gasoline engine emission performance
over the transient test, and analyzes reductions
achievable without oxidation catalysts.
24 "Analysis of Arthur D. Little Inc.'s Report to
the MVMA: A Review of the Heavy-Duty Gasoline
Engine Certification Test Cycles,'" ECTD Staff
Report, January, 1982.
In this report, ECTD staff analyze A.D. Little's
Report to MVMA, and conclude that A.D. Little's
analysis was in large part incorrect,
unsubstantiated by emission data, and far too
strict with respect to achievable accuracies and
cycle representativeness.
25 "Status of EPA/EMA Cooperative Test Progam," A.
Azary, EPA Technical Report, No., IV-A-1, March,
1982.
This report summarizes data collected by EPA and
the HD diesel engine industry in the first
serious attempt to establish correlation between
laboratories with the transient test.
26 "Preliminary Report on Statistical Analysis of
Heavy-Duty Diesel Engine Emissions Data," N.J.
Barsic, prepared for the Engine Manufacturers
Association/Technology and Methods Subcommittee.
The report analyzes emission data collected in
the EPA/EMA Round Robin correlation project,
using several statistical techniques.
B-88
-------
APPENDIX 4
ECTD's Analysis of A.D. Little's
Report to the MVMA
B-89
-------
ECTD Staff Report:
Analysis of Arthur D. Little inc's Report to the MVMA:
"A Review of the Heavy-Dutj Gasoline Engine
Certification Test Cycles"
January 1902
B-90
-------
Table of Contents
Page
I. Introduction 1
II. overview 1
A. Keavy-Duty Engine Test Procedures 1
B. EPA's Test Cycle Development 3
Methodology
III. Summary: ECTD's Analysis of the 4
A.D. Little Report
References 7
Appendix 9
B-91
-------
I. Introduction
On January 21,,1980, EPA promulgated Final Rules for the
control of gaseous emissions from 1984 and later model year
heavy-duty gasoline engines. A major component of these rules
was the adoption of a new transient emission test procedure.
The intent was to model over-the-road operation of an "average"
truck as closely as a single laboratory test procedure could,
thereby allowing over-the-road exhaust emissions to be
predicted with much greater confidence than the older
steady-state tests.
In April, 1981, Arthur D. Little, Inc. of Cambridge,
Massachusetts published a report entitled "A Review of the
Heavy-Duty Gasoline Engine Certification Test Cycles,"
commissioned and funded by the Motor Vehicle Manufacturers
Association (MVMA) of the United States, Inc. A.D. Little,
Inc. concluded that the EPA test cycles (both chassis and
engine) were unrepresentative of the on-road data base, and was
harshly critical of EPA's methodology and conclusions.
In this report, the ECTD staff analyzes the A.D. Little
report and the validity of its conclusions. Section II of this
report presents an overview of the heavy-duty engine
environment, and EPA's cycle generation methodology. section
III summarizes ECTD's point-by-point analysis of the A.D.
Little report. This point-by-point analysis is attached as an
Appendix.
II. overview
A. Heavy-Duty Engine Test Procedures; Design Intent,
Predictive Utility, and Inherent Compromise
Heavy-duty gasoline engines are subjected to a wide range
of operating conditions in the field. 'They operate in many
different traffic patterns, over many different terrains,
carrying a wide variety of loads and in a wide variety of
vehicles with varying numbers of axles, many different
commercial applications, and different numbers of transmission
gears. To meet the market's wide variety of needs, many
different engine types and characteristics are manufactured.
The major challenge in test cycle development is to compress
this diversity into a workable laboratory test procedure.
Intuitively, on-road emissions from an "average"
heavy-duty engine are best predicted a test procedure which
contains all operational parameters in their average on-road
proportion and severity. This point is made with the
recognition that there is probably no such thinq as an
B-92
-------
-2-
"average" truck or truck-trip, given the wide range of engine
operation in the field. Any single test cycle designed to
model the real world will be only one of a virtually infinite
number of possible cycles. This reality is nothing new. it
was stressed by the MVMA in their original comments to the
proposal containing the new test cycle:
"...Considering the diversity of design and use of
heavy-duty gasoline powered vehicles, it is probable
that no "representative" driving cycle exists...."[1]
In short, a single test cycle for heavy-duty engines must be an
approximation at best. This is fundamentally identical to the
circumstances surrounding the light-duty vehicle FTP, or the
current heavy-duty engine steady state FTPS. Each is a short
laboratory test attempting to model highly diverse environments
and operational conditions.
An alternative to a single engine cycle would be a
multitude of test cycles, each being representative of a
specific application. This alternative would lead to very
complicated certification procedures, and the engine
manufacturers would most likely attempt to certify one engine
configuration for as many different test cycles as possible.
The improved representativeness of multiple engine cycles does
not justify the added burden for both EPA ana the engine
manufacturers.
Another alternative is to use a chassis (vehicle) test
cycle. However, since many engine manufacturers do not
manufacture trucks, they obviously prefer an engine test
procedure. Also, the diversity of truck types and the large
number of driveline options for each truck type, would
necessitate certification procedures of such a complicated
nature that this alternative would be unworkable. The
heavy-duty engine industry has in fact argued against this
approach for these reasons.
Steady-state emission tests have historically been
recognized as simple procedures in concept and in operation.
Their predictive utility, however, is poor primarily because
gasoline engines are easily optimized to perform well on the
limited number of test modes. Secondly, they ignore
significant emissions-producing operational "modes of gasoline
engines such as transient operation ar.d cold starting.
Emissions reductions achieved in the laboratory, as measured by
steady-state tests, are not achieved in actual on-roaa
operation. This was noted early in the certification process
for light-duty vehicles when the original 7-mode steady-state
procedure was replaced by the transient CVS-72 procedure in
B-93
-------
-3-
1972. As technology was applied to meet stricter emission*
standards, i.e., engine emission control systems were designed
and optimised for best performance on a specific test
procedure, steady-state test results digressed further and
further from real-world emissions. Similarly, many studies and
actual testing data revealed the same shortcomings in the
heavy-duty gasoline engine 9-mode procedure.[1] in short, the
9-mode represents an unreasonable approximation of the real
world. For this reason, in 1972 EPA began work on a new, more
representative test procedure.
B. EPA's Heavy-Duty Test Cycle Development Methodology
It was recognized that any revised heavy-duty engine test
procedure would need to be based on actual on-road truck
operational data. In October 1974, in cooperation with the
industry-staffed Coordinating Research Council-Air Pollution
Research Advisory Committee (CRC-APRAC), EPA began the largest
on-road truck study to date (the CAPE-21 project) in New York
City, and shortly thereafter, in Los Angeles. Actual
commercial trucks were instrumented and driven through their
normal daily rounds by their own drivers, while engine and
vehicle operational parameters (speed, power, temperature,
etc.) were recorded. These recordings of the daily operation
of the 110 trucks formed the data base from which the
representative test cycles were to be generated.
This data base was input into a computer and used to
generate thousands of candidate driving cycles. This cycle
generation was accomplished by the computer simulating an
engine driving down a probabilistic road: the transition from
the present speed/power mode to the next speed/power mode in
the resulting cycle was determined by the observed frequency of
occurrence of the same transition in the CAPE-21 data base.
Those transitions most frequently observed in the field were
those which occur most frequently in the test cycles. The
generated cycles were then screened using statistical tests and
engineering judgment; those closest to the CAPE-21 data base
were selected as the final cycles.
Separate cycles were generated for each subset of the data
base determined to be sufficiently different to warrant its own
cycle. Practical limits to the number of subsets exist: each
division of the data base reduces the amount of data available
for subset cycle generation, whi."1' also lengthening the
eventual emission test. Possible subsets included geographic
factors, (New York vs. Los Angeles), road type (freeway vs.
non-freeway), engine temperature (cold vs. warm vs. hot
engine), engine type (gasoline vs. diesel), vehicle type
(2-axle vehicle vs. 3-axle vehicle,) etc. EPA concluded that
B-94
-------
-4-
only Los Angeles vs. New York, freeway vs. non-freeway, and
gasoline vs. diesel were significantly different enough to
warrant separate cycles, thereby creating 4 gasoline and 4
diesel engine cycles. (Four chassis dynamometer cycles were
also generated, but these have yet to be used in any rulemaking
concerning exhaust emissions.) The subsets chosen represented
EPA's judgment of the best available compromise.
These cycles (hereafter referred to as subcycles) were
then arranged end-to-end into a single cycle, and baseline
emission testing per the requirements of the 1977 Amendments to
the Clean Air Act (CAA) of 1970 was begun. Baseline emission
levels and emission standards were generated based upon the
transient test. The actual test procedure consists of the
cycle being run tv.ice—first after the engine has reached
equilibrium at ambient air temperature (the "cold" start),
followed by a 20-minute pause (hot soak), then a rerun of the
cycle (hot start). Emissions from both cycles are then
weighted (1/7 from the cold and 6/7 from the hot) into a
composite test result.*
Throughout the cycle and standard development process, EPA
documented and. published the results ana methodologies used.
These publications are listed as References 2-15, and form the
technical basis for the cycles.
Throughout the cycle development process, the end product
was always envisioned to be a workable and representative test
procedure. In the development of any test procedure, however,
there is an inherent tradeoff between procedure simplicity and
procedure representativeness. Arguments were made by A.D.
Little, Inc. that the compromises taken and the underlying
methodologies used in EPA's cycle development program were in
many cases inappropriate.
Ill. summary: ECTD's Analysis of A.D. Little, inc.'s Report to
the MVMA
General Conclusions
1. A.D. Little's criticisms of EPA's cycle development
methodology are based upon statistics and conjecture. no
emission data are presented or referenced to substantiate any
claims of engine or chassis cycle unrepresentativeness.
This is identical in methodology to the light-duty Federal
Test Procedure (FTP), with the exception of the numerical
value of the cold/hot weighting factors. The light-duty
weighting factors are .43 for the cold cycle and .57 for
the hot.
B-95
-------
-5-
2. The majority of A.D. Little's analysis of test cycle
validation is incorrect, being based upon incorrectly applied
statistical tests and incorrectly chosen comparative parameters.
3. A.D. Little's statistical analysis holds EPA to
unreasonable and impractical standards of accuracy. A notable
example is its analysis of truck sample size in which it argues
that in some cases over a thousand trucks are necessary to
accurately predict the percentage of operation in some
subcycles. A.D. Little's harsh criticism of the weakness in
EPA's methodology neglects to mention that the weakness was not
imposed by EPA error, but rather by fundamental limitations
which the input data placed upon the applicable statistical
techniques, and by the compromises required to compress the
vast data base into a workable laboratory test procedure.
4. A.D. Little's arguments were not only narrow in
focus (neglecting practicalities and ignoring emissions), they
were entirely negative. No single alternative to EPA's
methodology was advanced. No evidence was presented that EPA
neglected a more representative approach in lieu of the
methodology chosen. No specific suggestions were made about
how the cycles could be improved.
5. EPA sees nothing in the A.D. Little report to
warrant changing the conclusion that the engine cycles are
representative of the data base. The implications from all
emission data collected to date on the engine cycles are that
emissions are not exaggerated by the Monte Carlo technique.
Changes to the engine cycles to improve the agreement of their
operational parameter summary percentages with those of the
CAPE-21 data cause negligible or minor (less than 5 percent)
changes in measured emissions. This is true even for the most
operationally sensitive modes and emissions evaluated.
6. EPA has collected no emission data on the current
chassis cycle, the criticism of which comprised the majority of
A.D. Little's report. The chassis cycle has not been, nor is
it intended to be, used in any exhaust emissions certification
test procedures. A.D. Little's statistical arguments of
chassis cycle unrepresentativeness are generally unconvincing;
however, they are correct in pointing out that the percentage
cruise operation in the overall chassis cycle is somewhat less
than that indicated by CAPE-21. whether this represents a
measurable difference in emissions is unknown.
7. A.D. Little's statements concerning combustion
engine emissions are based upon a brief literature review, and
not upon actual "hands-on" experience with emission testing or
emissions control. A.D. Little, therefore, attributes many
B-96
-------
-6-
engine parameters {e.g., second-by-second thermal history,
cylinder wall temperaturer etc.) with larger than life emission
significance, whereas only a few {e.g., vehicle inertia weight,
air/fuel "ratio, frequency and degree of transient and full
power fuel enrichment, cold-start effects) will in actuality
dominate the results of a laboratory emissions test. as
substantiated by EPA's engine cycle emission data, A.D. Little
overestimates the emissions sensitivity of the test cycles to
the cycle generation methodology. Note that it was a highly
conjectural over estiiuat Ion, being based upon absolutely no
emission data whatsoever.
8. Finally, EPA has always maintained that the cycles
and test procedures are not perfect ana are subject to
modifications, if actual emission testing experience indicates
that modifications are warranted. A.d. Little and the
heavy-duty gasoline engine industry it represents have failed
to present such evidence, and failed to present any specific
alternative to the specific parts of the cycles or procedures
with which they disagree.
B-97
-------
-7-
References
1. Summary and Analysis of Comments to the nprm: "198 3
and Later: Model Year Heavy-Duty Engines, proposed Gaseous
Emission Reflations," U.S. EPA, OANR, OMS, ECTD, SDSB,
December 1979.
2. "Truck Driving Pattern and use Survey, Phase II,
Implementation Plan," William Smith and Associates, May 7, 1973.
3. "Heavy-Duty Vehicle Driving Pattern and use survey,
Final Report, Part I, New York City," Report no. APT. D-1523,
William Smith and Associates, May 1S73.
4. "Heavy-Duty Driving Pattern and use survey: Part II
Los Angeles Basin Final Report," Report No. EPA-460/3-75-005,
William smith and Associates, February 1974.
5. "Engine Horsepower Modeling for Diesel Engines,"
U.S. EPA, OANR, OMS, ECTD, SDSB, France, C., Report No. HDV
76-03, October 1976.
6. "Engine Horsepower Modeling for Gasoline Engines,"
U.S. EPA, OANR, OMS, ECTD, SDSB, Higdon, L., Report No. HDV
76-04, December 1976.
7. "Truck Driving Pattern and use Survey, Phase II,
Final Report, Part I," Report no. EPA-460/3-77-009, William
Smith and Associates, June 1977.
8. "Truck Driving Pattern and use Survey, Phase II,
Final Report Part II, Los Angeles," U.S. EPA, OANR, OMS, ECTD,
SDSB, Higdon, L., Report No. HDV 78-03, May 1978,,
9. "Analysis of CAPE-21 Horsepower Models," Systems
Control, inc., July 1978.
10. "Heavy-Duty Vehicle Cycle Development," U.S. EPA,
OANR, OMS, ECTD, SDSB, Smith, Malcolm, Report No. EPA
460/3-78-008, July 1978.
11. "Category Selection for Transient Heavy-Duty chassis
and Engine Cycles," U.S. EPA, OANR, OMS, ECTD, SDSB, France,
C., Report No. HDV 78-01, May 1978.
12. "Analysis of Hot/Cold Cycle Requirements for
Heavy-outi Vehicles," U.S. EPA, OANR, OMS, ECTD, SDSB, France,
C., Report No. HDV 78-05, July 1978.
B-98
-------
-8-
References (cont'd)
13. "selection of Transient Cycles for Heav^-Duty
Engines," U.S. EPA, OANR, OMS, ECTD, SDSB, Wysor, T., and
France, Report no. HDV 77-01, November 1977.
14. "Transient Cycle Arrangement for Heavy-Duty Engine
and Chassis Emission Testing," U.S. EPA, OANR, OMS, ECTD, SDSB,
France, C., Report no. HLV 78-04, August 1978.
15. "1969 Heavy-Duty Engine Baseline Program and 1983
Emission Standard Development," U.S. EPA, OANR, OMS, ECTD,
SDSB, Cox, T., Passavant, G., and Ragsdale, L., Report No.
EPA-AA-SDSB-79-23, May 1979.
16. "1972-73 Heavy-Duty Engine Baseline Program and NOx
Emission Standard Development," U.S. EPA, OANR, OMS, ECTD,
SDSB, Cox, T., et al., Report No. EPA-AA-SDSB-81-01, March 1981.
17. "Heavy-Duty Gasoline Engine Emission Sensitivity to
Variations on the 1984 Federal rest Cycle," Cox, T., SAE Paper
No. 801370, October 1980.
18. "The Application of a Three-Way Conversion Catalyst
System, to a Heavy-Duty Gasoline Engine," Hansel, J., et al.,
SAE Paper No. 810085, February 1981.
19. Draft Regulatory Support Document "Revised Gaseous
Emission Regulations For 1984 and Later Model Year Light-Duty
Trucks and Heavy-Duty Engines," Chapter II, section B, Prepared
by OMS, September 1981.
20. "Emissions From Heavy-Duty Engines using the 1984
Transient Test Procedures," Volumes I and II, urban, C. ana
Martin, S., Prepared for EPA under Contract No. 68-03-2603,
SwRI Report No. EPA-460/3-81.
21. "Emissions from Trucks by Chassis Version of 1983
Transient Procedure," Dietaman, H., et al., SAE Paper No.
801371, October 198D.
22. Introduction to Applied Statistics, A Course Outline
of statistics 402, Rothman, E.D., University of Michigan, July
1979.
23. "A Review of the Heavy-Duty Gasoline Engine
Certification Test Cycles," Final Report to the MVMA, by Arthur
D. Little, Inc., Cambridge, April 1981.
24. Introduction to Statistical Analysis, Third Edition,
Dixon, Wilfred J., Massey, Frank J., Jr., 1969.
B-99
-------
Appendix
Note: The issues and subject matter presented and
discussed in this Appendix are listed and numbered as they
appeared in the A.D. Little report to permit easy
referencing. A summary of the A.D. Little analysis for
each point is presented and followed by EPA's analysis.
Suffixes have been added to the A.D. Little numbering
system to differentiate between A.D. Little and EPA
analyses. The suffixes "L" and "E" will be used for A.D.
Little and EPA analyses, respectively.
1.0 Introduction
2.0 Summary and Conclusions
3.0 Engine Operational Parameters
3.1 L — All engine operational parameters must be carried
through on the test cycle and simulated by a reasonable
duplication to truly represent over-the-road operation.
3.3 L — Engine speed, load, speed/load relationships,
speed/load history, and engine thermal history must be
manipulated in the laboratory in a manner which will duplicate
the over-the-road state of the engine, although it i'S
recognized that some compromises must be made in the test cycle
in order to make it a feasible, cost-effective procedure.
3.4 L — During transient operation, critical engine
variables may be at off-design or non-optimum operating points
for much of the time. It is not necessary to follow an exact
over-the-road speed or load schedule so long as load excursions
are representative (i.e., do not exceed the magnitude or rate
of those experienced on-road), and are present for a
representative portion of total operating time.
3.6 L -- Speed/load relationships (including dwell times
at constant load) must be maintained. An unrealistic operating
condition is a change in engine speed without a change in
engine torque.
3.7 L — Engine thermal history is important relative to
emissions. A representative test must include cold operation
(including operational differences of a cold engine), unless
cold emissions represent an insignificant portion of the total.
3.8 L — Since thermal conditions are influenced by
previous speeds and loads, it is important that engine
speed/load history be simulated with some accuracy, although
B-100
-------
AP-2
this requires knowledge (and/or assumptions) of individual
vehicle, driver, road, and trip characteristics. it should to
possible, however, to define representative characteristics of
speed/load history for inclusion in a test cycle.
3.9 L — Chassis dynamometer cycles are subject to the
same criteria with respect to vehicle speed, and engine thermal
history.
3.1-3.9 E — A.D. Little conducted what was essentially a
literature review on the subject of combustion engine
emissions. For the most part, A.D. Little's limited summary of
pertinent literature correctly identifies sources of tailpipe
emissions, but at no time does A.D. Little discriminate between
major, minor, ana insignificant sources. A.D. Little's Chapter
3.0, "Engine operational parameters" is an academic review
containing neither actual emission data to support the authors'
contentions, nor any evidence that A.D. Little has any
"harids-on" or other experience with emissions testing,
measurement, and control.
Since February of 1978, EPA's Ann Arbor facility has
performed nearly 1,000 transient emission tests on heavy-duty
gasoline engines and over 300 on heavy-duty diesels. This
testing encompasses engines and technologies which include
uncontrolled, current technology, and advanced technology
prototypes. other tests explored emissions sensitivity to
variations in engine, cycle, and test procedure parameters.
EPA has acquired a substantial data base and substantial
"hands-on" experience with both the transient test procedure
and the data generated therefrom.[15,16,17,18,19] The
Southwest Research Institute's Department of Emissions
Research, under contract to EPA, has also been running
transient emission tests since June of 1978 on both HD gasoline
and diesel engines. (See References 20 and 21.} The above
experience, plus 10 years of light-duty emissions testing
experience, gives EPA confidence in its ability to identify
important emission parameters.
A.D. Little has identified and selectively addressed those
emission parameters they believe most associated with EPA's
cycle development methodology, inferring that "errors' in cycle
development affect the emissions measured over the resultant
cycle. We shall defer most of our discussion until later in
the Appendix where the particulars of cycle development
methodology are more thoroughly discussed; however, several
points need to be made.
First of all, A.D. Little presents no emission data to
support its conclusions. EPA has always maintained that if a
B-101
-------
AP-3
cycle is representative of over-the-road operation, the exhaust
emissions measured over that cycle will also be
representative. This by no means implies that every on-road
characteristic is critical with respect to emissions. This is
also true with respect to the proportional time any on-road
characteristic is present in the cycle. The A.D. Little report
makes several claims of EPA cycle non-representativeness on the
basis of idle times being too short, high-power cruises being
too short, and other operational distribution differences
between the individual subcycles and the CAPE-21 data. A.D.
Little then makes sweeping conclusions of unrepresentativeness,
yet makes no inferences and presents no data on the emissions
impact of these discrepancies. The cycle is, after all, used in
an emissions test procedure, and the only reasonable basis for
concluding non-representative emissions generation is to
demonstrate significant changes in emissions arising from the
discrepancies in operational parameters. unsupported and
unquantifiea inferences arising from a casual review of
combustion emission literature is insufficient proof.
Secondly, the majority of A.D. Little's criticisms
throughout their report focuses on the chassis cycles, as
opposed to the engine cycles. This is surprising, considering
that all heavy-duty engine exhaust emission test procedures and
exhaust emission standards are based upon the engine
dynamometer cycles. We shall, nonetheless, address the
criticisms since the cycle generation methodologies were
similar.
Furthermore, A.D. Little has identified several sources of
emissions and related these to the effects of specific engine
operation, i.e., how the engine operation in a test procedure
can affect measured emissions. At no time, however, are the
relative magnitudes of each emission source discussed. A
satisfactory test procedure can exclude all minor and
insignificant sources, so long as it includes all dominant
sources. The most significant determinant of engine exhaust
emissions is air/fuel ratio, both overall and locally within
the cylinders. Air/fuel ratio changes continually in
over-the-road operation, both as a function of load and from
increases in load (transient enrichment from the accelerator
pump), full power operation (power valve enrichment), motoring,
and from temperature effects. The degree of enrichment and
nominal air/fuel ratio calibration are determined by the engine
designer in response to on-road powe*: 3,"r< driveability needs.
These three major sources of emissions: transient loads, full
power operation, and cold starting, must be properly
represented in the test procedure if predictable on-road
control is to be achieved. For the chassis cycle, the vehicle
inertia weight and road load horsepower calibration of the
B-102
-------
AP-4
dynamometer (which ace independent of the cycle used) are also
major determinants of overall measured emissions.
Finally, identification of the real issues in this test
procedure controversy is warranted. The emission standards
generated by the Clean Air Act Amendment methodology were
derived on the full transient test procedure. if a systematic
bias in emissions from the "real-world average level" were to
exist,* that bias has been carried through to the emission
standards. If the test procedure generates higher emissions
the standards will be comparably higher. Emission penalization
of a manufacturer as a result of the certification test
procedure is not the issue. Representativeness itself is not
the all-important issue the heavy-duty gasoline industry
claims. The industry was perfectly content to certify engines
on the considerably less representative 9-mode for the last
eleven years. This is not said to deemphasise
representativeness—great pains were taken during cycle
development to assure maximum representativeness in a practical
procedure, and most of this analysis addresses that issue.
This is to say, however, that the industry is not so much
interested in achieving a perfect cycle, but rather they are
most concerned with ease of compliance (in -ernis of cost and
technical effort) and possible impact of resulting control
strategies on engine performance and durability. For our part,
the ECTD staff considers a test procedure to be a tool to
direct emission control technology towards achieving in-use
emissions reductions. To do so, it must adequately represent
in-use operation. The true test procedure controversy arises
when those modes which EPA considers most significant with
respect to emissions are also those which the industry
considers most problematic to control.
The transient test procedure as promulgated by EPA is by
no means unalterable. The regulatory support documents from
the Final Rulemaking openly stated that "...there should be no
problem in modifying procedures, changing validation and
dynamometer specifications, and making any changes deemed sound
if actual testing experience suggests that additional changes
to the test procedure are warranted. "11] The laboratory test
procedure was developed primarily at EPA's Ann Arbor facility,
using equipment and procedures which were technically
sufficient, but not necessarily the state-of-the-art or most
cost effective. Given the diversity of product lines,
This is by no means conceded. All data to date implies
that such an offset is unlikely, or quite small (see the
remainder of the Appendix).
B-103
-------
AP-5
laboratory equipment alternatives, and technical experience ana
opinion within the industry, there comes a point when
unilateral development of an emission test procedure must stop
ana a consensus with the industry reached. Attaining a
technical consensus requires constructive technical feedback,
including emission data, from the industry. This has not
occurred. to date, all feedback from the heavy-duty gasoline
industry, as typified by the A.D. Little report, has been
negative, characterising the transient test procedure as in
toto unrepresentative and technically deficient. EPA has been
presented an all or nothing proposition. The gasoline engine
industry, however, has not presented any emission data to
substantiate claims of unrepresentativeness, nor have they
suggested specific improvements or alternatives to the present
transient procedure, or even identified acceptable alternative
methods of test procedure development. The weight of present
technical evidence indicates that the transient test cycles and
procedure are by and large sufficiently representative of
in-use trucks. This by no means indicates that the cycles or
procedure are perfect, and are not capable of further
improvement.
4.0 Assumptions and Methodology Used in the Development of
Test Cycles
4.2 L — The Markov model, a second-by-second technique
which assumes that the likelihood of going to the next speed
and load depends only upon the present speed and load, neglects
speed/load history of more than one second. This approach did
not capture the smooth cruise, acceleration, ana deceleration
operation found over-the-road. in short, erratic cycles with
unrepresentative speed/load patterns were generated. The
transitions in speed and lead were random; holding times at
given speeds and loads were "memoryless" (i.e., were
independent of operation more than one second in the past), and
long-term cruising (i.e., steady-state operation) were not
included.
4.2 E — Several of A.D. Little's observations about the
Markov (Monte Carlo) model are correct. Characteristic
speed/load shift patterns do not appear in the cycles. Holding
times and transitions were independent of operation more than
one second in the past. Extended steady-state operation was
not included in the cycle. These facts are more easily
understood when one realises that th° Markov model makes no
assumptions about the data base; transitions from one
speed/load state in the cycle to the next occurs with the same
probability as the identical transition observed, in the
aggregate data base, i.e., all trucks and all truck trips
B-104
-------
AP-6
combined. implicit in the inclusion of a given speed/load
shift pattern into a cycle ate several assumptions, e.g., road,
trip, traffic, and driver characteristics, number and speed
ratios of transmission gears, accelerating or decelerating
operation, etc. Given the large number of trucks and truck
trips and the extreme variety of engine operations accumulated
on-the-road, the better statistical model is that which held
the fewest preconceived notions and made the fewest assumptions
about the data.
As mentioned earlier, the priia-e determinants of
representativeness are emissions measured over the cycle. A.D.
Little's analysis of thermal history implied that
uncharacteristic speed/load shift patterns generated
unrepresentative emissions. Several manufacturers have argued
that "erratic" and continually changing cycles, the "nervous
foot" syndrome, contribute to excessive transient enrichment
and, therefore, high HC and CO emissions on gasoline engines.
It has also been argued that unrealistic speed/load transitions
(e.g., a change in speed without a change in load) have been
incorporated into the cycle, thereby creating unrealistic
emissions. To date, no emission data have been advanced to
support these claims.
In reviewing the A. D. Little report, one of our major
criticisms is that it fails to differentiate between major and
minor emission sources. A case in point is its discussion of
engine thermal history. It is an established fact that cold
engine HC emissions are substantially higher than those of a
warm engine. The major reason for this is the effect of a cold
inlet manifold on the ability of gasoline to vaporize uniformly
in the inlet air. Cold manifold walls cause condensation of a
large portion of the gasoline. Cold enrichment by the choke
mechanism attempts to alleviate driveability problems, but
consequently results in excessive unbucned fuel etriissions. as
the intake manifold warms up, the problem gradually
disappears. A.D. Little has argued, however, that the Monte
Carlo technique is unrepresentative because it neglects engine
thermal history of more than one second, claiming
unrepresentative emissions because of the unrepresentative
second-by-second changes in cylinder wall temperature. But how
large are these potential emission differences? Are they
measurable, and if so, significant? A.D. Little cites
"experience"* to indicate that several seconds are required for
emissions at any given state to stabilize, yet fails to mention
that such time delays are primarily results of mixing and gas
residence characteristics of the exhaust stream reacting to the
change in operation, and of the finite response limitations of
the downstream emissions analyser. Given the fact that there
are almost 17 combustion events per second per cylinder in an
A personal conversation with a single industry engineer.
B-105
-------
AP-7
engine running at 2,000 rpm, one could argue that cylinder wall
temperature is much more responsive to second-by-second
operational changes than the downstream emissions analyzer. In
short, the effect of the Monte Carlo technique on engine
temperature-dependent emissions is most likely unmeasurable.
Intake manifold temperature, the major source of cold
emissions, increases slowly (over a period of several minutes)
from a cold start and is certainly unaffected by the Monte
Carlo technique.
This judgment, plus the observation that the Monte Carlo
"nervous-foot" characteristic does not significantly bias
emissions measured over the transient cycle, is implied from
emission data generated at EPA's Ann Arbor laboratory.
First of all, Table A-l presents comparative 9-mode and
transient emission data on fifteen 1969 model year heavy-duty
gasoline engines. These engines were uncontrolled with respect
to emissions. Although absolute correlation between the two
test procedures is unlikely, primarily because each procedure
exercises different emissions-producing mechanisms within the
engine, the average levels of emissions for both procedures are
highly illustrative. If a systematic, gross overstatement of
HC and CO emissions arising from throttle activity (or any
other parameter) were present, it should show up as a large
difference between uncontrolled transient and 9-mode
emissions. In fact, transient HC emissions are less than
9-mode, despite the fact that both cold-start and
transient-throttle operation are included in the transient
test. (Closed throttle motoring is somewhat less controlled.)
Transient CO is 24.7 percent higher, but we note that the
9-mode test does not include full-power operation, while the
transient does, and full-power Los Angeles Freeway (LAF)
operation accounts for 68.9 percent (from Table a-8 where the
contribution of the cold-start LAF is 12.6 percent added to the
hot-start LAF contribution of 56.3 percent) of CO measured over
the transient cycle on the highest emitting current technology
engines. In short, although the two test procedures are not
identically comparable, one would expect gross emission
differences to exist if the "nervous foot," or any other aspect
of the transient test, were grossly inappropriate.
Secondly, Reference 17 presented data detailing the effect
of dynamic throttle activity on transient emission results.
Throttle activity was first varied e^'" *-he overall effect on
emissions measured. The data show that past a certain
threshhold of throttle activity, HC and CO emissions increase
dramatically, as would be expected when transient enrichment
becomes excessive. Below that threshold, emissions are
essentially constant, even as throttle activity is decreased
B-106
-------
AP-8
Table a-1
Uncontrolled Emission Levels from Fifteen 1969
Model Year HP Gasoline Engines (in grams per borsepower~hour)
HC CO NOX
Engine
Transient
9-Kode
Transient
9-Mode
Transient
9-Mode
IHC 304
11.22
17 .0
127.8
119.9
6.70
6.04
Chrysler
318
7.96
8.31
87.0
50.5
7.60
9.75
IHC 345
7.12
7.84
75.5
47.0
6.46
9 .33
GM 350
6.21
7.25
126.1
131.5
5.36
4.21
Ford 300
7.81
9.79
233.4
235 .8
¦4 .91
4.35
IHC 345
6.41
8.99
94 .Q
39 .0
5.59
6.19
GM 366
8.59
9.25
187.9
122.3
5.32
5.65
Ford 361
14.12
10.19
22 8.4
180.0
5.43
4.83
Ford 360
7 .96
11.29
132.2
97.4
6.63
6.17
GH 292
a.54
8.52
172.8
95.1
5.14
7.16
Chrysler
318
8.82
9. 74
144.3
112.3
7.54
&. 25
Ford 361
9.57
10 .03
197.6
160.2
5.09
6.17
Ford 360
5.92
6.49
75.3
43,2
6 .88
6 .35
GM 350
e .64
9.81
150 .4
148 .2
4.58
3.96
Chrysler
361
12.63
10.27
168.7
133.5
6.CI
5.94
Mean:
8.77
9.65
146.8
117.7
5.95
6.30
STD Dev:
2.32
2.39
51.4
52.0
.96
1.73
% Difference,
Transient
; from
9-Mode:
-9.1%
+24.7%
-5.6%
B-107
-------
AP-9
well below that activity normally used in the transient
procedure, i.e., the cycle was mechanically smoothened. If the
test procedure were inherently overactive, one would expect
emissions to continue decreasing as throttle activity was
further dampened.
In summary, the Monte Carlo technique did not produce
characteristic speed/load shift patterns. However, the cycle
that was produced generates emissions which are representative,
as implied by the data above, and was produced without major
assumptions about the data, and hence is a statistically better
model. Again, this is argued with a complete lack of emissions
evidence to the contrary.
Finally, both A.D. Little arid the industry have argued
that unrealistic modes of engine operation were included in the
cycle by the Monte Carlo technique. Many of these comments
have already been addressed in Reference 1, but A.D. Little
argued that an increase in engine speed without a change in
load is unrealistic. A.D. Little presented an illustration,
reproduced here as Figure A-l, depicting its version of
realistic engine acceleration. A.D. Little argued that changes
in engine speed without changes in engine torque were
unrepresentative. we dispute this unsubstantiated
generalization. Note that engine speed can remain constant
while torque increases as a truck begins to climb a bill and
more throttle is applied to maintain vehicle speed. Note also
in Figure A-l that at any time after the intitial change in
torque, the engine speed increases while the engine torque is
constant. (That component of engine torque affecting
acceleration gradually decreases to zero, while that arising
from increased vehicle aer odynairdc drag and rolling friction
increases by the same amount.) Engine accelerations during
motoring are also realistic, when one considers that truck
drivers frequently downshift while braking with the engine.
What is not maintained in the test cycle, as discussed above,
are the familiar patterns of gear shifting during accelerations
and decelerations. All second-by-second operation in the
cycles, however, are physically explainable. A.D. Little's
hypothesized generalizations of possible truck operation do not
reflect what was actually recorded in the CAPE-21 project.
4.3 L — EPA's methodology of forcing accelerations at the
beginning and decelerations at the end of each 5-minute cycle,
was mandated by deficiencies in the model and resulted
in further unrepresentative characteristics.
4.3 E — A.D. Little's criticism of accelerations at the
beginning of a subcycle and decelerations at the end of a
B-108
-------
AP-10
Time
Source: Arthur 0. Little, Inc.
FIGURE A-1 .RESPONSE OF ENGINE SPEED TO A STEP CHANGE IN
ENGINE TORQUE AS A FUNCTION OF TIME
B-109
-------
AP-11
subcycle is unreasonable. In order to implement a test
procedure that is simultaneously representative of actual
operation ana practical in terms of time required to complete
the procedure, assumptions and compromises must be made.
Otherwise, the test procedure would be of the same duration as
the average truck's working day.
In the issue at hand, EPA sees no alternative to beginning
a cycle with an acceleration and ending a cycle with a
deceleration. If a truck starts at idle and ends at idle (a
reasonable assumption), then an acceleration at the beginning
of operation and a deceleration at the end are inherently
reasonable assumptions.
A.D. Little then claimed that the "forced" accelerations
and decelerations caused the resulting times spent in cruise,
acceleration, deceleration, ana idle to be unrepresentative.
Table 4-1 in the A.D. Little report lists the phase percentages
(cruise, acceleration, etc.) for the CAPE-21 data and the
chassis subcycles (NYNF, LAF, etc.). A.D. Little concluded
that the differences were too large and not representative of
the data. However, these conclusions were based upon A.D.
Little's incorrect selection of comparative parameters (see
5.3.1 E). A.D. Little's conclusions are invalid. (Were they
valid, no indication was given whether they would result in
measurable emission discrepancies.)
It should also be noted that A.D. Little's analysis was
based on a rather narrow focus. The chassis cycle is the
combination of three subcycles, so it would be more logical to
compare the combined segment inputs to the overall cycle phase
percentages.
4 .4 L — A.D. Little argued that the number of trucks and
truck trips was too small to produce a test cycle with the
accuracy and precision implied by the EPA and "...which would
normally be expected."
4.4 E — A.D. Little concluded that the number of trucks
used in the CAPE-21 project was too small to accurately
estimate the true values of the following operational
parameters: 1) mean and median truck speed, engine rpm and
engine power, 2) percent acceleration, 3) percent cruise, 4)
percent deceleration, and 5) percent motoring. The criterion
used was the number of trucks nec "^.ry, at a 95 percent
confidence level, to be within +2 percent (absolute error) of
the true value, or a relative variation of ten percent of the
sample mean. EPA will limit its analysis to the number of
trucks necessary to estimate the true value.
B-110
-------
AP-12
A.D. Little calculated that over 1,00C trucks would be
necessary to achieve a 2 percent absolute error for the N.Y.
Non-Freeway Gasoline engine subcycle mean rpm at a 95 percent
confidence level. Clearly such a requirement is absurd,
providing no allowances for real-world testing budgets. By
reducing the confidence level to 9G percent and the absolute
error to 10 percent, the number of trucks is reduced to 30!
Since 27 trucks were used, three more trucks would be necessary
to meet the more reasonable criteria. Table A-2 lists the
number of operating parameters for which there were too few
trucks tested to meet the specified criteria for confidence
level and absolute error.
Three criteria levels were evaluated in Table a-2: 1) the
A.D. Little criteria of a 95 percent confidence level with a 2
percent absolute error (EPA's original target criteria), 2) a
95 percent confidence level with a 5 percer.t absolute error,
and 3) a 90 percent confidence level with a 10 percent
absolute error.
The calculations for the number of trucks required to
achieve each criteria level were done for the six test
subcycles A.D. Little used in Tables 4-3 through 4-8 of their
report* These cycles include the: 1) Los Angeles Freeway
(LAF) chassis cycle, 2) Los Angeles Non-Freeway (LANF) chassis
cycle, 3) LAF gasoline engine cycle rpm, 4) LANF gasoline
engine cycle rpm, 5) LAF gasoline engine cycle power, and 6)
LANF gasoline engine cycle power.
For each cycle, the number of trucks required to achieve
the criteria was calculated for each of the seven operating
parameters: 1) mean, 2) median, 3) percent acceleration, 4)
percent cruise, 5) percent deceleration, 6} percent notoring,
and 7) percent idle. This resulted in six cycles with seven
operating parameters pec cycle, or 42 operating parameters for
each criteria level.
Table A-2 lists the number of these 42 operating
parameters which did not have the minimum number of trucks
necessary to meet the specified criteria.
twenty-two of the 42 operating parameters did not have the
minimum number of trucks required to meet the stringent A.D.
Little criteria that the sample values be within +2 percent of
the true value at a 95 percent confidence level.
Calculations for a more reasonable criterion of a 5
percent error at the same 95 percent confidence level shows
that only five of the 42 operating parameters didn't have the
minimum number of trucks. As mentioned previously, with a 10
percent error at a 90 percent confidence level, only one
parameter didn't have a sufficient number of trucks.
B-lll
-------
AP-13
Table A-2
Number of Parameters Requiring
Additional Truck Tests
to Meet
Criteria
Confidence Level
90%
95% 95%
Absolute Error
10%
5% 2%
No. of Parameters
1
5 22
(42 possible)
B-112
-------
AP-14
Considering the costs associated with achieving marginal
reductions in absolute error, EPA cannot justify further
testing to increase sample size. CAPE-21 was the largest
on-road truck usage study ever conducted. It is EPA's judgment
that the number of trucks tested was adequa-e, and that A.D.
Little's criticism was unreasonable.
4.5.1 L -- EPA's statistical filter was characterized as
incomplete and inconsistently applied. EPA's cycle acceptance
criteria neglected the possibility of incorrectly accepting an
unrepresentative cycle as representative. Modal percentage
measures (percent acceleration, deceleration, cruise, idle)
were not included in the computer's statistical filter. The
Kolmogorov-Smirnov test (K-S test) was not a sufficient test
for comparing cycle and input matrices because the data base
was too sparse. in fact, since the distributions of
acceleration, deceleration and cruise vary with speed, the K-S
assumption of independent identically distributed observations
is incorrect. Finally, where generated cycles could not pass
EPA's admittedly strict acceptance criteria, the criteria were
relaxed to accept the cycle. Therefore, in every engine and
chassis test cycle, there is at least 1 percentage measure that
differs from its input matrix by more than EPA's 2 percent
criteria.
4.5.1 E — The statistical filter, the theories behind it,
and the issues surrounding it are technically complex. The
reader is referred to the appropriate reports[10,13] in which
the filter's strengths and weaknesses are explained in detail.
To address A.D. Little's criticism of the statistical
filter methodology, it is best to review the real practical
constraints on cycle generation.
Firstly, the cycles are intended to model an extremely
diverse population of trucks, as a case in point, Figure a-2
shows the means of all observed percent rpm's for each
individual LA gasoline truck (freeway) varying from 40 to 85
percent. The distribution of truck operational parameters
about these means, and throughout the aggregate data base, is
continuous and decidedly non-normal. This widely distributed,
nonsymmetrical, non-normal distributional characteristic
severely limits the statistical tests which can be used to
quantitatively compare the population with an empirical
sample. If a quantitative measure of representativeness is to
be made at all (as opposed to a subjective verification using
intuition and engineering judgment), only a non-parametric test
is available, of which the K-S test is an applicable example.
This fact is independent of the cycle generation technique.
B-113
-------
AP-15
Number
of
Trucks
Mean of Mean %RPMs
All 2 Axle Trucks
Mean of Mean %RPMs
Tractor Trailers
Mean %KPM
3 Axle Trucks
Time-Weighted Input Data
Mean %RPM
Cycle Mean %RPM
60 65
Mean %RPM
Figure A-2: Mean %RPM Distributions of CAPE-21 Trucks
(LA Freeway - Gasoline)
B-114
-------
AP-16
Secondly, the longer a laboratory emission test procedure
runis, the greater the cost, the time, and the overall test
burden to accomplish it. The longest test EPA can reasonably
impose is on the order of one hour in length. (Note that this
is comparable in length to the light-duty Federal Test
procedure (FTP) and the current heavy-duty steady-state tests.
Note also that one hour does not include time required for test
preparation* vfehicle/engine soaking, or data analysis.) Given
an overall test procedure consisting of two separate cycles
(one cold start, one hot start) and one interitieoiate soak
(pause) period, nominal test cycle length is roughly 20
minutes j. Within these 20 minutes all significant truck
operational parameters must then be simulated to obtain
reasonably represehtative results. EPA's evaluation of truck
operational differences[ll,14] indicated that a minimum of 3-4
discrete subcycles are necessary within those 20 minutes for
maximum representativeness. In short, 5-7 minutes is the
maximum practical length of individual subcycles. if this
length implies gross unrepresentativeness, why then are 1-hour
tests sufficiently representative for light-duty vehicles, and
heavy-duty engines certified on steady-state tests?
Two alternatives exist for cycle generation. Manual
selection of truck operational segments from the data base can
be done using intuition and engineering judgment, followed by
subsequent screening and validation by a statistical filter.
On the other hand, a computer can replace the subjective human
generator, bringing with it the advantages of infinitely faster
sampling and calculation capabilities. In addition, a computer
is capable of sampling the data base in a statisticallly
representative way using the Monte Carlo technique, i.e.;
transitions are probabalistically incorporated into the cycle
according to their observed frequency of occurrence in the
actual data base. subjectivity in sampling is eliminated#
hence our conclusions above that the Mbnte Carlo technique is a
statistically better means of cycle generation. Consider also
that the computerized Monte Carlo method was capable of
generating over 10,000 candidate cycles per computer run from a
massive second-by-second data base. (For the LAF chassis cycle
alone,* the CAPE-21 data base was comprised of 1,015,316
individual seconds of operation). The longer the Monte Carlo
technique is allowed to probabalistically sample the data base,
the better the resulting cycle approximates the input data, a
problem arises here because of the practical limitation on
laboratory test procedure length. Required cycle times of 5 to
29 minutes ate very short compared to the time required for the
Monte Carlo technique to sample the entire input data base.
The probability of any one 5-minute cycle being representative
Of the input data is small; large numbers of cycles need to be
generated to assure that some of them will in fact be
B-115
-------
AP-X7
representative. (All of the above is more thoroughly discussed
in Reference 10).
Selection of the few representative cycles from this
multitude of generated cycles is the function of the
statistical filter; as discussed above, the K-S (or other
non-paranietr ic) statistical test is the only quantitative
alternative. The K-S test, however, is not without weakness.
The K-S test, as used in EPA's statistical filter, predicts
Type I statistical errors, i.e., the probability that a
representative cycle is rejected as nonrepresentative. (As
noted in Reference 10, these are acceptable errors, but result
in more candidate cycles having to be generated.) Accepting an
unrepresentative cycle as representative is referred to as a
Type II error, as a means of minimizing this potential error,
EPA's methodology deliberately increased the probability of
rejecting representative cycles (Type I error). The rationale
for this was simply that if one's statistical tests are made
stringent enough to reject an increasing number of
representative cycles, the probability of accepting
unrepresentative cycles is decreased. In addition, target
percentage difference criteria between subcycle and input
operational parameters were also selected; these target
accuracies were characterized by A.D. Little as "very strict."
The higher required probabilities of Type I error resulted
in a larger number of cycles having to be generated so that at
least a few could pass all six separate k-S tests performed on
candidate cycles. As discussed in Reference 10, freeway cycles
were more problematic in passing the strict criteria,
presumably because of the inhibiting effect of 5-minute
subcycle length on comprehensive input data sampling on
freeway-type input. Type I probability criteria were in some
cases, relaxed to allow candidate freeway cycles to be accepted
for further screening at EPA.
Further screening was then conducted at EPA based upon a
comparative review of operational parameters (e.g., percent
acceleration, deceleration, idle, and cruise) for the input and
candidate subcycles. The candidate subcycles were ranked
numerically according to relative K-S test performance,
operational parameter closeness to the input data, and
subjective verifications of agreement of subcycle parameter
distributions with those of the input data. The best
all-around subcycle for each categor^ *"3 then selected.[13]
In summary, every possible effort was made to guarantee the
representativeness of the final cycles, given the practical
constraints of cycle length ana available quantitative
statistical tests.
B-116
-------
AP-18
The majority of A.D. Little's criticisiu of EPA's
statistical filter focused on "...its incompleteness and the
inconsistency with which the statistical criteria ace
applied," Relaxations of the Type I ana percentage difference
criteria were claimed by A.D. Little to reflect inconsistency
and to have resulted in the acceptance of non-representative
parameters for the final cycles. A.D. Little supports this
claim by referencing statistical tests elsewhere in its
analysis (5.2.1 and 5.3.4). A.D. Little's analysis is
incorrect. Their tests made the incorrect assumption that the
data were from a normally distributed population, and to
compound its error it used the wrong comparative summary
percentages for the input data (see 5,2.1 E, 5.3.1 E, and 5.3.4
E of this Appendix). secondly, A.D. Little's claim of filter
incompleteness is based upon the absence of percentage measures
for cruise, acceleration, and deceleration from the filter.
A.D. Little failed to point out, however, that these percentage
measures were indeed used as screening criteria by EPA in
selecting the best subcycles from those already generated by
the computer.
Another criticism by A.D. Little emphasized that Type II
errors were not accounted for in the statistical criteria. EPA
concurs that the Type II error probabilities were not
calculated, but EPA made efforts to minimize them by increasing
the Type I error probabilities by maintaining strict target
percentage measure criteria, and by subsequent screening and
selection of the final subcycle based upon maximum agreement
with the input data.
Two other criticisms made by A.D. Little are judged to be
immaterial. A.D. Little argues that since percentage measures
vary with engine speed, they are not independent distributions
and hence violate the "independent identically distributed
(i.i.d)" assumption, thereby rendering the K-S test
inapplicable. Secondly, A.D. Little argues that since the
cruise submatrix of the LAF subcycle is accepted at the .001
significance level, this implies that "...a cycle as aberrant
as the one accepted could be expected to occur only once in
every thousand cycles." This is deducing a Type II error
probability from a Type I error probability.
The K-S test was merely used as a filter to automate the
identification of cycles with parameter distributions similar
to the CAPE-21 data base, which it effectively did.
Engineering judgment, not statistics, was used in the final
comparison of parameter distributions and final selection of
test cycles. The question of the percentage measures being
from independent distributions is not relevant. The same logic
applies to the acceptance of the cruise submatrix for the LAF
B-117
-------
AP-19
subcycle at a .001 significance level. Despite the fact that
A.D. Little was incorrect in deducing a Type II error
probability from a Type 1 error probability without
justification, engineering judgment, not statistics, was the
method used to select the best cycle from the several candidate
cycles passed by the filter.
in summary, EPA judges A.D. Little's criticism of the
statistical filter to be unreasonably strict, although we agree
that the filter methodology contains inherent but unavoidable
weaknesses. EPA concludes that the final selected subcycles
are not as representative as those possible given a drastic
(and unacceptable in practice) increase in subcycle length.
EPA is confident that sufficient care was taken both in
filtering and final screening of the thousands of generated
subcycles to assure that those selected are sufficiently
representative. A.D. Little failed to substantiate the claim
that the subcycles are unrepresentative in the sense that
significant emission differences are attributable to the minor
discrepancies accepted by the filter.
4.5.2 L — EPA's use of medians rather than means was
incorrect with regard to the number of trips per day, cycle
length, and initial idle time. The mean (arithmetic average
value) is the more appropriate indicator.
4.5.2 E — A.D. Little's criticism of EPA't use of medians
rather than means, for determining cycle length, initial idle
time, and the appropriate fraction of operation characterized
by cold starts, was based on opinion, but they failed to
substantiate their opinion.
According to e.D. Rothtaan, Professor of Statistics at the
University of Michigan, "The median, , is a preferred measure
of location for skewed distributions arid coincides with the
mean for symmetrical distributions."[22]
A.D. Little's point that the means ana medians differ
indicates that the distributions are in fact skewed, and
therefore makes the median the preferred measure of location,
in accordance with accepted statistical practice.
EPA does not agree with Rothman's statement for all
situations, but it at least indicates that the median is
correct in some cases. The next st-'^ont represents a more
balanced approach:
"If a measure of central value is being chosen only to
describe a set of numbers, a choice between the mean and
the median is mainly determined by the interpretation put
B-118
-------
AP-20
on the concept of "central.¦ Thus, for example, in buying
a large number of bags of grain the mean weight of 100
bags is as useful to the buyer as the weights of the
individual bags. However, in a discussion of amounts of
money spent by families for milk it may be more useful
(from the over-all health viewpoint, if not from the
dairy's) to know a median, which specifically pinpoints
the dividing line between the upper and lower 50 percent
of the families, rather than the mean, which indicates the
total amount spent."[24]
Regarding the number of trips per day, cycle length, and
initial idle time, EPA prefers the median because it is not
biased by a small number of trucks which have substantially
different operating parameters than the majority of trucks.
The tendency of the mean to be unduly influenced by operation
unrepresentative of the majority of trucks makes it unsuitable
in this case.
In summary, A.D. Little's accusation that EPA was
incorrect in using medians is unjustified. The statistics
literature does not support their position.
4.5.3 L — EPA did not appropriately recognise the large
operational differences between cold and hot engine operation.
EPA concluded that cold and hot operation were essentially the
same/ although average operational statistics reveal much less
strenuous engine operation when it is cold. Finally, cold
engine operation represents an insignificant portion of total
truck operation.
4.5.3 E — Table A-3 was presented by A.D. Little as
support for their claim. This table indicates that there are
substantial differences between cold operation and the
operational parameters for the Los Angeles trucks.
It should be noted that A.p. Little characterised the cold
operation versus normal operation for the Los Angeles gasoline
trucks as having the "most prominent" difference. But EPA
selected the New York Non-Freeway (NYNF) cycle as the first
cycle after the cold start, rather than either of the Los
Angeles cycles. This allowed EPA to achieve reasonable
agreement between the cold operational parameters of the
CAPE-21 trucks and those for the transient test procedure
without necessitating the development of a seperate cold start
cycle.
The limited amount of cold-start data in the CAPE-21 study
would not support the development of a separate cold-start
cycle that would be reasonably representative. Furthermore,
B-119
-------
AP-21
Table A-3
Statistics Comparing CAPE-21
Cold Operation Versus Normal Operation (Los Angeles)
Cold Normal Operation After
Statistic Operation Cold Start warm Start Normal Start
Average
%
Power
15
37
51
35
Average
%
RPM
24
62
73
53
Average
%
mp a
9
32
38
24
Average
%
Idle
47
22
20
31
B-120
-------
AP-22
EPA's .report entitled "Analysis of Hot/Cold Cycle Requirements
fot Heavy-Duty Vehicles"[12 ] concluded that, except for the
idle period, there was not a substantial difference in truck
operation following a cold, warm, norma_, or hot start.
Therefore,
"...it is not necessary to generate unique cold start
cycles (engine or chassis) from a matrix containing cold
operation only. The sole requirement for a cold start
cycle is a longer than normal idle period (approximately
20 seconds) following engine startup."[12]
Table a-4 shows that the cold operation percent idle and the
NYNF cycle percent idle differ by only 6 percent. A.D. Little
failed to show that EPA's conclusions are incorrect, and that
the compromises cause an unrealistic emissions bias. More
importantly, A.D. Little fails to show how a separate cold
cycle could be developed from the sparse data matrix. Finally,
A.Di Little's opinion that cold emissions are insignificant
contradicts all emission test experience to date. Table A-5
shows that HC emissions are substantially higher on the cold
start portion of the cycle; light-duty vehicle experience
substantiates this conclusion:
5.0 Validation of Test Cycles
5.1 L — Both qualitative (visual inspection of truck
parameters) and quantitative (statistical "t-test") measures of
cycle representativeness were carried out by A.D. Little in an
attempt to validate the test cycles.
5.2.1 L — EPA conducted no tests of significant
differences between the truck data and cycle parameters. using
the CAPE-21 data and the "t-test" (a statistical test which is
valid for normally distributed populations) at a 95 percent
confidence level, A.D. Little's analysis indicated that seven
of nine chassis cycle operational parameters failed the test
for representativeness.
5.1-5.2.1 E — A.D. Little's claim that seven of the nine
chassis test cycle parameters are statistically
unrepresentative of the CAPE-21 data base was based upon the
"t-test." The "t-test," however, assumes that the data is from
a normally distributed population. A.D. Little admits that the
CAPE-21 truck parameters are not normally distributed, but:
"...the central limit theorem implies that their means are
normally distributed as sample size becomes large. The
sample size of 44 trucks is deemed large enough for the
parameter means to be normally distributed and tested
B-121
-------
AP-23
Table A-4
Statistics Comparing CAPE-21 Cold
Operation to HYNF Test Cycle Operation
Cola NYNF
Statistic Operation[23] Operation!12]
Average
%
Power
15
20
Average
%
RPH
24
12
Average
%
Idle
47
41
B-122
-------
Table A-5
Co 14 Start
Isj
U)
1
2
3
' 4
NVNF
LAHF
LAF
HYNF
1 |a |
24.73
17.11
15.10
5.55
IIIC 446
2 I b |
23.26
6.97
1.81
5.23
3lc|
.289
. 188
.166
.061
4 Ml
6. 7%
5.62
5. 02
1-82
1.
25.50
9.09
4.93
3.BO
1IIC 345
2.
64.84
5.51
0.78
5.20
3,
.40
.13
.07
.06
4.
17 . OX
5.52
3.0%
2 .62
I-
47.86
12. 73
5.95
2.61
CM 366
2.
94,7
6.24
0.BI
2.69
3.
.64
• 16
.08
.03
4.
2B.82
7.2X
3.6%
1.4%
1.
61.49
12.57
6.42
2.81
CM 350
2.
95.0
6.30
.92
3-06
3.
.86
.17
.08
.04
4.
33.5?
6.62
3.12
1.62
1.
32.91
16.16
14.67
6.68
F 400
2.
46.7 7
9.66
2.60
9.38
3.
.56
.26
.24
• 11
4-
11.71
5.4%
5. OX
2.3%
1.
20.11
8.13
7.39
I - 71
F 370
2.
52 .25
5.29
1.35
2-47
3,
,36
¦ 14
.13
.03
4-
10.9%
to
w
3.9%
.9%
Enf.ine by EnRiini Transient liC Emission Breakdown
20
Minute
Pause
Hot
Start
5
6
7
8
NYNF
LAN?
LAF
IWIF
9.66
11.14
12.53.
5. 71
10.65
4.61
1.50
5-36
.677
.736
.830
.3 73
20.4X
22.2%
2 5.02
11-22
4.82
6.29
4.b4
3.40
7.14
3.40
0. 75
4.38
.44
.54
• 42
.29
18.72
23.0%
17.92
12.32
4-96
4.69
5.07
2.62
5.69
2.28
0.69
2.74
.39
.36
.37
.19
17.62
16.2%
16.72
8.62
4.71
6.50
4.5b
2.74
5.56
3.16
.65
2.95
.37
.49
.35
.21
14-4%
19.12
13.62
8. 22
B.62
10.11
13.17
5.69
12.10
5.77
2.33
8.02
.87
.96
1.25
.55
18.1%
20.0%
26.02
11.52
8.05
7.65.
6.9b
3.46
13,37
4.78
1.26
5.02
.85
.76
.69
.35
25.7%
23.0%
20. bX
10.6X
Compos ite
Test
He sul t
3.3?
3.32
1Q0X
2.35
2.35
1002
2.22
2.22
1002
2.57
2.57
1002
4.60
4. BO
3.31
3-31
luox
High
Medium, or
Low Emi11erIe
M
-------
Table A.- 5(cont'd)
Engine by Engine Transient HC Emission Breakdown
1
NYNF
Cold Start
2
LANF
3
LAF
A
NYNF
20
Minute
Pause
Hot Start
6
LANF
7
LAF
8
NYNF
Compo6 i te
Test
Result
High
Medium, or
Lou Emitter
1.
8.56
7.18
8.22
3.63
10.23
6.A1
7.87
3.52
-
C 360
2.
13.11
3.A2
1.08
3.96
13.5A
2.99
I.OA
3.77
2.A5
M
3.
.11
.09
.10
.05
.80
.A7
.58
.28
2.A5
A.
A.5X
3.6%
A.0%
2.0%
32.7X
19.2%
23.7%
10. 2%
100%
1.
17.38
10.57
2A.67
7.76
10.25
9.32
22.22
9.10
-
C AA0
2.
20.12
A. 10
2.78
7.AI
11.32
3.65
2.AO
8.69
3.81
li
3.
.19
.11
.26
.08
.67
.57
1.37
.56
3.81
A.
5.OX
2.9%
6.8X
2. IX
17.6%
15.OX
36.0%
1A.7%
100%
1.
16.38
3.88
5.3A
1.68
A.9A
2.39
A. 95
1.57
-
GM A5A
2.
19.06
1.7A
.63
1.83
5.82
1.07
0.60
1.72
1.29
L
3.
.20
.05
.06
.02
.35
.17
.33
.11
1.29
A.
15.5X
3.9X
A. 7%
1.6%
27.1%
13.2X
25.6X
8. 5X
100%
1.
A 7. 31
A.33
2.08
1.6A
A.12
3.71
1.95
1.83
-
CM 292
2.
65.65
2.62
0.39
2.08
6.17
2.20
0.37
2.33
2.12
L
3.
.80
.07
.03
.03
• A3
.37
.20
.19
2.12
A.
37.7X
3.3X
1. AX
1. AX
20.3%
17.5X
9.AX
9. OX
100X
1.
AA.5A
15 A3
6.80
6.A3
CM A5A
2.
62.38
75
0.68
5.83
3.
.AA
.15
.06
.06
A.
17.9X
b.lX
2.AX
2.AX
11.85
11.24
.70
28. 5*
6.97
2.52
.39
15.9X
5.65
0.57
.33
13. AX
5.80
5.25
.33
13.AX
2.A6
2.A6
100X
-------
Table A- 5 (cont'd)
1
NYNF
CoIdS tart
2
LANF
3
LAF
EnRi.ne-bY-Enfti.ne Transient 1IC Emission Breakdown
Hot Start
4
NYNF
20-
Minate
Pause
5
NYNF
6
LANF
;
LAF
8
NYHF
lota L
Test
Composite
high
Medium, or
Low Imitter
1.
21 .04
6. 13
10. 39
3.69
3.66
5.01
9.51
3.71
-
CM 350
2.
31.48
3-53
1.71
5.34
5.18
2^ 93
1.57
4.51
2.66
M
3.
.34
.09
.16
.06
.34
.46
.87
.34
7. .66
4.
12.8X
3.AX
6. OX
2.3X
12. BX
17.3%
32.7X
12.8X
100X
Average lie
Emission Level
Average:
All
Eng ines
17. OX
4.8X
4.IX 1.9X
21.IX 18.4X 21.72 11.OX
100X
(12 engines)
2.7b
CO
I
h--
M
Ui
Average:
High HC
Engine s
Ave r age;
Neil, IIC
Engines
Average:
Low IIC
Engines
9. IX
17. IX
27.3 Z
4.5X
5. OX
U.bX
5.2% 1.8X
3.7X 2.2X
3.2% 1.5Z
20 . 5X 20.IX
21.4X 18.9X
27.OX 12. OX
20.3X 11.42
21.7X 15.6X 17-2X 8. 7X
100X H (4 engines) 3.81
IU0X H (5 engines) 2.5U
100X L (3 engines) l.btJ
>
~d
to
cr
M Total grams per subcycle.
lb] Crams per brake-horsepower-hour per subcycle.
[c] Stibcyclc contribution, in e f f cc t ive ly-ve Lgtited grams per brake -horsepower -hour, to the composite test result. tvolien
a
-------
AP-25
using a t-distribution. Therefore, the t-test is the
appropriate test used for testing the representativeness
of all chassis cycle parameters."
A.D. Little was incorrect in its application of the
central limit theorem. As Figures A-2, a-3 , A-4, A-5, A-6 and
A-7 indicate, the individual truck data is decidedly
non-normal. Secondly, the central limit theorem can only be
invoked for means; most of the data is in the form of
percentage measures—not mean values.
Even if the t-test were applicable, A.D. Little made
another mistake in selecting the wrong statistics for
comparison. A.D. Little used individual truck data, taking
means of each parameter for all 44 or 27 trucks. A.D. Little
neglected the fact that each truck was instrumented for varying
amounts of time, i.e., the aggregate data base consists of
effectively time-weighted individual truck data. Even with a
completely normal distribution, it would be no surprise if A.D.
Little discovered differences between their parameter and the
cycles, since the cycles were generated from a different
distribution.
Finally, the use of means for validation purposes is
insufficient. Any two distributions can be very different yet
have the identical mean value. It is the shape of the
distribution which is critical, especially with respect to
emissions. EPA realized this during cycle development and
included K-S tests and engineering evaluations of the cycle
distributions.
5.2.2 L — A qualitative review of the CAPE-21 percentage
cruise data indicates that the chassis cycles exhibit more
cruise than the input data warrants.
5.2.2 E — In this section, A.D. Little repeats its
criticism that the chassis cycle cruise statistics are
unrepresentative. As previously discussed, A.D. Little's
statistical tests were performed using an incorrect assumption
of normality; however, its observation that the percent cruise
for the three chassis cycle segments is less than the
corresponding percent cruise for the CAPE-21 trucks is
correct. The question is whether the operational difference in
percent cruise causes a significant difference in emission
levels. This question has yet to be ^ri^tessed, but it has no
bearing whatsoever on certification test procedures or
regulatory requirements.
5.2.3 L — The "t-test" at a 95 percent confidence level
is again used to assert that the cycle percentage cruise as
B-126
-------
12
11
10
9
8
7
6
cr
J
4
3
2
1
0
f1'J guru A-3
DJst ribirt i'6-n o£ Speed' & Dec e lie tab.i'o>v
(AA LA Non-Freeway 'f if licks):
>
l
to
ON
2-S 7.5 12.5 17. S 22.S 1! V vi
prncfMT QfcrL
-------
12
li
10
9
ot 8
*r
% ^
Ck
0 5
1 4
at.
3
2
1
0
-------
Distribution oE Power % Deceleration
(2 7 LA Freeway Gasoline Trucks)
S LO 15 2D
ffflCfNl OfCCL
>
ns
ro
CD
-------
Distribution of Power % Cruise
(27 LA Freeway Gasoline Trucks)
o
~
JJ
in 20 an un no eo 70
rrncrin cnuiaF
-------
Figure A-7
Distribution of Power % Acceleration
(27 LA Freeway Gasoline Trucks)
15.0
14.0 .
13.0 .
12.0 ..
11.0 .
„10.0 .
S 9 0 .
2 8 0
ta
>
i 5.B..
4.0 .
o
3.0 .
2.0 .
1.0 .
0.0 _
0 5 ID 15 20 25
prncfin nccrL
-------
AP-31
developed by EPA for the chassis cycles were not
representative. Standardized values of 5.1, 1.6, 2.0 for each
cycle were exhibited, when the 95 percent confidence level
requires that the standardized value be 1,96 or less.
5.2.3 E — A.D. Little stated that statistical tests were
presented In Section 4.2, but none were found, so we assumed
they were referring to the statistical tests presented in
Section 5.2 of their report. If so, EPA's analyses, 5.2.1 and
5.2.2, are relevant responses to A.D. Little's claims.
5.2.4 L — The "t-test" is again used to assert that the
individual chassis cycles are too short to be representative
primarily because of cold-start implications for short cycles.
"...While it is recognized that these trip lenths sometimes
combine freeway and non-freeway driving, and that the finalized
cycles are 20 minutes (as opposed to 5), the individual cycle
segments should be representative of the data."
5.2.4 E — A.D. Little claimed that the chassis subcycle
lengths are too short to be representative, the major
difficulty being that "...emissions and engine operational
differences which result from.cold or thermally unstable engine
operations are too high a percentage of total operation."
The issue of cold vs. normal operational differences has
been addressed in Section 4.5.3 E of this analysis and will not
be repeated here. Concerning the issue of trip length, A.D,
Little calculated confidence bounds for NY and LA gas trucks.
Again, A,D- Little's statistics are in error! Their
calculation is based on the assumption that the data are
normally distributed, but as Figures a-8 and A-9 (from
Reference 14) show, the distribution of the data does not even
remotely resemble a normal distribution. Therefore, their
confidence bounds are incorrect.
In addition/ even if their 95 percent confidence liirits
were correct, the chassis c^cle length is within their
confidence limits for N"5f gas truck average trip length, and
only 0,28 minutes or 17 seconds less than their incorrect 95
percent confidence limits for LA gas truck average trip
length. Also, in order to comply with their suggestion that
the subcycle lengths should be equivalent to the CAPE-21 trip
lengths, the chassis trip length would have to be expanded to
the equivalent of six consecutive ^r-nrk trips. This would
sextuple the test burden, but no emissions data were supplied
to justify the additional burden.
5.3 L — The gasoline engine cycles were also determined
to be unrepresentative, by virtue of the fact that the cycles
B-132
-------
F.I gure A-8
UHE f»rn trip orimir fuhctjom Fpn mt busoiini!
50 70 00 110 130
HO 00 00 100 120 140
hue irn imp (inhutrsi
-------
Figure A-9
ISf Pf H "1 HiP OffMSITY fUSiCIJOff TOR LR GfiaOLIflf
DO
in
rrrt .a.
200 ISO 35 20 S
135 70 22 10 9
f JHf Pfn 1RJP I 111 NUT f iJ)
:r- i r. r
a 3
3 3
-------
AP-34
were tjoo short, mean and median rpm values for the LA Freeway
cycle were too high; the correlation of percent power with
percent rpm for the LANF cycle is not consistent with the
CAPE-21 £ata, the LAF rpm exhibited higher variability than the
data base, and both LA cycles contained unrepresentative
parameterss
5.3 E — EPA's analysis of the engine subcycle length is
virtually identical to that presented earlier on the chassis
subcycies. in general, the overall test cycle is a nominal 20
minutes in length; within that 20 minutes the
freeway/non^freeway operations are appropriately weighted.
With respect to the individual subcycies^ a 5-minute length is
only a problem if the Monte Carlo technique was unable to
sample the entire data base within that time, and the
statistical filter was unable to screen out the
unrepresentative subcycies, subcycle lengths of 5, 10, and 20
minutes were generated during the CAPE-21 cycle generation
program to evaluate this very issue; subcycle lengths beyond 5
minutes minimally improved the agreement of subcycle parameters
with the input data parameters. Finally, it is nonsense to
argue that a predictive engineering test procedure must be as
long in time as the phenomenon it is intending to model. The
overall test procedure is, nevertheless, very close in length
to the trip length of the average truck (as derived by median
trip length).
5.3.1 L — "The LA Freeway engine cycle appears to have a
disproportionate percentage of high power and high RPM's,...
thus the rpm profile of the LA engine cycle was not
representative of the individual truck, data."
5.3-5.3.1 E — We can summarize A-.D. Little's arguments
that the LAF engine subcycle contains disproportionate
percentages of high power and high speed by saying that their
analysis is incorrect and incomplete. EPA's own review
indicates that the LAF engine subcycle contains about 12
seconds more WOT/WOT-equivalent power operation than the
CAPE-21 data base indicates. The impact on composite cycle
emissions- is small, representing a maximum overstatement of- CO
emissions of approximately 5 percent, and a maximum
understatement of NOx emissions of 3 to 6 percent. This small
percentage emissions discrepancy arises from the iflost
©per.ati©r*ally sensitive mode of engine operation.
A.D. Little used mean rpiii's of the individual LAF trucks
as, the comparative indicator of subcycle agreement with the
data base.. A.l>.> Little argued that since "...only 6 out of the
27 gasoiine trucks had mean* RPM values exceeding the cycle mean
Epifti va>Mes-,,»..the rpm profile of the LA engine cycle was not
B-135
-------
AP-35
representative of the individual truck data." Table A-6
presents summary data for the 27 LA gasoline trucks. Note
immediately that nine exceed the mean cycle percent rpm,
instead of A.D. Little's six. More significantly, note that
A.D. Little ignored the fact that the aggregate data base not
only included all 27 trucks, but also varying amounts of data
foe each truck, i.e., the trucks were instrumented for varying
amounts of time, and each spent varying amounts of time on the
freeway. (The hcurs of measured operation per truck are also
presented in Table A-6). The subcycle was generated from the
aggregate data base, effectively time-weighting the individual
truck mean percent rpm values. Time-weighting each individual
truck mean yields a mean percent rpm value of 62.82 (as opposed
to the unweighted mean of individual truck means of 59.5.) In
short, the higher speed trucks spent more time on the freeway,
or conversely, more time on the freeway led to a greater
proportion of higher speeds being recorded. Figure A-2
presents the non-normal distribution of mean percent rpms for
the individual trucks (un-timeweighted). Note that there is an
approximately 1 percent difference between the subcycle mean
percent rpm (63.52) and that of the aggregate data. In a
population whose mean percent rpm's vary from 42 to 85, a 1
percent difference between the subcycle and CAPE-21 means is
trivial. However, a mean or median value is only a gross
indicator of a population's overall distribution. A.D. Little
referenced a comment by International Harvester that the LAF
engine cycle contains "...a disproportionate percentage of high
power and high percent rpm's." EPA concurs with this
observation, but believes it to be inconsequential to the
overall representativeness of the test procedure.
In the EPA report!13J in which the candidate engine cycles
were screened and reviewed and the final cycles adopted, it was
recognized that within the final LAF subcycle "...there is
slightly more operation in the higher power range than might be
desirable." This is evident in Figures A-10 and A-ll in which
the %Power and %rpm are presented. Also presented, based upon
the subcycle length of 316 seconds, are the times spent in each
segment of operation for the actual subcycle, and for a
theoretical subcycle identical to the CAPE-21 data.
We note immediately from Figures A-10 and A-ll that the
additional time spent at high speeds and high power is small
relative to the entire subcycle length (9 seconds, or 2.8
percent of the total 316 seconds, f: l speeds greater than 97
percent; 15 seconds, or 4.7 percent, for engine torques greater
than 85 percent) . if this small additional time were
reallocated into more appropriate pacts of the distribution,,
there would have to be a drastic difference in emission rates
between the original speed or torque and the new speed or
B-136
-------
AP-36
Table A-6
Los Angeles Freeway Truck Data
Hours of Measured
Gasoline Truck Data Operation Per Truck
Truck No. (Type) Freeway Mean %RPM Combined Freeway Only
5
(2
axle)
62.8
7.0
.7
7
(2
axle)
55 .8
17.1
.8
10
(2
axle)
52 .6
14.0
5.4
11
(2
axle)
64.1
(x)
22.7
14.1
13
(2
axle)
58.8
15 .4
.2
14
(2
axle)
47.5
15.6
8.1
17
(2
axle)
78.1
(X)
6.5
2 .4
19
(2
axle)
51.5
12.6
6.1
21
(2
axle)
54.3
12 .5
4 .2
24
(2
axle)
49.1
10 .4
2.9
25
(2
axle)
47 .1
16.0
8.8
26
(2
axle)
41.9
9.2
3.8
28
(2
axle)
61.7
8.4
2.0
30
(2
axle)
45.6
8.2
.5
32
(2
axle)
59.1
6.0
3.3
36
(2
axle)
47.5
1.7
.2
40
(2
axle)
66.0
(X)
9.5
4.9
42
(2
axle)
56.5
6.1
2.7
47
(2
axle)
46.1
14.7
7.7
35
(3
axle)
76.7
(x)
15.6
4.6
2
(tractor-trailer)
52.1
.9.0
1.3
3
(tractor-trailer)
64 .6
(x)
10.5
3 .4
8
(tractor-trailer)
74.6
(x)
15.7
6.0
12
(tractor-trailer)
78 .3
(x)
20 .0
13 .4
39
(tractor-trailer)
85 .4
(x)
20.4
13.4
43
(tractor-trailer)
68.1
(x)
7.8
5.2
8
(tractor-trailer)
59.6
11.3
3.7
Summary
LA Freeway Engine Cycle Mean RPM: 63.52
Time-Weighted Freeway Truck Mean RPM (or, Mean RPM of the
Aggregate Gas LA Freeway Data Base): 62.82
Mean of individual Truck Mean RPM: 59.50
* Calculated from the ratio of freeway records to combined
records, multiplied by total hours of operation,
(x) Denotes trucks with mean RPMs higher than LAF engine cycle
mean RPM.
B-137
-------
¦ J. A* UUU 1' l ay
296644805
Figure A-10
I.. A. Cas Freeway
(316 Seconds Total. Length)
Key: Subr.ycJu
Cape-21 Input
A-Subcycl e Time Ln Mode (seconds)
B—CAl'L;-21. Input Data Equivalent Time in Mode (seconds)
C—13 j fference (A-li) (seconds)
A
28
y
.16
- JLil
16
16
16
25
32
38
4 7
63
U
.3(1-
1 3
1 1
_ .. JA .
__L2
21
27
30
25
30
40
55
C
-2
...-1
H-3
-4
-3
-5
-11
-5
+7
+8
+7
H-H
K
20% ~
.15%
17.5/
1 21
. 9.5%
1
£ '
i
L
10%
.1.2. 5%
8.5%
— — — — ¦
9%
6.5%
r _ _
r ~ ~
9.5%
_ i
5%
4.5% ,
6%
8Z
HZ
—J — - -
5%
—<—1
5%
, t
1
rz
(
AZ-
,
3 X
— 1
5%
—1
1
1
1
j
h-
Motoring 0* JO** 20 30 40 50 60 70 80 90 100
(Ncgat1vcs)
% I'ower
* 0 <_ Z < 5
** 5 < % < 15
-------
\' V
I.. A. t;.i§ t itiOVrty
2966/iA BOS
i'igur i; A- 11
50
1.. A. Gas Freeway
(j 1 6 Seconds 'J'QLii1!. LeivgLli)
Key:
Subcycle
C;vpo-21 Input
'.0
A-Subcycle Time in Mode; (seconds)
JJ-CAP.E-21 Input Data Equivalent Time in Mode (seconds)
C-l):l fTerence (A-JJ) (seconds)
30
Cd
¦I
H
LO
vO
20
10
"M-
H-+
CWb
Hf-WTR-
I
I^CI1
-t-
-20 -14 -8
10 16 22 26 3't AO 46
-I-
52
-I-
50
-4-
70
76
82
>
hd
I
oj
CD
1 6
8
A
in
5 :
IS
ri%
ia
az
—I—
—t-
—i-
1—i—1
^r.l.5%
"I—t ¦¦ t I—I ¦ t |~1—f
on 9'» 100 106 112 110 12U30 Ufi Vh'l 14
HJ'M
-------
AP-39
torque for the reallocation to produce a measurable difference
in overall subcycle emissions. somewhat greater differences
would be required to produce a measurable change in composite
test cycle emissions.
This analysis will assume that the misproportion of rpms
will have insignificant effects on composite cycle emissions,
while torque misproportions are potentially large enough to
justify further study. The rpm assumption is based upon three
facts: 1) earlier published data[17] indicated that composite
test cycle emissions were minimally affected by proportional
changes in the RPM schedule, 2) no theoretical engineering
reason exists to explain why a drastically different brake
specific emission rate would exist at higher speeds, and 3)
observation of steady-state emission data presented in Table
A-7. (Note in Table a-7 that emission rates do change from
speed-to-speed, but not to the extent which would cause the 9
seconds of differential emissions—representing 2.8 percent of
the subcycle and 0.8 percent of the composite cycle—to
measurably alter composite cycle emissions.) on the other
hand, torque significance is assumed on the basis of three
observations: 1) earlier published data[17] indicated that
composite test emissions were measurably affected by
proportional changes in the torque schedule, 2) theoretical
reasons for drastic emission changes do exist, i.e., power
valve fuel enrichment above 90 percent torque, and 3)
significant differences are in fact observed in the
steady-state data presented in Table a-7. Changes in emissions
rates between wide-open throttle (WOT) (100 percent torque) and
torques less than 90 percent (non-WOT loads), as approximated
from Table a-7 are:
BSHC
BSCO
BSNOx
Average Factor
of Increase
2.5
16.6
0.44
Maximum Factor
of Increase
11.0
43.8
0.81
Minimum Factor
of Increase
0.8
7.9
0 .29
Note that numbers less than 1.0 represent a decrease in WOT
emission rates relative to non-WOT rates.
The remainder of the analysis will concentrate only on
BSCO emissions because: 1) as indicated above, they are
affected most by WOT operation, and 2, o indicated in Tables
A-5 and A-8, the LAF subcycle is the major contributor of BSCO
emissions to the composite test results, whereas it is a minor
contributor of HC emissions. in short, if the misallocated
torque points in the LAF subcycle cause a net change in
composite emissions, that change will be most strongly observed
in composite BSCO emissions.
B-140
-------
AP-40
Table A-7
Steady-State Brake Specific
Emissions of a Current Technology hdg Engine*
BSHC
* *
•RP M/%Load
3600
3-100
2.600
•100%
.33
.26
.22
75%
.28
.06
.03
50%
.03
.17
.07
25%
.08
.04
.26
Average: .27
Ma ximum Va-1 ue: .33
Mini'muiti Value: .22
.11***
.28
.03
BSCO**
3600
3100
2600
81.5
68.5
59.1
1.86
2 .00
3.61
2.14
4.33
6.08
4.00
6.45
7.44
Average: 69.7
Maximum Value: 81.5
Minimum Value: 59.1
4.21
7.4 4
1.86
360 0
31.00
2600
3.22
3.99
4 .62
BSNOX
* *
11.2
10.6
9.70
9.17
10.2
9.07
7.56
6.92
5.67
Average: 3.94
Maximum Value: 4.62
Mini-mum Value: 3.22
8.90
11.2
5.67
***
1978 California IHC 404.
Grams/brake-horsepower hour,
Represents the average of
rates.
all non-100 percent emission
B-141
-------
Table A-8
EnRi.ne-by-EriRi.ne
Transient CO
Emission
B reakdown
Cold
Start
20-
Hot
Sta rt
Compos ite
High
1
2
3
4 Minute
5
6
7
8
Test
Medium,
NYNF
LANF
LAF
NYNF Pause
NYNF
LAM
LAF
MY HF
Kesu 11
Low Enil
Ha]
236.4
245.2
774.8
127.2
123.3
200.0
708.1
122.7
_
IliC 446
21 bj
222.4
99.9
93.0
119.9
135.9
82.7
84.7
115.1
92.88
11
3 [ c 1
2.79
2. 73
8.63
1.42
8.73
13.3
47.1
8.20
92.58
4 |d |
3.02
2.?X
9.3X
1.5X
9.4X
14. rx
50. 7X
b.«Z
100X
I.
90.3
84. 7
153.2
60.0
39. 7
60. 2
150. b
56.1
-
IIIC 345
2.
229.6
51.4
24.1
79.2
58.9
32.5
23.2
72.3
32.8
L
3.
1.40
1.24
2.24
.88
3. 70
5.29
13.11
4.94
32.B
4.
4.3X
3.8X
6.8%
2.7X
11.31
16.IX
40.OX
15. IX
100X
1.
143.7
140.7
187.7
86.1
88.0
113.3
167.0
88.2
-
cm 366
2.
284.3
69.1
25.5
68 .8
100.B
52.9
22.8
92.1
41.9
L
3.
1.9
1.8
2.4
1.1
6.9
8.7
12.5
6. 6
41.9
4.
4.5%
4.3X
5. 7%
2.6X
16.5X
20.8X
29. BX
15. 8X
100X
1.
171.2
155.2
404.5
102.6
111.6
130.2
376.6
95.3
-
CM 350
2.
264 .5
77.7
57.8
116.6
131.9
63.4
53.9
102.5
67.8
h
3.
2.4
2.1
5.3
1.4
9.3
10.1
29.7
7.4
67.6
4.
3.5%
3.11
7.8X
2.IX
13. 7X
14. 9X
43.8X
10.92
100X
1.
222.9
2.4
620.6
130.6
103.5
161.6
582. 3
127.3
-
F 400
2.
316.7
97.0
L09.9
183.5
145.4
92.2
103.0
179.6
113.2
H
3.
3.8
2.6
10.0
2.1
10.6
15.6
56.2
12.3
113.2
4.
3.41
2.3X
8.8X
1.9Z
9.4X
13.8X
49.6X
10.9X
1002
l.
85.2
106.6
230.7
21.4
38.B
80.1
206.7
40. J
-
F 370
2.
221.5
69.4
42.1
30.9
64.4
50.0
37.4
58.4
45.0
L
3.
1.5
1.8
3.9
.4
4.2
8.1
21.U
4.1
45.0
4.
3.3X
4. OX
B.7X
. 9X
9.3X
18.OX
46.7X
9. IX
100X
-------
Table A-^-8 .(cont'd)
15 nRine-by-"Engine Transient CO Emission Breakdown
Gold
Start
1
2
3
4
;NV-NF
LAHF
LAF
NYNF
1.
107.5
144.7
868.6
61.2
C 360
2.
164,7
69.8
113.8
66.7
3.
1.4
1.8
10.8
.8
4.
1.5.X
2.02
11-7%
. 92
1.
228.3
203.6
1262.0
100.6
C 440
2.
264.3
78.9
142. 1
96.0
3.
2.5
2.1
13.1
1.0
4.
2.22
1.82
11.32
. 9X
1.
250.3
86.2
769.6
65.2
CM 45 4
2.
291.3
38.7
91.3
71.1
(Short
3.
3.1
1.0
9.0
.8
Block)
4.
3.8X
1.22
11.02
1.02
1.
315.0
115.7
159.4
64.7
CM 292
2.
437 .1
69.9
30.2
81.9
3.
5.6
2.0
2.7
l.l
4.
10.2X
3.62
4.92
2.02
1.
204.8
175.6
376.1
144.6
CM 454
2.
286.9
65.5
37.9
131.1
(Tall
3.
2.1
1.7
3.6
1.4
Block)
4.
3.82
3.02
6.42
2.52
20-
Mi nute
Pause
5
NYNF
Mot Start
.6
LANF
7
L#VF
8
NYfeF
Composite
Tea t
Kesul t
High
flediiuro, or
Low Emitter
56.6
127 -D
783.0
58.5
-
76, 3
59.5
103.3
62 .6
92.0
11
4.6
9.6
58.6
4,4
92.0
5.OX
10.42
63.72
4.6X
T00X
75.2
161.1
1217-2
94,1
-
83.0
63.0
131.7
89.9
115.6
it
5.0
10.0
75-9
5.9
115,6
4.32
8.72
65. 72
5. IX
-
86.9
102.8
714.1
69.7
-
102.4
45.8
87.2
76.2
Bl.9
M
6.4
7 .2
4 9.5
4.9
b 1.9
7.82
8.82
60.42
6.OX
IUUX
89.4
111.0
161. 5
70.0
-
133.7
65.9
30.5
89.1
55.0
L
9.6
11.2
16.3
7.1
55.0
17.52
20.42
29.62
12. 92
1U0X
153.9
157.1
366,2
13B. 2
-
146.1
56.7
36.9
124.9
55.9
L
9.3
9.0
2U. V
7.9
55.V
16.72
16.12
37.4%
14.12
1002
-------
Table A-8 (cont'rt)
CM 350
Cold
Sta rt
EnRine
-by-Enj>irie
20-
Transient
CO Emission breakdown
Hot Start
Total
Higti
1
2
3
4
Minute
5
6
7
B
Test
Medium, or
NYNF
LANF
LAF
NYMF
Pau6fi
NYNF
LANF
LAJ
hYNF
Compos ite
Low Emitter
1.
196.1
108.9
805.5
68.1
92.1
104. B
640.8
64.8
-
2.
293.3
62.6
132.21
98.6
130.1
61.4
106.0
78.8
101.5
tl
3.
3.2
1.7
12.1
l.l
8.5
9.7
59.2
6.0
101.5
4.
3.22
1.72
CO
1.12
8.42
9.62
58.32
5.92
1002
Average HG
bmissioh Lev
Average
All
Engines
3.92 2.83! 9.52 1.7*
10.7* IU.iX 47.9* 9.9X 100X (12 engines)
75.7
Average
CO High CO 4.
£»
,Eng ines
Average
Mo'I CO
Enj- i ncs
2.62 2.1% 12.62 1.21
3. 72 2.22 9.42 1.62
7.U 11. U 56.32 7.0% 10UX (5 engines)
10.8* 11.9% 52.IX 8.52 100* (2 engines)
IU5.5
>
I
*>
14.9 o
Average
I-ow CO
Engines
5.22 *.72 6.52 2.12
14.32 18.32 36.72 13.42 1002 (5 engines)
46.1
[a|Totnl grams pc. ouu^ycle.
(b) Crams per brake-horsepower-hour per subcycle.
(cj Subcycle contribution, in effectively-weighted grams per br-afce-iiorsepouer-hour , to the composite test result. (When
added together, all subcycle contributions add up to the composite teGt result). Far methodology, see Reference 17,
pp. 4-5.
[flj Relative percentage of subcycle contribution (:) to the total composite test result.
[ej In grams per brake-horsepower-hour: High (li)^>90
90^ medium (M^> 60
Low (L)^60
-------
AP-42
Since the power valves of HDG engines are typically
calibrated to provide fuel enrichment somewhere above 90
percent torque, we shall be conservative and denote all torques
above 90 percent as WOT-equivalent, with an average value of 95
percent torque. All other torques within the subcycle shall be
represented by a constant emission rate (y grams per hour) at
an average value of 50 percent torque. Referring to Figure
A-ll, we note that approximately 12 more WOT-equivalent points
are included in the subcycle than the CAPE-21 data indicate.
W'e also assume that 95 percent torque/WOT-equivalent points
produce 40 times the emissions (40y grams/hr) than non-WOT
emissions. This represents a 21-fold increase in brake
specific emission rate, somewhat higher than the average
observed for BSCO in Table A-7, over that of the nominal
non-WOT/50 percent torque emission rate. Inherent in the
analysis is the assumption that all non-WOT modes exhibit
effectively equal brake-specific emission rates. (This is not
actually the case, but differences are small enough relative to
WOT-equivalent emissions that non-WOT brake-specific emission
rates can be approximated as equal for our purposes.)
Anticipated percentage reductions in BSCO emissions are
calculated in Exhibit A-l, assuming substitution of 12 non-WOT
(50 percent power) points for the 12 extra WOT-equivalent
points.
Exhibit A-l predicts a theoretical 4-7 percent decrease in
BSCO emissions by resubstitution of the 12 WOT-equivalent
points. Similar calculations predict a 1.2 percent reduction
for BSHC (assuming an average 5-fold increase in grams/hour HC
emissions for WOT-equivalent points, and an average LAF
subcycle contribution of 25.8 percent from Table A-5), and a
1.0 percent increase in BSNOx (assuming an average grams/hour
decrease of 28 percent, an assumed contribution of 25 percent).
These predictions can be tested experimentally by
referring to Reference 17, wherein regression equations were
developed to model observed behavior of HDG emissions over the
actual transient test procedure when the composite cycle torque
schedule was proportionally changed. Referring to Exhibit A-l,
STEP 2, resubstitution of the 12 points results in a drop in
subcycle brake horsepower-hour of 2.7 percent. Note that a 2.7
percent decrease in all 316 subcycle torques will reduce a
certain number of torques originally above 90 percent to below
90 percent. In fact, all torques between 90 and 92.5 percent,
originally producing WOT-equivalent emissions will now lie
Below 90 percent and produce non-WOT emissions. Referring
again to Figure A-ll, the number of torques in the subcycle
originally between 90 percent and 92.5 percent is approximately
12 (one-fourth of 47). (We must point out that the power valve
may actually be activated at 93 percent torque, as opposed to
B-145
-------
AP-43
Exhibit A-l: Calculated Changes
in Composite Cycle Emission Results
Step 1: Calculate Mass Emission Reductions
y = grams/hr, non-WOT torques(average value = 50%)
40y = grams/hr, WOT-equivalent torques(average value = 95%)
LAF Subcycle: Total grams = y t31^" 87) + 40y (,
of emissions
where 87 seconds (from Figure A-10) represent the present
number of WOT-equivalent points, 316 represents the total
subcycle length in seconds, 3600 represents seconds per hour.
Therefore, total grams of emissions = 1.030y grams
Substituting 12 seconds of non-WOT for 12 seconds of
WOT-equivalent yields a new emission rate:
New Emissions, _ „ .316 - 87 + 12, ,n. ,87 -12,
in Total Grams Y 1 3600 ' 4uy 1 3600 '
= .9Q0y grams
Step 2: Calculate Reduction in Brake Horsepower-Hour
x = nominal maximum engine horsepower
= .0548x Horsepower-Hour
Substituting 12 seconds of non-WOT, 50% nominal torque for
12 seconds of WOT-equivalent, 95% torgue:
New Brake _ q. , 87 - 12. 316 -. 87 + 12
Horsepower-Hour ~ 'yDX 1 3600 ; + 'oux { 3600 ;
= 0.533x Horsepower-Hour
Step 3: Calculate changes in Subcycle Brake Specific Emissions
B«k|USpecUic Emissions = 75^ 9tans per hotsepower-hout
= 18.80 £
A
B-146
-------
AP-44
Exhibit A-l (cont'd)
New Emission Rate = *Q533X grams per horsepower-hour
= 16.89 ^
%, Dif ference = 16 '8iB~8q8 *8° x 100%- » -10 .2%
Step 4: Calculate composite Test cycle Emission Effects
From Table A-8, LA Freeway BSCO emissions contribute, on
average, from 43.2 percent to 68.9 percent of the total
composite cycle BSCO emissions for the largest to highest
CO-emitting engines, respectively.
Therefore, a 10.2 percent reduction in LA Freeway subcycle
emissions will result in an approximate 4.4 percent to 7.0
percent reduction in composite cycle BSCO.
B—147
-------
AP-45
our assumed 90 percent, and in a linear fashion as opposed to
our assumed step change. However, it is reasonable to assume
that demanding proportionally less power across the subcycle
w.ill proportionally decrease the amount of time the power valve
is operating, and hence, change the power valve dependent
emissions. This likely explains most of the observations in
Reference 17. We consider this a reasonable means of
estimating the emission effects of decreasing the time spent at
power enrichment in the subcycle.)
The regression equations from Reference 17 are presented
in Exhibit A-2; calculated changes in composite emissions are
also calculated. Note that they are then recorrected for the
relative contributions of the LAF subcycle, i.e., the equations
assume a percentage decrease in brake-horsepower-hour across
the entire composite cycle, whereas we are concerned with the
net effect of reallocating a limited number of points in the
LAF subcycle alone. (Note that Reference 17 concluded that
sensitivity emission trends were identical for each subcycle.)
These regression equations, based upon emission data collected
from three heavy-duty gasoline (HDG) engines, indicate that one
can expect negligible changes in BSHC, less then 4 percent
decrease in BSCO, and less than 2 percent increase in BSNOx by
decreasing the power valve operation to an extent equivalent to
reallocating the 12 misallocated points in the LAF subcycle.
The observed trends in BSCO and BSNOx are consistent with
theory. Operation at torques below the power valve generally
occurs at marginally rich and at essentially constant A/F
ratios. (We assumed earlier that brake specific emissions
within this constant air/fuel (A/F) range were essentially-
constant.) Operation above the power valve increases fuel
flow; less oxygen available for complete combustion is the
reason CO emissions (partially burned fuel) increase. The
additional fuel cools combustion somewhat, leading to observed
reductions in BSNOx emissions at WOT conditions.
in summary, EPA concurs that there is somewhat more high
speed and load in the LAF subcycle than the CAPE-21 data
warrants. Both theoretical analyses and an analysis based upon
actual HDG emission data indicate that the emission effects of
these additional points are probably minimal (less than 5
percent). Given that the baseline standards were developed
with the present cycle (i.e., any inherent biases in the
procedure are carried through into the emission standards), and
given that the emission effects of tliio small additional high
power operation are minimal anyway, EPA sees no reason
whatsoever to characterize the LAF subcycle as unrepresentative
based upon the additional high-power operation.
B-148
-------
AP-4 6
Exhibit A-2: Use of Regression Equations
for Three HDG Engines to Predict Changes in Composite
Emission Levels as a Function of Changes in LAF BHP-hr
Step 1: Calculate Change in BS Emissions, Given Change in
BHP-hr of -2.7%
BSHC* = -.0,062 (-2 .7) + 2 .812 = 2 .829 ( + .6%)
BSHC = -.004 (-2.7) + 4.837 = 4.848 (+.2%)
BSHC = .0028 (-2.7) + 10 .07 = 10 .06 (-0.1%,)
BSCO = .978 (-2.7) + 49.5 = 46.86 (-5.3%)
BSCO = 1.895 (-2.7) + 79.4 = 74.28 (-6.4%)
BSCO = 2.97 (-2.7) + 144.8 = 136.9 (-5.5%)
BSNOx = -.174 (-2.7) + 13.8 = 14.27 (+3.4%)
BSNOX = -.138 (-2.7) + 6.2 = 6.573 (+6.0%)
BSNOX = -.310 (-2.7) + 11.9 = 12.74 (+7.0%)
Step 2: Correcting for % Contribution of LAF Subcycle to Total
Cycle
Effective Contribution Corrected % Change to Total
Cycle
BSHC 25.8%** negligible
BSCO 57.4%** -3.7 to -3.0%
BSNOX 25.0%*** +1.0 to +2.0%
* Expressed in grams/KW-hour.
** Average, all engines from Tables A-5 and, A-8.
*** Assumed.
B-149
-------
AP-47
5.3.2 L — The LAF engine cycle rpm standard deviation was
22.9, while the weighted average of the individual truck
standard deviation was 17.2. "Thus, it appears that the cycle
variation was excessive."
5.3.2 E — A.D. Little's assertion that the standard
deviation of LAF rpm indicates that excessive variation is
present, is judged by EPA to be explainable and
inconsequential. Firstly, as indicated in Figure A-2, the
actual truck mean rpm's vary so widely that we cannot see how
small differences in standard deviations can be meaningful in a
practical engineering sense in the first place. Secondly, A.D.
Little again neglected the differential operating times of
trucks, i.e., their time-weightings. Finally, referring to
Figure A-ll, we infer that an increased standard deviation of
the LAF subcycle rpms relative to that of the input data is
explainable by the above-mentioned extra time spent at higher
speeds. As also discussed above, the emission impact is
minimal.
5.3.3 L — The correlation of percent power cycle percent
rpm for the engine cycles indicated that both the LAF and NYNF
compared favorably with the CAPE-21 data, while the LA
non-Freeway is "definitely aberrant."
5.3.3 E — A.D. Little pointed out that the correlation
coefficient of percent power with percent rpm for the LANF
engine subcycle numerically lies at the low extreme of the
distribution of individual truck correlation coefficients
(i.e., the subcycle value is less than all but one of the
truck's values.) From this, A.D. Little deemed the LANF
subcycle "aberrant."
EPA notes that each candidate subcycle was subjected to
three discreet evaluation criteria as compared to the input
data: 1) the degree of agreement of overall summary statistics
(means, medians, percent acceleration, etc.) for both rpm and
power, 2) the degree of agreement of the rpm and power
distribution functions, as determined by the K-S test, and 3)
the degree of agreement of the statistical parameter density
plots, as determined by engineering judgment. The final LANF
subcycle met the above criteria not only satisfactorily but
also better than all other candidate subcycles. The
correlation coefficient was not used for subcycle screening
purposes; the fact that the coef f \ ^i.«nt for the selected
subcycle lies at the extreme end of the input distribution does
not nullify the agreement of all other subcycle parameters with
the CAPE-21 input. EPA can only speculate as to the practical
engineering significance of the correlation coefficient. The
correlation coefficients for the individual trucks are evenly
B-150
-------
AP-4 8
distributed within their ranges of variat-on, although the
ranges ao differ apparently with freeway or non-freeway
operation. A.D. Little's arguments infer that the LANF, as
measured by the correlation coefficient, is too similar to
freeway operation, although this is not borne out when all
other statistical parameters are reviewed.
In summary, EPA agrees with A.D. Little's observation that
the LANF correlation coefficient does not represent an average
value of the individual truck coefficients. EPA doubts,
however, that this discrepancy constitutes any practical
emissions significance, especially given the subcycle's strong
agreement with the input data as measured by the other
statistical parameters.
5.3.4 L -- Again using a "t-test" and a 95 percent
confidence level, 5 out of 48 gasoline engine cycle parameters
were non-representative.
5.3.4 E — A.D. Little again used t-test on distinctly
non-normal populations (see Figures A-2, a-5, a-6, and A-7),
but provided no valid justification for its use. Application
of the central limit theorem is incorrect for samples which are
not made up of means. They also erred by using individual
truck data instead of making allowances for the actual
truck-hours included in the data base. Ignoring these
mistakes, however, EPA notes that A.D. Little concluded that
LAF rpm mean, median, and standard deviation (discussed and
discounted above), LAF power percent acceleration, and LANF
power percent motoring were deviant. For each engine subcycle
(NYNF, LANF, and LAF), there are two specified operational
parameters, rpm and torque/power, each with 8 discreet
statistical parameters (mean, median, standard deviation,
percent acceleration/ percent deceleration, percent motoring,
percent cruise, ana percent idle.) In short, there are 48
(three subcycles multiplied by two operational parameters
multiplied by eight statistical parameters) total parameters to
be evaluated. In short, 5 of 48 parameters failed A.D.
Little's test. Conversely, 43 of 48, or 90 percent, of the
engine cycle parameters were representative of the individual
truck data at a 95 percent confidence level, as determined by
A.D. Little. Despite 90 percent agreement at a 95 percent
confidence level using their own tests, and with no data nor
even an educated judgment as to the potential emissions impacts
of any of the remaining minor differences in summary
percentages/statistics, A.D. Little reached a sweeping,
all-inclusive judgment that the engine cycles were
unrepresentative. EPA regards this as unreasonably and
inappropriately strict.
B-151
------- |