SUMMARY AND ANALYSIS OF
COMMENTS TO THE N P R M:
"1983 AND LATER MODEL YEAR
HEAVY-DUTY ENGINES
PROPOSED GASEOUS EMISSION
REGULATIONS "
December,1979
-------
Summary and Analysis of Comments to the NPRM
"1983 and Later Model Year Heavy-Duty Engines
Proposed Gaseous Emission Regulations"
December, 1979
-------
Table of Contents
I. Introduction ii
II. List of Commenters iii
III. Part I: Analysis of Major Issues 1
A. Test Procedure 1
B. Redefinition of "Useful Life" 105
C. In-Use Durability Testing 115
D. Allowable Maintenance 126
E. Parameter Adjustment 154
F. Idle Test and Standards 161
G. Leadtime 164
H. Economic Impact 186
I. Technological Feasibility 217
J. Selective Enforcement Auditing . 246
K. Nonconformance Penalty 258
L. Diesel Crankcase Emissions Control 259
M. Numerical Standards/Standards Derivation. .... 264
N. Fuel Economy 280
IV. Part II: Analysis of Minor Issues 302
-------
I. Introduction
The Environmental Protection Agency (EPA) published a Notice
of Proposed Rulemaking (NPRM) on Tuesday, February 13, 1979,
proposing new heavy-duty engine (HDE) emissions regulations for
1983 and later model years. The proposed rule prescribed more
stringent hydrocarbon and carbon monoxide emission standards, and
established an assembly-line testing program and nonconformance
penalty system for 1983 and later model year heavy-duty (HD)
gasoline-fueled and diesel engines as mandated by the Clean Air Act
Amendments of 1977. Substantial changes were also proposed to the
emission test procedures, the definition of useful life, and the
procedures used to verify the durability of emission control
systems over their useful life.
This document presents a summary and analysis of the comments
received in response to the NPRM. The comments have been grouped
into two parts. Part I addresses the major issues which, for the
most part, are the proposed changes to the HDE regulations that
were listed in the Preamble to the NPRM. The analysis of these
major issues leads directly to final recommendations on the pro-
posed changes. Part II supplements the discussion of the major
issues by supplying the technical details. These details do little
to offset EPA's final recommendation on the respective major issue.
-------
II. LIST OF COMMENTERS
1. Professor Philip S. Myers
2. American Trucking Association, Inc. ATA
3. IVECO Trucks of North America IVECO
4. Motor Equipment Manufacturers Association MEMA
5. Caterpillar Tractor Company CAT
6. Chrysler Corporation Chrysler
7. Council on Wage and Price Stability COWPS
8. Cummins Engine Company, Inc. Cummins
9. Engine Manufacturers Association EMA
10. Ford Motor Company Ford
11. General Motors Corporation GM
12. International Harvester Corporation IHC
13. Perkins Engines
14. Mack Trucks, Inc. Mack
15. Mercedes-Benz of North America, Inc. M-B
16. Motor Vehicle Manufacturers Association MVMA
17. Spector Freight Systems, Inc.
18. Garrison Motor Freight
19. Environmental Action of Michigan, Inc.
20. United States Department of Commerce DOC
21. United States Department of Interior DOI
22. Detroit Diesel Allison DDA
23. Connecticut Construction Industries Association
24. University of Waterloo
tit
-------
PART I
Analysis of Major Issues
-------
A. Issue - Test Procedure
1. Summary of the Issue
EPA has proposed that the steady-state test procedures pre-
sently used for the certification testing of heavy-duty engines be
abandoned. In their place, the use of a transient test has been
proposed. This transient procedure was developed by statistical
analysis of on-road operational data collected in the New York and
Los Angeles urban areas. The transient test covers the full range
of engine operation; the steady-state tests are limited to specific
speeds and loads.
2. Summary of the Comments
Heavy-duty engine manufacturers were unanimous in their
criticism of the transient test, and in their desire to retain the
steady-state procedures.
a. Justification
First of all, they claimed that no substantive justification
for its promulgation has been advanced by EPA. EPA purportedly has
not proven that concrete air quality improvements will result from
implementation of the new test procedure. Furthermore, EPA pur-
portedly has not presented "hard" technical data supporting its
contention that the current test procedures inadequately predict
future on-road emissions. It was claimed that this lack of "hard"
technical data constitutes regulatory action on the basis of
conjecture. Furthermore, the commenters maintained that this
approach does not satisfy the judicially-determined requirements
(International Harvester vs. Ruckelshaus) that the Agency "bear a
burden of adducing a reasoned presentation supporting the relia-
bility of its methodology." The Agency has been expressly accused
of being "arbitrary and capricious" in its methodology and its
reasoning.
In the case of diesel engines, it was claimed that EPA's
justification for the transient test is especially weak, based
primarily upon regulations and standards which have yet to be
proposed (future particulate and NOx standards). A diesel engine
does not possess transient enrichment devices (chokes, accelerator
pumps), has HC and CO emission levels conceded by EPA* to be "quite
close to the 90 percent reduction level," and at that level,
manufacturers claimed that emissions are adequately predicted by
the current steady-state procedure. The industry argued that
future and unknown requirements for NOx and particulate control are
unsubstantiated justification for a transient test, and represent
an abridgement of the industry's rights to comment on regulatory
methodology.
In NPRM Draft Regulatory Analysis, pp. 132-133.
-------
For both gasoline and diesel engines, manufacturers commented
that EPA's arguments attempting to prove the inadequacy of the
current test procedure (the lack of transient operating modes,
limited number of engine speeds, absence of cold start operation,
improper and unrepresentative weighting factors, and lack of
correlation with on-road data) constitute insufficient justifica-
tion for a transient test. It was claimed that EPA has presented
no evidence that a transient test will alleviate these shortcoming
nor that these shortcomings need be alleviated to obtain on road
emission reductions. It was argued that modifications to the
current test procedure could adequately correct these problems at a
much lower cost. In fact, data does exist (California in-use
vehicle surveillance study) that purportedly shows 9-mode testing
does correlate with a proven transient cycle (LA-4).
The manufacturers' claimed that the change in test procedure
unreasonably hinders the industry for no substantiated benefit. It
was argued that the substantial data base and technological exper-
ience acquired through steady-state testing will be rendered
useless; the lack of identifiable operational modes in the trans-
ient test will make the design of new emission control technology
difficult. The cost of the procedural change purportedly outweighs
its proven benefits.
It was unanimously suggested that the Agency implement stan-
dards for 1983 based upon the steady-state procedures.
b. Representativeness
The specific transient test cycles proposed by EPA also came
under considerable attack. The industry almost unanimously claimed
that the proposed cycles are unrepresentative of actual truck
operation, having been generated from a questionable data base
using a questionable methodology. Specific problems with the
Cape-21 data and the proposed cycle which were cited by the manu-
facturers include:
i) The use of transient manifold vacuum and rack positions
to approximate transient engine torques and horsepower was tech-
nically incorrect and resulted in erroneous and physically unreal-
istic acceleration rates.
ii) The cold/hot weightings in the proposed cycles are
different from those indicated in the Cape-21 study.
iii) The overspeed in the cycles is highly unrepresentative
and indicative of the questionability of the data.
iv) The Monte Carlo technique, used as it was without time
sequencing of more than one second, resulted in erratic cycles with
unrepresentative speed/load patterns.
-------
v) A Large percentage of the Cape-21 data is "spurious," due
to vehicle vibration exciting the optical encoder used to measure
vehicle speed, and to electrical noise present in the rack position
indicators. The transient cycles were developed from this highly-
spurious data base, and thus are inherently unrealistic.
vi) All engines in the data base were naturally-aspirated;
since future engines will be almost exclusively turbocharged, the
data base is highly unrepresentative of the future fleet.
vii) The proposed cycles have average idle and cruise time
different from the data base.
viii) The actual trucks in the data base were unrepresentative
of current and future truck populations both in number and GVW, and
in some cases were in questionable mechanical condition.
ix) The distribution of cycle acceleration rates are dif-
ferent from those in the individual truck data.
x) Engine inertia was ignored in developing acceleration
rates for the proposed cycle, resulting in an overstatement of the
power an engine is capable of delivering.
xi) Only two cities were used in the data .base, and only
urban driving was represented.
xii) Horsepower models for Cape-21 data analysis were devel-
oped from extremely limited data bases.
xiii) Motoring torques were not measured during the Cape-21
study; those present in the test cycle are the result of guesswork
and inherently unrealistic.
xiv) Evidence was presented at the Public Hearings that no
engine could follow the cycles without help of a motoring dyna-
mometer. It was claimed no vehicle could follow it at all.
Aside from Cape-21, Chrysler Corporation maintained that
loadings on the transient tests were unrepresentative of those
found on smaller GVW trucks - a majority of Chrysler's production.
Chrysler suggested EPA recognize the noncommercial usage (e.g.,
motor homes) of some of the heavy-duty engines they manufacture,
and recommended the use of a chassis procedure using an appropriate
road load. In short, shouldn't there be separate certification
requirements for non-commercial HDV's?
The comment was also received that the operation of engines in
speed control mode on EPA's and SwRI's transient dynamometers was
unrepresentative of actual on-road operation and would result in
artificial emissions. Torque control is a more logical and tech-
-------
nically-correct strategy. EPA's transient data base is exclusively
speed control-generated and thus claimed to be highly unrepresen-
tative.
c.
Validation
Another major concern of the industry pertained to the lack of
knowledge about the correlation of emissions measured on the
transient test with those in the real world. The assertion was
made that EPA has failed to show that the proposed procedure
correlates with real world emissions any better than the current
procedures. Furthermore, should better correlation be established,
it was argued that it must be of a sufficiently superior degree to
justify the capital expenditures necessary to adopt the transient
test. It was claimed that the only method which can establish this
correlation is an actual on-road testing program. Most commentors
supported the validation approach outlined by Professor Philip
Meyers at the May Public Hearings.
EPA used an on-road program (San Antonio Road Route) to
conclude that the current test procedures are unrepresentative.
The industry claims it is only logical that the same procedure be
applied to conclusively demonstrate the validity of the new proce-
dure. Yet it was claimed that EPA has noC even demonstrated
emission correlation between laboratories testing on the new
procedure, let alone correlation with the real world. Furthermore.
in light of the purportedly questionable methodology used in
developing the cycle, and in light of the legal decision in Paccar,
Inc. vs. National Highway Traffic Safety Administration (i.e. , that
administrative agencies are obliged to test whether their standards
and procedures perform in the manner to which they are designed)
EPA is obligated to validate the proposed procedure on the road
prior to its implementation.
d. Evaluation of Alternatives
The manufacturers also claimed that the Agency failed to
adequately explore alternatives to the complex and expensive
proposed procedure. Industry claimed that errors in the Cape-21
methodology are serious enough to warrant review of the data and
to regenerate alternate test cycles. The main thrust of their
arguments, however, claimed that EPA violated its expressly stated
intention and its legal responsibility under the CAA to consider
alternatives; EPA's sole rationale for this approach was a lack of
time. It was advanced that nowhere in the regulatory support
documents is there evidence that EPA considered simpler, more
cost-effective procedures.
A case in point of EPA's rejection of alternatives involved a
simpler transient procedure submitted to the Agency by MVMA on
February 13, 1978. It is claimed that this proposal was rejected
-------
without adequate consideration of its merit (i.e., based upon EPA's
limited resources available to evaluate the proposal).
e. Inability to Comment
A legal issue raised in the comments involved the industry's
claim that they have been deprived of the due process of law. The
majority of the industry does not have experience with transient
engine testing, nor do they have the facilities needed to acquire
this experience. The few manufacturers who are running transient
cycles have only been operational for a short time. This scarcity
of data and experience with the proposed procedure has purportedly
curtailed the industry's ability to knowledgeably comment on the
proposed procedure and its feasibility. It was claimed that this
has effectively resulted in a deprivation of due process and the
industry's right to comment under the law.
f. Technical Adequacy
Further comments pertained to the questionable ability of the
proposed procedure to accurately and repeatably measure emissions.
The opinion was expressed that there may be significant emission
variability problems. To operate on different equipment at the
full range of the permitted validation specifications may result in
excessive variability and serious correlation problems between
laboratories.
g. Alternative Cycles (Caterpillar Cycle)
Caterpillar suggested that slight modifications to the pro-
posed cycle would allow diesel manufacturers to retain their eddy
current dynamometers for transient testing, thus saving time and
money- A specific cycle capable of being run on an eddy current
dynamometer was proposed.
h. Technical Details
Several manufacturers questioned the need for a CVS system,
citing both technical problems with bagged NOx measurements and the
existence of cheaper and more readily available alternatives.
Diesel manufacturers claimed that no need existed for a
12-hour cold soak for diesel engines; in fact, cold start diesel
emissions were characterized as equivalent to those measured on a
hot start. A cold start requirement for diesels was described as
overly burdensome and technically unjustified. Gasoline manufac-
turers also cited the fact that a 12-hour soak requirement unneces-
sarily tied up dynamometers and recommended consideration of a
forced cool-down. Should the cold cycle be reweighted to a smaller
degree, as also advanced by the industry, a cold start would be
even less technically justified.
r
-------
The remaining comments on the test procedures were highly
technical and detailed in nature. Resolution of these issues is
expected to have a minimal effect on the outcome of the test
procedure issue as a whole. These comments will be addressed in
Part II of the Summary and Analysis of Comments.
3. Analysis of the Comments
a. Justification
The argument of "arbitrary and capricious" is easily refuted.
The explicit goal of the Agency from the early 1970's onward has
been to develop a representative certification test procedure for
heavy-duty engines. A consistent and deliberate progression
towards a transient procedure ensued, based upon the Agency's early
judgment that a transient test was inherently more representative
of in-use operation. The allegation that EPA's methodology and
technical judgment were "arbitrary and capricious" is rhetorical
and not based on fact. A brief synopsis of the history of the
proposed test procedure will serve to illustrate that the develop-
ment process was technically sound, consistent through time, and
based upon a reasonably perceived need. The final justification
for a change in test procedures relies upon proof of the current
procedure's inadequacy, quantification of the incremental air
quality benefits, and an evaluation of the overall cost effective-
ness of the proposed procedural change.
The Agency's dedication to development of a transient proce-
dure is well documented in Agency records. The exact methodology
used in deriving the transient test was identified as early as
1972:
"Most of the activity on this project will be directed toward
the long range development of the realistic test procedure. This
effort will involve several elements, the most important ones being
1) the acquisition of truck-operating data in New York and Los
Angeles through the APRAC-CRC CAPE-21 project, 2) the use of
a computer program to process the road data and generate represen-
tative engine duty cycles, 3) the development of a technique for
determining mass emissions of HC, CO, (and) NOx.... emissions from
gasoline and diesel engines, the refinement of a suitable test
procedure and 4) the preparation of regulations."* Table A-l
highlights the events and decisions which followed.
It is significant to note that the fundamental methodology
cited above was not arrived at by EPA alone. This approach was
arrived at through cooperative interaction between EPA and the
regulated industry in the form of a jointly-funded on-road data
*Summary of" ECTD program plans for FY 1973, Project II.2., "HDV
Revised Exhaust HC, CO, NOx, and Smoke."
-------
acquisition program. The New York phase of Che Cape-21 project was
managed by committee with representatives from both industry and
EPA. (EPA withdrew from the committee following the New York study
due to concerns about conflict of interest implications of joint-
ly-funding programs with a regulated industry (see Exhibit A-l)
The Agency carried out the Los Angeles half of the study on its
own, but using identical methods and instrumentation as in the
cooperatively conceived New York study. (Further discussion of the
industry's awareness and involvement in this test procedure devel-
opment process appears later.)
The Agency's early decision to push for a transient cer-
tification procedure was initially brought on by observations of
light-duty emission control. The manufacturer's ability to selec-
tively optimize emission control systems to pass simplistic test
procedures (in this case, the 1968 California 7-mode) was recog-
nized early on; this resulted in the eventual selection of the
transient LA-4 driving cycle for light-duty certification.
Furthermore, the "Ethyl Study"** comparison of steady-state and
on-road tests vs. transient cycles concluded that transient tests
are inherently superior predictors of actual in-use emissions.
Whereas the duty cycles for light-duty passenger vehicles were
relatively easy to model, the tremendous variety and interchange-
ability of heavy-duty engines, driveline, and vehicle applications
presented a serious logistical problem in deriving a driving cycle
representative of an "average" truck. The next three years were
spent in the accumulation of a vast data base from which urban
truck usage patterns could be modeled. This was followed by two
years of data editing and the generation of driving cycles. At the
same time, various EPA contracts (summarized in Table A-2), studied
the relationships of various transient, modal, and on-road tests.
The data collected in these contracted studies led EPA to the
following conclusions:
i) At intermediate levels of emission control, emissions
reductions measured on steady-state procedures were somewhat
related to on-road reductions, and formed the basis for interim
emission standards.
ii) No combination of steady-state modes or modal weighting
factors could consistently predict transient emission reductions,
i.e., transient tests were inherently superior. (See Exhibits A-2
and and A-3.) Furthermore, based upon this and light-duty expe-
rience, at higher levels of emission control current test proce-
dures failed to correlate with the on-road data.
** Study 2, Table A-2.
-------
Based upon conclusion i), and recognizing Che fact that a
transient procedure was far from finalization, interim emission
standards based upon the steady-state tests were promulgated in
1977. Note, however, the following excerpt from the Environmental
Impact Statement accompanying these interim regulations (August 4,
1977):
".. . [C]urrent test procedures are inadequate predictors of on
the road GO and NOx emissions* (i.e., a given reduction in emis-
sions measured...results in a much smaller reduction...on—the—road
. . . Development of a new heavy-duty engine test procedure is
currently under way...In the interim, modifications to the current
test procedure...will improve...[their]...accuracy...and result
[in] greater reduction in on-the-road emissions than is now being
obtained."
The transient test procedure development continued with the
selection of candidate cycles, upgrading dynamometer control
systems, and actual baseline testing.
This transient baseline work established the practical feas-
ibility of the proposed transient procedure, during which time the
test procedure was refined, and allowed in-laboratory comparisons
between transient and modal procedures.
This in-house testing provides the "hard data" demonstrating
the inadequacy of the 9-mode gasoline procedure for prediction of
in-use emission reductions at the future levels of control. Table
A-3 summarizes the current technology and prototype engine testing
to date. The current technology summary indicates that even at
today's level of control, the 9-mode underestimates emissions
measured over a transient test. (For HC, this underestimation is
on the order of a factor of 6, for CO a factor of 3.5.)
An examination of applied catalyst technology (that level of
technology universally conceded by manufacturers in their comments
and testimony to be necessary to comply with the proposed standards
on either test procedure - see Summary and Analysis: Standards,
Standards Feasibility) reveals even larger discrepancies. Note
that the GM 400 and the Ford 351 were originally certified in
light-duty vehicles and consequently the 9-mode results exceed the
proposed standards. The addition of an air pump to the 400, while
reducing the 9-mode emissions to virtually insignificant levels,
reduced the transient emissions to a much lesser extent. In fact,
transient CO emissions were virtually uncontrolled. This, plus the
* Following the accumulation of actual transient test data in
1978, it was determined that all gaseous emissions are inadequately
predicted by steady-state tests. See data presented below in this
analysis.
-------
data from the remaining retrofit heavy-duty engines, prove that
certification of gasoline engines on a steady-state procedure
would result in serious underestimation of on-road emissions.
Engines equipped with catalysts sized merely to pass the 9-mode
test would emit CO at virtually uncontrolled levels under transient
and high power operation on the road.
This in-house data supports EPA's early judgments based on
light-duty experience, supports the results of the on-road studies
outlined in Table A-2, and graphically contradicts the manufac-
turer's claim that a 9-mode can be representative at future levels
of emission control.
The only study referenced by the manufacturers proving
9—mode correlation with a transient test was "Surveillance Testing
of Medium-Duty Gasoline-Powered Vehicles" (California Air Resources
Board Report No. VE-78-021, April 1978). This study showed equiv-
alent percentage reductions, reflecting the increasing stringency
of standards through the years of various model year vehicles,
measured both on a chassis 9-mode and the transient LA-4 light-duty
test. However, no trucks with a GVW greater than 8,500 Ibs.
were tested and in the words of the report, "...these vehicles
are primarily vans and pickup trucks which are functionally
light-duty vehicles..." The fact that the 9-mode predicts percen-
tage reductions for light-duty vehicles operated on a light-duty
procedure does little to assure the same degree of effectiveness on
heavy-duty vehicles and gives further evidence of the 9-mode's
inadequacy for heavy-duty. Light-duty tests results are not indi-
cative of heavy-duty working emissions. The light-duty LA-4 test
procedure was derived from on-road vehicle speed traces and repre-
sents an "average" trip for an urban passenger car. For a given
trip a passenger car engine "transports" less . than 6,000 Ibs.; a
heavy-duty engine "transports" vehicles and cargoes totaling into
the tens of thousands of pounds. Furthermore, for a given vehicle
acceleration, a heavy-duty engine undergoes many more transients
due to the higher number of gear ratios. For a cruise at a given
speed, heavy-duty engines work substantially harder to overcome
higher windage and rolling resistance. In short, the operational
characteristics of heavy-duty engines are so different from
light-duty engines (a point brought out in all of the commenters'
submissions which pointed out the severity of the heavy-duty
working environment), that separate test procedures for heavy and
light-duty were developed in the past. The emissions gener-
ated over the procedures are not comparable, regardless of the
level of technology. The fact that the test procedures were not
comparable, and the fact that the 9-mode seriously understates
transient emissions was conlusively demonstrated at the EPA labor-
atory when a General Motors light-duty van with a 1979 prototype,
catalyst-equipped 400-CID engine was first tested on the light-duty
LA-4 and Highway Fuel Economy chassis cycles. The engine was then
removed, set-up on the engine dynamometer and run through a tran--
-------
slant teat. The results are presented below:
Proposed Transient 9-Mode FTP LA-4 HWFE
Cycle (g/lb fuel) (g/lb) (g/lb) (g/lb)
HC 3.53 1.77' 1.51 0.420
CO 210.50 100.20 31.70 32.48
NOx 3.71 5.66 3.05 3.40
The transient light-duty cycle is not comparable to heavy-duty
cycles, nor are the measured emissions.
The proposed transient procedure contains several modes of
operation not present in the steady-state tests;* these were cited
in the Draft Regulatory Analysis as rationale for a transient
procedure on the logical presumption that the more accurately a
test procedure reflects operation in the real world, the more
accurately emissions measured on that test procedure will reflect
real world emissions. If the proposed cycles do indeed contain all
significant modes of heavy-duty engine operation in their proper
sequence and proportion, then it is only logical to assume that
emissions generated over the cycle are also representative.
Furthermore, given the choice of two test procedures, each yielding
different results, is it not logical to presume that the emission
results generated through the more representative test would more
accurately model the real world, especially when actual on-road and
laboratory data substantiate the inaccuracy of the less-represen-
tative test? It is concluded that not only is the 9-mode unrepre-
sentative, but its use would result in gross and unacceptable
errors in the prediction of on-road emissions, especially when
incorporated with catalyst technology. A 90 percent reduction in
on-road emissions utilizing the 9-mode test procedure is impos-
sible to guarantee.
The only remaining alternative would be to restructure the
9-mode test. Necessary modifications would entail the inclusion of
100 percent power modes to include WOT and power valve operation
which would more accurately simulate the power levels present in
the LA-freeway, and more accurately subject the catalyst to power
'levels present in the real world and ensure its proper sizing.
Furthermore, as illustrated by actual test results presented in
Figure 1, cold start HC and CO emissions will have significant
effects on measured emission levels. Note that hot start hydro-
carbons for this engine lie well below the proposed standard, yet
properly weighted cold start emissions result in a composite test
result exceeding the standard. With the catalyst properly sized to
handle the CO emissions, the same effect occurs (see Figure A-2).
Therefore, any restructured steady-state test would also require
* Transient operation, a full range of engine speeds and loads,
cold start operation, representative weightings of test modes.
-------
cold modes to adequately simulate the real world. However, no
steady-state gasoline procedure has ever demonstrated an ability to
correlate with on-road and transient procedures at lower levels of
control on anything but an engine specific basis (see Report 7,
Table A-2). The presence of transient enrichment devices, the
operation of which are completely ignored in any steady-state
procedure, operate a large percentage of the time in the real
world.*
Figure A-2 presents the test results of a catalyst-equipped
engine which meets the proposed standards.** Note that GO levels
on the highly transient, yet low horsepower New York non-freeway
segments, are significant. (Mass CO emissions were 34.5 grams for
high-powered Bag 7(LA Freeway), 12.0 grams for highly-transient Bag
8, New York Non-Freeway)). Therefore, not only are high power
modes necessary to model the real world, but transient modes are
also significant in their effects on measured emissions at catalyst
levels of technology. Finally, the industry has demonstrated an
ability to optimize emission control performance on any given test
procedure. Caterpillar stated this fact rather bluntly in their
written comments:
"This circumvention may be unintentional, but manufacturers
have no choice but to design engines to meet whatever test is
prescribed. For this reason the cycle must be as represen-
tative as possible of real world operation." (Emphasis ours).
Furthermore, many comments were received claiming that the proposed
cycle defies modal analysis and would be very difficult to design
to. ECTD interprets this to mean that it is easier to design
around a test procedure than it is to design a clean engine.
Based upon this ability, the historical inability of any
steady-state procedure to consistently correlate with any transient
operation, and a rational engineering judgment firmly supported by
the available data that the inherent transient carburetion charac-
teristics of gasoline engines have a significant effect on emis-
sions, the ECTD technical staff concludes that no steady-state
procedure will be better than the proposed transient test for
gasoline engines.
The case for the transient diesel test rests on the cost-
effectiveness of additional HC control, to some extent on consid-
eration of future regulations, and on proof of the inadequacy of
the 13-mode for prediction of HC emissions at the 1.3 g/BHP-hr
level.
Test results for diesel engines tested at Southwest Research
* Cape-21 data indicated that 53.4 percent of all gasoline truck
operation was transient in nature: 26.2 percent accelerations and
27.2 decelerations.
** 1979 GM 292 1-6 retrofit with a single Englehardt catalyst.
-------
Institute are summarized in Table A-4. On the average, 13-mode HC
understates the transient HC by a factor of 2,5, CO Levels are low
enough to conclude that a change in test procedure is not justified
for CO control.
To assess the benefits of the proposed test procedure change
for diesels, this analysis makes use of 1979 diesel certification
data, industry testimony at the July 16-17, 1979 Public Hearings,
and engineering judgments and assumptions explained below.
To predict the incremental HC reductions over the average
diesel engine's useful life due to a change in test procedure, the
following information must be known, or predicted to a reasonable
degree of accuracy:
i) Current (1979) HC emission levels for all diesel engines,
as measured on the 13-mode test (these are available from
certification records).
ii) Projected 1979 sales for all HD diesel engines (confi-
dential certification records).
iii) All available transient HC data for 1979 diesel engines.
iv) An average BSFC for diesel engines (for our purposes, we
shall assume 0.430 lb/BHP-hr).*
v) Average on-road diesel fleet fuel economy (for our
purposes, we assume 5.6 mpg).**
To achieve a 1.3 g/BHP-hr transient standard, assuming a 10
percent AQL, Chapter 7 of the Regulatory Analysis determined that a
production target of .89 g/BHP-hr will be necessary. This target
level represents the low mileage production emission mean levels
necessary to assure with 90 percent confidence that 90 percent of
all engines selected for a Selective Enforcement Audit (based upon
a sample size of three engines) will meet the 1.3 g/BHP-hr HC
standards.
Table A-5 summarizes transient vs. 13-mode HC emissions on
engines tested at SwRI and Cummins.
Caterpillar has quantified the transient/13-mode ratio, based
upon SwRI and their own limited data, at 2.65. Furthermore,
Caterpillar also notes that this ratio increases as transient HC
decreases (i.e., at lower levels, the 13-mode increasingly under-
* Based upon SwRI transient data (13-mode data is comparable).
** 1978 sales-weighted fleet average (see "Cost Effectiveness,"
Chapter 7 of Regulatory Analysis).
-------
states transient HC). On this basis, if engines with transient HC
less than 0.8 (as presented in TabLe A-5) are excluded, the revised
transient/13-mode ratio becomes 2.40. Within the range of emis-
sions which will most likely have to be reduced for compliance with
the 1.3 g/BHP-hr transient standard, this appears to be a reason-
able estimate.
The final consideration is the emission reductions which would
have occurred if the 13—mode were retained and the standards
derived from the gasoline steady state test.. A 90 percent reduc-
tion from sales-weighted 9-raode data acquired in the 1969 baseline
program indicates a HC standard of 1.0 g/BHP-hr- For an accurate
assessment of reduction due solely to implementation of the trans-
ient procedure, any reduction needed to bring 13-mode HC below 1.0
g/BHP-hr could not be counted. Assuming a slightly smaller produc-
tion distribution as compared to that for the transient standard,
the manufacturers' likely target goal would be approximately 0.7
g/BHP-hr.* Table A-6 contains the tabulated summary of this
analysis.
Total sales-weighted grams HC/BHP-hr per truck is 0.318.
Based upon a density of 7.1 Ib/gallon for diesel fuel, an average
diesel fleet fuel economy of 5.6 mpg, an average useful life of
496,000 miles, and an average BSFC of .430 Ib/BHP-hr, then the
sales-weighted average lifetime reduction attributable only to a
change in test procedure equals .316 g/BHP-hr/.430 Ib/BHP-hr x 7.1
Ib/gallon x 1/5.6 mpg x 475,000 miles x 1/454 g/lb x 1/2000 Ib/ton
= .49 tons.
Based upon our analysis of the expected benefits versus the
cost of compliance for HC using a transient test, the cost effec-
tiveness of this strategy for diesel engines attributable solely to
the change in test procedure is $77/ton, assuming a 10 percent
AQL.**
Other considerations cited in the Draft Regulatory Analysis
for the implementation of a transient procedure for diesel engine
certification still apply. Due to the anticipated difficulty in
attaining forthcoming NOx and particulate standards, a transient
test will be even more important in assuring compliance with these
regulations. The legitimacy of EPA's "future regulations" argu-
ment was severely questioned. There is merit in the manufac-
turers' arguments that unproposed regulations cannot justify
present proposals; justification must rely on the technical merit
and cost effectiveness of the proposed procedure. However, it
would be technically nearsighted for the Agency to ignore the
potentially significant impact of future proposals either mandated
* This assumes that the standard deviation of production emis-
sion means is proportional to the level of the emission standard,-
as testified by various manufacturers.
** See Economic Impact Analysis.
13
-------
by Congress (75 percent NOx reduction in 1985) or currently under
preparation by the Agency (particulates). It is not responsible
for EPA to leave in place a procedure we know full well to be
inadequate now, and which is expected to be even worse in the
future. Furthermore, diesel engines are increasingly being
equipped with turbochargers; it is well known that turbocharger
performance is influenced by transient engine operation. Increased
dieselization of the future fleet due to fuel economy pressures is
also cause for ECTD's concern, especially in the market segment
consisting of smaller, higher-speed diesels. Caterpillar testified
in their supplementary written statement to the July 16, 1979
Public Hearings that smaller, higher-speed diesels tend to produce
higher levels of HC. In relation to their 3208, Dina family 3
engine, Caterpillar stated:
"...This engine has a greater surface-to-volume ratio in the
combustion chamber because of a smaller displacement and
stroke-to-bore ratio. Finally, and most importantly, this
engine operates at higher speeds than other engines in our
product line. The fact that smaller high speed diesel engines
produce higher HC emissions was also voiced by Cummins Engine
Company in their testimony at the July public hearing.
Such engines are generally used to replace gasoline engines
(which operate at high speeds) to achieve significantly
improved fuel consumption. Small high speed diesel truck
engines represent a significant fraction of all diesel truck
engines manufactured in this country."
Not mentioned was the fact that economic pressures demanding higher
fuel economy will increase the market share of smaller diesels,
effectively increasing their impact on overall air quality prob-
lems.
In summary, continued use of the 13-mode for certification to
the proposed standards would result in an understatement of the
hydrocarbons generated during transient operation in an urban
environment; this is the basis for EPA's claim that use of a
transient procedure would result in greater confidence of a true 90
percent reduction. The fact that additional HC control will
actually occur was vividly illustrated by the testimony and com-
ments of several manufacturers; it was claimed that several of each
manufacturer's engine family's would not meet the proposed trans-
ient HC standard (i.e., additional emission control is required).
Finally, it has been demonstrated above that the additional emis-
sion control attributable only to the change in test procedure will
be cost effective.
The question now arises: can an equivalent degree of emission
control be predicted by a more stringent standard based upon the
13-mode procedure? Data from SwRI and Cummins is presented in
Figure A-3. Correlation between transient and 13-mode BSHC emis-*
-------
sions is weak. In their testimony in the July 16-17 Public Hear-
ings on the proposed standards, Cummins made the statement that
they have been unable to derive a correlation between the two test
procedures. The predictive value of the 13-mode test procedure
disappears at emission levels at and below the proposed standard,
and consequently is not a viable test procedural alternative.
The only remaining alternative is a new steady-state procedure.
The only data available to the Agency at this time (Report 7, Table
A-2) (Exhibit A-3) suggest that this is not acceptable, especially
in light of the increasing stringency of other standards.
To summarize the Agency's analysis of comments pertaining to
justification of the transient test, on the basis of engineering
judgment, on-road studies, and laboratory testing, it is believed
that current steady-state test procedures are overly simplistic and
inadequate predictors of actual in-use emissions. It has been
demonstrated that concrete air quality benefits will result from
implementation of a transient test, and that the change in test
procedure is cost effective and economically justifiable.
b. Representativeness of the Proposed Test Cycles
Inherent in the justification for any test procedure is
the requirement that the procedure be representative of real world
usage. EPA's derivation of the transient procedure from the
Cape-21 data base was harshly criticized as producing an unrepre-
sentative result.
The manufacturers claim that the engine torque estimation
techniques (i.e., using manifold vacuum, rail pressure, and rack
position) were invalid during transient operation, over 53 percent
of the Cape-21 data base.
First of all, it must be noted that at the time of the Cape-21
study, the use of manifold vacuum, rail pressure, and rack position
were generally accepted methods of engine load estimation within
the heavy-duty industry. Secondly, this very approach was recom-
mended by the CRC-EPA Joint Committee, primarily in light of the
fact that no practical alternative existed.
The major fault of these estimation techniques is the inherent
time delay between a change in the measured parameter and a change
in engine output load, the extent of that delay depending on the
rapidity of the change. For example, the use of manifold vacuum
for gasoline engines anticipates by some fraction of a second the
output shaft torque; opening of the throttle plate immediately
changes the vacuum level in the carburetor throat, yet an increment
of time is necessary for that change in pressure to manifest itself
in increased fuel flow, travel of the new charge through the intake
manifold into the cyclinder, compression of the charge, and smooth
power transmission during the combustion expansion cycle. To
-------
measure instantaneous engine flywheel torque, the accuracy of the
measurement within time depends upon the proximity of the measured
parameter to the output shaft.
The ideal situation is the actual measurement of shaft torque
by use of a shaft torquemeter. This approach was rejected for
Cape-21 as being impractical and cost prohibitive. For a truly
representative sample to be obtained in Cape-21, observation of
actual in-use commercial trucks was a necessity. In order for EPA
to acquire, instrument, and deinstrument these trucks, they .had to
be taken out of service; one of the conditions of the usage agree-
ments with the vehicle's owners was the fact that the vehicles'
time out-of-service be limited to overnight. Installation of
driveshaft torque meters, entailing custom driveshaft modifications
for .each individual truck, would have been time-consuming and
expensive, and thereby would have limited those trucks available
for EPA study. Furthermore, as outlined in Report 6, Table A-7, p.
30, commercially available torquemeters had a tendency to oscillate
under transient conditions on the road, resulting in technical
problems as well.
The most technically correct compromise involves measurement
of a load factor parameter as close as possible to the output
shaft. No measurable parameters were closer in time to the output
shaft than manifold vacuum for gasoline engines, and rail pressure
and rack position for diesels. (It could be argued that direct
measurement of instantaneous fuel flow rate would have resulted in
a slightly smaller time delay for diesels. ECTD does not view this
as a significantly better alternative, however, due to the absence
of accurate fuel flow instrumentation which could be readily
installed on a vehicle. In fact, the parameters measured are
themselves excellent indicators of instantaneous fuel flow.) For
gasoline and naturally-aspirated diesel engines in on-road vehi-
cles, the' time delay associated with these measured load factor
parameters is on the order of less than one second. ECTD does not
consider these small time lags (and resulting reference torque
overestimations (see below)) as large enough to invalidate the data
base, and ECTD considers the load factors measured as reasonable
estimates of shaft torque.*
* Cape-21 pressure transducers were installed on gasoline
engines at the EPA Motor Vehicle Laboratory, and a continuous
record of manifold vacuum vs. shaft feedback torque was taken. In
the case of a radical step change of engine speed and load, shaft
torque lagged manifold vacuum by as much as two seconds; however, a
. step change on a dynamometer is considerably more drastic service
than that seen on the road. During running of the highlyrtransient
New York Non-Freeway cycle, indicative of transient, in-vehiele
operation, at no time did feedback shaft torque lag manifold vacuum
by more than three tenths of a_ second.
-------
In the case of turbocharged engines, however, longer time
delays are present due to the added inertia of the turbocharger
itself. This had been observed at SwRI by the apparent "slug-
gishness" of turbocharged diesels over the transient cycle, and by
the lower totals of integrated brake horsepower-hour achievable
over the cycle while remaining within the valid regression limits.
It will be useful at this point to detail what was actually mea-
sured in Cape-21, and its resultant impact on the proposed cycle.
Consider an instrumented Cape-21 turbocharged diesel truck
cruising with constant engine parameters. The driver opens the
rack to affect a vehicle acceleration. The instrumentation records
the open rack as an increase in shaft torque although the increase
in actual shaft torque is delayed. Yet an actual vehicle accelera-
tion, as recorded by a change in vehicle speed, must wait until
that actual shaft torque is available. For any given acceleration,
opening of the rack preceded shaft torque and vehicle acceleration
by a constant time increment. This effectively anticipated shaft
torque has been carried over into the transient diesel cycle. Note
the percent torque commanded at 25, 26, 214, 215, 321,. 322, 377,
558, 927, 928, 1,116, and 1,117 seconds of the proposed diesel
cycle. In these cases, acceleration from idle is preceded by a
full one to two seconds of open rack—precisely what would happen
in real life, until the engine/vehicle accelerates. The cycle
performance criteria, however; do not penalize an engine for this
real lag. That the reference cycle itself overestimates the torque
achievable is indicated by the high integrated brake-horsepower-
hour levels. Yet, what actually occurs during an emission test
over the proposed cycle, as has been observed at SwRI, is that the
engine undershoots the integrated brake-horsepower-hour target by
as much as 15 percent (the validation limit) per the performance
capability of the engine, while the torque regression line sta-
tistics approach the upper limit of validation (reflecting non-
penalization of the engine for "sluggishness" and the accumulation
of high torque points at relatively high power non-turbocharger or
power lag affected points). The net affect is a "validation
window" within which the engine performance is permitted to fall.
It is ECTD's technical opinion that although the proposed reference
cycle may overestimate transient torque available—especially for
turbocharged diesels—the actual cycle the test engine will 'follow
in response to this command cycle is indicative of what was ob-
served on the road in Cape-21, and is not unrepresentative due
solely to turbocharger or nonturbocharger lag. (The relative
amount of turbochargers present in the Cape-21 data base is evalu-
ated later in this analysis.)
It was claimed that the cycle hot/cold weighting factors were
unrepresentative of the data base. This claim is made based upon
Report 11, Table A-7, which when subjected to Ford Motor Company's
and MVMA's analyses, yielded significantly lower cold start weight-
ings. It is our judgment that Ford's analysis misinterpreted the"
'7
-------
data. Were the report to be read more carefully, it would be
ascertainable that the 1.6 percent of gasoline operation quoted by
Ford and MVMA represented that portion of operation which was
"truly cold." Based upon the warm-up characteristics of the
Cape-21 trucks, there was 100 percent certainty that engine temp-
erature had not stabilized (i.e., 1.6 represents a lower bound of
cold operation for gasoline trucks); 3.0 percent was the lower
bound for diesels. Also in the report and overlooked by Ford and
MVMA. was an upper bound above which there was 100 percent certainty
that the engine temperature had stabilized, and as such represented
the upper limit of percent cold operation. (The upper bounds were
7.4 percent for gasoline and 11.1 percent for diesels.) The
report's objective was to quantify these ranges, knowing that the
average percent of cold operation lay somewhere between. This
study served to verify the accuracy of the weighting factors which
were derived in another report (Report 13, Table A-7), and also
served to characterize the fact that operational differences
between cold and hot trucks were negligible. The applicable
Cape-21 data from that report is summarized below:
Median Trips Median Trip
Per Day Length (Mins.)
Sample Size Sample Size
LA NY Weighted Avg. IA NY Weighted Avg.
Diesel Trucks 5.93 2.92 4.43 27 26 26
Gasoline Trucks 7.83 10.38 9.06 12 8 10
Assuming a nominal engine warm-up time of 5 minutes, as did
Ford in their analysis, the average percent cold operation derived
from the above data is:
Gasoline: 100% x 5 min. x 1/(9'.06 x 10) = 5.5%.
Diesel: 100% x 5 min. x l/(4.43 x 26) '= 4.3%.
The proposed test procedure, assuming again a 5-minute warm-up
time, yields:
Gasoline: 100% x 1/7 (5 min. x 60 sec/tain)/ 1167 » 3.7%
Diesel: 100% x 1/7 (5 min. x 60 sec/min)/ 1199 - 3.6%
In point of fact, EPA's cold weighting for the transient test
procedure is very close to that observed in the Cape-21 data, and
contrary to the comments received, slightly understates the cold
emissions.
Manufacturers claimed that the proposed cycles contained
unrepresentative overspeed. However, the overspeed in the cycles
-------
is indicative of the overspeed actually observed on the roads of
Los Angeles and New York. The EPA technical report on cycle
selection (Report 12, Table A-7) explained the meticulous screening
process for candidate cycles which was performed to ensure that
their speed and power distributions accurately resembled the data
base. Figures A-4 to A-15 are reproduced from this report and show
the actual distribution of speeds and powers for both the derived
cycles and the Cape-21 data base input. (A candidate cycle, the
New York Freeway, did exhibit excessive amounts of overspeed; this
was eventually traced to an erroneously recorded rated speed for
the engine in NY Truck 9.* The cycle was discarded.) ECTD finds
no basis for the claim of existence of unrepresentative speeds in
the proposed cycles.
The industry harshly criticized the cycle generation tech-
nique, whereby pattern sequencing of greater than one second was
ignored (i.e., use of the one-second Monte Carlo technique).
It is acknowledged that the Monte Carlo cycle generation
technique did not result in the characteristic rpm/load traces
normally seen during vehicle accelerations as the driver upshifts
through the several gears. The proposed engine cycles were deve-
loped in the following way: Each % rpm/% power pair observed
during the Cape-21 study was assigned a "transition probability"
for every % rpm/% power pair in the data base (i.e., the change
from Engine Condition A to Engine Condition B was assigned a
definite probability of occurrence). These probabilities reflected
the frequency of occurrence observed in the Cape-21 data base. The
cycle generation technique produced a continual progression of
engine transitions which accurately represent the actual transi-
tions and their frequency of occurrence observed in the 88 trucks
in New York and Los Angeles. Additional statistical tests were
performed on the numerous cycles generated to determine which
cycles were "closest" to the data base in terms of several other
parameters. (See Report 12, Table A-7 for a more detailed discus-
sion of the final selection procedure). The cycles eventually
selected are statistically equivalent in composition to the data
base, and every pertinent engine state and engine state transition
observed in Cape-21 is accurately represented. On the overall
question of use of the Monte Carlo technique for cycle generation,
Malcolm Smith in his June 7, 1979 submission to J. Hafele of
Caterpillar declared "...unrealistic rpm excursions have a very
small frequency of occurrence and hence have a very low probability
of being selected during the cycle development process."
The claim was made that a large percentage of the recorded
* 6.9 percent of the New York non-freeway cycle data base stems
from New York Truck 09. Assuming a 20 percent error in rated speed
carried through 6.9 percent of the cycle results in less than 1.4
percent error throughout the entire cycle.
-------
Cape-21 data was "spurious," arising from electrical and vibra-
tional "noise" and not actual truck operation. In particular,
Professor Meyers questioned the accuracy of the vehicle speed and
rack position indicators, and argued that "spikes" observable
on rpm versus vehicle speed plots were indicative of erroneous
instrumentation. EPA was also criticized for purported failure to
calibrate the instrumentation after each truck day.
First of all, the instrumentation and methodology developed
for the Los Angeles Cape-21 was identical to that used in New York,
and developed by the industry and EPA. Secondly, in EPA's two
year's experience with the instrumentation, the only occasion that
highly erratic, "spurious" engine/vehicle parameters were recorded
was the occasion outlined in the appropriate report (Report 9,
Table A-7, p. 40) (i.e., the vehicle speed encoder at zero speed).
Professor Meyers testified that he believed that this encoder gave
off erratic signals at all speeds. This was not the case. The
report clearly states the problem and its resolution:
"The optical encoder, however, because of vibration when the
vehicle was idling, sometimes recorded speeds as high as 15 mph
when the vehicle speed was zero. That is, a vehicle might stop
where the target was partly in view of the encoder window. As a
result of vehicle vibration, the target would oscillate into and
out of the window, generating an input signal to Channel 5. The
tach generator was, therefore, put back in the system and its
output fed to Channel 10. Thus, when the vehicle was moving,
Channel 5 gave an accurate measure of vehicle s»eed. When the
vehicle was stopped, the tach generator gave a reliable zero-speed
signal in Channel 10. Therefore, whenever Channel 10 showed zero
speed, Channel 5 was zeroed."
This "fix" was also accomplished in the LA study (Report 1,
Table A-7, p. 21). There is no basis for the assertion that the
chopped wheel, which was coupled directly to the speedometer drive,
gave erratic signals when rotating (i.e., the vibration produced
speed voltages only at zero speed). Finally, Report 8, Table A-7
summarizes Olson Laboratories' report to the MVMA concerning EPA's
Cape-21 data collection techniques. Conclusion 7, p. 1-3 states
"...The vehicle speed measured with the EPA instrumentation cor-
related well with the speeds measured with a 5th wheel..."
On the overall question of sensor integrity, adequate pre-
cautions were taken to ensure accurate recordings of vehicle
parameters. Refer to Report 6, Table A-7, pp. 103-104, in which
the four separate instrumentation validation procedures were
described. These procedures included visual checkout of transducer
output every half hour during truck operation, and three distinct
checks for unusual signal variation during and after the raw data
tape transcription process. Anything out of order was immediately
corrected or thrown out after-the-fact. In the New York study,
-------
instrumentation was calibrated both before and at the end of each
truck day (Report 6, Table A-7, pp. 34-38) and also prior to
installation in a new truck (Report 6, Table A-7, p. 39). Cali-
bration at the end of each truck day was not performed in LA due to
time constraints (the vehicles had to be returned to their owners
immediately after testing); however, transducer calibrations were
monitored from truck-to-truck. No truck-to-truck changes were
observed, leading ECTD to believe that no significant calibration
discrepancies occurred during a single day.
Finally, both Professor Meyers and Caterpillar attempted to
quantify the "spurious" content of the Cape-21 data. They assumed
that all out-of-range points were "spurious," and that these
out-of-range points deviated by a constant percentage from the true
signal. Extrapolating this constant percentage throughout the
entire data base, an estimate of total spurious points was made.
A review of the data collection and editing processes is
necessary to adequately respond to the above analysis. Prior to
on-road data gathering, each truck was instrumented and run on a
chassis dynamometer. On the dynamometer, the engine was "mapped,"
(i.e., at 250 rpm increments across the entire range of engine
speeds). Fifteen levels of engine torque were measured at each
speed. The result was the matrix graphed in Figure A-16 (referred
to in Cape-21 as an EVSL matrix). As explained in Report 9, Table
A-7, p. 49, 100 percent load for any truck was approximated as a
linear function of speed. Measured on-road load factors were thus
"normalized" during the editing process, i.e., expressed as per-
centages of this approximated 100 percent load. (The same was also
done for the zero-load function.) This linear interpolation of
maximum load was not without error. In actuality, a maximum load
function is curved; this linear interpolation therefore approxi-
mates the maximum torque available at all speeds. It is not
surprising that actually measured on-road load factors sometimes
exceeded this 100 percent approximation (see Figure A-16).
Table A-8 presents the actual edited output for a particular
truck. This is the output analyzed by Professor Meyers in esti-
mating the "spurious" content of the signal. A step-by-step review
of the editing record is warranted:
i) RPM above MAXIMUM: This represents the number of points
exceeding 150 percent rated speed. (150 percent was an arbitrarily
chosen cut-off point. As explained above, it is believed that
overspeed in the data is indicative of overspeed actually occuring
on the road). 150 percent was intended as a gross check on the
data; note that no points were deleted.
ii) RPM between 0 and 300: 300 rpm was arbitrarily chosen as
the minimum speed at which continuous engine operation was reason-
able. This by no means biased the cycle since actual time at a
-------
given speed determined the probability of the speed ending up in
the cycle. Furthermore, speeds measured below 300 rpm (7 points
for this summary) are reasonable; engine starts and stops were
included in the data base and it's only logical to presume that
some engine speeds were measured during starting and stopping.
iii) RPM below 0: This would be a fair indication of "spur-
ious" data. Note that there are none.
iv) Load factor above 100 percent plus 13mV: To allow for
small transducer calibration variations, transducer feedback of 13
millivolts in excess of 100 percent were allowed to be retained in
the data base. For this edit summary, 71 points were measured in
excess of this upper limit. As outlined above, ECTD believes that
these out-of-range points were actually achieved on the road and
were only considered excessive due to the conservatism of the
linearly interpolated EVSL model. Another possibility for on-road
out-of-range maximum loads was inherent in the chassis dynamometer;
any tire slip at maximum power would result in less torque being
measured at the engine for that speed. The chassis dynamometer
therefore may have slightly underestimated maximum available
torque. These underestimations are not serious, however, as
evidenced by the small number of on-road points actually measured
in excess of the model (.4%).
v) Load factor between 100 percent and 100 percent plus 13
mV: See D.
vi) Load factor between 0% load and 0% load minus 13 mV:
Similar to the maximum load method, 13 millivolts were also allowed
for transducer variation below the minimum load factor line.
vii) Load factor below 0% minus 13 mV: Both F and G reveal
the number of on-road torque parameters (i.e., manifold vacuums,
rack displacements, etc.) measured to produce less than zero power.
Should random spurious signals exist as claimed, it is only logical
to presume they would be evident here also. Note that none exist.
vii) Speed above 70 MPH: It was arbitrarily decided that
.truck operation over 70 mph would not be used for cycle development
purposes. Any points measured in excess of 70 mph were discarded.
ix) Speed negative while.moving: This would also be a good
measure of random "spurious" signal noise, yet only one point was
measured. It is reasonable to assume that for at least one second
during the entire day, the truck actually rolled backwards (e.g.,
while engaging the clutch from stop on an uphill grade).
x) Delta speed exceeding AMAX: EPA developed a theoretical
maximum acceleration model for use as a check on the on-road speed
data. It was highly conservative, assuming low vehicle mass and
-------
maximum engine torque, and was meant to eliminate obviously impos-
sible accelerations in the development of chassis dynamometer
cycles. (Decelerations were not characterized, nor worried about,
in the data base simply because decelerations are not solely
attributable to engine motoring, but primarily to braking.)
Only 0.6% of all accelerations were thrown out as unrealistic.
(This assumes 36 discarded points, 19,680 total records, and 30.2%
of the driving time - the LA gasoline truck average - spent in
accelerations.)
xi) Speeds to be zeroed: As discussed above, this is simply
the number of speed points registered as positive due to vibra-
tional excitation of a stopped optical encoder, and registered as
zero by the tach generator; this occurred only when the vehicle was
at rest.
The remainder of the edit output in Table A-8 relates which
points were interpolated or eliminated. (The interpolation and
elimination process is described in Report 9, Table A-7, p. 41.)
As explained above, Professor Meyers and Caterpillar assumed
out-of-range points were generated by random signal noise. In
point of fact, this "noise" was never observed during sensor
calibration on the chassis dynamometer when significant vehicle
vibration was present; all sensors when observed at steady-state
engine conditions yielded steady and repeatable results. (Elec-
trical ignition noise was an initial problem, however, on gasoline
trucks. This noise was eliminated by the use of capacitors and
spark suppressors in the ignition system, and the use of a separate
power supply for the data logger and support instrumentation.)
In summary, no random "spurious" signal noise was observed during
calibration, and as presented above, out-of-range data was in all
likelihood physically real and represented the slight inaccuracy of
the load factor models. ECTD has confidence in the accuracy of the
recorded data and rejects any claim of significant spurious con-
tent. Engine operational parameters used for cycle development
were those actually measured on the road.
The comment was made that future diesel engines will be
exclusively turbocharged, and that the proposed cycles were devel-
oped from a non-turbocharged data base. In point of fact, some
turbocharged engines were included in the data base, yet it must be
conceded that the percentage of turbocharged engines in the Cape-21
study (21 percent) was significantly less than that anticipated in
the future. EPA is faced with the practical problem of studying a
constantly changing fleet; given a finite time interval required to
analyze data and produce cycles, the present and future fleet will
always be different from that which was studied. ECTD believes
that the test cycles are representative of the data base studied.
ECTD has no data, however^ which shows that turbocharged truck.'s
usage patterns are different enough from non-turbocharged vehicles
-------
Co warrant abandonment of this transient test. However, future
modification of the test cycles to incorporate the performance
characteristics of new technologies is by no means precluded
by retention of the proposal at this time.
It has also been claimed that discrepancies exist between
operational mode distributions in Cape-21 and the distribution in
the proposed diesel cycles. Table A-9 details the time distribu-
tion of the four major modes.
These comments arose due to erroneously compiled data pub-
lished in the Draft Regulatory Analysis. Table VI-C, p. 122 of the
DRA simply presented the arithmetic mean of the tabulated modal
data from Cape-21, without regard for freeway/non-freeway weighting
factors. Reports 12 and 13, Table A-7, present an in-depth discus-
sion of the cycle selection and weighting procedures; however, data
from these reports are presented in Table A-10, and summarized in
Table A-9. The Cape-21 weighting factors were derived from the
amount of time individual trucks spent in freeway/non-freeway
operation; the actual cycle weighting factors were derived from
the length (in time) of each segment relative to the total cycle
length. Upon examination of Tables A-9 and A-10, it should be
obvious that the only gross discrepancies between modal percentages
arise from the steady-state procedures. The proposed cycle ade-
quately represents the operational mode distribution observed in
the Cape-21 data. (The largest deviation' between Cape-21 and the
proposed cycles lies in the 5.1 percent difference in idle percen-
tages on the diesel cycle. This corresponds to .051 x 1,199 = 61
seconds of additional idle. For the dirtiest (by a factor of two)
diesel engines tested to date at SwRI, a.1978 Caterpillar 3208 with
an idle HC emission rate of 38 grams/hour, the total effect of the
additional 60 seconds of idle is insignificant—approximately .027
grams/BHP-hr*-less than the tests' round-off error.
It was claimed that engine power was overstated due to lack of
consideration of engine inertia in computing acceleration rates.
In point of fact, engine inertia was ignored in the study and
resulting cycle development. Engine crankcase torque is composed
of four components:
i) Torque necessary to accelerate inertial masses (including
vehicle, wheels, and drivetrain, but not the engine**);
ii) Torque due to gravitational effects on vehicle mass while
on gradients;
*38 grams/hr x 60 seconds x 1/3600 hours/second = .63 ad-
ditional grams. . .63/23.3 BHP-hr (transient integrated BHP-hr for a
1978 Cat. 3208) = .027 g/BHP-hr difference.
** Torque arising due to engine inertia can only be observed at
the driveshaft when the engine is being driven, not when the engine
is driving.
-------
iii) Torque necessary to overcome driveline and tire-to-road
friction, and
iv) Torque necessary to overcome aerodynamic friction.
Measurement of the engine load parameter at the engine re-
sulted in measurement of the sum of these four components; the
magnitude of any individual component was irrelevant. Acceleration
rates for the cycle were determined by the transition probability
matrix from the accumulated Cape-21 data (i.e., the probability of
a given change in engine speed is directly dependent upon the
observed frequency of occurrence of that change in speed in the
Cape-21 data base).
Engine inertia and available power pertain more directly to
operation of a given engine over the proposed cycle itself.
Comments were made that engines would be unable to follow the
speed cycle without the help of a motoring dynamometer; engine
inertia was too high and available horsepower too low.
Engine driveability over the transient cycle is engine-
specific. Based upon observations and cycle performance summaries
of engines run on the transient cycle at EPA's laboratory, most
engines are capable of generating sufficient torque during speed
accelerations to produce residual positive torque at the drive-
shaft. Some engines, however; when subjected to a step change in
throttle position (i.e., a step acceleration command, tend to
stumble, develop insufficient torque, and are motored up to speed
by the dynamometer. This occurs on a minority of engines and at
worst 10-15 places during a test. Assuming 3 seconds per stumble,
this corresponds to less than 2 percent of the test. The remaining
98 percent of the time the engine drives properly and represen-
tatively. Not only is this true for carbureted gasoline engines,
but also for the most sluggish turbocharged diesels observed at
SwRI.
The prime cause of this performance lag phenomenon is most
likely the inherent driveability characteristics of individual
engines. It must be reiterated that it is an average cycle. The
transition probability matrix used in cycle development determined
a probable transition from one speed/torque pair to another, based
upon frequencies of occurrence of such transitions in the data
base. The data base was varied (i.e., different engines were used
in different vehicles in different applications). This is further
complicated by the effects of vehicle speed and gear ratios, up and
downhill operation, and by the presence of many different engines
in many different vehicles. The speed/load transitions in the test
cycles accurately represent the frequency of those transitions'
occurrence in the Cape-21 study. It is quite probable, however,
that transitions arising from an engine in a lightly-loaded truck
in low gear in the Cape-21 fleet are present in the cycles, and
-------
quite frankly overstate the performance capability of another
engine on a test stand trying to follow the accelerations and
demanded torques under its own power. This has been complicated by
the use of throttle parameters as engine power approximations
during the Cape-21 study. As discussed earlier; however, the
inherent time lag between throttle movement and a torque increase
is represented in the cycle itself (as evidenced by engine speed
accelerations lagging torque accelerations by one to two seconds),
and is not so drastic as to prohibit the majority of engines from
following the cycle under their own power. The few which can't are
motored for 2 to 3 seconds several times during the test. For a
test whose most significant emission contributions come from the
high-power, low-transient LA-freeway. these few points (at worst
less than 2 percent of the cycle) are insignificant contributors to
the final emission test results. It is emphasized, however, that
the test procedure validation criteria do not penalize an engine
for inability to develop full torque during steep accelerations.
Caterpillar testified that certain portions of the proposed
cycle violated laws of mechanics, in particular where 100 percent
power occurred simultaneously with an engine acceleration, the
point being made that no power remains to cause an acceleration.
This is indicative of confusion in the industry between road load
horsepower and engine horsepower. Operation at 100 percent road
load horsepower would indeed curtail desired accelerations;
EPA's proposed cycle, however, is designed around engine horse-
power. Consider a vehicle traveling at an intermediate speed at an
intermediate level of power. Calling for wide-open throttle at
this point results in 100 percent engine power, a portion of which
continues to nullify road load resistance and the remaining portion
represents the inertial power leading to vehicle acceleration. As
clarified above, the proposed cycle makes no distinction between
the four components of engine torque. Assuming 100 percent engine
power operation, an increase in the inertial power component (i.e.,
an acceleration) results simply from a decrease in one or more of
the remaining road load components (e.g., coming over the crest of
a hill).
In summary, ECTD recognizes the fact that some engines may
have a more difficult time developing torque during certain accele-
rations; this is inherent in the technical compromise resulting
from development of an average cycle. As mentioned in the earlier
discussion on cycle development history, this average cycle per-
formed on an engine dynamometer is the only practical certification
method available. Those speed/load transitions asked for in the
cycle are representative of those seen in real life. ECTD has
modified the test validation criteria to forestall penalizing any
engine incapable of following these few accelerations. Most
importantly and contrary to comments received, at no times are
physical laws of mechanics violated nor does the dynamometer drive .
the engine through accelerations, except for the few cases outlined
above.
-------
Certain comments were directed at EPA's selection of only two
cities for the road study, and the use of only urban driving. It
can only be reiterated that time and resource constraints limit the
scope of any research project, be it marketing research, election
voter polls, or on-road truck studies. It should be noted that the
industry-staffed Coordinating Research Council, in conjunction with
EPA, selected New York and Los Angeles as two cities representing
both the worst air pollution problems and the two most diverse in
terms of traffic flow and usage patterns. (Refer to Report 7,
Table A-7, for further discussion). Furthermore, mobile source
pollution is an urban problem and requires urban characterization.
The choice of equal weightings of New York and Los Angeles data was
a judgement made in the absence of any data or rationale to the
contrary.
Furthermore, the truck samples observed by EPA in New York and
Los Angeles were claimed to be unrepresentative of the urban and
national truck populations with respect to vehicle GVW. ECTD
argues that this claim is inaccurate. Tables A-ll and A-12 depict
Cape-21 vs. U.S. yearly production GVW distributions for both all
trucks and diesel trucks only. Agreement between the sample and
the population percentages is good. Furthermore, Reports 2, 3, 6,
and 7 (Table A-7) document the truck population research and the
resulting sampling plans derived for Cape-21 for both New York and
Los Angeles. The truck sampling plans incorporated all relevant
truck characteristics as defined by the population studies (e.g.,
two axle, three axle, tractor-trailer^ Reports 2 and 3, Table
A-7), and the actual sample followed the sampling plans with little
deviation (Reports 6 and 7, Table A-7).
Several commenters criticized EPA for failure to include
trucks of GVW less than 10,000 pounds, claiming they represent a
majority of truck population. In point of fact, most trucks under
10,000 Ibs. GVW are classified as light-duty, and it was not EPA's
intention to generate driving cycles for light-duty vehicles. The
truck percentages of heavy-duty vehicles between 8,500-10,000 Ibs.
GVW are presented in Table A-ll. The physical difference between
trucks rated at 8,500-10,000 Ibs. and those rated 10,000-14,000
Ibs. GVW are small; in many cases they are identical vehicles. At
the time of the Cape-21 study, this weight class of trucks repre-
sented a small percentage of the total. (One reason why the per-
centage of HDV's less than 10,000 Ibs. GVW has increased in recent
years is that many light-duty trucks were rerated into the heavy-
duty class to escape the light-duty certification test procedure.)
Several vehicles rated exactly at 10,000 Ibs. GVW were included in
the Cape-21 study; these were included in the 10,000-14,000 Ibs.
class.
In summary, EPA went to great lengths to assure that the
sample observed in the Cape-21 study was as representative as
possible of the overall truck population. Furthermore, to assure
-------
that the survey would be even more representative, actual in-use
commercial vehicles driven by their owners in normal day-to-day
operation were used. ECTD is confident that the trucks sampled in
Cape-21 were highly representative of the real world.
One manufacturer ventured to comment that the distribution of
cycle acceleration rates in the proposed cycle was different from
those observed in individual truck data. These claims were made on
the basis of one truck's plotted operational data; it is no sur-
prise that any given truck's operational data could apppear dif-
ferent from an average cycle derived from the data contributed by a
total of 38 trucks. It has already been shown that the overall
percentages of operational mode distributions for Cape-21 and the
proposed cycle agree closely, and that the cycle's second-by-second
transitions accurately reflect the frequency of such transitions'
occurrence in the data base.
The horsepower models used by EPA in translating Cape-21 data
to power levels used in the proposed cycles came under attack. The
main criticism claimed inaccuracy of the mode-Is and too small a
data base from which the models were derived. In Report 8,* Table
A-7, Olson Laboratories reported to the MVMA that:
i) "The EPA finding for Cummins diesels that percent power
is a function only of percent fuel pressure, and is independent of
engine rptn, was validated by the engine-dynamometer data for the
two sample engines." (p. 1-2)
ii) "The EPA finding that percent power for gasoline engines
is equal to percent load factor (manifold vacuum), computed at each
engine rpm, was essentially validated by the engine dynamometer
data for the two sample engines." (p. 1-3)
The only discrepancy with the Cape-21 models discovered by
Olson in their analysis was the percent rack travel.procedure used
for'Detroit Diesel engines. Based upon dynamometer data for two
DDA engines, Olson concluded that the EPA model—derived from a
single DDA engine—understated percent power to a significant
degree. The error was on the order of 10 percent at 80 percent
power, and increased as the actual level of power decreased.
ECTD's original assumption in development of these models was
that a single model could be used for all engines utilizing the
same load factor parameter (i.e., rail pressure, manifold vacuum,
and rack position). This assumption was validated in the cases of
rail pressure (Cummins) and manifold vacuum (gasoline engines).
However, the engine-to-engine differences in rack position (DDA)
engines appears significant enough to cast doubt upon the universal
validity of any single model. At the time, however, ECTD's model
This study was initiated and funded by the MVMA.
-------
was based upon the best data available. (In fact, the data used by
ECTD was collected and supplied by DBA.) Due to engine-to-engine
differences, there is no assurance that the engines, tested by Olson
for MVMA are any more representative than the engines tested by
DDA, yet due to this variability, ECTD recognizes that some error
in the horsepower model may exist. It is, however, a predictable
error, occurring only at lower horsepowers where emissions are
impacted least and does not diminish the overall validity of the
test procedure. The error is further minimized by 50 percent of the
data base consisting of the accurately modeled Cummins engines. On
the whole, horsepower models used in converting Cape-21 data to
power levels for cycle development were accurate and relatable to
all engines; where inaccuracy occurred, its impact was minimized by
the factors listed above.
Finally, since the absolute level of motoring torque was never
measured in the Cape-21 study, its inclusion in the data base was
described as "guesswork." While motoring torque was never mea-
sured, we are confident when it occurred. (Each truck was instru-
mented to indicate throttle position.) The test procedure will be
proposed such that any cycle point described as motoring actually
means closed throttle or any negative torque command necessary to
achieve closed throttle.*
To summarize ECTD's analysis of comments pertaining to repre-
sentativeness of the proposed test cycles, it is acknowledged that
certain compromises were made, and indeed had to be made, in the
data collection and cycle generation processes if any practical
on-road study of truck usage patterns and cycle development pro-
grams were to be accomplished in a reasonable time at reasonable
cost. Many of these early compromises in data collection (e.g.,
manifold vacuum and rack position as approximations for flywheel
torque) were made by the industry-staffed Coordinating Research
Council; for the industry to demand that EPA do the impossible or
the impractical while they themselves claim "unreasonable burdens"
or "impossibility" is self-contradictory. Consider a statement
made by MVMA in its February 13, 1978 letter to J. DeKany (EPA-
former ECTD Division Director):
"...Considering the diversity of design and use of heavy-duty
gasoline powered vehicles, it is probable that no "represen-
tative" driving cycle exists. Heavy-duty vehicles include
ambulances, school buses, pickup trucks, cement mixers,
delivery vans and tractor/trailer hauling rigs. Each of these
classes of vehicles have drivetrains and usage patterns
peculiar to their design function..."
* See Summary and Analysis of Comments - M. Numerical Stan-
dards/Standards Derivation. An exception to this closed throttle
mode is the use of -10 percent torque command during motoring modes
for gasoline engine tests.
-------
"...No single cycle can accurately characterize the emissions
of these...types of vehicle usage."
MVMA went on to propose an alternate procedure, substantiated
by no on road data whatsoever, which it claimed:
"...Would encompass all emission related transient conditions
of engine operation...and be representative of real world
heavy-duty engine operation."
Recognizing the difficulty of establishing true "represen-
tativeness," and constrained by the requirements of a workable
certification program, ECTD has been convinced from the beginning
that use of a single certification test cycle required development
from an actual on-road data base to assure the maximum represen-
tativeness possible. ECTD is confident that in any subsequent
review, be it technical or legal, it can be proven that the judg-
ments and decisions made were sound, were based upon practical-
ities and the resulting data represents the most comprehensive and
the highest quality data available. Although a virtually infinite
number of transient cycles are possible, ECTD believes that those
chosen adequately represent the data base.
In most other cases, manufacturers' criticism of the data
collection and cycle development processes arose from a lack of
information, or misinterpretation of data or the processes them-
selves.
It is ECTD's judgment that the proposed transient cycles were
the best attainable within the resources of the Agency. The
Cape-21 survey was the largest and most ambitious road survey of
heavy-duty truck usage patterns in history. The resulting cycles
were not perfect and as such are subject to change; however, they
adequately represent usage of the average truck, and will predict
in-use emissions significantly better than any of the available
alternatives.
Aside from the cycles themselves, ECTD's method of dynamometer
control was also attacked as unrepresentative (i.e., use of speed
control versus the use of torque control). This will be elaborated
on in greater detail later (see "Caterpillar cycle"); suffice it to
say here that ECTD does not consider this dynamometer control
strategy a threat to representativeness, but rather a more expen-
sive option for the diesel industry.
Chrysler Corporation's comments pertaining to the chassis
testing of non-commercial, low-GVW HDV's have merit. The Agency,
however, is not in a position to adopt Chrysler's proposed re-
solution of this matter. Following the decision to pursue engine
dynamometer testing (see "Evaluation of Alternatives,' later in
this analysis).. ECTD diverted resources towards development of an
-------
engine procedure. No emission work has been performed to char-
acterize chassis vs. engine dynamometer cycle emission differences.
Furthermore, all transient data to date was derived from engine
testing. Two alternatives exist, however. See §86.083-27(b)(1) of
the Proposed Rules in which any manufacturer who feels the proposed
procedure is unsatisfactory for a given engine may prescribe a new
test procedure by written application to, and subject to approval
by, the Administrator. Furthermore, current regulations allow
optional certification on the light-duty chassis procedure for all
trucks rated at 10,000 Ib. GVW or less. This test is equivalently
stringent, yet consists of the lighter duty cycle for which Chry-
sler has argued.
c. Validation
Manufacturers raised the issue of on-road validation of the
transient cycles. It was argued that EPA could not conscionably
promulgate a transient test without this on-road study.
No on-road or chassis dynamometer study of emissions from late
model trucks has been performed, nor is one forthcoming to address
this issue.
The manufacturers have argued that the transient cycle must be
conclusively proven on-road. This is an acceptance criteria never
required for promulgation of the 9- or 13-mode. Furthermore, the
manufacturers recommended retention of the steady-state tests, one
of which (9-mode) has been proven to be grossly unrepresentative of
future levels of control both in-lab and on-road (Tables A-2 and
A-3), and the other (13-mode) deemed to be highly questionable at
less stringent emissions levels now, and expected to be even worse
in the future.
To argue for retention of unrepresentative, invalidated test
procedures on the premise of a need for on-road validation is
logically contradictory. It is the ECTD technical staff's judgment
that on-road validation is really a question of representativeness
and based on the following premise: the closer to real life
operation an engine is operated in the laboratory, the closer the
emissions measured in the laboratory will be to those actually
found on the road. Heavy-duty truck operation, and therefore
heavy-duty emissions, are application-specific. The objective of
CAPE-21 was to arrive at an "average" duty cycle for an "average"
urban truck. Any given application wouldn't necessarily correlate
with emissions measured on the average cycle. To choose an on-road
application identical in duty cycle to the test procedure itself
would prove nothing. A road validation of the average cycle would
require a CAPE-21-sized study, but of a significantly higher
complexity due to the need to measure emissions both on controlled
and precontrolled vehicles. Such an endeavor would delay promul-
-------
gation of a transient procedure for years, while retaining steady-
state procedures which ineffectively measure transient emissions,
which can easily be designed around, which guarantee no real world
emission reductions, and which are grossly inadequate for assuring
attainment of the mandated 90 percent reductions.
Furthermore, EPA has always recognized that on-road validation
of any single test cycle would be virtually impossible. For this
reason, the Agency took great pains to assure that the cycles
developed were based upon extensive on-road data and utilized
meticulous cycle generation processes to assure that emissions
generated over these cycles would be representative of in-use
vehicles.
In summary, while no actual on-road testing of the proposed
cycles has been performed, to do so would require a massive effort
of several years duration. To delay promulgation of a transient
procedure would assure non-attainment of the legislated 90 percent
reductions in the time desired. Moreover, the technical staff
believes the proposed cycles are sufficiently representative for
the reasons discussed above to guarantee that emission reductions
measured on the proposed procedure would be repeatable in real
life.
d. Inability to Comment
Another major issue raised by the commenters was the fact that
industry lacked transient testing experience and facilities; this
effectively hampered their ability to comment on the proposed
rules.
Inherent in this argument is the claim that the industry was
surprised and unable to respond adequately. Such is not the case;
the development of the transient test procedure has been well
publicized to the industry for the last seven years. From the
initial cooperation on CAPE-21 to the publication of the February
13, 1979 NPRM, the communicative interaction between industry and
EPA has been open, comprehensive, and deliberate:
i) Over 6 years prior to the February 13, 1979 NPRM, EPA and
the industry-staffed Coordinating Research Council (CRC) managed
a jointly-funded on-road truck usage study for the specific purpose
of deriving representative test procedures.
ii) Over 4 years before the NPRM, the then Deputy Assistant
Administrator, Eric Stork, chaired a meeting between the EPA staff
and the Engine Manufacturer's Association (EMA) in Ann Arbor,
Michigan on November 20, 1974. Exerpts from the meeting's record
include the following statements by EPA:
- "Advanced test procedures including representative urban
-------
engine cycles are under development and expected to be
proposed..."
"Gasoline and diesel engines are being treated separ-
ately in cycle development, however, both will be evaluated
against common standards..."
"Transient duty cycles are likely to result but EPA is
open to a modified steady-state test if transient operation
of diesel engines is not a significant influence on emis-
sions. "
- "Equipment associated with transient duty cycles will
likely be DC-motor generator dynamometers with multi-
channel tape control. Possibly, an eddy current absorber
with AC motor, also tape controlled would suffice."
iii) Over 28 months before the NPRM, a detailed briefing of
the industry covering all aspects of CAPE-21 and the test procedure
development programs was given by EPA staff members in Ann Arbor,
Michigan on September 30, 1976. In attendance were representatives
of the entire heavy-duty industry.
iv) Twenty-three months before the NPRM, at the March 17,
1977 EMA meeting in Ann Arbor, Michigan, EMA was briefed on the
status of the cycle development. EPA requested that EMA member
companies evaluate the transient control capabilities of dynamo-
meters at their own facilities. (This was eventually followed-up
by limited transient testing at Cummins Engine Company during the
summer of 1977.)
v) Twenty-one months before the NPRM, RFP No. CI 77-0147 was
released in May 1977, to solicit bids for a baseline testing
contract. The RFP's detailed Scope of Work outlined the procedural
details and the equipment needed to perform transient testing of
heavy-duty gasoline and diesel engines. At least one company,
General Motors, obtained a copy. The actual Scope of Work of the
eventual contract was sent to EMA in the September 7, 1978 letter
of R. Nash (EPA) to D. Carey (EMA).
vi) Twenty months before the NPRM, a June 21, 1977 letter
from the Deputy Assistant Administrator Eric Stork, to Thomas Young
of EMA declared, "The benefits of a transient procedure are suffi-
ciently attractive to commit us to the development and study of
transient cycles through a baseline emission program....[T]he
intent of this plan is to develop a transient procedure so that it
may be used in future regulations."
vii) Over 15 months before the NPRM, EPA requested production
statistics and sales data for 1969 MY gasoline engines from the
Motor Vehicle Manufacturers' Association (MVMA) in an October 20,
1977 meeting. EPA's explicitly declared intention was to use the
data in the design of a transient baseline testing program from
which emission standards would be derived.
-------
viii) Over 11 months before the NPRM, EPA began transient
baseline testing of 1969 and current technology gasoline engines in
March of 1978. Not only were manufacturers contacted for assis-
tance in providing engines and repairing malfunctioning parts, but
several manufacturers sent representatives to the MVEL to observe
the actual transient testing and testing facilities.
ix) Six months prior to the NPRM, the Draft Recommended
Practice for the proposed transient test procedure was published in
August 1978. The Foreword of the Recommended Practice stated,
"These procedures are expected to form the basis for
new test procedures that will be implemented concurrently
with the new, more stringent emission levels for 1983 model
year HD vehicles....A Notice of Proposed Rulemaking (NPRM)
incorporating the new standards and transient test procedure
will be forthcoming."
The procedure published in this draft document was practically
identical to that published in the NPRM. No comments from the
industry pertaining to the Recommended Practice were received.
In summary, the industry has no right to claim that it was not
informed of test procedural developments, nor that it has not had
the ability to comment, criticize and provide inputs to the proce-
dure development process. EPA has freely shared all transient
emission data acquired by the Agency, and its contractor. Any
inability of the industry to comment based upon the inadequacy of
their own facilities is largely self-imposed. As shown above, in
March of 1977 EPA requested EMA to evaluate transient dynamometer
capabilities, followed-up by limited transient testing at Cummins.
Since that time, the majority of the industry* has done little to
acquire transient capability.
Secondly, the lack of transient testing ability was not
necessarily a deprivation of due process.
The gasoline engine industry received the detailed information
and data acquired during all transient testing at EPA's laboratory.
Furthermore, the process of standards derivation was throughly
documentd and distributed to the industry before the Public Hear-
ings and well before the close of the comment period. (See Report
14, Table A-2.) Enclosed in this report were detailed information
* It is interesting to note that of all the companies affected
by this package, Cummins was the first to acquire transient testing
capabilities, and has been the most progressive company in terms of
emission control. This is reflected in the fact that 93 percent of
Cummins 1979 unit engine sales already comply with the proposed 10
percent AQL production targets necessary for compliance with a 1.3
g/BHP-hr HC standard, as compared to a 36 percent industry average.
-------
on the transient test procedure and associated equipment, and a
summary of EPA's accumulated experience with the procedure. In
terms of feasibility analysis, (the area in which an inability
to conduct transient testing would be most restrictive), the
gasoline engine industry has unanimously realized for some time
that catalyst technology will be required. The cycle information
included in the NPRM (which gives the specific power levels re-
quired by the transient procedure), the availability of transient
light-duty catalyst data, and the dissemination of transient
catalyst data resulting from testing at the EPA laboratory all
allowed the industry to make a well reasoned extrapolation of what
their product lines require to achieve compliance. In fact, the
major criticism embodied in the feasibility comments and the major
feasibility problem is not test procedure related, but the issue of
in-use durability of heavy-duty catalysts. In short, the gasoline
industry did have a basis for comment on the proposed rules, and
these comments have been addressed.
The diesel industry was also given the results of EPA's
transient data, which was acquired at SwRI. Furthermore, both
Cummins and Caterpillar were able to submit data during the comment
period which was acquired through actual transient testing at their
own facilities. As with the gasoline industry, the major area in
which an inability to test would restrict the ability to comment
would be feasibility analyses. Data acquired at SwRI allowed
a reasonable comparison of transient results relative to the
industry's steady state data base, and as stated above, two diesel
manufacturers were conducting transient tests on their product
lines. For those manufacturers who were not running transient
tests, data was made available to provide a basis for comment.
In short, the development of the transient test for heavy-duty
engines is a logical extension of the earlier rationale for requir-
ing a transient test for light-duty vehicles. The Agency has
freely communicated this intention since 1972. The Agency has
shared all available data, and openly broadcasted its intentions
and solicited comments through the years. In all fairness to the
diesel manufacturers, however, it has been EPA's position through-
out the cycle development process that should a steady-state
procedure yield comparable results with the transient, the simpler
procedure would be used. Transient diesel data has only recently
become available, resulting in a relatively late decision to retain
the transient test. However, diesel manufacturers were alerted in
the NPRM and we can only reiterate our position of having openly
disseminated all data upon its availability, and having openly
announced our desire through the years to implement a transient
procedure if one were warranted.
e. Evaluation of Alternatives
Most commenters criticized EPA for purportedly ignoring
sr
-------
Executive Order 12044 by failing to consider simpler and less-
expensive alternatives to the complex transient procedure. ECTD
disputes this contention; Figure A-17 and the discussion below
outline ECTD's thought processes while evaluating test procedural
alternatives.
Use of a given test procedure entails an inherent compromise
between practicality and representativeness. For heavy-duty engine
emission testing, the ultimate procedural simplification is the use
of a steady-state test. Based upon several studies, in-house
and contractual testing, and reasonable technical judgments based
upon the preceding data and light-duty experience, EPA reached the
conclusion that the current steady-state procedures at present and
especially future levels of control are unjustifiably simplistic.
(See arguments for justification, earlier in this analysis.) In
short, the present tests are unacceptable alternatives.
Questions then arise as to the viability of modifying the
steady-state procedures (i.e., adding and/or reweighting steady-
state modes to produce a more representative test. A 23-mode test
procedure was performed in 1972 on 8 gasoline and one diesel
engines (Table A-2, Report 3, Part III); it was concluded that the
additional modes did little to improve the test. Furthermore, the
results of the "Sensitivity Study," (Table 2, Report A-7) indicated
that reweighting of any steady-state test could not achieve con-
sistently correlatable results with emissions generated over a
variety of transient tests (i.e., the probability of a reweighted
9- or 13-mode proving viable was minimal (see Exhibits A-2 and
A-3).)
Based upon this and upon experience with the industry, EPA
concludes that no steady-state test can remain valid as technology
progresses. Technology is developed based upon the test procedure,
and an overly simplistic test becomes Less realistic as more
technology is applied to certify on it. (A case in point is the
9-mode itself: Compare 9-mode vs. transient emissions for pre-
controlled engines and then current technology engines. The
discrepancies between the two procedures increased dramatically
with increasing control technology. Furthermore, both Cummins and
Caterpillar in the Public Hearings expressed no basic disagreement
with the concept of a transient test. In the evaluation of test
procedural alternatives, the Agency chose to use most of its
resources in that area where the highest probability of success
existed (i.e., a transient procedure based upon real world opera-
tional characteristics of trucks). To further pursue the steady-
state alternatives when those alternatives seemed inadequate
and represented approaches which were not likely to succeed, would
in all likelihood have delayed development of the representative
test procedure.
Having judged that a transient test is preferable, remaining
-------
alternatives to be considered are modification of the proposed
cycles, or the substitution of alternate cycles.
It is sufficient to state here that modifications to the
proposed cycle and procedure based upon comments received are being
responsibly considered in this Summary and Analysis, and if feas-
ible and meritorious, will be adopted. EPA is therefore addressing
its legal responsibility under Executive Order 12044 to evaluate
simpler and more cost-effective alternatives. (Specific modifi-
cation and procedural details are addressed later in this analy-
sis.) The question of alternate and presumably simpler transient
cycles (including chassis cycles) now arises.
Considerable internal debate within ECTD occurred over the
viability of a transient chassis procedure.* Performance of an
actual emissions test on a chassis dynamometer would be simple
(much like a light-duty LA-4), but the actual certification program
would by necessity be extremely complex and much more costly due to
the large number of engine/vehicle configurations and the need for
very large chassis dynamometers. A manufacturer would be required
to certify an engine for several different applications and vehicle
configurations. The MVMA realized this fact early, and in a
January 16, 1974 Discussion Paper presented to EPA on April 11,
1974, strongly advocated an engine dynamometer test.
Furthermore, EPA also considered the alternative of cer-
tification based upon numerous cycles, to be selected based upon a
vehicle's application. This, however, would have entailed a much
more burdensome certification program, and to minimize burdens on
the industry, EPA consolidated the cycles as much as was justi-
fiable.
Finally, in a February 13, 1978 letter to John DeKany of EPA,
MVMA proposed a simpler transient procedure. This letter repre-
sented the only in-depth transient procedure alternative submitted
to EPA during the entire cycle development process. It came at a
time when EPA was already running 1969 baseline tests (i.e., after
EPA had already devoted considerable money, time, and manpower to
development of the Cape-21 generated cycles, and to the Congres-
sionally-mandated uncontrolled baseline test program). ECTD's
formal response (letter of May 17, 1978, C. Gray (EPA) to H.
Weaver, MVMA) to this proposal follows:
"...(ECTD)... staff has reviewed MVMA1s proposed alternate
test cycle. It is their opinion that, even with appropriate
restructuring, it would not be representative of actual truck
operation. EPA recently ran a test program where transient
chassis cycles were "linearized." That is, steady state
* Issue papers from both the Heavy-Duty Group and the Cer-
tification Division were written, and are on file.
3?
-------
accelerations, cruises, and decelerations were used to dupli-
cate, to the extent possible, the full transient cycles.
Results were disappointing, with the linearized cycles giving
significantly less HC and CO. One must remember that, even
though the cycles were linearized, the engine was run through
several gear shifts with resulting rapid changes in speed and
torque. The MVMA cycle would be much milder, with smoothly
changing torque and speed. Based on our experiment, it is not
logical to assume that MVMA's cycle would yield results
equivalent to the fully transient cycle while maintaining the
long mode times MVMA desires. The MVMA cycle appears to be an
improvement over the 9 mode, but we are concerned that a cycle
with longer modal times might be easier to "design around"
than a fully transient cycle. Finally, EPA does not have
either the resources or the time (due to Clean Air Act re-
quirements) to investigate the alternate MVMA cycle."
Acceptance of MVMA's offer of alternate cycle development
would have effectively delayed promulgation of any transient test
without any assurance that the alternative cycle would yield
representive results.
In summary, EPA had neither the time nor resources to inves-
tigate every possible test procedure alternative. Regardless of
this fact, the ECTD staff believes that only a transient procedure
is technically correct, and that an engine dynamometer test is the
most practical. As to the specific transient cycle, there are an
infinite variety of possible transient tests, and it is categor-
ically impossible to evaluate all possibilities. EPA never pursued
the option of deriving a simplified transient procedure based upon
the Cape-21 data base; it was believed that any simplification
significant enough to result in substantial cost savings would also
be substantial enough to seriously compromise test representative-
ness.* ECTD is, however, fulfilling its legal obligation by
evaluating and addressing all comments received pertaining to the
NPRM, and where possible, simplifications will be made to the
procedure.
f. Technical Validity
Many comments were also addressed to the technical validity of
the test procedure itself. The concern was raised that when the
full range of dynamometer calibrations, validation statistics,
throttle actuator performance, and transient control strategies are
utilized, correlation between laboratories will be very hard to
achieve. The resulting certification program will be confused and
technically impossible to work with.
In general, the transient procedure as run in the past by EPA
*Exception: see "Caterpillar cycle," later in this analysis.
-------
and SwRI was cumbersome at times and slow to produce results, due
primarily to inexperience, equipment debugging and refinement, and
general efforts at test procedure development. On the other hand,
test results have been repeatable and not indicative of gross
emission variability within the range of validation statistics.
Furthermore, correlation on transient gasoline testing between EPA
and SwRI has been reasonably established (see Table A-13), although
some variability remains. General Motors has recently begun
tentative transient testing on a 1979 Chevrolet 350-CID V-8 origi-
nally tested at EPA. Using a completely different emission samp-
ling system (fuel based mass measurement integration versus EPA's
CVS dilute sampling), GM achieved comparable results:
BSHC BSCO
EPA* 3.14 _+ .70 118.1 +_ 6
GM* 3.05 + .13 107.5 + 2.4
* Average of three tests.
Furthermore, Cummins has achieved comparable emission results
with SwRI on an NTC-350:
BSHC BSCO
SwRI 0.74 4.51
Cummins 0.82 5.27
(Note: Cummins uses an integrated sampling technique for CO, HC,
and NOx. Dilute bag sampling of NOx appears to be technically
deficient insofar as unexplained losses of measurable NOx occur;
the values presented above for NOx are bag values.)
In short, every lab for which transient emission data is
available achieved some degree of correlation. Manufacturer's
fears of gross correlation difficulties appear unfounded.
At any rate, we recognize that the test procedure itself
requires fine tuning and streamlining. These problems, however,
are not inherent in a transient test, and will be addressed in this
document for inclusion in Final Rule. Actual certification tests
are four years away; as industry experience is gained with the
procedure, 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. As was
shown in the 1969 Baseline Technical Report, repeatability with the
transient test was equivalent to the steady-state; ECTD finds no
reasonable grounds for concluding that the test procedure is
technically unsound.
g. Alternative Cycles - "The Caterpillar Cycle"
Diesel manufacturers suggested that the proposed diesel cycle
be modified so that eddy current absorption dynamometers could be
retained, resulting in substantial cost and time savings. Both
EPA's and SwRI's transient diesel controllers use direct current
motoring dynamometers, as do Cummins' and Caterpillar's prototype
facilities. All transient diesel data to date have been obtained
on motoring systems; we are forced to consider this suggestion in
the absence of eddy current data.
Each dynamometer system would control the engine in different
ways. Motoring dynamometers to date have operated in "speed
control" modes (i.e., the dynamometer follows the speed cycle
independent of engine performance, while the engine is "driven"
over the torque cycle by automatic manipulation of the throttle or
rack position). On the other hand, eddy current dynamometers
operate in ''torque control" modes (i.e., the dynamometer is simply
a source of electric friction which loads down the engine over the
torque cycle while the engine is driven through the speed cycle by
operation of the rack.* These two control strategies could effec-
tively result in different second-by-second speed/torque pairs at
highly-transient portions of the cycle, possibly resulting in
differences in the measured emissions. These emission differences,
if any, are uncharacterized at this time.
Modifications to the proposed cycle would be aimed at elim-
inating highly transient and motoring portions where the two
dynamometer systems would operate differently. Caterpillar has
proposed exactly such a cycle in which 18 percent of the speed/
torque pairs have been slightly modified, based upon the inertia/
power characteristics of a single engine (see Figure A-18), i.e.,
the Caterpillar cycle is engine-specific.
* Motoring dynamometers are capable of operating either in
torque or speed control. EPA's original rationale for testing
in speed control was an expedient, based upon the judgment that
transient control of an engine dynamometer in speed control would
be simpler and more guaranteed of success than the available
options. Furthermore, there is a compelling safety reason for
speed control: the engine is always restrained by the dynamometer
and engine "runaways" are avoided.
-------
Each change to the cycle in Figure A-18 is Labeled to give the
purpose of the change. The key to the change symbols is as fol-
lows :
A - speed cycle changed to allow coastdown during motoring.
B - torque cycle changed to eliminate motoring during a
constant speed condition.
C - speed cycle changed to make acceleration compatible with
a turbocharged engine and the prescribed inertia.
D - torque cycle changed to increase the torque to achieve
the corresponding acceleration in the speed cycle.
The change to the torque cycle labeled E in Figure A-18 is only
apparent in the unnormalized cycle. This change occurs in the
cycle between 216 and 219 seconds.
ECTD's analysis of the Caterpillar proposal has taken two
forms; a statistical analysis identical to the test validation
regression analysis to determine how different from the proposed
cycle the Caterpillar cycle actually is, and actual comparative
emission tests on SwRI's motoring dynamometer -
The statistical analysis is presented in Table A-14. (For
purposes of this analysis, the following engine parameters were
assumed: idle speed = 700 rpm., rated speed = 2200 rpm, and maximum
torque at all speeds = 100 ft-lbs.) All of the statistical para-
meters met the cycle validation criteria.* Based upon this anal-
ysis, the Caterpillar cycle is similar enough to the proposed cycle
to qualify as a statistically equivalent cycle. The comparative
emission tests run at SwRI produced somewhat similar results, which
are shown in Table A-15. Of the two engines tested, one produced
identical emissions, while the other produced 14 percent less HC
and 8 percent less CO on the Caterpillar cycle. Cycle performance
statistics for each test were comparable, as were total integrated
BHP-hr's.
From this limited data, the ECTD technical staff reached the
following conclusions:
* This analysis regressed Caterpillar's command cycle against
EPA's command cycle. The analysis assumed all "motoring points"
represented closed throttle, and as such were thrown out of the
regression calculation. Inclusion of a -25 percent motoring
command unrepresentatively biased the regression, particularly
where Caterpillar eliminated motoring points. Even so, a regres-
sion including motoring commands of -25 percent produced only one
out-of-bounds statistic, the torque y-intercept (+26.0 ft. Ibs.).
-------
i) If in fact the Caterpillar cycle can be accurately run on
an eddy current dynamometer (as Caterpillar has claimed to have
done manually, and if in fact the modified cycle eliminated por-
tions of the proposed cycle which were unachievable on the eddy
current machine, then operation of the modified cycle on SwRI's
motoring dynamometer should not be significantly different from the
operation achievable on eddy current systems). Emissions measured
on either system should be comparable, but not necessarily iden-
tical.
ii) Statistically, the cycles are similar enough to be deemed
equivalent, at least insofar as emissions are concerned. Further-
more, the cycles are close enough that emissions generated over the
proposed cycle can be estimated by operation of the proposed cycle,
or even a slightly modified cycle, on eddy current equipment.
Based upon the above data, correlation between both dyna-
mometer systems should be feasible. The implications this holds
for certification is contingent on other factors, however. First
of all, EPA and SwRI use motoring dynamometers. To date, manufac-
turers have consistently copied EPA's equipment for certification
testing (e.g., Clayton dynamometers),* with an extra expense
providing insurance or interlaboratory correlation. Every diesel
manufacturer has already placed orders for at least one motoring
dynamometer and CVS system, so there is no indication of a change
in trend.
Secondly, although the limited data available indicates that
correlation between motoring and eddy current systems is probable,
this correlation is by no means guaranteed. Different control
systems with different response characteristics will be used, and
the degree of emission sensitivity to these differences is a
legitimate technical question. Furthermore, the differences
between diesel torque control and speed control have never been
established. (SwRI is unable to run in torque mode; EPA's facility
is still being debugged.) Correlation between eddy current and
motoring dynamometers will not be established before Final Rule-
making. (No eddy current transient emission testing facility
exists, nor can one be built up in less than six months.) Finally,
EPA's proposed test regression tolerances, by demanding excellent
speed statistics, essentially require a speed control test.
This lack of data will contribute to the manufacturers'
avoiding the exclusive development of eddy current dynamometers if
both the cycle and EPA's own certification facility remain the
same. It appears that the proposed cycle is sufficiently similar
to the modified cycle that it can be run on eddy current equipment
closely enough to achieve comparable emissions relative to a
motoring facility. In essence, there is no need to change the
In the light-duty vehicle testing area.
-------
proposed cycle if both motoring and eddy current dynamometers are
anticipated. It is unreasonable to expect, however, that manufac-
turers will accept this and rush headlong into the development of
exclusively eddy current facilities. Were ECTD to agree to
either develop an eddy current facility, run the motoring facility
in torque control mode (possibly voiding all SwRI data to date and
in the near future), or establish definitive correlation between
eddy current torque control and EPA/SwRI speed control, then
exclusive development of eddy current dynamometers would be likely,
resulting in the anticipated cost and leadtime savings. Torque
control (or use of eddy current dynamometers) for certification
would also require revision of the cycle validation statistics,
most likely tightening the torque specifications and loosening the
speed; in the absence of eddy current/torque control experience,
such a revision would be no better than an educated guess. Due to
the potential "runaway" engines, always a possibility when running
under automatic control, torque control is riskier from a safety
aspect. Finally, although instinctively one would expect no major
difficulties running in torque control, it has actually never been
done and at best remains unproven.
In summary, there is a lack of eddy current data. A decision
to modify the cycle and pursue eddy current/torque control facil-
ities could void all existing data, and a test procedure and
control system known to produce repeatable and reasonable results;
changes to the procedure at this time would be based upon conjec-
ture and absolutely no data or experience. It is probable that the
proposed cycle can be run on eddy current dynamometers with suf-
ficient emission accuracy to allow characterization and development
work, although manufacturer's certification facilities would almost
certainly model EPA's and consist of motoring dynamometers. There
is sufficient economic incentive for manufacturers to explore the
validity of the eddy current option for development while develop-
ing motoring certification facilities, although their behavior in
this respect is by no means certain. The viability of the eddy
current development option is not guaranteed; it is our judgment
that correlation between the two dynamometer systems on the pro-
posed cycle is probable. Based upon this judgment, no modification
to the proposed diesel cycle would be warranted.
One alternative would be to delay promulgation of the diesel
transient test until definitive eddy current data is available.
The timely acquisition of eddy current data would depend on sub-
stantial contributions by the industry, and the heavy-duty industry
has never been in a hurry to regulate itself. EPA would be ac-
cepting certain delays when a viable, although certainly more
expensive, transient procedure already exists.
Another alternative would be directed simply to allow addit-
ional time for investigation of the eddy current option. Optional-
use of either the transient or the 13—mode test would be allowed
-------
for one model year, Che first model year for which the proposed
regulations take effect. This effectively allows one additional
year for the industry to aggressively investigate the viability of
the eddy current option. This option should not be construed as an
admission that EPA considers the 13-mode to be technically compar-
able with the transient test. EPA would be willing, however, to
accept this compromise of technical validity for the sole and
explicitly stated purpose of possibly reducing the financial
burdens placed upon the diesel industry. The optional 13-mode
standard would be derived to reflect the approximate stringency of
the transient 1.3 g/BHP-hr standard (i.e., there is no relaxation
in relative standard stringency, merely a test procedure option).
A final alternative can be considered. Should EPA modify the
cycle to allow the use of eddy current dynamometers and run its own
certification facility in torque control, then diesel manufacturers
would not be compelled to invest in motoring facilties, resulting
in a substantial cost savings without substantially impacting the
air quality improvements attributable to a transient procedure (as
evidenced by the agreement between the two cycles at SwRI). This
has certain negative ramifications, however:
i) The data base at SwRI could be adversely impacted and
possibly voided. This is tempered by the fact that only a few
1979 diesels have been tested to date and the loss of data would
not be sizeable. More importantly, however, SwRI"s equipment is
not readily convertible to torque control. The particulate and NOx
baseline work would be delayed by as much as two months.
ii) Eddy current dynamometers have never been characterized
or proven for transient control. (Caterpillar has claimed other-
wise.) It could be reasonably presumed that eddy current dyna-
mometers are fully capable, if modified, for accurate and respon-
sive transient control, but this is by no means a certainty.
iii) EPA is faced with the decision of how to modify the
cycle. It has been shown above that the proposed cycle and Cater-
pillar's modified cycle are relatively similar. The proposed cycle
could be maintained, or Caterpillar's engine-specific cycle ac-
cepted at face value with the explanation that the arbitrary cycle
modification is justified on cost considerations above. ECTD could
also invest engineering time (with or without the cooperation of
the industry) in effecting its own modification to the proposed
cycle. This would be time consuming if actual testing were re-
quired (due to lack of facilities); if a theoretical analysis were
deemed sufficient or expedient the task could be accomplished in a
matter of weeks.
iv) Cycle validation statistics for eddy current or torque
control machinery would necessarily be a best guess. Due to the
considerable test procedural refinements anticipated as necessary
-------
prior to real certification testing, this is not a problem.
The most salient point of this discussion is the fact that no
emission data or performance capabilities of eddy current machines
have ever been characterized. To promulgate Final Rules based upon
uncertainty is not sound regulatory practice. ECTD is investigat-
ing^ the eddy current system as best it can prior to Final Rule-
making (e.g., running gasoline engines in torque control, running
small diesels in torque control on the gasoline dynamometer, and
attempting torque control vs. speed control comparisons on the
diesel dynamometer); all efforts will be made to resolve this
question, yet a definite resolution of the question of eddy current
viability in time for Final Rules must remain in doubt.
h. Test Procedural Details
Several commenters argued that use of a CVS for emission
sampling during the transient test was unnecessary, and in one
case, technically incorrect. Both Caterpillar and Cummins argued
that use of their home-built sampling systems be permitted.
Cummins presented data which indicated that bag sampling of
diesel exhaust results in NOx measurement errors. In a bag, there
appears to be a 15-20 percent loss in measurable NOx. (The same
type of problem lead to the continuous sampling of HC.) Resolution
of this discrepancy is anticipated to be continuous heated sampling
of NOx by integration. There is a possibility, however, that EPA
will be criticized for establishing an interim NOx standard based
upon a test procedure which understates NOx, while proposing a
procedure which does not for Final Rules. The interim NOx stan-
dard, however, is already so lax that no difficulty in its attain-
ment using either sampling system is anticipated.*
Since the precise measurement of NOx is not a critical ques-
tion in view of the lax NOx standard, the ECTD Technical Staff
believes that use of either bag or continuous dilute sampling
systems should be permitted for 1984. It is understood that
technical problems exist in bagged NOx measurements with diesels,
yet its use in 1984 should be permitted in 1984 due to its minimal
impact (i.e., lax standard), and to preclude criticism that adop-
tion of the dilute continuous measurement technique represents
increased standard stringency. It should be understood, however,
that dilute continuous NOx measurement will be adopted as part of
the forthcoming 1985 NOx NPRM, and bag sampling of NOx will be
abandoned.
The issue of alternate sampling systems is presently addressed
in the proposed regulations. Any sampling system adequately
* See Summary and Analysis of Comments "Feasibility of Compli-
ance" Chapter.
Vr
-------
demonstrated to yield equivalent results is permissible. (See
§86.1309-83(a)(l) of 44 FR, Part 86, February 13, 1979 NPRM.)
The final major test procedural issue pertains to the need for
a, 12-hour cold soak. This requirement ties up dynamometers and
drastically reduces test rates. It is judged that due to the
inherent cold start emission characteristics of gasoline engines
and catalysts, a cold start for gasoline engines is technically
justified. For diesels, however, the manufacturers are correct in
pointing out that a cold start requirement is less necessary.
Diesel emissions observed at SwRI are considerably more stable and
less sensitive to engine temperature (see Table A-16). ECTD is
concerned, however, that when the Congressionally-mandated NOx
reduction and the forthcoming particulate standards take effect in
conjunction with the proposed HC reduction, cold start emissions
may reach significant levels. It is not desirable to promulgate a
test procedure which may be adversely affected by these definite
future standards. The major criticism of the cold start involved
the 12-hour soak, and it is here that changes are warranted.
Recent data acquired by EPA in-house and from Cummins indicate
that the 12-hour soak requirement can be abandoned in favor of a
forced cool-down technique. The forced cool-down technique is
currently being refined, by no major difficulties are expected. It
is anticipated that a single engine temperature parameter (e.g.,
oil temperature) will replace the 12-hour soak requirement, and
define the point at which a cold start may begin. (Gasoline
engines will also have a catalyst temperature parameter.) There
will be general restrictions on the cool-down procedure (e.g.,
coolant type, coolant temperature, etc.), however, the EPA labora-
tory has achieved cold-start soak times of less than two hours.
This approach should eliminate all of the adverse effects of a
12-hour soak, and allow transient testing to be accomplished within
the time currently needed for the steady-state tests while retain-
ing the cold start cycle.
4. Recommendation
a. Retain the transient test for both diesels and gasoline
engines.
b. Retain the proposed cycle for diesel engines. Allow
optional certification on the 13-mode for the first applicable
model year (1984) at an approximately stringent standard. Test
cycle modifications may be made in the future, pending industry
investigation of eddy current dynamometer capabilities.
c. Substitute a forced cool-down procedure in place of the
12-hour cold soak for both gasoline and diesels.
d. Allow both continuous dilute and bagged NOx measurement
systems for diesel engines in 1984.
-------
e. Modify the proposed procedure in the case of many tech-
nical details (discussed in Part II of Test Procedure, Summary and
Analysis of Comments) to effectively streamline the procedure,
clarify intentions, and eliminate unnecessary requirements.
-------
Table A-l
Heavy-Duty Test Procedure Development
Decisions
1. Methodology for transient
procedure development estab~
lished and begun.
2. 9-tnode procedure to be
retained for future interim
standards.
Events
1. Ethyl study (1967).
2. 9-mode proposed as cer-
tification test for 1970 MY.
3. 23-mode test evaluated.
4. EPA/CRC joint contract
awarded to William Smith &
Associates (4/72).
1973
5. Negotiations completed
with Olson Laboratories for
computer development of test
cycle (1/22/73).
6. Final report: 25 con-
trolled gasoline trucks run on
San Antonio road route (2/73).
7. NYC Cape-21 data collection
begun (11/73).
1974
3. Engine dynamometer proce-
dure selected as top priority
over a chassis procedure.
4. Conflict of interest con-
siderations lead EPA to plan
and implement Los Angeles
CAPE-21 survey on its own
(see Exhibit 1).
8. Diesel engine certification
on 13-mode test begun (1974 MY).
9. Contract awarded to Olson
Laboratories for heavy-duty
cycle development.
10. Final report: 10 diesel
engines—SARR vs. 13-mode
chassis tests (8/74).
11. NYC CAPE-21 data collection
completed (10/74).
1975
12. LA CAPE-21 data collection
begins (1/75).
-------
Table A-l (cont'd)
Heavy-Duty Test Procedure Development
Year
Decisions
Events
13. Final report: 10 pre-
controlled gasoline trucks
on SARR vs. chassis 9-mode
(3/75).
14. LA CAPE-21 data collection
completed (5/75).
1976
15. Interim heavy-duty regula-
tion NPRM (5/24/76).
16. Final report: comparative
data on various transient and
modal chassis dynamometer tests
(5/76).
17. Formal industry briefing on
CAPE-21 and transient procedure
status (9/76).
1977
5. Based upon Events #1, 6,
10, 13, and 18, the 9-mode
test was rejected as a future
test procedure alternative;
the 13-mode was deemed
que s t ionable.
6. Based on CAAA, uncon-
trolled baseline programs
were initiated.
18. Final report: "Sensitivity
Study" (1/77) (arising from
data published 5/76).
19. EPA requests GM and Cummins
to attempt transient dynamometer
operation at their own facilities
(2,3/77).
20. RFP to SwRI for transient
baseline testing (5/77).
21. Cummins attempts transient
dynamometer control through
EMA's Transient Dynamometer
Evaluation Committee (7/77).
22. Clean Air Act Amendments
of 1977 (8/77).
23. Final Rulemaking, 1979 MY
Interim Standards (9/13/77).
-------
Year
Table A-l (cont'd)
Heavy-Duty Test Procedure Development
Decisions Events
24. Candidate heavy-duty
driving cycles selected
(11/77).
1978
7. M\TMA's alternate approach
to transient cycle develop-
ment rejected by the Agency
(see text).
8. In-house prototype gaso-
line engine tests reveal
significant discrepancies
between 9-mode and transient
procedures; the 9-mode is
further discredited.
25. EPA's transient gaso-
line dynamometer operational;
1969 baseline study begins
(2/78).
26. MVMA submits alternate
cycle development plan (2/13/78)
27. MVMA's alternate approach
rejected (5/17/78).
28., SwRI' s transient gasoline
dynamometer operational (6/78).
29. Recommended Practice for
the transient procedure pub-
lished (8/78).
30. SwRI's transient diesel
cell operational (10/78).
1979
9. Level of diesel emission
reductions achieved relative
to 13-mode plus lack of
correlation makes transient
procedure for diesels attrac-
tive, especially in light of
future NOx and particulate
standards.
31. SwRI begins transient
diesel baseline (2/79).
32. NPRM for transient pro-
cedure and 90 percent reduc-
tions published (2/13/79).
33. 1969 baseline program
completed (5/79).
34. Final standards proposed
(6/79).
35. Anticipated Final Rule-
making (12/79).
-------
Table A-2
Comparative Studies
Study Title/Date Study Summary/Conclusions
1. "Survey of Truck and Bus Operating Over the road data collection (RPM
Modes in Several Cities," June 1963. and manifold vacuum).
2. 'Exhaust Emission Analysis and Development of transient chassis dyna-
Mode Cycle Development for Gasoline mometer cycles from Study 1. Emissions
Powered Trucks," September, 1967. measured on transient chassis cycles,
(The Ethyl Study")„ road routes having same average speeds
and same operational mode distribution,
and the then in-use California steady-
state test (equivalent to the manifold
vacuum 9-mode).
Correlation between transient and
matched road emissions were excellent.
None of the transient cycles compared
well with the California 9-mode.
3. "Exhaust Emissions from Gasoline
Powered Vehicles Above 6,000 Lb. Gross
Vehicle Weight," by SwRI under EPA
Contract, April 1972.
Part I Four heavy-duty, 1969 MY gasoline.
engines run on chassis and engine
dynamometers under steady-state and
transient conditions.
For these uncontrolled vehicles, "steady
state conditions, including motoring
at closed throttle, can adequately
represent the emissions."
Part II Over-the-road emission data collected
over stop-and-go operation, and var-
ious cruising speeds, for four heavy-
duty 1969 MY trucks (gasoline).
Over-the-road HC was directly propor-
tional to the amount of transient
operation, NOx inversely, and CO
relatively stable.
Part III Nine 1970 and later MY trucks run on
experimental 23—mode test (engine and
chassis dynamometers).
-------
Table A-2 (Cont'd)
Comparative Studies
Study Title/Date Study Summary/Conclusions
4. "Mass Emissions from Trucks Oper-
ated Over a Road Course, Part I," by
SwRI under EPA Contract, February 1973.
5. "Mass Emissions from Diesel Trucks
Operated Over a Road Course," by SwRI
under EPA Contract, August 1974.
6. "Mass Emissions from Ten Pre-
controlled Gasoline Trucks, and
Comparisons Between Different
Trucks on a Road Course," by SwRI
under EPA Contract, April 1975.
Twenty-five gasoline trucks (1970-73
MY) tested over chassis 9-tnode and
San Antonio road route (SARR).
Regression analysis of chassis results
(Y) vs SARR (X):
SE
Slope Y-Intercept
HC
CO
NOx
.660 2.14 .499 -3.96
.813 44.3 1.176 -70.96
.837 2.67 1.411 -1.57
Regression analysis indicates that as
emissions decrease, the agreement be-
tween test methods disappears, nor is
correlation at these levels acceptable.
Ten diesel trucks (1970-73 MY) tested
over chassis 13-mode and SARR):
Regression analysis of chassis results
(Y) vs SARR (X):
R
SE
Slope Y-Intercept
HC
CO
NOx
.778 1.26 1.012 -.254
.644 4.57 .305 6.604
.786 5.04 2.90 -17.09
Concluded two test methods agreed some-
what at these levels of emissions; agree-
ment was questionable as emissions
decreased.
Ten gasoline trucks (1965-69 MY) tested
for emissions over chassis 9-mode and
SARR:
-------
Table A-2 (Cont'd)
Comparative Studies
Study Title/Date
7- "Heavy-Duty Fuel Economy Program -
Phase I, Specific Analysis of Certain
Existing Data," by EPA, January 1977,
("Sensitivity Study).
Study Summary/ Conclusions
Regression analysis at EPA:
2
SE
HC .795 3.25
CO .461 33.69
NOx -.124 1.99
Slope Y-Intercept
.794
.441
.399
6.43
134.33
2.67
Concludes that poor correlation exists
between the two test procedures at this
level of emissions; no correlation exists
at lower levels.
Eighteen gasoline and twelve diesel truck
analyzed for emissions over:
- Eight different steady-state tests,
- Three sinusoidal cycles, and
- Four average speed transient cycles.
Conclusion: no reweighting of any steady
state modes achieved consistent correlati
with any of the transient tests.
-------
Table A-3
9-Mode Versus Transient Emissions
Current Technology Engines
BSHC
Engine 9-Mode
1979 GM 292 0.42
1979 GM 454 0.39
1978 IHC 404 0.63
1979 GM 350 0.79
1979 IHC 446 0.42
1979 GM 366 0.50
1979 IHC 345 2.73
1979 GM 350 0.59
1979 Ford 400 2.15
1979 Ford 370 1.20
1979 Chrysler 360 1.18
1979 Chrylser 440 0.83
1979 GM 454 0.47
Transient
2.12
2.30
3.98
3.14
3.27
2.16
2.44
2.48
4.89
3.51
2.67
3.83
1.31
BSCO
9-Mode
26.86
17.33
18.07
14.62
24.28
17.40
17.68
20.40
53.16
37.12
21.38
10.47
20.11
Transient
54.98
51.55
54.56
118.07
90.40
43.43
34.44
64.76
112.43
47.75
98.14
112.38
78.49
Catalyst-Equipped (Prototype) Engines
BSHC
Engine 9-Mode
1979 GM 400 I/ 0.81
1979 GM 400 2/ 0.06
1979 Ford 351 I/ 0.97
1980 Chrysler 360 3/ 0.11
1979 GM 350 4/ 0.21
y Certified in light-duty
2/ Same as I/, but retrofit
3/ California package with
Transient
2.21
1.00
1.24
1.16
2.29
vehicle; equipped
with air pump.
production catalys
BSCO
9-Mode
45.91
2.46
70.86
0.31
0.18
with catalyst
t.
Transient
131.80
99.24
99.86
96.58
89.54
4/ Heavy-duty engine retrofit with catalyst.
-------
Table A-4
13-Mode Versus Transient Emissions
SwRI
Engine
BSHC
13-Mode
BSCO
Transient 13-Mode Transient
1978 Caterpillar 1.71
3208
3.37
3.34
3.79
1976 Cummins
NTC-350
0.24
0.68
2.20
4.99
1978 DDA 6V-92T 0.56
0.78
2.54
3.15
1979 Cummins
NTCC-350
0.32
0.86
3.30
2.62
1978 DDA 8V-71N 0.84
(#1 Fuel)
1.49
6.43
3.75
1978 DDA 8V-71N 0.69
Fuel)
1.30
8.00
4.35
-------
Table A-5
Transient Versus 13-Mode HC Emissions
Engine
1978 Caterpillar 3208
1976 Cummins NTC-350
1978 DDA 6V92T
1979 Cummins NTCC-350
1978 DDA 8V-71N
(#1 Fuel)
1978 DDA 8V-71N
(#2 Fuel)
A *
B
C
D
E
F
G
H
1979 Caterpillar 3208
Transient
3.37
0.68
0.78
0.86
1.30
1.49
0.99
0.76
0.72
0.86
1.83
2.22
1.25
0.55
1.96
13-Mode
1.71
0.24
0.56
0.32
0.69
0.84
0.36
0.38
0.27
0.43
0.68
1.14
0.27
0.14
1.20
Ratio, Transient
Lab 13-Mode
SwRI
SwRI
SwRI
SwRI
SwRI
SwRI
Cummins
Cummins
Cummins
Cummins
Cummins
Cummins
Cummins
Cummins
Caterpillar
Average:
1.97
2.83
1.39
2.69
1.88
1.77
2.75
2.00
2.67
2.00
2.69
1.95
4.63
3.93
1.63
2.40
Cummins' and Caterpillar data extracted from comments submitted.
-------
Table A-6
All Values in Grams/BHP-hr
(a)
(b)
(c)
(d)
(e)
(f)
Engine
Manufacturer Family
(g)
Sales-Weigh ted
Certification Transient Reduction Due Less 13-Mode Total Sales Transient Grams/BHP-hr
13-Mode Transient Target to Transient Reduction Percent Reduction Reduction
GM
GM
GM
GM
CM
GM
GM
GM
GM
CEC
CEC
CEC
CEC
CEC
CEC
CEC
CEC
CEC
CEC
IHC
IHC
I HC
Mack
Mack
Mack
Mack
Mack
Cat
4L-53T
6L-71N
8V-71N
6V-71NC
8V-71NC
6V-92TA
8V-71TA
8V-92TA
6L-71T
091
092A
092C
092E
172A
J72C
192B
193
221
222
DT-466B
9.0-Liter
DTI 466B
8
9
10
,11
SIB
3
0.83
0.84
0.82
1.27
0.80
0.58
0.51
0.50
0.55
0.38
0.32
0.26
0.26
1.20
0.53
0.30
0.38
0.79
0.69
0.64
1.38
0.56
0.31
0.76
0.12
0.58
0.87
1.20
1.99
2.02
1.97
3.05
1.49*
1.17*
1.22
1.20
1.32
0.91
0.77
0.62
0.86*
2.88
1.27
0.72
0.91
1.90
1.66
1.54
3.31
0.81*
0.74
1.82
0.29
1.39
2.09
1.97*
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
1.
1.
1.
2.
0.
0.
0.
0.
0.
0.
0
0
0
1.
0.
0
0.
1.
0.
0.
1.
0
0
0.
0
0.
1.
1.
10
13
08
16
60
28
88
33
31
43
99
38
02
01
77
65
42
93
50
20
08
0.
0.
0.
1.
0.
0
0
0
0
0
0
0
0
1.
0
0
0
0.
0
0
1.
0
0
0.
0
0
0.
0.
31
34
29
37
19
20
22
63
14
41
82
1.54
3.52
1.32
0.91
2.37
8.37
1.98
5.06
1.23
0.15
11.36
5.06
13.55
0.35
0.28
0.04
0.06
0.04
0.75
6.61
0.66
0.44
0.40
7.92
0.04
6.11
0.22
11.02
0.79
0.79
0.79
0.79
0.36
0.28
0.33
0.31
0.43
0.02
0
0
0
0.79
0.38
0
0.02
0.79
0.77
0.65
0.79
0
0
0.79
0
0.50
0.79
0.26
.012
.028
.010
.007
.009
.023
.007
.016
.005
.000
0
0
0
.003
.001
0
.000
.000
.006
.043
.005
.0
0
.063
0
.031
.002
.029
-------
Table A-6 (Cont §d)
All Values in Grams/BHP-hr
(a)
(b)
(c)
(d)
Engine Certification Transient Reduction Due
Manufacturer Family 13-Mode Transient Target to Transient
(e) (f) (g)
Sales-Weigh ted
Less 13-Mode Total Sales Transient Grams/BHP-hr
Reduction Percent Reduction Reduction
Cat
Cat
Cat
Cat
Cat
Cat
Cat
Cat
Cat
Cat
4
9
10
11
12
13
14
15
16
17
0.21
0.23
0.34
0.53
0.15
0.68
0.22
0.63
0.30
0.37
0.50
0.55
0.82
1.27
0.36
1.63
0.53
1.51
0.72
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0
0
0
0.38
0
0.74
0
0.62
0
0
0
0
0
0
0
0
0
0
0
0
0.20
0.02
2.25
1.00
0.31
1.54
0.30
0.21
1.98
0.36
0
0
0
0.38
0
0.74
0
0.62
0
0
0
0
0
.004
0
.011
0
.001
0
0
Total Sales- = 0.316
Weighted Grams/
BHP-hr per Truck
* Based upon actual transient emission data.
(a) 1979 certification records.
(b) 2.40 x a.
(c) Target g/BUP-hr
(d) b-c.
(e) (a - [13-mode target g/BHP-hr]) x (either 2.40 or actual transient/13-mode ratio)
(f) d-e.
(g) (sales/total sales) x F.
-------
Table A-7
Cycle Development and CAPE-21 Studies and
Technical Reports
Report Number Re port/Summary
1 "Truck Driving Pattern and Use Survey, Phase
II, Implementation Plan," by William Smith and
Associates, May 7, 1973.
This report outlines a sampling and instrumenta-
tion 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 William 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 Final Report," Report No.
EPA-460/ 3-75-005, by William Smith and Associ-
ates, 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 diesel engines used
in the CAPE-21 study.
5 "Engine Horsepower Modeling for Gasoline Engines,"
EPA Technical Report No. HDV 76-04, by L. Higdon,
December, 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 William Smith and Associates, June,
1977.
-------
Table A-7 (Cont'd)
Cycle Development and CAPE-21 Studies and
Technical Reports
Report Number Re port/Summary
This report summarizes the sampling plan, instru-
mentation, and data collected in the New York
phase of CAPE-21.
7 "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 sample plan, instrumen-
tation, and the data collected in the Los Angeles-
phase of CAPE-21.
8 "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.
9 "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 dynamo-
meter test cycles.
1° "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 statis-
tical comparative procedures used to meaningfully
combine truck categories to simplify the CAPE-21
data base.
H "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 extra-
polates the amount of cold operation present in
the CAPE-21 data base.
-------
Table A-7 (Cont'd)
Cycle Development and CAPE-21 Studies and
Technical Reports
Report Number Report/Summary
12 "Selection of Transient Cycles for Heavy-Duty
Engines," EPA Technical Report No. HDV 77-01, by
T. Wysor and C. France, November, 1977.
This report summarizes the statistical methodology
used in selecting the final test cycles from
the several cycles generated in Report No. 9.
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 Techinical
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 disoveries
gained in an actual transient test program.
-------
Table 8
.EDIT OUTPUT FOR LA - TRUCK 17,
DAY 1
——-FOITING SUMMARY •—
TOTAL NUMBER DF RECORDS REQUI R ING., EO I T ING ' = 1324-
NUMBER HF ZEROED RECORDS GN INPUT TAPE = 24-3
NUMBER OF ENGtNE-OFF RECORDS = 881
NUMBER OF RECORDS WITH PUT-OF-PANGE VALUES = 200
NUMBED OF RECORDS WITH:
RPM ABOVE RPM MAX IMUM(3950.J = 0
RPM BETWEEN 0 AMD 300 l, = 7
RPM BELOW 0 - 0
LF ABOVE L100+13 . = 71.
LF BETWEEN L100 AND L100+13 = ' 85
LF BETWEEN LF C/T AND LF C/0-13 = 0
LF BELOW LF C/0-13 ='• 0
SPEED ABOVE 70 = Q
SPEED NEGATIVE WHILE MOVING = 1
OFLTA-SPEEO.EXCEEDING AMAX = 36
SPEED TO BE'ZEROED = 216
[NOT AN OUT-TF-RANGE VALUE)
TOTAL NUMBER OF INTERPOLATED RECORDS = 166
RPM ABOVE RPM MAXIMUM INTERPOLATIONS = 0
RPM BETWEEN 0 ANO 300 INTERPOLATIONS = 1
RPM BELOW Q INTERPOLATION'S = 0
LF ABOVE LLOO*13 INTERPOLATIONS = 0
LF BEL.OW C/0-13 1NTEPPOLATICNS = 0
SPEED ABOVE 70 INTERPOLATIONS = 0
NEGATIVE SPEED INTERPOLATIONS = 1
SPEED INTERPOLATIONS DUE TO AMAX CRITERION = 27
ZERCEO RECORD INTERPOLATIONS = 137
TOTAL NUMBER OF RECORDS ELIMINATED DURING EDIT - 1072
RPM ABOVE RPM MAXIMUM = 0
RPM BETWEEN 0 AND 300 = 6
RPM BELOW 0 =0
LF A80VF L100 + 13 = 70
LF BELOW C/0-13 = 0
SPEED ABOVE 70 =0
NEGATIVE SPEED = 0
AMAX CRITERION = 9
CONSECUTIVE ZEROED RECORDS - 106
CONSECUTIVE ENGIME-CFF RECORDS = 881
TOTAL NUMBER np NON-ZEROEDT ENGINE-ON RECHRDS ELIMINATED
TOTAL NUMBER OF ZEROED RECORDS ON OUTPUT TAPE
THOSE DUE TO TIME DISCRE0 ANC IE S
TOTAL NUMBER OF RECORDS NGT REQUIRING EDITING
TOTAL NUMBER CF GOGO RECORDS ELIMINATED
TOTAL NUMBER OF RECORDS ON INPUT TAPE
TOTAL NUMBER OF RECORDS ON OUTPUT TAPE
* 69
1
= 13356
5
= -19680
= 18672
FND OF CONVERT ANH ECIT Fr:R LA TRUCK 17 »
-------
Cape 21
Proposed Cycle
9—mode
Cape 21
Proposed Cycle
13—mode
Table A-9
Summary Percentages*
GASOLINE
Acceleration
27.2
26.5
0
DIESEL
Acceleration
28.1
25.8
0
Deceleration
26.2
25.4
0
Deceleration
26.8
26.8
0
Cruise
20.6
20.3
76.8
Cruise
13.8
11.5
80.0
Idle
26.1
26.6
23.2
Idle
31.5
36.6
20.0
* From Table 10.
-------
w
Table A-IO
Hodal Percentages; CAPE-21 vs. Proposed Cycles
Freeway /Non-
Cape-2! Modal Freeway Cycle
Percentage Weighting
A D C I Factors
UaBo 1 ine :
NYNF 23 23 14 40 0.44
NYF 33 31 26 10 0.06
UNF 31 28 16 25 0.30
UK 29 29 40 2 0.20
Total CAPE-21
Composite:
Diesel:
NYNF 21 21 7 51 0.41
NYF 32 32 17 19 0.09
IANF 30 26 10 35 0.24
1-AF 36 35 27 2 0.26
Total CAPE-21
Composite:
Desired Weighted
Modal Percentage
A
10.1
2.0
9.3
5. 8
27.2
8.6
2.9
7.2
9.4
28.1
0
10.1
1.9
8.4
5.8
26.2
8.6
2.9
6.2
9.1
26.8
C
6.2
1.6
4.8
8.0
20.6
2.9
1.5
2.4
7.0
13.8
I
17.6
.6
7.5
.4
26.1
20.9
1.7
8.4
0.5
31.5
Modal Percentage Actual Actu
Generated Cycles Time-Weighting Moda
A D C I Factor A
23 22 12 41 0.47 10. B
33 31 24 12 0 0
31 29 14 26 0.26 6.1
28 28 41 2 0.27 7.6
Proposed
Cycle; 26.5
19 21 6 55 0.50 9.5
32 31 12 25 0 0
28 29 10 34 0.25 7
37 36 24 2.3 0.25 9.3
Proposed
Cycle: 25.8
al Time-Weighted
1 Percentages
D C
10.3 5.6
0 0
7.5 3.6
7.6 II. 1
25.4 20.3
10.5 3
0 0
7.3 2.5
9 6
26.8 If, .5
I
19.3
0
6.8
.5
26.6
27.5
0
8.5
0.6
36.6
-------
Table A-ll
CAPE-21 Study vs. U.S. Trucks and Buses Subject to HD Regulations
(U.S. Production) - (Exports) + (Imports from Canada)
6,000- 10,000- 14,000-
Year 10,000 14,000 16,000
1977 18.8% 6.4% .7%
1976 17.6% 10.9%
1975 13.8% 5.8% 1.9%
1974 7.4% 1.8% 1.7%
1973 7.2% 9.9% 1.6%
1972 6.3% 11.9% 2.1%
CAPE-21 0% 13.5% 2.1%
16,000- 19,500-
19,500 26,000
1.1% 34.3%
2.2% 37.5%
4.1% 45.0%
5.0% 44.3%
6.9% 37.4%
7.7% 36.7%
14.6% 22.9%
26,000- 33,000
33,000 and over
6.9% 31.8%
5.8% 25.9%
7.3% 22.1%
6.7% 33.1%
7.6% 29.2%
8.3% 26.9%
11.5% 35.4%
Yearly
Totals
100%
100%
100%
100%
100%
100%
100%
Source: MVMA data from Draft Regulatory Analysis, CAPE-21 Records.
-------
Table A-12
Diesel Usage in Heavy-Duty Vehicles
vs CAPE-21 Study
0-
Year 6,000
1977
1976
1575
1974
1973
1972
CAPE 21* -
6,000- 10,000- 14,000- 16,000- 19,500-
10,000 14,000 16,000 19,500 26,000
.5% - - 7.0%
1.3% - - - 5.3%
.2% 6.2%
2.2%
.2% 2.4%
.2% 2.8%
5.9% 2.9% 5.9% 5.9%
26,000-
33,000
11.0%
8.7%
13.4%
7.6%
10.2%
9.4%
14.7%
33,000
and over
81.4%
84.8%
80.2%
90.1%
87.2%
87.7%
64.7%
Totals
100%
100%
100%
100%
100%
100%
100%
Source: MVMA data, CAPE-21 records.
* based upon number of trucks, not truck days.
-------
Table A-13
EPA/SwRI Transient Emission Correlation
Test Data Test Data
Engine
1969 IHC
304
I
1969 IHC '
304
1978 IHC
404
1978 IHC
404
I
1978 IHC
404
1979 GM
454
1979 GM
454
1979 GM
454
1969 Ford
360
1969 Ford
360
(SwRI) (EPA)
6/78
10/78 1/79
11/78 11/78
1/7,
2/79 5/79
— 10/78
: 2/79 —
— 6/79
— 3/79
6/79 —
Comments
*SwRI-No Scats
SwR'I EPA
HC .HC
11.64
8.75 10.49
*SwRI-No Stats i 3.01 3.86
*SwRI-No Stats
Stats
—
_
^B^
—
SwRI EPA
CO CO
127.4
121.86 126.4
72.91 54.1
SwRI EPA
NOx N'OX
6.07
6.24 7.65
5.50 5.01
4.40 — i 76.5 —
6.3
3.85 3.72 57.5 73.3 5.0 4.42
i !
— 2.26
2.50 —
— 2.36
— 5.92
6.14 —
— 48.7
51.0 —
— 55.4
— 75.32
97.3 —
— 6.92
6.5 —
— 6.55
— 6.88
5.26 —
* Ac this time SwRI was incapable of data acquisition necessary to perform cycle per-
formance regression analyses. The test data is therefore qualified.
-------
Table A-14
Regression Analysis: Caterpillar Cycle (Y)
Versus Proposed Cycle (X)
Speed
Cycle
Segment
1
2
3
4
Total
Cycle
Segment
1
2
3
4
Total
Cycle
Segment
1
2
3
4
Total
Standard
Error
44.3 (rpm)
99.4
17.7
44.4
60.9 (rpm)
Standard
Error
1.3 (%)
9.4
5.3
1.3
5.6 (Z)
Standard
Error
1.6 (%)
6.6
2.9
1.6
3.8 (Z)
Slope
0.979
0.957
1.00
0.979
0.987
Torque
Slope
1.00
0.955
0.971
1.00
0.986
Horsepower
Slope
0.983
0.960
0.981
0.983
0.984
2
IT
0.991
0.972
0.997
0.991
0.990
2
IT
0.998
0.909
0.976
0.998
0.971
R2
0.991
0.939
0.991
0.991
0.981
Y-intercept
9.08 (rpm)
37.0
-3.4
9.04
5.85 (rpm)
Y-intercept
-0.40 (ft-lbs)
28.0
23.9
-0.41
11.2 (ft-lbs)
Y-intercept
0.09 (Bhp)
8.30
5.70
0.09
2.98 (Bhp)
-------
Table A-15
Emission Data: Caterpillar Cycle
Versus Proposed Cycle
Cummins NTCC
350 (1979 MY)
Proposed Cycle
D4-10 D4-11
BSHC
BSCO
BSNOx
BSpart
BSHC
BSCO
BSNOx
BSpart
CS HS T CS
0.72 0.73 0.73 0.88
2.32 2.19 2.21 2.37
5.06 4.92 4.94 5.12
X 0.42 X 0.44
DAA
Proposed Cycle
HS1 HS2 HS3
1.39 1.54 1.39
4.20 4.11 4.38
4.96 5.38 5.29
1.05 1.09 0.78
HS
0.83 0
2.22 2
4.95 4
0.41 0
8V-71N
HS4
1.42
4.25
5.21
0.85
T Mean
.84 0.79
.24 2.22
.97 4.96
.41 0.41
(1978 MY)
Mean
1.44
4.24
5.21
0.94
Caterpillar Cycle
D4-13 D4-14
CS
0.86
2.24
4.87
0.42
HS1
1.29
3.77
5.06
0.78
HS T CS HS T Mear
0.69 0.71 X 0.70 X 0.7"
2.19 2.20 2.23 2.18 2.19 2.2(
5.11 5.08 4.98 4.78 4.81 4.9:
0.41 0.41 0.43 X X 0.4.
Caterpillar Cycle
HS2 HS3 HS4 Mean
1.20 1.17 1.32 1.25
3.73 4.12 3.98 3.90
5.55 5.24 5.30 5.29
0.72 1.00 0.77 0.82
X: voided results
-------
Table A-16
Cold/Hot Diesel HC Emissions
(g/Bhp-hr)
Test
Cold
Hot
Test
Cold
Hot
Test
Cold
Hot
Test
Cold
Hot
Test
Cold
Hot
Caterpillar 3208
Dl-5 Dl-6 Dl-7 Dl-8 Dl-11
3.95 3.80 3.66 3.43 4.07
3.18 3.26 3.23 2.96 3.53
Cummins NTC-350
D2-7 D2-9 D2-10 D2-12 D2-13
1.12 1.14 1.11 1.02 1.00
0.64 0.70 0.69 0.67 0.61
DDA-6V92T
D3-4 D3-5 D3-7
0.76 0.81 0.77
0.79 0.75 0.70
Cummins NTCC-350
D4-4 D4-7 D4-13
0.74 0.93 0.86
0.79 0.94 0.69
DDA-8V71N
D5-1* D5-2*
1.01 1.26
1.26 1.31
Dl-13 Dl-14
3.98 3.89
3.53 3.35
D2-15 D2-16
1.01 1.10
0.60 0.60
(w/ #2 fuel)
(Caterpillar Cycle)
Tentative,
-------
Figure A-l
Ht-AVI U)l» dAbULIM? OH, ;Nj|^C ICil
TRANSIENT TEST RESULTS / IDLE TEST REPORr
i>oit»
lints
i-Af, i nntf cs
°Af. 2 '.A'JF
P\1 1 \\-
PAG 5 NlT'JF US
I'Ar, 7 LAP
I VfiF
f. CS
t US
TflTAI TEST
CMS/
CMS/
KW-HH
MC
CMS GMS/HI
/ en
CMS/ CMS/
flHPHR KW-HR
CMS CHS/".I
GMS/ GHS/
3HPHR KW-HR CMS CMS/HI
/ F.E.
L8S/ GMS
GAL. LBS. BHPHR ' KW-HR
12.9* 24. 56
7.47 I . R 4
1.20 0.09
1.00 0.75
2.40 1.79
1. ?? n. o|
1 .0') 1 .'!?
1.64 ).'i8
T.7? ?.77
..in (i.p,
1 ."P 1.14
21.24 3S.06
4.17 3.16
7.06 1,77
0.70 1.19
l.n> 3.31
2.08 1.61
6.41 1.61
0. 54 1.91
•5. IS
1. 60
2. in
239.56
3fl. 58
118.55
49.17
44. 13
?6.0fl
116. Vi
40. 26
107.08
07.14
«9.91
178.64 154.52
28.77 63.51
08.40 700.76
36.67 34.37
32. 9J 33. J2
20.12 45.92
CW..OO 6116.49
30.02 33.50
79.85
64.91
67.05
284.13
49.32
175.30
50.32
6J.95
35.66
171. 72
56.98
140.51
124.52
127.95
4.34
6.38
6.15
4.51
7.62
5.42
5.39
_
5.69
5.75
5.74
3.23
5.13
4.06
4.50
3.36
5.60
4.04
4.02
4.24
4.29
4.28
2.80
11.33
32.20
4.30
3.37
12.97
51 ,'Jl
4.50
5.14
0.80
6.06
7.29
6.23
10.07
7.90
7,63
7.89
0.22
8.17
0.14
0.21
0.63
O.ll
0. 12
0.21
0.62
0.12
1.08
1.07
2.15
0.84
1.27
3.87
0.69
0. 76
1.32
3.83
0.72
6.67
6.63
13.30
1.300 439.7
0.772 261. 1
0.655 221.5
0.904 332.3
1.020 345.0
0.773 261.5
0.651 220.2
0.059 290.6
0.749 253.3
0.723 244.6
0.727 245.9
Transient Gasoline Engine (w/Catalyst) Test Results
-------
Figure A-2
HD-H0016S
03-lb-79 TIME! 1<»?45800 Ho-800165
CO'iPOSIIE C*.
COMPOSITE HS
TOT/-.L Tfisr
PCTF.S
(V4S/ fiMS/
HHPHK KK-HR
RAG 1 NYNr C<; 36.'
RAft 2 LANF
HAG 3 LAF
BAG 4 NY MI-
BAG S NYN*-" US
BAT, 6 LANt
BAG 7 L<>F
BAG fl NYNF
IJASOI.INK HAG ENGINE TEST
IKST KfSilLTS KK.HOHT
OAIE8 03-15-79 Tint! 14845510 Hn-800165
(.0
HHPHt? r *l-rtH
GMS/MI
/ / __ NOA —
G.MS/ (ins/
BriPHH KW-HR >'>MS
L85/ UMS
LBS. BHPHR KW-HM
0.3S 0.26 0.5B 0.'*S
n.n^ o.o? n.|2 o.p i
O.I) 1 0.01 0,fl3 0..i'4
1. 3« i.oo n.fid i.-.->4
o.?l o.i ft n.:»s o.^7
0,0 « 0.0? A. 17 0.. *
O.OH 0.06 0.06 0.11
l.'ll ?.?.S 3. tft
o.H n.n o.i
O.H-< (l.'«l (l.'/S
2B5.30 .12. /b
4.SS 1.>9
6.?7 /..r,/
15. S; ll.i-l
37.')*i 2«..>2
4.V'i ,U:>7
ft.S7 4. >0
15. If ll.^rt
29. HI ?|..i.1
9. SO /.ob
12.2S '*.J4
Ib7.ljb 327.89
7.4h 5.«n
3?.^ I H«li
12. "4 20. 4»
25.00 43.71
rt.-il 6.?^
.H.bH H.6'*
12.00 20,. 1ft
37.16
1 ?. . .14
15.89
2. Id 1.63 i.44 2.51 O.J3 0.82 1.252 423.5
7,o6 5.71 12.56 S<.76 0.19 1,17 0.712 240.8
/.o6 5.71 3^.72 9.9^ 0.48 2.97 0.5V3 193.8
5.<-6 4.07 4.22 7.17 0.10 0.60 0.772 261.1
.1,72 2.78 ,.45 4.2« 0.09 0.57 0.869 293.9
7.65 5.71 1<^.79 9.93 0.19 1.15 0.688 232.7
4.«a 5.98 4tr.ll 80. 5J 0.48 2.96 0.564 190.8
S.J9 4.02 4.27 7.2S 0.10 0.60 0.762 257.7
7. U2 5.23 b.99 0.90 5.56 0.673 227.6
''••"> 5.49 9.56 O.f6 5.29 0.632 213.8
7.31 5.45 9.48 1.76 10.85 0.638 215. «
Transient Gasoline Engine Test (w/Catalyst, Meeting Standards)
-------
1.50 -,
Proposed Transient Standard (1.30 g/BHP-hr)
1.25
1.0
Transient
BSHC
0.75
Line of Perfect
Correlation
0.50
0.25
SwRI Data
Cummins Data
—I
0.25
,—
0.75
0.50
13-Mode - BSHC
Figure ^J. Transient Versus 13-Mode BSHC Emissions
1.0
-------
60
50 --
30 --
20
10 .-
Figure A-4
N,Y. Gas Non^Freeway
8410263
Key:
Cycle
Input
-20-14 -8 -2 4 10 16 22 28 34 40 46 52 58 64 70 76 82
94 100 106 112 118 124 130 136 142 1*8 15/i
-------
60 --
50 --
Figure A-5
L.A. Gas Non-Freeway
203887989
Key:
Cycle
Input
30 '
20 •
10 '
1 L 1 1 l*-i -
•"1 •( 1 II ^1
-20 -14 -8
I
-2
«* «• * "— --
— _ f"^- ~~~-— .
—,.""" -«* *•
w^»-»
«=i — ^trrrii-.....^
1 1 1 1 1 1 • j I1 '-I 1 1 L\ 1 1 1 1 1 1 1 1 "r flT"! 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
4 10 16 2228 34 40 46 52 58 64 70 76 82 88 94100106112118124130136142148154
RPM
-------
60 4-
50 4-
40 4-
M %
30 4-
20 4-
10
Figure A-6
L.A. Gas Freeway
296644805
Key; Cycle
Input
4-
i
-20 -14 -8 -2 4 10 16 22 28 34 40 46 52 58 64 70 76 82 88 94 100106 112118124130136142148154
/£ RPM
-------
60
50 "
40
si % 30 '-
20 --
10 --
+•
4-
Figure A-7
N.Y. Gas Non-Freeway
8410263
Key:
Cycle
Input
"h-TrT
I
Motoring 0*
(Negatives)
10** 20 30 $ 40 50
% Power
60
70
80
90
100
-------
60
50
Figure A-8.
L.A. Gas Non-Freeway
203887989
Key; Cycle
Input
40 4-
30 +
20 +
104-
-4-
Motoring 0*
(Negatives)
10**
20
30
40 50
% Power
60
70
0 < % < 5
80
90
100
-------
60
50 -•
40 -
30 --
20 -•
10 --
Figure A-9
L.A. Gas Freeway
296644805
Key: Cycle
Input
Motoring 0*
(Negatives)
10**
20
30
40 50
% Power
60 70
* 0 < % < 5
80
90
100
-------
60
50
30 --
20 --
10 --
Figure A-10
N.Y. Diesel Non-Freeway
2114147447
Key; Cycle
Input
^ --- ~""p»iy»t-.^J _ , _ 4
m i rn i rnri \ \ \ FH \ 1
•ri-rm-*-
i i • i i i i i
-20-14-8 -
2 4 10 16 22 28 34 40 46 52 5864 70 76 82 88 94 100106112118124130136142148154
A y RPM
-------
60 •-
50 -
40 --
30
20 ..
10 --
Figure A-l1
L.A. Diesel Non-Freeway
2110248101
Key:
Cycle
Input
T"I i < i I
rt"t I I I i I I I I -I I I
-20-14-3 -2 4 10 16 22 28 34 40 46 52 58 64 70 76 82 88 94 100106112118124130136142148154
^ RPM
-------
60 -
50 . .
40 --
30 .
20
10 .-
Figure A-l? ,
L.A. Diesel Freeway
1599345415
Key;
Cycle
Input
4-
rn i.i'M i i M i M IT-
-20-14-3 -2 4 10 16 22 28 34 40 46 52 58 64 70 76 82 88 94 100106 112118124130136142148 154
-------
60
50 +
40 +
30 4-
20 4-
10
Figure A- 13
N.Y. Diesel Non-Freeway
2114147447
Key. Cycle
----- Input
3^
Motoring 0*
(Negatives)
10**
20
30
50
60
70
80
90
^100
% Power
* 0 <_ % < 5
5 < % < 15
-------
60 -
50
Figure ~A-14
L.A. Diesel Non-Freeway
2110248101
40 --
% 30 --
20
10 --
, . . 1 ...
>
1 I
Motoring 0*
(Negatives)
Key: Cycle
Input
• ~~ — — — — „ 1 ~T — ' i TT~ 1
1 1 1""~1 " \
I I I I I I I I I I I I I I I '
1 • - 1 | « > 1 I | la i t I •• 1 | | 1
10** 20 30 40 50 60 70 80 90 100
% Power
* 0 5 % < 5
** 5 ^ % < 15
-------
60
50 +
40 4-
% 30-1-
10
Figure A-16
L.A. Diesel Freeway
1599345415
Key:
Cycle
Input
20 Hz.
• . . .. .
Motoring 0*
(Negatives)
I
10**
I
20
1
30
~~ *• « .-_,
1
•
'
40 50
% Power
I
60
70
•-•
80
" • 1
90
*
100
* 0 <_ % < 5
** 5 < 7, < 15
-------
Linearly Inter-
polated Function 100-
100% Load
RPM
75-
50-
25'
• • » • •
» * • *
•
» •
I
I*
r
r
'i
i
» *
• « •
• • •
• I
'I
• ••
» » * • •
* » » » • » «
» » •
»•*»
••*
i
I •
90
» *
T
Maximum
ri-'i' " •'•>'•
% Torque Parameter
Minimum
Linearly Interpolated
Function - Zero Load
Figure A-16 Graphed EVSL Matrix
-------
Figure A-17
Test Procedure Alternative Decision Flow Chart
Heavy-Duty Emission Test Procedure
New steady-states?
Are old ones viable?
•San Antonio road route
studies (1972-1975) said
No!
-Ethyl Study (1967)
said Nol
-In-house and contracted
testing of current and
prototype engines (1978-
1979) said No!
•No evidence or factual data
has been found by EPA or
presented to EPA by the
industry which demonstrates
the viability of current
procedures for heavy-duty
vehicles of the future.
Conelusion: The present 9-
and 13-modes are not accept-
able alternatives.
-Experimental 23-mode test
was evaluated (4/72); was
determined to be no better
than 9-mode or 13-mode.
-Sensitivity Study: no re-
weighting of steady-state
modes consistently cor-
related with any of several
transient tests i.e., steady
state tests in general are
dubious predictors of tran-
sient emissions.
•Engineering judgements
based upon light-duty
experience.
•Concern: Can any steady-
state remain valid through
progressing technology?
oo
Conclusion: The success
of a new steady-state test
is doubtful. If transient
test is representative and
cost-effective, then use it.
•Chassis cycles were
developed from CAPE-21
speed data. A certi-
fication procedure would
be similar to LD-LA4.
•Certification on a chassis
cycle, due to a large num-
ber of truck applications
and component variations,
would be impractical (See
text-MVMA letter) and ve-
hemently opposed by the
industry.
Conclusion: Due to logis-
tical problems of chassis
certification, this is not
a viable option for the
present rulemaking.
•An engine cycle is the
most practical of all
alternatives. For this
reason the CAPE-21 gen-
erated transient engine cy-
cles were proposed in the
NPKM.
•Further modifications
to the proposed cycles?
(See text)
•Further modifications
to the proposed pro-
cedures? (See text)
Conclusion: An engine
dynamometer test is the
most practical; a tran-
sient test is the most
representative. Modi-
fications to the pro—
posed cycles and pro-
cedures will be made
where possible. (See
recommendations and
text.)
-------
Figure A-18
CHANGES TO EPA PROPOSED CYCLE TO FORM THE CATERPILLAR MODIFIED CYCLE.
SEE TEXT FOR MEANING OF CHANGE SYMBOLS.
-PROPOSED CYCLE
MS
•MODIFIED CYCLE
1199
-------
Figure A-18 (cont.)
CHANGES TO EPA PROPOSED CYCLE TO FORM THE CATERPILLAR MODIFIED CYCLE.
SEE TEXT FOR MEANING OF CHANGE SYMBOI&
1600
TORQUE
(N-m)
0
-400
PROPOSED CYCLE
VS
MODIFIED CYCLE
2400
SPEED
IRPM)
400
-600-
• SEGMENT
•899H
-------
Figure A-18 '(cont.)
CHANGES TO EPA PREPOSED CYCLE TO FORM THE CATERPILLAR MODIFIED CYCLE.
SEE TEXT FOR MEANING OF CHANGE SYMBOLS.**
1600
TORQUE
(N-ml
0
-400
2400
SPEED
(RPM)
400 _
300
PROPOSED CYCLE
VS
MODIFIED CYCLE
SEGMENT 1L-
59'
-------
Exhibit A-l
Excerpt from Science Magazine
August 24, 1973
-------
?CIEI:C£ ^ Au.gT-ist, 2^, 1973
Auto Pollution: Research Group
One of the first issues that Russell
Train, the nominee for administrator of
the Environmental Protection Ageucy
(EPA), will have to decide if and when
he takes office, will be what to do about
that agency's role in automotive pol-
lution research. Train's predecessor,
William Ruckelshaus, promised Con-
gress that he would reassess some of
the agency's close research ties with
the auto and oil. industries it regulates.
At issue is EPA's participation in a
key research organization, called Co-
ordinating Research Council-Air Pol-
lution Research Advisory Committee
(CRC-APRAC), which has sponsored
much of the research that has been im-
portant to federal regulation in the
battle to clean up ths nation's air.
CRC-APRAC is supported by the auto
industry, ihe oil industry, and the
EPA.
However, a few months ago Ruckels-
haus promised Congress:
If it [EPA participation in CRC-APRACj
Sives the appearance to you and possibly
to others that this has compromissd our
position, we will have to .cease this asso-
ciation. . . .
An internal review is under way at
EPA, and a report is due soon.
Because three-fourths of the $23
million that the group has spent to date
has come from the American Petro-
leum Institute (API) and the Motor
Vehicles Manufacturers' Association-
(MVMA), with only the remaining
fourth from the government, CRC-
APRAC has been accused by public
interest lobbyists and members of Con-
gress as having a pro-industry bias.
Moreover, because it puts the regulated
Industries in bed with the agency that
regulates them, the arrangement, says
the: pollution guru of Congress, Senator
Edmund Muskie (D-Me.). poses a
serious conflict of interest for EPA.
Tne APRAC group is one wing of
CRC, a major trade organization
which, for over half a century, has
been a. vehicle for getting -the oil and
engine suppliers together on some com-
mon problems. Tne APRAC group is
''Unusual to CRC and to other trade
research organizations in general be-
cause it receives large amounts of fed-
eral funding and routinely has fed-
eral officials participating in its deci-
sions. The arrangement grew up in
the late 1960's, when auto pollution
was first becoming recognized as a,
national issue and when research funds
for EPA's predecessor in ths field, the
National Air Pollution Control Ad-
ministration (NAPCA), were scarce.
Now, however, critics argue that EPA
should be 'pursuing a "Caesar's wife"
policy and keep itself above sus-
picion in its regulation of the auto
industry, and that the CRC-APRAC
tie is compromising.
The alleged conflict of interest which
Muskie and others see in EPA's tie
with CRC-APRAC, however, may be
only the tip of the iceberg. Almost
without exception, when a research
scientist is funded by CRC-APRAC, hs
is already taking money from both
the industry being regulated and the
regulator. But this potential con-
flict is further tangled by the fact
that many of CRC-APRAC's contrac-
tors, separately, depend on the auto
or oil industry for a major share of
their business. Soms take money not
only from the industry, but from EPA
too. What emerges is not a clear-cut
line between scientists working for EPA
and those working for industry, but,
instead, a murkier set of in-j;roup rela-
tionships. Small wonder then, that,
after '5 years of national effort, many
upjjarently ^impiL- technical (['.'csrions
relating to auto emissions control re- '
main hotly disputed.
Of CRC-APRAC's foes, the bat-
known is Muskie. In hearings last
April on the EPA postponement of
the 1975 cm:;..sions control dsuJiins
th;it was imposed by ths 1970 Clean
Air Act, the Mains Democrat chal-
lenged the objectivity of studies don*
by a researcher who has dons much
of CRC-APRAC's work, on the health
effects of carbon monoxide (CO),
Richard D. Stewart of the Medi-
cal College of Wisconsin in Milwaukee.
Stewart had found evidence that the
average level of carbcxyhemoglobin—
an indicator of CO poisoning—in the
blood of nonsmokers across (he coun-
try was beiow 2 percent, which is
the safe limit now used in federal regu-
lation. (Stewart also found carboxy-
hemogiobin in the blood of smokers
to be higher than that in nonsmokers.)
Muskie, illustrating why CRC-APRAC
researchers are accused of b:as, pointed
oat that Stewart's work had been over-
seen, by a typical CRC-APRAC panel,
headed by a man from the General
Motors Corp. (CM), with people from
"PhJIlios Petroleum Co., Marathon Oil
Co.,""another GM man, and one EPA
representative, who, Muskie added sar-
castically, was "slightly outnumbered."
Muskie also waved a full-page Chrys-
ler Corp. ad publicizing Stewart's re-
sults,, and he said, "Chrysler is the one
automobile manufacturer which has at-
tacked ths health basis of the 1975
stardnrds. It is that information which
is going to be peddled around the
country ... for the purpose of attack-
ing the basis of ths 1970 Act."
(In fact, Stewart's findings, as writ-
ten up by Associated Press and carried
in newspapers across ths country, were
interpreted as evidence of the heavy
influence of smoking in CO poisoning,
a finding which other researchers on
health effects—such as John Gold-
smith of iha California State Health-
Department—believe may be valid
but nonetheless distracting from the
mnin point: that susceptible people, in-
voluntarily exposed to CO from auto
exhaust, suffer adverse -health eiTects.)
Muskie listed o;her panels of CRC-
SCIENCE. YOU. 1S«
-------
'^i? }>-~y- '•'•''• ic ic b;^ auto and oil com-
p:uiics,arc ^norously represented, while
EPA sn.-;p!u>e-:i are outnumbered-: —
sorr.otirnes by 12 to 1. lie argued that
the auto companies take advantage of
EPA's support oi" CRC-APRAC to give
its work credibility, and then publicize
their own interpretations of it.
Ths siuiiuw of EPA's involvement can
be us-.-ci r.nd v.ill be used to give die iiura
of credibility.. oilkial credibility, to state-
ments mads by Chrysler like this, chal-
lenging the- health basis of ths act. ... I
say the answer is to provide adaquate
funds and not Isan upon industry to do
the job.
Whether or not EPA is really an
equal partner in CRC-APRAC hinges
on the extant to which it exerts an in-
fluence on the group's deliberations.
The CRC-APRAC's full-time officers,
general manager Milton K. McLeod
and project manager Alan Zcngel
stated in interviews that most of the
group's decisions are made by the
APRAC committee, which has 6 EPA
•representatives out of a total of 21
members." The APRAC committee de-
cides, without formal outside review,
what work shall be undertaken, and who
shall be appointed to the many- sub-
pands, such as the one Muskie listed
during the hearings, which supervise
the research work itself. As to the gov-
ernment officials being outnumbered,
McLeod and Zengel admitted • (and
EPA .oiTicials confirmed) that the
panels often make decisions by voting,
and that sometimes EPA people vote
one way with the industry people vot-
ing the other.
However, not only do the oil and
auto companies appear to dominate
much of CRC-APRAC's decision-mak-
ing, but the group».which CRC-APRAC
selects to perform its research, in turn,
depend for their livelihoods on business
with these same industries. The most
obvious example is that part of CRC-
APRAC's work is funded by fuel com-
panies and performed by fuel com-
panies, and deals with matters in which
they have a vital interest. CRC-
•The APRAC gre'.ip coniii'.i of: C M. Huincn.
Girysto Corp.. chairman: 1. W. Blattenborser.
Cuia Service Oil Co.; D: L. Block of Ford Motor
Cs.; C, E. Burke of American Motor Corp.;
R. A. Coit 01 S'.wil Oil Co.: R. E. Edchar.it of
Esso Rocarwh nttil Engineering Co.; E. F. Fort
of iptc.-nanVnal It.ir^c^icr Co.: D. G. l.cvine
of E.-^i RswrwxJi inJ Enrrinc^tina Co.: C. D.
M.iriuuJc of i:c:J Mo'.cr Co.; C. II. Mos.-r of
Tv^:i.'J F:!-.: E- H. S.-ott of S:ar. given -.1 loUl oi approxi-
mately SI million to '.i'.a-e oil com-
panies: Esso Research :md Engineerinij
Co., a subsidiary of Standard Oil of
New Jersey, which is studying ths ef-
fectiveness of two well-known emission
control devices, thermal reactors, and.
dual catalysts; Ethyl Corp., where
changes in fuel volatility, a suggested
means for lowering harmful emissions,
are under study; and Phillips Petrol-
eum Co. Ons of the major decisions
EPA must make is whether shore-term
measures, such as altered fuels, and
add-on gadgets, such as the dual ca-
talyst, can be substituted by Detroit
for a major switch to a" new type of
auto engine with new fueling require-
ments, ^i^
In addition to funding oil companies
directly, CRC-APRAC supports, other
contractors who, in turn, depend on oil
and auto companies for a major share
of their business—a. situation that
again raises the question of their
stake in the outcome of the research.
The largest CRC-APRAC contractor
is TRW Systems, which has gotten S3..3
million from that group. Despite its
reputation among scientists as an aero-
space firm, the parent company, TR\V
Inc., in fact does approximately 40
percent of its worldwide business (its
annual sales are 51.6 billion) making
and marketing vehicle parts. Thus, it
is very much an interested party in
federal regulations affecting the auto
industry. TRW Systems, the research
arm of this giant, has studied all as-
pects of vehicle maintenance and in-
spection for CRC-APRAC. The issue
of vehicle maintenance 'and inspection
has been a bone of contention between
the industry and the government ever
since the 1970 act passed Congress. Ac-
cording to Charles Heinen of Chrysler
Corp., and CRC-APRAC's chairman,
the auto manufacturers have been ar-
guing that strict maintenance and in-
spection policies to keep existing auto
antipollution equipment clean, would
serve to meet emission standards. But
EPA standards setters have countered
that such a policy, emphasizing mainte-
nance, would de-emphasize the need
to improve the quality of the original
equipment- installed in the car. They
have said that this would therefore
shift the burden of the clean car from
the manufacturer to the owner or his
garage mechanic.
The second largest recipient of CK.C-
APRAC money has been Scott Re-
search Laboratories, Inc., one of the
country's loading makers of air poltu-
lion nu::i.v.ir]ng equipment, hi the last
3 years, Scott has done about hil: hi
business, or about S3.S million, with
auto and fuel companies tuid ilicir .
trade associations. Additionally, CRC-
APRAC over the same period has
spenf an added .$1.1 million at Scott.
One of Scott's major projects for
CRC-APRAC has been studies of ve-
hicle use patterns, or what EPA regu-
lators term "driving cycles." A driving"
cycle is a package of information on
when.- and where various types o£ ve-
hicles-:— trucks, cabs, cars, and .others
— are used, at what speeds they are
run, at what temperatures, and so
forth. Data on actual vehicle use,
which in turn go into making up the
EPA driving cycle, has been a,
central issue to many ongoing dis-
putes' over emissions control, since one
of EPA's standards setting jobs is to
determine the driving cycles, which in
turn .determine the performance stan-
dards that manufacturers must make
their engines meet. According to Mal-
colm Smith, one of Scotfs principal
investigators on the vehicle use studies,
at the termination of the CRC-APRAC
sponsored work, the auto industry took
the data to EPA and used it to argue
that existing federal "driving cycles"
be reexamined, but EPA refused.
~ O,ne of the largest contractors to CRC-
APRAC-has been the Stanford Re-
search Institute at Menlo Park, Cali- "
fornia, which has received $1.3 million
in the last 5 years. John Eikelman, SRI-
coordinator of environmental research
says that a major portion' of SRI's in-
dustrial environmental research has
been with the petrochemical industry.
including measuring pollutant damage
to vegetation, identification of crude
oil in spills, and other work. SRI has
also worked on catalytic emission con-
trol systems and auto parts for various
oilier industry sponsors. For CRC-
APRAC one principal researcher, Harris
Benedict worked on a nationwide assess-
ment of damage to crops attributable
to air pollution; but even this work
illustrates how the thrust of CRC-
APRAC .research, despite its intrinsic -
interest and merit, keeps coming back
to regulatory issues in svhich EPA is.
involved up to its ears.
The SRI researchers surveyed dollar
value losses to corn, ci'.rus, and otiier
food crops and to ornamental plants.
indexed them geographically, and came
up with an ovcral: annual loss csii.T.ati
oi $132 million, far less tlir.ti a previ-
ous estimate of $500 million. The
finding that air pollution doesn'E do .
AUGUST 1973
T5J
-------
..rsuch domea 10 cmp'—- *hich after
I irr rsc«ly f.i rural p.irts of ihs
^xuitr*—M h^d been fcjred his
-rsvcd u-rfful in arguing lyiinst
::ca.-=o; tp every iin-'t; auiomobiie
Jw f*-iad of thme who wam a ;si>
mprtc naiional pollution control strat-
•77 U.-nJ(cd la urb-ui arm. where air
oi!ution ii wont. Another
wudy found thai VH!
natural ~sink~ or ab-vorbcr o/ CO, TJii^
ii a finding «hich"clsar!y nfCccu the
debate ov«r whether overall CO l=vefs
arc increasing or dccrcniin-.j, ami. hence,
over the urjzncy of man'* need to re-
duce ihirm. Both sludivi. ircn have a
link, albeit indirect, to EPA's rcsulv
wry ro'e.
C KC- AP R AC'S research program
mtm be viewed in lijht of the fact
that v>m* of it is •performed by the oil
companies [hsnuclvci. some hy groups
who depend or have depended heavily
un oil ami auto companies for their buii-
ness ' both of \*>hich. have some stake
in .(he regulatory «me. A third partem
ainonj CHC-APP.AC contractors, and
one that further muddles toe issue of
who works for whom, ii that many
of the smaller CRC-APRAC contrac-
tors also take money from the Amen*
can Petroleum Institute and the Motor
Vehicle Manufacturers' Association
directly, from the oil and auto indus*
tries* and in some cawr, from the
government too. An alUn-the family
pattern appears to characterize th*
winning and losing of polluiioii re-
search contracts. For example. Smjtht-
at Scott Uboraiones, noted that alter
EPA cedmed to accept the .industry's
interpretation of its survey* of vehicle
use to rcexamine the tlrivtn j cycles, Scots
was ?blc to continue tKr work through
MVMA sponsorship anyway. Another *
case wa» that of Wilbu "
dates, an intcmatiooal
consulting firm, which had research and
development contracts simultaneously
with EPA and with CRC-APRAC,
According to one of the researchers
there, Wilbur Smith Associate* has sub-
contracted a part of its work to a
Bedford. Masu, aerospace firm CCA
Corp.. which, oddly enough has la
addition held itt own contract direct-
ly with CRC-APRAC Many of CM
• principal investigators interviewed fe-
marked that these 'overtopping, inter-
locking contract awards were typical
of the auto emission research butuicu,
and some added chat it was also a
characteristic of the acru\pacc-dcfcm«
department business in which many of.
thcsa investigator! previously worked.
In fact, several major univ«n»y centcrr
for air potluthm work are comptcuous*
ty absent from the lisi of -40-odd CRC-
^\PKAC contractors, whereas about 1*
'of the contractors are firm prominent
ia the acrosoactf field. Many of the ia*
vc>ti^ators iiitcrview«; only thing worse than an un-
employed aerospace cn^fnccr." ha
quipped. "U aa unemployed ucrospiu
ecgincer who has gone to work oa-
th* emdcoament,"
• Interviewed about the soundness of
policies which appear m cncourafi*
rcusrcfsn ro a«« moo*?
EPA and -the auU and oil hu
maay of tha invesci^un
"Haw ds* would you da ii?"
pointed ou* dut just jr*iaj
morwy to EPA— with A provna iha
EPA gst oui o/ CRC-APRAC— «hisii
is> what NTu»*i**s jtaJ is conaiiicnitj
doing— wouki not »otv« (h« probfcm.
sine* EPA has u much ira«« ia the
Outworn* o( chs rcuana u t&2 ia6nuy
which athecs echoed; that
ante pjvemnieBt body, wrtinj ta cf<
feet as a thtcA paR)* to tl» ^sairarcny.
become die prim« sponMT of auio
emnsMm mcarctL ~l*m amazed ttiu
paru of IIEVV fihc Dtywaol of
Healih. Hducaticn. and
been over^ook«i us all
shouldn't they buiZ4 up a
the N£cHi [Nationl Imanju of EC-
vironmecial Health Sciences!? * . .
They're wod=, They'd, be ideal. - . . iu>
they've been
-------
Exhibit A-2
-------
DATE:
SUBJECT:
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
October 13, 1976
Prediction of Heavy Duty Gasoline Trucks' Transient Emissions from
Steady State or Sinusoidal Test Procedures
FROM: jt^net Becker, CAB
TO:
Gary Rossow, SDSB
THRU: Marcia Williams, Chief, CAB
The intent of this memo is to explore the possibility of predicting
transient cycle emissions for heavy duty gasoline trucks from emissions
as measured over a steady state or sinusoidal procedure. Of particular
interest, of course, is the currently used 9-mode FTP composite, or some
reweighted version of it.
Although it was not clear that analysis had been done to indicate that
there is a difference between single-axle gas trucks and other gas
trucks, SwRI included only the nine single-axle trucks in their analysis
used to determine the predictability of transient emissions from the 9-
mode FTP, or some reweighted version of it.
When the 15 mph and 20 mph average speed transient cycles were linearly
regressed on the 9-mode FTP composite emissions, the best predictive
relationships overall for HC, CO, and NOx were derived from the 15 mph
cycle when the trucks were empty. For this case, as for the other
speed-load combinations, prediction of HC transient emissions was fairly
accurate (correlations between transient HC and 9-mode FTP emissions.
ranged from .93 to .96). Although NOx could be predicted fairly accurately
for the 15 mph, empty truck case, prediction was not acceptable over any
other speed-load combination. Transient CO could not be predicted
accurately via the 9-mode FTP composite over any of the six speed-load
transient cycles.
To determine a. reweighted version of the 9-mode test which would hope-
fully achieve better predictability of transient emissions, a linear
programming technique was used. This technique found the set of modal
coefficients (weights for each of the none modes) which would best
predict transient emissions. The method of least squares was applied.
(At this point in time, SwRI and EPA are jointly evaluating the accuracy
of this approach). The set of modal coefficients was subjected to the
following constraints before the prediction relationship was developed.
1. All modal coefficients had to be non-negative.
2. The sum of the modal coefficients had to equal 1.
EPA
1320-4 i
v. 3.74)
-------
-2-
The magnitude of the various modal coefficients (weighting factors) used
in the resulting prediction equations vary with pollutant, driving cycle
(transient vs. sinusoidal), speed, and load. For example,
1. The closed throttle mode is significant in the prediction of
transient HC at all speed-load combinations, but is insignificant
in the prediction of transient CO and NOx at all speed-load com-
binations.
2. The idle mode generally is significant in the prediction of
transient NOx, but is insignificant in the prediction of sinusoidal
NOx.
3. The 19" mode is insignificant in the prediction of transient
CO, but is highly significant in the prediction of transient NOx.
(However, when the wide open throttle mode is substituted for
the 3" mode, the 19" mode becomes the most significant mode for CO).
On the basis of examples such as these, it appears that a unique set
of modal coefficients which will successfully predict transient HC,
CO, and NOx does not exist. However, no formal anlaysis was done to
support this statement.
SwRI's conclusion that the correlation between transient cycle emissions
and the reweighted 9-mode emissions is generally higher than the cor-
relation between transient cycle emissions and the 9-mode FTP composite
could be fallacious for several technical reasons (including possible
problems with the constraints applied to the system, possible correla-
tion among the supposedly independent variables (the modes), and SwRI's
using the same data used to determine the new weighting factors to decide
whether correlation improved). Also, the fact that some of the reweighted
correlations are smaller than the original correlations leaves the linear
programming technique open to criticism.
Substituting a wide-open throttle mode for the 3" vacuum mode of the
9-mode FTP did not improve the ability to predict transient emissions
for the 15 and 20 mph average speed cycles, according to SwRI's analysis.
With respect to the question as to whether percent change in emissions
as measured over the 9-mode FTP composite can predict percent change
in emissions as measured overa a fully transient cycle, four levels
of emission control were defined: pre-1970, 1970-1973, 1974-1975
Federal, and 1975 California. The average pre-1970 gas truck emissions
for HC, CO, and NOx were used as the bases for the percent changes
calculated. The percent change in emission (by pollutant) for each
of the 18 gasoline trucks was calculated. SwRI used the 10 mph and 20 mph
average speed transient cycles with the trucks half full for this part
of the analysis.
-------
-3-
A linear regression of percent change in transient emissions on percent
change in FTP emissions was performed. For the 1970-1973 level of
emission control, the linear relationships at 20 mph average speed/cycle
yielded better accuracy of prediction than did the 10 mph cycle. How-
ever, accuracy using the 10 mph cycle was significantly worse only for
HC. 'ihe regression equations for the 20 mph average speed cycle are:
HC: y = .95x - 17.11 R2 = .982,
CO: y = .92x + 6.92 R2 = .994
NOx: y -1.82s - 84.30 R2 =1.000
where y = percent change as measured over the transient 20 mph cycle
with trucks at half load,
and x = percent change as measured over the 9-mcde FTP composite.
2
Although the R values are high, suggesting that accurate prediction
via these equations is possible, the practical use of the equations
is limited by at least two considerations:
1. Only four trucks were used to determine this equation
and a priori one would expect the R" to be hi?h,
and
2. These equations could be used only over the limited range of x
values covered by the four trucks in the sample. For example,
one can not logically predict an 84.30% decrease in transient
emissions, given that there was no change in FTP emissions.
(This corresponds to setting x equal to zero and solving
for y).
For the 1974 level of control, the linear regression relationships
between the two transient cycles and the FTP for percent changes
in emissions were accompanied by R^ values below .5, which is un-
acceptable for prediction purposes. These relationships suggest
however, that percent decrease as measured over eicher the 10 or
20 mph transient cycle is greater than percent decrease as measured
over the 9-mode FTP-
It was of interest to determine if emissions measured over sinusoidal
driving cycles could be used to predict emissions as measured over the
fully transient cycles, the idea being that a sinusoidal test would be
simpler and less expensive to operate than a fully transient test.
Linear regression relationships of emissions as measured over the 20 mph
average speed transient cycle on emissions as measured over the 20+5
mph sinusoidal cycles were developed for HC, CO, and NOx. The trucks
-------
-4-
were at half load for this analysis. Accuracy of prediction was poor,
with R values ranging from .2 for HC and NOx to .3 for CO. For CO,
emissions measured over the transient cycle are always greater than
those measured over the sinusoidal cycle for the case studied. This
implies that significant CO on-the-road emissions could occur even if a
vehicle complied with a sinusoidal test for CO.
To determine if emissions on deceleration and acceleration can be
predicted from a steady state test, emissions as measured over the
sinusoidal cycles at 30+5 mph and 40+2 mph (at half payload)
were linearly regressed on the corresponding steady state emissions.
The 40+2 mph equations yielded the better predictions:
HC: y = .995x - .252 R2 = .385,
CO: y = .939x -3.138 R2 = .812,
NOx: y = .940x + .615 R2 = .852,
where y = emissions (gm/min) as measured over the sinusoidal cycle,
and x = emissions (gm/min) as measured over the steady-state cycle.
For this case, accuracy of prediction is fair for HC and NOx, but unacceptable
for CO. The ability of the 30 mph steady-s-tate cycle to predict the 30
+ 5 mph sinusoidal cycle emissions was poor for all three pollutants.
The sinusoidal tests produce lower HC and CO levels than the steady
state tests for the ^0+2 mph case. This fact would indicate that
there are not significant HC and CO emissions during transient maneuvers,
provided that the 40+2 mph cycle can be considered to contain transient
maneuvers.
-------
Exhibit A-3
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
SUBJECT: Transient vs. Steady-State Test Procedures for
Measuring Emissions of Heavy Duty Diesel Trucks
FROM: Janet Becker, CAB 1:
TO: Gary Rossow, SDSB
THRU: Marcia Williams, Chief, CAB
DATE: September 28, 1976
Currently, the Federal Test Procedure used to measure heavy duty diesel
vehicle emissions is a 13-mode non-transient engine procedure. This non-
transient engine procedure is being used to establish the initial compliance
of new heavy duty vehicles with federal emission standards. Since heavy
duty vehicles usually operate in transient cycles, the question Immediately
arises as to the comparability of emission factors obtained via a non-transient
test procedure and those which might be obtained via a transient driving
cycle which represents the driving patterns of a typical heavy duty truck.
Olson Laboratories, under contract to EPA, is in the process of analyzing
data which will be used Co determine such a driving cycle.
Due to the unavailability of test data over the final driving cycle, it is
clearly impossible to answer this comparability question precisely at
the present time. However, it is possible to move ahead on the general
question of the comparability of the 13-mode test results and test results
derived from various transient tests; and thus to anticipate the degree
of comparability between the 13-mode test and the forthcoming representative
driving cycle. Since it appears from preliminary data collected by Olson
Laboratories that a 15-20 mph average speed transient cycle will be selected
as the representative driving cycle, the transient cycles at these average
speeds are emphasized in the present memo. The transient driving cycles
used in this analysis were developed by EPA on the basis of preliminary
CAPE-21 data on a small sample of trucks.
Emissions data on 12 heavy duty diesel and 18 heavy duty gasoline trucks
were collected and analyzed by Southwest Research Institute (under Contract
Nos. 68-03-2147 and 68-03-2220). Data were collected for a variety of
chassis dynamometer tests, including the 13-mode*, steady-state operation,
sinusoidal driving patterns, and completely transient driving cycle
operation. Each test was carried out when the trucks were empty, at 50%
payload, and at GVW. Each test, except the 13-moae, was carried out at a
variety of average speeds.
*It was assumed that the 13-mode test would give the same results on
a chassis dynamometer as on an engine dynamometer.
EPA Form 1320x4 (R»». 6-72)
-------
The question as to whether or not the 13-mode FTP composite can be used
to predict the emissions over the 15 mph transient or 20 mph transient
cycle was addressed via regression analysis. Prediction equations of the
form:
y = ax + b,
where y = emission (gm/min) over the transient driving cycle in question,
and x = 13-mode FTP composite (gm/min) were developed for HC, CO, and
NOx for empty trucks, half-full trucks, and full trucks. Associated with
each such equation is a "coefficient of determination" (R ) , which is
used to help decide the accuracy of the prediction equation. For the
15 mph and 20 mph transient cycles, equations for half-full and full
trucks yielded larger R~ values than did the equation based on empty
trucks' emissions. The R values are given below.
Half-full trucks Full trucks
15 mph HC .440 .446
transient CO .537 .691
cycle NOx .850 .843
20 mph HC .493 .348
transient CO ,,513 .656
cycle NOx .878 .841
2
The R values associated with HC and CO indicate that prediction based
on a linear, equation of the form y = ax •+• b is not very accurate. For
NOx, the R~ value is large enough to use the prediction equations with
more confidence. The two prediction equations for NOx which were provided
by SwRI are:
Equation 1: y = .674 + .641x, R2 = .878,
where:
y =» emissions (gm/min) measured over 20 mph transient cycle
(half load)
and
x = emissions (gm/min) measured via the 13-mode FT? composite
Equation 2: y = .671 + .467x, R2 = .850,
where:
y = emissions (gm/min) measured over 15 mph transient cycle
(full load)
loo
-------
and
x = emissions (gm/min) measured via the FTP composite.
These equations indicate that prediction equations vary with different
speed-load combinations. However, the contributions of speed and
load can not be factored out without the other two equations (20 mph,
full load and 15 mph, half-load), which SwRI did not provide.
Since a linear equation using the currently defined weighting of the 13
modes (resulting in the 13-mode FTP composite) does not predict transient
emissions well for all three pollutants, the 13-modes were reweighted via
a linear programming technique. The objective was, for each load-average
speed transient cycle, to find a linear combination of the 11 modes,
(although there are 13, 3 are idle and these modes were combined for this
analysis so as to decrease the number of estimated parameters) which would
equal the emissions measured over the given transient test. This objective
was subjected to the constraints:
1) All modal coefficients must be non-negative,
and
2) The sum of the coefficients must equal one.
These constraints could prove to be very limiting if, for example, the
emissions over the transient cycle were consistently higher than over
any of the modes. Constraint 2 is particularly questionable from a
mathematical viewpoint, although perhaps is supported from the engineering
angle. Since all modes were included in the regression, another questionable
assumption is the independence of the modal tests correlation among the
independent variables affects the estimates of the modal coefficients.
The results of-this linear programming technique were equations approximating
a given transient cycle's emissions on the basis of the reweighted 13 nodal
values. The half-load equations provided the best overall prediction for
all three pollutants.
15 mph, half-loaded trucks (modes are in parentheses and are explained below)
HC : y = .180 (1/7, 13) + .507(2) + .170(5) + .144(6) + .002(3), R2 - .345
CO' : y = .271 (1,7,13) + .202(4) + .243(5) + .021(6) + .067(8) + .97(10).
R2 = .773
NOx: y - .351(1,7,13) + .433(2) + .164(5) + .010(8) + .042(9). R2 - .«17.
where:
y = emissions (gm/min) over the transient cycle at
JO/
-------
Explanation of modes
(1,7,13) Idle
(2) 2% peak torque at peak torque speed
(3) 25% peak torque at peak torque speed
(4) 50% peak torque at peak torque speed
(5) 75% peak torque at peak torque speed
(6) 100% peak torque at peak torque speed
(8) 100% rated horsepower at rated speed
(9) 75% ratedhorsepower at rated speed
(10) 50% rated horsepower at rated speed
(11) 25% rated horsepower at rated speed
(12) 2% rated horsepower at rated speed
Keeping in mind the fact that 12 observations are being used to estimates
11 parameters, these R values are surprisingly low. Undoubtedly the
correlation among modes is influencing the R values, as well as the
constraints that were put on the system. Until further analysis is
done, these equations are not recommended for use in predicting transient
emissions.
SwRI considered the question of whether percent change in emissions over
transient cycles can be predicted from the percent change in emissions
over the 13-mode FTP composite by looking only at the 10 mph and 20 mph
transient cycles. The diesel trucks were divided into two groups determined
by method of emission control: pre-1974 models and 1974-1975 models. A
pre-1974 average emission (gm/min) was calculated for each test cycle.
For each of the five 1974 and 1975 model-year trucks, a percent change
in emissions was calculated on the basis of the pre-1974 average level.
The percent changes observed over the transient tests at 10 mph and 20
mph average speeds were linearly regressed on the percent changes observed
in the 13-mode FTP composite, and the following relationships resulted:
Transient, 10 mph avg. speed
HC : y = .llx—16.52 R2 » .012
CO : y = 1.03x + 8.54 R2 = .531
NOx: y = l.llx + 2.23 R2 = .795,
where:
y = percent change over transient cycle,
and
x = percent change over 13-mode FTP composite.
-------
5
Transient, 20 mph avg. speed
HC: y = -.04x - 17.89 R2 = .002
CO: y - 1.58x + 8.15 R2 = .700
NOx: y = 1.06x + 2.86 R2 = .865,
where:
7 = percent change over transient cycle,
and
x = percent change over 13-mode FTP composite
From a statistical viewpoint, only for NOx could the percent change in
the 13-mode FT? be used to pradict percent change over the transient
cycles with any degree of accuracy. One caution is that from looking
at a scatter plot for NOx, it appears that the relationship might be
curvilinear as opposed to linear, so possibly the prediction equation needs
a quadratic term.
A sinusoidal driving cycle is a compromise between a steady-state cycle
and a fully transient cycle. One question of interest is how well the
sinusoidal predicts the fully transient emissions. SwRI used 20+ 5 mph
sinusoidal and 20 mph average speed fully transient half-load data to
derive the following linear relationships.
HC : 7 - .519x + .368 R2 = .686
CO : 7 - .635+5.465 R2 = .118
NOx: y - .923x + 2.501 R2 - .946
where:
7 = emissions (gm/min) over 20+ 5 mph sinusoidal at half-load
and
x = emissions (gm/min) over 20 mph transient cycle at half-load.
For CO, sinusoidal emissions should not be used to predict transient
cycle emissions. For HC, prediction of transient from sinusoidal
is borderline, and for NOx, predictability looks good.
To see if the accelerations and decelerations that are missing from a
steady-state cycle can be accounted for via a linear relationship,
linear equations were developed to predict sinusoidal emissions (30+ 5mph
and 40+_ 2 mph from steady-state emissions at 30 and 40 mph at half-load.
The prediction was better using the 40 mph cycle, and these equations are
103,
-------
HC: y = l.lSx + .363 R2 = -686
CO: y = 1.074x + .926' R2 = .569
NOx: y = 1.791x -r .665 R2 = .971.
Where:
2
y = emissions (gm/min) over 40imph steady-state cycle at half-load,
and
x = emissions (gm/mia) over 40 niph steady-state cycle at half-load.
Only for NOx is the prediction relationship good enough to use with
confidence. The equations for KG and CO are poor-to-borderlina predictors.
Moreover, the steady-state emissions are less than the sinusoidal emission
levels, indicating that significant emissions occur for all pollutants
during transient maneuvers.
In conclusion, it appears that only for NOx can a version of the test
(either the FTP composite or a reweighted version) be used to accurately
predict transient cycle emissions. Also, sinusoidal emissions are
greater than steady-state emissions for all three pollutants implying
that significant emissions occur during acceleration or deceleration.
Thus, it seems possible that manufacturers could design emission control
systems which would satisfy a version of the 13-mode FTP (a composite
of steady-states), but still would permit unacceptably high emission
levels when trucks are on the highways in transient operation.
-------
B. Issue - Redefinition of "Useful Life"
1. Summary of the Issue
In the February 13, 1979 NPRM, EPA proposed that the current
definition of "useful life" be changed for heavy-duty engines.
Currently in the regulations useful life is interpreted as approx-
imately half of the service seen by a typical heavy-duty engine;
specifically, for gasoline-fueled engines, 5 years, 50,000 miles,
or 1,500 hours of use, whichever occurs first and for diesels, 5
years, 100,000 miles, or 3,000 hours.
The proposal extends this "useful life" period to the "average
period of use up to engine retirement or rebuild, whichever occurs
first." The manufacturers would themselves determine this average
value for each engine line they manufacture. In no case, however,
may the useful life of any heavy-duty engine be less than 50,000
miles or 5 years nor less than the basic mechanical warranty on the
engine. For most engines, this change more than doubles the useful
life period and thus has significant effects on durability testing
and warranty obligation.
2. Summary of the Comments
A large number of comments dealt with EPA's justification for
changing the useful life definition. Concern was expressed that
Congressional intent was violated and that the divergence from past
regulatory experience was unwarranted. Second, that inherent
quality of a full-life useful life which requires lifetime emis-
sions compliance was seen as an increase in the stringency of the
emission standards. Third, commente I s criticized the "average"
aspect of the useful life definition for causing unnecessary
problems. Finally, EPA received a range of comments that all
revolved around difficulties in actually determining a useful life
value for a given engine line. The following paragraphs briefly
expand on these four areas of comment.
The comments that were directed at Congressional intent cite
portions of the legislative histories of both the 1970 Clean
Air Act (which first addressed "useful life") and the 1977 Amend-
ments to that Act (which made modifications to the language of
useful life provisions). Often quoted were the passages which made
clear that the legislators indeed understood that the 50,000-mile
"lifetime" they chose for durability and warranty purposes in 1970
approximated only half of the expected life of a light-duty vehi-
cle. Thus, said the commenters, Congress explicitly wove the
half-life concept into the Act. Also, an excerpt from a Senate
report preceding the 1977 Amendments explains the rationale behind
the wording of §202(d)(2) (which allows the Administrator to
lengthen—but not shorten—the useful life for non-light-duty
vehicles) as providing greater flexibility in defining the dura-
/OS-
-------
bility of trucks. Chrysler in particular saw in this excerpt
"no intention to abolish the half-life concept." Finally, several
comments contended that Congress had meant for there to be a
"balance" between the treatment of light-duty vehicles (LDVs) and
heavy-duty engines (HDEs) in useful life matters.
A related area of comments involved the implications for
heavy-duty engines of the recently decided Court of Appeals
action relating to motorcycle useful life._l_/ This court pro-
ceeding was going on as Congress considered the 1977 CAA Amend-
ments, and the record shows that the legislators knew of and
responded to the useful life controversy. Their answer was to
remove motorcycles from §202(d)(2) and create §202(d)(3) to specif-
ically allow a shorter useful life than 50,000 miles/5 years for
motorcyles. Some commenters argued that by leaving the language
affecting other non-light-duty vehicles and engines (including
HDEs) unchanged, Congress "implicitly continued to recognize the
half-life concept" for these vehicles (MVMA). Additionally,
Mercedes Benz pointed to the Court's opinion in which the court
argues that, according to the legislative history of the Amend-
ments, the approach used in the 1970 CAA for LDVs and HDEs was
"reasonable."
Most of the commenters noted that the half-life concept has
been a part of vehicle emissions regulations since the 1966 (HEW)
rules applying to 1968 model year vehicles, preceding the earliest
statutory mention of useful life. In that rulemaking, 100,000
miles was identified as the basis for "lifetime emissions." Under
the assumption that emissions deteriorations would be linear, HEW
established a procedure for calculating average lifetime emissions
at the approximate half-life (50,000-mile) point. All subsequent
regulations for light-duty vehicles, light-duty trucks, and heavy-
duty engines have used half of the expected life as the useful
life. The comments imply that the average lifetime emissions
concept has embodied the intentions of Congress through the years
and that for EPA to now change that concept for HDEs is unwar-
ranted.
A final area of comments which strikes at the actual concept
of a full-life useful life argued that it acts to increase the
stringency of the emissions standards. As the heavy-duty regu-
lations are presently constructed, manufacturers must design their
engines so that during approximately the first half of their
lifetime their emissions do not deteriorate past the level of the
standards. This situation requires that the emissions of a new
engine be somewhat below the standard in order that deterioration
may be accommodated. The proposed full life concept would require
lifelong emissions compliance and hence a still lower initial level
of emissions. This is the "increased stringency" referred to in
the comments. Some of the commenters went on to claim that EPA is
in effect requiring a reduction in emissions in excess of the 90
percent minimum set by Congress.
fof
-------
The two remaining major areas of comment were not directed at
the basis of the full-life concept so much as at specific problems
which might be expected to arise from EPA's proposed application of
the concept. The first of these is the language of the NPRM that
requires the manufacturers to determine an "average" period of
engine use for each engine line. The comments hinge on the impli-
cation that half of the engines subject to an average useful life
will require rebuild or retirement before they reach that useful
life. While some commenters implied that an emission warranty
claim would result in each case, most said that a flurry of claims
could be expected to result from decay in emission-related compo-
nents toward the end of the useful life. Also, there was consid-
erable concern that the emissions-related warranty would be con-
fused with the engine warranty, sparking warranty conflicts, and
that the full life warranty on emissions would require coverage of
related parts beyond commercially-sound limits. Finally, Ford
urged a labeling change that would make it clear that the useful
life number was given "for the sole purpose of the emission-system
warranties required by the Clean Air Act."
The last set of comments were procedural in thrust and revolve
around the difficulties that the manufacturers would expect in
defining a useful life number under the proposed full-life concept.
First, data concerning actual engine usage periods is largely
unavailable at this time. Additionally, the lack of specificity in
the "retirement or rebuild" useful life limit drew comment since the
decision of when to retire or rebuild is reached by the user on
largely economic—as opposed to mechanical—grounds. Thus, manu-
facturers would find it difficult to arrive at an average period
for this event for an engine. The problem would be further com-
pounded by the wide range of vocational applications seen by many
engine families which makes the rigor of duty a quite variable
entity. Ford is probably a worst case example because over 65
percent of their gasoline trucks are sold as incomplete chassis;
this, they claim, prevents them from knowing the end uses of their
engines.
A treatment of comments relating to several additional useful
life issues, more minor in scope than those above, may be found in
Part II of this document.
3. Analysis of the Comments
The same order in which the issues were summarized in the
previous section will be followed as the issues are discussed and
analyzed below.
a. Congressional Intent/Regulatory Precedent/Stringency
The language and the legislative history of the Clean Air Act,
-------
as amended, support the proposed changes to a full-life useful life
definition.
Nowhere does the Act place a half-life useful life restraint
on heavy-duty useful lives. Quite to the contrary, Section 202(d)
(2) clearly provides the Administrator with the discretion to set
the useful life for a duration or mileage greater than that set by
Congress for light-duty vehicles if he determines that a greater
duration or mileage is appropriate. Given the need to create an
incentive for manufacturers to build emission control components as
durable as the rest of the traditionally long lasting engine parts
and the significant air quality benefits that will be realized if
the proposed definition of useful life is adopted (see discussion
below), adoption of the proposed definition is certainly "appro-
priate" and well within the descretion explicitly granted to the
Administrator by the Act.
Nor does the legislative history evidence a Congressional
commitment to impose a half-life restraint on setting the useful
life for heavy-duty vehicles. While the final outcome of the 1970
Clean Air Amendments was a conscious defining of light-duty useful
life to be half of the expected actual life, it seems that decision
was a result of forces that were present at that specific time and
to that specific class of vehicles. A 100,000 mile/10-year re-
quirement was seriously considered by a Senate committee _2/, but
was halved largely as a compromise response to the light-duty
vehicle industry reaction against any sort of performance warranty
: any
ies).
(versus "parts and labor" warranties
There was no similar commitment to the half-life concept with
respect to heavy-duty vehicles and engines; nor was there any
indication that Congress intended to "balance" the treatment of
LDVs and HDEs in the manner suggested by the commenters.
There is no reason to believe, as some commenters suggest,
that when Congress removed motorcycles from the vehicles affected
§202(d)(2) and created §202(d)(3), they meant to endorse the
half-life concept that was then being applied by EPA to heavy-duty
engines. Rather, in creating a separate provision for motorcycles,
Congress was simply interested in retaining the 50,000 mile/5-year
minimum for those "other motor vehicles," while at the same time
expressly authorizing EPA to adopt a useful life of less than the 5
years/50,000 minimum set by §202(d)(2). Had there been a desire to
place a half-life constraint on EPA, Congress could have easily
inserted such language at that time.
The Court in Ear ley-Davidson v- EPA. 598 F.2d. 228 (D.C.
Cir, 1979) did indeed agree with Congress that motorcycles should
be treated differently from heavy-duty and light-duty vehicles, but
not on the basis of half-life vs full-life useful life. Rather the
Court concurred with Congress that a useful life of less than ~5
-------
years or 50,000 miles for motorcycles was appropriate, but the 5
year/50,000 mile minimum should be retained for other mobile
sources.
Turning now to the comments which implied that past regulatory
practice should constrain future rulemaking, the staff takes a
somewhat different view. The regulations promulgated by EPA must
be the best attempt possible at that time to fulfill the wishes of
Congress within the context of feasibility, cost, and other fact-
ors. Indeed, in 1970, early in the history of vehicle emissions
regulations, HEW established a useful life concept which allowed
heavy-duty engines to exceed the standards for the greater part of
their lives. This interpretation may be evidence of an uncertain
regulatory climate during that time frame, but the provisions were
clearly not mandated by Congress. Further, the only other regula-
tions involving useful life since the 1970 CAA, those pertaining to
motorcycles and aircraft, apply the full-life concept.
The final area of comment which affects the full-life concept
itself is the stringency issue. This idea might be better treated
in the broad context of how it fits into the total full-life useful
life plan. If the Administrator has the discretion to adopt a
full-life useful life, then a lower zero-mile emission level is
simply a practical result of applying that requirement to the
certification process. Thus, we agree that, in a narrow sense, the
design-goal emission level is more stringent under a full-life
useful life concept. But, in the larger perspective, the standards
themselves are not more stringent; they are simply met for the
lifetime of the engine. The staff cannot accept the stringency
issue as an argument against the full-life useful life. In any
event, Congress asked for standards representing a reduction of "at
least 90 percent", provided they are technologically feasible
(emphasis added).
The staff position on the idea of a. full-life useful life as
formulated before receipt of the comments remains largely un-
changed. That position is summarized as the following:
Because of the extended periods of use seen by HDEs, continued
functioning of emissions systems is vital. The present "half-life"
useful lives in reality represent something less than half of the
actual lifetimes of most engines. Thus, to assure the air quality
benefits for this package are realized—and that the consumers get
their money's worth—it is necessary for the emissions systems to
function close to the full life. In no instance is this more clear
than in the case of gasoline-fueled engines, which will be equipped
with catalyst technology for the first time. Absent an incentive
to design appropriate durability into these components, one would
expect a congregation of catalyst failures around the minimum
useful life point. Similar logic holds as well for diesel manufac7
turers as they improve the durability of their emission-related
-------
components. It is the incentive for durable design, then, that is
the crux of the staff's argument for a full-life requirement.
In light of our previous discussions, concluding that neither
the stringency issue, existing regulations, nor Congressional
intent constrain EPA to a half-life policy, we continue to support
the idea of a full-life useful life concept, subject to practical
improvements in its application.
b. "Average" Useful Life and Problems in its Determination
The remaining discussion will deal with the practical dif-
ficulties associated with the full-life useful life concept. The
first of those is the proposed requirement that the useful life
value supplied by the manufacturer be the "average" for that engine
family. The staff has considered alternative methods of estab-
lishing the useful life number, though no commenter offered a
suggestion along these lines. For example, the alternative of
allowing complete latitude in defining the useful life is likely to
encourage unrepresentative values. Depending on whether a manufac-
turer places emphasis on quick durability programs and few warranty
claims or rather would favor a lengthy durability program to delay
the use of an in-use df, a manufacturer might either gravitate
toward the lower useful life limit or place the useful life too
high, respectively. (One present lower limit as proposed is the
basic engine warranty, which is based on economic grounds and can
be expected to undershoot the actual period the engines are on the
road in most cases.) Another alternative could be for EPA to
establish that some percentage of the period of use or some per-
centile of the retirement/rebuild distribution be used instead of a
straight average. This option, however; suffers from a complete
lack of data to support any specific numbers.
The staff must conclude that specifying that an "average"
useful life be determined is the best way under a full-life useful
life plan of balancing fairness to the industry with some measure
of assurance for EPA that the chosen "period of use" is accurate.
Regarding the flood of warranty claims anticipated by the
commenters, the staff disagrees that under the proposed rule half
of the manufacturers' engines will require emissions warranty work.
Although it is clear that half will reach their individual retire-
ment/rebuild points, this does not necessarily mean an emissions
violation will exist in every case. Certainly there will be a
number of additional warranty claims attributable to the extension
of the useful life period. The Agency does not, however, expect
this number to be excessive. Additionally, the proposed regula-
tions imply that the manufacturer would be responsible for post-
rebuild emissions compliance.
110
-------
The staff has recommended changes in the proposed rule which
speak to each of these issues. First, new provisions are included
which define the end of an engine's useful life as the average
period of use or the point at which the engine needs rebuilding,
whichever is reached first (provided that the 50,000 mile/5 year
minimum has been passed). Thus, the cost of the rebuild, as well
as all subsequent repairs, will be borne by the owner—not by the
manufacturer. Similarly, the problem of warranty conflicts and
misleading labels are answered, we think, by the recommended
policy of allowing the manufacturer to define useful life values
for separate service applications (again, see Conclusions and
Recommendations). The last point made by manufacturers in the area
of warranty was that a full-life emissions warranty will force
emission-related parts to be guaranteed beyond the point of eco-
nomic wisdom; i.e., for the full useful life of the engine. The
staff can only respond that the purpose of an emissions warranty is
to assure that emission control components on in-use vehicles and
engines are operating. This necessarily requires manufacturers to
design emission control components with adequate durability, even
if the costs incurred might be considered too large by standard
marketing criteria.
EPA does not believe that the costs associated with Section
207(a) warranties will increase significantly when the extended
useful life is implemented. (As we stated under Issue F. - Idle
Test and Standards, Section 207(b) warranty regulations have not
yet been implemented; any costs associated with such regulations
will be treated in future emission performance rulemaking pack-
ages.) Because costs have been included in our economic analysis
to cover the increased durability of emission control components we
have at least partially accounted for the costs which might be
incurred if a manufacturer chooses not to work toward more durable
components and as a result suffers warranty claims. Moreover, any
effect on the aftermarket industry is likely to be minimal since we
expect a comparatively small number of additional warranty claims
to result from the redefinition of "useful life."
The final area of comment regarding useful life pertains to
the anticipated problems with accurately specifying an average
useful life value. While we do recommend that an average be
required, we agree that such difficulties need to be considered in
order to achieve the best design for a full-life policy.
Our agreement with the commenters does not in all cases extend
to the severity of the problems. It is not surprising to us, for
example, that a broad base of engine-life data does not exist in a
convenient form. But since the regulations would give the manufac-
turers a reason to collect this kind of data, we believe that it
could be obtained relatively easily once the effort is made
(through telephone surveys, for example). The staff is convinced
that such information would be accessible even for engines sold SLS
-------
incomplete vehicles. Also, it is the average lifetimes of the
engines themselves, not their emission controls, that is of in-
terest; so the comments about the data gathering being complicated
by the advent of new control technology apply only for those
engines whose basic mechanical deterioration is affected by the
controls. We expect this to be a rare situation. The upshot is
that useful life data is available now for all engines which are
not to be radically changed for 1983 certification. Further, the
staff believes that the acquisition of such information will
facilitate the approximation of the useful lives of new engines as
well.
The staff has been able to directly address the remaining
problems (i.e, the lack of specific retirement/rebuild criteria and
the variation in engine vocation within an engine family) through
recommended changes in the regulations, as explained in the Con-
clusions and Recommendations section below.
4. Staff Recommendations
On the basis of the comments and their analysis above, the
staff recommends that the useful-life provisions as proposed
be retained largely intact. Three significant changes are offered,
however, which respond to a wide range of comments.
As we concluded during the Discussion above, the staff be-
lieves that the full-life useful life concept should remain
a part of this Rulemaking. Within this context, we advocate that
the language "average period of use" be kept intact for the sake of
practicality. Since the manufacturers will be setting the useful
life values, EPA's requiring that value to be an average appears to
be the most reasonable method of encouraging accurate useful lives.
Several of the difficulties associated with an "average"
useful life, however, will be reduced or eliminated if certain
staff recommendations are adopted. Specifically, we support 1) a
set of more objective criteria for determining when rebuild is
necessary, 2) a manufacturers' option to supply for the owner
alternate expected useful lives depending on service application,
and 3) modifying the "useful life" definition to be less restric-
tive of the manner in which the useful life is determined.
The first of these suggested changes is the most significant
and would remove much of the uncertainty in defining an "average
period of use up to engine retirement or rebuild." The major
criterion for determining whether an engine is due for a rebuild
would appear on the label and would be, for the purposes of this
rulemaking, a compression test, along with a measure of oil
consumption and of bearing failure. Those tests will cover
nearly all mechanical situations which normally signal the need for
a rebuild. Since the actual test values will be determined by the
-------
manufacturer for each engine family, establishing the average
useful life should be easier and more accurate. Another impli-
cation is that an "actual useful life" will exist for each indi-
vidual engine; there will be a measurable endpoint to the manufac-
turer's obligation for an engine with respect to both durability
testing and the emissions warranty. Thus, the regulations clearly
will not require post-rebuild emissions compliance.
The second recommendation amounts to allowing a qualifying
statement on the label to indicate to the owner that the useful
life of this particular engine can be expected to vary from the
"average" due to a lighter or heavier service application. The
label could also direct the reader to the operator's manual for
information about vocation-specific average useful lives, about how
the emissions-related warranty differs from the mechanical war-
ranty, etc. The purpose of the label change is to promote user
understanding of the "average useful life" concept and hence to
reduce the threat of warranty conflicts.
The final staff recommendation is to remove from the defin-
ition of useful life the restriction that for new engines the
useful life be determined from durability testing. We see this
provision as an unnecessary complication of the process of estab-
lishing a useful life value.
Some of our recommendations, particularly the first two, will
to a. certain extent add to the complexity of portions of the
regulations and the certification process as compared to the
original proposal. However, the staff is firmly convinced that by
making these adjustments to the proposal, EPA will not only answer
a range of reasonable comments but will improve the workability,
versatility, and fairness of the full-life useful life concept. We
urge the adoption of these provisions.
112
-------
References
_!_/ Harley-Davidson Motor Company, Inc. v. EPA, U.S. Court of
Appeals, D.C. Circuit, No. 77-1104, March 9, 1979.
2J Legislative History of the Clean Air Act Amendments of 1970,
Senate Public Works Comm. Print No. 1, August 25, 1970,
§207(c).
-------
G. Issue - In-Use Durability Testing
In order Co becter respond to comments on the proposed in-use
durability testing procedure and to optimize all components of the
program, EPA is delaying the finalization of the in-use durability
testing requirements. Further analysis of the design of the
durability program will continue and finalization of the program
is expected to occur on the same time line as the statutory NOx
reduction. The summary and analysis of comments on this component
of the proposal are not included in this document. Instead, they
will be addressed when the in-use durability regulations are
finalized.
Beginning in 1984, and continuing until finalization of a
revised durability testing procedure, the burden of durability
testing will be shifted to the manufacturers. Under this concept,
the manufacturers will determine their deterioration factors in
programs which they design. EPA will not approve the programs
which the manufacturers design but will require that they: 1)
describe their durability testing program in the certification
application, 2) certify that their durability testing procedures
account for deterioration of emission related components and other
critical deterioration processes, and 3) adhere to the maintenance
requirements as applicable specified in the allowable maintenance
regulations. These requirements are the same as those proposed for
the determination of the preliminary deterioration factor.
Manufacturers are encouraged to begin small-scale in-use
durability programs in the near future so they can gain some
meaningful experience with in-use durability testing. This will
benefit the manufacturers and EPA in that they could generate
in-use durability data which could verify the feasibility of and
need for an in-use type durability testing program.
EPA has chosen to finalize its proposal of multiplicative
deterioration factors for all heavy-duty engines. The comments
received concerning this aspect of the proposal are summarized and
analyzed below.
1. Summary of the Issue
EPA has proposed that multiplicative deterioration factors be
used for both gasoline-fueled and diesel heavy-duty engines.
2. Summary of the Comments
Multiplicative Deterioration Factors
Diesel engine manufacturers commented that there is no support
for the use of multiplicative DFs for diesel engines. The manu-
facturers further commented that the use of a multiplicative DE
I/T
-------
unduly penalizes engines with low initial emissions and makes Che
design goal standard more stringent regardless of the actual
deterioration properties of the engine. Gasoline-fueled engine
manufacturers provided no conclusive evidence that catalyst-based
technology should not require a multiplicative DF. Only Ford Motor
Company supplied data to support their claims, but this was based
on Non-Methane Hydrocarbon data.
3. Analysis of the Comments
As stated in the regulatory analysis which supports this
rulemaking action, the use of multiplicative DFs over additive
DFs or vice-versa is not unequivocally supported by theoretical or
empirical considerations. EPA's position on the multiplicative DF
issues is discussed in three sections below: (a) the importance of
using the correct type of DF, (b) multiplicative DFs for gasoline-
fueled heavy-duty engines, and (c) multiplicative DFs for heavy-
duty diesel engines.
(a) The importance of using the correct type of DF.
Given two durability fleets, one called "clean" and one
"dirty":
+ = additive DF value; X = multiplicative DF value
Std.: emission standard
4K Point: emission level at the 4,000-mile (125 hr) emission
test
UL: emission level at the end of the useful life
Case 1 — "Clean Durability Fleet"
Std.: 1.3 g/Bhp-hr
4K Point: .5 g/Bhp-hr
UL: 1.0 g/Bhp-hr
+ = .5 g/Bhp-hr X=2
With these DFs, emission data engines can have 4K emissions
as high as .8 g/Bhp-hr with an additive DF, but only .65 g/Bhp-hr
with a multiplicative DF.
Case 2 — "Dirty Durability Fleet"
Std: 1.3 g/Bhp-hr
116
-------
4K Point: .5 g/Bhp-hr
UL: 1.2 g/Bhp-hr
+ = .7 g/Bhp-hr X = 2.4
With chese DFs emission data engines can have 4K emissions as
high as .6 g/Bhp-hr with an additive DF, but only .54 g/Bhp-hr with
a multiplicative DF.
Thus, in either of the cases discussed above, the additive DF
would give an emissions deterioration cushion for the manufacturers
which would be more desirable for certification carryover and SEA.
EPA's concern is best understood by studying the impact of
using the results of the durability testing for one family to apply
to all calibrations within that family. Using the data provided in
cases 1 and 2 shown above, the cases below show the impact when an
additive DF is used when a multiplicative DF is appropriate and
vice versa.
Case 3 — "Clean Durability Fleet"
+ = .5 g/Bhp-hr
X = 2
If an additive DF is allowed when a multiplicative DF is
really appropriate, then the error could be: (1.3 g/Bhp-hr)
computed - (1.6 g/Bhp-hr) actual = -.30 g/Bhp-hr, or actual deteri-
oration to 1.6 g/Bhp-hr when the standard is 1.3 g/Bhp-hr.
If a multiplicative DF is allowed when an additive DF is
really approprite, then the error could be:
(1.3 g/Bhp-hr) computed - (1.15 g/Bhp-hr) actual = .15
g/Bhp-hr, or actual deterioration to 1.15 g/Bhp-hr when the stan-
dard allows 1.3 g/Bhp-hr.
Case 4 — "Dirty Durability Fleet"
+ = .7 g/Bhp-hr
X = 2.4
If an additive DF is allowed when a multiplicative DF is
really appropriate, then the error could be:
(1.3 g/Bhp-hr) computed - (1.44 g/Bhp-hr) = -0.14 g/Bhp-hr,
or actual deterioration to 1.44 g/Bhp-hr when the standard is 1.3-
g/Bhp-hr.
-------
If a multiplictive DF is used when an additive DF is more
appropriate, then the error could be:
(1.3 g/Bhp-hr) computed - (1.24 g/Bhp-hr) actual = +.06
g/Bhp-hr, or actual deterioration to only 1.24 g/Bhp-hr when the
standard allows 1.3 g/Bhp-hr.
It can be seen from the cases shown above that:
1. a multiplicative DF allows the maximum air quality
protection because it yields a buffer when applied in a
situation and which may be additive and is the correct meth-
odology in a multiplicative situation;
2. an additive DF in a multiplicative situation could allow
an engine to exceed the emission standard, but has no effect,
positive or negative, in a situation which is really additive;
3. the effects of interchanging DF determination method-
ologies decreases as the actual amount of deterioration
from the same starting point increases.
(b) Multiplicative DFs for gasoline-fueled heavy-duty en-
gines.
The catalyst-based technology anticipated in heavy-duty
gasoline-fueled vehicles supports the use of multiplicative DFs
because catalysts reduce the engine-out emisions by the same
percent regardless of any slight variability in the engine-out
emissions .
EPA's rationale for the use of multiplicative DFs for catlayst
equipped engines is outlined below.
The condition that must be satisfied for the use of multipli-
cative deterioration factors, rather than additive factors, to be
appropriate is that differently calibrated engines in the same
engine family experience the same percentage increase in emissions
over a given interval of service accumulation. This can be shown
to be the case under suitable assumptions.
Let: E (M) = engine-out emissions of a pollutant as a
function of mileage M;
EC (M) = tailpipe emissions of the same pollutant as a
function of mileage M;
e(M) - 1 - _ = catalyst efficiency as a function,
of mileage M,
-------
M.= initial reference mileage;
M = final reference mileage;
(1) represent engine 1 of the family; and
(2) represent engine 2 of the family
Assume that engine 1 and engine 2 are calibrated differently
and therefore have slightly different engine-out emissions, but
that the efficiency of the catalyst as a function of mileage is the
same for the two (i.e., that the catalysts are equivalent when new
and that the slightly different engine-out emissions do not signi-
ficantly affect catalyst deterioration). Then,
E^1>2)(M.) = E(1'2)(M.)(1 - e(M.))
tp i eo i i
EJ1>2)(MJ = E(1'2)(MJ(1 - e(Mj)
tp r eo f f
Etp'2)(V = E(1eo)(V(1 " e(Mf)}
. .
tp i eo i
If it is further assumed that for both engines the deteriora-
tion in engine-out emissions is negligible, then
E(l,2) E(1,2)(M
eo f eo i
and
Ov
EtP Mi
= E(2)(MJ
tp f
E(2)(M.)
tp i
This is the condition which must be satisfied for a multiplicative
deterioration factor to be appropriate.
This conclusion has been reached using three fairly reasonable
assumptions: 1) engine-out emissions deteriorate very little; 2)
catalysts on differently calibrated engines deteriorate identical-
ly, and 3) catalysts can be modeled as proportional reduction
devices .
//f
-------
(c) Multiplicative DFs for heavy-duty diesel engines.
In order to unequivocally demonstrate that an additive DF is
preferential to a multiplicative DF, one must first define the
conditions which demonstrate these two situations.
Additive DF: like engines with different initial emission
levels deteriorate the same absolute amount.
Multiplicative DF: like engines with different initial emis-
sion levels deteriorate in an amount directly proportional to their
initial emission levels.
The best possible example to illustrate the deterioration
nature of heavy-duty diesel engines would be to have emission
results from two engines within the same family which have both
been tested for emissions durability (1,000 hr). Unfortunately,
the current durability testing program only requires one engine
from each family to meet the durability requirements.
As a possible alternative to this approach, EPA has studied
the emission results from 37 heavy-duty diesel engines which
underwent emissions durability testing for 1979 certification.
Figures C-l, C-2, C-3, and C-4 which follow this discussion are
plots of 125 hr emission levels versus DF. Specifically they are:
Figure C-l: HC 125 hr vs. Additive DF
Figure C-2: CO 125 hr vs. Additive DF
Figure C-3: HC 125 hr vs. Multiplicative DF
Figure C-4: CO 125 hr vs. Multiplicative DF
The lines drawn on each of the figures represent the best fit
straight line through the data points.
Although the best fit lines appear to adequately represent the
data points, no statistical significance can be found. The R^
values for these regressions were:
Figure C-l: 0.00190
Figure C-2: 0.04116
Figure C-3: 0.01410
Figure C-4: 0.01140
In addition, and perhaps more importantly, the data points as
shown in the figures do not conclusively support an additive or
multiplicative DF for HC or CO when compared to the definitions
given earlier. In short, deterioration does not appear to be the
same absolute amount when initial emission levels are different nor
does deterioration appear to increase as the initial emission level
increases. Thus, neither type of DF is conclusively supported..
-------
^SCATTER VAR=£,1 CAsES = ALL INTEHVAL= ( - .24 . .84) I < 0 . 1 . 35) HEAD=10 HC.OP VS HC.J35HR>
SCATTEK PLOT HC.OF VS HC.12E.HR
N= 34 OUT OF 37 3.HC.DF VS. 1.HC.1Z5HR
HC.DF
.84000 *
FIGURE C-l
.72000
us HR EMISSION; LEVEL vs.
He
OP
.60000
.48000
_ .36000 «
.24000
.12000 *
-.13876-15+
u » « • a
-.12000
-.24000
0.
.15000
.___«____^____,___««_•__+_-__*--_-*----*----»----»----»----»----»----+-___»
.30000 .60000 .90000 1.3000 HC.J25HR
.45000 .75000 1.0500 1.3500
-------
SCATTER PLOT CO.OF VS C0.12SHH
N= 36 OUT Of 37 4.CO.OF VS. 3.C0.125HR
CO.OF
4.2000 «
FIGURE C-2
3.6000 *
US HR. EfllS-SlOK)
VS, ftDDlTl\/& DP
CO
3.0000
2.4000
l.BOOO
1.2000
.60000
-.27062-14+
«» a «
2 • •
•2
-.60000 *
-1.2000 *
4
0.
2.6000
1.3000
3.9000
5.2000
7.8000
6.5000
9.1000
10.400
C0.125HR
11.700
-------
SCATTER PLOT HC.MULTOF VS HC.12SHW
N= 37 OUT Of 37 12.HOULTOr VS. l.HC.12bH«
HCMLiLTOF FIGURE C-3
1.6667 » « """
1.1383 *
.96221 *
.78609 *
.60998
.43386 *
.25775 *
KR
VS. HULTl PL\cAT\\/E
1.4906 *
HC.
.81633 -1+
1
0.
.30000
.60000
.90000
1.2000
.1SOOO
.45000
.75000
1.0500
HC.125HR
1.3500
-------
-jv-Hiii-r\ ,Mn-iti.j <-n jc J-MUL. i iv i c i\ » «i_- i i u t i i t i / nr-Hu=iu i.u.nuuiur
SCAlTt'H PLOT CO.MULTOF VS C
N= 37 OUT OF 37 1 <*.COMULTOF
COMULTDF
1.H689 « «
. 3.CU.12bHW
FIGURE C-4
1.5825 *
1.^*393
1.2962
1.1530
/Z5 HR EMISSI6/0 LE\)E"L \l$.
OF
1.7257
1.009B
u o »
.B6664
.72347
.58029 +
4
0.
2.6000 S.2000
1.3000 3.9000 6.5000
7.8000 10.400 CO.125HR
4.1000 11.700
-------
As is well known to Che manufacturers, EPA will soon propose
particulate standards for heavy-duty diesel engines. Preliminary
data available to EPA from both internal testing and manufacturers'
comments indicate that the trap oxidizers expected to be used to
meet this particulate standard also act as a proportional reduction
device for gaseous emissions and would thus support a multiplica-
tive DF.
In conclusion, since The data available to EPA does not
conclusively support the use of an additive DF and since a multi-
plicative DF always provides greater air quality protection, EPA's
technical staff concludes that a multiplicative DF is the more
appropriate choice for diesel engines. This policy brings heavy-
duty diesel engines under the same type of DF as is used for
light-duty diesel vehicles.
4. Recommendations
Retain the multiplicative deterioration factor procedure as
proposed for both gasoline-fueled and diesel heavy-duty engines.
-------
D. Issue - Allowable Maintenance
1. Summary of the Issue
Included in the pending NPRM are newly-proposed provisions to
limit the amount of maintenance which can be performed on heavy-
duty durability-data engines. Emission-related maintenance must be
technologically necesasry and must have a reasonable likelihood of
being performed by owners in the field. specific minimum main-
tenance intervals are proposed which EPA has determiend to be
technologically feasible. Additionally, "emission-related main-
tenance" and "non-emission-related maintenance" are defined. These
provisions will help ensure that in-use engines do not exceed the
emission standards as a result of control technology which requires
more frequent maintenance than the users will actually perform.
2. Summary of the Comments
The most significant comments relating to allowable main-
tenance will be summarized and treated in Part I of this document:
the remainder, in Part II. The three categories into which the
major comments fall are 1) questions of EPA's justification, from
both a legal and a logical standpoint, 2) criticism of certain of
the proposed maintenance intervals, and 3) comments relating to the
four criteria for assuring "a reasonable likelihood of maintenance
being performed in-use."
Beginning with the legal issues, several commenters questioned
EPA's authority to establish "technologically feasible" intervals
for maintenance. Several commenters1 interpretations of §§207(c)
(3)(A) and 206(d) of the amended Clean Air Act (CAA) (cited in the
NPRM as the basis of the provisions) differed from the interpreta-
tion of the Agency. Mack Trucks and International Harvester (IHC)
believe the intent of the law is simply to require that the main-
tenance on certification engines is not more frequent than that
specified in the operator's maintenance instructions and to assure
that the instructions are "comprehensive and comprehensible."
A distinct legal issue forwarded by Caterpillar claimed
that EPA is in violation of the company's First Amendment rights by
requiring minimum maintenance intervals. The argument is based
on the assumption that maintenance information is a form of "com-
mercial speech" and as such cannot be limited or regulated.
Caterpillar cites legal precedents to support the assertion that
the "time, place or manner" of such communications may be regu-
lated—but not the content.
In addition to these legal criticisms, commenters questioned
the logical and factual basis of EPA's proposed revisions. First,
the claim was made that the profit-making aspect of most heavy-duty
applications has led to good in-service maintenance practices, but"
-------
no supporting data was provided. Also, EPA was criticized for not
adequately addressing the weaknesses of the current maintenance
requirements. Second, if inspection/maintenance facilities for
heavy-duty engines are established, said one commenter, they would
provide the necessary stimulus for the owners to do the proper
maintenance.
The final criticism along these lines centers around the
tendency of the extended maintenance requirements to force some
manufacturers to improve the durability of certain emission-related
components. Specifically, the claim is that market pressures have
the effect of extending Che component durabilities to the maximum
that the first-cost increases will allow. The decisions about how
much durability and required maintenance should be designed into a
component have traditionally rested in the hands of each manufac-
turer and have been based solely on economic criteria; the comments
support a continuation of this state of affairs.
A substantial volume of comment material was directed at the
more technical issue of the proposed intervals themselves. Only
four maintenance intervals were singled out as being unreasonably
long. For gasoline engines, comment concentrated on the intervals
proposed for spark plug and catalyst replacement. For diesels, the
comments addressed the turbocharger and injector maintenance
intervals. These interval-related comments, in contrast with the
comments summarized above, were often accompanied with supporting
test data.
Generally, commenters expressed their concern that the pro-
posed maintenance intervals are too long and that EPA's factual
basis for the changes is inadequate or nonexistent. Also, the
Motor and Equipment Manufacturers Association (MEMA) and Ford
presented an argument that the proposed requirements would ad-
versely affect competition among independent parts manufacturers
and dealers. More emission-related warranty repairs will be
required as a result of the extended intervals, say the commenters,
and these repairs will take place at the engine manufacturer's
repair outlets using the manufacturer's parts. In this scenario,
the business of the independents would be expected to suffer. MEMA
then suggests that, if EPA decided to follow through with the
proposed intervals, manufacturers be permitted to recommend to the
owners shorter maintenance intervals. The longer required inter-
vals would be applied during durability testing to promote low-
maintenance designs, but warranty claims would perhaps be less
frequent among the presumably better-maintained in-use population.
Moving now to the specific intervals, the proposed maintenance
requirements for gasoline engine spark plugs received considerable
comment. EPA's apparent extrapolation from LDV spark plug exper-
ience is criticized for several reasons, all relating to the
deterioration of the electrode gap.
-------
First, because of the higher N/V ratios in HDV's compared to
LDV's (i.e., HDV's are geared 1.5 to 2.5 times lower), a greater
number of ignition events occur for a given distance traveled.
Thus, on a mileage basis, spark plugs in HDE's would be expected to
deteriorate more quickly than similar plugs in LDVs. Second,
combustion temperatures are characteristically higher in heavy-duty
gasoline engines than in light-duty engines. Both of these situa-
tions will tend to erode the spark plug gap in HDE's more rapidly
than in LDV's. Additionally, IHC mentioned that oil consumption
can contribute to combustion chamber deposits and spark plug
fouling (as distinguished from gap erosion).
General Motors discussed at some length the several problems
that might accompany erosion of the gap. Primarily, there is more
probability of misfire. Faulty ignition will reduce power, worsen
fuel economy, and greatly increase HC emissions with the accom-
panying threat of catalyst overheating. GM presented data showing
catalyst bed temperatures reaching the critical range (above
1600°F, they say) in a heavy-duty vehicle with a 10% intermittent
misfire. (The catalyst specifications were not included.) Also,
gap erosion and increasing voltage requirements create greater
dielectric stresses in ignition parts (as discussed below).
GM attempted to show that the effect of a shift to unleaded
fuel on spark plug life is not as great as EPA implies. Delco Remy
measured ignition voltage requirements of spark plugs in extended
LDV service using lead-free gasoline. After 50,000 miles, service
"less severe than 30,000 miles of heavy truck operation," the
required ignition voltage had increased by 46%.
Finally, IHC attributes California's decision not to extend
heavy-duty spark plug maintenance intervals to a lack of data upon
which to base such an extension.
The comments reacting to the proposed 100,000 mile catalyst
maintenance interval were more voluminous. For the most part, the
criticism was of the limited data from which the extended interval
was derived.
General Motors presented the most complete analysis of cata-
lyst durability. Their discussion concentrated on the aspects of
heavy-duty engine operation which they felt would cause more rapid
catalyst deterioration than would be expected in light-duty usage.
These characteristics are greater fuel and oil consumption, high
exhaust gas temperatures, and the existence of high-speed closed-
throttle motoring operating modes.
When compared to LDVs, HDEs burn more gasoline' per mile
traveled and consume more oil as well (the latter effect is par-
tially due to the reduced oil viscosity at high temperatures which--
permits more leakage past the rings and valve seals). The lead in
-------
unleaded gasoline and Che phosphorus in gasoline and motor oil are
Che primary causes of chronic cacalysc poisoning. GM's analysis
estimated the ratios of HD to LD poisoning races from lead and
phosphorus. By making assumptions of the contaminant concentra-
tions in fuel and oil and of the relative poisoning effects of Pb
and Ph, GM calculated an estimated catalyst deterioration ratio for
trucks 3.42 times greater than passenger cars. Also, computer
modeled 100,000 mile deterioration factors of 5.4 for HC and 35 for
CO were projected from assumed fuel- and oil-economies for a 1979
Chevrolet 350 CID engine. Finally, projeccion of in-progress
durabilicy testing data yields a 100,000 mile deterioration
factor for HC of 3.5.
Next, GM stated that EPA's conclusion that manufacturers will
be able to solve overheating problems lacks support. They go on to
report their experience with catalyst temperature during testing.
Catalyst bed temperatures on a GM development engine have reached
1500°F to 1680°F during wide-open throttle, though CO limits were
exceeded on the transient test. Since further control will be
needed, GM anticipates that additional air injection will push the
temperatures as high as 1800°F to 1900°F due to further catalysis
of the CO. (Alumina substrates begin to experience phase changes
above 1600°F, according to GM, and such temperatures can eventually
destroy the structural integrity. Simultaneously, catalyzing
efficiency suffers.)
The last issue discussed by GM is the catalyst temperature
problem associated with high-speed closed throttle motoring. A GM
on-the-road test showed catalyst bed temperatures which exceeded
1800°F after the truck climbed a hill and began down the other
side.
Comments received from both GM and Ford bear on how the
over—temperature problems might be addressed. Ford performed an
EPA proposed transient test in which exhaust gas temperature was
measured at four locations--4, 20, 42, and 63 inches from the
exhaust manifold. Analysis of the data reveals the various times
from test start to 600°F (representative of catalyst light-off),
peak temperatures, and the distribution of time spent at various
temperatures. Temperature traces were very similar in shape but
progressively cooler the further back from the engine the thermo-
couples were. "Light-off" occurs during the cold-start test within
1 minute at 4 inches but takes more than two minutes at 63 inches.
Peak temperatures varied from 1600°F (4") to 1400°F (63"). At 4
inches the exhaust gas spent 90% of the time above 600°F as opposed
to 60% at 63 inches. Temperatures in the first three minutes were
slightly higher during the hot-start portion of the test.
Ford pointed out that catalyst peak temperatures would be
somewhat higher since additional air is added to the exhaust during
Che high-power/fuel-enrichment (high temperature) modes; cemper-:
-------
atures "well in excess of 2000°F" are expected. Substrate melt
temperature, according to Ford, is 2650°F. Ford expects there to
be a high probability of substrate melting during severe vehicle
operation. Finally, mention was made of the anticipated trade-off
between moving the catalyst and achieving light-off in a reasonable
amount of time.
General Motors suggested the use of "power-modulated air
injection systems and high-speed-overrun converter bypass sytems"
to address the anticipated catalyst temperature problems. Also,
regarding the high-speed motoring difficulties, GM reported three
attempts to develop carburetors and fuel-injection systems that
could shut the fuel supply off during engine motoring. The at-
tempts have heretofore failed because fuel that remains in the
manifold evaporates and creates jump in emissions since it is too
lean a mixture to ignite; violent backfires can also occur.
Similarly, upon restoration of the fuel flow the manifold is wetted
and combustion is intermittent for several seconds. GM states that
they have begun new efforts to develop such fuel shut-off systems
but are not hopeful that the combustion problems can be solved.
Moving now to the comments relating to diesel engine main-
tenance requirements, the intervals proposed for turbochargers and
injectors received the most attention. The thrust of the comment
was that the manufacturers should be permitted to recommend what-
ever maintenance they consider to be required on their engines.
EPA's basis for requiring the proposed cleaning, rebuilding, and/or
replacement intervals for injectors, injector tips, and catalysts
was challenged.
Mack Trucks recommended turbocharger and injector cleaning at
50,000 mile intervals and their replacement as needed "to maintain
representative performance." Thus, EPA's proposed intervals are
double and quadruple those recommended by Mack for their present-
technology equipment.
Caterpillar objected that the proposed intervals for diesel
injector tips were the same as those for gasoline injector tips.
Caterpillar went on to suggest that EPA assumed identical injector
operating environments for both gasoline and diesel engines.
The final area of comment was directed at the four satisfac-
tion criteria which were proposed to assure maintenance performance
in the field. Criteria (A), (C), and (D) were criticized for vague
or confusing language and criterion (B) for being illegal.
If the only option available to a manufacturer is criterion
(B), then it would be required to pay for the maintenance. Ford
suggests that such a requirement contradicts §207(g) of the CAA
by placing the maintenance burden on the manufacturer rather than
the owner.
IZO
-------
Ford also commenced on aspects of two other criteria (in
addition to the use of vague terminology). Criterion (C), they
say, will not be applicable to a situation where the only change in
the recommended maintenance is to adjust the interval. Also, Ford
reads criterion (D) to mean that when a signal is used to encourage
maintenance performance, the signal must be removed after survey
data has been collected. The data would be of "doubtful utility"
in such a case.
3. Analysis of the Comments
This section presents the EPA staff's discussion and anlysis
of the comments summarized above. The comments will be treated in
the same order that they appear in the Summary of Comments.
This section begins with an overview of EPA's position on allowable
maintenance in general to provide a context for the discussion.
By restricing the amount of emission related maintenance
allowable during durability testing, EPA is primarily trying
to encourage an effort on the part of the manufacturers to reduce
the amount of owner attention that their emission systems require.
This encouragement fits into the larger strategy of sustaining the
air quality benefits of regulatory actions as the vehicles/engines
are actually used. Indeed, both the U.S. General Accounting Office
and the Automobile Association of America have recently pointed to
increased light-duty vehicle emission system durability as an
approach to better in-use emission performance in those vehicles
(1, 2).
Certainly a functioning network of inspection-and-maintenance
programs would help achieve proper maintenance in the field, but
such a network does not yet exist for heavy-duty engines. Like-
wise, the providing of clear maintenance instructions to the user
will also help to some extent. Again, this in itself is not a
total solution because the nature of emission control systems is
often such that the operator is not aware that maintenance is due
or that it is necessary. Thus, manufacturers have a real opportun-
ity to help ensure in-use emission-system performance by pursuing
long-lived designs that require little attention. EPA expects that
once resources are directed toward these design goals, manufac-
turers will be able to reduce required maintenance well below
that necessary for current technology components.
The staff analysis of the comments will begin with the issue
of legal authority. Section 206(d) of the 1977 Clean Air Act
Amendments (CAA) directs that "[t]he Administrator shall by regu-
lation establish methods and procedures for making tests under this
section," (i.e., tests to determine emission compliance). It is on
the basis of Section 206 that EPA's entire certification and
durability programs have been built, as well as the Selective-
Enforcement Auditing regulations.
-------
The commenters are concerned that there is no specific Con-
gressional mandate for EPA to establish minimum technologically
feasible maintenance intervals for durability test engines.
However, the proposed maintenance requirements easily fall within
the rather broad wording of §206. (Even certification and dura-
bility testing as they appear in present regulations are not
specifically described in §206, yet they have never been succes-
sfully challenged.) The requirement for the design of a certifica-
tion program is that vehicles and engines be tested "in such a
manner as he [the Administrator] deems appropriate". The "appro-
priateness" of the proposed changes is discussed later in the
context of the "factual basis" comments.
Section 207(c)(3)(A) of the CAA requires vehicle and engine
manufacturers to provide owners with maintenance instructions which
"correspond to regulations which the Administrator shall promul-
gate." We challenge Mack's narrow interpretation that the manu-
facturer's responsibility consists solely of providing the owner
with "comprehensive and comprehensible" maintenance instructions.
The legislative history of §207(c)(3) supports a broader inter-
pretation.
Among the responsibilities of the Agency under §207(c)(3), we
believe, is to make certain that the maintenance provided to owners
is no more than that necessary to assure emission compliance. A
manufacturer should not be allowed to avoid its warranty obliga-
tions by requiring excessive maintenance that is not performed
widely in the field. This would result in the voiding of many
warranties because of a failure to properly perform the mainte-
nance, even though such maintenance was not actually necessary to
keep the vehicle or engine in compliance. Therefore, except under
special circumstances, the maintenance required of the owners to
retain their warranty should not be more than that performed during
the certification testing. The conclusion, then, is that the
maintenance instructions should be based on the maintenance done
during §206 durability testing.
The logical and factual basis for establishing technologically
feasible maintenance intervals was challenged from several direc-
tions, but little information to substantiate the claims was
provided. The staff agrees that, to some extent, better mainte-
nance habits should accompany the commercial aspects of heavy-duty
engine usage. We are not convinced that the degree of maintenance
required to maintain emission compliance is widely performed,
especially when component designs require frequent attention and
when performance of the maintenance does not improve driveability
or fuel economy.
Clearly at the lower emission levels proposed in this package
proper maintenance is a key part of an overall in-use emissions
control plan. The weakness of the present regulations is the lack
-------
of incentives for the required maintenance to actually get done.
The regulations address one facet of the problem by encouraging all
manufacturers to use the best technology components possible from a
low-maintenance requirement standpoint.
As we pointed out earlier, inspection-maintenance programs for
heavy-duty vehicles/engines can provide another important part of
an in-use control strategy. However, only a few localities have
heavy-duty I/M programs in place and the suggestion that I/M might
obviate the need for allowable maintenance seems premature. The
staff believes that widespread implementation of I/M in the future
will increase the relative importance of allowable maintenance
requirements. By clarifying the warranty-related maintenance and
improving its chance of being performed properly, these regulations
will make it easier for the owner to obtain warranty repairs.
Straightforward requirements and easy repairs are a prerequisites
to public acceptance of I/M.
The staff views the argument regarding market pressures and
component durabilities to be somewhat misdirected. If lower
maintenance in some components indeed provides a powerful competi-
tive advantage, then the market should be an important factor in
encouraging reduced maintenance and longer lasting component
designs beyond today's technology. Generally, however, we do not
believe that the market pressures for improved durability in
emission-related components is strong. (The durability of emission
controls has not been widely stressed in advertising, for example).
The staff is also concerned with the implication that manufacturers
would be willing to trade off improved maintenance characteristics
and durability (and hence, a degree of better maintenance in the
field) for commercial purposes. We cannot accept the argument of
the existence of market pressures as a rationale for allowing more
frequent maintenance than present technology has been shown to
require. Conversely we do hope that the pressures will, in the
future be a strong factor in encouraging continuing reductions in
the amount of maintenance required on emission-related components.
Several comments were directed in a general way at the pro-
posed minimum maintenance intervals, both challenging EPA's
factual basis and expressing concern about possible effects of
EPA's actions. We will address these general comments before
moving on to the technical treatment of the individual intervals.
The factual justification for the intervals proposed for spark
plugs in gasoline engines and diesel turbochargers and injectors is
sound. We present our technical rationale for these intervals
later in this section. The case of catalysts for heavy-duty
gasoline engines is unique in that the specific technology has not
been completely developed. Neither EPA nor the industry can
exactly predict the durability and the maintenance requirements of
1984 heavy-duty catalysts; a best estimate must be derived from the"
-------
information available. EPA's response to this situation should not
be, however, to provide no guidance or incentive for the design
process regarding durability. In the absence of a requirement some
manufacturers can be expected to devote less effort to durability
considerations, resulting in unnecessarily short replacement
intervals. Available data indicates that the specified interval
for catalyst replacement has a very high probability of being
achieved. Thus, we conclude that EPA is acting properly in encour-
aging the development of catalysts that will last to (or nearly to)
the useful life of the engine.
The comments of MEMA and Ford that independent parts suppliers
and manufacturers will suffer as a result of the proposed require-
ments is based, we believe, on a faulty assumption. That assump-
tion is that the number of claims for emission-warranty repairs
will necessarily increase because many owners will find that their
emission-related equipment no longer functions properly, despite
the performance of "recommended" maintenance. This assumption
implies that Ford and MEMA do not expect that improvements will be
made by manufacturers in the level of maintenance required in order
to bring them into line with the best available technology. The
purpose of the requirements is to encourage just such improve-
ments. In the event that manufacturers choose not to work toward
lowermaintenance components and the warranty claims do occur, the
manufacturer should be liable. There might be a potential for a
market shift in aftermarket parts replacement if EPA required
unrealistically long maintenance intervals. Parts would fail more,
requiring free manufacturer replacement under the emission warranty
and depriving the independents of those sales. 'However, the
maintenance intervals required here are realistic and represent a
level of technology which will be reasonably easily achieved by the
manufacturers (as we will discuss shortly). A market shift due to
an increased number of warranty claims should not result from these
regulations.
On the other hand, the very fact that maintenance intervals
are being increased will mean that parts will be replaced less
often and aftermarket parts sales may drop. We are convinced,
however, that the benefits which will be seen by the consumer in
vehicle durability and in cleaner air outweigh such an adverse
economic impact on the aftermarket industry. The Clean Air Act
certainly does not encourage manufacturers to design into their
products a high degree of maintenance; rather, it simply requires
that owners be allowed to perform all non-warranty maintenance at
establishments of their choosing.
Emission-related maintenance (as defined in Subpart A, Section
86.084-2) on engines, subsystems, or components used to determine
the deterioration of emission controls will be limited to that
which is technologically necessary. EPA has established minimum
technologically necessary intervals for a number of emission-re—
-------
lated components. This maintenance is also that which will be
recommended to the owner in the operator's manual. The manufac-
turer may recommend more frequent maintenance, as long as the
instructions for such additional maintenance are clearly differ-
entiated (in a format approved by the Administrator) from the
emission-related maintenance approved under Section 86.084-25(c).
Performance of this additional maintenance may not be made a
prerequisite to emission warranty coverage. It may be appropriate
for a manufacturer to require additional maintenance as a precon-
dition to warranty coverage of such maintenance is necessary to
offset the effect of severe and abnormal operating conditions.
These issues are a proper subject to be considered in the course of
developing performance warranty regulations under Section 207(b) of
the Act. Permitting additional "recommended" maintenance addresses
MEMA's concern that manufacturers be able to recommend maintenance
in addition to that performed during durability testing. Also, the
provisions should answer Caterpillar's concern about accurate
communications with their customers.
The issue of spark plug maintenance intervals will now be
addressed. The staff has reconsidered the analysis used in the
report "Emission-Related Maintenance Intervals for Light-Duty
Trucks and Heavy-Duty Engines". We feel that improvements in that
methodology are possible and have adjusted the proposed 30,000-mile
interval on the basis of the new analysis.
This analysis will calculate the improvement in light-duty
vehicle spark plug change intervals between 1974 and 1978 due to
the change to unleaded gasoline and then apply the percentage
improvement to 1974 heavy-duty spark plug intervals. In 1974,
domestic LDV intervals ranged between 12,000 and 15,000 miles. By
1978, the range was 22,500 to 30,000. Comparing the endpoints of
the ranges yields increases of 188% and 200%; we will use the
simple average of these increases, or a 194% increase in recom-
mended LDV plug change intervals as a result of the fuel change.
This is probably a conservative estimate since additional spark
plug longevity is likely to result from design improvements as
well.
Thus, we have constructed a basis on which to compare 1974
heavy-duty spark plug intervals (when leaded fuel was used) to
anticipated 1984 intervals (following the introduction of unleaded
fuel). By proceeding in this manner, it is possible to "cancel"
the effects of differences between LDV and HDE operating charac-
teristics and conditions, differences which drew much of the
comment. Our analysis will assume that there will be no signifi-
cant changes in N/V ratios, combustion temperatures, and oil
consumption between 1974 and 1984 heavy-duty gasoline-powered
vehicles and engines. Thus, the introduction of unleaded fuel will
be the major variable affecting spark plug life.
-------
Spark plug replacement intervals recommended for HDE's in 1974
were 12,000 miles for GM and IHC, 16,000 for Ford, and 18,000 for
Chrysler. The high Chrysler number probably relates to the lighter
vehicles which predominate their heavy-duty fleet; concentrating on
Ford, GM, and IHC, a representative interval is 14,000 miles.
Increasing this interval by the 194% we expect as a result of the
fuel change, we arrive at 27,160 miles as a projected heavy-duty
spark plug interval.
We have relied in this analysis on intervals specified for
vehicles subject to a 50,000-mile useful life. Since later in
their lives engines tend to burn more oil and hence deteriorate the
gap more quickly, we should adjust our projected replacement
interval to allow more frequent maintenance over a 114,000-mile
useful life. To provide this additional cushion, we have rounded
the interval down to 25,000 miles, which is very appropriate ~for
heavy-duty engines employing unleaded gasoline.
As the preceding discussion suggested, a range of comments are
addressed to some degree by the revised analysis. Certainly the
higher N/V ratios and cylinder temperatures of HDEs might be
expected to have a greater effect on HDE spark gap erosion as
compared to that in LDVs. Yet, 1974 replacement intervals for
spark plugs were in the same range for both HDEs and LDVs. This
fact implies that the engine speed and combustion temperatures are
not major contributors to gap deterioration, at least, during
12,000 to 15,000 miles of use with leaded gas. No comments sug-
gested that after the introduction of unleaded fuel the relatively
harsh conditions seen by heavy—duty plugs would have a greater
effect on durability relative to light-duty plugs than before the
fuel change. Therefore, there seems to be validity to the assump-
tion in the proceeding analysis that both HDE and LDV spark
plug intervals will experience similar relative improvements when
the shift to unleaded fuel occurs.
Regarding the mention of oil comsumption as a contributor of
the fouling of spark plugs, the staff see the use of "hot" and
"cold" spark plugs as a means of addressing this problem, as well
as the electrode erosion problem. For certain applications,
manufacturers will certainly be able to recommend plugs using
different alloys to either heat up quickly and burn off oil and
carbon residues (from low-speed/idle operation) or stay cooler to
reduce eroding of the gap (for high-load applications).
Next, while we understand that the effects of misfiring spark
plugs on catalysts can be catastrophic, we do not agree that a
large increase in the occurrence of misfiring due to plug deterio-
ration will accompany these regulations. The technology for a
25,000-mile spark plug exists and hence- gross misfiring should not
be an issue.
lid
-------
The Delco Remy LDV data shows that even with unleaded fuel
spark plug voltage requirements increased significantly after
50,000 miles (equivalent to 20,000 to 33,000 miles of truck opera-
tion from the standpoint of ignition events, according to GM). The
data indicates that gap erosion will indeed occur in extended spark
plug service even after the introduction of lead-free fuel into the
heavy-duty gasoline fleet. The severity of the problem for heavy-
duty is not as clear. Heavy-duty ignition systems often operate at
higher voltages than do light-duty systems. Additionally, GM does
not seem to imply that higher voltage requirements in the 50% range
toward the end of a spark plug's life will necessarily be accompa-
nied with frequent misfire. The staff interprets the data pri-
marily to indicate a possible effect from dielectric wear on the
ignition parts resulting from higher later-life voltage require-
ments or higher voltage systems. GM (and other commenters)
did not provide any information to indicate that such problems
would be difficult to overcome, if indeed they occur in heavy-
duty engines.
The staff sees no reason to alter its previous analysis
showing a 25,000-mile spark plug to be reasonable for 1984 heavy-
duty gasoline engines.
Finally, a conversation with a representative of the Califor-
nia Air Resourses Board* seems to refute IHC's theory regarding the
reasoning behind CARS's decision not to extend heavy-duty spark
plug maintenance intervals. Their actions were not aimed at
heavy-duty engines at all, and thus they had no reason to investi-
gate heavy-duty spark plug life. It was not a lack of data that
led to a continuation of the current heavy-duty maintenance inter-
val but rather a complete lack of effort in that direction.
We now turn to an analysis of the comments relating to the
100,000-mile catalyst replacement interval proposed in the NPRM.
The comments generally took issue with EPA's extrapolation of
light-duty catalyst technology to heavy-duty applications.
General Motors presented a useful methodology for estimating
the relative effects of lead and phosphorus catalyst poisioning in
LDVs vs. HDEs. However, the staff disagrees with several numerical
values which GM used in the analysis. First, they compared 100,000
miles of heavy-duty service to only 50,000 miles of light-duty
service. Second, with regard to fuel economy, the staff believes
that the analysis will be improved by the use of different miles-
per-gallon numbers. GM uses a predicted 1983 fleet average light-
duty value of 25 mpg. , But since our interest here is to judge the
difficulty in applying catalyst technology to HDE's it is more
reasonable to observe a "worst case" light-duty application (i.e.,
one that would come closest to exposing a catalyst to heavy-duty
Bob Weiss, Certification Department, August 7, 1974.
-------
type treatment). We looked at data from several 1979 Cadillac
certification vehicles with large, catalyst-equipped engines
(425-CID) and found an average fuel economy of 13 mpg. This
is the value we use in our analysis. Similarly, while we believe
that 6.5 mpg is a better estimate of the average heavy-duty gaso-
line engine fuel economy, GM's 5 mpg figure will be used for our
worst case analysis. The following table compares GM's numbers,
adjusted for 100,000 miles of light-duty operation, with EPA's.
100,000 Mile
Passenger Car
EPA
13
7,692
38.46
.77
2,500
40
45.4
38.5
46.2
GM
25
4,000
20
0.4
2,500
40
45.4
20
45.8
100,000 Miles
Heavy-duty Truck
GM and EPA
84.7
65.8
Fuel Economy (MPG)
Gallons of fuel
Grams Pb (.005 gr/gal)
Grams Fuel Ph (.001 gr/gal)
Oil Economy (mi/qt)
Quarts of Oil
Grams Oil Ph (.16 wt%)
Total Pb - grams
Total Ph - grams
Total Pb & Ph Contaminants
5
20,000
100
2
1,250
80
90.8
100
92.8
192.8
Applying GM's estimated phosphorus-to-lead poisoning ratio of
9:1, one arrives at an expected rate of poisoning 2.16 times
greater for heavy-duty than for light-duty (using 25 mpg) or 2.06
(using EPA'S 13 mpg). Then, making the adjustment for catalyst
sizing, the factor for GM becomes 1.56 and for EPA becomes 1.49.
Our analysis indicates that GM's comparison of 50,000 miles of LDV
service to 100,000 miles of HDE service to arrive at a poisoning
ratio of 3.42 somewhat exaggerated the poisoning problem. The
effect of changing the LDV fuel economy in the analysis is rela-
tively small.
GM went on to multiply their HD:LD poisoning ratio by an
average 1980 light-duty deterioration factor (DF) to arrive at
a high predicted heavy-duty DF. The EPA staff finds two aspects of
this final step to be questionable. First, since 1979 average LDV
DFs are around 1.2 for both HC and CO, GM's "average" of 1.35 seems
high. This may be due to the introduction more three-way catalysts
for 1980 or some other hidden factor; but since all but a handful-
-------
of 1979 LDV's are equipped with oxidation catalysts we believe
their DFs provide an adequate baseline. Second, the expected
effect of poisoning on a catalyst is to affect the conversion
efficiency. GM applied the poisoning ratio directly to the LDV DF,
again exaggerating the results in their favor. The correct way to
make this calculation is as follows, using the DF of 1.2 and the
poisoning ratio of 1.49 from above.
A DF of 1.2 implies that, over the 50,000 miles of LDV extra-
polation, emissions will deteriorate 20 percent. Assuming that
continued deterioration to 100,000 miles is linear (although it
usually levels off with time), one would expect that twice the
deterioration, or 40 percent would occur- Beginning with a
100,000-mile LDV DF of 1.4, then, one can apply the HD-to-LD
poisoning rate ratio to the percent deterioration. Thus, an
expected heavy-duty DF of (1 + [(.4)(1.49)]) = 1.60. That is, the
percent deterioration of a light-duty catalyst is increased by
nearly 50 percent due to the increased poisoning expected in
heavy-duty applications.
Since the DF calculated above only takes poisoning into
account, we should make some adjustment for the loss of conversion
efficiency occuring as a result of higher average temperatures
to be expected in heavy—duty catalysts. Temperature excursions
beyond 1800°F can begin to cause a phase change in alumina sub-
strates or monolith washcoats, reducing their large surface areas
with the resulting loss of active catalyst sites and hence conver-
sion efficiency. The correlation between temperature exposure and
loss of efficiency has never been established to our knowledge; a
GM representative said, however, that they assume an effect on
emission deterioration due to heat exposure of the same magnitude
as the effect due to poisoning. With the higher noble metal
loadings necessary for heavy—duty catalysts, more active sites are
available and loss of some is not so crucial as in light-duty
catalysts. This point, coupled with the catalyst cooling measures
that we expect to be used for heavy-duty (these measures are
discussed later in this section), lead the staff to the conclusion
that a heat-related deterioration effect which is half the poi-
soning related effect is reasonable. We will adjust our heavy-duty
DF calculation by using a combined poisoning-plus-heating ratio of
1 + [.49 + 1/2 (.49)] = 1.74. Then the revised heavy-duty DF to be
expected with catalyst-equipped engines is (1 + [(.4)(1.74)]) or
1.70. It is clear that the difficulties which GM anticipated in
achieving low enough 4,000-mile emission values are not so great as
their high DFs had implied.
GM also submitted the results of a computer modeling program
which also predicted very high rates of emission deterioration
for heavy-duty catalyst-equipped engines. We suspect that the
complex model may have incorporated some of the poor assumptions
discussed above. In a conversation with an author of the model,
-------
we learned that only the poisoning mechanism was simulated with
any real basis of information; heat effects were assumed. No
post-50,000 mile data, of course, was available to include in the
model, resulting in more judgements. The number of opportunities
for error to be introduced into the model through both the inputs
and the operations themselves is phenomenal; the staff has little
faith in the results, especially when they vary so much from our
reasoned analysis above.
The only durability data on heavy-duty catalyst-equipped
engines that was provided appeared in GM's submission. On the
basis of five test points ranging from 0 to 500 hours of dynamo-
meter operation (equivalent to 0 to 15,000 miles), GM extrapolated
a high HC DF. This data is too limited to be conclusive and is
drawn from a current-technology catalyst system. GM's linear
extrapolation directly contradicts their comment (regarding dura-
bility testing) that catalyst deterioration is not linear. And the
expected improvements we expect in catalyst durability further
reduce the strength of GM's conclusions.
We have discussed above the expected long-term effects of
chronic poisoning and infrequent temperature excursions on catalyst
efficiency. We wish to pursue now in more detail the problem of
short-term high temperature transients—their effects on catalyst
structure and methods of avoiding their occurence. GM and Ford
provided limited data on various aspects of this issue, and tests
recently performed by EPA provide a further base of information.
In the next paragraphs, we will describe the EPA tests, analyze
data resulting from both these and industry test programs, and draw
conclusions about the threat of catalyst overheating. The discus-
sion begins with a presentation of the staff position on the
effects of high temperatures on catalysts.
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 unaccessible due to the loss of
porosity; this process effectively reduces the number of sites
available for catalysis and hence lowers the efficiency of conver-
sion. 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.
The published "safe operating temperature" for gamma alumina
substrates is approximately 1700"F (which contrasts with GM's
"critical temperature" of 1600°F). Gamma-to-alpha phase change can
be expected to occur between 1750°F and the alumina melt tempera-
-------
ture of around 2900"F. The staff will assume that below 1700°F, no
change in the structure of gamma alumina takes place.
The staff is unaware of any research which would indicate how
substrate structure and loss of efficiency are affected by time and
temperature. (A representative of Englehard Industries Division
reported in a telephone conversation that when they have exposed a
catalyst to 1800'F in a steady-state bench test, efficiency losses
of 40-602 have been observed after twenty-four hours of exposure).
It is known, however, that heavier noble-metal loadings, like those
expected in heavy-duty catalysts, bolster the durability of the
conversion efficiency. This is because more active sites are
available initially, allowing a cushion if some are lost through
agglomeration or reduced substrate surface area (or even poison-
ing) .
We expect that any temperature excursion beyond 1700 or 1750°F
will probably cause a limited amount of the alumina to change
phase. Yet for a well-loaded catalyst (in the range of 40 g/ft ),
temperature excursions between 1750 and 2000°F lasting less than a
minute should not cause major losses in conversion efficiency.
However, because of the cumulative nature of the effects, the
frequency of such events would have to be minimized, an issue which
is addressed in detail below in the context of catalyst heat
reduction. (The staff consulted Dr. Ray Ober of Englehard Indust-
ries Division and References 3, 4 and 5 in the formulation of the
foregoing position).
EPA has recently conducted a test program which investigated
catalyst temperatures during several types of engine operation. A
GM 454-CID "tall block" heavy-duty engine with dual exhausts was
equipped with two 260-CID GM catalysts. While the total volume if
the two catalysts exceeded the displacement of the engine, the
noble metal loading was only 10 g/ft in each catalyst (compared to
40 g/ft expected by the staff for heavy-duty catalysts).
The primary purpose of the tests was to observe the sensiti-
vity of maximum catalyst temperatures to the distance between the
exhaust manifold and the catalyst. Distances of 38, 68 and 116
inches were tried (nominally 3, 6 and 9 feet). For each catalyst
position, we operated the engine over the proposed transient test,
through a series of high-power, high-speed steady-state modes, and
finally under closed throttle motoring conditions. We immediately
followed each steady-state mode with a motoring mode, the motoring
being done at the same speed as the preceding steady-state portion
(the exception was that if the catalyst tempertures did not sta-
bilze below 1700°F during the steady-state run, we omitted the
motoring). Catalyst bed temperatures were recorded continuously.
The maximum catalyst bed temperatures reached during the
transient portion of the EPA tests are presented below for each of
-------
the three catalyst positions. Hot-start-segment temperatures
were not available for two of the eight tests, as indicated by
dashes. Also, the estimated catalyst light-off times, taken
as the time from the beginning of the test to the time at which the
catalyst temperature reaches 575°F, appear in the table.
Transient Test Results
38" 68" 113"
Distance from Cold Hot Cold Hot Cold Hot
Exhaust Manifold Start Start Start Start Start Start
Maximum Catalyst 1592 1586 1519 1534 1455
Temp. (°F) 1597 1597 1539 1511 1451 1451
1537 1534 1531
Light-off Time (sec)
1st Catalyst 140 196 285
112 220 256
164 181
2nd Catalyst 220 256
192 315 300
238 231 256
As the catalysts were placed further from the exhaust mani-
fold, a progressive drop in maximum temperatures occurred. The
magnitude of the effect was approximately 70°F for each additional
distance of 3 feet (A possible explanation for the lower maximum
temperatures seen in the third test at 38" is offered later in this
discussion). This pattern compares with Ford's exhaust temperature
investigation, which showed that three feet of travel in the
exhaust pipe cooled the exhaust flow over 100° F. It follows,
then, that while much of the heating of the catalyst is a result of
the exothermic oxidation reactions, a significant portion of the
heating results from the exhaust gas and hence is sensitive to
catalyst placement.
The time required for the catalyst to reach the 575°F "light-
off" temperature was increased each time the catalysts were placed
further back from the manifold. We calculated the average light-
off time for the catalyst pair during each test and then averaged
all of these single test values for each catalyst distance.
The effect of each three-foot shift of the catalyst was shown
to be a loss of 40 to 50 seconds in light-off time. Again, compar-
ing Ford's data, the time for the exhaust gas to reach the 575°F
catalyst threshold temperature varied by about 70 seconds over
three feet of distance; this observation supports the EPA data
regarding the effect of catalyst position on light-off time.
-------
Turning Co the steady-state runs which preceded the motoring
tests, stabilized catalyst temperatures appear below. A tempera-
ture was defined to be "stabilized" when it was not changing by
more than 1°F every 5 seconds. For reference, rated horsepower
for the test engine was 225 hp.
Stabilized Steady-State Catalyst Temperatures (°F)
Catalyst Distance 38" 68" 113"
RPM Torque HP
1500 330 94 1440
2500 205 98 1335 1243 * 1172*
3000 200 114 1318 * 1247*
2500 255 121 1317 *
2500 275 131 1347*
3500 200 133 1480
3000 250 143 1500 1385 * 1317*
3300 250 157 1438 * 1410*
3500 250 167 1530
3000 300 171 ** 1525
3500 300 200 ** *** 1575
3690 320 225 *** *** ***
* Average of both catalyst temperatures.
** Temperature did not stabilize below 1600°F.
*** Staff assumed that temperature would not stabilize below
Despite the fact that each speed-torque combination was
not attempted for each catalyst position, it is possible to see a
trend in these data. It appears that, in general, the stabilized
catalyst temperature at a given speed-torque setting becomes lower
as the catalyst is moved back. This is most obvious in the in-
stance of the higher power modes during which the temperature
stabilized only in the 113" position. Thus, even during high-
powered operation of extended duration, the effects of catalyst
placement on catalyst temperature may be observed.
For several of the steady-state modes that reached a stable
temperature, we closed the throttle and let the dynamometer
drive the engine at the same speed. At the 38" catalyst distance,
the motoring was unfortunately continued for only 1.5 minutes.
However, at the 68" and 113" position, the engine was usually
motored until the temperature peaked and began to fall. Figures
D-l plot the temperature profiles, beginning with the stabilized
temperatures from the steady-state modes, for the three catalyst
positions. Plotted values represent the average temperatures for
the two catalyst for the 68" and 113" graphs. For 38", only one.
catalyst is represented.
-------
2000.
1800.-
Li_
CD
1600.-
LU
C£
1400.
UJ
Q_
z:
1200.
1000.
0.0
Figure D-l(a)
0.5
CATALYST
POSITION-38 IN
1 - 3000 RPM
2 - 3500 RPM
3 - 2500 RPM
1.0 1.5 2.0
TIME CMINUTESD
2.5
3.0
-------
2000.7
1800.-
u_
(D
Q
*—/
UJ
ID
UJ
Q_
1600.-
1400.
1200.
Figure D-l(b)
CATALYST
POSITION-68 IN
3300 RPM
3000 RPM
3000 RPM
2500 RPM
2500 RPM
(3
1000.
0.0
0.5
1.0 1.5 2.0
TIME CMINUTESD
2.5
3.0
-------
2000.
1800.-
U_
LU 1600.
LU
C£
ZD
LU
O_
21
LU
1400.-
1200.;
1000.
Figure D-l(c)
CATALYST
POSITION-113 IN
3300 RPM
2500 RPM
3000 RPM
3000 RPM
2500 RPM
1.0 1.5 2.0
TIME CMINUTESD
-------
The striking rise in temperature during the one motoring test
at 68" which began at 1440°F was due to the stopping of the motor-
ing in response to high temperatures. We assume that the initial
burst of fuel when the throttle opened was not combusted and caused
the catalyst temperature (primarily of one catalyst) to begin to
rise very quickly. Thus the rapid rise did not actually occur
during motoring as the figure suggests. In the way of further
explanation of the data, the steady-state runs which resulted in
stabilized temperatures greater than 1500°F were not followed by a
motoring test. Finally, data at the 30 second point of motoring
was taken only at the 113" position.
The general pattern illustrated in Figures D-l is that the
catalyst temperature rises rapidly at first, then peaks and
begins to drop. Since the motoring speed and throughput of
raw fuel should be constant, it seems that a short-lived event
occurs soon after the closing of the throttle that pushes the
temperature up for a minute or two. That event may be the result
of the remaining mixture in the intake manifold burning poorly and
subjecting the catalyst to a burst of hydrocrabons. This partic-
ular catalyst-heating situation would not persist because the
remnant manifold mixture would be followed by a still leaner
mixture from the motoring process itself. Thus an immediate but
temporary rise in catalyst temperature would be expected. A
conculsion of this reasoning is that it is probably not motoring
itself but the transition to motoring that threatens the catalyst
with overheating, even though the catalyst sees some raw fuel
during motoring.
The effect of catalyst placement is less perceptible in the
motoring data than in the transient and steady-state testing.
There seems to be a pattern within the data at the 68" and 113"
positions that indicates that higher stabilized starting tempera-
tures result in greater temperature rises during motoring. So, to
the extent that placing the catalyst further back reduces these
starting temperatures, the motoring heat rise can perhaps be
addressed simultaneously. Also, the data appear to indicate that
more distant catalyst placement leads to a quicker arrival at the
maximum temperature; we can presently offer no explanation for this
pattern.
A final pattern that is discernable from the data is that
lower speeds result in lower maximum temperatures. Reduced
HC throughput is probably the explanation for this phenomenon.
Because of the various torque levels that preceded the motoring
runs, however, motoring speed and maximum temperature are not
directly comparable.
An insight into the effects of heat spikes on the efficiency
of a catalyst with a light noble metal loading is possible from the
motoring investigation. An inadvertant excursion above 1800°F for
If?
-------
cases peaks and reverses after 1 to 3 minutes. The magnitude
of the temperature increase is such that if the motoring is
begun while the catalyst is sufficiently hot, the substrate or
washcoat can reach damaging temperatures (1750°F and above).
On the basis of the EPA and industry data discussed above
and within the context of the previously described staff position,
we are prepared to make several conclusions about the catalyst
heating issue. We will separate from the discussion for the moment
the special problems associated with closed throttle motoring.
The staff is convinced that several straightforward design
approaches exist which individually or in combination can greatly
reduce the threat of catalyst damage through overheating. The
first approach is through catalyst placement, taking advantage of
the relative freedom in catalyst positioning afforded by heavy-duty
vehicles as compared to light-duty vehicles. Because of their size
and construction, heavy-duty chassis usually make it possible to
situate a catalyst rather far from the engine.
Similar approaches which could also reduce catalyst tempera-
tures take advantage of heavy-duty chassis as well. Cooling fins
on the catalyst and the exhaust pipe would improve heat transfer
from the converter. Reducing or removing the insulation familiar
to light-duty catalyst containers (and present on the catalyst used
in the EPA tests) would facilitate further heat transfer. Accord-
ing to a representative of Engelhard Industries Division, removal
of insulation can affect catalyst temperatures by as much as
100-200" F. The less space-restrictive and temperature sensitive
characteristics of heavy-duty chassis (relative to light-duty)
which reduce the need for insulation would also allow the use of
screens, cages, or similar types of open shielding if such protec-
tion is necessary. Still another approach which might be useful in
some applications is the installation of wind deflecters to
improve the air flow across the catalyst during highway operation.
When such measures are taken to cool the catalyst, light-off
time becomes more of an issue. More efficient removal of heat
from the catalyst can be expected to compound the effects of
catlayst positioning, which was demonstrated above to significantly
increase light-off time. The tradeoff which Ford pointed to
between lower catalyst temperatures and light-off clearly exists.
The magnitude of this tradeoff problem is somewhat exaggerated
by the commenters. The cold-start/hot-start weighting applied to
the results of the proposed transient test tend to minimize the
impact of emissions during the early part of the cold start segment
(The reader is urged to consult the Feasibility of Compliance
chapter of this document for an in-depth discussion of this topic).
We wish not to encourage manufacturers to forfeit cold-start
emission control but rather to point out that some increase in
-------
light-off time should not be a determining factor in achieving a
given design-goal emission level.
In addition, engine-out cold-start emissions can be reduced
directly, diminishing the need for early conversion efficiency.
The design of choke operation and carburetor flow characteristics
to minimize cold-start HC and CO are two examples. A final approach
is to use monolithic catalysts, which, because of their lower mass,
heat up and light off more quickly than do pelleted designs.
The obvious exception that we have made in the discussion
thus far is the motoring mode. While the catalyst-cooling approa-
ches suggested above should greatly reduce the frequency of occur-
ance of high-temperature excursions, the staff believes that their
elimination is not likely. Because a relatively small number of
such events will begin to degrade the structure and efficiency of
current alumina substrates, we believe that some method of avoiding
the temperature spikes will be necessary. A mechanism to shut off
the fuel flow during closed-throttle motoring and an air-pump shut
off to stop the oxidation process in the catalyst (or both) are
suggested as ways of protecting the catalyst. (See the Staff
Recommendations).
An underlying assumption of the entire catalyst heating
discussion above is that gamma alumina will continue to be the only
catalyst support material. In the event the new potential market
for heavy-duty catalysts spawns the development of more tempera-
ture-resistant substrates, the conclusion of our analysis would
change greatly. Conversations with Engelhard Industries repre-
sentatives have indicated that such substrates are being developed
and may well be available by 1984.
Substrates which could withstand greater temperatures without
losing their surface area would reduce the need for distant cata-
lyst placement and extensive cooling measures, making light-off
less of a problem. Additionally, the temperature spikes associated
with motoring would possibly no longer threaten the catalyst,
eliminating the need for special protection during this mode.
Even if alumina remains as the primary substrate material, we
conclude the technology exists for heavy-duty catalyst systems that
will function for 100,000 miles. Deterioration of catalyst effi-
ciency may be slightly more rapid than that seen in current light-
duty systems, but not to such a degree that the feasibility is
compromised. Additionally, since the deterioration curves of
catalyst systems generally flatten out as time goes on, the signi-
ficant loss in efficiency is expected to occur in the first half of
the catalyst life. If catastrophic failure from gross heat effects
or intentional poisoning is avoided, continued functioning beyond
100,000 miles is very possible, even to the estimated 114.,000
miles average heavy-duty gasoline engine useful life.
-------
several seconds during a motoring experiment seems to have resulted
in a loss in efficiency. The experiment fell between the second
and third transient tests at the 38" catalyst positions; on the
basis of a comparison of the emission values from those two tests,
a decrease in conversion efficiency of about 15% was observed.
This loss may explain the lower maximum catalyst temperatures seen
in the transient test following the temperature excursion.
The very light noble metal loading of these catalysts is
probably most responsible for the loss of efficiency. Because
of the relatively small number of active catalyst sites available,
a small loss in subtrate surface area would be expected to appre-
ciably reduce the conversion efficiency. The heavier loading of
heavy-duty catalysts will improve this situation.
It is important to note that the value of the entire EPA
temperature analysis is compromised to some degree by the inconsis-
tant treatment of the dual catalyst system. The temperature
characteristics of the two converters were appreciably different,
but they were not always treated independently. For example, the
transient test maximum catalyst temperatures as recorded represent
only one catalyst, the one that was the hottest. Similarly, the
temperatures during the steady-state and motoring study sometimes
correspond to only one catalyst.
Although we assumed that the GM catalysts were equal in
mass and in flow specifications, it appears that they differed
to some degree; the temperatures sometimes differed by 100°F.
Because the transient emission tests combined the exhaust flows for
analysis, it is not possible to separate the emission components
contributed by the individual banks of cylinders under the influ-
ence of the individual catalysts. Therefore, for instance, since
one of the catalysts may for some reason have been more sensitive
to heat effects, it alone may have caused the efficiency loss.
Detailed conclusions drawn from the temperature analysis study,
then, are of limited value.
However, the EPA data reveals several general trends which
lead to the following conclusions:
1) Catalyst placement has a marked effect on the maximum
converter temperatures reached during operation over the proposed
transient cycle and in high-speed steady-state modes. Temperatures
during motoring can possibly be controlled by lowering the initial
temperatures through catalyst placement.
2) Catalyst light-off times progressively increase with
catalyst distance from the exhaust manifold.
3) The transition to closed-throttle motoring is accompanied
by an immediate catalyst temperature rise which in at least some-
-------
The final area of comment regarding maintenance intervals is
specific to diesel engines. (General comments not directed at
specific intervals were treated previously.) Looking first at
turbochargers, it is important to note that the introduction of
crankcase emissions into the turbocharger inlet is no longer being
considered by EPA. If EPA decides to require control of crankcase
emissions from turbocharged diesels, a method will probably be
recommended which allows the turbocharger to be bypassed.
Mack Truck's comments recommend turbocharger maintenance at
50,000 miles, their current practice. However, they submitted no
information to indicate what differences exist between theirs and
Caterpillar's 200,000-mile-interval turbocharger that would explain
Mack's requirement of more frequent maintenance. The technology
obviously exists for a more maintenance-free turbocharger. Mack is
either presently using a design which requires more attention than
technology would dictate or they are simply recommending more
maintenance than their turbocharger requires. In either case
Mack's objection is based not on technology but on a concern about
EPA's justification for establishing required minimum intervals.
The issue of justification was treated earlier in this section.
EPA's proposed intervals for cleaning diesel injector tips
were not based on an assumption of similar operating environments
for both gasoline and diesel injectors. The basis was simply the
observation that 100,000-mile and greater cleaning intervals are
recommended now on some engines. In the absence of any submitted
information that would indicate that this low-maintenance tech-
nology is inappropriate to other heavy-duty diesels, we reaffirm
the feasibility of the proposed intervals. This concludes the
staff analysis of the interval-specific comments.
We recommend that EPA delay the requirment that manufacturers
must demonstrate "a reasonable likelihood" that proper maintenance
will be performed in-use. Our recommendation arises not from
specific comments about these proposed provisions but from a belief
that such a requirement is not necessary at this time. It appears
to us that the manufacturers would reasonably easily be able to
show that required maintenance was indeed being performed on the
emission-related components which these regulations will require.
With respect to the forthcoming NOx regulations, however, the
situation is different. It is possible that three-way catalyst
technology will be used, in which case oxygen sensors will control
the feedback systems. It is for this type of component that the
staff believes some sort of assaurance of in-use maintenance will
be necessary.
At such a time that these provisions are reproposed, EPA will
analyze the comments received with this package as well as any new
comments.
-------
4. Recommendations
The staff has concluded that the proposed maintenance require-
ments (Section 86.083-25 of Subpart A) should be retained in their
proposed form with the following exceptions:
1. The technologically necessary spark plug change interval
should be revised from 30,000 miles to 25,000 miles.
2. The technologically necessary catalyst replacement
interval should remain 100,000 miles. The preamble of the final
rule, however, should make it clear that for gamma alumina catalyst
substrates, EPA expects that an air pump shutoff capability and/or
a motoring mode fuel shutoff will be necessary for motoring condi-
tions exceeding 15 seconds in duration in order to protect the
catalyst. We further recommend that such a system be specifically
exempted from being classified as a defeat device.
-------
References
_!_/ "Better Enforcement of Car Emission Standards — Away to
Improve Air Quality", Report by the Comptroller General of the
U.S. General Accounting Office, Report IICED-78-180, January
23, 1979.
2j American Automobile Association letter to Rep. Henry Waxman,
August 13, 1979 (EPA Central Docket Section #OMSAPC-78-4).
_3_/ "Comparison of Catalyst Substrates for Catalytic Converter
Systems", J.L. Harned and D.L. Montgomery, General Motors
Corporation, SAE Paper #730561.
4/ "Active Aluminas as Catalyst Supports for Treatment of Automo-
tive Exhaust Emissions", Harry E. Osment, Kaiser Aluminum and
Chemical Corp., SAE Paper #730276.
_5_/ "Ceramic Substrates Technology for Automotive Catalysts,"
~ Maxwell Teague, Chrysler Corp., SAE Paper #760310.
-------
E. Issue - Parameter Adjustment
1. Summary of the Issue
Briefly stated, the issue is this; Does the available evi-
dence justify the proposed regulation of heavy-duty parameter
adjustment.
In the NPRM, EPA proposed to amend the certification process
to permit the Administrator to adjust previously identified engine
parameters (i.e., idle fuel-air mixture, idle speed, initial spark
timing, and choke valve action) to settings anywhere within the
physical limits of adjustment for the parameter(s) in question.
The proposed requirements were identical to those recently promu-
lgated for light-duty vehicles and will encourage manufacturers to
design heavy-duty engines which are less susceptible to in-use
maladjustment. Such maladjustment is capable of causing in-use
emissions to be substantially higher than allowed fay the standards.
2. Summary of the Comments
Many commenters expressed their concern that EPA had no data
which showed that heavy-duty vehicles are being maladjusted in
the field. They point out that EPA states in the preamble to
the NPRM that "EPA does not have test data which indicate how
serious (maladjustment) may become for heavy-duty vehicles with the
use of advanced emission control technology. However, there is no
reason to believe that the degree of heavy-duty in-use maladjust-
ment will be much different than has been the case with catalyst
equipped light-duty vehicles and light-duty trucks."
The commenters contend that since EPA admittedly has no data
to support heavy-duty parameter adjustment, EPA is using the
"crystal ball" approach which was struck down in court in Inter-
national Harvester vs. Ruckelshaus.
Several commenters cited an Oregon Department of Environ-
mental Quality inspection and maintenance program report which
included 4600 gasoline powered trucks, all weighing more than 8500
Ibs GVW. That report includes a paragraph stating that heavy-duty
vehicles have less of a problem with the inspection and mainte-
nance program than do light-duty vehicles. The report goes
on to hypothesize that since heavy-duty vehicles are working
(i.e., commercial) vehicles they probably receive better overall
maintenance than general passenger vehicles.
Many commenters expressed their belief that since heavy-duty
vehicles are primarily commercial vehicles, they are well maintain-
ed by professional mechanics and. therefore, maladjustment should
not be a problem. However, no commenter provided any data ta
support this belief.
-------
Diesel engine manufacturers stated that even if the available
light-duty data on parameter maladjustment could be extrapolated to
heavy-duty vehicles, resultant conclusions would have to be limited
to gasoline-fueled heavy-duty vehicles because the light-duty data
did not include any diesel engines. One commenter cited an EPA
contract report^!/ which involved the testing of 12 used heavy-duty
diesel engines. He claimed that the "data certainly indicates
little if any, deterioration or tampering with the emission
control systems considering the normal possible spread in emission
levels".
Diesel manufacturers also stated that since diesel engines
are used in a wide variety of applications they require wide ranges
of parameter adjustment. Specific examples of parameters which
need a wide range of adjustment or the width of such ranges of
adjustments were not given.
Some comments other than those summarized above were received.
These additional comments are considered secondary in nature and
are treated in Part II of this document.
3. Analysis of the Comments
The major issue raised by cotmnenters was a lack of supporting
data used by EPA to justify the need for parameter adjustment
regulations for heavy-duty trucks. EPA's technical staff believes
that available evidence is sufficient to demonstrate the need for
heavy-duty parameter adjustment. That evidence is discussed in
this section.
Light-duty in-use maladjustment studies^/ .3_/ ,4/ ,5_/ have shown
conclusively that parameter maladjustment is the source of signifi-
cant in-use emissions which are above standard. Heavy-duty ve-
hicles have the same engine parameters which perform the same
functions as those paramete-rs found to be most commonly maladjusted
for light-duty vehicles. For example, both light-duty and heavy-
duty gasoline-fueled engines have adjusting mechanisms on the
carburetor which control the idle air-fuel mixture and the idle
speed. Also, both heavy-duty and light-duty gasoline-fueled
engines have chokes to facilitate cold start driveability and both
have spark plugs whose firing must be properly timed. Maladjust-
ment of these various parameters can arise from simple failure of
the operator to have proper maintenance performed, improperly
trained mechanics, or in some cases deliberate maladjustment in an
attempt to improve performance (at the expense of emissions). All
of these causes would exist for heavy-duty vehicles as well as for
light-duty vehicles. Heavy-duty inspection and maintenance studies
done in Oregon and New Jersey show that heavy-duty vehicles fail at
rates essentially identical to those found in light-duty I/M
programs. The light-duty I/M failures are primarily due to par--
ameter maladjustment and there is no reason to believe that the
-------
majority of heavy-duty I/M failures are due to anything but par-
ameter maladjustment.
EPA's Restorative Maintenance Program2/ showed that parameter
maladjustment is a significant problem with light-duty vehicles
less than 12 months old. Idle mixture maladjustment was found on
37.7% of the vehicles tested. Results also showed idle speed to be
maladjusted 25% of the time, choke to be maladjusted 10.4% of the
time and initial spark timing to be maladjusted 19.0% of the
time. Table E-l summarizes this study and shows the significant
increases of in-use emissions due to the various maladjustments.
Other studies, 3/ ,4_/ ,5/ have given similar results with the ad-
ditional conclusion that the older vehicles become the more likely
they are to be maladjusted. It is clear from these studies that
parameter maladjustment is occuring to a wide extent on light-duty
vehicles and the resultant increase in emissions is substantial.
Again, heavy-duty gasoline-fueled vehicles have the same
engine parameters that were found most likely to be maladjusted on
light-duty vehicles. These engine parameters serve the same func-
tion whether on light-duty vehicles or heavy-duty vehicles. It is
reasonable to expect that heavy-duty maladjustment of these para-
meters will be similar to the light-duty experience.
However, several commenters expressed the belief that since
heavy-duty vehicles are commercial vehicles they are better main-
tained than light-duty vehicles. They stated that better mainte-
nance should mean that parameter maladjustment is not as big a
problem for heavy-duty vehicles. No direct data was presented to
substantiate this rationale. However, an Oregon Department of
Environmenal Quality report6_/ was cited as evidence. In that
report 4600 heavy-duty gasoline vehicles had been tested as of
February 1978 as part of the Oregon Inspection and Maintenance
program. The preliminary conclusion reached was that heavy-duty
vehicles were having less of a problem with I/M than light-duty.
This was hypothesized to be due to the commercial nature of heavy-
duty vehicles and the predicted better maintenance these vehicles
might be receiving.
An examination of this report reveals that heavy-duty vehicles
were failing the I/M test at a 37% rate while light-duty vehicles
were failing at a rate of 40%. This difference of 3% is certainly
not a major one. Since that report, an additional 7000 heavy-duty
gasoline vehicles have been tested and the failure rate has risen
to the 40% level. This means that heavy-duty gasoline-fueled
vehicles are now failing Oregon's I/M test at almost exactly the
same rate as light-duty vehicles.
Furthermore, a New Jersey Department of Environmental Protec-
tion studyT/ showed that the degree of commerciality of heavy-duty.
gasoline vehicles has no effect on I/M test failure rates. The
-------
Table E-l — Summary of Results from
Restorative Maintenance Study
1975/76 St.nd.rdll 1.5/lJ/J.l
H.liJjuit-
C»raoctcr mint Hole
Ulc Mixture 31.11
>0.5I CO
Idle Spctd 251
>IOO RfH'fuc lit
»00 HTM llov 101
• - »
>3'^dvmcL'4 11.31
M'tUtMJcJ 9.71
OioVe 10.41
Rich 5.31
L<»<\ 4.3;
Co*i>nrl>on Uecweon VchlcUi*
All Other Vililcteo
UC CO K0»
All other 0.81 1.16 2.64
Viih HiUdjuilmcnt 3.16 41.6 3.78
tl66I 44651 -11
All other 1.26 19.77 2.77
Vltli H«tfl*
rropcrly Adjusted 0.75 M6 2.J6
Hol/iil ju« tc11 vchlclo)
IIC CO I0>
At rcctlvtd 1.31 20,27 ).M
Dlioblcratnt
tlnlne >nJ 1.25 18.44 J.6J
choke1
Idle mliturc
»nd Idle 0.90 8.11 2.69
ipccd
Comploto 0.87 7.6S 1.31
Koit or*c Ion
Hot AviU.bl,
Hoc Aviitiblt
Alttr
Cfltetlvt Ktlpirforvinct
(Ch«n(c It it f ri«lou»
pinlnj tcit)
IIC CO Id
Inrlcbtd to
eltiilc Ion >8St Ollt -41
bcif Idle
Hot AviUibU
AJvinccd 5' O4I 461 «|J(
Enriched 1 *]){ «!0t «15I
notchti
'VihicUi in at received condition.
-------
study split the trucks into two groups; Large fleet trucks (i.e.,
those fleets which had more than 29 trucks) and small fleet
trucks (i.e., fleets having less than or equal to 29 trucks). It
was expected that the large fleet trucks would have a smaller
failure rate because the fixed costs of a periodic maintenance
program can be spread over a larger volume. The results of the
study showed no significant difference between the I/M test failure
rate of the large fleet trucks versus the failure rate of the small
fleet trucks. The actual failure rate achieved by both groups was
about the same as the failure rate reported for heavy-duty gaso-
line-fueled vehicles in Oregon's I/M program.
In summary, for gasoline-fueled vehicles, the data from EPA's
Restorative Maintenance program_l/ shows that parameter maladjust-
ment is the biggest reason for in-use light-duty vehicles' failure
to pass the FTP as well their failure to pass I/M tests. Heavy-
duty gasoline-fueled vehicles have the same engine parameters which
serve the same function as those engine parameters found most
likely to be maladjusted on light-duty vehicles. The same basic
causes for maladjustment of those parameters exist for heavy-duty
vehicles as for light-duty. Since heavy-duty gasoline vehicles
are failing I/M tests at rates identical to light-duty, it is
reasonable to conclude that heavy-duty engine parameters are being
maladjusted to a degree comparable to light-duty vehicles.
The heavy-duty diesel engine manufacturers' rationale that
since EPA has no data on diesel parameter maladjustment, diesel
engines should be excluded from this part of the regulation is not
acceptable. The NPRM did not list any parameter that would be
subject to adjustment by the Administrator for diesel engines
because there is admittedly a lack of data concerning diesel
maladjustment. EPA will not subject any diesel parameters to
adjustment until such time as evidence shows that adjustability
should be limited and the Administrator gives manufacturers
sufficient lead time. The fact that EPA has no direct data
which shows diesel engines to be maladjusted in the field does not
mean that EPA does not have a legitimate concern regarding misad-
justment of diesel parameters or that it does not have a sufficient
basis to create the mechanism for evaluating and regulating para-
meters which at some later date are identified as being maladjusted
in-use such that emissions are significantly affected. Certain
diesel parameters such as fuel injector timing are adjustable and
can affect fuel economy and driveability (as well as emissions) and
could be maladjusted in-use.
Diesel engine manufacturers' were also concerned that these
regulations would limit the adjustment of parameters to an
extent whereby the range of adjustment would be so small that
the engine could not be used in all of the applications in which it
had been used previously. The proposed parameter adjustment..
regulations would not disallow an engine from having a wide range
ITS
-------
of adjustment for a parameter. It may be that wide ranges of
adjustments are necessary because diesel engines are used in a
wide variety of vehicles and applications. However, no matter
what application the engine is used for that engine must still meet
the emission standards. Whether the parameters of an engine are
adjusted for urban delivery usage or line-haul usage makes no
difference as to the applicability of the emission standards. EPA
is not attempting to limit the number of different applications in
which an engine can be used but rather EPA wishes to insure that
the engine will meet emission standards in all its potential
applications and that parameter settings which could cause exces-
sive in-use emissions are eliminated. The question of the possible
need for a wide range of adjustment for any particular engine
parameter would be addressed at the time that parameter was
identified by EPA as being subject to adjustment under the regu-
lations.
4. Recommendations
EPA's technical staff recommends no major changes to the
proposed heavy-duty parameter adjustment regulations. Minor
changes to the NPRM, for clarification purposes, are addressed in
Part II of this document.
-------
References
JY Study of Emissions From Heavy-Duty Vehicles; EPA-460/3-76-012,
"" May 1976.
2/ An Evaluation of Restorative Maintenance on Exhaust Emissions
~ of 1975-1976 Model Year In-Use Automobile, EPA 460/3-77-021,
December, 1977.
_3_/ 763-1975/76 Model Year Surveillance Test Program Report,
Vehicle Surveillance Section, California Air Resources Board,
June 1977, Preliminary Draft.
kj Tune-up: Its Effect on Fuel Economy, Emissions and Perfor-
mance - Results of the 1975/76 Test Program Conducted by
Champion Spark Plug, Champion Spark Plug Company
_5_/ The Incidence of Tampering on Cars in New Jersey During 1975,
Mobile Source Enforcement Division, June 22, 1976.
6/ See item IV.G. 27 in the Docket (Docket #OMSAPC-78-4) .
7j Summary Report on New Jersey Department of Environmental
Protection Analysis of Heavy-Duty Gasoline-Fueled Truck
Emissions, New Jersey Department of Environmental Quality,
Trenton, New Jersey.
-------
F. Issue - Idle Test and Standards
1. Summary of the Issue
EPA has proposed separate certification standards and test
procedures for the idle mode for both gasoline and diesel engines.
2. Summary of the Comments
Manufacturers unanimously criticized the proposed idle test as
redundant. The proposed transient test procedure, as do the
current steady-state tests, already contains substantial portions
of idle (approximately 25 percent), and was claimed will adequately
characterize the contributory emissions of the idle mode.
Secondly, EPA was criticized for failure to document air
quality benefit or needs associated with the idle standard. It was
asserted that factual evidence has not been advanced by EPA to
support the CO "hot spot" - "street canyon" problem referenced in
the Draft Regulatory Analysis. It was also argued that the pre-
sumed need for a heavy-duty idle standard is diminishing over time
with the increased stringency of light-duty standards, as evidenced
by diminishing central city eight- and one-hour CO violations. The
contributory effect of heavy-duty vehicles was characterized as
negligible. In summary, the industry declared that the EPA is
legally obligated under the Clean Air Act to quantify air quality
needs and benefits associated with promulgated standards, and EPA
has purportedly failed to do so for the idle test.
Third, diesel manufacturers complained that diesel engines
have inherently low HC and CO idle emissions. Required idle
certification testing would purportedly add to testing expense with
no corresponding impact on air quality. Furthermore, Cummins
submitted data showing that any diesel engine failing the idle test
would certainly fail the transient test.
Finally, one manufacurer commented that an idle standard will
act as an unnecessary design constraint; design flexibility is
needed to "tradeoff" emissions between various operating modes.
3. Analysis of the Comments
The Cape-21 heavy-duty vehicle operational characterization
study indicated that over 45 percent of truck operational time in
the New York City urban area was spent at idle. Based upon this
fact, it is reasonable to presume that some degree of ambient CO
problems can be attributed to idling heavy-duty vehicles. Also,
high CO ambient readings are commonly found associated with con-
gested, rush-hour traffic. It is also well established that such
congested traffic situations contain high percentages of idle
operation.
-------
Heavy-duty gasoline vehicles subject to certification on
the transient procedure almost certainly will utilize catalyst
technology. Based upon data collected from prototype, catalyst-
equipped engines in the EPA laboratory, it has been observed that
catalysts which are sized to adequately handle the high-power,
high-speed portions of the transient cycle will be large enough to
have the capacity to virtually eliminate idle emissions on a
certification test. The idle standard will provide assurance that
this capacity will be used to control idle emissions. They very
argument made by one of the manufacturers that an idle standard
will act as an unnecessary design constraint eliminating needed
flexibility to "tradeoff" emissions between idle and other oper-
ating modes - is the prime reason the idle standard is needed. The
idle standard would assume that idle emissions are not traded for
emissions in other modes.
Positive air quality benefits can also be attributed to use of
the idle standard in conjunction with the implementation of Section
207(b) of the 1977 Clean Air Act Amendments. The identification of
failed in-use catalysts, air pumps, fuel metering and related
components, and their subsequent replacement/repair would have the
net effect of reducing the number of gross emitters in the heavy-
duty class, and therefore would have the net effect of improving
air quality. Furthermore, use of the idle standards would allow
lower cut points for the Inspection/Maintenance program, and would
make I/M more publicly acceptable since manufacturers' warranties
may be invoked to pay for maintenance. The allowance of lower set
points enables the test to better discriminate in identifying
failed and properly-operating catalyst systems (more so than in the
case of light-duty I/M engines). Overall, the idle test is quick,
simple, cheap, and an effective indicator of failed catalytic
converters.
The staff notes that although the existence of this short test
will make it easier to implement the 207(b) warranty with respect
to heavy-duty vehicles, it will not, of itself, implement that
warranty. Rather final emission performance warranty regulations
are required. The Agency has proposed emission performance war-
ranty regulations on April 20, 1979 44 FR 23784. As proposed,
heavy-duty vehicles could be subject to the warranty. However,
because of comments received asserting that the warranty, as
proposed, was inappropriate with respect to Heavy-Duty vehicles and
engines, the Agency is considering omitting them from the final
rule and reproposing new regulations to cover them.
In any case, it will be future warranty regulations that can
actually implement this warranty for heavy-duty vehicles and
engines. Therefore the staff has not calculated the cost that
would be associated with implementation of the 207(b) warranty with
respect to heavy-duty vehicles and engines. These costs will be
figured in future rulemaking packages on the emission performance-
1*2.
-------
warranty. The staff would like to point out, however; that even if
heavy-duty vehicles and engines are included in the soon to be
promulgated emission performance warranty package, it believes
that the warranty costs for heavy-duty vehicles and engines
would be small. The economic analysis prepared in response to
the 207(b) warranty (see Public Docket EN-79-6) concluded the
additional cost to new light-duty vehicles and light-duty trucks
would be less than $5.00 per vehicle. The Agency received no data
from heavy-duty manufacturers or other parties demonstrating that
the costs would be significantly different for heavy-duty vehicles
and engines.
Furthermore, costs of compliance with the certification idle test
are minimal. As discussed above, catalysts effective on the
transient certification procedure can easily be made to meet the
certificaton idle standard as a matter of course, therefore,
requiring no additional development costs. The only attributable
costs to the idle test procedure are those associated with per-
formance of the actual test, (i.e., insignificant on a cost per
engine basis).*
Diesel engines, however, emit minimal idle emissions (today's
diesels are well under the proposed idle standards), will not be
equipped with emissions sensitive catalyst systems, and evidence to
date indicate virtually no deterioration. Use of an idle test
procedure for diesels, with or without in-use compliance testing,
is expected to have little or no effect.
In summary, the high percentage of time urban trucks spend at
idle, the ease of an in-use idle procedure, the relative effec-
tiveness of an in-use idle procedure in detecting failed catalyst
systems, and the virtually nonexistent costs of a certification
idle standard support its promulgation. No compelling data
at this time, however support implementation of any idle standard
for diesel engines, and a delay in its promulgation is warranted.
This decision could be reconsidered in the future, should the need
become more evident.
4. Recommendations
Retain the idle CO standard for gasoline engines. Delete the
idle tests requirements for diesel engines; delete the idle HC
standard for gasoline engines.
* Reference Chapter 5 of the Regulatory Analysis, "Economic
Impac t."
-------
G. Issue - Leadtime
1. Summary of the Issue
In brief, this issue can be stated as follows: what is the
earliest model year by which heavy-duty engine or vehicle manufac-
turers can comply with the proposed regulations? The Glean Air Act
calls for the establishment of the 90 percent standards for heavy-
duty engines in 1983.
In the NPRM, EPA indicated that manufacturers of both gaso-
line-powered and diesel engines would be able to comply with the
proposed regulations in time for the 1983 model year. This belief
was based upon an analysis of information then available concerning
leadtime for test equipment procurement and checkout, control
technology development, and engine certification. The EPA time-
table for gasoline-fueled engines included 10 months for procure-
ment of CVS systems and dynamometer modifications, 14 months for
technology development, and 12 months for certification. For
diesel, the procurement phase was extended an additional 10 months
and development time was reduced to 4 months. Comments on these
times were requested in the NPRM.
2. Summary of the Comments
All commentors who discussed leadtime claimed that there was
insufficient leadtime to comply with the proposal by the 1983 model
year. Responses varied as to when the proposal could actually be
implemented. Most manufacturers' timetables indicated 1984 as an
attainable goal. Some (IHC for gasoline and Mack) indicated that
1985 was the earliest attainable year, while others (Detroit
Diesel, Caterpillar, and IHC for diesel) indicated that compliance
was not possible before 1986.
Manufacturers estimated equipment procurement and checkout
times extending into mid-1981 (as opposed to late 1980 under the
EPA timetable). Gasoline-fueled engine manufacturers estimated
development times of 3-6 months longer than EPA's original esti-
mates (14 months). For diesel engine manufacturers, these develop-
ment times ran from 1-1/2 to almost 3 years, as contrasted with the
4 months contained in the EPA proposal.
Commentors offered a wide variety of suggestions as to what
EPA should do in response to these leadtime problems. They can be
broadly categorized as those which advocated withdrawing most of
the original proposal and substituting 90 percent reduction stand-
ards based upon the current test procedures, and those involving
use of some form of a transient test procedure with leadtimes
extended beyond 1983. These suggestions, and EPA's responses, are
detailed in Part II of this Summary and Analysis of Comments.
(if
-------
The timetables developed by the individual manufacturers are
presented in Figures G-l (gasoline-fueled) and G-2 (diesel) and
reviewed below. It should be noted that in the interest of clarity
Figures G-l and G-2 do not necessarily include all of the detail or
keep the same terminology used by each manufacturer.
a. Gasoline-Fueled Engine Manufacturers
Ford
The Ford comments on leadtime analyzed the time required for
what Ford considered the three main components of a compliance
program: a) developing transient testing capability, b) engine
development and certification, and c) catalyst development. Ford
concluded that 1983 was not attainable because the required timing
of these three elements was not compatible. To determine what
model year would be feasible using Ford's timing, the Technical
Staff has combined these elements into an overall schedule.
The Ford timetable shown in Figure G-l indicates compliance
for the 1984 model year. Ford has had limited transient test
capability since March 1979 in one cell. Equipment for two full
cells has been ordered, and they are expected to be operational by
May 1980. The remaining cells to make up the full compliment
of 12 cells which Ford feels it needs are projected for June 1981.
These times are based upon 9-10 month leadtimes for equipment and
1-2 month installation and checkout. The schedule assumes that the
catalyst design program will begin in January 1980 (after issuance
of Final Rule). This program has been timed by Ford to allow a
minimum of 44 months leadtime before start of vehicle assembly
operations. The schedule also assumes the availability of a
"forced-cooling" procedure for an accelerated testing rate (see
"Test Procedure" issue and durability testing based upon a 50,000-
mile useful life.
Ford's catalyst development program is based upon preliminary
test data which "clearly indicates that existing light-duty cata-
lyst technologies are inadequate..." Ford also states that the
catalyst devleopment schedule "assumes that no major problems
occur which would necessitate new catalyst designs during the
catalyst development program."
General Motors'
The timetable supplied by General Motors also indicates 1984
as an attainable model year. Partial testing capacity in two cells
is expected by November 1979. Full testing capacity (with a CVS
sampling system) in two cells is expected by December 1980.
At that time, GM' s engine development program would begin.
Final certification would be in time for the 1984 model year.
Combined delivery and checkout time is estimated by GM as one'-
-------
Figure C-l
Gasoline Fueled Engine Leadtlmee
FOltD
<,.H.
CIIKYSI.KI
r.u.c.
1979
1 f 1
2 Partial 12 Full
Cells 1 Cells
2 Partial Dynes • 1
1 — "
1
1980
i i i
(Cell 3 I Cell;
' „ :i
1 Advanced 1500-1
Hour Deyelj
4 Full Dynos —
irdware 1 i
, 1 '
'al.
2 Cells 1 4 Cel
1981
1 • 1
4-12 1 Englii*
Practice
Cell 1 . I Develoj
Cell ° • 1
Cell 3
Cell
1982
1 1 f
I*"" llr
Cert 1
Development
inrnt 01 I
| , Diirahlllly ,
1983
1 > '
I9U4 Cert 1
1984 Cert i
IQS4 1
1
lOB/i Cert II
-r-4
1—1985 Cert 52
1 1 . I9fl"> C
'11, ' It
II
1984
* I . |
IQflS rnvt^ I
-------
year. GM has allowed 6 months for equipment installation and
checkout, for either dynamometer control systems or CVS systems.
Production tooling leadtimes in the GM schedule are based upon
assumptions that could cause considerable delay if not borne out.
Leadtime for catalysts is based upon use of cores already being
tooled for "certain passenger car applications." As discussed
elsewhere in the evaluation of feasibility, GM feels that there may
be significant durability problems necessitating larger catalysts.
If this were to occur, leadtime would be extended.
The GM timetable is also predicated upon the assumption that
their quadrajet carburetor will not have to be redesigned or
replaced. This carburetor, which uses air valve secondaries and is
found on a majority of GM heavy-duty gasoline-fueled engines, has
been identified on engines tested by EPA as a possible source of
high transient emissions.
GM has indicated the potential for delay in this schedule in
other areas as well. Their dynamometer equipment is a non-standard
General Electric design featuring aluminum core armatures. While
it appears that these dynamometers are suitable for transient
operation, there is a possibility that the armatures may not be
capable of handling peak field currents. That would necessitate
the installation of new dynamometers and reduce available develop-
ment time and tooling leadtime. GM's estimates are also based on
the assumption that separate emission systems for California
(requiring separate certification) will not be required.
Chrysler
Chrysler is in a distinct position among gasoline-fueled
engine manufacturers in that current heavy-duty certification is
being done using eddy current dynamometers. The leadtime to
purchase new dynamometers is significantly longer than that to
convert existing equipment. However, Chrysler has already begun
construction of 2 new cells, which are expected to be fully oper-
ational by August 1980. Two existing electric dynamometer cells
will be converted and will be operational by April 1981. In their
submittal, Chrysler has indicated that their full development and
certification process would have to begin 2-1/2 years prior to the
first effective model year. This would require a minimum of 4 test
cells. Thus, the first year for which Chrysler could certify
engines under its proposed schedule would be 1984.
International Harvester
IH was the only gasoline manufacturer indicating more than
four years of leadtime required to certify to the transient proce-
dure. The schedule submitted by IH indicates completion of their.
first dynamometer system by June 1981, with the remaining three
IC7
-------
systems being staged at three-month intervals (based upon IH
manpower limits). Allowing 18 months for development and 12 months
for certification for each engine family results in a program
stretching to September 1984 (in time for the 1985 model year).
As was the case with GM, the IH schedule is predicated on the
assumption that IH will not have to develop engines to meet separ-
ate emission standards for California. This requirement would add
additional time.
b. Diesel Engine Manufacturers
General Motors
An outline of the timetable submitted by GM and other diesel
manufacturers is given in Figure G-2. GM projects one research
test cell for September 1979. This would be used for some pre-
liminary engine characterization and as a basis for developing
specifications for the 12 cells needed for GM for engine develop-
ment. These 12 cells would be procured in 1980 and installed by
September 1981. They would be installed in a new diesel test
laboratory which is now being built. This facility was already
being built at the time of the proposal, and was initiated for
product development independent of regulatory requirements.
Technology assessment would begin on the research cell. Upon
completion of the installation of the 12 cells, a two-year and
9-month period of engine development and pre-certification testing
would begin. Certification testing could begin in July 1984. As
stated by GM, implementing the transient test would be "very
ambitious for the 1985 model year."
Caterpillar
The main emphasis of the Caterpillar leadtime comments con-
cerned development and validation of a revised test procedure
using eddy current dynamometers. Based upon the presumptions that
a revised cycle could be easily "validated" and that emission
standards would be set at or above the levels of current production
engines, Caterpillar projected that a revised test procedure could
be implemented for 1983. Delays in validation or the need to
reduce engine emissions would extend the Caterpillar schedule 3-5
years. Further consideration of the Caterpillar timetable will be
based upon the assumptions that the test procedure promulgated by
EPA will not require a validation program and that reductions in
engine emissions will be required.
The timetable which Caterpillar would follow under the above
conditions is the longest of all manufacturers' estimates. Cater-
pillar estimates that compliance with the proposal might not be
possible before as late as the 1990 model year. The earliest model"
-------
Figure G-2
Diesel Engine Leadtlmes
1979
, |
Pilot
(;elli_rr
st
2 Cells
Cater--
nlllar
~~ Cuiimiins
Inlerpretgt Ji>ns)
Hack
I.II.C.
19BO
1 i i
Ira. I
nil en
-H
10 Cells (Modi Hi
Procure and Tnsta]]
:haracterizatlon & Devci
1 Cell 2
I — -t'rocedi
1981
I
1
lugy Autumn.' ' •
1
J Cycle) f-~—l= s
Gqulp. 1
.
Baaellnes
r*1 Pq'iip. , r Pynn
l)y
1982
I j |
tievelopm
i ... i
t\ .-I, , , ppurtlnpmcnl-
1 I
10 02 |
1983
i 1 '
1 Pre-Cert
nc l-i-b years)
•f
1 — Pre~Cert Ihirablil
Family Hi 1
Family 02 1
i F<
1964
f * f
•85
TeuLliiB ' Cei L
'86-'
I C"Tt
nC (3-5 years) and Cert
5
rt 1 Toolinp-J
vH
— — .15 Cert 11
1 • '86 CULL 112 "
1 |— '86 Cert
tit 1 (—'86
nJly 05 1
1985
CBfi?)
_M
18
fa'W *"
'3 J
3 -|
Cert >-
J86 Cert^-
-------
year in which Caterpillar projected compliance with the proposal
was 1986 (assuming it was possible to retain existing eddy current
dynamometers and that an engine development program to meet the
final standards would be required).
Caterpillar presented the various elements of their timetable
at different stages of their testimony and did not tie them to-
gether into one overall schedule. In fact, the. various aspects
of the Caterpillar timetable, when assembled together in Figure
G-2, reveal certain discrepancies regarding the development of
transient test capability. During oral testimony at the May 14-15
Hearing, Caterpillar indicated that two transient test facilities
were on order and would be operational by the end of 1979 (page 118
of transcript). Caterpillar further indicated that another 20
months would allow them to have their desired quota of ten facil-
ities operational. This later time was expressly dependent upon
the assumption that some modified cycle permitting the use of
existing eddy current dynamometers would be acceptable to EPA.
Caterpillar's written submission following the Hearing modified
these times without explanation. On page 17, the "end of the year"
from the oral presentation becomes "early 1980". Page 26 changes
the 20 months needed to have 10 operating facilities which projects
to September 1, 1981 to "the beginning of 1982". That presentation
indicates that all ten cells would be obtained through modifica-
tions to existing cells, and does not appear to consider the two
systems already on order. Since the testimony presented at the
Public Hearing was more specific as to the timing of the availa-
bility of the 10 needed cells than the somewhat generalized phra-
sing of the written submission, the September 1981 projections will
be used for this analysis.
The Caterpillar presentation relied heavily on the possibility
of retaining existing eddy current dynamometers. Caterpillar
indicated that a full 3-1/2 years would be required to add facil-
ities capable of running the unmodified transient cycle. Telephone
discussions with Mr. Joseph Haefele of Caterpillar revealed that
this time was necessitated by the need to construct a new wing on
their technical center to house the test cells. It included design
of the facility as well as actual construction. Mr. Haefele
indicated that construction of the facilities would be completed by
January 1983. This would be followed by a six-month installation
period for test equipment, and an anticipated three-month debugging
period. The cells would then be fully operational by the fourth
quarter of 1983. This time is three months longer than presented
in the Caterpillar written submission (corresponding to the debug-
ging period). Mr. Haefele indicated that Caterpillar's recent
experience in constructing a new wing to its technical center
confirms such a. timetable. That project took over 3-1/2 years to
complete.
-------
Turning to engine development, Caterpillar presented a general
outline for "any significant modification to a product line" which
the company indicated would take from three to five years to
successfully complete. The approach was not based upon considera-
tions of particular technologies to be evaluated or engine design
changes which might need to be made to reduce emissions.
Cummins
Cummins' presentation regarding leadtime was very brief. They
estimated 1-1/2 years for equipment installation and followed that
with the position that "we will not have time both to implement a
transient test system and to develop new emission control technol-
ogies based on that system for the 1983 model year." Cummins did
not indicate how much development time was appropriate, nor what
model year they could first expect to certify to. However, coupled
with Cummins' repeated call for a 4-year leadtime which they
believe is required by the Clean Air Act, the preceding wording
suggests the 1984 model year for Cummins. Cummins already pos-
sesses some transient testing capability and has generated sub-
stantial test data. Their current capability surpasses all other
diesel manufacturers. The EPA technical staff interprets the
materials supplied by Cummins as indicating the possibility of
certifying for the 1984 model year. The associated schedule is
given in Figure G-2. Cummins cautioned that CVS delivery times
which they used could be delayed. They believe that equipment
suppliers may not be able to meet industry demand. In addition,
they feel that delays could result from equipment changes resulting
from development of the EPA particulate procedure (e.g., heat
exchangers).
Mack
The timetable proposed by Mack indicates that 1985 would be
the earliest model year for which Mack could certify their engines.
They expect to have two cells operational by May 1981. After that
follows a 2-1/2 year period which is divided into three phases:
one year to "establish baselines," six months for development of
control technology, and 12 months to assure engine durability. EPA
staff interprets the baseline phase as including current engine
characterization and some assessment of technologies - a category
corresponding approximately to what other manufacturers describe as
advance development or technology assessment. The development and
durability phases correspond to what other manufacturers generally
have included in the single category of development. The Mack
timetable concludes with six months to obtain certification and six
months to introduce minor production tooling changes for the 1985
model year.
In separate testimony from that presenting the above time-
table, (p. 18 of June 27, 1979 comments) Mack has indicated that
-------
they expect to have their first complete cell operational in
"early 1980." This fact does not appear to be reflected in the
above timetable.
International Harvester
The IH timetable calls for 66 months of leadtime for diesel
engines to reach production. In presenting this schedule IH
indicated that they "would be reluctant to invest in any extensive
number of dynamometers and the control equipment for the proposed
transient test cycle until the Final Rules for particulate measure-
ment have been published in the Federal Register." Therefore, the
"leadtime for diesel engines can begin no sooner than the Final
Rule" for particulates, which they assumed would be June 1980.
IH would not project certification of any engines before the 1985
model year; and it would not certify all engines before the 1986
model year.
IH used the 20-month leadtime for equipment acquisition which
had been estimated by EPA in the NPRM. Stating an inability to
project the required amount of engine development time pending
knowledge of the yet-to-be proposed particulate standard, IH
"assumed" a period of 18 months for this phase of its program.
Near the end of the development program, durability testing would
begin. This durability testing would also overlap the initial 2
months of certification,, as shown in Figure G-2. Total time for
acquisition of the first test cell to completion of certification
of the first engine family is 34 months.
As was done with their gasoline engine program, IH would plan
to stage the installation of test cells for its second through
fifth engine families at 3-month intervals. IH would therefore
certify its last engine line for the 1986 model year.
Others
The Engine Manufacturers Association commented that the
20-month period estimated by EPA for procurement and installation
of test cells is "unrealistic and unsupported by the record" and
that the 4-month development time for diesel engines is "totally
insufficient." EPA's suggestion that equipment might be purchased
in advance of Final Rule was also rejected. The Motor Vehicle
Manufacturers Association similarly rejected the idea that manufac-
turers should already be working toward achieving the reductions
specified in the Act.
Perkins Engines commented that the 20-month leadtime for
equipment estimated by EPA was adequate. Because of insufficient
development time after that period, Perkins recommended that the
transient procedure remain optional for the first two years.-
Perkins presented no specific timetable for compliance.
-------
IVECO Trucks of North America (importer of heavy-duty diesel
vehicles) commented on the special hardship which they foresaw for
smaller manufacturers. Shorter leadtime was seen by IVECO as
forcing increased competition between manufacturers for limited
supplies of test equipment. Being unable to afford to pay premium
prices that larger manufacturers could absorb would put smaller
manufacturers at a time disadvantage. In addition, smaller manu-
facturers could not readily afford to order equipment in advance of
Final Rule because of the financial risk if substantial changes
were made in the Final Rule requirements.
3. Analysis of the Comments
To provide a basis for comparing and analyzing the various
schedules which have been submitted to EPA, the information will be
used to revise the original timetable proposed by EPA in the NPRM.
This will be done separately for gasoline-fueled and diesel en-
gines.
a. Gasoline-Fueled Engines
The gasoline timetable proposed by EPA included ten months for
procurement of test equipment, fourteen months for technology
development and twelve months for certification. Each one of these
areas can now be updated.
The limiting factor considered in EPA's procurement phase was
the delivery of equipment. Gasoline-engine manufacturers (except
for Chrysler which has advance ordered new dynamometers already) do
not need to purchase new dynamometers, so that the time limiting
factor becomes the emissions sampling system. EPA's original time
estimate, as well as the estimates supplied by the manufacturers,
was based upon use of a constant volume sampler (CVS) with cri-
tical flow venturi (CFV) flow regulation. This is the type of
system now being used by EPA. One principal vendor of CVS systems
has indicated to EPA that delivery times of 6-7 months are cur-
rently being quoted (including backlogs). Ford in its submission
used 9-10 months, which is significantly longer than the above
estimate. However, they used a correspondingly short time for
installation and checkout as we shall see below.
An alternate approach employs a positive displacement pump
(POP) for flow measurement. Delivery times for PDF systems would
be approximately 6 months as compared to the 7 months for CFV
systems.
One possibility has been considered that has the potential of
eliminating the delivery time problem to get an early start on
perhaps one cell. This involves using two light-duty CVS systems
in parallel. Assuming systems were already available from the-
manufacturers light-duty work means no delay awaiting system
-------
delivery. However, it must be noted that delivery times for
dynamometer control systems to convert electric dynamometers to
transient control are estimated by the techincal staff as approx-
imately 6-7 months. In addition, all manufacturers indicated in
their submittals that procurement for one or two advance cells had
already begun, the latest of which would be available by the end of
August, 1980. This would allow the early beginning of work,
especially for items with long leadtimes. It also makes the
attempt to use light-duty CVS systems in parallel unnecessary.
Following procurement, either of the alternate measurement
systems (GFV or PDF based) would require a period for installation
and checkout. Recent EPA experience indicates that 3-6 months is a
reasonable estimate for this time. The 3000 CFM CVS delivered to
EPA in January, 1979 took six months to be fully operational.
However, this unit was developmental in nature and had several
unique features that had to be checked out. It therefore repre-
sents maximum installation time. Ford estimated installation would
take 1-2 months. In the case of multiple dynamometer instal-
lations, the EPA staff estimates that the first system could be
operational in approximately 3 months, and all cells in 6 months.
Ford estimated an installation time for 10 cells of approximately
10 months. GM used 7 months for 4 cells. These schedules probably
contain time for unforseen delays.
The overall equipment acquisition and installation phase can
be determined by combining the above estimates. The result is
approximately January 1981 to have all cells operational. For
comparison, manufacturers estimates of this time are as follows:
Procurement and Number of
Manufacturers Installation Time Cells
Ford July, 1981 12
GM Jan., 1981 4
Chrysler April, 1982 4
IH April, 1982 4
GM's projections meet the EPA timetable. Chrysler presents no
specific analysis of their procurement and installation activities,
and it appears reasonable to believe Chrysler could accelerate
their program sufficient to meet the January 1981 date. The
additional time projected by Ford is due to extended delivery time
estimates and the large number of test cells (12) Ford expects to
have. While Ford may choose to have 12 transient cells, this does
not appear to be necessary for their projected 8 engine families.
EPA's staff estimate, based upon historical ratios of dynamometers
to engine familes for heavy-duty manufacturers, indicates that 9
test cells should suffice for Ford (one cell per family plus one
additional). The January 1981 date may be difficult for Ford, but
should be feasible.
Pt
-------
Amongst the manufacturers, IH stands out as departing in a
major way from the EPA estimate. For this reason alone the
figures might be considered as unnecessarily long. IH indi-
cated during testimony at the May 14-15 Hearing (pages 573-574 of
transcript) that one gasoline cell and one diesel cell had been
ordered and would be installed "give or take about 11 or 12 months"
from that time. In a telephone conversation with Mr. Bill Martin
of the IH gasoline staff on July 20, 1979, Mr. Martin indicated
that the new gasoline cell had not actually been ordered yet, but
would be very soon thereafter. Mr. Martin indicated that this cell
should be operational by September 1980.
If the equipment for the remaining three cells is ordered in
January 1980, the second cell should be available by the end
of September 1981. IH had planned to stage follow-on cells at
3-month intervals. This was to conserve manpower requirements.
However, as the second and subsequent cells go through instal-
lation, it is reasonable to expect the installation time estimated
for the first cell to be reduced through gained experience.
Therefore, EPA believes that cells 3 and 4 could be brought on line
by IH by January 1981.
The second phase of an overall timetable is control system and
engine development, which, .as was indicated in the NPRM, can be
broken down into work involving use of vehicles (e.g., catalyst
environmental factors and durability assessments) and work done in
test cells. EPA is aware that manufacturers have already begun
vehicle testing of 'catalyst systems to evaluate their ability to
function in the heavy-duty environment. Dynamometer testing can
begin with engine characterization and technology assessment after
the advance procurement of 1 or 2 cells by the manufacturers, and
be followed by engine development as later test cells become
available. EPA had originally estimated that 14 months would be
sufficient for development of gasoline-fueled engine control
systems. This estimate was based largely on the assumption that
"for the most part, the HD manufacturers will be able to utilize
the catalyst control technology currently used on light-duty
vehicles and light-duty trucks" (44 FR 9471, February 13, 1979).
Testimony submitted during the comment period indicated that
manufacturers were encountering significant durability related
problems in applying these systems (see the analysis of feasi-
bility). These problems have stemmed from the higher loads and
higher power requirements for heavy-duty engines as well as from
closed-throttle motoring. Larger, more heavily loaded catalyst
systems protected during motoring by air-pump or fuel shutoffs will
probably be needed.
EPA is aware that Chrysler has developed a catalyst based
system for possible use in meeting the 1980 California standards
for its 360 cubic inch engine. Chrysler had not actually certified-.
this engine at the time of this analysis, but data has been submit-
-------
ted from testing under current test procedures for durability
engines which indicates that the engine could be certified.
Durability 1500-hour data on the 9-tnode procedure was as
follows: HC = 0.64 g/BHP-hr, CO = 20 g/BHP-hr, HC + NOx = 3.9
g/BHP-hr. EPA has had the opportunity to test this same config-
uration as one of the current technology engines being evaluated on
the transient procedure. Transient test results were as follows:
HC = 1.39 g/BHP-hr, CO = 136 g/BHP-hr, NOx =2.33 g/BHP-hr. This
data indicates the impact of the change in test procedure on the
catalyst system performance. The major contributors to the high
transient emissions from this engine were the cold-start bag and
the Los Angeles Freeway portion of the cycle.
In reviewing the test data submitted to EPA on the durability-
data engine, it appeared that as the engine approached the 1500-
hour point, the catalyst either failed, or was about to fail
because the emission rates began to raise rapidly. Clearly, the
system would not be applicable in meeting EPA requirements without
significent changes.
EPA has also tested two heavy-duty engines in catalyst config-
urations certified for use in light-duty trucks. These two engines
(a GM 400 and a Ford 351) both experienced very high CO levels, in
excess of 100 g/BHP-hr, as had the Chrysler engine.
Based upon the data now available, the EPA technical staff
believes that somewhat more development time than the 14 months
originally estimated will be necessary. Although the durability
problems are real, the EPA staff has already identified approaches
which could be used to cope with them. In addition, EPA also
recently demonstrated the feasibility of the target emission levels
associated with these regulations. (For more discussion of these
areas see issue D - Allowable Maintenance, and I - Technological
Feasibility.) Therefore, only a modest increase in development
time over that originally estimated will be required. Eighteen
months of development time will be used as a conserative estimate
for gasoline-fueled engines.
The final phase of EPA's overall timetable as proposed con-
sisted of one year for the certification process. A review of the
steps in this process indicates that less than twelve months is
required for certification, but that it must be keyed to the start
of production, which for gasoline-fueled engines (for those manu-
facturers who also make light-duty engines) is late summer or early
fall. The current process consists of three steps: the Part I
application review, testing of durability and emission data en-
gines, and the Part II review. These steps take approximately 1
month, 5 months, and 1 month, respectively, for gasoline-fueled
engines. An abbreviated certification process could eliminate the.
first step. However, with full-life useful life and the need to
-------
establish deterioration factors comparable to in-use values, it is
likely that an increased amount of time will be needed by manufac-
turers to establish durability. Therefore, this analysis allows 7
months for certification. Based upon issuance of a certificate 30
days in advance of a September 1 production start date, the certi-
fication process would have to begin by January 1 of the model year
previous to that being certified. This confirms the original EPA
estimate for gasoline-fueled engines.
The elements of equipment procurement and installation,
develoment, and certification can now be combined into an overall
schedule, as shown in Figure G-3.
The timetable as given misses the deadline for start of 1983
engine production by approximately two months. It is concievable
that if all went well, certification might be possible for 1983.
Alternatively, start of 1983 production might be delayed a couple
of months toward the end of 1982. However, the EPA schedule as
developed above has already been based upon optimum timing esti-
mates. No time was allowed to develop CVS specifications or
negotiate sales contracts before ordering test equipment. Delays
in delivery, installation, or in time for personnel training on
the test equipment of several months, although they cannot be
specifically identified by their nature, can be considered likely.
If there were unusual problems in control system development they
might also extend the development time. In addition, this schedule
has not made any specific allowances for tooling time for catalysts
and possibly other engine parts. Tooling leadtimes have been
estimated by manufacturers as follows:
Manufacturer
Ford
Ford
GM
GM
IH
Lead Time
26 months
32 months
15 months
Comments
36 months
21 months
Catalyst.
Major equipment parts.
Catalyst. Based upon
use of catalyst already
being tooled for other
applications.
Redesigned carburetor.
Some of these times are probably longer than would actually be
necessary. If we use the IH estimate of 21 months for redesigned
-------
Figure G-3
EPA Projected Gasoline - Fueled Engine Compliance Schedule
1980
i t i
Technology
Advance Cell LAssmnt.—
— Procurement Cell 2 -i- A;J^ _
1 Cells
.. .. -Pir-lH Dni-lfi-M 1 Mr ..— 1. ..
1981
*
1982
i i |
Start '83 Prod j
11 ORA Port- — - 1
1983
111
*Note: Completion of installation for some
cells could further overlap with development.
OQ
-------
components, then catalyst and carburetor designs would have to be
done by early 1981 in order to certify for 1983, which seems
unlikely under the timetable of Figure G-3.
Both GM and IH indicated in their submissions that their
projected timetables included the assumption that separate emission
systems for California would not be required. At this time that is
probably a reasonable assumption. If there were a potential for
delay arising from separate California standards it could be raised
at such time as California were to apply for a waiver from the
Federal standards.
In support of their contention that EPA was underestimating
leadtime, GM included a schedule projected by EPA in a memo of
August 19, 1975, ("Scheduling of Final Heavy-Duty Vehicle Regula-
tions", D.A. Finley to Ernest S. Rosenberg, Chief, Regulatory
Management Staff). That schedule projected a total time of 5 years
and 9 months from final rule to certification as compared to the 2
year 10 month leadtime developed above. There are many aspects of
the 1975 projections that could be detailed to explain this discre-
pancy. However, it is more appropriate to realize that this memo
is simply a very early projection of what was at that time a
largely unknown future process. The memo itself states that "at
the present time, scheduling for this program remains speculative"
and indicates that the timetable developed should be considered an
"outside estimate". Therefore, those early projections cannot be
used to challenge current estimates.
b. Diesel
The timetable for diesel engines proposed by EPA included 20
months for equipment procurement and installation, 4 months for
development, and 12 months for certification. Data now available
to EPA indicates that some changes in those projections are in
order.
The 20-month's leadtime for equipment acquisition was based
upon the purchase and installation of electric motoring dynamomet-
ers to replace the eddy current dynamometers currently being used
by diesel manufacturers. This time period remains accurate
EPA contacts with dynamometer vendors and delivery times quoted by
manufacturers in their submission indicate 12 months for equipment
delivery. The EPA staff estimates that five months for installation
and checkout is adequate. This 17-month period would correspond
to the first cell being operational, with a three-month additional
period for bringing multiple cells on-line. All major manufac-
turers have indicated that procurement of one or two advance cells
with electric dynamometers is already underlay. If the remaining
number of required test cells were ordered in January 1980, they
would become available for testing in June - August 1981. With the
exception of Caterpillar and IH, all manufacturers' projections
were within a couple of months of these estimates.
-------
Caterpillar indicated in their testimony that conversion to
electric dynamometers would require the addition of a new testing
facility to house the equipment. Caterpillar further indicated the
time to do this would be three years for construction of the
facility and six months for equipment installation. The EPA
technical staff believes that the construction period could be
significantly reduced, but that availability of test cells would be
delayed considerably even under the most optimistic assumptions.
For example, if the new facility could be built in 1-1/2 years,
test cell installation would not be completed before early 1982.
Therefore, Caterpillar would probably have to house the new equip-
ment, at least temporarily, in their existing test cells. While
this might not be desirable from Caterpillar's viewpoint, it would
be feasible. Caterpillar's intention to remodel if eddy current
machines were retained indicates that the CVS systems can be
accommodated in the current cells. New dynamometer systems should
also be able to be accomodated. For example, one type, the regen-
erative brush type of electric motoring dynamometer, requires no
more space than an eddy current dynamometer. There would be extra
expense associated with installing the test equipment in remodeled
cells and later moving it to a new facitlity, but this might be
necessary for Caterpillar to keep pace with the rest of the indus-
try.
IH has indicated their intention to delay the aqcuisition of
new test equipment pending the issuance by EPA of final particulate
standards. Although the concern raised by IH is understandable, it
is clearly unreasonable to delay a current compliance program for
the sake of an as yet undefined future rule. Rather at such time
as a particulate standard were proposed, that proposal would have
to consider the impact of new standards or possible test procedure
changes on any ongoing programs as one effect of the proposal. If
there were any need to establish new timetables, it would be
determined at that time. The impact of changes in test procedure
upon equipment investments already made would also be evaluated at
that time.
In addition to the issue of the starting date, the EPA tech-
nical staff believes that the timetable presented by IH is longer
than necessary. In the IH verbal testimony at both the May 14-15
Hearing and the June 16-17 Hearing, IH indicated that a single
diesel cell was ordered. The May 14-15 testimony indicated that
the system would be operational in approximately one year. Based
upon this information, availability of the first IH dynamometer
system can be accelerated from September 1, 1981 to June, 1980. In
addition the installation of the second and subsequent systems can
be accelerated. The 20-month leadtime for equipment acquisition
used by IH was based upon the EPA estimate used in the NPRM.
However, as indicated above, that time is considered sufficient for
installation of multiple dynamometers, rather than just the firs't
dynamometer as used by IH. The EPA staff estimates that the second
-------
IH system (in addition to its advance ordered system) should be
available by June 1981 and the remaining three systems by September
1981.
Data submitted during the comment period (principally by
Caterpillar) on alternate transient cycles indicates the possibil-
ity that an eddy current dynamometer-based system may be capable of
reproducing the transient cycle, or a somewhat smoother version of
that cycle, close enough to produce comparable emission results to
those obtained on an electric motoring dynamometer. However, there
is insufficient data at this time to resolve the question of how
close it may be able to come. A principle cause for caution
arises from the fact that the torque and speed response relation-
ships on an eddy current dynamometer (which operates basically in
what is called "torque control" mode) are different from those of
an electric dynamometer (operating in a "speed control" mode).
These different system characteristics could give somewhat differ-
ent emission results, even if the same cycle tolerance specifica-
ions were met. Small differences could become important for an
engine whose emissions were close to Che standard. Therefore, any
manufacturer wishing to retain his eddy current dynamometers rather
than purchase new electric dynamometers may feel it necessary
to first undertake a pilot program to establish test correlation
with EPA results. If this were successfully accomplished, the
manufacturer would then be able to use eddy current dynamometers.
The EPA staff estimates that it would take approximately six
months to obtain the dynamometer control system to convert an eddy
current dynamometer to transient operation. If the CVS system from
the manufacturer's advance cell could be used for sampling, the
cell might be on line in an additional four months. Two months of
testing to establish correlation would then follow, making a total
of approximately one year. If correlation were successful, the
manufacturer could then proceed to convert his remaining eddy
current machines. This could be done on the same timetable as
developed previously for gasoline-fueled engines (9 months for the
first cell to become available plus three months for remaining
cells).
The net result of this process would be a timetable some
four months longer than the timetable for replacement of eddy
current machines with electric machines (both alternative time-
tables are given in Figure G-4). This delay, coupled with the risk
Chat eddy current dynamometer-based measurements may not in fact
correlate well enough to be usable, would make ic unlikely that any
manufacturer would try that approach.
Because of the potential cost savings involved if eddy current
machines could in fact be used, it seems desirable to Che EPA
staff to attempt to reduce the risk facing manufacturers wishing t-o
explore this option. Since Che program Co esCablish correlacion
-------
Figure G-4
EPA Projected Diesel Engine Compliance Schedule
1980
i i i
n _ , , I Installjjest
~~l)yno Control-! — ,, . ~t~vii '
if I • 1r 1
Advance .
Cell '" 1
, Procurfiinfint
I Characterize
* Assessment
1981
1 | i
Using F.ddy Current
Procurement //2 1 All
n- ! i |Cells
Tprnnol nrrw • 1
Assessment
Usin.T F.lectric Motoring
^ 1 Mil . . *
1 *
1982
1 * '
nvnanorifiterfi
A
\~
Dynamometers
Start 1983 Prod.
I— Durability
1983
1 1 1
1
1
*Note: Completion of installation for some
cells could further overlap with development,
-------
would require approximating one year, this goal could be accom-
plished by allowing some form of optional certification on the
13-mode procedure for the first year of implementation of the new
standards. The optional standard would be derived from comparison
of transient and 13-mode emission levels. It would be established
with the intent of approximating as closely as possible the bene-
fits to be realized from the transient test. This might involve
some loss of air quality benefits for that year because of the fact
that 13-mode emissions cannot accurately quantify in-use transient
emissions. However, the fact that manufacturers would have to
certify to the transient procedure in the second year should
preclude any significant loss of benefits. Manufacturers will
avoid the need to modify engine lines twice in such a short
period of time.
As test cells become available, engine testing will begin. In
contrast to the situation with gasoline-fueled engines, little
diesel work can be done before acquisition of transient testing
capability. This is because the problem facing diesel manufac-
turers is not so much one of in-use durability of systems as one of
determining the transient performance of various technologies or
engine changes which might be used to reduce emissions. Advance
cells would be used for initial engine characterization plus some
technlogy assessment. As more cells become available, engine
family development would begin. Engine development time was limited
to four months in the EPA proposal. Manufacturers have made a
valid case that this time is insufficient. In another section of
this document (see the "Test Procedure" issue), estimates have been
made of those engine families which would exceed the 1983 stan-
dards, based upon the target values supplied by manufacturers in
their submissions. These estimates indicate that approximately 70%
of diesel engine families will need development work. For some
manufacturers (GM and IH), all families may need emission reduc-
tions. However, many of these families will exceed target values
by relatively small amounts, and should be easily brought into
compliance. The EPA technical staff estimates that diesel engines
should require less development time than gasoline engines.
Fourteen to 16 months should be sufficient.
Development times submitted by manufacturers exceeded these
estimates considerably in some cases:
Estimated
Manufacturer Development Time
GM 27 Months
Caterpillar 3-5 Years
Cummins No specific estimate
Mack 18 Months
(after characterization)
IH 18 Months
-------
GM presented no basis for its estimate of 27 months. In fact,
their testimony indicated a lack of information on which engine
families would need correction or by how much. The EPA staff
estimates of reductions needed for GM diesel engines (see Table-6
of analysis of "Test Procedure" comments)indicates that several GM
families will probably need only small reductions. It is also
unlikely that GM would require more time to develop its diesel
engines than its gasoline engines, which GM has indicated would
consume 19 months (excluding 5 months of durability).
In its testimony concerning feasibility, Caterpillar estimated
that 4 of its 11 currently certified engine families fail to meet
the standards. This agrees with the EPA staff estimates which also
indicates that only one of thse families would require what could
be considered a large reduction. The development time of 3-5 years
proposed by Caterpillar was not related to the estimates of engine
families needing work and appears inconsistent with those esti-
mates. With one complete test cell expected to be available by the
end of 1979 or early 1980, work on the single family needing the
most work could begin. Work on remaining engine lines could be
done within the time frame estimated above by the EPA staff.
Near the end of the development process, the manufacturers
will be able to begin durability testing. Since engine useful
lives will be approximately double the values currently being used,
it will be important to assess emissions durability. The current
accelerated diesel durability testing takes an average of 3 months
to complete, so an estimate of 6 months will be used to assess the
extended useful life. Manufacturers may well desire more time than
this to adequately assess the in-use performance of engines for
long useful lives.
The current certification process for diesel engines is
similar to that for gasoline-fueled engines. The three steps and
required time intervals are as follows: the Part I application
review, 1 month; testing of durability and emission data engines, 4
months; the Part II application review and issuance of certifica-
tion, 1 month. An abbreviated certification process such as is
now being implemented for light-duty and heavy-duty certification
could eliminate the Part I review. Durability testing has already
been accounted for in the previous paragraph. Since manufacturers
will establish their own preliminary deterioration factors, there
will be no "official" durability testing as part of certification.
Emission-data engine testing should take approximately 1 month at
the maximum of two emission data engines per family. Allowing for
issuance of certification 30 days before the beginning of produc-
tion, which for diesel engines is the beginning of the calendar
year, this means that certification must begin by October 1 of the
year prior to the applicable model year.
-------
All the elements of a diesel engine compliance schedule are
combined in Figure G-4. Shown in that Figure are two timetables.
The first is based upon establishing test correlation on eddy
current dynamometers, while the second is based upon the procure-
ment of electric motoring dynamometers. The first schedule shows
that if adequate correlation were demonstrated, certification to
the transient procedure should be possible for model year 1984
using eddy current dynamometers. However, if the attempt to
correlate were not successful, the manufacturer would have incurred
an approximate 1 year delay and would then have to follow the
schedule for procurement of electric dynamometers. That would
delay certification until 1985. Therefore, the EPA technical staff
does not believe any manufacturer would choose this course.
If a situation should arise where the need to estabish correlation
was elimated, then one year would be gained on this schedule and
certification for 1983 would be possible.
The second timetable, that for procurement of electric motor-
ing dynamometers, indicates that certification for model year 1984
would be feasible with an approximately 7-month "cushion." This
time could be used by the manufacturer to increase development
time or cover unforseen delays. The second timetable also allows
leadtime for tooling of any major engine changes before start of
production. Estimates for tooling leadtimes were given by GM and
IH as 24 and 22 months, respectively. The timetable of Figure G-4
would allow at least six months of development before long leadtime
tooling commitments would be required. Additional time would be
available through use of the advance purchased ceil. For those
engine families requiring little or no development work, certifi-
cation for 1983 would be possible.
4. Staff Recommendations
The EPA technical staff recommends delaying implementation of
these regulations until 1984. For the first year of implementation
the staff also recommends the use of an optional 13-mode standard
for diesel engines.
This analysis has revised the EPA projected compliance time-
table based upon manufacturers comments and other new data avail-
able to the EPA staff. The results indicate that for gasoline-
engines there is some possibility that certification could be
accomplished for the 1983 model year, but that the risk of missing
that deadline would be high. For diesel engines, some engine
families could meet a 1983 certification deadline, but those
requiring significant emission reductions could not. For diesel
engines there is a possibility that eddy-current type dynamometers
might be retained, at considerable cost savings. This possib-
ility is most likely to be realized if diesel manufacturers were
allowed to certify to an optional 13-mode standard for 1984. This
standard would be derived to, as far as it is possible, give
similar reductions Co those required by the transient procedure.
-------
H. Issue - Economic Impact
1. Summary of the Issue
The U.S. EPA has proposed a comprehensive control strategy for
1983 and later model year heavy-duty engines.
For both gasoline-fueled and diesel heavy-duty engines.
this strategy includes a new test procedure, more stringent
HC and CO emission standards, a new useful life definition, a
revised durability testing program, allowable maintenance provis-
ions, parameter adjustment, selective enforcement auditing at a 10%
AQL, and an idle test with idle emission standa£d£ for HC and CO.
In addition, the control strategy also includes a diesel
crankcase emission standard for heavy-duty diesel engines.
In the proposal, the EPA technical staff estimated a per
engine cost of $204 for a gasoline-fueled heavy-duty engine with
discounted operating costs of $1,016.
For diesels, the average per engine cost will be $185; with no
expected increase in operating costs.
The rulemaking strategy as a whole was estimated to cost $2.54
billion dollars with $2.382 billion for gasoline-fueled heavy-duty
engines, and $158 million for heavy-duty diesel engines over the
first five-year period.
2. Summary of the Comments
The comments will be summarized according to the major compo-
nents of the rulemaking strategy. For both types of engines, the
costs of the following will be directly addressed: test proce-
dures, development and emission control hardware, certification,
allowable maintenance, useful life definition, parameter adjust-
ment and selective enforcement auditing. The cost of diesel
crankcase emission hardware will also be addressed separately.
To the extent possible, the comments in the costs area will be
addressed on a manufacturer-by-manufacturer basis, but flexibility
in format is a necessity.
A. Test Procedures
1. Gasoline-Fueled Engines
Comments on test procedure related costs were received only
from General Motors and Chrysler Corporation.
-------
General Motors
General Motors Heavy-Duty Gasoline Engine Transient
Emission Test Facility Detail of Estimated Cost
Description Cost
Dynamometer Room No. 1 (Four Single-Ended Dynos)
Constant Volume Sampler (2) and Installation $ 436,000
Dual Bag Emission Bench (2) and Installation 277,000
Additional Computer Facilities (HP-21 MXF) and Installation 295,000
Computer Interface Modification 78,000
Dynamometer and Controls Rework 18,000
(Improved response and heaters)
Miscellaneous Transducers, Propshafts and 91,000
Dynamometer Room Equipment
Rearrangement (Control Room Revisions, 222,000
Equipment Room Revisions, Dynamometer Room
Revisions, Relocate Equipment, etc.)
Subtotal $1,417,000
Dynamometer Room No. 2 (One Double-Ended Dyno)
Constant Volume Sampler (1) and Installation 218,000
Dual Bag Emission Bench (1) and Installation 133,000
Additional Computer Facilities (HP-21 MXF) and Installation 100,000
Computer Interface Modifications 42,000
New 300 H.P- Dynamometer (includes controls, 390,000
MG set, installation)
Miscellaneous Transducers and Propshafts and 33,000
Dynamometer Room Equipment
Rearrangement (remove existing dynamometer, 42,000
MG set, misc. rework)
Subtotal $ 958,000
Dynamometer Room No. 3 (Two Single-Ended Dynos)
Dynamometer and Controls Rework $ 18,000
(to allow computer control)
Computer Interface Modifications 36,000
Miscellaneous Transducers, Propshafts and 57,000
Dyanmometer Room Equipment
Test Cell Rearrangement and Rework 41,000
Subtotal $ 152,000
-------
Building Revisions for Additional Space Requirements
Remove and Construct New Walls 150,000
Rearrangement 60,000
New Mezzanine 125,000
Mezzanine Lift 10,000
Environmental Control for Engine Cold Soak 40,OOP
Subtotal $ 385,000
Outside Engineering Fees $ 100,000
Total $3,012,000
Chrysler Corporation
Total Cost Unit Cost
Two new test cells $2,320,000 $1,160,000
Two cells modified 314,000 157,000
Renovate development cells 140,000 70,000
Develop controller 30,000 7,500
Emission Test Equipment:
CVS systems 500,000 125,000
Emission carts 335,000 83,750
Bag carts 280,000 70,000
TOTAL $3,919,000
2. Diesel
Detailed cost estiamtes were received only from General
Motors. Caterpillar gave a percent error comment without revealing
specific figures. Cummins gave only a cost per engine range for
the test procedure and certification. Mack submitted a total cost
figure.
General Motors
Total Cost Unit Cost
Building Shell $ 5.2 M
Building - service equipment 7.49 M
Test Cell Support Systems
Dynmometers and Controllers:
Electric 2.1 M 300,000
Eddy Current 350 K 50,000
Equipment
CVS 2.1 M 150,000
Emission Instruments 2.1 M 150,000
Data Acquisition Systems 700 K 50,000
Quick Change Engine Mounts 350 K 25,000
Misc. Test Instruments and Equip. 420 K —
Special Contingency 5.2 M —
TOTAL $26,010,000
-------
Caterpillar's comments on certification testing facilities
stated that these would be 23.41% of their total expected expendi-
ture. Mack commented that their total test procedure costs were
expected to be about $5,216,000.
B. Development and Emission Control Hardware
1. Gasoline-Fueled Engines
Comments were received from General Motors, International
Harvester, and Chrysler -
General Motors
Component
Dual Monolithic Converters
Pipe Insulation and Chassis Shields
Large AIR Pump
AIR Modulation System
Decel Fuel Shut-off and Electronics
Temp Compensated Accelerator Pump
Mechanical/Vacuum Power Enrichment
Electric Choke
Adjustable WOT Fuel Curve
Tamper Resistant Idle
Tamper Resistant Choking Valve
Cold Start Ignition Modifier
Engine Modifications
Filler Neck, Gas Cap, Labels
Tamper Resistant Distributor
Assembly Tools
Assembly Labor
TOTAL
Cost
$220.00
30.00
25.67
7.00
10.90
2.09
1.95
4.50
.70
.50
6.50
4.00
13.00
2.50
.61
10.00
$340.00
Tooling Cost
$ 2M
?
11M
7
1.85M
1.2M
25 K
250 K
7.75M
204 K
6.1M
200 K
10.211M
600 K
$45 M*
* Total estimated.
Chrysler
Component
Cost Estimate
Underbody 3-Way Monolithic Catalyst $160.00
Close Couple Twin Monolithic Catalyst 165.00
Electronic Spark Advance and
Feedback Carburetor 95.00
Increased Thermal Protection, 2.0g SHED
Test Hardware, and Air Switching 75.00
EGR Maintenance Warning and OSAC Valve -30.00
TOTAL $465.00
-------
IHC
Component Cost Estimate
Catalytic Converters, Dual
Stainless Steel Mufflers and Pipes $250.00
Filler Neck Restrictions and Decals 50.00
EGR and Converter Warning Systems 40.00
Wiring, Brake Piping, Choke and Heat Shields 60.00
Capital and Engineering Development 70.00
TOTAL $470.00
2. Diesel
Diesel engine manufacturers as a group were unable or chose
not to comment specifically on the emission control hardware
necessary to meet the HC and CO emission standards.
General Motors - could not comment.
Cummins Engine Co. - range of costs from $0-$600 (variable injec-
tion timing).
Caterpillar - 59% of their total cost is approximately $275 per
engine.
International Harvester - no comments received.
Perkins - no specific comment.
Mack - no specific comment.
Iveco - no specific comments.
Daimler-Benz - no specific comment.
C. Certification
Estimates of actual increases in certification costs were
received from only two manufacturers, GM and Chrysler.
1. Gasoline
General Motors - GM provided cost estimates for the following
components of certification:
125-hour emission data engine test $23,000
Pre-production durability testing 122,400
125-hour test prior to durability fleet use 24,000
Emission testing during durability testing 43,000 (6 tests)
From this data GM detrained a certification cost of $4.00 per
engine.
-------
Chrysler - Estimated total cost at $2 million because they would
run twice as many durability data engines and emission data engines
prior to certification. Chryslers estimate also included in-use
durability testing costs.
2. Diesel
Only General Motors commented, and gave specific estimates for
each phase of the certification.
General Motors -
125-hour emission data engine test $24,000
Pre-production durability testing 124,400
125-hour test prior to durability fleet use 15,000
Emission testing during durability testing 24,000 (6 tests)
Caterpillar - stated that certification would be 2.35% of the total
cost.
Cummins - estimated preliminary deterioration factor testing costs
at $62,000 per engine family.
D. Useful Life Redefinition
All manufacturers stated that they would incur substantial
warranty cost increases with the new useful life definition.
Although these costs were not detailed, some commenters gave a
rough cost based on their interpretation of the regulations.
General Motors - $100 per engine.
Caterpillar - $89 per engine.
International Harvester - $150 per engine.
Chrysler - $200-400 per engine.
These increased costs arise primarily from expected warranty
claims associated with engine rebuild, catalyst failure, turbo-
chargers, injectors, carburetors, etc.
E. Parameter Adjustment
Only Ford, Chrysler, and General Motors commented specifically
on the costs of parameter adjustment.
Ford - commented that they envisioned tamper resistant idle speed,
spark timing, air/fuel ratio and choke bimetal adjustment. Costs
were estimated at $90 for electronic idle speed controllers and
$5-10 for a timing "lock."
-------
Chrysler - stated costs would be similar to those for light-duty
vehicles.
General Motors - expected to use tamper resistant idle, choking
valve, and distributor at a cost of about $7.50, with tooling costs
of $6.3 million.
F. Selective Enforcement Auditing (SEA)
Comments on the costs of SEA were divided into three major
areas; facilities and equipment, formal SEA testing costs, and
production line audit costs. In general, all manufacturers thought
EPA had underestimated these costs.
1. Test Facilities and Equipment
General Motors - CM presented an elaborate analysis of SEA facility
costs based on four possible configurations and a set of assump-
tions. Of the four configurations presented, the second seems to
be the closest to what EPA envisioned, so this configuration will
be discussed. The facility costed as Configuration 2 in the GM
comments assumes the following:
a) Service accumulation can be expected to need up to 125
hours and would take about six days to complete.
b) Service accumulation can be done on absorbing only dyna-
mometers and the test cells would need temperature control.
c) Emission testing would be done on d-c dynamometers and
the test cells would need temperature and humidity control.
d) The 12-36 hour cold soak would occur in the emission
testing test cell.
e) The facility would not be connected to an existing
exhaust emission test facility.
The facility costed includes:
a) Control room for dynamometer control consoles and emis-
sion equipment.
b) Electrical equipment rooms for m-g sets and constant
volume samplers (CVS).
c) Fire protection system for the test cells, control rooms
and electrical equipment rooms.
d) Fuel tank farm, fuel cells, and a distribution system.
-------
e) Engine and pallet storage area.
f) Parts and equipment storage area.
g) Shipping and receiving area.
h) Span gas storage, liquid nitrogen supply, and distribu-
tion system.
i) Master gas storage area.
j) Particulate filter weighing room.
k) Equipment maintenance and calibration room and standards
laboratory.
1) Technician laboratory.
m) Machine shop.
n) Engine prep and build area.
o) Offices.
p) Conference room.
q) Restrooms.
r) Locker rooms.
s) Lunch rooms.
A building of this type has been estimated to cost about $200 per
square foot....The size of each area is estimated as follows:
a) Test Area (includes penthouse) 85,000 sq. ft.
b) Support area (includes penthouse) 36,000 sq. ft.
c) Office area 15,000 sq. ft.
Total (building only) 135,000 sq. ft.
($27.2 million)
Test Equipment Configuration 2:
a) 14 dynamometers (with control consoles) $5,000,000
b) 6 emission benches and CVSs 3,000,000
c) 10 computers 1,000,000
d) Miscellaneous equipment 1,100,000
Total (equipment only) $10,100,000
/13
-------
All of these costs are summarized below:
Building costs
Equipment costs
$27,200,000
$10,100,000
$37.3 million
Chrysler - Chrysler estimated that the SEA regulations would cost
about $1.8 million in additional new facilities and equipment.
1 Dynamometer $190,000
2 Collectors 140,000
2 Diagnostic carts 240,000
2 Control consoles & software 200,000
1 Dynamometer cart receiver 60,000
8 Transporters 72,000
4 Gopower receivers 120,000
4 Gopower dynamometers 80,000
1 Gopower test cell 400,000
1 Main dynamometer cell renovation 100,000
1 Soak room (20' x 20') 140,000
$1,742,000
International Harvester - IHC estimated only that two additional
cells would be necessary in addition to those considered by EPA for
a total of six cells. The two additional cells would be required
with a 40% AQL. With a 10% AQL IHC stated they would need a total
of 11 emission cells plus 5 "run-in" cells at a total building and
equipment cost of $22,535,000.
Caterpillar - Caterpillar estimates SEA facility costs would be
5.29% of their total expenditures.
2. SEA Testing
The comments and cost estimates below cover the actual costs
expected with SEA testing. These are based on the number of audits
per manufacturer outlined in the draft regulatory analysis.
General Motors - GM estimated annual costs of $3.3 million for
personnel related expenses and $130,000 for expenses to ship
engines and components to the test facility. These costs can be
broken down as:
66 employees at $50,000 per annum;
$30,000 per year to ship diesel engines;
$100,000 per year to ship gasoline-fueled engines and compon-
ents .
-------
Chrysler - Chrysler estimated manpower and testing costs at
$320,000 per year.
Mack - Mack Trucks estimated that actual SEA audit costs would be
near $25,000 for manpower, supplies and spare parts.
Caterpillar - Caterpillar estimated SEA testing would be 2.94% of
their total cost.
3. Production Line Audits
The production line audit costs are those associated with a
10% AQL. These costs cover primarily additional testing of produc-
tion engines.
Mack - Mack Trucks estimated self auditing costs of $225,000 per
year, but gave no additional estimate for new facilities or break-
down of the cost of the audit tests.
Chrysler - Chrysler estimated that production audits would cost
$320,000 per year; but this figure includes formal EPA audits.
International Harvester - IHC estimated they would audit 6.6% of
each years production at $1000 per test. In addition, they stated
they would require 11 audit cells plus five run-in cells to perform
an audit test program.
Cummins - Cummins Engine Co. stated that production testing at a
moderate rate would cost $60 per engine produced but gave no cost
breakdown.
G. Diesel Crankcase Emissions
Although most commenters discussed the impact of the diesel
crankcase emission proposal, only two manufacturers gave any cost
estimates.
General Motors - GM estimated that $200,000 would be required to
develop a closed crankcase system.
Caterpillar - Caterpillar Tractor Co. estimated costs of $10 for
their 3208T, $135 for their 3306 and 3408 engines. No further cost
breakdown was included.
H. Allowable Maintenance
No comments were received on the specific costs of the allow-
able maintenance provisions, except for the warning systems mention-
ed under hardware.
-------
3. Analysis of the Comments
The EPA technical staff was disappointed in the quality and
quantity of the comments received on the economic impact of the
proposed regulations. Few commenters replied in the detail
requested in the NPRM and only minimal supporting cost breakdowns
were given by those who provided specific cost comments.
For those manufacturers who chose not to comment on a specific
item in the proposal the EPA technical staff has no option but to
assume that the cost estimates in the draft regulatory analysis
were correct for that manufacturer.
The costs which EPA will ultimately consider chargeable to
these regulations are only those which are necessary to meet the
requirements imposed by these rules and not necessarily the total
which the manufacturers stated they might spend.
This discussion section will be formatted in a manner similar
to that of the Summary of Comments section with flexibility in
format being used when appropriate.
A. Test Procedure
1. Gasoline-Fueled Engines
Since only General Motors and Chrysler gave specific comments,
only their cost figures will be addressed.
General Motors - The General Motors test procedure cost estimate is
outlined below with the EPA technical staff's analysis.
As can be seen from the description below, GM's cost estimates
can be divided into hardware expenditures, computer and computer-
related costs, and miscellaneous equipment, rearrangement, rework,
and construction. A discussion of these costs for each of the four
dynamometer rooms together with EPA's estimate of the cost is shown
below.
Cost
GM EPA
Dynamometer Room No. 1 Estimate Revision
(Four Single-Ended Dynos)
Constant Volume Sampler (2) and installation $ 436,000 300,000 ll
Dual Bag Emission Bench (2) and installation 277,000 133,000 2j
Additional Computer Facilities (HP-21 MXF) 295,000 0 _3/
and installation
Computer Interface Modifications 78,000
-------
Dynamometer and Controls Rework 18,000 —
(Improved response and heaters)
Miscellaneous Transducers, Propshafts and $ 91,000 0 4/
Dynamometer Room Equipment
Rearrangement (Control Room Revisions, 222,000 — 5/
Equipment Room Revisions, Dynamometer Room ~
Revisions, Relocate Equipment, etc.)
Subtotal $1,417,000 $751,000
JY Based on EPA and Chrysler data
_2/ GM should not require more than one emission bench per CVS.
_3_/ A realtime computer system not required by these regulations.
_4/ GM probably has this equipment already.
5J An absolute maximum, probably more than required by these
regulations.
Cost
GM EPA
Dynamometer Room No. 2 Estimate Revision
(One Double Ended Dyno)
Constant Volume Sampler (1) and installation 218,000 150,000 \J
Dual Bag Emission Bench (1) and installation 133,000 —
Additional Computer Facilities (HP-21 MXF) 100,000 0 2/
and installation
Computer Interface Modifications 42,000 —
New 300 H.P. Dynamometer (includes controls, 390,000 175,000 _3_/
MG set, installation)
Miscellaneous Transducers and Propshafts 33,000 0 4/
and Dynamometer Room Equipment
Rearrangement (remove existing dynamometer, 42,000 —
MG set, misc. rework)
Subtotal $ 958,000 $542,000
I/ Based on EPA and Chrysler data.
2/ A real-time system not required by these regulations.
3/ Hawker Siddley Electric Dynamometer and Ultra Electronics
Controller.
4/ It is likely GM already has most of this equipment.
Cost
GM EPA
Dynamometer Room No. 3 Estimate Revision
(Two Single-Ended Dynos)
Dynamometer Control Rework (to allow $ 18,000 —
computer control)
Computer Interface Modifications 36,000 —
(1?
-------
Miscellaneous Transducers, Propshafts and
Dynamometer Room Equipment
Test Cell Rearrangement and Rework
Subtotal
57,000
41,000
0 I/
$ 152,000 $95,000
I/ It is likely GM already has this equipment.
Cost
Building Revision for Additional
Space Requirements
Remove and Construct New Walls
Rearrangement
New Mezzanine
Mezzanine Lift
Environmental Control for Engine
Cold Soak
Subtotal
Outside Engineering Fees
Total
GM EPA
Estimate Revision
$ 150,000
60,000
125,000
10,000
40,000
0 IJ
0 2/
$ 385,000 $285,000
$ 100,000 0 3/
$3,012,000 $1,673,000 4/
_!_/ This cost is probably already included in rearrangement costs above.
2J Not required,.due to forced cool down provisions.
_3_/ This cost is probably already included in construction cost above.
4_/ EPA considers this the absolute maximum GM would spend.
Based on the analysis above EPA liberally estimates GM's
facility costs due to these regulations to be about $1,673,000.
This exceeds EPA's original estimate by $421,000 due primarily to
unanticipated construction costs. If these construction costs can
be minimized then the two costs estimates should be reasonably
close.
Chrysler Corporation - Chrysler estimated their total test proce-
dure related costs at $3,919,000. A breakdown of these costs shows
the purchase of two new test cells, the modification and renovation
of two development cells, and other test equipment.
In 1979 dollars, most of Chrysler test procedure related costs
seem reasonable. The $2.32 million for new test cells is higher
than anticipated by EPA, but Chrysler's comments indicated that
this was the actual cost of two new test cells which were comman-
deered from light-duty testing for use in heavy-duty development.
-------
The cell renovations and modifications expected are estimated
by Chrysler to cost $454,000 as compared to EPA's estimate of
$435,000.
Since Chrysler will certify only 2-3 engine families and will
use only two emission test dynamomters only two CVS systems, at a
cost of $250,000 will be necessary.
Assuming Chrysler's other emission test equipment costs are
necessary, Chryslers actual expected costs should be near
$3,419,000.
2. Diesel
General Motors estimated diesel test facility costs at
$26,010,000. The breakdown for this figure is provided again below
for the sake of the discussion. EPA originally estimated a total
cost of $11,928,000. The analysis below will assume nine engine
families.
Building Shell
Building - service equipment
Test Cell support systems
Dynamometers and Controllers
Electric (7)
Eddy Current (7)
Equipment
CVS
Emissions Instruments
Data Acquisition Systems
Quick Change Engine Mount
Misc. Test Instruments & Equipment
Special Contingency
Unit Cost
300,000
50,000
150,000
150,000
50,000
25,000
Total Cost
5.2M
7.49M
2.1M
350K
2.1M
2.1M
700K
350K
420K
5.2M
$26,010,000
General Motors began construction of a new diesel test lab
back in 1975, long before these regulations were proposed. EPA
cannot accept GM's estimate of any costs associated with new
building structures or cell support systems nor can EPA accept GM's
cost estimates for new eddy current dynamometers. As stated
previously, any economic impact analysis should consider only
incremental cost increases and most certainly not past expen-
ditures.
With nine engine families the EPA technical staff believes GM
will require five to seven remodeled eddy current dynamometers,
motors, and controllers, for development. EPA concurs with GM's
estimate of seven new DC electric dynamometers and controllers for
emission testing.
-------
EPA believes that GM will need only one CVS system for every
two of its emission certification cells and thus will require only
four CVS systems. GM may require some minor facility modifications
in association with CVS installation and some minor modifications
to the pre-production development dynamometers (eddy current).
EPA Revision of General Motors Cost Estimates
Cost Per
Item Cell Total
Building Shell 0 0
Building Cell Support Systems 0 0
- Dynamometers & Controllers
Electric (7) 300K 2.1M (7)
Eddy Current (7) 0 0
Eddy Current Motors 85K 595K (7)
and Controllers (7) _!_/
- Equipment
CVS 150 ..000 600K (4)
Minor Facility Modifications _!/ 80,000 320K
Emissions Instruments^/ 150,000 1.8M
Data Acquisition Systems 50,000 600K
Quick Change Engine Mount 3_/ — 0
Misc. Test Instruments — 0
and Equipment 4/
Special Contingency 5/ — 0
TOTAL 6~015M
y EPA estimate.
2j GM probably already has this equipment, but lacking other
data the EPA technical staff will assume these are necessary.
_3_/ Not necessary to comply with these regulations due to forced
cool down provisions.
4_/ It is very likely that General Motors already has most of
this equipment.
5/ Not supported in the manufacturers comments.
Since no other diesel manufacturer commented in a detailed
enough manner to allow an analysis the EPA technical staff assumes
that the cost estimaes in the draft regulatory analysis were
correct.
-------
b. Development and Emission Control Hardware
1 Gasoline Engines
The EPA technical staff appreciates the manufacturers comments
on the emission control hardware which they believe will be required
to meet the standards. EPA is not in the business of dictating
what emission control strategy manufacturers may use or what they
may charge for their development and hardware, but is interested in
determining the approximate cost and nature of the hardware which
will be required to meet the revised standards.
Type of Hardware - Based on EPA's own technical analysis and the
information provided by the commenters. the following emission
control hardware seems to be necessary to meet the revised emission
standards and other provisions of the proposal.
Dual Monolithic Oxidation Catalysts
Chassis Heat Shields (2)
Stainless Steel Exhaust (2)
Engine Modifications to allow Unleaded Fuel Usage
Catalyst Durability Hardware
Unleaded Filler Restriction and Decal
Parameter Adjustment Modifications
Air Pump Improvements
Air Modulation
Electronic Ignition
EGR
This system is similar to that outlined by General Motors
except for exhaust pipe "insulation" and minor carburetor modifi-
cations which may be required.
In comparison to Chrysler's submittal. the major differences
lie in the catalysts and the evaporative emissions hardware which
Chrysler included. The EPA technical staff does not foresee the
feasability or need for start catalysts to handle cold start
emissions, nor does it see the need for the use of a 3-way catalyst
and feedback carburetor over an oxidation catalyst. The proposed
NOx standard is not stringent enough to require a 3-way catalyst.
The evaporative emissions hardware should not have been included in
a system to meet the exhaust emission standards.
International Harvester foresees the need for EGR and con-
verter warning systems which EPA does not believe will be necessary
even with the longer useful life expected.
The cost of a system which the EPA technical staff believes
will be necessary to meet the emission standards is outlined in the
regulatory analysis which supports this rulemaking action. The
-------
control strategy ultimately chosen by each manufacturer and the
price which it will ultimately charge for this hardware is cont-
rolled by each manufacturer.
2. Diesel
The fact that most diesel manufacturers chose not to comment
on the cost of meeting the emission standards indicates that they
could not comment meaningfully because of their lack of transient
testing capability and, thus, they could not be certain of the
magnitude of the task.
An EPA analysis of the diesel transient test data currently
available to EPA (see Test Procedure Issue) shows that 14 of the
current engine families already meet the target emission levels for
HC and CO. These engine families represent 36.3 percent of 1979
projected sales.
An additional 14 of the families (38 percent of sales)
are within easy range of meeting the target reductions with only
minimal changes to injectors or other calibrations.
The final ten engine families will require some work to meet
the target emission levels. This would include the minor changes
to injectors and calibrations discussed above plus possibly combus-
tion changer redesign, turbocharging and after cooling, pre-chamber
injection, variable injection timing, and the addition of EGR on
some diesel models.
The engines which appear to require the largest emission
reductions seem to have one or more of the following characteris-
tics:
High-rated speed, low-rated BHP.
Naturally aspirated.
Two-stroke engines.
- High surface-to-volume ratio.
Larger than average sac volume.
Turbocharged but not intercooled or aftercooled
The actual average per engine cost which EPA estimates for
diesel engines can be found in the regulatory analysis. This cost
will primarily be applied toward the ten families which will need
the most work since 74 percent will be able to meet the target
emission levels with little or no development work.
C. Certification Costs
1. Gasoline
General Motors provided certification cost estimates in the
same category as EPA's but these were unsupported by any further
-------
breakdown in detail. Chrysler provided no detail in their cost
estimate.
EPA has identified two major areas in certification:
a) Pre-production durability testing (Deterioration Factor
Assessment)
b) 125-hour emission data engine test
(a) Pre-Production Durability Testing (Preliminary Deterioration
Factor Assessment)
Assume the current EPA procedure is used, and allow 10 percent
of the manpower cost to cover overhead and miscellaneous.
Thus, the costs of this program would be:
Set up
Map
Test
Remove
16 phr
8
6
4
34 x 14 = 476 person hours
Service Accumulation +3000 person hours
3476 phr x $30/hr = $104,280
20 # x 1500 hr. x 1 gal, x $1.00 = $ 4,934
hr- 6.08# gal.
Engine Cost Estimate = $ 2,000
Certification Overhead and Miscellaneous = $ 10,000
Total = $122,000
(b) 125-hour Emission Data Engine Test
Set up 16 phr
Map ; 8
Test 6
Remove 4
Service Accumulation 250
284 phr at $30/hr. = $ 8,520
20 # x 125 hr- x 1 gal, x $1.00 = $ 411
hr. 6.08# gal.
Engine Cost Estimate = $ 2,000
Certification Overhead and Miscellaneous = $ 2,000
Total = $ 13,000
-------
Using these figures, EPA concludes that the following are
reasonable costs for certification testing.
Pre-production durability testing $122,000
125-hour emission data engine test $ 13,000
2. Diesel
General Motors estimated diesel certification costs in the
same categories as EPA. The categories are the same as stated
above for gasoline-fueled engines. EPA's estimates for these costs
are different than those cited by General Motors.
General Motors' cost estimates are not supported with any
detailed breakdown but EPA expects GM anticipated higher manpower
costs during service accumulation. EPA's cost breakdown for each
category is given below.
(a) Pre-Production Durabilty Testing (Deterioration Factor Assess.)
Assume the current . EPA procedure is followed, and allow 10
percent of the manpower costs to cover overhead and miscellaneous.
The costs of this program would be:
Set up 20 phr
Map 10
Test 8
Remove 6_
44 phr per test
10 Tests 440 phr
Service Accumulation +2000 phr
2440 phr at $30/hr. = $ 73,200
Estimated Engine Cost = $ 7,000
Certification Overhead and Miscellaneous = $ 7,300
Total = $106,000
(b) 125 hour Emission Data Engine Test
Set up 20 phr
Map 10
Test 8
Remove 6
Service Accumulation 250
294 phr at $30/hr. = $ 8,820
-------
140 * x 125 hr x l gal" X Mi - $ 2 220
hr. ° X 7.09# gal. * ^'2ji0
Estimated Engine Cost = $ 7,000
Certification Overhead and Miscellaneous = $ 1,800
Total = $ 20,000
Using these figures, EPA concludes that the following are
reasonable costs for certification testing for diesel engines.
Pre-production durability testing $123,000
125-hour emission-data engine tests 20,000
D. Useful Life Redefinition
The EPA technical staff has no basis by which to analyze the
manufacturers' comments on their costs associated with the redefi-
nition of useful life. The manufacturers all expected most if not
all of these costs to be associated with warranty claims but did
not provide any detailed analysis of where these costs would be
incurred. Since the proposed regulations are not warranty regula-
tions, warranty related costs cannot be included in this analysis.
The EPA technical staff does expect the useful life redefi-
nition to affect two other components of this proposal: emission
standards and hardware.
With a longer useful life, the target levels for HC, CO, and
NOx will have to be lower thus requiring more research and develop-
ment to meet the target levels.
Since the emission standards must be achieved for the full
useful life, the hardware which is required to meet the standard
must be more durable in addition to being more efficient. This
will require increased costs.
E. Parameter Adjustment
Ford stated that the parameter adjustment provisions would
cost $90-$100 per engine. General Motors stated costs which sum to
$7.50 in hardware plus $6.3 million in tooling costs. Assuming 5
year sales of 2 million GM costs become $10-$11 per engine. An
interpretation of Chryslers statement that "...the impact on
Chrysler's heavy-duty vehicles will be the same as |or our light-
duty vehicles," would give costs of about $3-$8 per engine.
* Parameter Adjustment Regulations - Summary and Analysis o'f
Comments, October 2, 1978.
-------
Based on the comments from GM and Chrysler; the cost of the
minor changes necessary to adopt the parameter adjustment regula-
tions is in the range of $3-$ll.
Practically all light-duty truck engines are also available as
heavy-duty engines. Since light-duty truck parameter adjustment
regulations are already in place, the cost of implementing heavy-
duty parameter adjustment should be limited to making the necessary
hardware modifications to all engines produced for heavy-duty
vehicles.
This cost, with profit, would surely not exceed $5.00 per
engine and without profit would probably be nearer $3.00 - $4.00.
Only the incremental changed hardware would be required. The
engineering and production development will already be done in
association with light-duty trucks.
G. Selective Enforcement Auditing (SEA)
When analyzing comments on the costs of SEA it is important to
remember the goal which the industry must achieve. In the case of
SEA this goal is to pass a formal EPA audit at or below the esta-
blished Acceptable Quality Level (AQL).
The goal of passing these audits can be achieved through at
least three means: 1) research and development aimed at reaching
lower target emission levels; 2) production line quality control
procedures, and; 3) post production emissions testing (self
audits). The degree to which these three methods must be imple-
mented depends on the stringency of the standard, the stringency of
the AQL, and the degree of confidence the manufacturer desires in
it ability to pass a formal EPA audit.
EPArs analysis of the comments on this issue will be based on
the factors described above.
1 . Test Facilities and Equipment
The number of test facicilites and the amount of accompanying
test equipment necessary for SEA will be dictated by either the
formal EPA SEA audit rate or the manufacturers own production line
auditing program.
a. Facilities and Equipment for Formal EPA SEA
EPA's formal SEA testing requirement is two actual audit tests
per day. If the manufacturer's sales are less than 30,000 per year
only one test per day is required. Based on a statistical analy-
sis, an average sample size of twelve engines per audit is expec-
-------
ted.* This assumes a 10 percent non-compliance rate in the con-
figuration being audited.
The analysis to determine the number of facilities required
will assume the following:
1. Engine installation and removal takes 4 hours each time
(2 hours to install and 2 hours to remove).
2. Engine "break-in" will be conducted on eddy current
dynamometers but the engines will undergo emissions
testing on a DC electric dyanmometer.
3. Forced cool down would take 2 hours, in place of a
natural soak of 12-36 hours and would occur in the
"emission testing" cell after engine mapping.
4_ Once formal SEA emissions testing has begun 2 tests per
day must be completed if sales exceed 30,000 per year,
otherwise, only one test per day is required.
5. Engine break-in, including installation, service accumu-
lation, and removal uses 48 hours in a
"break-in" cell.
6. Engine break-in takes 24 hours but is not conducted for
more than 16 hours per day.
7. Emissions testing, including installation, mapping,
and removal takes 11 hours.
Under these assumptions 1 "break-in" cell could provide 3
engines per week, and 2 cells could provide an average of 6 engines
per week for emissions testing. Therefore 4 break-in cells would
provide the average of 2 engines per day which would be necessary
to comply with EPA's formal audit requirements. For smaller volume
manufacturers only 2 break-in cells will be necessary.
a. Test Facilities for Self Audits
Production line auditing test facility and equipment needs
would depend on the self-audit rate chosen by the manufacturer and
the manufacturers annual production volume.
The current industry-wide light-duty vehicle self-audit rate
at a 40 percent AQL is approximately 0.2 percent. A self-audit
rate of 0.6 percent seems reasonable at a 10 percent AQL for
* Analytical Development of Sampling Plans for Selective En-
forcement Auditing, Sylvia G. Leaver, MSED, December, 1978.
-------
heavy-duty engines with new emission standards and a new SEA
program. The self audit rate is completely at the manufacturers
discretion, so the 0.6 percent figure assumed by EPA is admittedly
subjective but probably a little high. It is very likely that the
self audit rate will drop quite substantially in future years as
the manufacturers gain more confidence in their SEA compliance
efforts and produce engines to meet the same emission standards for
several years.
A self-audit rate of 0.6 percent means that 3 engines in every
500 will be tested. Using the facilities required for formal SEA
audits and a total break-in period of 16 hours, large-volume
manufacturers could test as many as 1000 engines per year and
small-volume manufacturers could test as many as 500 per year. At
a 0.6 percent audit rate, 1000 engines would support production of
166,000 per annum and 500 engines would support production of as
many as 83,000 per year.
The tables below give EPA's estimates for manufacturer's
facility needs based on the analysis presented above and the sales
projections prepared by EPA for the regulatory analysis:
SEA Facilities - HP Gasoline-Fueled
Break-in Cells Emission Cells
GM* 4 2
Ford* 4 2
IHC* 4 2
Chrysler* 4 2
SEA Facilities - HP Diesel
Break-in Cells Emission Cells
GM* 4 2
Caterpillar* 4 2
Cummins* 4 2
Mack* 4 2
IHC* 2 1
Others(9)* 2 each 1 each
* Facility needs dictated by formal EPA audit requirements.
As shown in the tables above, none of the heavy-duty manu-
facturers would require more test cells than those necessary
for the formal EPA SEA audits.
-------
The number of test cells estimated for each manufacturer is
very dependent on the assumptions used in this analysis. If the
break-in period or self audit rate were changed then each manufac-
turer's facility needs would also change. The EPA technical
believes the facility estimates presented above to be reasonable
and very close to the actual manufacturer needs.
c. Facility Costs
Having now estimated the number of complete facilities re-
quired by each manufacturer the task becomes to estimate the cost
of these facilities.
These costs will be estimated by assuming heavy-duty gasoline-
fueled engine manufacturers buy all new equipment and test facili-
ties. This assumption is extremely conservative.
For heavy-duty diesel manufacturers, EPA assumes that all new
facilities and equipment will be purchased with the following
exceptions. EPA assumes that all of the larger volume manufac-
turers will use the eddy current dynamometers removed from their
certification facilities as the break-in dynamometers for SEA.
Secondly, EPA believes that no small volume manufacturer would buy
SEA facilities but would use certification facilities if an audit
were conducted. Thus, EPA assumes one half the costs of the
certification facilities would be attributable to SEA.
In all cases, it will be assumed that these facilities would
be placed near the production facilities to minimize production
self audit costs and to allow the common use of other support
facilities.
This analysis will assume one CVS per emission test cell and
uses estimates provided by vendors, manufacturers, and EPA exper-
ience. Manufacturers would probably buy one CVS per emission test
cell to assure their ability to meet formal audit requirements.
Cost per Complete Emission Test Site with CVS:
Gas & Diesel
Diesel Gas
D.C. electric dynamometer _!_/ $120K
Dynamometer installation 2/ 20K
Computer control 3/ 35K
CVS and installation 2j 180K (150K)
Analytical system 4/ 150K
Computer interface 4/ 40K
Structure and other support 500K
functions and hardware
(2500 sq.ft. at $200 4/
per sq.ft.) 2j
$1.045M $1.015M
-------
_!_/ Hawker Siddeley Electric Dynamometer.
Y/ EPA estimate.
3/ Ultra Electronics Incorporated.
4~/ GM estimate.
Cost per Break-in Site: Diesel - Large Volume Manufacturer
D_iese_l_
Dynamometer installation \J $ 5K
Receiver 2/ ~ 30K
Transporters 2J 25K
Dynamometer Control _3_/ 20K
Structure and Other Support 400K
Functions (2000 sq.ft. at $200
per sq.ft.) _3_/
$48 OK
$530K (if new dynamometer required)
\J Conversation with Eaton Inc.
21 Chrysler estimate.
_3_/ EPA estimate.
Cost per Break-in Site: Gasoline-Fueled
The cost per break-in site for gasoline-fueled engine manufac-
turers would be the same as for diesels except a dynmometer esti-
mated by GM and EPA to cost at most $50K would also be necessary.
So the cost per break-in site would be $530K.
Small volume heavy-duty diesel manufacturers would use certi-
fication facilities for SEA and the only costs incurred would be
one half the certification facility costs. These certification
costs are based on a worst case assumption that each manufacturer
would have to purchase or'modify all equipment.
In closing this discussion of facility and equipment costs, a
discussion of the manufacturers estimates in this area would be
useful.
General Motors
GM estimated total equipment and facility costs of $37.3
million dollars and EPA estimated a cost of about $8.4 million
dollars. The basic difference between the two estimates lies in
the underlying assumptions in three areas: break-in period length,
cold soak requirements, and support facilities. The EPA technical
staff believes that a 16-hour "break-in" period is much more
realistic than a 125-hour "break-in" period. Manufacturers will
try to minimize their "break-in" periods to protect the resale"
-------
value of these engines. A forced cool down would negate the need
for a long soak period and facilities for this soak. Finally, EPA
believes that support facilities GM claims are necessary would, in
fact, be already available in other co-located facilities. Support
facilities which are necessary should be available in the cost for
32,500 square feet allowed in this analysis ($6.5 million).
Chrysler
The estimate of Chrysler and EPA differ because Chrysler
assumed the use of current facilities with some additions and
modifications. EPA accepts all of Chryslers cost estimates except
the soak room at $140,000. A forced cool down would eliminate the
need for this soak room. Therefore, EPA modifies Chryslers esti-
mate to $1.602 million dollars.
International Harvester
IHC estimated they would need 11 audit cells plus five run-in
cells to perform an audit test program at a 10 percent AQL. EPA
estimates IHC could meet their needs with two emission testing
cells and four "run in" cells. The apparent discrepancy between
these figures is due to a very high self audit rate anticipated by
IHC (6.6%) plus only a 232 day per year work period. If IHC were
to assume a 300 day work period they would require seven emission
cells and 14 run-in sites cells to audit their production at 6.6
percent. At a 0.6 percent audit rate, which EPA assumed and
believes is probably high , the number of extra facilities would
decrease to zero. Therefore, based on EPA's analysis the number of
audit cells required by IHC could ultimately be dictated by the
formal EPA SEA test rate.
2. Formal SEA Testing
The actual formal SEA testing costs are a function of a number
of different elements which will be discussed below. Costs will be
computed on a cost per formal audit basis.
a. Selection and Transport - For gasoline-fueled heavy-duty
engines both the engines and the dressup components must
be shipped to the point where formal SEA testing occurs.
In some cases this will be at the vehicle assembly point
but it may not be in all cases. The EPA technical staff
assumes that a round trip cost of $400 would amply cover
these items on a per engine basis. On a per manufacturer
basis this is conservatively high.
For diesel engines, no dress up components are necessary
for shipment in addition to the engines, so only $30
round trip selection and transport costs is necessary.
-------
In the majority of the cases EPA believes that SEA
testing will occur at the engine plant because many
diesel manufacturers do not also make vehicles for their
engines.
b. Break-in Costs - For this analysis we have assumed that
manufacturers will use a 24 hour break-in period in a
procedure similar to that used on an durability or
emission data engine.
For gasoline fueled engines this assumes an average fuel usage
of:
hr. 6.08 gal.
For diesel engines this assumes a fuel usage on the average
of:
24 hr. x x —^-i x —: = $427
hr. 7.09 gal.
In addition to the fuel, each engine would incur break-in
costs of about $600 in association with manpower. These costs are
attributable to break-in (24 hr) , set-up (6 hr) and removal (2
hr). The $600 figure assumes one technician per every two engines
during break-in. So final break-in costs become $679 for gaso-
line-fueled engines and $1,027 for diesel engines.
Emission Testing Cost
The emissions testing costs are primarily associated with
manpower. The manpower required can be roughly divided as shown
below:
Gasoline* Diesel*
Set up 6 phr
Map 4
Test 6
Remove 2
Cool Down 0
6 phr
4
7
2
0
18 person hours (diesel 19)
* These figures assume more experienced technicians and a more
efficient procedure than that used initially during certification.
At thirty dollars per hour, the manpower cost is $540 for-
gasoline engines and $570 for diesel engines. Fuel cost is $5 for
diesels and $10 for gasoline.
a/1
-------
Engines Per Audit
Based on the statistical analysis by Sylvia Leaver of MSED the
average sample number would be 12 engines until a pass/fail deci-
sion is made. This assumes 10 percent of the engines are in non-
compliance.
Miscellaneous
Some small cost per test expenses for overhead, supervision,
electricity, water, air conditioning and other items are inevitable
but the actual amount is difficult to assess. EPA shall assume a
per test cost of 10 percent of the manpower costs or about $115 per
test.
To summarize:
Miscellaneous Cost
Engine
Cost Selection &
Transport t Break-in
Engine Engine
HD Gas = $400 +
HD Diesel = $30
Finally:
Cost Cost
Audit Engine
$679 +
+ $1027
Engines
Audit
Engine
$550 + $1
+ $575 +
= $1745 x
Emission Testing
Engine
15 = $1745
$115 = $1745
12 = $21,000 pe
In closing this discussion of SEA testing costs, a comparison
of these cost estimates with those of the commenters would be
appropriate. However, most of the manufacturers, except Mack
Trucks, estimated total costs of both self audits and SEA so direct
comparison is not possible.
Using EPA's estimate of Mack sales in the mid 1980's and an
audit rate of one audit per 30,000 engines sold, Mack would be
subject to 1 to 2 formal EPA SEA audits per year. This cost would
be $21,000 to 42,000 per year. Mack estimated costs of $25,000 per
year.
3. Production Line Audits (Self Audits)
Production line audit costs on a manufacturer-to-manufacturer
basis are very difficult to estimate. Some manufacturers may audit
as much as 1 percent of their production and some may do little more
than spot checking.
The EPA technical staff believes that on an industry-wide
basis a self-audit rate of about 0.6 percent will prevail in the-
first year of production. However, as the manufacturers gain more
2/3
-------
experience with SEA and produce engines to meet the same standards
for several years, the audit rate should drop to 0.4 percent in
three years for HD gasoline and 0.4 percent or less for HD diesel.
In addition, because self-audits may be conducted on new or
existing facilities at engine or vehicle assembly points and are a
manufacturer's tool designed primarily to meet the manufacturers'
needs, the costs of self-audits should be substantially less than
those for formal audits.
Because self-audits will be conducted near engine or vehicle
assembly points, shipping and handling costs for engines and
components should be small, say $30 (one person-hour). The other
major difference in cost would be a decreased "break-in" period.
EPA believes the manufacturers would minimize their "break-in"
periods to protect the engine's resale value.
A reasonable criteria would be that each "break-in" dynamom-
eter provide one engine each 16-hour "day." Thus, using a set-up
and removal time of 4 cell hours, the break-in period becomes 12
hours.
Using these assumptions, the cost per engine of a self-audit
can be computed at $1,072 for gasoline-fueled engines and $1,274
for diesel engines. These computations are shown below. For
comparison, IHC expected self-audit costs of $1,000 per engine for
either gasoline-fueled or diesel engines.
Gasoline-Fueled Self-Audit Costs
Category Cost
Selection and Transport I/ $ 30
Break-In Manpower:
Set-up and remove • 6 phr.
Break-in 6 phr. 2/
12 phr. x" $30/phr = 360
Fuel:
12 hr x 20# x 1 gal x $1.00 = 39
hr 6.08# gal
Emissions Testing: 3/
Manpower 540
Fuel 10
Miscellaneous: 10 percent of manpower 93
TOTAL $1,072
_!_/ Assumes one person-hour.
2J Assumes one person to monitor two engines,
_3_/ Same as for formal SEA testing.
-------
Diesel Self-Audit Costs
Category Cost
Selection and Transport _!_/ $ 30
Break-In Manpower:
Set-Up and Removal 6 phr.
Break-In 6 phr. 2/
12 phr x $30/phr 360
Fuel: 12 hr x 140# x 1 gal x $0.90 = 213
hr 7.09 # gal
Emissions Testing _3_/
Manpower 570
Fuel 5
Miscellaneous; 10 percent of manpower 96
TOTAL $Tl74
_!_/ Assumes one person-hour.
_2/ Assumes one person to monitor two engines.
_3_/ Same as for formal SEA testing.
G. Diesel Crankcase Emissions
Although few comments were received on the cost of controlling
diesel crankcase emissions the comments from Caterpillar and
General Motors can be used to make a gross estimate of the costs of
a system.
First and foremost, the EPA technical staff agrees with
Caterpillars estimate of $10 to control diesel crankcase emissions
from non-turbocharged engines.
For turbocharged engines, the issue becomes more complex
because of'the high maintenance interval for turbochargers and the
tendency of "oily" crankcase emissions to foul the turbocharger.
Based on the fact that GM estimated R&D costs to close the diesel
crankcase at $200,000 and Caterpillar estimated the actual cost at
$135-$145, it seems likely that at least in the view of GM and
Caterpillar, controlling diesel crankcase emissions is technolog-
ically feasible. Indeed, diesel crankcase emissions have been
controlled in marine engines for years.
The question is not one of technological feasibility but one
of technolgical practicality. The system used on marine diesel
engines is not practical for diesel truck and bus engines.
The system envisioned by EPA would include a large pump,
pressure regulator, oil separator and some tubing which would
remove the crankcase emissions and reintroduce them after the
turbocharger. In addition to these two modifications EPA foresees.
obvious engine redesign and assembly costs which would be neces-
-------
sary. The EPA technical staff is convinced that controlling crank-
case emissions from turbocharged diesel engines is technically
feasible.
In total, EPA envisions a system costing $75 - $100 which
would cover materials, labor and recovery of engineering and
development costs.
H. Allowable Maintenance
Although no specific comments were received on the costs of
the allowable maintenance provisions, it is obvious that some work
will be necessary by vendors of catalysts, etc. to assure the
durability of their product. It is quite difficult to estimate
these costs, but in most, if not all cases, an approximation is
possible.
1) 100,000-mile catalyst - cost of increased noble metal
loading and larger catalyst volume plus engineering development.
2) 30,000-mile spark plug - may be technologically feasible
using current technology and unleaded gasoline.*
3) 200,000-mile turbocharger - technologically feasible with
little or no cost, already available from Caterpillar.*
4. Recommendat ions
The final cost figures used to compute the economic impact of
these regulations should be reevaluatad based on the manufacturers'
comment. The cost ultimately included should be that which is
required by these regulations and not necessarily that which the
manufacturer might spend. In some cases, more money may be spent
than is required, but this would not be done if there was not an
overall benefit to be derived, such as greater operating efficiency
or manpower savings.
See Allowable Maintenance Issue.
-------
I. Issue - Technological Feasibility
1. Summary of the Issue
EPA has proposed HC and CO emission standards representing 90
percent reductions from the uncontrolled baseline level as mea-
sured on the transient test procedure. Assuming certification on
the transient procedure, are the proposed standards technically
feasible?
2. Summary of the Comments
The comments received can be broken down into two areas of
primary, relevancy: the feasability of diesel engines complying
with the proposed HC standard of 1.30 grams/BHP-hr; and the
feasability of a catalytic converter for gasoline engines maintain-
ing effectiveness throughout 100,000 miles.
a. Diesel Engines
Diesel manufacturer's took issue with EPA's assertion in the
Draft Regulatory Analysis that no significant development work
would be required to allow diesels to comply with the proposed HC
standard.
Caterpillar claimed that 4 of their 11 currently certified
engine families would require "significant" design modifications.
Cummins also submitted data substantiating the fact that several of
their engine families were above the proposed standard. It was
argued that the majority of diesel engine families on the market
today would require some work to attain production emission targets
attributable to the stringency of the 10 percent AQL. Caterpillar
claimed that compliance for at least one engine family was impos-
sible. Cummins claimed that new and unproven control technologies
would be necessary.
Diesel manufacturer's also claimed that their ability to
evaluate technical feasibility was severely hampered by lack of
transient experience. All but Cummins (which has limited transient
capability) relied upon an assumed 13-mode/transient ratio derived
from SWRI's limited diesel baseline work in their feasibility
analyses. Mack declined to comment at all, however, claiming
inadequate experience and deprivation of due process.
Furthermore, Diesel manufacturers complained that facility
modifications necessitated by adoption of the transient procedure
would curtail already critically short leadtime, further aggrava-
ting the technical difficulties of compliance.
Future and unknown NOx and particulate standards were cited by
diesel manufacturers as limiting factors on future hydrocarbon
2./1
-------
control. As yet unknown, these factors contributed to the high
degree of uncertainty over eventual compliance.
Finally, no diesel manufacturers expressed concern over
compliance with the proposed GO and interim NOx standards. Cummins
flatly declared that CO and NOx compliance would be no problem.
b. Gasoline Engines
Gasoline engine manufacturers' harshly criticized any standard
requiring use of catalytic converters on heavy-duty trucks, claim-
ing it represented a "disservice to our customers", and unanimously
denied the feasibility of a 100,000-mile catalyst under heavy-duty
conditions. Use of lead-free fuel has the impact of decreasing
valve and engine durability. Data was submitted purportedly
illustrating state-of-the-art catalyst technology and alleging that
present catalyst technology is inadequate for assuring the proposed
emission reductions over the full useful life of the engine.
The major problem was characterized as one of catalyst dura-
bility in the heavy-duty environment. Light-duty catalysts operate
in less extreme thermal and vibrational environments, and are
required to last only half as long. Sustained exposure to the high
temperatures in the heavy-duty environment and exposure to pro-
longed motoring during engine-braking were characterized as fre-
quent and probable causes of cataclysmic catalyst system failure;
it was claimed that present technology catalysts cannot survive
under the full range of heavy-duty environments.
General Motors suggested on upward revision of the standards
to levels achievable by non-catalyst technology, and also claimed
EPA had no factual evidence that the proposed standard can be met
for the entire life of the engine.
In summary, all gasoline manufacturers except Chrysler
declared that a durable catalyst was impossible. Chrysler, how-
ever, maintained that a single catalyst system was indeed possible,
and in fact, under development and in production for sale in
California in 1980.
Like the diesel manufacturers, the gasoline engine industry
cited future NOx standards as contributory factors on limitations
on achievable reductions.
Ford maintained, aside from the durability question, that
attainment of a 15.5 g/BHP-hr CO standard on the proposed transient
test would be impossible, even with the best catalyst efficiences
observed to date.
27?
-------
3. Analysis of the Comments
a. Diesel Engines
It would be useful at this point to analyze the level of
technology present in today's diesel, and the magnitude of emission
reductions required by the proposed rules.
Table 1-1 presents actual and extrapolated transient emission
data for all heavy-duty diesel engines certified in 1979. Table
1-2 presents various engine design parameters and applied emission
control equipment. Table 1-3 presents a breakdown by manufacturer
of the percentage of total engine sales extrapolated to meet the
low mileage emission targets.
Extrapolated transient emissions for diesel engines for which
no transient data exist were derived using a transient/13-mode HC
ratio of 2.40. This ratio was derived in the Summary and Analysis
of Comments pertaining to the Test Procedure Issue, and represents
the average ratio of all engines tested on the transient procedure
to date. It cannot be emphasized too strongly, however, that no
predictive correlation between the two test procedures has been
found by any of the laboratories running transient tests. It
was testified by Caterpillar that this ratio tends to increase as
13—mode HC emissions decrease, i.e., the predictive value of the
13—mode, already dubious at HC levels around and above the stan-
dard, becomes worse at lower emission levels. Furthermore, it has
been demonstrated that the attainment of emission standards entails
designing to a given test procedure; for this reason steady-state
test procedures have been historically invalidated by the applica-
tion of technology to certify upon them. It should not be con-
strued that the use of an average ratio relating HC emissions
observed on both procedures constitutes admission of a correlation
existing today with current technologies. More importantly, the
application of future technologies to certify on the steady-state
test would result in emission results observed on future engines
exhibiting even less correlation. Lack of comprehensive transient
data is the sole rationale for use of an average ratio; that such a
ratio is the observed average of all engines actually tested is the
justification for its use. It is conceded that extrapolation of
transient emissions from certification 13—mode results on an engine
by engine basis entails some error, in some areas perhaps signifi-
cant. For the purposes of this technology assessment, however, it
is believed that use of the 2.40 ratio constitutes a reasonable
guess at the level of emissions observable on today's diesels. For
certification to stringent emission standards, however, a "guess-
timate" of representative emissions - as would be done if the
13—mode were retained - is technically unacceptable.
Using this 2.40 transient/13-mode ratio as an estimation.
technique, examination of Table 1-1 indicates that many engine
-------
Table 1-1
Anticipated Diesel BSHC Reductions
1979 13-Mode Anticipated HC
Engine Certification Transient HC Reduction Percent of
Mfr. Family HC Emission I/ Emission 2/ Necessary 3/ Company Sales
GM 4L-53T 0.83
GM 6L-71N 0.84
GM 8V-71N 0.82
GM 6V-71NC 1.27
GM 8V-71NC 0.80
GM 6V-92TA 0.58
GM 8V-71TA 0.51
GM 8V-92TA 0.50
GM 6L-71T 0.55
CEC 091 0.38
CEC 092A 0.32
CE.C 092C 0.26
CEC 093E 0.26
CEC 172A 1.20
CEC 172C 0.53
CEC 192B 0.30
CEC 193 0.38
CEC 221 0.79
CEC 222 0.69
IHC DT-466 0.64
IHC 9.0 Liter 1.38
IHC DTI-466B 0.56
Mack 8 0.31
Mack 9 0.76
Mack 10 0.12
Mack 11 0.58
Mack SIB 0.87
Cat 3 1.20
Cat 4 0.21
Cat 9 0.23
Cat 10 0.34
Cat 11 0.53
Cat 12 0.15
Cat 13 0.68
Cat 14 0.22
Cat 15 0.63
Cat 16 0.30
Cat 17 0.37
* Actual transient results.
I/ Includes deterioration factors
dure) .
11 Using transient/13-mode ratio
or actual transient data when
1.99
2.02
1.97
3.05
1.49*
1.17*
1.22
1.20
1.32
0.91
0.77
0.62
0.86*
2.88
1.27
0.72
0.91
1.90
1.66
1.54
3.31
0.81*
0.74
1.82
0.29
1.39
2.09
1.97*
0.50
0.55
0.82
1.27
0.36
1.63
0.53
1.51
0.72
0.89
(determined
of 2.40 (see
available .
3/ Based upon a transient production target of
1.10
1.13
1.08
2.16
0.60*
0.28*
0.33
0.31
0.43
0.02
0
0
0*
1.99
0.38
0
0.02
1.01
0.77
0.65
2.42
0*
0
0.93
0
0.50
1.20
1.08*
0
0
0
0.39
0
0.74
0
0.62
0
0
5.9%
13.4%
5.0%
3.5%
9.0%
31.8%
7.5%
19.2%
4.7%
0.5%
35.3%
15.7%
42.1%
1.1%
0.9%
0.1%
1.9%
0.1%
2.3%
85.7%
8.6%
5.7%
2.7%
53.9%
0.3%
41.6%
1.5%
57.4%
1.0%
0.0%
11.7%
5.2%
1.6%
8.0%
1.5%
1.1%
10.3%
1.9%
per 1979 proce-
Text)
0.89
J
g/BHP-HR.
-------
Table 1-2
1979 Dieael Engine Family Certification Data
Mfr.
CM
CM
CM
CM
CM
CM
CM
(ill
CM
CliC
CEC
CEC
etc
etc
CEC
CEC
cec
CEC
etc
me
UK;
HIC
Engine Engine
Engine Family Cycle
4I.-53T 2
6L-7IN 2
8V-7IN 2
6V-7IHC 2
8V-7INC 2
6V-92TA 2
8V-7ITA 2
8V-92TA 2
6L-7IT 2
091 (Nil 230,250) 4
092A 4
092C 4
093E(NTC 350,400) A
I72A(VTB 903,350) 4
I72C( ) 4
I92B (NT 450) 4
193 (KTb 600) 4
221 (V555) 4
222 (VT225) 4
DT-466 4
9. 0-Liter 4
DTI 466B A
Turbo-
Charger^
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Inter- After- Inj.
Cooled Couled Timing
a*. 10'
13*. 15*
12', 13*
to*, ir
12*.I3'
X 10'. 14'
X 12*. 14*
X 11', 13*
II*. 14*
19'
19'
X 19*
21'
18.5*
X 18.5'
22*
22'
16'
16'
X 15'
Cornpreafl ion
Ratio
18.7
18.7
18.7
18.7
18.7
17.0
17.0
17.0
17.0
15. a
15.0
14.3
16.6
15.5
14.5
17.0
16.2
16.3
19.1
16.3
Sac
Volume
0 7mm3
0. 5ouu~
O.Viuu,.
0. 7uuu.
0 . 7mul
i
0. 7uuu.
0 . 7mm
0.7uua3
3
0.9mm3
1
0 . 6mm.
0.6ouu.
0.6muu
0.6uin
0.6mm
3
0.32mmf
0.32mm
CID
212
426
568
426
568
552
568
736
426
855
855
855
903
903
1150
1150
555
555
466
551
466
No. of
Cylinders
4
6
8
6
a
6
a
8
6
6
6
6
8
6
6
8
8
6
8
6
Bated
Sjiee.1
2500
2300
2300
2100
2100
2100
2100
2100
2100
2100
1900
2100
2100
2100
2100
3300
3000
24-2600
2800
2600
Hated
BMP
155-170
184-239
248-316
160-190
230-270
300- 35
350
435
260-275
220- 40
293
400
350
275
450
600
216
225
210
180
210
Surface/
Volume Ratio
9.0
8.6
8.6
8.6
8.6
7.2
8.0
7.2
8.0
11.2
10.8
9.9
15.1
10.7
9.9
13.7
15. A
2.5
16.33
12.5
ECS*
FM
—
—
TO, SPL
TD, SPL
TD, SPL
TD, SPL
TD, SPL
TD, SPL
—
AFC, SPL
SPL, AFC
SPL
AFC. SPL
AFC, SPL
AFC, SPL
—
—
FM, SPL
PCV
FM, SPL
-------
Table 1-2 (Cont'd)
1979 Ilieaet Engine Family Certification Uala
Hti .
Hack
Hack
Hack
Mack
Mack
Cat
Cal
Cal
Cal
Cal
Cai
Cat
Cal
Cal
Cal
Cal
Engine Engine
Engine Family Cycle
8 (ETZ 1005)
9 (ENUT 676)
10 (ETAZ(B)I005A>
II (ETZ 675)
SIU (ETZ 4T71I)
3 (3208)
4 (3306)
9 (3406)
10 (3406)
II (3406)
12 (3408)
13 (3208)
14 (3306)
IS (3408)
16 (3406)
17 (3408)
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
Turbo- Inler- After-
Charger cooled cooled
X
H
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Inj.
Timing
18'
19*
17*
18'
15'
16*
12'
10*
10'
28*
II'
16'
8.5'
28*
26.5*
28'
Compression
Balio
15.0
14.99
17.0
17.0
15.5
16.5
17.5
16.5
16.5
14.5
15.3
16.5
17.5
14.5
14.5
14.5
Sac
Volume
0 32UUU1
0.35mm
0. 32mm
0.35mm
0.5mm J
0.24mm"1
25.4ium?
25.4mia^
22. lium
1 .Imm^
• IWlfl ^
0.24mm
25.4nuu
^
I.I mm
1 . 1mm
l.lm»
CID
998
672
998
672
475
636
638
893
893
893
1099
636
638
1099
893
1099
*
Cycle
8
6
8
6
6
8
6
6
6
6
8
a
6
8
6
a
Bated
Speed
2100
ia-2ioo
2100
2100
2400
2800
2200
2100
2100
2100
2100
2800
2200
2100
19-2100
2100
Baled Surface/
BI1P Volume Ratio
354
283-315
392
235
210
160-210
250
325
375
300- 25
450
200
245
400
350- 80
450
10.66
8.825
11.84
10.178
6.528
11.2
9.6
9.6
9.6
9.3
11.0
11.2
9.6
9.4
9.3
9.4
ECS*
SPL
SPL
SPL
SPL
SPL
AFKC,
AFRC.
AFBC,
AFRC,
AFBC.
EGB
AFBC,
AFBC,
AFRC,
AFRC.
SPL
SPL
SPL
SPL
SPL,
SPL
SPL
SPL
SPL
EGB = Exliaual Gau Recirculalion
HI = Fuel Modulator
TO = Throtlle Delay
AFC •* Air Fuel Control
AFHC " Air Fuel Ratio Control
PCV " Positive Crankcaue Ventilation
SPL - Smoke Puff Liaiitar
ECS ° Eiuiuuon Control Systdm
-------
Table 1-3
Percent of Total Company Sales (# Engines Sold)
Expected to Meet Proposed HC Standard Target*
Company Percent of Engine Sales
Caterpillar 28.1%**
Cummins 95.5%
Mack 3.0%
me 5.7%
Detroit Diesel 0%
* Based upon data from Table 1-1.
** This percentage jumps to 66 percent when Caterpillar
engine family 3, the high volume 3208, is discounted.
-------
families will require varying degrees of modifications to achieve
compliance. It also becomes apparent that greater than a third of
both 1979 engines (36%) and 1979 engine families (37%) should
already meet the 1984 production transient target levels (10%
AQL). Actual transient data presented in Table 1-4 more than
substantiate this claim. 72% of all Diesels tested on the tran-
sient procedure exhibited BSHC levels below the standard; 44% had
HC levels below the production targets.
The degree of present compliance is manufacturer specific, as
illustrated in Table 1-3. Cummins Engine Company is well ahead of
the rest of the diesel industry in terms of emission control.
Sixty percent of their engine families and 95.5 percent of their
projected 1979 unit sales should meet the 10 percent AQL production
target level (0.89 g/BHP-hr)* associated with the 1.3 g/BHP-hr
transient HC standard. In their written submission, Cummins
claimed that some development work was necessary on several of
their families to achieve total compliance; specifically noted were
efforts to develop a new technology, variable injector timing. Of
the four engine families needing work (see Table 1-1), the grossest
HC emitter - family 172A - has already been dropped for 1980.
Furthermore, Cummins identified several control strategies in their
written submission (Table 1-5) and included each strategy's known
effects on emissions.** However, Cummins' declined to submit
detailed data using the explanation that "detailed data on the
emissions capability of future technologies are of a proprietary
nature...." ECTD interprets this to mean that Cummins believes it
has technological and competitive advantages over the rest of the
industry; this interpretation is substantiated by Cummins' distinct
predominance in the field of emission control. Cummins major new
technology, variable injector timing, has already been incorporated
on at least one test engine (identified as engine D on Table 1-4);,
transient HC emission data for this engine (.86 g/BHP-hr) met the
production target emission level. In summary, Cummins is quite
close today to 100 percent compliance, has identified and is
familiar with several control strategies for the future, and is
capable today of transient testing and development work. Cummins
has testified that their avowed corporate policy for the last
several years has been to develop low emission engines; the facts
substantiate this claim, and imply that a consistent application of
technology can effectively reduce engine emission levels.
Caterpillar Tractor Company is a. distant second in terms of HC
* See Chapter 7 of the Regulatory Analysis, "Cost Effective-
ness ."
** Cummins qualified this submission, due primarily to the fact
that the test data were acquired on the 13-mode test, and not
necessarily relatable to the transient procedure.
*** For which actual transient data is available.
2.2
-------
Table 1-4
Actual Transient Diesel HC Emissions
Engine BSHC
1978 Caterpillar 3208 3.37
1976 Cummins NTC-350 0.68
1978 DDA 6V92T 0.78
1979 Cummins NTCC-350 0.86
1978 DDA 8V-71N 1.30
(#2 Fuel)
1978 DDA 8V-71N 1.49
(#1 Fuel)
1979 DDA 6V92TA 1.17
(#1 Fuel)
1979 DDA 6V92TA 1.09
(#2 Fuel)
Lab
SwRI
SwRI
SwRI
SwRI
SwRI
SwRI
SwRI
SwRI
Below
Standard?
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Below 10%
AQL Target?
No
Yes
Yes
Yes
No
No
No
No
1979 IHC DTI-466B
A*
B
C
D (w/variable injector
timing)
E
F
G
H
1979 Caterpillar 3208
0.81
0.99
0.76
0.72
0.86
1.33
2.22
1.25
0.55
1.96
Total Percent
Total
SwRI
Cummins
Cummins
Cummins
Cummins
Cummins
Cummins
Cummins
Cummins
Yes
Yes
Yes
Yes
- Yes
No
No
Yes
Yes
Caterpillar No
Below Standard:
Percent Below
72%
10% AQL Target:
Yes
No-
Yes
Yes
Yes
No
No
No
Yes
No
44%
Cummins' and Caterpillar data extracted from comments submitted.
2.2. sr
-------
Table 1-5
Trends in Emissions with Design Changes as Submitted
by Cummins (Based on 13-Mode Data)
Injection System Changes
Increase # Spray Hole
Increase Spray Hole Area
Increase Spray Hole Angle
Advance Injection Timing
Tighter Injector Setting
Higher Clearance Injectors
Faster Injection Profile
Piston Changes
Increase CR
Increase Piston to Head
Clearance
Increase Bowl Diameter
Deeper Bowl Piston
Miscellaneous
Smaller Turbine Casing
Aftercooling
Increase Intake Restriction
Increase Coolant Temp.
Increase Fuel Temp.
BSHC
decrease
decrease
7
decrease
decrease
increase
increase
decrease
decrease
decrease
decrease
Key: - No Change
? Indeterminate Change
-------
emission control, due primarily to the high HC emissions of their
high-sales volume 3208 (DINA Family 3) (representing 57.4 percent
of Caterpillar's 1979 projected unit sales). Caterpillar's
lengthy supplemental written submission on the technological
feasibility of the proposed HC standard addressed only the 3208 -
Family 3 engine. Caterpillar bluntly stated that "...the current
version of the 3208 does not comply with the proposed HC standard
...[and] available technology would not bring this engine into
compliance." ECTD takes issue with the latter assertion, primarily
on the basis of certification data from Caterpillar Family 13 -
also a 3208. Referring to Tables 1-1 and 1-2 and to all data
presented in the 1979 Part I submission to EPA's Certification
Division, the only difference between Families 3 and 13 is the
presence of EGR on 13. Note that 13-mode HC (on which the extra-
polated transient emissions are based) on Family 13 is almost 50
percent less than that on Family 3. Caterpillar, however, made no
mention of this fact in their written submission. Yet telephone
conversations with Caterpillar representatives revealed that the
EGR, traditionally employed for NOx control, was added to Family 13
also for the hydrocarbon control required by more stringent Cali-
fornia standards. (In Part I, Volume I of 1979 Certification
Records, p. 1A-IX, 6-1.0, Caterpillar states, "... Family 13 is
intended to comply with a standard more stringent than the...
(1979)...Federal Standard.") It should be noted, however, that
Family 13 also does not meet the 1984 production targets, but is
substantially closer, based upon the assumed transient/13-mode
ratio of 2.4. (Actual transient data for Family 3 reveals this
ratio to be only .1.48; were this to be consistent on the almost
identical Family 13 - for which no transient data is available -
then Family 13 would have transient HC emissions of 1.00 g/BHP-hr,
well under the proposed standard and only marginally exceeding the
production targets. This marginal difference could be eliminated
by the several techniques Caterpillar did mention.)* Given the
additional fact that neither 3208 Family is turbocharged or after-
cooled (the two major methods of HC control), ECTD can only con-
clude that the current version of the Family 3 3208 incorporates
little emission control technology. (Table 1-2 reveals that the
3208 is also the only Caterpillar engine which does not utilize an
air fuel ratio control system).** Nonapplication of technology in
the past is by no means a persuasive argument that technology is
inapplicable for the future, especially with several proven options
available, one of which - EGR, has already been applied on a
production basis (Family 13, 3500 unit projected 1979 sales). ECTD
concludes that the Family 3 3208 can achieve compliance; the
excessive emissions on the current version arise primarily from the
absence of control technology.
* Decrease in sac volume, increased compression ratio, timing
advance.
** AFRC limits the injection of fuel during accelerations to that
for which enough combustion air is present, thereby reducing smoke
and most likely transient HC emissions.
2J2.7
-------
Aside from the Family 3 3208, 70 percent of Caterpillar's
remaining engine families are anticipated to meet the production
targets.
Of the remaining engines which do not comply, highly inform-
ative comparisons can be made with those which do. For example,
compare Family 15 and Family 17 - both 3408 engines - on Table
1-2. Family 15 is anticipated to exceed the production targets
(Table 1-1), while 17 complies; Family 17, however, is virtually
identical to 15. The manifolds, valves, and injection systems are
the same. Different model turbochargers and air fuel ratio control
systems are used, but most importantly, Family 17 is equipped with
an aftercooler. Furthermore, both engines have sac volumes con-
siderably larger than those achieved on other engines, the re-
duction of which facilitates HC control. Therefore, no compelling
reason exists to presume the future noncompliance of Family 15.
Family 11, which does not comply, and Family 16, which does, are
also identical except for the fact that 16 is aftercooled.* Both
engines also have high sac volumes.
In summary, with the exceptions of Families 3 and 13, for
every Caterpillar engine family which is observed to emit high HC,
there is a virtually identical engine which does not. It is
construed to be significant that Caterpillar declined to comment
on any engine family except Family 3. As discussed above, Families
3 and 13 both exceed the targets, while addition of EGR to the
virtually identical 13 achieved a significant HC reduction.
Furthermore, both engines lack turbochargers and aftercoolers.
ECTD cannot dispute the contention that some redesign and develop-
ment work will be necessary for Caterpillar, and it is recognized
that Caterpillar is in a less than optimal situation by virtue of
the fact that their highest sales engine is probably their dir-
tiest. Yet it is also one of the least controlled engines on the
market, viable technologies exist today, and four years leadtime
for development is available.
International Harvester produces three diesel engine families,
one of which (DTI 466B) has been tested at SwRI over the proposed
transient test and easily complied with the production targets.
The 466B** is a higher technology (i.e., intercooled) version of
the DT 466 high-volume engine} and is primarily manufactured for
* Injector timing on Family 16 is more retarded than on Family
11 (26.5° BTC vs. 28" ETC). Since injection timing retard tends to
increase HC, ECTD must conclude that HC control on Family 16 is
even higher than that presumed at first glance. ** IHC has
declared that the DTI 466B is being modified for 1980 sales in
California, reflecting an increase in the stringency of the Cali-
fornia HC standard. Design changes will occur primarily in
the injector system and combustion chamber, and indicate that even
tighter HC control is achievable.
-------
sale in California. It is readily apparent, however, that the
application of this additional control technology to the 466 has
already been accomplished on a production basis. IHC's remaining
engine, the low-sales volume 9.0-liter version, is the engine on
which the greatest HC reductions will be required. It is also
neither turbocharged nor intercooled, and has a number of char-
acteristics common to high HC engines (high rated speed, low rated
horsepower, very high surface/volume ratio).
Mack's product line has yet to be tested over the transient
cycle, but extrapolated transient emissions in Table 1 predict
three of five engine families will exceed both the target and the
standard. Family 9 is one of these although already turbocharged
and aftercooled. Mack declined to comment on technological feas-
ibility claiming lack of data; this lack of data also constrains
this analysis. Furthermore, each of Mack's engine families are
different; this precludes comparison of emissions and applied
technologies between comparable engines. ECTD can only draw
inferences from the degree of compliance of other manufacturers,
the effectiveness of control strategies as evidenced by engines on
the market today, and the lack of evidence that Mack's engines are
fundamentally different in some way from other diesel engines.
Based upon this and the fact that two out of five* of their engine
families presently comply, no compelling evidence exists to indi-
cate that Mack's engines cannot be brought into compliance given
four years to do so.
Detroit Diesel is the only major diesel manufacturer whose
entire product line exceeds the HC target levels. Particularly
dirty are those engine families which are naturally aspirated. DBA
has indicated that families 4L-53T, 8V-71NC, 6L-71N, 8V-71N, and
6V-71NC will not be produced after 1982, however, and need not be
addressed in this analysis. Of the four remaining engines, all are
relatively close to the standard. Data submitted by Caterpillar**
indicated that a decrease in injector sac volume will decrease
hydrocarbon emissions. Figure 1-1 depicts this Caterpillar data
and corresponding sac volumes and emission levels for the four DDA
engines. Presuming a comparable emission trend with sac volume
reduction, three of the four DDA's can be brought below the target
level with a reduction in sac volume from the present . 7mm-1 to
.24 mm^- The 6L-71T can be brought relatively close by sac
volume reduction. Addition of an aftercooler, as already done on
the 6L-71TA, easily brings it within the low mileage targets. When
coupled with other injector optimizations presented in Table 1-5,
achieving compliance with these four families should be relatively
easy and inexpensive.
* This represents only 3 percent of their unit engine sales,
however.
** Table VI, "Effect of Nozzle Sac Volume...," of Supplementary
Statement to July 16, 1976 Public Hearings.
-------
In summary, a large percentage of diesel engines on the market
today meet the 1984 target EC levels for a 10 percent AQL. Fur-
thermore, proven strategies have been identified which will allow
the vast majority of the remaining engine families to comply; many
of these strategies have already been incorporated on production
engines already on the market. Compliance with the transient HC
standard with the 10 percent AQL for diesel engines can be accom-
plished by all manufacturers by 1984.
All diesel engines will easily comply with the 15.5 g/BHP-hr
CO standard; no engine tested at SwRI has exceeded 5.0 g/BHP-hr,
and none are anticipated to do so. This is due to the diesel
engine's inherently low levels of CO emissions.
Several manufacturers claimed that the 10.7 g/BHP-hr interim
NOx standard would be difficult to achieve in conjunction with the
reduction in HC. ECTD takes issue with this claim, primarily on
the basis of:
i) The highest NOx observed at SwRI on any 1979 engine has
been 5.91 g/BHP-hr,* only 55 percent of the proposed standard; HC
emissions for this engine were below the production target. Data
from Cummins was somewhat higher (due somewhat to a different
measurement technique); Cummins flatly stated, however, that
compliance would be no problem. Several manufacturers did comment,
however, that NOx measurements at SWRI were technically suspect.
Investigation of the equipment at SWRI revealed deficient water
traps in the 13-mode NOx analyzer, yet SWRI's 13-mode NOx measure-
ments were never used for regulatory action and standard setting.
The 13-mode tests were used both for assuring the operational
integrity of the engine and for comparitive purposes with transient
tests. Errors in SWRI's 13-mode NOx measurements therefore have no
impact whatsoever on this regulatory action.
ii) Manufacturers based their estimates on 13-mode data. The
transient test procedure generates less NOx'than the 13-mode.
Therefore, the industry's projections are overly pessimistic (see
Table 6.)
iii) Discrepancies have arisen between NOx measurements
using the CVS-bag technique and a dilute integration technique. Up
to 25% lower NOx is measured on the bag technique due to unexplain-
ed chemical reactions in the bag itself. Yet even a 25 percent
increase in SWRI's bagged NOx (Cummins utilizes the integration
technique) fails to come close to the 10.7 g/BHP-hr standard for
any of the engines tested. The difference between measured tran-
sient NOx at SwRI and the proposed NOx standard is so great that it
renders the issue of bagged vs. integrated NOx an academic question
for the purposes of this technology analysis.
1979 IHC DTI-466B (California calibration).
ilo
-------
1.50 -r
1.3 Standard
1.25
1.0
.89 Target
0.75
HC (g/BHP-hr)
0.50 -L
0.25
Caterpillar Data
» : DDA Engines
0.25
0.50
0.75
1.00
1.25
Sac Volume
(mm )
Figure i-i Projected Effects of Sac Volune Reduction on
DDA Engines.
-------
Table 1-6
Transient vs. 13-Mode NOx
Transient NOx
Engine
1979 Cummins
NTCC-350
1978 DDA 8V-71N
#1 Fuel
#2 Fuel
1979 DDA 6V-92TA
#1 Fuel
#2 Fuel
1979 IHC DTI 466B
A (Cummins data)
B
C
D
E
F
G
H
g/BHP-hr
4.91
5.40
5.69
5.83
5.91
5.56
8.94
8.45
7.82
5.07
6.99
5.69
6.94
7.24
Transient
Sampling
System
Bagged
Bagged
Bagged
Bagged
Bagged
Bagged
Integrated
Integrated
Integrated
Integrated
Integrated
Integrated
Integrated
Integrated
13-Mode NOx
g/BHP-hr
8.7*
7.10
7.03
7.28
7.58
5.70
9.12
8.66
7.98
4.59
7.58
6.30
8.20
8.14
Lab
SwRI
SwRI
SwRI
SwRI
SwRI
SwRI
Cummins
Cummins
Cummins
Cummins
Cumin-ins
Cummins
Cummins
Cummins
EPA Certification Data.
-------
In short, the NOx standard is so lax that compliance will be
easily achievable. There is enough slack between measured levels
and the standard that tradeoff with hydrocarbons is possible
(i.e., incorporate HC control techniques which result in higher
NOx.)
Finally, to allow an HC + NOx standard would be tantamount
to decontrolling HC simply because of the laxity of the NOx stan-
dard; this option is not recommended.
b. Gasoline Engines
We now turn to gasoline engines and the issue of catalyst
feasibility which is comprised of two questions: is the durability
of 100,000 miles achievable, and will future catalysts be suffi-
ciently efficient to allow compliance with the proposed standards?
Detailed discussion of the issue of catalyst durability is
presented in the Summary and Annalysis of Comments pertaining to
Allowable Maintenance intervals - the 100,00 mile catalyst. It
will suffice here to summarize those arguments and conclusions.
First of all, in—use catalyst durability data is limited to
50,000 miles, light-duty applications. Evidence was presented by
the manufacturers showing that a viable, 100,000 mile heavy-duty
catalyst would be difficult to design, primarily because of higher
poisoning rates and higher temperatures experienced in the heavy-
duty environment. ECTD's technical analysis*, however, points out
several viable strategies and finds no compelling arguments to
suggest that the full life catalyst would not be feasible.
The issue of catalyst efficiencies is impacted primarily by
the production target levels required by deterioration factors and
AQL level. Furthermore, the very mechanisms by which catalyst
overheating is precluded tend to delay catalyst light-off, thereby
increasing the impact of cold start emissions.
Discussions pertaining to projected catalyst deterioration
factors and reductions necessitated by a 10 percent AQL can be
found in the Allowable Maintenance and the Cost Effectiveness
Analyses, respectively. Based upon these analyses, the probable
target emission levels for catalyst - equipped engine are 0.5
grams/BHP-hr HC and 5.9 g/BHP-hr CO**. (NOx has been effectively
decontrolled and should not impact catalyst feasibility in any
way.) It is based upon these levels that the feasibility of
compliance will be evaluated.
Oxidation catalyst efficiencies are functions of catalyst
* Regulatory Analysis, "Allowable Maintenance - Summary and
Analysis of Comments".
** Several manufacturers concurred with these projected targets.
233
-------
sizing (i.e. cubic inch displacement), substrate and noble mate-
rials, catalyst density (i.e. the internal surface area available
for catalysis), noble metal loadings (i.e., grams/ft^ of cata-
lytically active metals within the catalyst), catalyst temperature,
and the amount of residual oxygen present in the raw exhaust
(usually introduced into rich running engines by means of air
injection.)
It has been observed during catalyst testing at the EPA
laboratory that the Los Angeles Freeway segment of the transient
test will dictate the final catalyst design. High speed, high
power performance during this segment creates the highest flow of
exhaust gases and a richer fuel/air mixture due to power enrichment
devices operating at the higher loads. Here the catalyst is
subjected to a combination of higher exhaust volumes, shorter
residence times, and higher concentrations of pollutants due to
power enrichment. Any catalyst capable of cleaning LA Freeway
emissions will easily eradicate the emissions over the remainder of
the cycle (with the notable exception of cold start emissions.)
It has also been observed in experiments with catalyst-
retrofit engines that HC control is a by-product of CO control,
i.e. if CO emissions are adequately controlled by a catalyst, then
HC control follows as a matter of course. Any catalyst designed to
handle CO over the LA Freeway should control HC over the entire
test despite the high cold start HC emissions. CO control is
therefore an reasonable measure of compliance capability.
Tables 1-7, 1-8, and 1-9 present emission data from two
heavy-duty engines retrofit with catalysts and tested at the EPA
lab.
Table 1-7 present data taken from a 1979 GM292 1-6 engine
retrofit with dual Englehardt catalysts 50 grams/ft3, 2:1 ratio
of platinum/pollidiuiE), and in standard, non-catalyst configura-
tion. The 292 represents one of the smallest truck engines and was
a logical choice on which to first test the effectiveness of
currently available catalysts. Note that while hydrocarbons were
reduced co a level close to the carget of .50 g/BHP-hr, CO was
reduced to under the standard but not to the target levels of 5.9
g/BHP-hr. (This substantiates the claim that HC is easily control-
led, despite the preponderance of cold start HC emissions*.) It
should be noted that no additional air injection was used, for
optimal performance on the transient test. General Motors re-
cognized the need for greater air injection in their written
* Cold start emission impact on composite test results were
calculated according to the following equation:
% Cold Start = 1/7 (grams : Bag 1) * 100 %
1/7 (Total grams: cold cycle) + 6/7 (Total grams: hot cycle)
-------
Table 1-7
Emission Data: 1979 GM 292
BSHC
Standard
Cycl
1.
2.
3.
4.
5.
6.
7.
8.
Cold
Hot
Test
e Segment Configuration
NYNF
LANF
LAP
NYNF
NYNF
LANF
LAP
NYNF
Cycle:
Cycle :
Composite :
65
2
0
2
6
2
0
2
6
1
2
.65
.62
.39
.08
.17
.20
.37
.33
.56
.38
.12
with Conversion
Catalyst Efficiency
36
0
0
0
1
0
0
0
3
0
0
.76
.35
.02
.03
.33
.21
.03
.08
.01
.18
.58
Test Total*
Standard Configuration:
Cata
lyst Version
Cold Start
Integrated
16.95
16.72
emissions as
BSHC
BHP-hr
a percentage
44%
87%
95%
99%
78%
90%
92%
97%
54%
87%
73%
Test
BSFC
.655
.638
of the test
BSCO
BSCO
Standard
Configuration
437
69
30
81
133
65
30
89
77
51
54
.09
.90
.23
.94
.68
.85
.50
.14
.59
.21
.98
with
Catalyst
285.
4.
6.
15.
37.
4.
6.
15.
29.
9.
12.
30
55
27
57
98
79
57
13
01
50
25
Conversion
Efficiency
35%
93%
79%
80%
72%
93%
78%
83%
63%
81%
78%
composite**
Standard Configuration: 38% 9.7%
Catalyst Version : 72% 26 %
* Sum of Cold start and Hot start.
** See footnote in text for derivation.
-------
Table 1-8
Emission Data: 1978 IHC 404 - Dual Air Pumps
Standard
Cycle
1.
2.
3.
4.
5.
6.
7.
8.
Cold Cyc
Segment Configuration
NYNF
LANF
LAF
NYNF
NYNF
LANF
LAF
NYNF
le:
Hot Cycle :
Test Composite:
36.32
9.31
0.80
6.69
17.15
4.80
0.75
5.73
5.42
2.90
3.26
BSHC
with
Catalyst
9.22
0.66
0.07
0.12
1.33
0.25
0.07
0.14
0.83
0.18
0.28
BSCO
Conversion
Efficiency
75%
93%
91%
98%
92%
95%
91%
98%
85%
94%
91%
Test Total
Integrated
Standard
Catalyst
Standard
Catalyst
Configuration:
Version :
Cold Start
Configuration :
Version :
23.784
23.055
Emissions as
BSHC
12%
33%
BHP-hr
Test
BSFC
.689
.708
a Percentage of the Test
BSCO
4.2%
12 %
Standard
Configuration
253.
90.
58.
69.
92.
62.
59.
64.
79.
62.
65.
31
87
30
06
62
68
85
70
59
80
22
with
Catalyst
103.
4.
12.
3.
2.
0.
10.
3.
17.
7.
8.
18
45
86
13
00
0
63
39
05
64
98
Conversion
Efficiency
59%
95%
78%
95%
98%
100%
82%
95%
79%
88%
86%
Composite
-------
Table 1-9
Extrapolated Emissions From IHC 404:
Fourfold Increase In Air Injection Volume Over Certified Configuration
1.
2.
3.
4.
5.
6.
7.
8.
Segment
NYNF
LANF
IAF
NYNF
NYNF
LANF
IAF
NYNF
Cold Cycle:
Hot Cycle :
BSHC
Conversion Efficiency
Air Pump Over Standard
Catalyst Configuration
6.63
0.41
0.05
0.37
2.74
0.36
0.05
0.25
0.60
0.28
82%
96%
94%
94%
84%
93%
93%
96%
89%
90%
BSCO
Conversion Efficiency
Air Pump Over Standard
Catalyst Configuration
81.07
0.26
2.72
0
12.02
0
3.43
0
7.55
3.10
68%
99%
95%
100%
87%
100%
94%
100%
91%
95%
Test Composite: 0.32 90% 3.74 94%
Cold Start emissions as a percentage of the test composite
Standard Configuration
Air Catalyst Configuration:
BSHC
12%
26%
0.27
-------
submission, claiming a need for "power modulated air injection
systems" (i.e. air injection which could increase with loads, and
not merely engine speed.) With regard to overall efficiences,
several manufacturers claimed that catalyst efficiences observed to
date have been too low to effect the required reductions. The most
significant conclusion which can be drawn from Table 1-6 is the
fact that observed catalyst efficiencies over the transient test
were 73-78 percent despite small air injection volumes and engine
calibrations optimized for a radically different test procedure and
non-catalyst emission control.
Tables 1-8 and 1-9 present data taken from a 1978 IHC 404 V-8
engine. The certified configuration included a single air pump;
ECTD personnel retrofit the engine with a second air pump supplied
by IHC (pump capacities are shown in Figure 1-2). Vacuum fit-
tings for the pumps' air divert valves were blocked off. The
engine was then retrofit* with four 113 CID Englehardt monolithic
catalysts (50 grams/ft-^ platinum-palladium, 4:1 ratio).
Test data presented in Table 1-7 represent a standard tran-
sient emission test with the engine in the above configuration.
(The left air pump delivered air to the left exhaust manifold, the
right pump to the right manifold.) The emission reductions a—
chieved in this configuration are striking. Transient CO was
reduced 86 percent from the certified configuration level; tran-
sient HC was virtually eliminated. A transient CO level of 8.98
g/BHP-hr is well below the standard, but still not quite at the
target emission level. At this point a valuable clue is found in
the catalyst efficiences for the LA Freeway; aside from the cold
start, catalyst efficiencies on the LA Freeway are the lowest
observed over the entire test. This implies a need for leaner
power enrichment or increased air injection.
In an attempt to provide this additional air injection, the
outputs from both air pumps were diverted into the right manifold
and only emissions from the right cylinder bank were measured. To
derive a total emission measurement for the engine, an additions.!
test was run with both air pumps diverted into the left bank from
which emission were measured. Total emissions were then approxi-
mated based upon the sum of emissions measured from each bank.
Emissions derived from this configuration are presented in
Table 1-9. As is readily apparent, emissions generated over the
transient cycle were virtually eliminated; transient BSCO was
reduced well below the target levels. Gold start emissions were
not great enough to threaten compliance.
* Two catalysts per cylinder bank; catalysts for a single bank
were mounted in parallel. Catalysts were mounted approximately
four feet from exhaust manifold, two feet ahead of the two dually-
mounted mufflers.
-------
Mil
Pump Data
IHC Part No. 446746-C92
461369-C91
Drive Pulley Ratio:
Pump rpm
Crank rpm ].5:1
Pump Output: 7.21-8.30 CFM
(3 1000 pump rpm and 1.6 in. Hg
Backpressure.
•H
0)
25
HI
It]
Left Pump
Right Pump
Total Capacity
0
5:00
1000 12£0 I £00 I7S0 2000 2250 2OT 27S0 3000 32S0 3S00 37S0 H000
Figure 1-2 IHC 40/c Air Iniection Capacity,
-------
Some additional observations can now be made with respect to
engine brake horsepower, fuel economy, and catalyst temperatures.
Figure 1-3 presents the maximum horsepower curve for the IHC 404
when equipped with a varying number of air pumps. Based upon the
above emission data, this particular engine requires between two to
four times the air injection volume of the certified configurtion
to achieve the low mileage emission targets. With double the air
injection producing 3.98 g/BHP-hr of CO and four times the air
producing 3.74 g/BHP-hr, a tripled air injection volume (approxi-
mately 60 CFM) could be reasonably presumed to allow emission
target compliance for this engine. Based upon the horsepower
curves presented in Figure 1-3 and the fuel consumption curves in
Figure 1-4, tripling the air injection volume can be presumed to
increase engine BSFC by 8.4 percent. It cannot be emphasized too
stongly, however, that this engine was calibrated to achieve
emission reductions on a radically different test procedure without
catalysts, i.e. performance and fuel economy were sacrificed to a
certain extent as the combustion process itself was altered to
achieve lower emissions. Emission reductions with catalysts,
however, require less engine calibrations and combustion modi-
fications, i.e., catalyst technology allows engine optimization for
both fuel economy and performance while also reducing emissions.
The proof of this lies in observations made of the light-duty
fleet between model years 1974 and 1975. A switch to catalyst
technology between these model years resulted in a fleet-wide
16.7 percent increase in fuel economy. There is no reason to
believe that the heavy-duty fleet will behave differently. There-
fore the increase in fuel consumption necessitated by the increased
air injection will be more than offset by decreases in fuel con-
sumption due to the combustion process optimizations permitted by
catalyst technology*. Furthermore, a single air pump delivering 60
CFM is certainly more efficient than three air pumps each deliver-
ing 20 CFM. Smaller pumps exhibit efficienty losses due to bound-
ary lays effect on the smaller blades. Conversion to a single pump
would certainly decrease the increaded fuel consumption to some
extent. In short, it is the Technical Staff's firm belief that no
fuel economy penalty will be experienced. In all likelhood, or
discussed in the Fuel Economy Summary and Analysis of Comments, a
net fuel economy benefit will result-
The final consideration relating to increased air injection is
the catalyst operating temperature. Temperature data for emission
test presented in Table 1-7 (dual air pumps) are presented in
Figure 1-5, for the test in Table 1-8 (the equivalent of four air
pumps) in Figure 1-6. Temperatures for catalyst on both the
* For a more detailed discussion of fuel economy effects,
including comparisons with the light-duty fleet, see "Fuel Eco-
nomy." Regulatory Analysis, Summary and Analysis of Comments."
-------
200
175
7.1% Drop in Drake
Horsepower
150
125
100
75
50
25
No air pumps
.4-
500
1000
One air pump
(certified
configuration)
air pumps
Extrapolated third
air pump (equiva-
lent to tripled
air infection
volume) .
1500 2000 2500 3000 3500
Engine RPM
Figure 1-3 WOT Horsepower. Curve for IHC 404 (no catalysts) Showing Horsepower
Losses Due to Increased Air Injection.
-------
Fuel Flow
(Ib/hr)
125
119
100
s.
75
50
25 '
HOT Fuel Consumption:
Certified Configuration: 119 lb/hr/169 IIP =
.704 Ib/HP-hr.
Tripled Air: 119 lb/hr/156 HP = .763 Ib/HP-hr.
Conclusion:
.763 Ib/BHP-hr represents an 8.4 percent increase
in fuel consumption at wide-open throttle.
1000
1500
2000
2500
3000 3500
RPM
Figure 1-4WOT Fuel Flow, IHC 404 (none, one and two air pumps).
-------
i U00
! 700
1G0Q
1LC0H
°F IHLI0
IH00
I :-iii0
l& IIUILJ
^ Q00
GUH
300
7J00
IH0
t!
a :
K- :
Inlet, right-cylinder bank catalyst.
Inlet, left-cylinder bank catalyst.
Outlet, right-cylinder bank catalyst.
Outlet, left-cylinder bank catalyst.
Twenty-minute
soak
Hot Cycle
30 as: H0
Test Time (minutes)
— \
£0
llff E0 55 E0
Figure 1-5 Catalyst Temperature Data: 20..CFM AIR per Cylinder Bank
-------
I GOD
HQJ]
200
JDDD
Baa
HQ0
A: Inlet - Right Cylinder Bank Catalyst (No Air)
D: Outlet - Left Cylinder Bank Catalyst (Air)
0: Inlet - Left Cylinder Bank Catalyst (Air)
• : Oil Temperature
Twenty
Minute
Soak
£.0
2ET
30
MB:
5:5:
E0
-------
right and left cylinder banks are presented. Catalyst temperatures
on the high power, hot cycle LA Freeway barely exceeded 1500°F,
regardless of the air injection volume. At least for this engine,
additional air injection did not raise the catalyst operating
temperature to a critical level.
In summary, a limited test program conducted at EPA's Motor
Vehicle Laboratory achieved emission reductions exceeding those
required by the proposed standards on one of the larger gasoline
engines in less than two weeks time. The industry, on the other
hand, has four years. Based upon this experience, we can only
conclude that technology is available to allow compliance with the
low-mileage emission targets. As discussed in the "Allowable
Maintenance," section of the Regulatory Analysis, a full life,
100,000-mile catalyst is presumed achievable.
4. Recommendations
The proposed standards are attainable for both gasoline and
diesel engines.
Retain the proposed standards for Final Rulemaking.
-------
J. Issue - Selective Enforcement Auditing
1. Summary of the Issue
In brief, this issue can be stated as follows: What Accep-
table Quality Level (AQL) should be promulgated in the final rule?
The AQL represents the percentage of heavy-duty engines (HDE)
within a given population which will be allowed to exceed the
emission requirements. The Clean Air Act does not specify the AQL
to be applied to an assembly-line testing program like SEA.
A 10% AQL reflects EPA's view that the statute requires
every engine to be warranted to meet the emission standards
while allowing 10% for measurement error and inevitable quality
aberrations. A 40% AQL assures that, for an engine population
assumed to have a skewed-normal distribution, engines within this
population will comply with standards on the average.
EPA promulgated a 40% AQL for its light-duty vehicle (LDV) SEA
program because at the time the regulation went into effect, the
LDV industry was building vehicles to meet previously established
standards on the average. In order to have brought the light-duty
vehicle engine families into compliance with a 10% AQL, manufac-
urers would have had to add additional emission control equipment
to retain their certificates of conformity- EPA's intent in
promulgating a 40 percent AQL for its light-duty vehicle SEA
program was to provide light-duty vehicle manufacturers the time
and flexibility to bring all their vehicles into conformance with
the standards on a reasonable schedule. This schedule is to
parallel efforts to improve fuel economy.
In the HDE Notice of Proposed Rulemaking (NPRM), the Agency
proposed a 10 percent AQL as part of the total compliance strategy
outlined in the proposal. EPA indicated that the 10 percent AQL
could be met within costs not unreasoanbly burdensome to the
manufacturers. Comments on costs associated with meeting a 10
percent AQL were requested in the proposal.
2. Summary of the Comments
The manufacturers, and other organizations which responded,
were practically unanimous in their opposition to the implemen-
tation of a 10 percent AQL. Most of the comments concerned reasons
why a 10 percent AQL should not be promulgated. Very little data
were provided relating to the actual technological and economic
considerations associated with meeting this AQL level or, in many
cases, even the much preferred 40 percent AQL. Many manufacturers
attributed this lack of data to their inability to run the newly-
proposed transient test procedure for heavy-duty engines.
-------
Ail commentors made one or more of the following three
major points: The 10% AQL is contrary to the "Congressional
intent of averaging",, the 10% AQL effectively makes the standards
more stringent than a 90% reduction from baseline, and the 10% AQL
is inconsistent with the 40% AQL currently in effect for the
light-duty vehicle SEA program. In addition, commentors gave
various other reasons why a 10% AQL should not be put into effect:
it will cause penalties in fuel economy; it will cause increased
costs in the areas of test facilities, production testing, emis-
sions hardware, and fuel consumption, it provides no important air
quality impact: and it conflicts with certification requirements.
which imply "averaging". One commentor suggested that a 40%
AQL should be promulgated, after which it could be lowered in the
future as was suggested for the LDV SEA program. Another corn-
mentor stated that various combinations of emissions standards and
AQL should be investigated for cost-effectiveness.
Most manufacturers stated that, due to the above reasons,
the AQL should be revised to 40% in the final rule. Some gave
examples of a 40% AQL or "averaging" sampling plan that EPA could
adopt for the final rule.
a. Gasoline - Fueled Engine Manufacturers
General Motors (also manufactures diesel engines)
GM stated that it is opposed to a 10% AQL sampling plan for
SEA. GM believed that Congress did not intend, in the 1977 Clean
Air Act Amendments, to impose a more stringent AQL than that
required for the LDV program (40%). G.M. asserted that Congress
intended averaging for production line testing.
In its SEA discussion, GM directed most of its arguments
towards supporting the concept of averaging for production line
testing. These arguments included statutory language ("...regu-
lations shall contain standards which require a reduction of a
least 90 percent ...from the average of the actually measured
emissions..."in Section 202(a)(3)(A)(ii) of the Act); the Draft
Regulatory Analysis discussion which GM claims to be based on
averaging; ambient air quality considerations; various past Con-
gressional Committee reports and statements of EPA Administrators;
consistency with certification requirements; and the analysis of
the baseline testing program.
GM also stated that the emissions design target necessitated
by the 10% AQL would have to be more than 50% below the target
necessary to satisfy the "90%" reduction from baseline. No anal-
ysis was provided to support this statement.
-------
Ford Motor Company
Ford stated that it favored a 40% AQL for SEA because that AQL
is consistent with Clean Air Act requirements, with certification
requirements, and with the current LDV SEA program. Ford also
stated that a 40% AQL was needed because of the many new HDE
compliance strategies contained in other parts of the NPRM.
This manufacturer believes that technology is not currently
available and will not be available by 1983 to enable it to design
its engines to comply with a 10% AQL. Ford stated that the 10% AQL
SEA is more stringent than certification, which is based on aver-
aging, and that Congressional action and environmental studies have
not demonstrated a need for a 10% AQL. Finally, Ford contended that
an SEA program in any form would not result in an air quality
improvement; however, if such a program was to be conducted,
Ford felt that the 40% AQL is the only logical cost-effective
alternative.
Ford developed an AQL sampling plan similar to the one pro-
posed by EPA and incorporating a 40% AQL. It suggested that this
plan be adopted in the final rule. In the area of economic impact
of the SEA regulation, Ford stated that the 10% AQL would require
100% production testing "to ensure an adequate probability of
meeting an SEA test order," although Ford did not analyze the
relationship between AQL, production testing rate, and the proba-
bility of passing an SEA. Ford contended that the 100% production
testing would in turn impose additional test facilities, equipment,
and plane modifications. The manufacturer stated that a 10% AQL
may also reduce assembly line speeds and require more emissions
hardware. No analysis of the extent of these effects was
provided by Ford.
Chrysler
Chrysler commented that the 10% AQL represents a "considerable
increase in stringency" ever the 40% AQL and would result in a
greater than 90% reduction from baseline. This manufacturer stated
that the Clean Air Act does uot require every engine to meet
emission standards throughout its useful life. Rather, Chrysler
believed the Act "compels11 averaging because of the Section 2Q2(a)
(3)(A)(ii) statement requiring reductions "...from the average of
the actually measured emissions,..11. It stated that deterioration
factors are determined through an averaging methodology. Chrysler
also pointed out that several statements in the NPRM documents
suggest that EPA itself was viewing emissions on an average basis.
The HDE NPRM Preamble stated that a 40% AQL was instituted for
LDV SEA "...to avoid an unreasonable economic impact on the indus-
try." Chrysler stated that the "several SEA failures and many
close calls" at the 40% AQL level in the LDV SEA program were"
-------
evidence that Che industry is not advanced enough in its production
practices to contend with a 10% AQL and that the 10% AQL was
therefore unwarranted and unsubstantiated.
In the area of cost impact, Chrysler mentioned two factors
associated with a 10% AQL: additional emissions hardware (possibly
an expensive 3-way catalyst) and additional manpower for compliance
surveillance. No analysis was presented to support these cost
increments.
Chrysler advocated adopting a 40% AQL because of economic and
technological considerations and because it would satisfy the
Congressional intent of averaging.
International Harvester Company (also manufactures diesel
engines)
IHC stated that the 10% AQL proposal should be withdrawn until
properly evaluated on a cost-benefit basis because of the enormous
cost involved. It estimated an SEA at 10% AQL would cost $180 per
IHC engine audited for SEA purposes.
IHC believed that Congress did not intend that a 10% AQL be
promulgated because that AQL imposes emission standards more
stringent than those required by certification. IHC stated that
the legislative history of the Clean Air Act indicates that ve-
hicles need only meet standards on the average. In addition, IHC
felt that the imposition of a 10% AQL would have an adverse eco-
nomic impact on the heavy-duty industry, so it should be relaxed as
it was in the light-duty vehicle SEA regulations. Specifically,
IHC envisioned that the 10% AQL would impose a large in-house
quality audit program that would dwarf the costs for test facili-
ties and EPA audit testing. IHC provided estimates of the costs of
a building, equipment, and testing.
b. Diesel Engine Manufacturers
Cummins Engine Company
Cummins advocated a 40% AQL because of what it claimed to be
the significant variation (possibly 20%) in test-to-test and
engine-to-engine variability shown on the current steady-state test
procedure. With the proposed decrease in standards, Cummins felt
that variability will increase. (Cummins seated that it had done
analyses of these variabilities, but they were not provided. It
did cite two studies on NOx instrument error and did suspect
that varibility could be 60% or more of the proposed standards.)
This manufacturer stated that the 40% AQL would allow for this
increased variability, which is presumed to also show up on the
transient test. dimming suggested that the AQL could be reevalu-
ated after the new standards go into effect.
-------
Caterpillar Tractor Company
Caterpillar stated that the 10% AQL is not consistent with the
previously established concept that heavy-duty engines must meet
emission standards on the average during their useful lives.
Caterpiller believed the proposed 10% AQL would now require
that almost all engines comply with standards. Caterpillar recom-
mended that a 40% AQL be adopted to approximate averaging and
thereby retain the original compliance concept. Caterpiller felt
that to adopt the 10% AQL would be to require emissions reductions
in excess of the 90% from baseline required by Congress in the 1977
Clean Air Act Amendments. This conmentor suggested a methodology
that could be used to establish a revised standard such that this
standard, in conjunction with a 10% AQL, would give average engine
emissions representing a 90% reduction from baseline.
Mack Trucks, Inc.
Mack stated that the 10% AQL is not consistent with Congres-
sional intent and with past EPA policy of requiring SEA test
vehicles to meet standards on the average (in the LDV SEA program).
Mack asserted that a 40% AQL is more representative of annual
production and more cost-beneficial. In addition, Mack argued Chat
there is a substantial fuel economy penalty incurred in going from
a 40% to a 10% AQL because of lower NOx design targets. Mack did
not explain why the change in the NOx design target would be
required.
Mercedes-Benz of North America
Mercedes-Benz, a subsidiary company of Daimler-Benz AG,
stated that the 10% AQL proposal is without merit. In view of the
statement about "...the average of the actually measured emis-
sions,." in Seer-ion 202(a)(3) (A)(ii) of the Clean Air Act, Mer-
cedes-Benz considered the 10% AQL to be contrary Co Congressional
intent, which is that standards are to be met on the average. In
the LDV area, Mercedes-Benz noted, EPA has recognized that a 40%
AQL corresponds to an averaging concept and, therefore, Mercedes-
Benz felt that the 40% AQL should be adopted to determine compli-
ance with heavy-duty engine emission standards.
The Perkins Engines Group (England)
Parkins stated that Clean Air Act enforcement provisions are
the same for both the LDV and HDE categories. To impose a 10% AQL
for the HDE class when a 40% AQL is currently in effect for the LDV
SEA program represents a double standard, in Perkins view, which is
all the more arbitrary in view of the fact that LDVs are a greater
source of overall ambient emissions than HDEs. This commentor
stated that a 40% AQL should be adopted to ensure compliance
on the average and co ensure comparability with the LDV SEA
-------
program. Perkins suggested that the AQL could then be tightened in
the future for both classes.
c. Other Commenters^
Motor Vehicle Manufacturers Association
MVMA" s main comment was that the 10% AQL is not consistent
with an averaging concept for determining compliance with stan-
dards. Its arguments were based on the legislative history of the
Clean Air Act, statements of past EPA Administrators, the averaging
concept embodied in certification regulations, ambient air quality
studies based on averages, the statutory language in Section
202(a)(3)(A)(ii), and the inclusion of averaging concepts in the
Regulatory Analysis for the NPRM.
Engine Manufacturers Association
EMA stated that the intent of Congress was that production
engines should meet standards on the average. EMA contended that
the present 40% AQL for the LDV SEA program approximates averaging
and thus conforms to Congressional intent. EMA indicated that EPA
has shown no rational basis for imposing the much more stringent
10% AQL and urged adoption of the 40% AQL.
U.S. Department of Commerce
The Department of Commerce (DOC) commented that there is no
rationale in the NPRM for a 10% AQL. DOC stated that the 10% AQL
is "exceedingly stringent" relative to the 40% AQL in the LDV SEA
program. The 10% AQL will cause a drastic increase, DOC believed,
in the stringency of the standard to account for unavoidable
production variations. DOC stated that a 10% AQL may not be
technologically feasible, would adversely affect fuel economy, and
would cause substantial cost increases. DOC did not, however,
orovide any data or analysis to support these claims. A 40% AQL
should be adopted, DOC concluded, to ensure meeting standards on
the average.
U.S. Council on Wage and Price Stability
COWPS recommended that a cost-effectiveness study of the 10%
AQL be performed to evaluate its economic and social worth as part
of the NPRM. COWPS suggested that an estimate of the emission
reduction in going from a 40% AQL to a 10% AQL can be made, given
the statistical distribution of assembly-line engine emissions.
COWPS also asserted that the average emission levels neces-
sitated by the 10% AQL would be appreciably lower than those
required by the 90% reduction, i.e., than the 40% AQL levels which
approximate average emissions, but they provided no analysis to
support this conclusion. If EPA desires only a 90% reduction in
-------
average emissions, COWPS suggested thac Che numerical standards be
raised so that the average emission level with a 10 percent AQL
coincides with a 90 percent reduction from the baseline average.
Regardless of the combination of numerical standards and AQL that
is ultimately promulgated, COWPS felt that combination should have
satisfied the test of cost effectiveness.
3. Analysis of the Comments
Since many of the manufacturers and organizations responding
made similar comments on the AQL issue, each of the major comments
will be discussed under a separate heading in this section for
purposes of clarity. For further information relating to the 10%
AQL issue, reference is also made to the cost-effectiveness studies
in Chapter VII of the Regulatory Analysis and to the discussion of
the technological feasibility of the emission standards in the
Summary and Analysis of Comments.
a. The 10% AQL is Not Contrary to Congressional Intent
When reviewing the comments to the NPRM on SEA for light-duty
vehicles in 1976, the EPA Office of General Counsel (OGC) reached a
finding that "...Congress intended that, eventually, every car
coming off the assembly line should meet the emission standards
established under Section 202." A copy of the memorandum con-
taining this finding is available in the Public Docket for this
Rulemaking. OGC acknowledged that a phasing in of this requirement
was appropriate to avoid implementing SEA in an unreasonably
burdensome manner, so long as the ultimate goal of full compliance
is not abandoned. As explained in the LDV SEA preamble (41 FR
31474, July 28, 1976), auto manufacturers argued that implemen-
tation of a 10% AQL would have a disastrous economic impact on the
industry, since it would result in a loss of certification for a
majority of engine families. A 40% AQL was therefore established
to implement SEA in a manner not unreasonably burdensome to the
affected manufactures. This approach was designed to "provide
manufacturers Che time and flexibility to bring all their vehicles
into conformance with the standards on a reasonable schedule" (41
FR 31475).
Authority for SEA testing of heavy-duty engines is the same as
for LDVs ^Section 206(b) of CAA). EPA maintains the position that
there is a specific legal basis for requiring every HDE coming off
the assembly line to meet standards. The full text of the EPA
General Counsel memorandum, mentioned above, explains how, in fact,
the language of the Clean Air Act and the relevant legislative
hiscory support an "every car" approach to compliance with emission
standards.
The ultimate goal of every vehicle and engine complying with
emission standards is also supported by the U.S. General Accounting
-------
Office (GAO) . The GAO did not take issue with EPA's legal inter-
pretation of the Clean Air Act on this matter and recommended that
the current LDV SEA program be revised to "...require a Federal
emission standard compliance rate more indicative of the current
rate for car configurations tested, which is well in excess of the
60% passing rate required." (GAO Report GED 78-180, p. 28).
b. The Relationship Between the Standards and a 10 Percent
AQL
Section 202(a)(3)(A)(ii)'of the CAA states, in pertinent part,
"...regulations... applicable to emissions from vehicles or engines
manufactured during and after model year 1983, in the case of HC
and CO, shall contain standards which require a reduction of at
least 90%... from the average of the actually measured emissions...
during the baseline model year." Pursuant to this requirement, EPA
conducted a test program on 1969 model year heavy-duty gasoline
engines (the last model year before imposition of HC/CO standards
for heavy-duty engines). Using the sales-weighted average emission
levels obtained during this program, the standards were then set by
multiplying these levels by 10 percent, i.e., a 90 percent reduc-
tion. These numbers once identified, then became the required
standards. The 10 percent AQL does not change the values of the
standards; it merely requires that every production engine must
comply with the established standards. This is consistent with
EPA's finding as discussed in Section 3(a), that every produc-
tion engine must comply with standards established under Section
202 of the Clean Air Act.
EPA has performed an analysis which indicates that a 10% AQL
can cause a manufacturer to design to lower target emission
levels than those required by a 40% AQL. However, the magnitude of
the difference between the target levels depends on several fac-
tors, some of which are within the manufacturer's control. One of
the most important of these factors is the variability of identical
production engines ("width" of emissions distribution) at each
design level. By increasing quality control and minimizing other
variations in the manufacturing and assembly process, the manufac-
turer may reduce variability and raise the target emission levels
which he needs to be meet. In practice, the Agency believes that
each manufacturer will trade off to one degree or another lower
design targets vs. stepped-up quality control to obtain the most
cost-effective approach towards the 10% AQL goal.
c. The Consistency of the 10% AQL With the 40% AQL Currently
In Effect for the Light-Duty Vehicle SEA Program
The 40% AQL was established for the LDV SEA program to imple-
ment the program in a manner not unreasonably burdensome to the
affected manufacturers. At the time LDV SEA was proposed, several
auto manufacturers stated that they built the average production
-------
vehicle to meet the standards. It is important to note that
the situation regarding LDVs and the 10 percent AQL is different
from that relating to heavy-duty engines. As discussed in the
Preamble to the LDV SEA regulations:
"The approach taken here, then, of not setting the AQL at 10%
will provide manufacturers the time and flexibility to bring
all their vehicles into conformance with the standards on a
reasonable schedule. Such a schedule can be compatible with
their parallel efforts to improve fuel economy and which does
not expose them unduly to the risk of loss of certification
while they are learning to bring their production vehicles
into compliance with the law." (41 FR 31475, July 28, 1976)
The circumstances under which the HDE SEA program is being
promulgated are significantly different than those in the LDV
case. Within the constraints of the CAA, EPA is authorized to set
the standards for the HDE industry. The Agency can, therefore,
take the effect of a 10% AQL into account when considering whether
revised standards or more stringent statutory standards should be
set. Moreover, the affected industry has 4 years leadtime before
the standards and the SEA program go into effect, so that manufac-
turers can plan their design targets so as to have all production
engines in compliance with the law starting in 1984. EPA's ap-
proach in both the LDV and HDE cases is a consistent one: The
Agency has endeavored to implement an SEA program consistent
with its legal interpretstion that every vehicle or engine must
meet standards and in a practical manner that does not place an
unfair or unreasonable economic or technological burden on the
affected industry. In the HDE case, the Agency has determined that
the final standards, in combination with a 10% AQL, are not
unreasonably costly to the affected manufacturers and are tech-
nologically attainable within the 4-year timeframe.
d. The Effect of the 10 Percent AQL on Fuel Economy
Based on assessments of technological feasibility by EPA's
Office of Mobile Source Air Pollution Control (OMSAPC), there will
be no fuel economy penalty in designing to meet the 10% AQL emis-
sions targets. The 10% AQL imposes a NOx target level already
being bettered by most heavy-duty diesel engines, so EPA does not
accept Mack's contention that a lower NOx design target would be
required ;hat would in turn result in a substantial fuel economy
penalty. In fact, EPA analysis indicates no loss in HDD engine
fuel economy and a 4-9% benefit in fuel economy for HDG engines.
In the case of heavy-duty diesel engines, no fuel economy
penalty is expected due to modifications to meet the required
design targets. The slight penalty resulting from use of exhaust
-------
gas recirculation controls are expected to be offset by more fuel
efficient emission control techniques, such as aftercooling and
improved injector design.
For heavy-duty gasoline engines, OMSAPC anticipates that
high-efficiency catalysts will be developed to comply with the
regulatory requirements. The use of catalysts allows engines to be
tuned for fuel economy, as opposed to non-catalyst equipped en-
gines, where the engine must be tuned to comply with the emission
standards which could cause possible fuel economy penalties.
e. If a 40% AQL is Promulgated, It Could Be Lowered in
Future Model Years As Was Suggested in the LDV SEA
Program
As discussed in 3.a., the Act has an established legal basis
for promulgating a 10% AQL. As discussed in 3.c, the Agency has
the opportunity, in this rulemaking, to set standards and an AQL
such that no unreasonable burden will be placed on the HDE manufac-
turers in terms of their ability to comply with all aspects of the
total regulatory strategy. Perkins Engines Group stated that the
AQL should be the same for both the HD and LD classes. It is not
EPA's intention to ensure absolute comparability between the two
classes, but rather to set an AQL consistent with its legal
interpretation of the Clean Air Act and the production capabilities
of the affected industry at the time that both emission standards
and AQL go into effect.
EPA has determined, based on available information and anal-
ysis of its own and manufacturers' data, that a 10% AQL can be
implemented in 1984 on a cost-effective basis. Therefore, EPA
does not see the need to first promulgate a 40% AQL and then lower
it in future model years to the legally required 10% level.
f. Air Quality Impact of a 10 Percent AQL
EPA has performed an analysis of the reduction in emissions to
be obtained in going from a 40% AQL to a 10% AQL in the SEA pro-
gram. This analysis appears in Chapter VII of the Regulatory
Analysis. The findings of this analysis indicate that by imple-
menting a 10% AQL, HDG HC emissions will be reduced an average of
0.04 tons per vehicle over the vehicle's lifetime, EDG CO emissions
will be reduced 0.5 tons, and HDD HC emissions will be reduced
0.24 tons.
As shown on Tables VII-1 and VII-2 in the Regulatory Analysis,
there reductions represent a positive reduction in HC and CO for
HDG and HDD engines which EPA analyses have shown can be achieved
in a cost-effective manner. On the basis of dollars spent per ton
of emissions removed the 10% AQL compares favorably with other
emission control strategies.
-------
g. Relationship of a 10 Percent AQL Program to Certification
Requirements
Several commenters indicated that they felt the present
certification program embodied an averaging concept which con-
flicted with the concept of a 10 percent AQL. They argued that
consistency required use of a 40 percent AQL so that essentially
the average engine emission level would meet the standards.
The staff does not agree with this contention. The purpose of
the certification program and an SEA progam are complementary and
do not conflict. Section 206 of the Clean Air Act, "Motor Vehicle
and Motor Vehicle Engine Compliance Testing and Certification",
authorizes a certification program (206(a)) and an assembly line
testing program (206(b)). If a new motor vehicle or engine design
demonstrates compliance with Section 202 standards throughout its
useful life, a certificate of conformity will be issued under
206(a) regulations. The certificate is issued with respect to
Section 202 regulations, i.e., regulations establishing emission
standards. Since the function of the assembly line testing program
is "to determine whether new motor vehicle's or engines being
manufactured do in fact conform with regulations with respect to
which the certificate of conformity was issued" the program will
determine compliance with emission standards.
In summary, the EPA certification and SEA programs attempt to
accomplish different but related objectives. Through certifica-
tion, a manufacturer demonstrates that it has the capability to
design a vehicle or engine that will comply with standards through-
out its useful life under conditions simulating actual use. Once
these prototype vehicles or engines demonstrate compliance, SPA
issues the manufacturer a certificate of conformity allowing it to
actually manufacture vehicles or engines similar to the prototypes
for distribution into commerce. Then SEA requires the manufacturer
to demonstrate that newly manufactured vehicles or engines
will also comply with standards thoughout their useful lives.
h. Cost Impact of the 10 Percent AQL on Heavy-Duty Engine
Manufacturers
There is a cost component attributable to the 10% AQL, as
there is to all other compliance options in the regulatory package.
A heavy-duty engine manufacturer will actually incur a 'VI10% AOL"
cost in those cases where it experiences difficulty in attaining
the target emission levels, i.e., when the manufacturer must spend
more money in going to the 10 percent AQL target level from some
other (higher) level, and where it decides to step up its in-house
quality control programs in response to a 10 percent AQL.
A cost-effectiveness analysis has been performed in conjunc-
tion with Che evaluation of this regulation. One option examined
-------
was Che cost of the proposed SEA program at a 40 percent AQL versus
its cost at a 10 percent AQL. The analysis indicated that the 10
percent AQL SEA program is the more expensive option, but that the
cost of moving to the 10 percent AQL is small relative to other
options in the regulation,' and also, in view of the benefit in air
quality, the 10 percent AQL has a favorable cost effectiveness
ratio that is in line with the other compliance strategies in this
regulation.
4. Staff Recommendations
It is recommended that a 10 percent AQL be promulgated in the
Final Rule. An SEA program with a 10 percent AQL is consistent
with EPA's legal interpretation of the Clean Air Act, does not
place unreasonable cost burdens on heavy—duty engine manufacturers,
results in a positive reduction in emissions, has no impact on fuel
economy- and is technologically feasible, given the emission
standards to be promulgated.
-------
K. Issue - Nonconformance Penalty
1. Summary of the Issue
This issue concerns the system for production compliance
auditing (PGA) and nonconformance penalties (NCP) proposed in
the NPRM. Since the proposed emission standards were considered
feasible for all manufacturers to meet, nonconformance penalties
were not made available in the NPRM.
As described elsewhere in this Summary and Analysis of Com-
ments (see I. "Technological Feasibility"), the EPA staff still
believes that the standards are attainable by all manufacturers.
However, to provide for isolated instances when compliance may not
be attained due to unforeseen circumstances, the staff recommends
that the PCA/NCP system be made available. Since the NPRM did not
contain either a proposed "upper limit" different from the stan-
dards, or the marginal cost component of the general penalty
formula, reproposal will be required. Therefore, the PCA/NCP
system should proceed as a seperate rulemaking and no detailed
analysis of the comments received will be done at this time,
2. Recommendations
The PCA/NCP portion of the original proposal should be separ-
ated from the final rulemaking and reproposed. It should be
reproposed with the addition of upper limits on certification and
the marginal cost component of the general penalty formula. An
opportunity for comments will be provided for all aspects of the
reproposed PCA/NCP regulations.
The preamble to the final rulemaking, in describing the
removal of PCA/NCP from the package, should make it clear that
EPA's intent is for all manufacturers to comply with the standards,
and that manufacturers should proceed in that fashion. It should
be explained that the availability of a nonconformar.es penalty is
more of s. "safety valve" for unforeseen complications than a route
:o a less stringent emission standard (via designing for a higher
emission rate and planning on paying a penalty for all engines).
Reproposal and finalization of PCA/NCP at a later date does not
relate in any way to any statutory leadtime requirements for
finalizing the heavy-duty engine emission standards.
-------
L. Issue - Diesel Crankcase Emissions Control
1 Summary of the Issue
The proposed regulations require that "[njo crankcase emis-
sions shall be discharged into the ambient atmosphere" from 1983
model year (and later) heavy-duty diesel engines. A similar
requirement has been in effect for gasoline engines for a number of
years; this is the first time EPA has proposed to regulate diesel
crankcase blowby.
2. Summary of the Comments
The proposed crankcase controls for diesel HDE's drew con-
siderable adverse reaction. Both EPA's justification and feasi-
bility issues were addressed by most of the commentors.
Several comments pointed to the low brake-specific hydrocarbon
and carbon monoxide emissions from HD diesel crankcases as evidence
of a lack of need for controls. Additionally, the information
quoted in the NPRM is inconclusive in establishing the presence of
nitrosamines in diesel blowby emissions.
The feasibility of controlling the crankcase emissions was
challenged on the basis that most HD diesel engines will be
equipped with turbochargers and intercoolers/aftercoolers by 1983.
The anticipated technical problems arise from the oily nature of
the blowby emissions which in a simple system would be introduced
into the inlet air supply. Although in a naturally-aspirated
engine the slight negative pressure of the manifold can draw
crankcase fumes into the combustion chamber, the manifold of a
turbocharged engine is under greater pressure than the crankcase.
Thus, unless it is pressurized, the blowby must enter the stream on
the inlet side of the turbocharger, allowing the oily emissions to
become deposited on the compressor wheel. Similarly, the heat
transfer surfaces of the heat exchangers can become coated with the
residues. Several of the commenters indicate that such events can
hamper the efficiency of both of these components. Loss of Curbo
efficiency can detrimentally affect performance, fuel consumption,
and emissions. The commenters also expressed a concern that turbo
durability will suffer and increased maintenance will be necessary.
Mack Trucks has tested a turbocharged engine equipped with
crankcase gas recirculation and observed a decrease in performance
and an increase in fuel consumption and smoke opacity. The inter-
cooler also became plugged appreciably.
Finally, Cummins Engine Company mentioned four means of
crankcase control which may have potential, none are developed to
the extent of assessing their feasibility. These four alternatives
follow:
-------
1) Duct gases to turbo-inlet by way of a pressure regulator
and oil separator (this method has resulted in severe loss of turbo
efficiency).
2) Draw gases through the regulator and separator and a pump
to the manifold, downstream of the turbo (an expensive alterna-
tive; still requires much development work to ascertain whether it
is satisfactory).
3) Aspirate and mix gases into the exhaust flow.
4) Pump the gases through the regulator and a separator into
the exhaust stream.
3. Analysis of the Comments
The heavy-duty diesel crankcase emissions data reported in
"Diesel Crankcase Emissions Characterization, Final Report of Task
No. 4, Contract No. 68-03-2196," referenced in the NPRM, have been
updated with one Cummins engine. The additional engine showed HC,
CO, and NOx crankcase emissions (g/BHP-hr) which were very compar-
able to thse reported previously and lend creedence to the earlier,
limited data.
Estimates of the cost effectiveness of requiring control of
the major crankcase pollutants may be calculated by dividing the
anticipated control system costs by the expected lifatime emis-
sions, in tons. Since all four major pollutants are controlled by
the system, it is appropriate to distribute the system cost equally
among the four pollutants. Sc for each pollutant one-quarter of the
system cost is divided by the tonnage of lifetime emissions, as
represented by the product of the emission rate and the total time
of engine operation. The resulting cost-effectiveness numbers can
be compared to those associated with other control strategies for a
measure of the acceptability of the costs.
The following crankcsse emission rates are taken from the
"Final Report" referenced above:
Emission Rate (grams per hour)*
HC
Cummins NTC-25G 0.63
DDA 67-71 #1 Q.,77
DDA 67-71 #2 (Std. Speeds) 0,69
DDA 67-71 #2 (Low Speeds) 0.39
Mean 0.62
CO
2.48
0.06
0.45.
0.44
0.86
NOx
1.15
0.04
0.39
Q.22
0.45
Part
1.02
0.75
2.11
1.30
* Grams/hour numbers were not available for the Cumins engine.
These estimated rates of emissions would be seen over a'n
average lifetime for HD diesels of approximately 475,000 miles.
Using 3,000 hours of operation to represent each 100,000 oiiles.
-------
the expected lifetime is 14,250 hours. (As a rough check, 475,000
miles at 40 mph yields 11,850 hours).
Finally, a control system cost estimate is needed. We have
very little information on which to base such an estimate, but
Caterpillar quoted in their comments the following anticipated
costs for PCV systems:
Engine Estimated System Cost
3208T (Naturally aspirated) $ 10
3306 (Turbocharged and aftercooled) 135
3406 (Turbocharged) 145
3408 (Turbocharged and aftercooled) 145
For the purposes of this cost-effectiveness calculation, we
will use $10 for naturally-aspirated engines and $100 for turbo-
charged engines. The $100 figure assumes a turbocharger-bypassing
system as described below (Caterpillar provided no information to
support their higher numbers).
Distributing these costs among the pollutants and coverting
grams to tons, we created the following table:
Cost Effectiveness of Control ($/Ton)
Pollutant Naturally Aspirated Turbocharged
HC $257/Ton HC 2,570/Ton HC
CO 185/Ton CO 1,850/Ton CO
NDx 354/Ton NOx 3,540/Ton NOx
Particulate 123/Ton Part. 1,225/Ton Part.
To have meaning, these cost-effectiveness numbers must be
compared with the cost effectiveness of other emission control
programs. Listed below are ranges of cost effectiveness covering
most of EPA's stationary and mobile source control programs:
Pollutant Cost-Effectiveness Range ($/Ton)
HC 70 - 800
CO 10 - 40
NOx 100 - 2500
Particulate 10 - 1000
It is clear that crankcase control on turbocharged engines is
not as cost effective as other control programs for any of the four
pollutants. However^ controls on naturally-aspirated engines fall
in the "acceptable range" for HC, NOx and particulates. Since the
costs were allocated equally among four pollutants, we could now
re-compute the cost effectiveness when the costs are distributed
-------
only among HC, NOx, and particulate. Performing this calculation
yields the table below:
Cost Effectiveness of Control ($/Ton)
Pollutant Naturally Aspirated Turbocharged
HC 342 3,420
NOx 472 4,715
Particulate 163 1,630
These final cost-effectiveness numbers are still in the acceptable
range, and the staff concludes from this analysis that crankcase
controls on naturally-aspirated engines are justified on the basis
of HC, NOx and particulate control. (Routing crankcase NOx emis-
sions through the combustion chamber will not actually elimi-
nate them. However, controls on naturally-aspirated engines remain
cost effective even when credited to HC and particulate control
alone).
On the other hand, if nitrosamines are found to be a signifi-
cant component in crankcase emissions, a considerably more costly
control system might be acceptable. Preliminary data from current
heavy-duty diesels indicates that nitrosamines may indeed be
present. The complete and reduced data will not be available to
EPA until late in 1979.
Finally, the staff has explored the feasibility question,
which to a large extent revolves around the compatibility of
crankcase controls with turbochargers and associated heat ex-
changers (intercoalers and/or aftercoolers). Of course, these
components are not present on naturally-aspirated engines and hence
there are no major technical problems with crankcase control on
these engines. Naturally-aspirated engines in 1979 comprise 23
percent of the heavy-duty diesel market*, a fraction that is
expected to rapidly drop even further as manufacturers use turbo-
charger technology to respond to fuel economy pressures. Still,
the ease of application and low cost of controls on naturally-
aspirated engines make controlling this portion of the market a
reasonable option.
Alternative #2 suggested by Cummins in their comments (Summary
of Comments above) seems very worthy of pursuit for turbocharged
engines. By allowing the turbocharger and heat exchangers to be
bypassed, the oil separator/ pump/pressure regulator configuration
would eliminate the excessive deterioration of the component
efficiencies. (The pump itself might be affected to some degree by
the oily emissions.) While it is clear that such a system has yet
* Eased on 1979 certification data.
-------
to be developed for diesel engines, the staff perceives no major
technical obstacles to impede the design; pump technology in
general is highly advanced. Notwithstanding the cost, which will
probably exceed $75, the staff is convinced that a pumped system is
a technologically-feasible option for 1984. (It is interesting to
note that a 1980 GM turbocharged engine is already equipped with
crankcase controls, though probably the turbocharger is not by-
passed) .
4. Recommendations
We conclude from the foregoing analysis that a crankcase
control system for naturally-aspirated engines is feasible and may
be justified on the basis of EC, NOx, and particulate control.
However, controls for turbocharged HD diesels are not expected to
be cost effective for HC, CO, NOx or particulate control. We do
expect that the forthcoming nitrosamine emission data will warrant
a serious reconsideration of the need for controls. Further, we
conclude that a system using a pump to bypass the turbocharger can
be developed, if necessary, for 1984 model year engines.
Our recommendation is that diesel crankcase control require-
ments be retained for naturally-aspirated diesels but that the
finalizing of control requirements be postponed for turbocharged
engines pending the completion of the nitrosamine research now in
progress. Significantly, we urge that the Preamble of the HD
Gaseous Final Rulemaking clearly make the following points:
1. EPA anticipates that crankcase controls on turbocharged
diesels will be necessary for the control of nitrosamine emis-
sions. The proposed requirements for these engines are not being
finalized at his time and will remain proposed until such a
time that new nitrosamine emissions data is available (probably in
late 1979). In the event that EPA decides to pursue a final
rulemaking for these provisions on the basis of the new data, the
Public Docket will be reopened for comment specific to this topic.
A public hearing will be held if requested.
2. EPA is convinced that a control system employing a pump
can overcome turbocharger and heat exchanger efficiency and dura-
bility problems and can be developed for the 1984 model year.
-------
M. Issue - Numerical Standards/Standards Derivation
1. Summary of the Issue
EPA has proposed new emission standards for heavy-duty
engines:
1.3 g/BHP-hr HC
15.5 g/BHP-hr CO
10.7 g/BHP-hr NOx
The HC and CO standards are based upon a 90 percent reduction
from an actually measured uncontrolled baseline; the NOx standard
was derived to reflect no greater stringency than today's stan-
dards. All standards were based upon the transient test procedure.
2. Summary of the Comments
a. Inability to Comment
All manufacturers claimed a distinct inability to comment on
the proposed standards derived from the transient procedure. First
of all, the industry argued that lack of experience with the
transient test and lack of equipment to gain this experience
essentially deprived them of their opportunity to meaningfully
comment. EPA purportedly restricted the industry even moreso by
failure to propose actual standards with the 2/13/79 NPRM. In
Mack's words, this delay in announcing standards until May 1979,
plus overall inexperience with the transient test, effectively
resulted in a "deprivation of due process."
b. Standard Stringency
The industry also argued that the proposed standards, both
above and in the context of the remainder of the proposed rules,
uere substantially more stringent than Congress intended. The SO
percent HC and CO reductions called for in the Clean Air Act
Amendments were characterized as "merely targets." 'fee in context
with a 10 percent AQL and a full useful life, HC and CO reductions
in excess of 90 percent will be required. Several manufacturers
also characterized the 10.7 g/BEP-hr NOx standard, intended to
reflect equivalent stringency with today's standard, as more
stringent than required.
c. Standard Derivation
The methods of standard derivation were also criticized, both
in concept and in implementation. Cummins argued that EFA has done
no health effects study to support these proposed standards and
advocated standard setting based upon health effects, and not
simply percentage reductions.
-------
HC and CO standards were actually derived by a 90 percent
reduction from a baseline of in-use 1969 gasoline engines. Spe-
cific criticism of this Baseline Program and the resulting stan-
dards derivation were:
(i) Twenty-three engines were too small a sample.
(ii) The engines tested were unrepresentative, based upon
inconsistent accumulated mileages, too small displace-
ments, inappropriate pre-test tune-up procedures, and
unrepresentative in-use applications.
(iii) No deterioration factors were computed into the mea-
sured baseline emission levels, nor were the allowable
maintenance criteria required by the proposed rules
followed in the Baseline Program.
(iv) Test validation criteria were relaxed to such an extent
that unrepresentative and unrepeatable emission results
were generated. EPA's inability to stay within toler-
ances is indicative of flaws in the test procedure.
v) Proposed humidity tolerances were consistently exceeded
in the baseline program. The need for these tolerances
was questioned.
(vi) Motoring at -10 percent of maximum torque in place of
closed throttle resulted in a significant underestimation
of emissions, thereby lowering the baseline levels and
the resulting standards.
(vii) A sales-weighted emissions average by definition is
lower than some of the emissions used to generate the
average. Therefore, a standard based upon a 90 percent
reduction from this average represents more than a. 90
percent reduction for many of the engines tested. It
was suggested that a 90 percent reduction from the 90
percentile be used in deriving the standards.
Finally, EPA's derivation of the interim NOx standard was also
criticized. Industry argued that EPA's derivation resulted in too
stringent a standard.
3. Analysis
a. Inability to Comment
A detailed discussion of the industry's ability to comment on
the proposed rules and standards is contained in the Test Procedure
section of the Summary and Analysis of Comments. In short, the
industry has been regularly advised and informed throughout the
seven years that the transient test was being developed. EPA has
-------
openly broadcast its intention to promulgate a transient procedure
for several years. Data from the 1969 Baseline Program was regu-
larly disseminated; manufacturer's representatives personally
witnessed transient tests at the EPA lab in eary 1978. An MVMA
task force with participation by EPA representatives initiated
prototype transient testing at Cummins in the summer of 1977.
Since then the majority of the heavy-duty industry has done little
to acquire transient capability. The industry's inability to
comment is largely self-imposed. EPA can only reiterate the fact
that all data available to the Agency was freely and openly dis-
seminated.
Furthermore, EPA believes that sufficient technical infor-
mation was made available to allow well reasoned and accurate
analyses on the part of the manufacturers. Second-by-second cycle
listings were provided in the NPRM, allowing exact computation of
the engine's required operational modes. Gasoline engine manu-
facturers acknowledged unanimously that catalyst technology would
be necessary; in-use catalyst durability is the most difficult
technical hurdle for the industry to clear, yet assessments of on
the road catalyst temperature sensitivity are not dependent upon
the ability to run the transient emissions test. Two diesel
manufacturers were able to submit actual transient data on their
engines (Cummins and Caterpillar); transient diesel emission data
collected at SwRI on several engines was distributed to the in-
dustry for their analyses - data from over thirty baseline and
current technology gasoline engines was made available. In gen-
eral, a given manufacturer's inability to run a given engine over
the transient cycle did not preclude the industry's ability to
comment. Sufficient data was available to allow characterization
of the present state of the art of emission control and to allow a
reasonable judgement as to the viability of compliance technolo-
gies.
EPA is well aware that final numerical standards were not
proposed with the 2/13/79 NPRM because the 1969 Baseline Program
was still underway. Upon publication of the final numerical
standards and the technical report outlining baseline testing
methodology and results in May of 1979, EPA fulfilled its legal
obligacion by allowing an additional two and one-half month comment
period and an additional Public Hearing. Furthermore, che final-
ized standards were extremely close to the NPRM1s "best estimate"
of 1.4 g/BHP-hr HC and 14.7 g/BHP-hr CO. The NPRM also explained
that EPA would not finalize standards less than .76 g/BHP-hr HC and
11.4 g/BHP-hr CO without reproposing, in the unlikely event that
baseline emissions would end up that low. In short, the industry
had six full months and two Public Hearings to submit their com-
ments and opinions on the Proposed E.ules. For the first three
months of this comment period, EPA published a lower limit of
emission levels below which the standards would not fall. For the
last two and one-half months the final numerical standard were
available and open to public comments and hearings. In summary,
-2,66
-------
EPA has more Chan discharged its legal obligation to allow comment
on both the proposed rules and standards.
Standard Stringency
Section 202(a)(3)(A)(ii) of the Clean Air Act Amendment of
1977 empowered EPA to establish HC and CO emission standards which
represent a reduction of "at least 90 percent" from uncontrolled
levels, provided such reductions are technologically and economi-
cally feasible. The interim NOx standard is intended to be no more
stringent than today's standard.
It is EPA's position that the 90 percent reductions of HC and
CO represent the laxest standards desired by Congress, providing
the resulting reductions were proven to be technologically feas-
ible, cost effective, and directly relatable to improvements of the
public health* and welfare. The standards in themselves are no
more stringent than those required. Furthermore, these minimum
reductions are feasible at reasonable cost, and result in concrete
air quality benefits.**
In the context of the rest of the proposed rules, ECTD
recognizes that additional emission control will be required, but
only to insure that the mandated reductions will actually be
achieved on the road.
Most commenters argued that adoption of the transient test
procedure resulted in significantly more stringent standards.
There is no doubt that more effective emission control is required
at the levels of the proposed standards for the transient pro-
cedure. This is not indicative of greater stringency on the part
of the transient procedure, however, but rather an indication of
the laxity and inadequacy of the current steady-state procedures at
lower emission levels. Note that both procedures yielded compar-
able HC and CO emission levels for uncontrolled engines in the
1969 baseline, yet at the lower levels of current technology
engines the steady-state procedures seriously underestimate emis-
sions expected to be seen in-use. That the transient procedure
appears more stringent is due solely to the defeatability and
laxity of the current procedures.
The 10 percent AQL requirement allows no more than 10 percent
of the engines rolling off the assembly lines to exceed the stan-
dards as opposed to the 40 percent allowed for light-duty vehicles.
Furthermore, the full useful life concept requires increased
durability of emission-related equipment. The manufacturer has the
* See below, Section, "Standard Derivation" - health effects.
** See Regulatory Analysis and Summary and Analysis of Comments
for the particular issue.
-2.47
-------
option in both cases to select a lower low-mileage emission target
to compensate for production variability and for higher deteriora-
tion factors. The manufacturer has other options, however: to
enhance the durability of the emission control equipment and to
reduce production variability. Either of these options relaxes the
need for reducing low mileage targets and will be used to some
degree. ECTD recognizes however, that the predominantly used
option will most likely be lower target levels. In effect, this is
not an increase in the stringency of the actual standards, however.
These additional restrictions on compliance represent stringency
over and above the numerical standards, and assure in-use compli-
ance. (See the pertinent analyses on the Summary and Analysis of
Comments - Selective Enforcement Auditing and the Redefinition of
Useful Life, for arguments pertaining to justification.)
With regard to the stringency of the numerical HC and CO
standards, ECTD can only claim that they represent a true 90
percent reduction from the uncontrolled baseline, as specified by
Congress. Furthermore, as discussed in the Summary and Analysis of
Comments relating to Technological Feasibility, the standards are
achievable within reasonable cost. Finally, concrete air quality
improvements have been proven to be directly relateable to these
standards. The remainder of the package is designed to insure that
the 90 percent reductions will actually be achieved in-use.
c. Standards Derivation
Cummins took issue with EPA's concept of standard derivation,
claiming that EPA's concentration on strict percentage reductions
from a baseline was narrow minded and not cognizant of the true
basis of any pollutant standard, i.e., the protection of public
health and welfare. Cummins argued that EPA should have assessed
the health effects of HC and CO arising only from the heavy-duty
source, and set standards for heavy-duty based upon the impact on
public health of this single source. Cummins argued that use of
the National Ambient Air Quality (NAAQ) standards was n.ot an
adequate substitution for the Congressionally mandated pollutant-
specific study. In short, Cunanins claimed that standards should
have been derived per Section 202(a)(3)(E) of the Clean Air Act
Amendment, and not per 2Q2(a)(3)(A) as was done by EPA. Beyond
this broad, philosophical interpretation of the derivation proce-
dure, however, Cummins did not identify specific details and
methodologies of practical implementation. Furthermore, Cummins
did not quantify nor try to quantify the specific impact of this
philosophical approach upon the actual 1.3 g/BHP-hr HC and 15.5
g/BHP-hr CO standards.
EPA takes strong issue with Cummins assertion that the Agency
has "resisted a mandated regulatory process"* by purportedly
Cummins' August 14, 1979 Supplemental Submission, p. 11.
-------
ignoring health effects during the standard derivation process.
While no single document with the specific title "Pollutant Spe-
cific Study - Heavy-Duty Vehicle Sources" was published prior to
this rulemaking all information which would have appeared in such
a document, however, was published with the Draft Regulatory
Analysis accompanying the proposed rules, or referred to therein.*
The adverse health effects of overexposure to HC, CO, and NOx are
well documented in the literature and well known throughout the
Congress and the industry. It is EPA1s present intent to submit
the information contained in the Regulatory Analysis to Congress to
satisfy this reporting requirement.
Contrary to Cummins' assertions, the driving force behind the
proposed standard is the health issue, defined by EPA in terms of
the NAAQ standards. The standards were derived specifically to
define the maximum level of pollutant concentration that people
could be exposed to without experiencing adverse consequences to
their health. It was shown in the Draft Regulatory Analysis that
103 out of 105 urban areas of population in excess of 200,000 were
in violation of the NAAQ standards and therefore represent a
potential health risk to over 100 million people.
The 1980's will be a decade in which no given source of
pollution will be singled out as the major polluter. Any improve-
ment in air quality will not be accomplished by eliminating a
single source (e.g., light-duty vehicle standards represent the
lowest achieveable with current technology), but through a con-
certed effort directed at all contributing sources. The contri-
buting source addressed here are heavy-duty vehicles. Based upon
EPA's analysis, heavy-duty emissions could be reduced by 100
percent and still not bring most urban areas into compliance with
the NAAQ's. Given the health derivation of the NAAQ's , this
implies that any percentage reduction of heavy-duty emissions, even
100 percent, would be "health effective."
This is the basis for EPA's approach: given the fact that
violation of the NAAQ's are commonplace, any standard with which
compliance is feasible and cost effective cannot help but be
"health-effective" if it results in a tangible air quality im-
provement. EPA recognizes the fact that a reduction of 90 percent
is not immutable and has extensively reviewed the standards for
feasability and economic impact. In the context of the rest of the
proposed rules, however, a 90 percent reduction from the uncon-
trolled baseline is close to the maximum reasonably achieveable
with current and future technology at reasonable cost. It is
significant to note that Cummins could not identify a difference
* See Chapter IV OF Draft Regulatory Analysis, "Environmental
Impact." Also see Chapter III of "Air Quality, Noise and Health,"
Report of a Panel of the Interagency Task Force on Motor Vehicle
Goals Beyond 1980, March 1976.
-------
between EPA's proposed standards and standards derived per their
suggested approach. The ECTD staff believes that any incremental
standard reduction is health effective, and the level of standards
are impacted solely by the question of technological feasibility.
The support documentation for this regulatory action adequately
outlines the health-based rationale fcr the proposed standards and
satisfies the Congressional intent that such an analysis predicate
any standard derivation.
The standards derivation process embodied within the 1969
Baseline Program* received several procedural criticisms.
(i) Twenty-three engines comprised the data base from which
the standards were derived. Many commenters argued that more
engines were necessary to adequately characterize uncontrolled
emissions.
ECTD disputes this claim on the basis of Figures M-l and M-2.
Here the sales-weighted average emissions are presented as they
evolved with each additional engine added to the baseline, along
with the percent change of the average with each additional engine.
The last seven engines tested changed the baseline HC and CO
averages by no more than _+_1.6 percent. The last two engines
changed the HC average by 0.0 percent. Testing of more than 23
engines for baseline purposes would have been redundant and would
have delayed promulgation for no valid reason.
(ii) The engines tested in the 1969 Baseline were character-
ized by the industry as an unrepresentative sample. ECTD takes
issue with this assertion.
The 1969 baseline sample was designed to incorporate all
gasoline engines marketed in 1969 except those of very small sales
(less than 1-2 percent). The data used in deriving this sampling
plan was submitted by the manufacturers. Table M-l presents the
sampling plan, including 1969 market shares and actual engines
tested. No significant engine families ware neglected. Weighting
and averaging of the elements of a sample t:o represent the relative
proportions of those elements in the population is a standard
statistical technique for estimating population means. This was
done by sales-weighting the emission data in order to place higher
weight on the larger sellers, i.e., those engines whose emissions
would contribute more to >he overall average. In short., the
sampling plan includes all significant engine families, appropri-
ately weights their emission according to their relative contri-
butions to the whole and represents an adequately sized sample from
* Rafer to the EPA Technical Report "1969 Heavy-Duty Engine
Baseline Program and 1983 Emission Standards Development," &y T.
Cox, et. al., May 1979, in which the actual standards derivation
was explained.
•2.70
-------
which Che average emission level of all engines sold in 1969 could
be estimated.
Commenters also criticized EPA's pre-test tune-up procedures,
along with the purportedly unrepresentative applications from which
the baseline engines was drawn. A review of the regulatory method-
ology and the degree of problems encountered in its implementation
will serve to counter this criticism.
EPA was required by the Clean Air Act Amendments to derive
standards "from the average of the actually measured emissions from
heavy-duty gasoline-fueled vehicles or engines, or any class or
category thereof, manufactured during the baseline model year."*
(Emphasis ours.) This was interpreted to require actual emission
testing of 1969 (the baseline model year) engines. It was debated
within EPA whether to use in-use engines, or to seek cooperation
from the manufacturers so that essentially new engines, identical
in design and components to 1969 engines, could be built from
scratch.
This latter option was rejected because strict reading of the
Clean Air Act requires the test engines to have been manufactured
in 1969. Furthermore, the chances of adequately representing 1969
engines by remanufacturing per 1969 specifications were slim, and
the program would have been prohibitively expensive. Original
equipment for 1969 engines would in many cases have been unavail-
able, as would the manufacturing facilities themselves (retooled
to meet present needs). Special manufacture of individual engines
would have been time consuming and expensive. Manufacturers
stated that build-up and delivery of representative engines could
not be guaranteed, i.e., EPA's test schedule would have been at the
mercy of the regulated industry and have proceeded at their
convenience. This was unacceptable.
The only remaining alternative was to test in-use 1969 en-
gines, i.e., engines which by the time the baseline began had been
on the road in continuous use for nearly 9 years. The average
gasoline truck on the road in 1973 (when the 1969 engines were
entering middle sge) averaged about 11,300 miles per year.**
Assuming no major changes in trend, after 9 years of service the
average 1969 engine traveled over 100,000 miles and wasn't much
good for either commercial work or emission testing. EPA re-
stricted its selection of engines to those which were in original
configuration, i.e., engines which had not been rebuilt in the
block, shafts, and valves, which were equipped with original
* Clean Air Act Amendment 1977, Title II, Part A, Section
202(a)(3)(A)(ii).
** Derived from data presented within the "Draft Interagency
Study of Post-1980 Goals for Commercial Motor Vehicles" by a
Federal Interagency Task Force, June 1976.
-------
carburetors and distributors, and which passed basic checks of
mechanical integrity. These selection criteria were essential if
representativeness with the original 1969 fleet was to be main-
tained. Given the 9 years of continued use, however, engines
meeting these criteria were understandably rare. Over two thousand
inquiries were made and three hundred engines were inspected in the
field, resulting in the selection of the 23 engines. (The field
inspection was performed only if the engine met the original
screening criteria.) In short, the greatest care was exercised
throughout the procurement process to obtain the best engines
possible for testing, recognizing of course that "perfect" engines
were impossible to obtain. Futhermore, should deterioration
be application specific, only at gross levels of deterioration
should significant impact on emissions occur, arid at those levels
the selection criteria should have precluded inclusion in the
baseline. In summary, all efforts were made to insure that only
mechanically sound original engines were used in the baseline to
accurately reflect 1969 engines per Congressional intent. Prac-
tical limitations resulted in low-mileage applications being
preferentially selected. No data whatsoever was submitted to imply
that application-related errors were incurred.
To ensure that Congressional intent was complied with as fully
as possible, the philosophy of the pre-test engine preparation
procedures was to assure that the engine met all operation speci-
fications as prescribed by the manufacturer for the 1969 version.
Only in this way could the Agency be certain that 9 years on-the-
road had not produced uncharacteristic and improperly adjusted
engines whose emissions would be unrepresentative of the 1969
Fleet. ECTD even went so far as to personally deliver several
emission related components (carburetor, distributors) to indi-
vidual manufacturers for check-out and restoration. When re-
placement parts were needed, only OEM pares were used. The end
result of this philosophy was a complete tune-up and check-out of
all operational engine components; all adjustments were made
exactly as the manufacturers recommended to their customers in
The applicable service manuals. To do otherwise could have allowed
maladjusted and unrepresentative engines into Che baseline.
To summarize EPA's position on the representativeness of the
baseline angines, within Che realm of the possible, EPA took those
steps and actions which minimized errors and maximised both com-
pliance with Congressional intent and the technical validity of the
data.
(iii) SPA was criticized for purported failure to include
deterioration factors and failure to .comply with allowable main-
tenance procedures, as outlined in the Proposed Rules, during the
baseline program.
Due to the elaborate tune-up procedures and checks on median-
-------
ical integrity, it has been assumed by ECTD that the baseline
engines were restored to an effectively "as new" condition and
deterioration factors were effectively zero. Furthermore, certi-
fication data for non-catalyst gasoline engines reveal that these
engines have inherently low deterioration factors, and support the
zero D.F. assumption.
The question of allowable maintenance procedures is an insig-
nificant point with regard to the baseline. For durability test-
ing, allowable maintenance provisions preclude unrepresentative
maintenance to permit accurate characterization of deterioration,
i.e., to allow deterioration to occur and be measured. For base-
line purposes, however, the objective was to eliminate deterior-
ation by engine tune-up so that "as new" emissions could be mea-
sured.
In short, the engines tested in the baseline were character-
ized as new engines, and therefore required no computation of
deterioration factors. The Baseline was not a durability test
program and required no allowable maintenance constraints.
(iv) EPA's relaxation of the transient test validation
criteria by no means implies flaws in the test procedure, and by no
means guarantees unrepeatable and unrepresentative results. The
vailidation criteria set forth in the NPRM were derived from
limited transient experience acquired very early in the Baseline
Program. It must be stressed that the EPA transient facility was
the first of its kind in the entire motor vehicle industry and test
procedure developments and refinements occured throughout the
entire baseline program. The validation criteria were relaxed
based upon a recognition of the limitations of current dynamo-
meter/engine control system technology.
Furthermore, all data acquired in the Baseline Program and in
subsequent test programs at both EPA and the Southwest Research
Institute indicate that the revised tolerances more than guarantee
test repeatability and lab to lab correlation. (See "Tesc Proce-
dure" Summary and Analysis of Comments for further discussion on
the technical validity of the procedure.) EPA does recognize,
however, that a deliberate attempt to optimize emissions by running
a certain cycle up to the limits of the revised criteria could
defeat the intent of the procedure. With this in mind, EPA fully
intends to tighten the validation criteria in the future as exper-
ience is gained and technical improvements are made in the tran-
sient control capabilities of engine dynamometers. This is an-
ticipated to happen well before certification testing of 1984
engines if necessary.
(v) EPA acknowledges that humidity specifications proposed in
the 2/13/79 NPRM were consistently violated during the 1969 Base-
line Program. Humidity effects on HC and CO emissions, however,
a? 3
-------
are generally regarded to be minimal, as evidenced by the fact that
no humidity corrections are required for light- or heavy-duty for
HC or CO. Yet the difficulty experienced in humidity control along
with the high cost of such control, lead ECTD to the conclusion
that the humidity requirements be dropped for the Final Rules. In
its place will be an appropriate NOx correction factor.
(vi) Motoring at -10 percent maximum torque was chosen by EPA
for those portions of the transient gasoline procedure where
negative torques are desired.
The EPA gasoline dynamometer facility incorporates several
safety features which prevent injury to personnel and damage to
equipment. These include continual monitoring of system operation
by the support software, which in the cases of overspeed (e.g.,
engine runaways) and overload (e.g., a greater torque than the
engine driveshaft and dynamometer can withstand safely), will
automatically shut the facility down. These shutdowns occur both
when the command signal asks for greater than maximum permitted
speeds and load excursions, and when the system actually exper-
iences such excursions. The maximum possible torque excursion
permitted was j^400 ft-lbs; this was the maximum safe load on the
driveshafts, the in-line torquemeters, and the General Electric
motoring dynamometer- At several times during the transient
gasoline test, the engine is commanded to deliver wide open throt-
tle torque followed almost immediately by a motoring condition.
For an engine capable of delivery 360 ft-lbs at wide open throttle
(normally observed in the larger gasoline gengines), a torque
excursion from wide open throttle to a motoring command of greater
than -10 percent (i.e., 360 + 0.1 x 360 = 396 ft-lbs) would result
in both a commanded and actual torque excursion greater than the
400 ft-lbs allowable on the test equipment, at which point safety
overrides would stap the test.
It was in recognition of this fact that the original decision
was made to use -10 percent as Che motoring command, as opposed to
a completely closed throttle.
The level of motoring to be usea during the transient test can
only be based upon judgement. The only practical instrumentation
available for measurement of load factor parameters during the
CAPE-21 project could detect the fact that motoring was occuring,
but not its absolute level. Data is available, however, which
indicates that part-throttle and completely closed throttle both
occured frequently in normal use.
The use of -10 percent for motoring essentially recognized
equipment limitations of EPA's laboratory. General Motors sub-
mitted much discussion and theoretical analyses showing differences
in hydrocarbon measurements between part- and completely closed
throttle operation at higher engine speeds. ECTD cannot dispute
-------
these claims; restricting complete closure of the throttle is a
recognized technique for controlling hydrocarbons during the
motoring portion of the simplistic 9-mode test. It can be argued,
however, that changing the torque level of any mode will influence
emissions. As mentioned above, part-throttle motoring was observed
frequently in the CAPE-21 data base and its inclusion in the test
procedure is hardly unrepresentative. Baseline levels are defined
by the test procedure used; the 1969 Baseline used in standards
derivation incorporated both part-throttle and closed-throttle
motoring. (At lower speeds, -10 percent is sufficient to close the
throttle.) The certification procedure for 1984 gasoline engines
will also incorporate -10 percent motoring; the gasoline industry
will be required to test in a manner completely consistent with how
the standards were derived. In short, the standards are defined in
terms of -10 percent motoring. (Motoring in diesel engines,
however, is a different case. Whereas motoring emission in gaso-
line engines arise from air/fuel mixtures too lean to burn, the.
fuel in diesel engines is shut off during motoring at closed rack.
Therefore, motoring emissions of diesel engines are relatively
insignificant. The diesel test facility at SWRI has been capable
of running at closed rack at all speeds. There are no compelling
reasons not to run diesel engine at completely closed throttle if
its possible to do so; running diesels at partly closed rack
may even overstate emissions by a small amount.)
(vii) Comments were received questioning the derivation of
standards from a sales-weighted average, claiming that 90 percent
reduction from on average is actually greater than a 90 percent
reduction for those engines in the baseline with emissions greater
than average. ECTD can't dispute this. ECTD does note, however,
that the suggestion to derive standards from the 90 percentile of
the baseline is contrary to all regulatory history and the exact
wording of the Clean Air Act Amendment, which precisely stated the
standards were to be derived from an "average of the actually
measured emissions."
Aside from the 1969 Baseline, ECTD's derivation of the NOx
standard was criticized as resulting in too stringent a standard.
In this particular instance, ECTD disputes the contention of
stringency. As discussed in the Analysis of Comments pertaining to
Technological Feasibility, a transient NOx standard of 10.7 g/
BHP-hr is so lax as to represent a decontrol of the pollutant. No
gasoline or diesel engine tested at SPA, SwRI, or any other lab-
oratory on the transient test procedure has ever exceeded 10.7
g/BHP-hr. Only one gasoline engine has shown NOx as high as 9.7
g/BHP-hr. In the case of diesels, if all engines certified in 1979
can comply with an HC + NOx standard of 10.0 g/BHP-hr on the
current procedures, which measures higher NOx relative to the
transient, it is a misrepresentation of the facts to claim that a
transient NOx only standard of 10.7 g/BHP-hr is more stringent.
-------
4. Recommendation
EPA's derivation of the proposed standards was technically
complete, competently performed, and within the express direction
of Congressional intent.
Retain the proposed standards.
-------
Table M-l
Chrysler
(9.3%)
318-3
318-1
361
383
413
225
Baseline Sampling Plan
Manufacturer Engine Sales
10,850
10,150
7,000
2,000
1,500
1,000
of Market
0.4
0.3
Sampling
Target Range
Total
Actual
Procurement
1
1
1
0
0
1
Ford
(33.5%)
330
360
361
300
391
477
390
534
50,200
21,300
17,300
14,200
6,700
2,600
2,300
2,000
14.4
6.1
5.0
4.1
1.9
0.7
0.7
0.6
Total
2
2
2
1
1
0
0
GM
(39.3%)
350-2
366
292
351C
250
307
305C
477
350-4
396
47 , 000
22,000
18,000
12,000
10,000
9,000
6,600
6,300
3,000
2,000
13.5
6.3
5.2
3.6
2.9
2.6
1.9
1.8
0.9
0.6
3-4
1-2
1-2
0-1
0-1
0-1
0-1
0-1
0-1
0-1
Total (9-10)
3
2
1
1
0
0
0
0
0
IHC
(14.7%)
V345 20,500 5.9 1-2
7304 17,300 5.0 1-2
V392 7,600 2.2 0-1
RD450 3,350 1.0 0-1
VS478 2,000 0.6 0-1
Total (3-4)
2
1
1
0
277
-------
SflLES-HEIGHTED BflSELINE TRRNSIENT
EMISSIONS HC(G/BHP-HR)
+9.7%
0%
I.DO
3.00
5.00
7.00
11.00 13.00 15.00
NO. OF ENGINES
17JJO
19.00
21.00
23.00
25.00
-------
M
SflLES-WEIGHTED BflSELINE TRRNSIENT
EMISSIONS CO(G/BHP-HR)
\ - -1 17
\ \ -1
\
+0.3%
-1.5% -0.6% +0%
11.00 13.00 tS.OO
NO. OF ENGINES
17.00 19.00 21.00 23.00
I
25.00
-------
N. Issue - Fuel Economy
1. Summary of the Issue
EPA has proposed more stringent HC and CO standards for
heavy-duty vehicles. The proposed NOx standard is not considered
any more stringent than the current NOx standard. The issue is:
What effect will these new standards and associated procedures have
on fuel economy of heavy-duty vehicles?
2. Summary of the Comments
Relative t:o other issues proposed in the NPRM, the volume of
comments about the effects of the provisions in the NPRM on HD fuel
economy were rather meager. The focus of these comments can be
summarized into five categories: (1) The effect of the regula-
tions will prevent fuel economy improvements that could potentially
be obtained if the proposed regulations were not in effect; (2)
The proposed regulations will cause a fuel economy loss (no speci-
fic cause ever cited); 3) The proposed regulation will cause a
fuel economy loss due to the NOx standard; 4) The proposed regu-
lations will cause a fuel economy loss due to a more stringent AQL
(10% over 40%); and 5) The proposed crankcase regulations will
cause a fuel economy loss on turbocharged diesel engines.
Comments that discussed fuel economy foregone most notably
the Council of Wage and Price Stability claimed past LDV data would
indicate a 5-10% fuel economy loss. Caterpillar predicts the
proposed emission regulations "... could have ..." around a
2.6% per year fuel economy improvement forgone due to shifting
resources from fuel economy improvement to emission control.
Ford estimate a 10-15% outright loss in fuel economy, and
Chrysler simply stated "... chat the adverse effects on fuel
economy will be sufficiently great to also provide grounds for a
revision of the standards."
Both GM and Cummins discussed fuel economy impacts in relation
tc California NOx standards, but not in relation to the proposed
provisions in the NPRM.
hack provided an analysis based on the current 13-taode diesel
test procedure indicating a 0.98% improvement in brake specific
fuel consumption by selecting a 40% AQL over a 10% AQL.
Caterpillar predicted a 1-3% fuel economy loss on turbocharged
diesel engines due uc crankcase emission controls.
3. Analysis of the Comments
The commenns on fuel economy cover both gasoline-fueled and
-------
diesel engines. Because of the different control strategies
anticipated, it will be easier to discuss these engines separately.
a. Gasoline-Fueled Engines
The comments on gasoline-fueled engine fuel economy could
generally be characterized as statements expressing opinion but
lacking supporting data. The Council on Wage and Price Stability
did provide some reference material to support their claim of a
5-10% fuel economy penalty. The documents cited include an out of
date 1974 CRC study, and manufacturers comments during public
hearings held in early 1977 and late 1978.
The Council of Wage and Price Stability claim that the light-
duty data cited would indicate "non-trivial" fuel economy penalties
of 5-10%. In making the transfer between light-duty vehicles (LDV)
and heavy-duty gasoline-fueled (HDG) vehicles the Council of Wage
and Price Stability apparently failed to look at the relative
differences in current emission control systems (between LDV and
HDG).
An analysis of more recent data indicates that it will be
entirely possible for the fuel economy from HDG vehicles to in-
crease as much as 17% with the application of catalyst technology.
A more conservative estimate would be a fuel economy increase
between 4 and 9%.
Since these statements are directly contradictory to the
comments submitted by the interested parties, a review of the
historical facts involving fuel economy effects of emission con-
trols on LDV fuel economy will be given. One of the most recent
papers on that subject (SAE Paper 790225) presents test data on
over 6,000 cars ranging from pre-controlled model year vehicles
through 1979 model year vehicles. The fuel economy data was taken
on the same test procedure (75 FTP) for all vehicles. Table N-l
tabulates this data.
Before discussing this data two assumptions should be dis-
cussed. One, the LDV city fuel economy values will be used to
compare to HDG transient test values. The reason behind this
assumption is that the city cycle exercises the LDV in a transient
manner more than the highway cycle, and therefore, would provide a
better comparison (of the two cycles) to the transient HDG cycle.
The second assumption is not necessarily an assumption, but a
selection of a reference point for analyzing the data. The 1974
LDV model year is selected as a point for initial comparison. For
light-duty vehicles, the 1974 model year represents the last model
year prior to wide spread oxidation catalyst (OC) usage. The 1974
model year could be characterized as "just before catalyst central
era". If it can be assumed that wide spread catalyst usage will
-------
Table N-l
Trends in Sales - Weighted
Fleet Fuel Economy, Passenger Cars I/
Pre-Control
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1975
Each
City
12.9
12.6
12.6
12.6
12.3
12.2
12.0
12.0
13.7
15.2
16.0
17.0
17.6
Year's
Hwy.
18.5
18.4
18.6
19.0
18.2
18.9
18.1
18.2
19.5
21.3
22.3
24.1
24.3
Weight
55/45
14.9
14.7
14.7
14.8
14.4
14.5
14.2
14.2
15.8
17.5
18.3
19.6
20.1
Mix:
Avg.
Test Wt.
3812
3863
3942
3877
3887
3942
3969
3968
4057
4060
3943
3649
3508
1974
City
12.5
12.4
12.7
12.3
12.1
12.2
12.0
12.0
14.0
15.5
15.6
15.4
15.3
Weight
Hwy.
17.4
17.6
18.8
18.5
17.9
18.9
18.1
18.2
19.9
21.8
22.1
22.4
21.2
Mix:*
55/45
14.3
14.3
14.9
14.5
14.1
14.5
14.1
14.2
16.1
17.8
18.0
18.0
17.5
Average Test: Weight = 3968 Ib.
-------
occur on HDGs with the proposed standards, then the 1979 HD interim
standards and associated emission control technology is analogous
to the 1974 LDV model year technology.
Table N-2 provides a comparison of historical fuel economy
data. The city fuel economy data for LDVs comes from the 1974
weight-mix category in Table N-l. The 1974 weight-mix category
compares the fuel economy from all LDV model years on a constant
weight basis. The comparison of LDV fuel economy on a constant
weight basis is considered more representative of heavy-duty
vehicles since reducing vehicle weight would not generally be an
option for heavy-duty vehicles. It would be assumed that any HD
vehicle weight reduction would be taken up by increased payload.
In addition to the LDV-HDG comparison, light-duty truck. (LOT)
data shows similar fuel economy improvement trends. Data from
recently completed LOT baseline testing (1969 and 1973 model years)
is presented in Table N-3 along with LOT data from SAE Paper
790225._!/ The SAE Paper does not calculate a constant weight mix
for LDT's as the paper did for LDV's (Table N-l). Therefore, the
baseline fuel economy is presented as the baseline sales-weighted
inertia weight (IW) versus the same specific weight class for the
later model LDT's presented in reference _!_/. Inspection of Tables
N-4 _3_/ and N-5 kj indicates that this is a reasonable assumption.
Returning to the LDV-HDG comparison, it is evident that by
going from pre-emission control technology to pre-catalyst control
technology (pre-'68 to 1974), a fuel economy penalty of 4% was
incurred by LDVs (Table N-2). Test data from pre-controlled 1969
HD engines (Table N-6)_5_/ and pre-catalyst 1979 engines (Table
N-7)_6_/ indicate approximately the same order of fuel economy
penalty was incurred by HDVs. It should be pointed out here that
the fuel economy penalty incurred by HDGs is over estimated and
should be somewhat smaller than that indicated. This overestima-
tion occurs due to the fact that contrary to popular opinion,
vehicle fuel consumption improves with age (see references 1 and
2). Since the pre-controlled 1969 engines were tested at signifi-
cantly higher mileages than the pre-catalyst 1979 engines, correct-
ing the fuel economy of the pre-controlled engines to the pre-
catalyst mileage values would increase the brake specific fuel
consumption (BSFC) of the pre-controlled engines, and thereby,
decrease the difference between the two categories. It should be
pointed out that all the LDV test data was corrected to a 4,000-
mile fuel economy value.
After 1974, and once LDV catalyst technology was introduced,
the fuel economy of light-duty vehicles increased rapidly. Table
N-l does show some fluctuation in the increase. However, during
these years ('76, '77, and '78) certain automobile manufacturers
used ambiguities in the test procedure in order to obtain higher
fuel economy results. The somewhat lower mileage values for the
-------
Table N-2
Fuel Economy Comparison*
LDV
HDG
Model Year City MPG** % Change
Pre-control 12.5 —
1974 12.0 -4.0%
1975 14.0 +16.7%***
1979 15.3 +27.5%***
Model Year BSFC+ Z Change
Pre-Control++ .688 —
1979+++ .721 -4.8%
1984 ( ) ( )
* Sales Weighted Average.
** Constant Vehicle Weight Basis, 1975 LDV FTP.
*** Based on 1974 mpg.
+ Transient Engine Test (Ib/hp-hr) constant KP/weight basis
++ 1969 Baseline._5_/
+++ 1979 Baseline.6/
-------
Table N-3
Trends in Light-Duty Truck
Fuel Economy (LPT) vs Emission Standards
Fuel Economy (75
Weight Class 4000 4500
Model Year
1969
1973
1975
1976
1977
1978
1979
11.89_b/
-
13.83 12.01
15.86 12.81
16.70 14.85
16.01 13.91
15.42 13.58
FTP) a/ Standards (75 FTP)
5000+ HC CO NOx
11.04_c_/
10.02_d/ 2.0 20 3.
11.17 2.0 20 3.
10.73 2.0 20 3.
15.96_d/ 2.0 20 3.
16.85_e/ 1.7 18 2.
1
1
1
1
3
a_/ MPG 1975-1979 values from reference IJ.
b/ 1969 Baseline, sales-weighted IW = 4680._3_/
7/ 1973 Baseline, sales-weighted IW = 4917.4_/
d/ 1975-1978 LOT class excludes vehicles greater than 6000 Ibs.
GVW, 1978 mpg reflects some dieselization.
e/ 1979 LOT class includes vehicles up to 8500 Ibs. GVW, 5000+
class reflects some dieselization.
f/ g/mi.
-------
Table N-4
1969 Light-Duty Fuel Economy Baseline 3/
Vehicle
Number
404
428
441 •
618
607
418
444
601
419
427
450
602
421
425
473
491
610
613
Sales
Weighting
Factor
(%)
1.32
1.32
11.31
8.54
34.18
1.99
1.99
0.84
2.35
2.35
2.35
2.35
4.85
4.85
4.85
4.85
4.85
4.85
Inertia
Class
5500
5500
4500
5000
4500
4500
4500
5000
4500
4500
4500
4500
5000
4500
5000
5000
5000
5000
Weighted
Inertia
Class
66.00
66.00
508.95
427.00
1538.10
89.55
89.55
42.00
105.75
105.75
105.75
105.75
242.50
218.25
242.50
242.50
242.50
242.50
4680.90
Fuel
Economy
(mpg)
12.03
13.68
13.22
11.17
12.76
11.81
13.26
10.93
12.00
12.03
10.84
11.21
12.39
7.61
11.73
11,08
9.65
10.80
Weighted
Fuel
Economy
(mpg)
0.1587
0.1806
1.4952
0.9539
4.3614
0.2350
0.2639
0.0918
0.2820
0.2827
0.2547
0.2634
0.6009
0.3691
0.5689
0.5374
0.4680
0.5238
11.8914
-------
Table N-5
1973 Light-Duty Fuel Economy Baseline _4/
Vehicle
Number
612
637
634
631
629
628
632
608
486
609
627
605
630
620
624
625
635
611
Sales
Weighting
Factor
(%)
6.51
6.51
6.51
6.51
6.51
2.93
1.52
8.03
3.76
3.76
3.76
3.80
8.00
8.00
8.00
5.75
8.90
1.19
Inertia
Class
5000
5000
5000
5000
5000
5000
5000
5000
4500
5500
5500
4500
5000
4500
4500
5000
5000
5000
Weighted
Inertia
Class
325.5
325.5
325.5
325.5
325.5
146.5
76.0
401.5
169.2
206.8
206.8
171.0
400.0
360.0
360.0
287.5
445.0
59.5
4917.3
Fuel
Economy
(mpg)
10.93
11.49
11.58
11.58
10.75
12.15
14.17
9.28
11.76
9.43
10.45
10.66
11.94
11.06
11.08
11.45
10.71
10.15
Weighted
Fuel
Economy
(mpg)
0.7116
0.7480
0.7539
0.7539
0.6998
0.3560
0.2154
0.7452
0.4422
0.3546
0.3929
0.4051
0.9552
0.8848
0.8864
0.6584
0.9532
0.1207
11.0373
-------
TARLE N6:
SALFS-WEIGHTtu 1H4NSILNT ENGINE EMISSIONS(G/UHP-HHI
IV(->9 HASELINE EM(ilNEtS)
ENGINE
01 FW 225H 2994 032
225 1
02 V392 658417
392 1
03 391-JW
391 1
04 V304 64804B
304 1
05 F330 9AN505S-
330 1
06 GM351 24B3434
351 1
07 F330 9UN505S
330 1
08 GM350 V0512XI
350 2G
09 D318 PM 31BR
318 3
10 V345 31960C
345 1
11 r.M 350 2 LJPN
350 20
12 F300 1
300 1
13 V345 719456
345 1
14 nM36fi ARKUCKLE
366 1
15 F361 SHOE
361 1
16 F360 EGG1
360 1
17 OM292 RACKET
IB 0318 EGG2
318 1
19 F361 BLE 19
361 1
20 F360 EGG3
360 1
21 GM350 TENNIS
350 2G
22 D161-3 SLUG
361 1
23 GM366 SKR1
366 1
C 1 7C
j I r E-
FACTOR
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.00368
0.02699
0.02331
0.06135
O.OM834
0.04417
0.08834
0.05521
*
0.03804;
,
0.03620
0.05521
0.05031
0.03620
0.03865
0.03067
0.03742
0.06380
0.03558
0.03067
0.03742
0.05521
0.02454
0.03865
2
3
2
5
5
2
3
2
3
2
3
3
3
3
2
3
2
3
3
2
3
2
3
HC
7.20
6.35
13.55
11.21
28, 1 )
9.72
34.16
9.40
7.9,
7.12
6.21
7.81
6.41
8.59
14.12
7,96
8.5'.
0.82
9.57
5.92
8.64
12.63
8.53
r-wrtus /
CO
52.20
178.47
179.34
127.76
157.15
111.51
224.37
170.77
H6.97
76.53
126.13
233.38
94,02
187.92
228.39
132.1"
172.86
144.26
1^7.55
75.32
Ib0.36
168.68
134.87
' KHr'-HK
MOA P/l
8.46
4.24
5.83
6. 70
7.H9
8.80
6.25
4.H2
7.60
6.4b
5.36
4.91
5.59
5.32
5.43
6.63
5.14
7.54
5.H9
6.88
4.58
6.111
4.66
Rf HC
0.026
0.171
0.316
0.688
2.485
0.430
3.018
0.519
0.303
0.25b
0.343
0.393
0.232
0.332
0.413
0.298
0.545
0.314
0.294
0.221
0.477
0.310
0.330
ViEIGHlEO G/MHP-HR
CO
0.192
4.818
4.181
7.H38
13.883
4.926
19.822
9.429
3.308
2.770
6.U64
11.741
3.403
7.263
7.006
4.947
11.029
5.133
6.060
2.819
B.302
4.119
5.213
HOX PART BSFC
0.031
0.11'.
0.136
0.411
0.697
0.389
0.553
0.266
0.2P9
0.234
0.2^6
0.247
0.202
0.206
0. 167
0.248
0.328
0.268
0.156
0.258
0.253
0.148
0.180
0.6390
0.777T
0.6465
0.682B
0.727?
0.6520
0.7503
0.66HO
0.5993
0.7110
0.6157
0.6940
0.6131
0.7 J73
0.7795
0.6553
0.7615
0.6603
0.68V1
0.6355
0.6443
0.6850
0.6651
*— WF 1 (tHTFO
^ WLlv'^lf
MSFC
0.00235
0.02099
0.01507
0.04189
0.06424
0.02BBO
0.06629
0.03688
0.02280
0.02574
0.03399
0.03491
0.02220
0.02850
0.02391
0.02452
0,04859
0.02350
0.021 15
0.02178
0.0 J55H
0.01681
0.02572
SALES-WEIGHTED GAS BAG TOTALS?
90* REDUCTION FROM BASELINE!
12.74 15S.IH 6.08
1.27 15.52 0.61
0.0
a
-------
ENGINE
09 IHC446
MV-8 5
02 V345C 79HLE-2
V-345 3
03 GM366 79HLE-3
114 I
04 OM350 79HLF-4
113 3
OS F400 79dlE-b
6.6L "t"9-73J
06 F370 79HLt-6
07 C360 79t)LE-7
LA-1 CAl-4
08 C440 799LE-8
PDM CR3-2
10 GM454 79HIE-10
us i
05 GM?92 125 CTE5
112 1
02 GM<»54 CTE2
114 3
6 GM350 CTE6 •
113 1
D TKANSltNT ENUlME tMlSSIOMS (G/MHP-HH)
IV79
31
PEFERENCE
0
0
0
0
0
0
0
0
0
0
0
0
WTG.
FACTOR
O.OHH2
0.02786
O.OB415
0.04277
0. 1244R
0.1 OJ 17
0.1 183P
0.09755
0.06189
0.03205
0.00338
0.22529
C 1 7C
b 1 / r_
3
3
3
3
3
2
3
3
3
1
3
1
HC
3.27
2.44
2.16
2.4H
4.89
3.5.1
2. (.7
3. 83
1.31
2.12
2.36
2.66
CiKrtMS
CD
90.40
34.44
43.43
64. 76
112.43
47.75
96. 10
1 12.38
78.49
54.98
55.36
114.02
/ HHM-HR
mix pr
5.^8
6.46
8,42
6.62
4.29
5.54
4.36
4.48
6.23
9. 74
6.b5
6. 58
iKT HC
0.265
0.068
0.182
0. 106
0.608
0.355
0.316
0. 373
O.OR1
0.068
0.008
0.599
WEIGHTED
CO
7.333
0.959
3.655
2.770
13.995
4.83H
11.369
10.962
4.857
1.762
0.187
25.689
G/HHP-HR
MOX PART
0.445
0. 180
0.708
0.283
0.534
0.561
0.516
0.437
0.385
0.312
0.022
1.482
BSFC
0.7160
0.6500
0.7190
0.7167
0.7463
0.7795
0.6890
0.6813
0. 7653
0.6550
0.6677
0.7270
^ ""
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
Wf tr.HTFO ->
HC.iVj"MC.L/ r
HSFC
05808
01811
06050
03065
09290
07886
08151
06647
04737
02099
00226
16379
SALES-WEIGHTED GAS BAG TOTALS:
3.03 6B.37
5.87
0.0
0.721<«8
-------
1979 model year (and to some extent for 1978) reflect the correc-
tion of these ambiguities to ensure that these manufacturers would
perform the test pxoperly. The important aspect of the data,
however, is that the trend between 1974 and 1979 was one of con-
tinued improvements in LDV fuel economy.
The LDV fuel economy increased over these years in spite of
the fact that more stringent emission standards were enacted over
the years. Table N-8 compares the effect between pre-catalyst
technology and catalyst technology on emissions and fuel economy.
As Table N-8 indicates LDV HC and CO were reduced substantially
between 1974 and 1975. During this emission reduction fuel economy
improved approximately 17%, Comparing the HDG HC and CO reduc-
tions, and the allowable HDG NOx increase to the LDV historical
data, it certainly seems reasonable to expect that HDG fuel economy
will increase with the proposed standards.
Some may claim that the LDV fuel economy improvements are due
to effects other than improved engine control technology. Such
factors as vehicle streamlining, power train optimization, and
changes in power to weight ratio do in fact account for a portion
of the change in LDV mileage, especially the 1979 figures. How-
ever, there were very few of these changes between the 1974 and
1975 LDV model years. Therefore, it must be assumed that most of
the almost 17% improvement in LDV fuel economy was directly attri-
butable to improved engine control technology.
Others may claim that it isn't proper to compare just the
standards (Table N-8) between pre-catalyst to catalyst technology,
or pre-catalyst HD emission levels to catalyst standards without
considering pre-controlled levels, and the potential effects of
such issues s.s the proposed changes in useful life, a 10% AQL
Selective Enforcement Audit (SEA) limit, etc. Table. N-9 provides
this comparison. The comparison between LDV and HDG is based on
emission levels from various test programs. An attempt was made to
compare emission levels based on similar service accumulation. For
instance, the average service of the LDV test vehicles was 70,000
miles while for HDG it was 60,000 miles. The service accumulation
for the comparison of the ore-catalyst and catalyst technology
emission levels is based on the interval used for emission data
vehicles or engines. In this manner, the impacts of durability and
SEA can be evaluated as low taileage target {LMT) levels for emis-
sion data engines. The derivation of LMT levels for HC and CO is
discussed in the Cost Effectiveness Chapter of the Regulatory
Analysis document (Chapter VII). A similar procedure was used to
derive the LMT for NOx.
Comparing the stringency cf the HD proposal (Table N-10) to
past LDV daca, it is apparent that the estimated HD LMT levels
represent an increase in the reduction of HC by approximately II
percent and 23 percent for CO over that experienced by LDVs in the
2.10
-------
Table N-8
Comparison of Fuel Economy and
Emission Reduction versus Control Technology
a/
b/
c/
d/
I/
f/
Pre-
Catalyst
HC LDV 3.0 a/
HDG 3.03 c/
CO LDV 34 a/
HDG 88 c/
NOx LDV 3.1 a/
HDG 5.87 £/
F.E. LDV 12.0 e/
HDV .721 f/
1974 LDV Standard of 3.4/39/3.
1975 FTP results, g/rai.
1975 LDV Standard, g/mi .
1979 Transient Baseline, g/BHP-hr
Proposed 1984 Standard, g/BHP-hr.
Catalyst
Technology
1.5 b/
1.3 d/
15 b/
15.5 dV
3.1 b/
10.7 d/
14.0 e/
( }
0 expressed
.
Percent
Change
-50.0
-57.1
-55.9
-82.4
0.0
+82.2
+ 16.7
( )
as approximate
Constant weight basis mpg, reference _!/ .
1979 Transient Baseline, Ib
fuel/BHP-hr,
reference 6/ .
-------
Table N-9
Comparison of LDT and HDG
Emission and Fuel Economy Trends n/
Catalyst Technology
HC
CO
NOx
F.E.
LDV
HDG
LDV
HDG
LDV
HDG
LDV
HDG
Pre-control
8.74 a/
12.74 b/
86.5 a/
155.18 b/
3.54 a/
6.08 b/
12.5 i/
.688 _!/
Pre-catalyst
3.08(-64.8) c/
3.03(-76.2) dj
35.92C-58.5) c/
88.37C-43.1) d/
2.90(-18.1) c/
5.87C-3.5) d/
12.0C-4.0) j/
.72K-4.8) m/
LDV Emission
Factors/HD LMT
1.32C-84.9) e/
.50C-96.1) £/
22.92C-73.5) e/
5.9(-96.2) Jf/
2.44C-31.1) e/
7.0( + 15.1) f_/
14.0(+12.0) k/
( - )
Standards
1.5(-82.8) g/
1.3C-89.8) h_/
15C-82.7) g_/
15.5C-90.0) h/
3. K-12.4) g_/
10.7C+76.0) h/
-
a_/ Surveillance Test Data, g/mi,, 1965-67 LDV, avg mileage 70,000,
Reference _8_/.
_b/ 1969 HD basline, g/BHP-hr, avg mileage 60,000, Reference _5/ •
_c_/ Surveillance Test Data, 1974 LDV, avg mileage 6,000, Reference
!/•
d_/ 1979 HD baseline, g/BHP-hr, emission-data engines, Reference
!/•
e/ Surveillance Test Data, g/mi, 1975 LDV, avg mileage 8,000,
Reference _8/.
_f/ Estimated low mileage target (LMT) emission levels for 1984
HD emission-data engines.
£/ 1975 LDV standards, g./mi,
h/ Proposed 1984 HD standards, g/BHP-hr.
if Constant weight basis LDV pre-control mpg3 Reference I/.
j/ Constant weight basis, 1974 LDV mpg, Reference _!_/.
_k/ Constant weight basis, 1975 LDV mpg, Reference _!_/-
I/ 1969 baseline BSFC, Ib/BHP-hr, Reference _5/•
ml 1979 baseline BSFC, Ib/BHP-hr, Reference _6/.
n/ Values in parenthesis represent change from pre-control value".
-------
Table N-10
Fuel Economy Effects and Comparison of
Stringency of LDV and Proposed HP Standards a/
HC
CO
NOx
F.E.
LDV
HDG
LDV
HDG
LDV
HDG
LDV
HDG
Precontrol
8.74
12.74
86.5
155.18
3.54
6.08
12.5
.688
Stringency of Emission
LDV Emission Factors/
HD LMT
-84.9%
-96.1%
-11.2%
-73.5%
-96.2%
^2277%"
-31.1%
+15.1%
46.2%
+12.0%
( - )
Reductions b/
Standards
-82.8%
-89.8%
- 7.0%
-82.7%
-90.0%
- 7.3%
-12.4%
+76.0%
+88.4%
_
aj Values taken from Table N-6.
b/ Catalyst technology.
-------
first year of catalyst control. The LMT HD NOx level actually
represents a decrease in stringency of about 46 percent compared to
LDV levels.
EPA test data shows that these LMT targets for HC and CO can
easily be obtained. The Summary and Analysis of Comments on
Technological Feasibility describes an experiment in which cata-
lysts were added to a 1978 404-CID engine. HC and CO emissions
could be incrementally eliminated by incrementally increasing the
flow in the AIR injection system. Test data indicates that in-
creasing the AIR injection rate cost about 2.5% to 4% loss in fuel
economy for every 20CFM increase in flow rate.
The LDV data (Table N-10) shows that the LDV fuel economy
increased by 12% on a constant weight basis between pre-control and
the first year of catalyst technology. This fuel economy increase
included the penalty incurred due to the addition of an air pump.
The HD experience indicates that the 11% and 23% HC and CO
differentials in emission reductions between LDV ani HD can be
accomodated by increasing the AIR injection rate by 20 to 40 CFM
over the LDV rate. Assuming a 4% fuel economy loss per 20 CFM
increase, a 4-8% fuel economy penalty for HD engines would be
incurred. However, since the data shows LDVs experienced a 12%
increase in fuel economy with similar emission reductions, the 4-8%
HD penalty would be subtracted from the 12% increase due to the
addition of catalyst technology. The result would be a 4-8%
increase in HDG fuel economy over the pre-controlled versions.
Table N-ll shows the emission results of the limited EPA
experiments on the 404-CID engine. The amount of engineering time
that went into the selection of hardware that resulted in these
substantial reductions was very minimal. Possibly 4 to 8 person-
hour's were involved in the selection of initial hardware, and in
the iterations from one modification to the other. Build up and
testing time consumed maybe another 20-30 hours over a 2- to
3-week time span. Other than the catalysts and air pumps added to
the engine, no other modifications were made. So even though this
testing showed a small fuel economy loss, in as little as 2 weeks
practically anybody would be able to modify this engine to show a
fuel economy improvement, probably up to 10% over the standard
configuration, and still meet the LMT levels.
Table N-12 presents data on another engine that EPA modified
with the addition of oxidation catalysts (no other modifications).
In this case, a comparison within the engine line could be made
from data repressnting the 1969 pre-coutrclled conditicn, the 1979
certified condition, and a modified 1984 condition. The data
(Table N-12) indicates that not only did the modified version
reduce emissions to less than the proposed standards, the engine
also produced nore horsepower and obtained better fuel economy than
the 1969 counterpart.
-------
Table N-ll
Fuel Economy and Emission Comparison a_/
404-CID HD 8-CyUnder Engine
1978 California Calibration b/ (EGR/AIR)
172 BHP @ 3748
guration
HC
3.98
CO
54.56
NOx
5.01
BSFC
.680
(+4.0) e/,
(20 CFM Air)
Modified Configuration 0.28 8.98 4.09 .708 d/
w/Ox Catalyst, 40 CFM Air) (-93.0) (-83.5)
Modified Configuration 0.32 3.74 c_/ 3.98 .765 e/
w/Ox Catalyst, Simulated (-92.0) (-93.1) (-9.0)
80 CFM Air
a/ Transient test, g/BHP-hr; BSFC, Ib/BHP-hr. Numbers in paren-
thesis represent change from standard configuration.
W 1978 California standards, steady-state test, 1.0 HC, 25 CO,
7.5 NOx, g/BHP-hr.
c_/ Emission reduction exceeds LMT of 5.9.
d/ Measured transient BSFC.
_e_/ BSFC extrapolated from the same day comparison of WOT BSFC vs.
varying air injection flow rates, see analysis in Summary and
Analysis of Technological Feasiblity.
_f/ Previous transient data (not same day comparison) in standard
configuration indicates a BSFC of .672 (+5.1).
3.?:
-------
Table N-12
Fuel Economy and Emission Comparison *
292-CID HP 6-Cylinder Engine Line
HP (3 rpm
Pre-Control
Configuration
20 CFM Air
1979 W/Catalyst, 114 @ 3760
20 CFM Air
HC
CO
0.58
12.25
NOx
7.30
BSFC
109 (3 3546
114 @ 3760
8.54
2.12
172.86
54.98
5.40
9.74
0.761
0.655
0.638
Transient test, g/BHP-hr; BSFC, Ib/BHP-hr.
-------
The previous analysis on both engines, did not include any
potential fuel economy improvements available through system
optimization due to the decrease in NOx stringency. Another aspect
not to be overlooked is that a 4-8% HD fuel economy improvement
over pre-controlled engines (1969) represents a 9-13% improvement
over pre-catalyst engines (1979). (Table N-2 shows that the
average pre-catalyst engine incurred a 4.8% penalty compared to
pre-controlled engines.)
Based on the previous discussion and the incredible ease in
which the emissions were reduced on the 404-CID engine (Table
N-ll), the promulgation of the 1984 HD emission standard will not
cause a fuel economy penalty even at the HDG LMT levels, and will
in fact probably allow an improvement in HDG fuel economy.
The final estimated amount of fuel economy increase can be
evaluated in two ways. One, would be to assume the full 17%
improvement between pre-catalyst and catalyst technology (1974 to
1975) LDVs could be obtained by HD engines. Then, the net HD fuel
economy improvement could be calculated by substracting off the
penalty incurred due to the additional AIR injection. However, not
all pre-catalyst HD engines (1979) experienced the full HD fleet
average fuel economy loss. These engines would tend to set a lower
limit on fuel economy improvement. To calculate the lower limit,
the estimated fuel economy improvement could be obtained by sub-
stituting the LDV pre-control to catalyst technology (pre-1968 to
1975) fuel economy improvement of 12% for the 17% figure, and once
again subtract off the penalty incurred by the additional AIR
injection.
Assuming many engines would require an additional 40 CFM AIR
injection rate, the range of estimated fuel economy improvement
resulting from the proposed 1984 HDG standards would be between 4%
(12-8) and 9% (17-8).
From the data presented some may argue that even though the
implementation of catalyst technology may cause a net improvement
in HDG fuel economy, the higher AIR injection flow rates necessary
for the catalyst technology rob power from the engine. In other
words, the additional AIR flow creates a fuel economy foregone
issue. While it cannot be denied that increasing AIR flow rates
does increase engine fuel consumption relative to the useable power
output, we suggest that this issue cannot be viewed from that
perspective.
We suggest that the alternatives involved in this issue are:
(1) propose a standard that forces catalyst technology and also
allows an HDG fuel economy improvement 4-9 percent, versus, (2)
staying with standards that do not force catalyst technology.
a??
-------
Alternative number 2 by necessity would rely on engine modi-
fications (more refined, but similar to those encountered on 1974
LDV's and 1979 HDG's) to control emissions. In this case, as in
Table 2, we could expect a decrease in fuel economy. So, we
suggest that the choice is between a standard that allows a fuel
economy improvement as well a providing significant ambient air
quality benefits, versus a standard that does neither.
b. Diesel Engines
It is interesting to note that not one manufacturer claimed
the proposed diesel emission standards would cause a fuel economy
loss. Those that claimed a fuel economy loss, claimed others
factors in the package such as crankcase emissions, and AQL limits
would cause the fuel economy loss.
The one exception to these comments is the Caterpillar claim
that in order to meet the proposed emission standards, Caterpillar
would have £o shift resources from fuel economy development to
emission control development. Tnis comment is a difficult one to
discuss because it is so subjective for diesel engines. Cater-
pillar provided the fuel economy improvement trends on one engine
line. Based on the Caterpillar data the yearly improvement in fuel
economy for this engine is leveling out in the better portion of
the normal range of diesel fuel consumption. For some unexplained
reason Caterpillar claims the leveling out trend will suddenly take
a sharp change in 1985. Further, Caterpillar qualified this claim
by "Figure 8 illustrates the anticipated impact that the proposed
emissions could have on the fuel economy trend." (Underlining
added for emphasis.)
Caterpillar did not expand on the reasons for the drastic
change in fuel economy improvement beginning in 1985. It is
assumed that if such a change were to take place, it would be the
result of new technology. Since none of the other diesel manu-
facturers directed comments to this area, we can only assume
that the marketplace will force continued development of new
technology resulting in improved diesel fuel economy.
Caterpillar also suggested that a 1-3% loss of fuel economy
would occur on turbocharged engines due to the crankcase emission
standards. Caterpillar apparently only evaluated one of several
alternatives for controlling crankcase emissions, that of ducting
the crankcase blowby int:o the turbocharger inlet. There are at
least three other options discussed in the Summary and Analysis of
Comments on Crankcase Emissions that Caterpillar apparently did not
consider. At least one of the other options, and possibly more,
need not cause a fuel economy penalty. However, considering the
affect of ducting blowby into the turbocharger inlet is a realistic
issue, the Summary and Analysis of Comments on Crankcase Emissions
has recommended that crankcase emissions be controlled only from
-------
non-turbocharged engines. The analysis does recommend, however,
that the issue of crankcase emission control on turbocharged
engines remain in the proposal stage until more data is gathered.
Therefore, the issue of potential fuel economy loss due to crank-
case emission controls on turbocharged diesel engines is deferred.
Mack claimed less than 1% fuel economy loss between a 10% AQL
and a 40% AQL. Mack based their analysis on a potential change in
NOx that might be required to meet the 10% AQL. The analysis used
data from the current steady-state 13-mode test.
A review of data from an ongoing diesel transient test pro-
gram?/ suggests that the potential change in NOx levels that Mack
claims will be required for the 10% AQL probably won't occur on a
fleet-wide basis. Chapter 7 of the Regulatory Analysis discusses
the impact of a 10% AQL as well as durability and other related
issues in terms of low mileage targets (LMT). Using the derivation
of LMTs from Chapter 7, an LMT for diesel engines with the proposed
10.7 g/BHP-hr NOx standard would be around 8 g/BHP-hr. A review of
the data from engines tested to date_7_/ indicate that most engines
are already well below the LMT NOx level. Therefore, most engines
would not experience a fuel economy penalty at a 10% AQL.
Another important point to note about the Mack comment is that
the only alternative apparently investigated by Mack was to lower
the low mileage target (i.e., low initial emissions). The option
of improving quality control of the product (engines) apparently
was not considered. If Mack were to improve the product quality it
is difficult to understand how a fuel economy penalty could be
incurred.
GM and Cummins both commented on a fuel economy loss due to
NOx control. But, they only discussed the loss in the context of
the steady-state California standards of 6 g/BHP-hr HC + NOx. As
stated previously most diesel engines appear to be well below the
proposed transient 10.7 g/BHP-hr NOx standard as well as the
estimated LMT NOx levels. Therefore, no fuel economy penalty is
expected to occur with proposed 1984 standards.
4. Recommendations
We conclude from the foregoing analysis that heavy-duty
gasoline-fueled and diesel engines will not incur a fuel economy
penalty if the proposed HC, CO, and NOx standards are promulgated.
Based on this analysis it is reasonable to expect that gasoline
engines with catalyst technology will obtain a fuel economy im-
provement. In the long run (2 to 5 years) the fuel economy im-
provement from gasoline fueled engines could be as high as 13%.
However, considering that initially the manufacturers will be on a
learning curve and must consider such issues such as low mileag'e
targets, audit quality limits AQL, etc., the potential fuel economy
27?
-------
fuel economy improvement could be reduced to 4% to 9% for the first
few model years after the proposed standards take effect.
For diesel engines a fuel economy improvement is also possible
due to the relaxed NOx standard. Even though diesel engines are relatively
closer to the proposed standards than current gasoline engines, they
will also be on a learning curve the first couple of years after imple-
mentation. Therefore, it is expected that fuel economy from diesel s
will remain stable.
Our recommendation is that the proposed standards be promulgated.
For the purposes of determining the economic impact of the proposed
regulations, we recomrend the following fuel economy impacts.
HDG 4% Fuel Economy Gain
HDD 0% Fuel Economy Gain
-------
References
_!/ "Light-Duty Automotive FueL Economy...Trends through 1979",
SAE Paper Number 790225, J. D. Murrell, U.S. Environmental
Protection Agency, February, 1979.
21 "Intake Port Deposits Improved Fuel Economy," Automotive
Engineering, October, 1979, pages 43-47, a magazine conden-
sation of SAE 790938, L. B. Graiff, Shell Development Co.,
October, 1979.
_3/ Data from EPA contracted test program on 1969 Light-Duty
Trucks, Contractor, EG&G Automotive Research Inc., San Anton-
io, Texas, testing completed in the fall of 1979.
_4/ Data from EPA contracted test program on 1972 and 1973 Light-
Duty Trucks, Contractor, EG&G Automotive Research Inc., San
Antonio, Texas, testing completed in the fall of 1979.
5j "1969 Heavy-Duty Engine Baseline Program and 1983 Emission
Standards Development", EPA/OMSAPC Technical Report, T. Cox,
G. Passavant, L. Ragsdale, May, 1979.
_6_/ Data from In-house EPA HD Transient Test Program on twelve
1979 Model Year Certified-Configuration Engines, Ann Arbor,
Michigan, testing completed in November, 1979.
II Data presented in Summary and Analysis of Technological
Feasibility.
8/ "Automobile Exhaust Emission Surveillance - Analysis of the FY
~~ 1975 Program," EPA Report No. EPA-460/3-77-022, NTIS No.
PB 279 355, December, 1977.
-------
PART II
Analysis of Minor Issues
-------
Issue - Test Procedure: Part II
1. Summary of the Issue
EPA has proposed a new transient test procedure. Aside from
the major issues concerning the Test Procedure as a whole, numerous
minor issues were raised pertaining to various technical points and
details of the procedure.
In short, are these technical details properly specified to
insure accurate test results, are they unduly restrictive, and can
the test procedure be made more flexible?
Subissue - Exhaust Sampling and Analytical Systems
Summary of the Comments
Many commenters argued that other sampling systems be per-
mitted besides those proposed in the NPRM. Cummins advocated usage
of continuously integrated dilute sampling system, pointing out
discrepancies between bagged and continuous NOx measurements.
General Motors argued that use of computers for sample recording
should be permitted. Numerous comments suggested changes in
analyzer specifications, or clarifications in optimization pro-
cedures and equipment. Clarifications were requested on required
equipment and components on the sampling system. In short, com-
ments pertaining to the following areas were received:
- non-methane vs. a total hydrocarbon standard;
- use of CVS vs. alternate procedures;
- accuracy of CVS parameter measuring equipment;
- CVS calibration equipment: type and accuracy;
- CVS equipment;
- CVS backpressure;
- HFID optimization and calibration procedures;
- HFID calibration gases;
- HFID fuel impurities;
- HFID response time;
- Calibration gas accuracies;
- NOx measurement systems;
- Heated analyzers for bag analyzer;
- Computer-assisted sample recording
Analysis of the Comments
For the most part, all comments listed above were incorporated
into Subpart N of the regulations.
One of the more significant issues raised pertaining to
exhaust gas analysis involved Fords' assertion that only non-
methane hydrocarbons should be regulated. Ford stated that -by
50 Z.
-------
calculating the hydrocarbon standard based on a measurement of
total hydrocarbons, including methane, EPA violated section 202(a)
of the Act because methane does not endanger the public health and
welfare. Ford made a similar challenge to the HC standard for
light-duty vehicles which is also based on a measurement of total
hydrocarbons. The U.S. Court of Appeals for the District of
Columbia Circut unequivocably rejected Ford's assertion and
concluded that the Clean Air Act authorized regulation of all
hydrocarbon emissions. (See Ford Motor Company vs. EPA, No.
78-2041 D.C. Cir. 1979) (The Court of Appeals had not reached its
decision when Ford submitted its comments on this matter). If
Congress authorized the regulation of all hydrocarbon emissions
from light-duty vehicles, it reasonably follows that regulation of
all hydrocarbon emissions from heavy-duty vehicles is not prohi-
bited by the Act.
Even if the standard were recalculated to measure only non-
methane hydrocarbons, the level of non-methane hydrocarbons control
would remain the same because the stringency of the standard would
be adjusted accordingly (Section 202(a)(3)(A) sets an HC standard
of at least 90 percent reductions; Congress contemplated a more
stringent standard if it was technologically feasible).
Recommendations
Retain the total hydrocarbon standard.
Allow the use of alternatives to the CVS concept if equivalent
results are obtainable. Specify how equivalency is to be demon-
strated.
Allow the use of metering Venturis, large radius nozzles,and
ASME flaw nozzles for CVS calibration, if traceable to NBS stan-
dards .
Add specifications for either a mixing box or dilution tunnel
for diesel sampling. (These specifications are consistent with
requirements for particulate sampling procedures).
Add decaileci specifications for the diesel HFID (e.g. probe
location in tunnel, response time, "overflow" calibration proce-
dure, etc.) for clarification and accuracy.
Add specifications for the continuous sampling of CO, CCL,
and NOx for diesel engines. Allow the option of continuous
sampling for gasoline engines, provided results are equivalent.
Subissue - Engine Cool-down Procedures
As discussed in Section A of the Summary and Analysis of.
Comments, the industry argued for adoption of a forced cool-down
303
-------
procedure in lieu of the 12-hour cold soak. The industry also
argued that the 36-hour cold soak limit was overly restrictive.
Analysis of the Comments
An EPA test program in conjunction with data submitted by
Ford and Cummins has established the viability of such a proce-
dure. The procedure as written in Subpart N of the regulations is
flexible, constraining only the coolants, coolant temperatures, and
methods of coolant application. A cold start emission test may
begin when the oil temperature reaches a designated range. (For
catalyst-equipped engines, a temperature range within which the
catalyst must be prior to a cold start is also specified).
The oil temperature range specification is 68°F to 75°F, as
opposed to an ambient range of 68°F to 85°F. This discrepancy
arises from EPA's forced cool down test program in which statis-
tically significant differences arose between engines tested after
a natural 12-hour cool down and those tested after a forced cool
down to oil temperatures exceeding 75"F. (A description of the EPA
test program can be found in a separate EPA technical report).
Recommendations
Adopt the forced cool down procedure as detailed in §86.1335
of Subpart N.
Drop the 36-hour maximum time limit on natural cold soak; drop
the 72 hour maximum time limit on the entire test sequence. There
is no apparent need for these requirements.
Subissue - Engine Mapping Procedures
Summary of the Comments
Diesel manufacturers commented that the proposed mapping
technique was inaccurate at lower speeds. Mack questioned the
definition of "measured rated speed", in particular for engines
which develop usable horsepower at speeds well above the speed at
which maximum horsepower occurs.
Gasoline manufacturers questioned the safety of engine opera-
tion at low speeds and wide open throttle. Use of a cubic spline
technique for maximum torque curve generation was also questioned;
linear interpolation was suggested as an alternative.
Analysis of the Comments
Transient diesel testing at SwRI has substantiated the inac-
curacy of the proposed diesel mapping technique. A transient-
mapping technique, a slow progression from minimum to maximum
3o V
-------
speeds at wide open throttle, has since been used with satis-
factory results. (Both the proposed technique and the transient
technique were originally recommended by an SAE subcommittee to
insure safety at lower engine speeds).
Macks' comments pertaining to "measured rated speed" are well
founded. The definition has been changed to allow emission testing
at speeds up to a "rated" speed specified by the manufacturer.
(See definition in Recommendations). This would assure emission
testing across the true full range of engine speeds, but at the
manufacturers option. As in the proposal, the speed of maximum
horsepower remains the lower limit of 100 percent speed.
At no time during the transient baseline work at EPA or SwRI
has trouble been experienced with steady-state mapping of gasoline
engines. The manufacturer's concerns are unfounded. However, EPA
recognizes that special cases may exist and an optional transient
mapping procedure is allowed for gasoline engines.
Further recognizing the existence of special cases, both
diesel and gasoline manufacturers are allowed to specify alternate
mapping procedures if the techniques specified in the Final Rules
are judged unsafe or unrepresentative. Advance approval of the
Administrator is required before this option may be taken.
Finally, use of a linear interpolation between points is
required to generate the maximum torque curve if transient mapping
techniques are used. Requirements that data points be collected at
least once a second during transient maps are no more stringent
than those- required to run the transient test itself.
Recommendations
Specify a transient mapping procedure for diesel engines,
utilizing linear interpolation between the mapping points.
Allow an optional transient mapping procedure for gaso.line
engines, utilizing linear interpolation between the mapping
points.
Redefine "measured rated speed" as the highest engine speed
at which maximum horsepower occurs, or a, speed specified by the
manufacturer provided that the specified speed is at least 100 RPM
greater than the highest speed at which maximum horsepower occurs,
and provided that at least 50% of the maximum horsepower occurs at
the specified speed.
Allow an "escape clause" for special case engines, i.e. those
with which the Administrator agrees that a safe and/or representa-
tive rasp cannot be performed per the Final Rules procedures.
-------
Drop the requirement that only 8 hours be allowed for practice
cycle runs after mapping. Experience at the EPA Lab has shown
this to be overly restrictive for certain control systems. No time
limit shall be specified.
Details of these incorporated recommendations are found in
§86.1332 of Subpart N.
Subissue - Engine Starting Procedures
Summary of the Comments
Several manufacturers argued that starter motors should not be
required. The dynamometer can start the engine just as easily and
the impact on emissions is negligible.
Cummins suggested that engine stalls during the hot start
portion of the test cycle not be permitted to void the entire
test; another hot soak and cycle should be allowed.
Caterpillar claimed that a limit of 15 seconds of cranking was
less than that recommended to their customers and argued for
longer permissible cranking times.
Analysis of the Comments
Starter motors have been used at EPA and SwRI from the first
transient test until the present. No problems have been encoun-
tered and problems with additional equipment (batteries, battery
chargers) were insignificant. Although emission impacts observed
to date have been minimal, the Staff is concerned with the compro-
mise in representativeness entailed in use of the dynamometer for
starting, especially at lower emission levels and for harder-
starting engines. Safety concerns of open electrical contacts in a
fuel-filled environment are recognized; proper use of judgement and
common sense in the location of batteries and switches, however,
more than alleviate these concerns.
Engine restarts during stalls or other voiding incidents in
the hot cycle portion of the test have been performed at EPA's lab
for some time. Emissions can be affected, however, if the engine
is shut down during the high power, high temperature LA Freeway
portion of the cycle. The engine would then be restarted at a high
temperature than normal. A single hot cycle restart should be
permitted, however, if the engine is shut down before the LA
Freeway begins (less than 580 seconds into the hot cycle).
Caterpillar's concern for greater cranking time is understood
and incorporated into the procedure.
Recommendations
Retain the starter motor requirement.
-------
Allow a hot cycle resoak and restart, provided that the
engine is shut down before 580 seconds into the hot cycle.
Allow a greater cranking period, if a manufacturers' re-
commendation to the customer is consistant with a longer cranking
time.
Subissue - Engine Testing Procedure
Summary of the Comments
Many commenters requested clarification of general test
requirements, i.e. temperature, humidity, and barometric pressure.
The statisitical validation criteria were also questioned.
Questions were raised as to how the cycle was to be run with
engines equipped with automatic chokes, or with engines designed
for use with automatic transmissions.
Analysis of the Comments
Test ambient conditions were clarified. Humidity specifi-
cations were relaxed, forbidding testing (i.e. requiring humidity
control) only when engine intake air humidity exceeds 90 grains per
pound of dry air. No humidity restrictions are placed on test cell
ambient air or CVS dilution air. Engine intake air and CVS dilu-
tion air are required to be temperature controlled (25°C + 5°C),
but test cell ambient air is not - provided that the emission
control apparatus on the engine are not temperature-effected. (This
is intended to preclude the need for high volume air conditioning
systems in test cells). Barometric pressure requirements remained
the same, but were clarified.
Inlet and exhaust restictions representing "normal" in-
vehicle conditions for diesel engines were specified.
Allowances were made for testing of automatic choke gasoline
engines, and for all engines used with automatic transmissions.
These changes were included in the cycle themselves and/or in the
validation criteria.
The statistical validation criteria were relaxed, per tran-
sient testing experience. An additional relaxation (over and above
those outlined in the 1969 Baseline Report) was made: the standard
error for Brake Horsepower was relaxed from 7 percent to 8 percent
of maximum. This is based upon transient diesel testing experience
acquired at SwRI. The statistical criteria was also changed co
reflect operation of automatic choke and/or automatic transmis-
sions .
307
-------
Recommendations
Relax the general humidity, temperature, and pressure require-
ments, per §86.1330 of Subpart N.
Relax the statistical validation criteria per §86.1341 of
Subpart N.
Allow for the use of automatic choke and automatic transmis-
sion, per §86.1333 and §86.1341 of Subpart N.
Subissue - Miscellaneous
Summary of the Comments
Various typographical errors and two mathematical errors were
pointed out in the proposal.
Analysis of the Comments
The assertions of error were checked and changed if true.
Recommendations
Change the errors in the procedure for Final Rule.
-------
B. Issue - Redefinition of "Useful Life"
1. Subissue - Transfer of "useful life" into market.
a.
Summary of the Comments
Cummins said that "EPA, by allowing each manufacturer to
provide its own yardstick to do its own measurement with no appro-
priate standard, will be effectively transferring the entire useful
life mileage question into the marketing arenas."
b. Analysis of the Comments
The associating of actual rebuild criteria with the definition
of useful life addresses Cummins' problem by placing most of the
emphasis on technical rather than market criteria. While we
obviously intended that the proposed manufacturer-determined useful
life alone would be based on technical aspects of the engine, we
agree that without the rebuild criteria some manufacturers may have
emphasized marketing criteria.
c. Staff Recommendations
We recommend no further changes in response to this comment.
2. Subissue - Encouragement for short useful lives.
a. Summary of the Comments
Cummins believes that to keep costs down, some manufacturers
will sell short-lived engines. Especially when these engines are
used past their useful lives, air-quality will suffer.
b. Analysis of the Comments
Currently the only incentive which EPA provides for longevity
is a useful life roughly one-half of the average lifetime of the
engines. It is toward the goal of durable emission controls that
the extension of the useful life (and the defining of minimum
maintenance intervals) is introduced. The "ambiguity" of the
useful life will be reduced by the establishment of rebuild cri-
teria as an end to the useful life. We expect that this package
will result in more, not less, incentive for manufacturers to build
in long-lasting emission control into their engines. The visibil-
ity of the "useful life" on the label should also provide an
advantage to the makers of durable engines.
c. Staff Recommendations
We recommend no further changes in the regulations.
30?
-------
3. Subissue - Congressional Authority
a. Summary of the Comments
Cummins points to three issues, as follows:
(1) Congress specifies that the Administrator must
determine the useful life, not the manufacturer.
(2) Congress specifies that the useful life must be "a
period" of use, whereas the proposal would result in widely varying
useful lives.
(3) If the Administrator were to disapprove a manufac-
turers' useful life, the timing would be such as to wreak economic
hardship by delaying certification.
b. Analysis of the Comments
(1) Our view is that EPA is indeed defining the useful
life as specified by the Act. The manufacturers determine for
their engines what the values of the useful lives are (or the
values of the rebuild criteria), but "useful life" is defined by
EPA in the regulations.
(2) Similarily, for any given engine family or indivi-
dual engine, a "period of use" exists and is called the "useful
life". Congress did not constrain EPA to one number for all
heavy-duty engines. Currently, in fact, two periods of use exist,
one for each fuel type.
(3) Section 86.083-22 (d)(2) of the proposal clearly
says that the Administrator does not approve a manufacturer's
useful life. The problems that concern Cummins cannot occur.
c. Staff Recommendations
We recommend that no further changes be made regarding the
useful life issue in response to these comments.
3)0
-------
C. Issue - In-Use Durability Testing
There are no subissues needing consideration in this Part
All pertinent comments have been addressed in Part I.
3!)
-------
D. Issue - Allowable Maintenance
1. Subissue - Certification at shorter intervals.
a. Summary of the Comments
IHC was concerned that the proposed maintenance provisions did
not permit more frequent maintenance during certification than the
EPA established minimum intervals. Owners would get a free re-
placement under warranty, though the cost of the engines would have
to go up to cover the replacement parts.
b. Analysis of the Comments
The primary intent of the new maintenance requirements is to
encourage low-maintenance emission control systems. To allow more
frequent maintenance than what is technologically necessary would
be contradictory to this purpose. In any event, it is usually more
costly for the consumer to pay for the replacement of a component
at the time of purchase than to pay for the designing-in of greater
durabilitry.
c. Staff Recommendations
We recommend that no changes be made in response to this
comment.
2. Subissue - EPA's language in certain passages of the
proposed regulations is vague or overly restrictive.
a. Summary of the Comments
Ford commented that defining "emission-related maintenance" as
that having a "substantial effect" on emissions was not adequately
specific for regulatory purposes.
In addition, Ford believes the EPA definition of "new tech-
nology" is too restrictive as it is used in §86.083-25(c)(1)Civ) of
the NPRM as a criterion for EPA to authorize more frequent mainte-
nance than that which is technologically necessary. Similarly,
Mack said that the proposed definition and application of "un-
scheduled maintenance" should be no more restrictive than their
current recommendation, which calls for maintenance when the
operator observes malfunction symptoms according to their trouble-
shooting guide.
b. Analysis of the Comments
We believe that the term "substantial effect on emissions",
while open to subjective judgement in an individual case, clearly
signals that a rather large effect must be present. If perfor-
3)2.
-------
mance of a certain maintenance item resulted in a small change in
emission, near in magnitude to the limit of accuracy of the analy-
sis instruments, EPA could hardly claim that the effect was "sub-
stantial". In any case 13 examples of non-emission-related main-
tenance are given in the regulations (§86.083-25(c)(1)(vi)). If
a manufacturer were honestly confused as to how EPA would catego-
rize a certain maintenance item, it is not uncommon to seek in-
formal advice early in the certification process as to what EPA's
interpretation might be.
Ford's concern about restricting "new technology" to that "not
found in production on any motor vehicle prior to the 1980 model
year" seems unnecessary. Certainly, EPA desires to allow more
frequent maintenance only in extreme cases, since we believe we
have adequately analyzed the maintenance requirements of current
technology components. However, for a new component, design,
system which has different maintenance requirements is introduced
it should not be difficult to show that this "new technology"
deserves special consideration. The specific terminology is
necessary to prevent, for example, a new arrangement of current
components from being introduced simply for the sake of a reduction
in maintenance requirements.
Finally, we believe that the proposed constraints on unsche-
duled maintenance are not overly restrictive and serve to assure
that maintenance is not casually or repeatedly performed if it is
noc scheduled.
c. Staff Recommendations
Based on the consideration of the comments, we recommend that
the language cf the instant provisions remain unchanged.
3. Subissue - Suggested Maintenance Criteria.
a. Summary of the Comments
Ford suggested the following as the only restrictions on
maintenance during certification:
(I) Maintenance (and the interval at which it is performed)
must be necessary for Che proper functioning of the emission
control system and the vehicle.
(2) Maintenance procedures must be performable by mechanics
in the field.
b. Analysis of the Comments
EPA's reasoning behind the specifying of minimum intervals is
313
-------
discussed in the major portion of this document. However, it may
be useful to explain why we believe the maintenance requirements,
while more restrictive, do fulfill Ford's criteria. First,
changes which will appear in the final rule will clearly relate the
maintenance done during certification to that required in use,
inherently requiring that maintenance be easily performed by
mechanics. Second, by requiring that maintenance be technologi-
cally necessary, we are in effect requiring no less maintenance
than has been shown to be adequate for the proper functioning of
current systems and vehicles.
c. Staff Recommendations
We recommend no further changes in the maintenance require-
ments in response to this comment.
4. Subissue - Combustion chamber opening during certifica-
tion.
a. Summary of the Comment
GM commented that requiring that the combustion chamber not be
opened during servicing is inconsistent with normal service prac-
tices .
b. Analysis of the Comment
We have not proposed to change the unscheduled maintenance
provisions in the current regulations. They require that addi-
tional unscheduled maintenance (other than for misfire, choke
maladjustment, or incorrect idle speed) may not, among other
things, require direct access to the combustion chamber. We do
not believe that removal of the cylinder heads should be a routine
practice during certification and support the current provisions.
c. Staff Recommendations
We recommend that the provisions relating to the opening of
the combustion chamber be retained.
-------
E. Issue - Parameter Adjustment
1. Subissue - Validity of the EPA Restorative Maintenance
Study.
a. Summary of the Comments
Some commenters challenged the use of the EPA Restorative
Maintenance Study (an Evaluation of Restorative Maintenance
on Exhaust Emissions of 1975-1976 Model Year In-Use Automobiles,
EPA 460/3-77-021, December 1977) as an basis for these regula-
tions. They claimed that the Restorative Maintenance Study does
not establish any correlation between in-use maladjustment and
failure to meet the emission standards.
b. Analysis of the Comments
The position advanced by these comments is in error. Similar
comments were received when parameter adjustment was proposed for
light-duty vehicles. The implications of the Restorative Main-
tenance study are discussed in the Summary and Analysis of Comments
which accompanied the final light-duty vehicle rule-making, at
pages 7-12. The Restorative Maintenance study clearly identified
parameter maladjustment as a major cause of excess in-use emis-
sions .
In issue E of Part I, a summary of the results of the Restor-
ative Maintenance program is provided (Table E-l) . That table
indicates clearly the impact of parameter maladjustment on emis-
sions .
Staff Recommendations
None.
2. Subissue - The language defining what parameters must be
identified is vague and ambiguous.
Summary of the Comments
This comment pertains to Section 86.083-21(b)(1)(ii), which
defines those parameters which must be identified as including
those which "may affect emissions." This definition was felt to
be too vague, and possibly subjecting manufacturers to undue
hardship and unexpected delays.
b. Analysis of the Comments
The staff recognizes the validity of this comment. The
wording of the section in question can be revised similar to that
wording proposed by Ford in its comments.
-------
Staff Recommendations
Revise the wording of Section 86.083-21 (b)(l)(ii) to clarify
the parameter to be reported.
3. Subissue - Timely determinations by the Administrator
a. Summary of the Comments
Commenters believed that the regulations should contain
provisions such that early submittals result in an early response
from EPA. EPA stated in the preamble that to avoid disruption of
certification, it intended to make early determinations of those
parameters subject to adjustment, the adequacy of limits, stops,
seals, etc. if the manufacturer submits the necessary information
early. However, the regulations do not require such early deter-
minations.
b. Analysis of the Comments
The staff recognizes the validity of this problem. Manufac-
turers should be given a definable timetable within which to work
in developing complying engines. In order for this to happen, the
manufacturer should have a specified maximum waiting period for
EPA determination.
c. Staff Recommendations
The regulations should be modified to provide a maximum time
interval of 90 days for EPA to review submitted information and
make its determinations. This time would be exclusive of time
involved in obtaining additional information needed to make the
determinat ions.
4. Subissue - The definition of what constitutes a "new
parameter".
a. Summary of the Comments
Commenters objected to defining a new parameter as one which
"was not present on vehicles of the same engine family in the
previous model year." The objectionable part of this definition
was the " same engine family" criteria, which commenters felt did
not accomplish the objective of the definition. This objective
they felt to be the distinguishing of a new parameter (which should
be introduced in a conforming condition) from existing parameters
(which need adequate lead-time to develop tamper-resistant de-
signs). A change in engine family, commenters felt, had no direct
bearing on whether a parameter was in fact a new one.
-------
F. Issue - Idle Test and Standards
There are no subissues needing consideration in this part
All pertinent comments have been addressed in Part I.
-------
b. Analysis of the Comments
The staff recognizes the validity of this objection. The
wording of the regulations can be revised to accomodate this
problem.
c. Staff Recommendation
The wording of Subsection 86,083-22(e)(1) (i) should be changed
to eliminate the "same engine family" criteria for identification
of a new parameter.
5. Subissue - Manufacturers risk relative to in-use mal-
adjustments .
a. Summary of the Comments
Comments on this issue contended that a determination by EPA
of the adequacy of limits, stops, seals, or other means to inhibit
adjustment of parameters constitutes a finding by EPA that the
manufacturer has adequately fulfilled his responsibility to
restrict adjustability. Consequently, any actual maladjustment
occurring in-use should be considered tampering instead of malad-
justment, and the manufacturer should bear no responsibility for
such occurrence (such as recall liability).
b. Analysis of Comments
It is the manufacturer's responsibility to do a thorough job
in designing non-adjustable parameters. EPA determinations during
certification are based upon information available prior to accumu-
lation of actual in-use experience. Notwithstanding the fact
that EPA has issued a certificate of conformity to produce an
engine family, the manufacturer remains liable for the performance
of his engines in-use. The final test of the adequacy of any means
to inhibit adjustability is the performance of that means in-use.
Thus, it may happen that although a vehicle is certified, it is
later recalled, because EPA may inaccurately predict in-use malad-
justments. EPA's position, dating from prior to even the light-
duty parameter adjustment regulations, is that the fact that an
in-use vehicle is adjusted to non-recommended settings does not
necessarily prove that it has not been properly maintained and
used. EPA has, therefore, asserted that maladjusted in-use ve-
hicles may be subject to recall action.
c. Staff Recommendations
None.
3/7
-------
G. Issue - Leadtime
1. Subissue - Statutory Timetables
a. Summary of the Comments
Many commenters indicated that the 1977 Clean Air Act Amend-
ments required EPA to promulgate final 1983 standards by December
31, 1978 and allow a four year leadtime for implementation by the
industry. Because EPA has failed to meet the December 31, 1978
deadline, commenters felt that the complete rulemaking could not be
imposed for 1983. Several alternatives were suggested as to how
EPA ought to deal with this situation. These included allowing
more time to implement the regulations, deletion of the transient
test in favor of the current steady-state procedure, or implemen-
tation of a modified transient test (the "Caterpillar" cycle).
b. Analysis of the Comments
EPA acknowledges that the timetable foreseen by Congress was
as outlined above. What neither the Congress nor EPA anticipated
at the time the 1977 Amendments were enacted was the amount of
additional time necessary to develop a test procedure designed to
insure that the significant reduction in HC and CO emissions that
Congress sought would actually be achieved. The Agency is approxi-
mately one year behind the schedule for implementation that Cong-
ress had expected. The decision to delay the year of implementa-
tion one additional year until 1984 maintains the four year lead-
time inherent in the Clean Air Act timetable. That decision, made
from considerations of feasibility alone, satisfies the coramenters
demands for a four year leadtime.
c. Staff Recommendations
In light of the decision to delay implementation until 1984,
no specific action is needed in response to this subissue.
2. Subissue - Interpretations of legal issues have been made
by technical and administrative personnel rather than by the EPA
Office of General Council.
a. Summary of the Comments
In its comments, General Motors indicated that in their review
of EPA documents obtained under a Freedom of Information Act
request, they found that no formal legal opinion had ever been
requested from the Office of General Council on any of the key
legal issues involved in the rulemaking. Rather, GM believed
that legal interpretations were being made by technical and
-------
administrative personnel. General Motors indicated the possible
operation of a "housekeeping perspective" leading to an unwarranted
desire to prematurely promulgate the transient test.
b. Analysis of the Comments
The idea that development of this rulemaking proceeded without
input from the Office of General Counsel is incorrect. Informal
contacts with representatives of that office were made as needed
when matters of legal interpretations arose. Representatives of
the Office of General Counsel also participated fully in the
internal review process involved in promulgating the initial
proposal (as well as these final regulations). The Office of
General Counsel is also one of the key offices which must formally
concur with a rulemaking before it is proposed or finalized. Such
concurrence is only made after a careful review of the action
involved.
c. Staff Recommendations
None.
3. Subissue - EPA determination of Technological Feasibility
a. Summary of the Comments
In its comments, GM contended that EPA had made a determina-
tion of technological feasibility for the proposed action at a time
when adequate transient testing had not yet been done and before
promulgating the final test procedure. This was believed by GM to
have been an arbitrary and capricious action denying manufacturers
knowledge of the standard and final regulations, opportunity to
request a revised standard, and four year leadtime.
b. Analysis of the Comments
EPA did not make any final determination of technological
feasibility until after a complete review of all comments submitted
during the comment period for this rulemakng. EPA stated in the
preamble to its- proposal (44 FR 9471, February 13, 1979) that
"manufacturers comments on the feasibility of meeting the proposed
standards will be considered in setting final standards. If
revisions to the statutory standards are warranted, they will be
made." Preliminary conclusions concerning feasibility had to be
made before the proposal could be published for comment. That the
engineering judgement behind those preliminary conclusions were
sound is born out by the finding in this final rulemaking that the
regulations are indeed feasible (see issue I - Technological
Feasibility). At no time prior to this did EPA make a final
determination cf feasibility under Section 202(a)(3)(C).
-------
c. Staff Recommendations
None.
4. Subissue - Committments of Resources by Manufacturers
Prior to Final Rulemaking.
a. Summary of the Comments
Several commenters believed that EPA was predicating the
feasibility of its proposed timetable for diesels upon the expecta-
tion that manufacturers could begin working toward compliance with
the regulations once the NPBM was published. These commenters
contended that manufacturers could not be required to make signi-
ficant resource committments prior to final rulemaking action by
EPA.
b. Analysis of the Comments
In the preamble to the NPRM, EPA did indicate its belief that
the diesel engine manufacturers could begin facility acquisition
before promulgation of final rules. At the time of publication
some companies had already taken such steps. Testimony supplied to
EPA during the comment period indicated that all domestic manufac-
turers had made advance committments to obtain some limited test
capability. However, EPA's belief in the feasibility of its
proposed 1983 compliance date was not predicated upon these commit-
tments as the commenters suggest. In the NPRM, EPA also stated
that "even assuming that the manufacturers do nothing until promul-
gation of final rules (assumed promulgation date is December 1979),
EPA concludes that the proposed emission levels are achievable with
already available emission control technology within the leadtime
existing." (44 FR 9471, February 13, 1979).
The leadtime analysis used in this final rulemaking makes use
of the information supplied to EPA concerning advance committments
of resources. These committments have been made voluntarily by
manufacturers to develop their own in-house testing capability.
Since such advance facilities are being procurred, it is appro-
priate to incorporate their availability into leadtime consider-
ations. EPA is not requiring that these actions be taken, but
simply recognizing that they have.
c. Staff Recommendations
None.
3ZI
-------
H. Issue - Economic Impact
1. Sabissue - Warranty Claims Associated with the Useful
Life Redefinition
a. Summary of the Comments
Based on the useful life redefinition which was proposed, "the
average period of use up to engine retirement or rebuild, whichever
comes first", most manufacturers stated that they would incur
increased warranty claims on as much as 50 percent of their sales.
The figures submitted by each manufacturer are given below:
Caterpillar $ 89 per engine
Chrysler $200 - $400
GM $100
IHC $150
Other manufacturers discussed increased warranty claims but did
not quantify the impact.
b. Analysis of the Comments
The revised useful life definition which is discussed in issue
B reads :
The useful life of a heavy-duty engine is reached when
one of two possibilities occurs:
a) the mechanical rebuild criteria are surpassed
b) the average period of use is reached.
However, in no case can this be less than 5 years or 50,000
miles whichever occurs first. With this definition, some of the
manufacturer jeopardy is removed. This is true primarily because
the useful life is much more specific than in the original propos-
al.
The manufacturers warranty liability should be limited inher-
ently in that they would be expected to build durable engine/con-
trol systems. In no case does EPA anticipate the manufacturers
paying for rebuilds on 50 percent of the engines it sells.
In conclusion, these regulations which are establishing a new
useful life definition are not warranty implementation regulations.
Increased warranty costs associated with a potentially longer
useful life should be addressed when the warranty regulations are
implemented.
c. Staff Recommendations
Increased warranty costs should be considered by the Office of
Enforcement when the warranty regulations are implemented. These
costs need not be included in this package.
322
-------
2. Subissue - Inspection/Maintenance (I/M) Costs
a. Summary of the Comments
Several cotmnenters stated that EPA had underestimated the
costs of implementing I/M for heavy-duty gasoline-fueled vehicles
in several areas: the I/M fee, vehicle downtime, and operator
time.
b. Analysis of the Comments
The actual I/M costs are dependent upon the means by which the
program is implemented. Two methods will be discussed, private
garages and public inspection points.
A fee of $5 per test seems reasonable in light of the current
range of fees being charged in light-duty I/M programs ($3-$12).
Average vehicle lifetime I/M costs should not exceed $40 per
vehicle. If the I/M program is run by private garages, this cost
could be less because the new vehicle dealer could conduct the
first I/M test as part of dealer preparation. This is often the
case with mandatory safety inspection programs.
EPA believes that an annual vehicle I/M check would be con-
ducted during a period of minimal vehicle usage, or concurrently
with other routine maintenance, so no vehicle operating time would
be lost. Therefore, the only additional cost might be associated
with the operators time spent during the I/M test. If the annual
I/M check is conducted by state owned facilities, then there would
be a cost tied up with the time spent driving the vehicle to the
inspection facility. In any case, this cost would cover only 60
percent of the heavy-duty gas fleet since only 60 percent are in
commercial applications. If the I/M checks are done by private
garages, then the checks could be done in association with routine
maintenance, and no other costs would be incurred. In situations
where a commercial vehicle would have to be driven to a state
operated I/M facility, the inspection would probably take about
one-half hour or $7 according to ATA estimates.
The regulatory analysis supporting the final rulemaking does
not include I/M as an absolutely essential part of the comprehen-
sive control strategy, so I/M costs need not be included as part of
the cost effectiveness calculations.
c. Staff Recommendations
None required.
3. Subissue - Replacement Catalyst
a. Summary of the Comments
32.3
-------
Several of the manufacturers of gasoline-fueled heavy-duty
engines stated that they might require replacement catalysts to
meet the new useful life definition. Base on the comments received
from GM, the cost to replace a catalyst would be about twice the
manufacturing cost.
b. Analysis of the Comments
The EPA technical staff is convinced that a full life catalyst
is technologically feasible for use on gasoline-fueled heavy-duty
engines. The basis for this decision is given in the allowable
maintenance issue. Therefore, the regulatory analysis need not
include the costs of a replacment catalyst.
c. Staff Recommendations
Do not include the cost of a replacement catalyst in the
economic impact since it will probably not be necessary.
4. Subissue - Modified or Additional Pumping Facilities for
Unleaded Fuel.
a. Summary of the Comments
The American Trucking Association (ATA) stated that an incre-
ased cost of this regulation would be the need for additional
unleaded gasoline pumping facilities in the public and private
sectors. Under their assumptions, they estimated a. lifetime per
vehicle cost of $52.
b. Analysis of the Comments
EPA concurs with the comments by ATA, and has included this
cost in the economic impact analysis. This cost has been included
by allowing an additional 0.5 cents in the leaded-unleaded price
differential. In other words, the expected differential was
increased from 2.5 cents to 3.0 cents per gallon to allow for the
amortization of these facilities. Over the 114,000 mile average
lifetime of each vehicle, this becomes about $57 per vehicle.
114,000 miles 1 gallon $.005 = $5? .,
lifetime X 9.9 miles X gallon * '
c. Staff Recommendations
The cost of unleaded fuel pumping facilities should be includ-
ed in the analysis.
5. Subissue - Incremental Cost/Benefit Analysis
-------
a. Summary of the Comments
Several commenters, especially the Counsel on Wage and Price
Stability, emphasized that EPA had not presented an adequate
incremental cost benefit analysis covering all of the options
considered in this rulemaking action.
b. Analysis of the Comments
The EPA technical staff recognizes the need for an in-depth
incremental cost benefit analysis. The Analysis which was prepared
for the final rulemaking action should be published in support of
the final rule.
c. Staff Recommendations
Publish the alternative actions considered and the incremental
cost/benefit analysis as part of the regulatory analysis supporting
the final rulemaking action.
6. Subissue - Increased Cost Associated with a Fuel Economy
Penalty
a. Summary of the Comments
Ford Motor Company stated that due to the cold start weighting
requirements a 10 to 15 percent fuel economy penalty should be
anticipated.
Cummins Engine Company estimated a 1-5 percent fuel economy
foregone penalty due to the durability testing program. Cummins
felt that durability testing would force them to delay the use of
new technology in the marketplace.
b. Analysis of the Comments
The EPA technical staff is in absolute disagreement with the
comments by Ford. Ford's "projected" fuel economy loss is based on
an invalid extrapolation of steady-state data in an attempt to
simulate the heavy-duty transient test.
Fords projected fuel economy loss of 10-15 percent is based on
an increase in brake specific fuel consumption (BSFC) on their
simulated transient test. They claim the fuel economy decreases
from .620 to .721 Ib/BEP-hr as the HC standard decreases.
The invalidity of Fords' approach is easily demonstrated by
data gene-rated during 1979 baseline testing at EPA. The same
engine type which Ford used in their simulated transient test
was tested on the transient test by EPA at the Ann Arbor Labor-
atory. Ford estimated a BSFC of .627 for their current tech-
-------
nology 6.1 L engine on the transient test and a BSFC increase of 16
percent to .721 with the new emission standards. When EPA tested a
similar current technology engine on the transient test, the BSFC
was .7795, a full 25 percent greater than Fords' estimated value
for engines designed to meet the same emission standard. Clearly,
Fords' simulated transient test is not capable of accurately
predicting fuel consumption on the EPA transient test.
EPA appreciates Fords' efforts to meaningfully comment on the
NPSM, but obviously cannot consider Fords fuel economy decrease
projections as valid.
Cummins Engine Company estimated a 1-5 percent fuel economy
penalty in Appendix I of their comments, but no supporting state-
ments or data were given. EPA expects that Cummins was trying to
draw a parallel to their statements concerning the 1980 California
emission standards for HC and HC + NOx. The EPA technical staff
does not foresee Cummins using variable injection timing to meet
the 1984 HC or NOx standards and does not believe Cummins has
presented a firm basis for any fuel economy penalty for their
engines.
c. Staff Recommendations
The final economic impact analysis should not include any
costs associated with a fuel economy penalty.
7. Subissue - Dislocation in the Gasoline Engine Market
a. Summary of the Comments
International Harvester Company (IHC) raised fears that the
proposed regulations would remove any marketing advantage currently
held by gasoline-fueled engines and replace the gradual trend
toward dieselization with a "stampede". Their major concerns are
in the first price increase differental and operating cost increase
differental.
b. Analysis of the Comments
The EPA technical staff is very concerned about the economic
and employment impact of the proposed regulations. IHC's basic
contention is that these regulations will rapidly decay the posi-
tion of the gasoline-fueled engine in the heavy-duty market.
One must consider several different facets of this concept
before drawing any conclusions.
The past history of the gasoline-fueled engine in the heavy-
duty market is shown below for the last twelve years:_!_/
326*
-------
Year Percent Gas Percent Diesel
1967 81 19
1968 79 21
1969 76 24
1970 75 25
1971 75 25
1972 76 24
1973 74 26
1974 72 28
1975 79 21
1976 76 24
1977 69 31
1978 68 32
The data clearly shows that prior to the oil embargo of 1973 and
1974 and the 1974 economic downturn the market split between
gasoline-fueled and diesel engines remained fairly constant at
about 3:1.
However with the rising fuel prices in the late 1970's the
market share for the gasoline-fueled engine began to decrease.
This can be seen especially in weight classes VII and VIII which
are the "heavy-heavies".21 From this data it is clear that natural
market pressures due to fuel economy concerns are forcing the shift
to dieselization. Most studies expect classes VII and VIII to be
almost 100 percent diesel by 1990._3/
What is the impact of these regulations on this sales mix
shift? The average first price increase expected by EPA, $394
gasoline-fueled and $195, diesel, will decrease the selling price
differential between the two engines by $200. In addition, un-
leaded fuel costs, less decreased exhaust system and spark plug
maintenance, will increase operating costs by $83, for a total
differential of about $300. However, EPA predicts a fuel economy
increase of at least 4 percent in gasoline-fueled engines which
should lead to decreased operating costs.kj
On this basis, the impact of these regulations on the selling
prices of heavy-duty engines will be examined.
Gasoline-Fueled Diesel
Engine
Selling Price: $3000
First Price Increase: 394
Operating Costs: 83
Fuel Economy Benefit: -788
Selling Price $3394
Operating Cost Change -705
-------
As can be seen in the example above, the average price differ-
ential of $3000-$4000 _5/ dollars is decreased only $199, about 5.7
percent. This first price increase should have only a minimal
impact on sales of heavy-duty gasoline powered vehicles (.91
percent to 1.8 percent decrease) ._4/
The anticipated net decrease in operating costs may actually
make the gasoline-powered vehicle more attractive than it pre-
sently is in some applications.
In conclusion, EPA does not believe that these regulations
will have any substantial impact on sales of heavy-duty gasoline-
fueled engines. Any orderly displacement in this market which is
being caused by fuel economy pressures will remain relatively
unaffected by these regulations. In addition, the full impact of
the mandated diesel particulate and heavy-duty NOx regulations may
ultimately have an influence on this market split which easily
outweighs the minor impact expected here.
c. Staff Recommendations
None required.
8. Subissue - Manufacturer and Dealer Profit
a. Summary of the Comments
The commenters did not concur with EPA's exclusion of manufac-
turer and dealer profit from the first price increase for gasoline-
fueled and diesel heavy-duty engines.
b. Analysis of the Comments
The EPA technical staff recognizes that prudent business
practice will force the manufacturers and dealers to seek at least
an average profit on the funds which they invest in emission
control technology.
Unfortunately, these additional profits substantially increase
the price for cleaner air. Some would argue that profits on
emission controls are really a transfer payment from one segment of
society to another and are not really a "cost" which should be
considered in the cost effectiveness calculations for the emission
control strategy under consideration. At the present time, EPA
will be conservative in their cost effectiveness calculations and
include profit at all levels in these figures.
Having determined that manufacturer and dealer profit should
be included in the economic impact analysis, the amount of this
profit must be determined. Based upon the vastly different nature
of the gasoline and diesel heavy-duty markets the EPA technical
32?
-------
staff believes that profit levels should be evaluated for each
segment of the market.
To determine these profits and, in addition, overhead figures,
EPA studied financial data for the major domestic manufacturers in
both segments of the heavy-duty market. For gasoine-fueled engines
EPA studied data for 1976, 1977, and 1978 from General Motors (GM),
Ford, Chrysler, and International Harvester (IH). For diesel
engine EPA studied data from GM, Cummins, Caterpillar (CAT), Mack
Trucks, and IH.
(i) Gasoline Engine Manufacturers
The data below represents the overhead and profit levels which
EPA estimated using financial data found in Moody "s Industrial
Manual.^/ The values below are the fraction of the costs of
sales which overhead and profit comprised for each of the manufac-
turers in 1976, 1977, and 1978. For example an overhead fraction
of .11 implies that overhead is equal to 11 percent of the cost of
selling the product. The same example can apply to the profit
figure.
1976 1977 1978
Ovhd Profit Ovhd Profit Ovhd Profit
GM .117 .145 .109 .141 .117 .129
Ford .113 .080 .099 .100 .103 .081
Chrysler .102 .049 .067 .023 .104 -0
IH .172 .075 .169 .074 .193 .056
In terms of overhead the values range from .067 to .193 with
an average cf .122. Profit values varied from less than zero to
.145 with an average of .079. The range en these values is too
large to be explained simply. All four of these manufacturers
produce heavy-duty trucks and light-duty trucks, but only three
produce light-duty vehicles. Two produce diesel engines, and one
produces other farm type equipment. The EPA technical staff does
not believe that the average profit figure cited above represents
the profits which the industry would seek on their investment. The
EPA technical staff believes that using the GM average figures for
the period (.114 overhead and .138 profit) would conservatively
estimate the highest expected overhead and profit figures. GM's
profit figures are the highest and their overhead figures the
second highest of the four corporations studied, A figure of .252
for manufacturer overhead and profit should be used.
(ii) Gasoline Vehicle Dealers
Dealers could be expected to seek a profit on their increased
3Z1
-------
investment when purchasing a new truck or bus for sale to the
ultimate vehicle owner. Dun's Review, !_/ a well known business
magazine, contained an article on dealer profit after taxes and
other expenses. This was estimated at 1.5 percent. To account for
taxes etc., EPA shall use a figure of 3 percent for dealer profit.
The use of emission control technology on gasoline-fueled
engines does not inherently cause an increase in dealer overhead.
No additional personnel or engine servicing is required. There
should not be any increase in dealer overhead.
(iii) Diesel Engine Manufacturers
The data below represents the overhead and profit levels EPA
estimated using financial data found in Moody's Industrial Manual.
The values below are the fraction of the cost of sales which
overhead and profit comprised for each of the manufacturers in
1976, 1977, and 1978.
1976 1977 1978
Ovhd Profit Ovhd Profit Ovhd Profit
GM .117 .145 .109 .141 .117 .129
IH .172 .075 .169 .074 .193 .056
Cat .121 .165 .113 .172 .109 .172
Cummins .308 .169 .335 .150 .336 .115
Mack .123 .055 .096 .067 .085 .099
The overhead values ranged from .085 to .336 with an average
of .167. Profit values ranged from .055 to .172 with an average of
.119. Although these numbers also have a large range, these is
more reason available to explain the range in the figures.
Of the five compaines listed, three make heavy-duty engines
and vehicles and two make only engines. GM produces light-duty
vehicles, light-duty trucks, and buses as well as a wide variety of
other motor vehicle related products. IH makes heavy-duty engines
(diesel and gasoline) as well as vehicles and other farm equipment.
Caterpillar (Cat) produces not only diesel heavy-duty engines for
over the road use, but produces a wide variety of off-road con-
struction equipment. Cummins produces only engines, and engine
related components for sale to other manufacturers. Mack is a pure
producer of diesel heavy-duty trucks, producing both diesel engines
and vehicles.
336
-------
The EPA technical staff believes that the average figures
cited above adequately represent the heavy-duty diesel industry and
suggests their use in the economic analysis.
(iv) Diesel Vehicle Dealers
The EPA technical staff realizes that diesel vehicle dealers
(retail franchises or manufacturer representatives) will try to
seek a profit on their slightly increased investment in a new
heavy-duty diesel engine. Although some new diesels are sold as
single units, as to independent owner-operators, a large percentage
of the sales are multiple units sales to large truck fleets or bus
companies. It is reasonable that any minor profit sought by the
vehicle dealer on the sale of a diesel engine with emission con-
trols would be lost in the final price negotiations on the purchase
of the vehicle. Heavy-duty vehicles with diesel engines often sell
for more than $50,000 a piece.
Best judgement dictates that due to the higher selling price
and the tendency toward multiple unit sales in heavy-duty diesel
vehicles, the minor profit on emission control technology would be
ulimately lost in the final price negotiations. In any case, the
first price increase estimates developed in the economic impact
analysis may have a larger estimate error than the profit sought.
c. Staff Recommendations
Based on the discussion above several recommendations are
presented.
For gasoline-fueled engines an overhead figure of .114 and a
profit figure of .138 should be used. A dealer profit of 3 percent
is also recommended. In total this becomes:
RPE » (Manufacturing Cost) (1 + .114 + .138) (1.03)
RPE = (MC) (1.29).
For diesel engines the nsnufacturer overhead and profit
figures recommended are .167 and ,119 respectively. Best judgement
dictates no further increase in dealer overhead or profit. So in
total, RPE = (MC) (1 + .167 + .119) = MC (1.29).
In conclusion, it should be noced that the manufacturing costs
cited above and used in the regulatory analysis contain 20 percent
overhead and 20 percent profit at that level in the production of
the hardware. This is taken from the Rath and Strong report as a
gross estimate._8_/ These 20 percent figures are realistic in the
cases where the parts are supplied to the engine manufacturers by
independent vendors, but are conservatively high for the parts
produced within the engine corporation.
-------
References
JY Based on MVMA data.
2J See Regulatory Analysis, Chapter III.
_3_/ See for example "Will Diesels Dominate", Neil M. Szigethy,
Fleet Specialist Magazine, May-June 1979.
4/ See Regulatory Analysis, Chapter V.
_5_/ Based on IHC written comment, June 14, 1979, Appendix A,
page 9.
_6/ "Moody's Industrial Manual", 1979, Volume I.
]_/ "Dun's Review", November 1978, Vol. 112, No.5 pp 119-121.
8J "Cost Estimates for Emission Control Related Components/
Systems and Cost Methodology Description, Leroy H. Lindgren,
Rath & Strong, Inc. March 1978, EPA -460/3-78-002.
332
-------
I. Issue - Technological Feasibility
There are no subissues needing consideration in this part
All pertinent comments have been addressed in Part I.
333
-------
J. Issue - Selective Enforcement Auditing
Subissues on this topic are addressed in a separate submission
to the docket prepared by the Office of Enforcement.
33V
-------
K. Issue - Nonconformance Penalty
There are no subissues needing consideration in this part,
All pertinent comments have been addressed in Part I.
335
-------
L. Issue - Diesel Crankcase Emissions Control
There are no subissues needing consideration in this part
All pertinent comments have been addressed in Part I.
-------
M. Issue - Numerical Standards/Standards Derivation
There are no subissues needing consideration in this part.
All pertinent comments have been addressed in Part I.
337
-------
N. Issue - Fuel Economy
There are no subissues needing consideration in this part
All pertinent comments have been addressed in Part I.
------- |