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.

-------