EPA-AA-IMS/80-8
      Derivation of 1981 and Later Light Duty Vehicle Emission Factors for
                       Low Altitude, Non-California Areas
                                 November, 1980
                                     NOTICE
This report  does not  necessarily  represent the  final EPA decisions  or  posi-
tions.   It is intended to present a technical analysis of  the  issue  using data
which are currently  available.   The purpose in the  release of such  reports is
to facilitate  the exchange  of technical information  and to inform  the  public
of technical  developments  which may  form  the  basis for a  final  EPA decision,
position or regulatory action.
                        Inspection and Maintenance Staff
                      Emission Control Technology Division
                 Office of Mobile Source Air Pollution Control
                       Office of Air, Noise and Radiation
                      U.S. Environmental Protection Agency

-------
                              Table of Contents

                                                                     Page

I.     Introduction                                                   1

       A.  Purpose and Coverage                                       1
       B.  Background on the Post-1980 Fleet                          1
       C.  Composition of the Fleet:  Technology Types                2
       D.  Summary of Results                                         4


II.    Description of the Data Base                                   6

III.   General Methodology                                           10

       A.  Introduction                                              10
       B.  HC/CO Methodology                                         10
       C.  NOx Methodology                                           11
       D.  Comparison to the Methodology Used in the
           Previous Analysis                                         12


IV.    Specific Unit Analyses - HC/CO                                13

       A.  Closed Loop vehicles designed to meet
           0.41 g/mi HC and 3.4 g/mi CO standards                    13
       B.  Closed Loop vehicles designed to meet
           0.41 g/mi HC and 7.0 g/mi CO standards                    25
       C.  Oxidation Catalyst vehicles designed to meet
           0.41 g/mi HC and 3,4 g/mi CO standards                    29
       D.  Oxidation Catalyst vehicles designed to meet
           0.41 g/mi HC ammd 7.0 g/mi CO standards                   33


V.     Specific Unit Analyses - NOx                                  35

       A.  Closed Loop vehicles designed to meet
           a 1.0 g/mi NOx standard                                   35
       B.  Oxidation Catalyst svehicles designed to meet
           a 1.0 g/mi NOx standard                                   45


VI.    Composite Emission Factors for the 1981 Federal Fleet         47


VII.   Composite Emission Factors for the 1982 Federal Fleet         53
VIII.  Composite Emission Factors for the 1983 and Beyond
       Federal Fleet                                                 58
IX.    Comparison to Previous EPA Emission Factor Estimates          63

-------
I.   INTRODUCTION

    A.  Purpose and Coverage.

    The purpose of this document is  to  describe  the  methodology used in revis-
    ing the  emission factor  equations  for  the  post-1980  Light Duty  vehicle
    fleet.  The available data  will  be  presented and discussed  and  the perti-
    nent  assumptions   and   analyses   will   be  outlined.    Low   Altitude,
    Non-California, gasoline-fueled Light Duty vehicles will  be  the  only cate-
    gory of vehicles  covered  in this report.   Light Duty  Trucks,   Light  Duty
    Diesels,  California vehicles,  and  High  Altitude vehicles are all  covered
    under  separate analyses,  although  the other analyses  will often use  this
    document  as a  source  analysis.  This document  is only concerned with  the
    non-I/M case.   The I/M  case will be discussed in a separate analysis  and
    will result  in substantially  different  emission factor  equations.   This
    document  is one contributor  to  a  larger  effort designed  to  revise  the
    entire Mobile  Source Emission Factors  Document  (EPA-400/ 9-78-005).

    B.  Background on the  Post-1980 Fleet.

    The post-1980  Light  Duty vehicle fleet merits a  separate  analysis from  the
    current  fleet  for several  reasons.  Beginning in 1981,  the  Federal  exhaust
    emission  standard for  oxides of  nitrogen (NOx)  drops from 2.0 g/mi  to  1.0
    g/mi.   The hydrocarbon  (HC) standard remains at  0.41  g/mi. in 1981 and  the
    carbon monoxide (CO) standard  drops to  3.4 g/mi for most vehicles.   It  is
    the change in  the NOx standard specifically  that is of significance.   The
    effect of  this change  in the standard  will  be to  lead  manufacturers  of
    most vehicles  to adopt  a  technology which utilizes what  has become  known
    as a  Three-Way catalyst.    It  is called a Three-Way  catalyst  because  it
    allows not only  the  conversion  of  hydrocarbons  (HC)  and carbon monoxide
    (CO) as with a conventional  Oxidation  or Two-Way catalyst,  it also  allows
    the catalytic  conversion of  NOx.    Thus  it provides a  new and  previously
    unused source  of NOx control.  To enable the  catalyst  to perform  these
    three  conversion functions simultaneously,  precise control of the air/fuel
    ratio  is  required.  This  is  most often  accomplished through the use of  an
    on-board  microprocessor which  receives  inputs  from  a  variety  of sensors
    (notably  the  oxygen  sensor, which  is located  in the  exhaust  stream and
    provides  an indication of  the air/fuel ratio), processes these inputs con-
    tinuously,  and then provides an output  signal   to  the  carburetor or fuel
    injectors  to  adjust  the  air/fuel  ratio.  The  system  thereby  provides a
    feedback  loop  and  is  therefore  also known  as  a Closed  Loop  system.   A
    microprocessor  can  be  designed to control  other  engine functions as well,
    such as  spark timing,  idle  speed, and  EGR  flow rate.   Thus the net  effect
    of the change in  the  NOx  standard will  be  to introduce  significantly
    different  technology  into  the  fleet in large  measure  beginning in 1981.
    Due to the differences  between this new technology and  the more conven-
    tional  technology of  the  past several  years, it is  to  be  expected that
    there  will  be  differences  in  the  in-use emissions  performance  of the

-------
post-1980  fleet.   These differences  in in-use  performance necessitate  a
separate analysis.   This  fact was  recognized in  the  previous edition  of
the Mobile Source Emission Factors Document  (EPA-40Q/ 9-78-005) which  this
document revises and updates.

Two other considerations that affect the post-1980  fleet need  to  be raised
here.   First,  beginning in  1981,  the  "Parameter Adjustment"  regulations
(44 F.R. 2960) will be  applied to the  fleet.   These regulations will cause
limitations in  the  adjustability of some  of the  basic  parameters  of  the
engine.   In  1981,  these  regulations  will  be  applied  idle  mixture  and
choke, and in  1982,  they  will be applied  to  timing.  While these  regula-
tions  have  less of  an  effect on most  Closed Loop  vehicles than  on  most
Oxidation catalyst  vehicles,  due to the largely  self-adjusting nature  of
the  former, the  Parameter  Adjustment  regulations will nonetheless  have  an
impact, especially on certain parts  of the fleet.  The effect  of  the Para-
meter  Adjustment  regulations  has been  taken  into account  throughout  this
analysis.

A second consideration  affecting the  post-1980  fleet  is the  presence  of
the  Clean  Air  Act  Section  202(b)(5)  waiver  fleet:   those  cars  that  re-
ceived a CO waiver  from 3.4 g/mi to 7.0 g/mi in 1981  and  1982.   In 1981,
and to a lesser extent in 1982, a portion of  the  fleet will  be  designed  to
meet a 7.0 g/mi CO  standard as a result  of these waivers.    These cars  can
be expected to have higher CO emissions in general due to the  higher stan-
dard,  and  the  impact of those higher  emissions,  although  small, has  been
figured into the fleet-average emissions in this  analysis.

C.  Composition of the Fleet;   Technology Types.

Before discussing the methodology in  depth,   the  projected  make up  of  the
fleet  needs to  be  discussed.   In terms of the emission  control systems  to
be  employed  there  will  essentially  be  three  different   systems  in  the
post-1980  fleet,  however  two of these  three  are assumed  to have  similar
emissions performance.  Thus,  this  analysis  distinguishes  only two  tech-
nology types with unigue  emissions  performance from among  the  fleet.   The
distinguishing characteristics of the  two technology types  will be  briefly
presented below.

     1. Closed Loop Vehicles.

     By far the  largest percentage  of the  fleet  will be comprised  of  this
     technology type.   The distinguishing  characteristics   of  this  tech-
     nology type are feedback control  of the  air/fuel ratio and the use  of
     a  Three-Way  catalyst.   In  reality,  as  was  alluded   to  above,  this
     technology type could be  further  broken  down into two  separate tech-
     nology types:   those vehicles equipped  with  an  Oxidation  catalyst,
     supplied with  air  by an  air  pump, following  the  Three-Way  catalyst
     and those vehicles without  the additional Oxidation catalyst  and  air

-------
pump.   Three reasons  were  responsible  for not  making this  further
division  of Closed Loop  vehicles.   First, the manufacturers  of most
vehicles  have  indicated  that  during  significant   failure  modes  of
their  Closed Loop vehicles,  the air flow  from  the air pump  will  be
diverted  to  the  atmosphere  (or to  the air  cleaner  for  silencing)
instead of  to the Oxidation  catalyst.    This is due  to concerns for
catalyst  protection.   As  will be discussed later,  it  is these signi-
ficant failure modes of Closed  Loop vehicles  that are assumed to make
large  contributions to the  overall fleet composites.   During  a  fail-
ure mode  of this type, without the benefit of air  being  supplied  to
the Oxidation catalyst, the in-use data indicates  that  these  systems
have HC  and CO emission  levels as high as systems  without  the  addi-
tional Oxidation  catalyst and  air  pump.  Thus,   there  is  no  need  to
distinguish  between  these   two  systems  during  significant  failure
modes.

Second, examination of in-use data from both types  of vehicles  (with
and without  the additional  Oxidation catalyst and  air  pump)  did not
reveal a  significant  and  consistent  difference between the  emissions
of  the two  systems when  operating  at other than significant  failure
modes.

Third, the  presense or absence of an Oxidation catalyst and air pump
is not assumed to effect NOx significantly.

In  summary  then,  the  two systems  will  be treated as  one  technology
type having a unique emissions performance.

One final point to be made regarding this technology  type  has  to  do
with Ford Motor Company vehicles in  the 1981-1983  timeframe.    Ford
has  indicated  that  it intends  to  certify a  large  portion  of its
1981-1983 fleet as  open loop vehicles  equipped  with  Three-Way  cata-
lysts.  That  is,  these vehicles  would  not employ an  on-board micro-
processor  with  a feedback  oxygen  sensor but   would   still  have   a
Three-Way catalyst  to enable some  catalytic  reduction  of NOx.   Due
to  a  lack of any in-use  data on  these  systems  at  the time of  this
analysis, as well as   to  uncertainty  as  to the  fraction of Ford  vehi-
cles which will have  open loop systems, these vehicles  were included
under  the Closed  Loop technology type.    In  1984,  any  Ford  vehicles
which  had been  open loop are assumed to  go  closed  loop  due to the
advent of the High Altitude  regulations.

2.  Oxidation Catalyst Vehicles.

Some  vehicles,  notably   small   foreign  vehicles  and  vehicles   with
unique engine configurations, will be able  to meet  the  1981  standards
without the  catalytic control  of NOx provided by  Three-Way catalyst

-------
     technology.  Instead,  they  will rely  on an  Oxidation catalyst,  air
     pump or  pulse-air  system,  and, in  most  cases, EGR.    These vehicles
     are expected to comprise a  relatively small  segment of the fleet.

D.  Summary of Results.

The results of  this analysis can best be presented by  discussing the vari-
ous failure modes expected to occur and  by  briefly characterizing the per-
formance of each  pollutant  (HC,  CO, NOx).  For  the majority of the fleet,
which consists of Closed Loop vehicles,  the principal  failure  mode result-
ing  in  significant  emission increases   occurs  through  the  loss of  the
closed loop capability  of the system and a resultant  rich mode  of  oper-
ation.   These failure  modes, while not  initially  very  numerous, have  a
large impact on the overall  fleet emissions for  HC and CO due  to the very
high  emissions  resulting for those two  pollutants.   Thus,  a relatively
small percentage of the  fleet contributes a disproportionately large  share
of the  final  composite  emissions for HC and  CO.   This is  especially true
for CO.  The data base  of in-use vehicles which  was relied  on in perform-
ing this analysis gave  significant  indication of this  type  of behavior for
Closed  Loop  vehicles.   NOx  emissions  for  those  cars  with an  open  loop
failure will decrease due to the rich operating condition.

For Oxidation Catalyst  vehicles, which  are designed to operate open  loop,
a  more  traditional  deterioration   pattern  is assumed  to occur.   Briefly
stated,  the regression methodology  used  for the  1975-1980 Light Duty  fleet
was revised to  represent 1981 and  later  Oxidation  Catalyst  vehicles.   The
revised  methodology  accounts for the effect  of the Parameter Adjustment
regulations and for the fact that  the  1981  standards are  more stringent
than the 1975-1980 standards.

A  graphical  comparison  of  the  new  emission  composites with  the emission
composites arrived  at  in the  1978 analysis  (Appendix  E  of the  Mobile
Source  Emission Factors Document,   EPA-400/9-78-005)  will be  presented  in
Section  IX.  To generally characterize that comparison  however, there is  a
slight  increase  in  the  HC composite at  50,000   miles,  a  definite increase
in the CO composite at 50,000 miles, and  a  slight  increase  in  the NOx com-
posite at 50,000 miles.

The following table presents the new emission  factor equations:

-------
                                    Table 1-1
                 Exhaust Emission Rates  for  Light Duty Vehicles
                For  All Areas Except California and High Altitude
                                  New Vehicle    Deterioration Rate
Pollutant          Model Year     Emission Rate  (g/mi/10,000 miles)
                                    (g/mi)
HC
HC
HC
CO
CO
CO
NOx
NOx
NOx
81
82
83+
81
82
83+
81
82
83+
0.39
0.39
0.39
5.60
5.21
5.00
0.75
0.75
0.75
0.19
0.19
0.19
2.75
2.76
2.76
0.15
0.15
0.15

-------
II.  DESCRIPTION OF THE DATA BASE

    A.  Introduction.

    The data  used  in this analysis come  from a variety of sources.  The bulk
    of the  data comes  from  EPA testing  programs  in Los  Angeles designed  to
    test  in-use Closed Loop  vehicles.[1,2,3]*   In addition,   data  from  50
    in-use Closed Loop  vehicles tested  in Portland,  Oregon were  used.[4]*  For
    each set  of data,   the effect  of  the Parameter  Adjustment regulations  was
    accounted for by removing from the  data  base those cars which showed evi-
    dence of removal of idle mixture limiting  devices where applicable,  a mal-
    adjustment of idle  speed of greater than 200 rpm**, or a timing  maladjust-
    ment of greater than +5°.   In some  cases,  data  from vehicles with  evidence
    of  maladjusted  parameters,  but  which  also had  evidence  of  other  non
    parameter-related  problems, were   retained in  the  calculations  of  the
    levels of pollutants  not primarily affected by  the maladjusted  parameter
    and/or in the  calculation of the  incidence of non  parameter-related prob-
    lems.

    A  final  preliminary  consideration  which  needs  to  be mentioned  at  this
    point has to do with  the  methane  correction factor for vehicles certified
    in California.   The California certification process accounts for  the fact
    that a  certain  portion  of  the  HC measured  in the  exhaust  is methane
    (Cfy).  Vehicles  being  certified   in California are  allowed to  claim  a
    "credit" for that  portion  of the HC" exhaust which is methane.    The  de-
    fault credit is  11% methane, however manufacturers  can  claim more  credit
    by demonstrating  that  their vehicles emit a higher  fraction of  methane.
    Since most of  the  vehicles  in  the   data base were certified  in California,
    yet this analysis is concerned with non-California  vehicles, this differ-
    ence needed to be accounted for.   Use of  the methane credit  has  the  effect
    of raising the  total .HC design standard  from the perspective of this ana-
    lysis.   For example, vehicles receiving  the default methane  credit  of  11%
    can emit up to 0.46 g/mi total HC  and still pass California  certification
    after the credit has been applied.   This analysis accounted  for  this rela-
    tive difference in effective design standards by applying a ratio  of those
    design standards (e.g. 0.41/0.46)  to the  HC emission  levels found  in  the
    data base, except in those cases where the emission levels were judged to
    be independent  of  the  design standard.   These exceptions  will  be pointed
    out as they  occur.

    The principal data used for analyzing the  emissions performance of the  two
    technology types  will  be  briefly presented  in the following sections.
    * Numbers  in brackets  indicate  references  listed at the end of the Section.
    ** At the  time the analysis was finalized,  idle speed  was included in the
    Parameter  Adjustment regulations as a parameter which would need to be de-
    signed to  be non-adjustable beginning in 1982.  That requirement has since
    been, lifted.   The ' analysis  was not  revisted to  account for  the change
    since it was  determined  to not have  a  significant  impact and  due  to the
    fact that  the composite emission factors had already been  finalized.

-------
B.  Closed Loop Vehicle Data Base.

The  vehicles  from the  following four  groups were  combined  to  create  an
overall data base to  be used in determining  the emissions  characteristics
of  Closed Loop  vehicles.   An  assortment of various  systems  and  engine
sizes are  represented.   This data base  is  assumed to  give  representative
emission values  for  Closed Loop vehicles.   The  four groups which make  up
the data base  represent the most recent  and advanced technology put  for-
ward by  the  various  manufacturers.   The simpler,  first  generation  systems
were excluded  from the  data base (e.g.  1978/ 1979 Ford Pintos,  1978/1979
Pontiac Sunbirds) due to being judged unrepresentative.

     1.  1979 Ford/Mercury 351 CID engine family  (3.8WBV2TT95x95)
     equipped with Electronic Engine Control  II (EEC-II).

          a.   A total of 97 in-use vehicles were tested.  82 vehicles  were
          tested  by   contractors  in  Los  Angeles,  and 15  vehicles  were
          tested in Portland, Oregon.  Six vehicles were eliminated  due  to
          Parameter  Adjustment  concerns.   This  engine  family  is equipped
          with an Oxidation catalyst  following the Three-Way catalyst and
          an air pump which supplies air to the Oxidation catalyst.

          b.    This  engine  family  uses  a digitally  based microprocessor
          which allows very sophisticated control  of  the engine.  It regu-
          lates not  only the air/fuel  ratio, but also spark  timing, EGR
          flow rate,  and the deployment of air flow from the  air  pump.

          c.  Of the 15  vehicles tested in Portland,  Oregon,  one was  from
          California  and was therefore designed to meet  the  1979  Californ-
          ia  emissions .standards   (0.41/9.0/1.5).   The  other   vehicles,
          while being sold  outside of  California,  were  nonetheless  assumed
          to be designed to meet the  1979 California standards for  HC and
          C0_ and  the" 1979 Federal Standard  for  NOx   (2.0 g/mi).  ThUf was
          done based  on the  technical assumption that the 1979 351 CID
          engine family was intended to be an  in-the-field test of  the EEC
          system.  As such, it  was  primarily designed to meet the  tighter
          California  standards to give a better indication of how well the
          system would  perform  under the eventual 1981 Federal  standards
          (0.41/3.4/1.0).    Thus the   basic   system  calibration  was  de-
          signed  to  meet  the  1979  California  standards rather  than the
          looser 1979 Federal  standards (1.5/15.0/2.0).  For the vehicles
          sold outside  California,  however,   (i.e.  the  vehicles  tested  in
          Portland)  it  is  assumed that the calibrations pertaining  to NOx
          (e.g.  spark  timing,  EGR  flow  rate) were  relaxed due  to  fuel
          economy and  driveability  concerns.  This   could easily be  done
          without significantly affecting  the primary  HC  and  CO calibra-
          tion.   The in-use data from Portland well  supports this  assump-
          tion: the average HC and CO levels  from  Portland are the same  as
          or lower than  the average  HC and CO levels from Los Angeles and
          the average Portland NOx levels are higher.

-------
          d. These vehicles received a 17% Methane credit.

     2.1980 General Motors California X-Body vehicles (01C2XCP
     01C2XC. and 02X2NC) equipped with the Computer Controller?
     Catalytic Convertor system (C-4).  151 and 171 CID.

          a.  A  total of  92 in-use  vehicles  were tested,  all  in  the Los
          Angeles  area.   9  vehicles  were  eliminated  due  to  Parameter
          Adjustment concerns.

          b.  These  vehicles employ a digitally based  control  system.  The
          C-4 system  in this  application  primarily controls the air/fuel
          ratio.   The  C-4  system  is  the  basic  system  projected  to  be
          employed by General Motors throughout the early 1980's.

          c. These vehicles received the standard 11% Methane credit.

     3.  1979 Toyota Celica Supra vehicles (4M-E) 156 CID.

          a.  A  total of  25 in-use  vehicles  were tested,  all  in  the Los
          Angeles  area.  Two  vehicles were  eliminated  due  to  Parameter
          Adjustment concerns.

          b.  These  vehicles are fuel  injected  and employ an analog based
          control system.

          c. These vehicles received.the standard 11% Methane credit.

     4.  1979 VW Audi:  5000.  131 CID.

          a.  A  total of  4  in-use  vehicles were  tested, all  in  the  Los
          Angeles  area.  No  vehicles  were eliminated due  to  Parameter
          Adjustment concerns.

          b. These vehicles are fuel injected.   They do not use  EGR.

          c. These vehicles received the standard 11% Methane  credit.

C.  Oxidation Catalyst Vehicle Data Base.

     1.  Data support for  this technology type, which  is similar  in most
     respects to current oxidation-catalyst-and-air-pump  technologies,  was
     obtained from  the  analysis completed  to revise  the emission  factor
     equations for  1975-1980  vehicles.   That  analysis was based  on in-use
     data obtained  from a  wide  array of  1975-1979 model  year  vehicles
     tested around the  country through EPA's Emission Factor Program.  The
     specific methodology involved in  applying  that  data to this  analysis
     will be presented later (Sections IV and  V).

-------
    D.  Miscellaneous Data.  1978-79 Volvos and Saabs equipped  with  the Lambda
    Sond system  (4CL, 6CL,  BT20CA  and BSI20CA).   A total  of  162  Volvos  and
    Saabs were tested in  programs  in Los Angeles and Portland.[1,2,3,4]   Data
    from these  vehicles  were used  chiefly  in determining  deterioration  rates
    for Closed Loop vehicles (see Section  IV.A.2.b.).

    This fleet has  a  much wider mileage spread than found  in the  major Closed
    Loop vehicle data base  and  could therefore be used in  determining  deteri-
    oration rates.  All  Volvo and  Saab vehicles  marketed west of the  Missis-
    sippi have the  same  engine  calibrations, so  the Portland and Los  Angeles
    vehicles  could be  analyzed  together  without  needing  to  take  different
    standards into account.

References for Section II.

1.  EPA  Contract  No.  68-03-2590 with Automotive Environmental  Systems,  Inc.
Results published in EPA  document EPA-460/3-79-004.

2.  EPA  Contract  No.  68-03-2889 with Automotive Environmental  Systems,  Inc.
Results published in EPA  document EPA-46Q/3-80-006.

3.  EPA  Contract  No. 68-03-2774  with  Automotive Testing Laboratories,  Inc.
Results not published at  the time of this  report.

4.  EPA Contract No. 68-03-2829, Test Group  No.  1,  with  Hamilton  Test Systems,
Inc.  Results published in EPA document  EPA-460/3-80-006.

-------
                                       10
III.  GENERAL METHODOLOGY

    A.  Introduction.

    This  section  presents  the general  methodology used  in constructing  the
    emission factor model for the post-1980  fleet.   Specific  methodologies  are
    presented in detail  in Sections IV  and  V.   The HC/CO methodology  and  the
    NOx methodology will  be discussed separately.   HC and  CO  have been ana-
    lyzed together due to a basic similarity in  the in-use failure modes which
    result in  high HC and  CO emission  levels,  whereas the  failure modes  to
    which NOx is sensitive are relatively  independent.   The  general methodol-
    ogy used in this  analysis is  similar in  concept to that  used in the 1978
    analysis (Appendix E)  as will  be discussed later in this  section.

    B.  HC/CO Methodology.

         1.   Unit of Analysis.  The  basic  units  of analysis  (i.e.,  the  unique
         subfleets  of similar vehicles for which unique analyses are performed
         and which  are  aggregated later into a  fleet-composite analysis)  are
         individual technology  types  certified  to specific  pairs  of  HC/CO
         standards.  Thus, each of  the  two technology types discussed  in Sec-
         tion I.e.  will comprise a  separate  unit of analysis  for  each of  two
         possible  combinations of HC/CO  standards,  for a  total of  four  units
         of analysis.  The  two possible combinations  of  HC/CO standards  are
         0.41 g/mi  HC with 3.4 g/mi CO,  and 0.41  g/mi HC  with  7.0 g/mi  CO.
         This second combination stems  from the CO waiver decision.

         2.   Categories.   Each unit of analysis  representing Closed  Loop  vehi-
         cles will  be divided  into  categories  (i.e.  groupings  of vehicles
         within the overall unit of analysis  which  have similar emission  char-
         acteristics).  Several different vehicle conditions  could be included
         under and contribute  to  the size of a  category,  but a single set  of
         emission  characteristics  for  each  pollutant  describes the category.
         Oxidation  Catalyst vehicles will not be  divided into  categories.

         3.    Emissions Performance  of  the  Categories.    The  average emission
         characteristics  of each category  for a  given pollutant are described
         by  two parameters:

              a.   Zero-mile level.    The  zero-mile   level is  the average emis-
              sion TeverTor~a~pollutant  at zero  miles.

              b.   Deterioration rate.  The deterioration rate  is  the amount
              of increaseInthe  emission  level  of  a pollutant  per  10,000
              miles.

         The  unit  of measurement  for the zero-mile  level is g/mi.  The unit  of
         measurement   for  the deterioration  rate is  g/mi/10,000  miles.    The
         deterioration rate  should  not  be  confused with the  "deterioration

-------
                                  11
     factor" which is the ratio of  the emission level  at  50,000 miles  to
     the emission level at 4,000  miles.   The concept of the  deterioration
     factor will be used later.

     4.   Migration between Categories.   Within  a given  unit  of  analysis,
     for example the Closed Loop  vehicle technology type designed to  meet
     a 0.41 g/mi  HC standard  and  a 7.0  g/mi CO  standard,  the size of the
     different categories will  change with  time.   As  would  be  expected,
     the general  movement  over  time  is  from categories representing  more
     well-maintained vehicles with  lower  emissions _to  categories repre-
     senting less well-maintained vehicles or vehicles with some  component
     failure with correspondingly  higher emissions.    The growth  rates  for
     the categories are expressed in units of percentage of the total  unit
     which   enters the  category  per   10,000  miles.   Some categories  of
     course will have negative growth rates.   That is, as some categories
     grow,  others must  decrease  in size.

     5.   unit-of-Analysis  Composite  Emissions.   The  emissions from each  of
     the categories  within  a  unit  of analysis  will  be weighted together
     every   10,000  miles to arrive  at  a composite emission characteristic
     for that unit.  This  will be  done by weighting the emission  levels  of
     each category  at  a given mile point  by the fraction of the unit  of
     analysis represented by  that category at the  given  mile point.   The
     net result  will be  a  table of  the  unit's  emission levels. at  each
     10,000 mile point.  This  table will reflect the relative contribution
     of the various categories based on the size of the categories as  well
     as  their individual emission  levels.  Examples  of this concept can  be
     found  in Sections  IV  and  V.

     6.   Fleetwide Composite Emissions.   To  obtain a  fleetwide  composite
     emission factor for a given  model year,  the emission composites  from
     each of the units of analysis will be combined according to  a weight-
     ing scheme similar to  that described above  for  combining categories
     within a  unit  of  analysis.,  The  separate  unit-of-analysis emission
     composites will be weighted  according to the respective  fractions  of
     that  model  year's  fleet  represented  by  those  units  of  analysis.
     These   fractions were determined  by examining  manufacturer's state-
     ments  and certification data.  The determination of how the  fleet was
     broken down into the various units of analysis will  be  discussed  in
     depth  for each model  year in  Sections  VI through VIII.

C.  NDx  Methodology.

The NOx methodology is essentially identical in concept to the HC/CO meth-
odology.  In the following subsections only the pertinent differences  will
be mentioned.

-------
                                   12
     1.   Unit of Analysis.   For  NOx,  there  are  only two  units  of  ana-
     lysis.   There  is  only  one applicable NOx standard,  1.0 g/mi, and NOx
     is not  assumed to be  affected by  the  different CO  standards intro-
     duced by the waiver decision.  Thus the two units of analysis are:

          a.  Closed Loop vehicles
          b.  Oxidation Catalyst vehicles.

     2.  Categories.

          a.  Closed loop  vehicles will have four  categories  of operating
          condition.

          b.   Oxidation  Catalyst vehicles will  not be divided  into  cate-
          gories.

     3.  Emissions Performance of the Categories. - No difference.

     4.  Migration between Categories. - No difference.

     5.  Uhit-of-Analysis Composite Emissions.  - No difference.

     6.  Fleetwide Composite Emissions. - No difference.

D.  Comparison to the Methodology used in the Previous Analysis.

     1.  The methodology used in  this  analysis  is similar to  that  used  in
     developing Appendix E  (Appendix  E of the  Mobile  Source Emission  Fac-
     tor Document  EPA-400/9-78-005).   Appendix  E  divided a unit of  ana-
     lysis into different categories  with  independent emission performance
     characteristics  (zero-mile  levels  and deterioration  rates)  as  has
     been done  in  this analysis.   Those  categories also  grew or  declined
     in size over time, and a unit  composite was obtained by weighting the
     various  categories and  adding them together.   The chief  methodologi-
     cal difference between  Appendix  E and this analysis is that  Appendix
     E used  a single  unit  of analysis to represent the  entire  fleet for
     all model  years  after  1980  whereas this  analysis  has four  distinct
     units of analysis for  HC and  CO and two units of  analysis  for  NOx.
     Appendix E  assumed one  technology type;   this  analysis  assumes  two
     different technology types.   Appendix E treated  only  one pair of  HC
     and CO  standards;  this analysis  has needed  to account  for a wider
     array of combinations  of HC  and CO standards  due  to the waiver de-
     cisions.  As will be apparent  later,  the  categories defined  for  this
     analysis are also very different  from the categories used in  Appendix
     E, as are the categories'  emission levels.

-------
                                       13
IV.  SPECIFIC UNIT ANALYSES - HC/CO

    A.  Closed Loop vehicles designed to meet 0.41 g/mi HC and
    3.A g/mi CO standards.

         1.   Category definitions.

              a.   Primary.   This category  is  comprised of  those  vehicles,
              designed to  operate closed  loop,  which have  either lost  their
              closed  loop  capacity,  or  have  had  it  severely  restricted.
              These vehicles  are  assumed  to then  operate in  a rich  mode.
              Possible  contributors  to  this  category  include  vehicles  with
              catastrophic  oxygen  sensor   failure,  microprocessor   failure,
              other sensor failures or  tampering with the  closed loop  system.

              It was  necessary to  select from  the  data  base  the  subset of
              vehicles which  would be relied  upon in predicting the  emission
              characteristics of  this  category.   This  was  done by  selecting
              all  vehicles  which  had  CO emission  levels  greater  than  50.0
              g/mi, on  the premise  that any  vehicle with  CO emissions  this
              high must be experiencing rich operation due to  a  failure of the
              closed loop control system.

              b.   Secondary.   This category  is  comprised  of all the  vehicles
              in the  fleet  not in  the  Primary  category  and not  in  the  Mis-
              fueling  category  (see  below).   Contributors  to  the  Secondary
              category include  vehicles  experiencing general  malmaintenance,
              failure or degradation of  sensors  not leading  to loss of closed
              loop operation,  tampering,  as  well  as a  large  percentage of
              vehicles experiencing  generally  good maintenance with resultant
              low emission levels.

              As with  the  Primary category,  a  subset of  the  data base  fleet
              was selected to be  used in predicting the emission characteris-
              tics of this category.  This was  done by selecting all  vehicles
              which had CO emission levels less  than 50.0  g/mi.

              c.   Misfueling.   This  category  is comprised  of vehicles  which
              have been misfueled  and have therefore had  their emission  con-
              trol systems  damaged.   Misfueling  refers  to the  use  of  leaded
              fuel in  vehicles  equipped with catalysts.   The leaded  fuel in
              effect "poisons" the catalyst and dramatically reduces its  abil-
              ity to convert HC, CO and NOx to harmless by-products.   Misfuel-
              ing also damages the oxygen  sensor  on Closed Loop vehicles.  It
              affects the output  voltage of  the  sensor  as  well as hindering
              the ability of  the  sensor  to respond  quickly  to changes in  the
              composition  of  the  exhaust  stream.   This  normally  results in
              more rich operation  and  consequently higher HC  and CO  emission

-------
                             14
     levels.   The  interaction of  this category  with  the other  two
     will be described in a later  section  dealing  with category sizes
     (Section IV.A. 3).

2.   Emission levels of the categories.

     a.   Primary Category.

     i.   Zero mile levels.   The  average   measured  HC/CO  emission
     levels and the average odometer mileage of the  vehicles  selected
     from the data base to represent  the Primary  category are:

          Average HC = 3.85 g/mi         (n  = 10)
          Average CO = 108.rf g/mi       (n  = 10)
          Average Mile = 9,163 miles     (n  = 10)

     (Note:   Of the ten  vehicles in  the data base which  met  the cri-
     teria to be selected  to  represent the Primary  category  (greater
     than 50.0  g/mi  CO)  five were  from  the General Motors  X-Body
     fleet and five were from  the Ford  351  fleet.)

     Applying the  deterioration  rates  (see below)  to these  average
     emission levels results in  the following back-projected  emission
     levels at zero miles:

          HC = 3.74 g/mi
          CO = 107.36  g/mi

     (Nate:    The  reader will  note  that  for the  Secondary  category
     discussed below,  the zero-mile CO level was adjusted  by  a ratio-
     ing factor of  3.4/9.0  to  account  for  the fact  that  the  vehicles
     in  the  data base  were designed  to  the California  CO  emission
     standard of 9.0  g/mi  rather than  to the 3.4  g/mi standard which
     applies to this unit of analysis.   The CO emission level  was  not
     ratioed  to  the  3.4  g/mi CO  standard for  the  Primary  category
     since  this  category  represents  vehicles with  a  major, failure
     mode during  which  CO emissions  are  assumed to  be  effectively
     independent of the CO design standard.  This same  logic  dictated
     that the HC  emission  levels need  not  be ratioed to  account  for
    . the various  effective HC  design  standards  resulting  from  the
     application of methane credits.   See Section  II.A.  for  a more
     complete discussion of this  issue.)

     ii.  Deterioration rates.   The HC/CO deterioration rates  for  the
     Primary  category were  obtained  by  adopting  the  deterioration
     rates  developed  for  the  Secondary category.   That  development
     will be  presented in  the  next section.  The deterioration rates

-------
                         15
for  the Secondary category were  adopted due  to a  lack  of data
describing  the deterioration of  Primary  category vehicles.  The
deterioration  rates arrived at are:

     HC= 0.12  g/mi/10,000 miles
     C0= 0.71  g/mi/10,000 miles

b.   Secondary Category.

i.   Zero-mile levels.   The   average  measured  HC/CO  emission
levels  and  the average odometer  mileage  of the  vehicles  in the
data base which represent the Secondary category are:
     Average HC = 0.32 g/mi        (n = 191)*
     Average CO = 5.47 g/mi        (n = 195)*
     Average Miles = 8,064 miles   (n = 191)

Applying  the  deterioration  rates (see  below)  to  these  average
emission  levels  to  back-project  the  levels to  zero miles  and
then  applying  a  ratio  of  3.4/9.0  to the  CO emission  level to
reflect  the  fact that  the vehicles  in  the  data base were  de-
signed to meet a  9.0 g/mi CO standard and not the 3.4 g/mi stan-
dard which is the focus of this analysis,  results in the follow-
ing Zero-mile emission levels:

     HC = 0.23 g/mi
     CO = 1.86 g/mT

ii.  Deterioration rates.

The development  of  deterioration rates  for  the  Secondary  cate-
gory was an involved process.  Before presenting  that process in
detail, several points can  be  made which  will give helpful  back-
ground on  why the  deterioration rates  were  developed  as  they
were.

First, the vehicles  in  the Closed Loop vehicle data  base do  not
have a large enough  mileage  spread to support a  credible  regres-
sion  analysis  (which  would  yield  a  deterioration  rate  and
zero-mile level).   The vast  majority of  vehicles  in the data
base have odometer mileages between  6,000 and 12,000 miles.
* The sample sizes used  to  calculate  the  average emission levels
for  the  two pollutants  differ due  to the  fact that  Parameter
Adjustment  concerns  eliminated  vehicles  for  one pollutant  but
not for the other.  See Section II.A.

-------
                        16
The next most  apparent  source  of data for examining  the deteri-
oration of Closed Loop  vehicles  is  the deterioration  observed in
the EPA  Certification process.  For  each engine  family with  a
unique emission control system produced in a  given  model year,  a
representative  vehicle  from  that  family must  demonstrate  the
ability to meet the applicable emission standards  out  to 50,000
miles in  order to  be certified.   While  the deterioration  ob-
served in  the  Certification  process  is  believed  to  be a  good
measure of the  relative durability of a given vehicle's emission
control system,  it  does not,  however,  give  an  accurate picture
of how the vehicle will perform out in the field.   It  has always
been the case  that  in-use  vehicles deteriorate at  a  faster  rate
than that observed  in Certification.   This is due  to  owner  mal-
maintenance  and  tampering,  harsh real-world  conditions,   and
other  factors.   This is  assumed to  still be  the case in  the
post-1980 timeframe.

Given  the fact then,  that  the  data  base  used  to  determine
zero-mile levels could  not  be  relied  upon to  predict  deteriora-
tion and  that  the  use of Certification .deterioration  factors by
themselves would  be  unrealistic,  it  was  necessary  to rely  on
other sources  of data to determine the deterioration  rates.   As
was  mentioned   in  Section  II.D.,  there  is  a  fleet  of  in-use
Closed Loop  vehicles,  the  Volvo/Saab fleet,  which does have  a
significant mileage  spread.   This  fleet is made up of  162  1978
and 1979 Volvos and  Saabs with a mileage  spread  of between 0 and
30,000 miles.   There are,  however, substantial  reasons why  the
in-use deterioration observed for  these  vehicles  could not  be
adopted outright  to  represent  the in-use  deterioration of  the
post-1980  fleet.   These  reasons  center  around  differences  in
technology between  the Volvos  and Saabs  and typical  post-1980
vehicles.  The  Volvos and  Saabs  are fuel  injected  as  opposed to
carbureted, have an  analog  based control  system as opposed  to  a
digitally  based control  system,  do   not use  EGR and  have  a
European  manufacturer.    These   differences  are  significant,
especially so  for NOx.  Since an approach was needed  which could
be applied consistently to  all  three  regulated  pollutants,  (HC,
CO, NOx), the  differences  in  technology  needed to  be  taken  into
account.

In  view  of  all  of  the  above-mentioned  considerations,   the
following procedure  was decided upon.  First,  a regression  was
performed on the Volvo/Saab fleet after removing  vehicles  due to
Parameter Adjustment  concerns.   Two vehicles  with  FTP  CO  emis-
sions  greater  than  50.0 g/mi  were also  removed from  the  fleet
since  these  vehicles  would be  classified as  Primary  category
vehicles.  The  regression  yielded  a zero-mile level and  a  slope
which represent in-use  deterioration.   Second, the  deterioration

-------
                         17
predicted  by Certification  for  the  1978  and  1979  Volvos  and
Saabs  was  quantified by  weighting  together  the  Certification
deterioration  factors (d.f. *s)  for  each  of the  possible model
year/engine  size  combinations  among  the Volvo/Saab  fleet.   The
weighting  technique  was  based  on  the  number of  vehicles  in the
fleet which  came  from the  various model year/engine size combin-
ations.   There were  some   HC  d.f. 's  from  some  of  the  model
year/engine  size  combinations  which  were  less  than 1.0  which
were raised  to  1.0  for the  purposes of calculating the  weighted
average.   Third,  a  ratio was formed using  the  in-use deteriora-
tion and  the Certification  deterioration.   This ratio  gives  an
indication of how Secondary  category  Closed Loop vehicles deter-
iorate  in  the  field  relative  to how  prototypes  of  the  same
models  deteriorate  in the  Certification process;  the ratio can
be  used  as  an  adjustment,  or  correction  factor,   to  predict
in-use  deterioration  of  Secondary   category  vehicles  given  a
figure  of  Certification  deterioration.   This ratio was  based  on
units  of  the  "deterioration factor minus  one"  (d.f.-l).   (The
in-use  deterioration  regression was easily  converted  to  a d.f.-l
by first finding  the  emissions  at 50,000 miles  and 4,000 miles,
taking  the ratio  of  those  two figures to obtain  a  d.f.  and  then
subtracting  one to  obtain the  d.f.-l).   The d.f.-l was  used  in
the ratio  since it  is the portion of  the d.f.  greater  than one
which represents  the  percent increase  in a pollutant over 50,000
miles.  To use  the  d.f.  by  itself in  the ratio  would have  been
misleading  since  the  d.f.  is itself a  ratio  (the  emissions  at
50,000 miles over the emissions at 4,000 miles).

Finally,  the ratio  described  above,  which  predicts  the  rela-
tionship between  Certification  deterioration and  in-use deteri-
oration, was applied  to  an  average Certification d.f.-l  for  1981
Closed  Loop vehicles  in  the Certification  process  (n =  89).
Applying  the ratio  yields  a figure  for  an  in-use  d.f.-l  which
can be  used to represent  the in-use  deterioration  of Secondary
category post-1980 vehicles.   The following equation illustrates
this relationship:
    x (d.f .-l)ynivn/Saah in-use = (d«f --98+  n-use
      (d.f.-l)volvo/Saab Cert.

The in-use d.f.-l arrived at  can  be converted to a d.f.  by add-
ing one.  This deterioration  factor  (d.f.)  can then be converted
to a  deterioration  rate (d.r.) by  using the  following  equation
and the average emission levels at  the  average mileage (approxi-
mately 8,000 miles)  of the  Secondary category vehicles  found  in
the Closed Loop in-use data base.

-------
                                      18
d.f.[in-use]  = Emissions at 50.000 miles   _  Emissions  at 8,000  miles  +  (4.2)(d.r.)
               Emissions at 4,000 miles    ~  Emissions  at- 8,000  miles-(.4)(d.r.)

              The only unknown in this equation  is  the deterioration  rate
              (d.r.) which can therefore  be  easily  solved  for.

              Two exceptions to this  procedure need to be  mentioned here be-
              fore presenting a table of  the values used in  the calculations
              and the deterioration rates arrived at.  First, for HC, the
              average Volvo/Saab Certification d.f. (1.02) was  so low that
              using the d.f.-1 to find the ratio between in-use and Certifica-
              tion deterioration resulted in an  unrealistically high  ratio
              (40.5).  In view of how low the HC d.f.  is and how it thereby
              unrealistically inflates the in-use/Certification ratio,  it was
              decided to regard the HC Certification d.f.  as an anomaly and  to
              adopt the ratio observed for CO (5.6) to be  used  for HC as
              well.   Given the fact that  HC  and  CO  emissions are both affected
              in the same way by relatively  similar malperformances for Closed
              Loop vehicles, this approach is judged to give a  better estimate
              for the HC ratio.

              The second exception to the general procedure  outlined  above has
              to do with the fact that for CO, the  vehicles  in  the data base
              were designed under a 9.0 g/mi CO  standard whereas this unit of
              analysis consists of vehicles  designed to meet a  3.4g/mi  stan-
              dard.   The following set of equations was used to translate from
              the in-use data point based on 9.0 g/mi  CO standard vehicles to
              arrive at a slope for vehicles designed  to a 3.4  g/mi CO  stan-
              dard.   These equations  incorporate the assumption of equal
              slopes (or deterioration rates) for vehicles designed to  meet
              3.4 g/mi or 9.0 g/mi CO standards  and the assumption that the
              Zero-mile levels for vehicles  designed to 3.4  g/mi or 9.0 g/mi
              CO standards are in the same ratio as their  standards (9.0/3.4).

                   1.  d.f.81 = a + 4.2(s)
                           01    a -  .4(s)

                   2.  Z9 = b - (.8)(s)

                   3,  23.4 = a - (.8)(s)

                   4.  Z9/Z3t4 = 9.0/3.4

                        d.f.si = in-use d.f. for post-1980 vehicles (derived
                        as described  above for the  general case).

                        a = CO emissions  at  8,000 miles for  Secondary vehicles
                        designed to meet  the 3.4 g/mi  CO standard (an unknown).

-------
                          19
           s = in-use deterioration rate (unknown).

           b = CO emissions at 8,000 miles for Secondary vehicles
           designed to meet the 9.0 g/mi CO standard (known from
           the in-use data).

           19 = Zero-mile CO emissions for Secondary vehicles
           designed to meet the 9.0 g/mi CO standard (unknown).

           23.4 = Zero-mile CO emissions for Secondary vehicles
           designed to meet the 3.4 g/mi CO standard (unknown).

 These equations can be combined given the in-use data point (b)
 and the calculated d.f. (d.f.31) to arrive at a CO deteriora-
 tion rate (s) for vehicles designed to meet the 3.4 g/mi CO
 standard.

 The following table presents the pertinent values used in calcu-
 lating the HC/CO deterioration rates.
                      Table IV-1
Input Values to the Calculation of Deterioration Rates
HC
CD-
Volvo
In-Use
d.f.-l
0.81
0.96
Volvo
Cert .
d.f.-l
0.02
0.17
Actual
In-Use/
Cert Ratio
40.5
5.65
Ratio Used
for In-Use/
Cert.
5.65
5.65
Average
1981 Cert.
d.f.-l
0.37
0.27
Resultant
d.f.-l
2.09
1.52
Resultant
1981+ d.f.
3.09
2.52
 Applying the resultant d.f.'s to the in-use data points  found  in
 the data base as described above gives the following  HC/CO
 deterioration rates:

      HC = 0.12 g/mi/10,000 miles
      CO = 0.71 g/mi/10,000 mile's"

-------
                        20
c.   Misfueling Category.

i.  Zero-mile levels.  The zero-mile levels for the Misfueling
category were obtained in  a two-stage process.   First,  the
amount of damage caused by misfueling was quantified.   This was
done by examining the results from two programs which measured a
number of vehicles'  emissions before and after  extensive mis-
fueling.  Both of these programs were performed under contract
to EPA.  One was performed by the California Air Resources Board
(CARB)[1] and the other was performed by an independent con-
tractor, Automotive  Testing Laboratories Inc. (ATL).[2]  These
two programs tested  a total of nine Closed Loop vehicles.  Three
of the nine vehicles were  removed from the analysis due to con-
cerns about engine problems the car might have  been experiencing
or due to the vehicle having only had a small amount of leaded
fuel run through it.  For  the remaining six cars,  the average
emission levels before and after misfueling were examined to
determine the average percent increase in emissions due to mis-
fueling.  Those average percent increases are:

     HC = 364% (n=6)
     CO = 128% (n=6)

These figures represent increases due to oxygen sensor  damage
and other engine-out effects as well as catalyst damage.

The second stage of  the procedure was to add these percent in-
creases onto the Zero-mile emission levels for  the Secondary
category.  Vehicles  in the Primary category are also expected to
be misfueled, at the same  proportional rate experienced by the
rest of the fleet, but the effects of misfueling (for HC and CO)
are assumed to be overshadowed by the effects due  to experienc-
ing a Primary category failure.  The rich operation assumed for
Primary category vehicles, results in essentially  zero  catalyst
efficiency for HC and CO by itself due to lack  of oxygen in the
catalyst bed(s).

-------
                              21
     The resulting Zero-mile levels for the MisfuRji.no oateoory are:

          HC = 1.05 a/ml
          CO =4.24 g/mi
     ii.  Deterioration rates.  The HC/rn deterioration rates for thp
     Misfueling category were ohtained by adoptino the deterioration
     rates developed for the Secondary catenory.  There is insuffi-
     cient data on misfueled Closed Loop vehicles to allow a senarate
     analysis of the deterioration nf these vehicles.   The develnn-
     ment of the deterioration rates for the .Secondary catennrv can
     be found in Section TV.A.2.h.

          HC = 0.1? o/mi/10.nnn miles
          CO = 0.71 o/mi/10,nnn miles

3.  Category Size.

The size of a category is described hy two narameters:   the size at
zero miles, expressed in percent nf the unit of analysis,  and the
growth rate of the category as expressed hv the nercent of thp ori-
ginal unit which "migrates" into the cateoorv oer in ,000 miles.

     a.   Primary.  The size parameters nf the Primary catennrv were
     estimated hy considering a number of separate sources of infnr-
     nation.  First, the incidence of vehicles from the in-nsp vehi-
     cle data base which would fall 5nto the Primary catpnorv was
     examined.  A total of ten Primary cateoorv vehicles was found in
     the fleet.  The total fleet size is ?n^ vehicles after taHna
     Parameter Adjustment concerns into account.*  I is i no these f in-
     ures results in an incidence of 5.n percent fnr the Primary
     cateoory .  This is at an averaoe mileane of °,l£? miles.
     The second piece of evidence used to determine  the  size  nara-
     meters of the Primary cateoory came from InoWno  at the  f.lppt nf
     162 Volvos and Saabs (described in Section TT.n.l^.  These
     * Note:   There are a total, of ?nl  cars used in  the Mr
     and 205 cars used for P.n since some cars were removed due to
     Parameter Adjustment concerns for  one nollutant hut not  for the
     other, as discussed previously (Spotion TT.O.I. 9n^ Was uspd a<
     the average.

-------
                        22
vehicles employ  the Lambda  Sond system:  a  closed loop, system
with ported fuel injection and a Three-Way  catalyst.   One of the
distinguishing characteristics of this fleet  is that  it  contains
a wider range  of vehicle odometer  mileages than the  Closed Loop
vehicle data base.  This  fact was  used to examine  the growth of
Primary category problems  with  increasing  mileage.   The  fleet
was divided into two groups  based  on mileage intervals,  and the
incidence  of   vehicles  with some  malperformance  of  the  Lambda
Sond  system was noted  (eg.  oxygen  sensor  failure, defective
electronic  control  unit).   The  first  group contained  vehicles
with between  0 and 10,000  accumulated miles.   This  group  con-
sisted  of  64  vehicles,  three of  which  had  a  Lambda Sond  mal-
performance indicated.    This  translates  to  an  incidence of 4.7
percent.   Those  vehicles with a Lambda  Sond malperformance had
average CO emission levels  over  400 percent  higher  than  the
group's overall  average.  The average mileage of the  first  group
was approximately 5,800  miles.   The  second group was made  up of
vehicles with  between 10,000 and 20,000 accumulated miles.   This
group consisted  of 65 vehicles,  five  of  which had  a  Lambda  Sond
malperformance.   This   translates  to   an  incidence   of   7.7
percent.   Those  vehicles with a Lambda  Sond malperformance had
average CO emission levels 390  percent  higher  than  the  group's
overall average.   The  average mileage of  the  second group was
approximately 14,400 miles.

In comparing these  two  groups  of vehicles, the chief conclusion
to be  drawn is  the trend towards  an increase  in  malperformance
of the  closed  loop  control  system  with increasing  mileage.   The
two groups examined are of essentially the same size, the  only
difference being between  their average mileages.   The difference
in  the incidence of  malperformance  for  the two  groups  is 3.0
percent.

TO arrive  at  the size  parameters  for  the Primary  category, the
preceding  findings were combined and supplemented  with technical
judgment.    In  applying  the  data from  the in-use  data   base,  a
figure of  5 percent was taken  as the incidence of  Primary  cate-
gory  vehicles  at 10,000 miles.   The  3%  difference between the
two mileage groups of Volvos and Saabs is  taken as  an indication
that  there will  be a  measurable  increase  in  Primary  category
failures with  increasing mileage.   Due to  the  small  sample  size
of  vehicles,  however,   and  due  to  differences  between  the
Volvo/Saab system and the variety  of systems to  be seen in the
post-1980  fleet,  this   figure does not necessarily indicate the
definitive  growth rate  for  1981 and  later vehicles.   This  ana-
lysis  assumes that the  growth rate will  be  somewhat  lower:
2%/10,000  miles.   This  is based on  the  assumption that due  to

-------
                         23
improved  assembly  line testing,  the  presence of  on-board diag-
nostics,  and  other  reasons,  the performance  in  the field of the
majority  of  the fleet will  be better  than  what was  seen among
the in-use Volvos and Saabs.

Combining  this  growth rate  with the  incidence  of 5%  at 10,000
miles resulted in the following size parameters:

     Initial Size = 3%.
     Growth Rate = 2%/10,000 miles.

b.  Secondary.   The size  parameters  for the  Secondary category
were  determined by  simply  taking  what  was left  of  the fleet
after  establishing  the sizes  of  the  Primary  category  and  the
Misfueling category (see below).

The size parameters for the Secondary category are:

     Initial Size = 89.24%.
     Growth Rate = -1.84%/10,000 miles

c.  Misfueling.   The  size of  the Misfueling  category  was deter-
mined by  adopting the rate observed in an EPA covert observation
study  performed  by  the Mobile  Sources Enforcement  Division.[3]
This  study  observed  the  fueling  practices   of   over  22,000
catalyst-equipped  vehicles  in 36  states  and  has the  largest
sample size of  any  study   of its  kind.  It observed  vehicles  of
various  model  years  and  manufacturers.    While   arguments  are
often  put forward  as  to trends  in  the misfueling  rate  with
regard to vehicle  age,  model  year, engine technology  etc.,  none
of  those  trends have  been  substantiated or  quantified  to  the
extent necessary to be used  in an  analysis of this  sort.  There-
fore the  observed rate of  8% was adopted for this  analysis as  a
best estimate.   The category  is not  assumed to grow  with time.
As was mentioned earlier,  however, vehicles which have  a Primary
category  type of failure  and which are  misfueled are  assumed  to
be  best  represented  by the  Primary  category  emission  levels.
Thus,   the overlap  of  the Primary  and Misfueling  categories  is
represented by  the Primary  category.    As  the.  Primary  category
grows with time  therefore, the percentage of cars  represented  by
the  separate  Misfueling   category  emission  levels  declines.
This growth of  the  overlap between the  two  categories,  although
small, has been  accounted  for  and  is  the source of the negative
growth  rate  below.   Thus,  the Misfueling  category  size  para-
meters are:

     Initial Size = 7.76%
     Growth Rate = -0.16%/10,000 miles

-------
                                       24
  0
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
         4.    Unit-of-Analysis  Composite.

         The table below presents the composite emissions of Closed  Loop  vehi-
         cles designed  to meet  standards  of 3.4  g/mi  CO  and 0.41 g/mi  HC.
         These composites  were arrived at  by weighting  together the  various
         categories as was described in Section III.B.5.  The following tables
         present  the composites and illustrate how the various categories were
         weighted together.  The  HC and CO  values are  presented in  g/mi  at
         10,000 mile intervals.
3.
3.
3.
4,
4.
 .74
 .86
 .98
 .10
 .22
4.34
4.46
4.58
4.70
4.82
4.94
0.03
0.05
0.07
0.09
0.11
0.13
0.15
0.17
0.19
0.21
0.23
                                      Table  IV-2
Miles   E**(Pri.)  S*(Pri.)  E**(Sec.)
HC
KSec.)
0.23
0.35
0.47
0.59
0.71
0.83
0.95
1.07
1.19
1.31.
1.43
Emissions
S*(Sec.)
0.89
0.87
0.86
0.84
0.82
0.80
0.78
0.76
0.75
0.73
0.71

E**(Misfuel.)
1.05
1.17
1.29
1.41
1.53
1.65
1.77
1.89
2.01
2.13
2.25

S*(Misfuel.)
0.08
0.08
0.07
0.07
0.07
0.07
0.07
0.07
0.06
0.06 '
0.06
                                                                      2.29
 * S = Size
** E•= Emission Level
                                     Table IV-3
                                    CO Emissions
Miles   E(Pri.)   S(Pri.)    E(Sec.)    S(Sec.)   E(Misfuel.)   5(Misfuel.)   Comp
0
1
2
3
4
5
6
7
8
9
10
107.36
108.0-7
108.77
109.48
110.18
110.89
111.60
112.30
113.01
113.71
114.42
0.03
0.05
0.07
0.09
0.11
0.13
0.15
0.17
0.19
0.21
0.23
1.86
2.57
3.27
3.98
4.68
5.39
6.10
6.80
7.51
8.21
8.92
0.89
0.87
0.86
0.84
0.82
0.80
0.78
0.76
0.75
0.73
0.71
4.24
4.95
5.65
6.36
7.06
7.77
8.48
9.18
9.89
10.59
11.30
0.08
0.08
0.07
0.07
0.07
0.07
0.07
0.07
0.06
0.06
0.06
5.21
8.02
10.83
13.65
16.46
19.27
22.08
24.90
27.71
30.52
33.33

-------
                                   25
     These composites yield the following regression equations:

     HC = 0.40 + (0.19)(m)
     00^5.21 + (2.81)(m)

     (m = miles/10,000)

     These  regression equations  were  not  used  in  determining  the final
     fleet composite  emissions, although very  little  difference results if
     they are.   They  are presented for those  interested  as an aid in com-
     paring the emissions performance of various units of analysis.

B.  Closed Loop vehicles designed to meet 0.41 g/mi HC and
7.0 g/mi CO standards.

The reader  will note  that  the definitions,  rates and levels  defined  for
this unit  of analysis borrow  extensively  from the  previous unit  of ana-
lysis.  This  is due  to  the  fact  that the only difference  between the  two
units of analysis is  the CO design standard.  This difference  will effect
only  certain  CO emission  levels.  While  the  emission control technology
used under these two  standards will differ (vehicles designed  to meet  the
7.0 g/mi  CO  standard will  not use a  back-up  oxidation  catalyst  and  air
pump in many cases),  this difference is not expected  to cause much differ-
ence in in-use emissions performance as discussed in Section I.C.I.

     1.  Category definitions.

     The categories  for  this unit  of  analysis are defined  to  be the same
     as those defined in Section IV.A.I.

     2.  Emission levels of the categories.

          a.  Primary Category.

          i.   Zero-mile levels.    The   zero-mile  levels  are the same  as
          those determined for the  Primary category of the  unit represent-
          ing Closed  Loop  vehicles designed to meet  0.41 g/mi HC  and  3.4
          g/mi CO standards  (Section IV.A.2.a):

               HC = 3.74 g/mi
               CO = 107.36 g/mi

          (Note: . The CO emission  level is not ratioed to  the  7.0  g/mi  CO
          standard  since the  Primary  category represents  vehicles with  a
          major  failure  mode  whose emissions  would  be independent  of  the
          design standard.   Similarly,  the HC emission level  is not  ratio-
          ed  to  take the  different standards resulting  from  the  Methane
          correction factor  into account, as  discussed in  Section  II.A.)

-------
                         26
ii.  Deterioration rates.  The HC/CO deterioration  rates are the
same as  those determined  for Closed  Loop vehicles  designed to
meet 0.41  g/mi  HC  and 3.4  g/mi  CO standards  (Section  IV.A.
2.b.).   This was done to follow the  principle that  similar tech-
nologies deteriorate at the same rate.   This  principle was wide-
ly  used  in the  1978 analysis (EPA-400/9-78-  005)  and  has been
widely  used  in  the  1980  revision  of  that   document  for  both
pre-1981   and   post-1980  vehicles.    The  deterioration  rates
arrived at are:

     HC = 0.12 g/mi/10,000  miles
     CO = 0.71 g/mi/10,000  mile?

b.  Secondary Category.

i.  Zero-mile levels.   The average HC/CO emission levels and the
average mileage of the  vehicles in the data base  which represent
the Secondary category are:

     Average HC = 0.32 g/mi        (n = 191)*
     Average CO = 5.47 g/mi        (n = 195)*
     Average Miles = 8,064  miles    (n = 191)

Applying the  deterioration rates  (see  below)  to  these average
emission levels to project them to zero miles and  then applying
a ratio of 7.0/9.0  to the  CO emission level  to reflect the fact
that the vehicles in the data base were designed to meet  a  9.0
g/mi CO  standard and  not  a  7.0  g/mi standard,  results  in  the
following Zero mile  emission  levels:

     HC = 0.23 g/mi
     CO = 3.83 g/mi

ii.  Deterioration rates.  The HC/CO deterioration  rates are  the
same as  those determined for  Closed Loop vehicles designed  to
meet  0.41  g/mi  HC  and  3.4  g/mi   CO  standards.     (Section
IV.A.2.b.).  This follows  the principle that similar  technolo-
gies deteriorate at  the same  rate.

     HC = 0.12 g/mi/10,000  miles
     CO = 0.71 g/mi/10,000  miles
*Thesesamplesizes  differ  due  to  the  fact  that  Parameter
Adjustment concerns  eliminated  vehicles  from  the  calculations
for one pollutant, but not for  the other.   See Section II.A.

-------
                              27
     c.  Misfueling Category.

     i.   Zero-mile levels.   The  Zero-mile  levels  of  the Misfueling
     category  for this  unit of  analysis  (Closed  Loop  vehicles  de-
     signed  to  meet standards of  0.41  g/mi HC and 7.0  g/mi  CO) were
     developed in  the  same  way as  for  the previous  unit of  analysis
     (Section  IV.A.2.C.).   The  figures  for  percent  increase  due  to
     misfueling were  simply added  onto  the  Zero-mile  levels  of  the
     Secondary  category.   The  Zero-mile  levels  arrived at   for  the
     Misfueling category are:

          HC = 1.05 g/mi
          CO = 8.73 g/mT

     ii.  Deterioration rates.  The  HC/CO deterioration  rates for  the
     Misfueling category are the same  as those developed  for Closed
     Loop vehicles designed  to meet  standards  of  0.41 g/mi HC and  3.4
     g/mi CO (Section  J.V.A.2.D.).   This  follows  the principle  that
     similar technologies deteriorate  at the  same rate.   The deteri-
     oration rates arrived at are:
          HC = 0.12 g/mi/10,000 miles
          CO = 0.71 g/mi/10,000 miles'

3.  Category Size.

The category  size parameters for  this  unit of analysis are  the  same
as  those  developed  for  the  previous  unit  of  analysis  (Section
IV.A.3.).

     a.  Primary Category.

     i. Initial Size = 3%.

     ii.  Growth Rate = 2%/10,000 miles.

     b.  Secondary Category.

     i.  Initial Size = 89.24%.

     ii.  Growth Rate = -1.84%/10,000 miles.

     c.   Misfueling Category

     i.  Initial Size = 7.76%.

     ii. Growth Rate = -0.16%/10,000 miles.

-------
                                       28
  0
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
  0
  1
  2
  3
  4
  5
  6
  7
  8
  9
  10
         4.  Unit-of-Analysis Composite.

         Tables IV-4 and  IV-5  below present the composite emissions' of Closed
         Loop vehicles designed to meet standards of 0.41 g/mi  HC  and 7.0 g/mi
         CO.   The  HC and CO values  are  presented in  units  of g/mi  at 10,000
         mile intervals.
Miles   ECPri.
3.
3.
3.
4.
4,
 .74
 .86
 .98
 .10
 .22
4.34
4.46
4.58
4.70
4.82
4.94
107.36
108.07
108.77
109.48
110.18
110.89
111
112
113
    60
    30
    01
113.71
114.42
SCPri.)

 0.03
 0.05
 0.07
 0.09
 0.11
            .13
            .15
            .17
           0.19
           0.21
           0.23
Miles   ECPri.)   S(Pri.)
 0.03
 0.05
 0.07
 0.09
 0.11
 0.13
 0.15
 0.17
 0.19
 0.21
 0.23
                                     Table IV-4
HC Emissions
ECSec.)
0.23
0.35
0.47
0.59
0.71
0.83
0.95
1.07
1.19
1.31
1.43

SCSec.)
0.89
0.87
0.86
0.84
0.82
0.80
0.78
0.76
0.75
0.73
0.71
Table IV-5
ECMisfuel.)
1.05
1.17
1.29
1.41
1.53
1.65
1.77
1.89
2.01
2.13
2.25

CO Emissions
ECSec.)
3.83
4.54
5.24
5.95
6.65
7.36
8.07
8.77
9.48
10.18
10.89
SCSec.)
0.89
0.87
0.86
0.84
0.82
0.80
0.78
0.76
0.75
0.73
0.71
ECMisfuel.)
8.73
9.44
10.14
10.85
11.55
12.26
12.97
13.67
14.38
15.08
15.79
                                                      SCMisfuel.)   Comp.
0.08
0.08
0.07
0.07
0.07
0.07
0.07
0.07
0.06
0.06
0.06
0.40
0.59
0.77
0.96
1.15
1.34
1.53
1.72
1.91
2.10
2.29
                                                      SCMisfuel.)   Comp.
0.08
0.08
0.07
0.07
0.07
0.07
0.07
0.07
0.06
0.06
0.06
              These composites  yield  the  following regression equations:

                   HC = 0.40  +  CO..
7.32
10.09
12.85
15.62
18.39
21.16
23.93
26.70
29.47
32.24
35.00
                   CO = 7.32  +  (2.77)(m7
                   Cm = miles/10,000)

-------
                                   29
C.  Oxidation Catalyst vehicles designed to meet
standards of 0.41 g/mi HC and 3.A g/mi CO.

     1.  Discussion of the analysis.

     Since the  units  of analysis which represent  Oxidation Catalyst vehi-
     cles are not broken down into  categories,  the approach used in deter-
     mining  their  emission  rates  will be  presented  in  a  much different
     format.   As was  discussed in  Section  I.D.,   Oxidation  Catalyst vehi-
     cles in the post-1980 timeframe  are  assumed  to have  an emissions per-
     formance similar  to current  Oxidation  Catalyst  vehicles.   Oxidation
     Catalyst vehicles  after 1980  are  not  assumed  to experience signifi-
     cant changes in basic emission control technology.  They  are  assumed
     to rely on EGR and  engine  modifications  for  the control of NOx and to
     be equipped with Oxidation catalysts and air  pumps for the control of
     HC and  CO.   This assumption  largely  determined what  data  could best
     be used to predict  the  emissions for these vehicles.   In essence,  the
     emission rates  developed  for  1980 Oxidation Catalyst  vehicles  from
     the  1980  revision  of  the Mobile  Source  Emission  Factors  Document
     (EPA-400/  9-78-005) were  adopted after  two  important factors  were
     taken into account:   the effect of  the  Parameter Adjustment  regula-
     tions and the  difference in CO standards.

     First,  the  effect  of  the   Parameter  Adjustment  regulations  was
     accounted for.  This was  done in the following  three  step procedure.
     First, the data  base assembled  to analyze  the  emissions  performance
     of current  technology  vehicles  (1975-80)  was  stratified  to  look  at
     only those vehicles equipped  with  Oxidation  catalysts and  air  pumps,
     i.e., those vehicles  equipped with the control  technology  most simi-
     lar to that which  is  assumed to be  used  in   the  post-1980  timeframe.
     Most of these vehicles  were  also equipped with  EGR, but the presence
     of EGR  was not  used as  a  stratifying criteria since  some  unique
     engine configurations are  assumed  to meet the 1.0 g/mi NOx standard
     without  EGR.   From this  subset of the data base,  the average emission
     levels  for  HC,  CO  and NOx  were  determined.   In the second  step,
     vehicles were eliminated  from the subset of  the data  base  described
     above if they showed evidence  of maladjustments  which  can  be expected
     to be prevented by  the  Parameter Adjustment regulations.   The  average
     emission levels  for HC,  CO  and  NOx were then calculated  for  these
     vehicles.   Comparing the average emissions before and  after removing
     those vehicles affected by the  Parameter  Adjustment  regulations  led

-------
                              30
to the calculation of  a "percent benefit"  figure  due to  the  regula-
tions,   (The  "percent  benefit"  approach was  taken rather  than  per-
forming  separate  regressions  on the  subsets of  vehicles  described
above, due  to  those subsets  having relatively  small sample  sizes.
For example, there were only 85 vehicles in  the  subset used  to  pre-
dict  the percent benefit  due to  the regulations  in the  1982  model
year).  This procedure was performed in  a stepwise  fashion to  account
for the  fact that the Parameter  Adjustment  regulations are applied to
idle mixture and  choke in 1981  and  then applied to  timing in  1982.
Thus, there is  a  separate  "percent benefit" for both  of  those years.
The  1981  "percent   benefit"  reflects   the  impact  of  limiting  the
adjustability of idle mixture  and choke  while  the 1982 "percent  bene-
fit" reflects  the  impact of limiting  the  adjustability of timing.
Thus, the total impact of the regulations  is staggered between  1981
and  1982.   The following  table presents  the calculated   figures  of
"percent benefit" for HC and CO.
                          Table IV-6
     Percent  Benefit  from Parameter  Adjustment Regulations

                           1981                  1982*

    HC                     16.2%                 16.7%

    CO                     25.0%                 27.3%
Finally,  these  figures of  percent  benefit were  applied  to the  zero
mile levels and deterioration rates determined  for  the  1980 low alti-
tude, non-CaTTTbrnia Light  Duty Vehicle  (LDV) fleet.  Table IV-7  pre-
sents the  zero  mile  emission levels and deterioration  rates  for  that
fleet before applying any benefits or  modifications.
*Cummulative benefit.

-------
                              31
                          Table IV-7
           1980 Light Duty~Vehicle "Emission Factors

                          Zero Mile       Deterioration Rate

      HC                     0.29               0.29

      CJD                     6.14               2.86

(Note:   The  zero mile  levels  and deterioration  rates from  the 1980
LDV fleet presented above incorporate  the  effect  of misfueling at the
same  misfueling  rate projected  for  the Closed  Loop vehicles  of the
post-1980 fleet).

Account must  also  be  taken  of the difference  in  design standards for
CO between 1980  and  1981.   The 1980 Federal  CO standard  is  7.0 g/mi
while it  is  3.4 g/mi  for  this unit of  analysis.   (The difference in
NOx standards will be  discussed  in a similar  way  in Section V.B.I.)
To account  for  this difference  in standards,  the  zero-mile  CO emis-
sion level for the 1980 model  year,  after  being modified  as discussed
above to  account for the effect  of the Parameter  Adjustment regula-
tions, was  ratioed by  a  factor of  3.4/7.0.   The  deterioration rate
was not  factored to  account for the difference in standards in order
to follow the principle that similar technologies  deteriorate  at the
same  rate.    This  principle  was  widely used  in  the  1978  analysis
(EPA-400/9-78-005) and  has  been  widely used in the 1980  revision for
both  pre-1981  and post!980  vehicles.   The  deterioration rate  was
modified to account  for the effect of the  Parameter Adjustment regu-
lations  since  those  regulations  represent  a change  in  technology
which will affect the rate of in-use deterioration.

Table IV-8 presents the final  HC and CO  zero mile emission levels and
deterioration rates  determined  for  Oxidation  Catalyst vehicles  for
1981 and 1982-and-beyond (1982+).

-------
                                       33

    D. Oxidation Catalyst vehicles designed to meet standards
    of 0.41 g/mi HC and 7.0 g/mi CO.

         1. Discussion of the analysis.

         This unit  of analysis represents Oxidation  Catalyst  vehicles granted
         a  CO  waiver  to  7.0  g/mi  under  Section  202(b)(5) of  the Clean  Air
         Act.   The  different CO standard  is  the only difference  between  this
         unit of analysis  and the unit  of analysis described in  the  previous
         section.   This difference  in CO  standards is not expected to  result
         in a significant  difference  in  the  basic nature of the  emission  con-
         trol technology  to be used.   Vehicles which  receive 'a CO waiver  to
         7.0 g/mi are expected to simply  use less stringent  air/fuel control
         and less efficient catalysts than vehicles  designed  to meet the  3.4
         g/mi standard.   The analysis for these  vehicles  is  therefore  essen-
         tially the same  as that  performed for the previous section.   The  one
         exception  being  that the 1980  Light Duty Vehicle CO emission  level
         does not need to be ratioed to account  for  a difference in standards
         since there is no difference.

         The resultant zero mile emission levels and deterioration rates  for
         this unit of analysis are presented  in Table  IV-10.

                                  Table  IV-10
       Oxidation Catalyst  Vehicle  Emission Factors (7.0 g/mi CO Standard)

                      1981                             1982+

           Zero Mile       Deterioration  Rate    Zero Mile    Deterioration Rate

HC            0.24                 0.24            0.24             0.24

CO            4.61                 2.15            4.47             2.08


         2.   unit-of-Analysis Composite.

         This section  presents  the  composite  emissions  of Oxidation  Catalyst
         vehicles designed  to meet standards  of 0.41  g/mi HC  and 7.0  g/mi CO.
         The HC and CO values are presented  in units of  g/mi at 10,000 mile
         intervals.   Table IV-11 is simply a tabulation  of  the equations shown
         in Table IV-10.

-------
                                       34
                                  Table IV-11
     Oxidation Catalyst Vehicle Emission Composites  (7.0  g/mi  CO  Standard)

                          1981                    1982+
Miles

  0
  1
  2
  3
  4
  5
  6
  7
  8
  9
  10
HC

0.24
0.48
0.72
0.96
 .20
 .44
 .68
 .92
 .16
 .40
1.
1.
1,
1.
2.
2.
2.64
CO

4.61
6.76
8.91
11.06
13.21
15.36
17.51
19.66
21.81
23.96
26.11
HC

0.24
0.48
0.72
0.96
 .20
 .44
 .68
 .92
1.
1,
1,
1.
2.16
2.40
2.64
C0_

4.47
6.55
8.63
10.71
12.79
14.87
16.95
19.03
21.11
23.19
25.27
References for Section IV.

1.   EPA  Contract No.  68-03-2783,  Work Effort  No. 2  with the  California Air
Resources Board.

2.   EPA  Contract No.  68-03-2693,  Work  Effort No.  4  with Automotive  Testing
Laboratories, Inc.

3.   EPA  Internal  Memorandum;  August  2,  1979; from Benjamin R.  Jackson,  Deputy
Assistant Administrator,  Mobile  Source and Noise  Enforcement, to  all  Regional
Administrators.

-------
                                       35
V.  SPECIFIC UNIT ANALYSES - NOx

    A. Closed Loop vehicles designed to meet
    a 1.0 g/mi NOx standard.

         1.  Category definitions.

         There are four categories of  vehicles  in  this unit of analysis.  Each
         category will be presented and briefly characterized.

              a.   EGR.   This  category  is  comprised  of  those vehicles  with
              inoperative EGR systems.  This includes  cases where  the  EGR  sys-
              tem is inoperative due to tampering,  in-use  failure,  or  manufac-
              turing defect.  It also  includes  cases of EGR  failure where  EGR
              operation falls under  computer control  and  is  therefore subject
              to computer malfunction.   Vehicles were  selected from  the  data
              base to  represent this  category  on  the basis  of a  diagnostic
              assessment of EGR function and measured  NOx  emissions.   Vehicles
              diagnosed  as  having  an  EGR  problem and   whose  NOx emissions
              exceeded their design standard were selected.

              b.  Misfueling.   This category  is comprised  of vehicles which
              have been misfueled, i.e. they have had  extensive damage done to
              their emission control  system through the  use of leaded rather
              than unleaded gas.   As  will  be  discussed  below,  misfueling  is
              assumed to overlap with  each  of the other three  NOx categories.

              c.  Low.   This category  is made  up of vehicles which would  fall
              into the  Primary  HC/CO  category  and which are  therefore assumed
              to have lower than normal  NOx levels.  This assumption  is based
              on the  fact  that rich  operation   (as  found  in Primary  category
              vehicles)  inherently implies  lower NOx production in the engine,
              as well as the  fact that rich  operation improves the catalytic
              conversion of  NOx  in  a Three-Way  catalyst.

              d.  Secondary.  As in the HC/CO analysis, this  category  includes
              the remainder  of  the  fleet.

         2.   Emission levels of the  categories.

         Before  discussing  how the  emission  levels  of  the  categories were
         determined,  several  preliminary   points  need  to  be made.    First,
         unlike for HC  and  CO,  the vehicles in the Closed  Loop  Vehicle data
         base were designed  to  meet several different  NOx  standards.   The 1980
         General Motors  X-Body  vehicles were designed  to meet  a  1.0  g/mi  NOx
         standard.  The  other vehicles tested  in Los Angeles were 1979  models
         (the Ford 351's, Celica  Supra's  and Audi  5000's) and were therefore
         designed to meet a 1.5 g/mi  standard.   A third group of vehicles  are

-------
                              36
the Ford 351's tested in Portland, Oregon.   As was  discussed in Sec-
tion U.S., these cars  are  assumed to have been designed  to meet the
Federal 2.0 g/mi NOx standard.  Since this analysis  is  concerned with
predicting NOx emissions under a 1.0 g/mi standard,  the NOx emissions
from all vehicles designed  under  the 1.5 g/mi  or  2.0  g/mi standards
were divided by 1.5 or 2.0  respectively before being used  in the ana-
lysis.  This was  done  due  to  the  fact  that NOx  emission  levels  are
determined  by  the  interaction of timing,  air/fuel ratio,  catalyst
design, and rate of  EGR,  all of which are balanced  in  order  to meet
the NOx design  standard.   A vehicle is  therefore  assumed  to perform
in direct relation to its design  standard.   The reader may note that
this same  principle  was used  in ratioing  the  CO emission  levels for
Secondary category vehicles to be applicable to  the analysis  of the
various CO standards.

The second preliminary point that  needs to be made has  to do specifi-
cally with the Ford 351's from Los Angeles.  During  the course of the
analysis it became  clear  that these vehicles  had  a  unique NOx prob-
lem.  Of the 73  Los  Angeles 351's not in the  Low  category,  17 vehi-
cles were  found  to have an EGR  problem with NOx emissions above the
standard and an additional  20  vehicles had NOx emissions  significant-
ly  above  standard  without  indication of an  EGR problem.   It appears
that the 351 system, which  controls the EGR flow  rate  by  command of
the on-board computer, had  an  operating problem resulting  from either
a  design  or manufacturing  flaw.    Since  these vehicles represent  a
substantial fraction of the data base and since the  size of the prob-
lem observed with the 351's was judged to be atypical of the 1981 and
later fleet, it was decided to reweight the various  categories within
the 351  fleet for  data analysis  purposes.   The  internal  weighting
from  the  General  Motors  X-Body  fleet  (roughly  the same size)  was
therefore applied to the  Ford  351 fleet to be able  to  use  this data
in  a  representative  way.   The  internal  weighting  of  the X-Body fleet
(without Low category  cars) was:   four  vehicles in the EGR category,
five  cars  without  an EGR problem  but  with  NOx emissions  above stan-
dard  and  70  cars with  NOx emissions  below standard.  The  average
emission levels  from the Ford 351's grouped by  these same  internal
divisions were reweighted to match the weighting of the  X-Body fleet.

The emission  characteristics of the various  categories will  next be
presented.

     a.  EGR Category.

     i.   Zero-mile  level.   The  average NOx  emission  level and  the
     average mileage of the vehicles  in the fleet selected to repre-
     sent the EGR category are:

-------
                          37
     Average NOx =1.71 g/mi        (n  = 9)*
     Average Miles = 9,510 miles    (n  = 9)

The  deterioration rate  (see  below)  was  then  applied  to work
backwards to arrive at a Zero-mile  level of:

     NOx =1.58 g/mi

ii.   Deterioration rate.   The deterioration  rate  for  the  EGR
category was obtained  by adopting  the deterioration rate devel-
oped for the  NOx Secondary category  (see  below).   This was done
due to a lack  of data describing the  deterioration  of EGR cate-
gory vehicles.

     NOx =0.13 g/mi/10,000 miles

b.  Misfueling Category.

i.   Zero-mile level.    As  was  mentioned  above,  the  Misfueling
categoryoverlapswith  the  other  three  NOx  categories.   For
example, some vehicles are assumed  to have an EGR problem and to
have been misfueled.  As could  be  expected,  each  of the possible
overlap  situations  is  assumed  to  have  a unique emissions per-
formance.  This might appear  to be  a different  approach than was
used for the HC/CO  analysis.   For  HC/CO the  overlap of the'Mis-
fueling category and the  Primary  category was handled  by assum-.
ing  the  emissions of  the overlap  vehicles were represented  by
the emissions of  the  Primary  category, rather  than  establishing
a  new,  distinct category.  In  that case  however, Primary cate-
gory vehicles, due to  the rich operation  inherent to  such vehi-
cles, are already experiencing  essentially zero catalytic activ-
ity  for  HC  and CO.  This  is  due  to the  lack  of oxygen  in  the
catalyst bed(s).  Thus,  misfueling, which primarily damages  the
catalyst, will have  a negligible  incremental  effect on  Primary
HC and  CO  emission levels.   NOx,   however,  is  a very  different
case.  None  of  the  operating  modes  represented by  the  other
three NOx categories are  assumed  to have  a comparable  effect  on
NOx  conversion  in  the  catalyst.   Misfueling  will  therefore
incrementally damage  the  NOx control  capability  of vehicles  in
each of  the  other categories.  In  essence then,  the  Misfueling
category is made up of three sub-categories:  Misfueled/EGR,
*  In  accordance with  the  explanation above regarding  EGR  fail-
ures among Ford 351  vehicles  in Los Angeles, this sample of "9"
is in  fact  composed of 4  X-Body vehicles,  17 Ford  351 vehicles
from Los Angeles collectively having the weight of  only 4  vehi-
cles and 1 Ford 351 from Portland.

-------
                         38
Misfueled/Low,  Misfueled/Secondary.   The  emission  characteris-
tics of  each sub-category were  developed by following  a common
procedure.    That  procedure will  first be  explained below  and
then applied to each sub-category.

The procedure for calculating the damage  resulting  from  misfuel-
ing parallels  the  procedure adopted  for  HC/CO.  The  same fleet
of six  Closed Loop  vehicles  described in  Section  IV.A.2.C  was
examined  to  quantify the  percent increase  in NOx  emissions  due
to misfueling.   The  average percent  increase observed  was  197%
(n = 6).

This percent  increase  was then  applied across  the  board  to  the
Zero-mile NOx  levels of the EGR, Low and Secondary categories.
Table V-l presents  the  resultant  Zero-mile  levels for  each  of
the misfueling sub-categories.
                     Table V-l
  Zero-mile  Levels of the Misfueling Sub-categories

Sub-Category             NOx Zero-mile level

Misfueled/EGR            4.69 g/mi
Misfueled/Low            0.80 g/mi
Misfueled/Secondary      1.72 g/mi

ii.   Deterioration rate.   The deterioration  rate used  for  each
of  the  Misfueling sub-categories  was obtained  by adopting  the
deterioration  rate  determined  for  the   NOx  Secondary  category
(see  below).   This  implies  that vehicles  with  badly  damaged
catalysts will deteriorate  at  the  same rate  as vehicles  with
relatively  undamaged  catalysts.    While  this  assumption is  not
perfectly  accurate,  the  deterioration  rate  for  the  Secondary
category was  adopted  due to  a  lack of  any  data describing  the
deterioration  of  misfueled  vehicles  and  because  the  stated
deterioration rate includes the deterioration due to  engine  wear
and other phenomenon rather than  just  due to catalyst deteriora-
tion.  The deterioration rate arrived at  is:

     NOx = 0.13 g/mi/10,000 miles

c.  Low Category.

i.   Zero-mile level.    The  average  NOx  level  and  the  average
mileage of  the vehicles in the  data  base selected to  represent
the Low category are:

-------
                          39
     Average NOx = 0.40 g/mi      (n = 9)
     Average Miles = 9,539 miles  (n = 9)

 (Note: One vehicle  from  the  GM X-Body fleet which had high HC/CO
 emission  levels  was not  used to calculate the  Low category NOx
 emission  level due  to  its having an inoperative EGR system.  Its
 emission  levels  were  therefore not representative of vehicles in
 the  Low  category.   Vehicles with this combination  of high HC/CO
 emissions  yet  with an inoperative EGR system  will be discussed
 in the next section on category sizes).

 The  deterioration  rate (see  below)  was  then used  to work back-
 wards and arrive at a zero mile level of:

     NOx = 0.27 g/mi

 ii.   Deterioration Rate.   The deterioration  rate  for  the  low
 category was obtained  by  adopting the  rate developed for the NOx
 Secondary category  (see  below).   This was done  due to a lack of
 data describing the deterioration of  Low  category vehicles.  The
 deterioration rate arrived at is:

     NOx =0.13 g/mi/10,000 miles

 d.  Secondary Category.

 i.   Zero-mile  level.    The   average  NOx  level  and the  average
mileage of  the  vehicles  in  the  data  base selected  to represent
 the OK category are:

     Average NOx = 0.68 g/mi      (n = 155)
     Average Miles = 7,814 miles  (n = 155)

 The  deterioration rate (see below) was  then  used to  work  back-
wards and arrive at a zero mile level of:

     NOx = 0.58 g/mi

ii.   Deterioration  Rate.   The  deterioration  rate   for  the  NOx
Secondary category was obtained  in essentially  the same way  as
was  done  for  the HC/CO  Secondary category (Section  IV.A.2.  b).
It will  therefore  not be discussed in  depth,  rather  the  perti-
nent numbers will  be  briefly  presented.   The NOx  d.f.  obtained
 from  the  in-use  1978-79 Volvo/Saab  fleet  was  2.56  and  the
1978-79 Volvo/ Saab  Certification fleet had an  average  weighted
d.f. of 1.27.  Thus  the  ratio of in-use to certification  d.f.-l
values is 5.78 (1.56/.27).   The  average Certification  d.f.-l  for
vehicles  being  certified  for  the 1981modelyear  was   0.16.

-------
                              40
     Applying  the  Volvo/  Saab  ratio to  this  figure and  adding  1
     results in a projected  1981 in-use  d.f. of  1.92.  Applying this
     d.f. to the data point  found in  the  data base (0.68  g/mi  NOx at
     7,814  miles)  and  converting  to a  deterioration rate gives  a
     deterioration  rate of:

          NOx = 0.13 g/mi/10,000 miles

3.  Category Size.

The sizes of  the various  categories are  described by  two  parameters:
the initial size,  expressed  in  percent,  and the  growth  rate  of  the
category  as expressed  in percent  of the  original unit  which  "mi-
grates" into the category per 10,000  miles.   These two parameters  are
developed for each  of the  four NOx categories.

     a. EGR.
     Several sources of information contributed  to  the  development of
     the size parameters of the EGR category.  Each will  be presented
     separately, followed  by a summary discussion.

     i.   1980 General Motor's X-Body  Fleet.

     The  first  source of  information examined  was  the  incidence  of
     EGR category vehicles  in the 1980  X-Body  fleet.  A total of  5
     vehicles was found in the EGR category out  of  a  total fleet size
     of 92 vehicles   This  results in an incidence of  5.4% at  an aver-
     age mileage of 5,500 miles.   It  should be  noted here that since
     we are  only looking  at the  incidence  of  a  particular  problem
     within the fleet, the effect of  the Parameter  Adjustment regula-
     tions need not be taken into account.   This  is especially  true
     for EGR  problems,  which will  not be affected  by  the  Parameter
     Adjustment regulations.

     ii.  1979 Ford  351 Fleet.

     The second  source of information examined  was the  incidence  of
     EGR category vehicles  in the  1979  Ford  351  fleet.    The  micro-
     processor used on this  engine  family  controls  the EGR  flow  rate
     through a  system of sensors  and solenoids.   As was  previously
     discussed, for an unusually large number  of the  vehicles tested,
     the EGR  system  was found to be  malperforming.   In  many  cases,
     there was either no voltage signal or an  inadequate  voltage  sig-
     nal coming  from  the  microprocessor unit  to the EGR-controlling
     solenoid.  20  vehicles  out  of  a total  fleet  size  of 97  were
     found to be in the EGR category.  (This  includes vehicles  which
     were removed from the  calculation of  NOx emission levels  due  to
     Parameter  Adjustment  concerns.)     This   gives  an  incidence  of
     20.6% at an average mileage  of approximately 11,000 miles.

-------
                               41
      iii. Surveillance Data on Current Technology Vehicles.

      The  final  source  of information  relevant  to  this  discussion
      deals with the magnitude of  EGR malperformance among the current
      in-use  fleet.   The information  comes from an  internal  EPA ana-
      lysis of  data from  the  EPA  Emission Factors Program  (EFP)  and
      the  1978  EPA  Tampering  Survey.[1]   The  analysis  examined  the
      initial incidence and growth  rate  of  EGR problems.   Data from
      vehicles from the 1973 through  1978  model  years were included in
      the analysis.   For  purposes of  this document however,  only  the
      1977-1978 model  year  vehicles will  be  considered.  (The earlier
      model years  show a significantly  higher  incidence  of  EGR mal-
      performance).  These  vehicles are equipped with  the most recent
      EGR systems.   It  is  assumed  that  these  EGR  systems  are  most
      representative  of  the  EGR   systems  to   be  employed  in  the
      post-1980 fleet.  The results of the study which  are  applicable
      to this analysis are presented below:
                            Table V-2
                       1977-78  Model Years
                                                  -^
                          Initial EGR
      Data Source         Failure Rate        Growth Rate

Emission Factors Program      2.77%          2.68%/lO,000 miles

1978 Tampering Study          5.15%          Q.92%/10,000 miles

      The size  parameters  of  the EGR  category  were obtained  through
      combining  the  preceding sources  of  information  with  technical
      judgement.  The  5.4 percent incidence  of EGR  category  vehicles
      from the  X-Body  fleet combined  with the  MSED Tampering  Survey
      figure  of  5.15  percent  were  combined to  establish a zero  mile
      incidence  of  5  percent  for  malfunctions in  the  EGR system  it-
      self.   An additional 2 percent was added onto this  base  level to
      account for the effect of computer malfunctions on  computer  con-
      trolled EGR  systems as  was  evidenced by  the high incidence  of
      this failure mode among  the Ford 351  fleet.    The growth  rate of
      the EGR  category was  obtained  by taking  the average of  growth
      rates  indicated by  the Emission  Factors  data and the MSED Tamp-
      ering  Survey data.  The  average of 1.8%/10,QOO miles  was  rounded
      up to  2%/  10,000 miles.  The  overlap of  the EGR  category  with
      the Misfueling category  was also taken into account as discussed
      in the  following section.   Thus,  the size parameters of  the  EGR
      category  are:

-------
                              42
            Initial  Size  =  6.44%
            Growth Rate = 1.84V10, OOQm lies

       b.   Misfueling.  As was  discussed in Section  V.2.b.  which  pre-
       sented emission  levels  for  the  NOx Misfueling  category,  this
       category  is  further  broken down into three sub-categories:   EGR/
       Misfueling,     Low/Misfueling,    Secondary/Misfueling.      These
       sub-categories  represent  the  overlap  of  Misfueling  with  the
       other categories.   Several helpful  points  can be  made  before
       presenting the  growth  rates of  the individual subcategories.

       First, this  analysis assumes that a  given portion of the  fleet,
       8%,  is misfueled at zero-miles and  remains misfueled throughout
       the  analysis.  (See  Section  IV.A.3.C.  for  a  more complete  dis-
       cussion  on why an 8% constant rate  was chosen).    The  analysis
       also assumes that this 8% is  spread proportionately throughout
       the  fleet.  That  is, the number of  cars which are misfueled and
       in  the EGR category  is directly proportional to the total  number
       of  EGR category vehicles in the  overall fleet, and likewise for
       the  Low and  Secondary  categories.  It follows therefore that as
       the  incidence  of  vehicles in  the various  NOx categories  (EGR,
       Low,  Secondary)   changes  with   increasing  mileage,   that   the
       sub-categories  which   represent   the  overlap  situations   also
       change in size in a parallel  fashion.   This concept can best be
       illustrated  by considering  the 8% of  the  fleet  which  makes  up
       the   Misfueling  category  as  a separate  fleet.   The Misfueling
       fleet experiences  the same internal  changes in its sub-category
       sizes as does the total fleet with its  categories, and  in  exact-
       ly  the same proportion.

       The  following table  presents the  size  parameters  of each of the
       Misfueling sub-categories which   were  calculated  in  the manner
       described above.
                            Table V-3
         Size Parameters of the Misfueling  Sub-categories

                      Initial  Size             Growth  Rate

Misfueled/EGR             0.56%             0.16%/10,000 miles
Misfueled/Low             0.16%             0.1256/10,000 miles
Misfueled/Secondary       7.28%             -0.28%/10,000  miles

-------
                              43
     c.  Low.

     The Low  category  represents  those vehicles with Primary category
     HC/CO  emission levels  and  a  correspondingly  low  NOx  emission
     level.   Thus,  the  size of  the  Low  category  is inherently  and
     directly related to the size of  the Primary HC/CO category.   One
     consideration  that  needs  to  be taken into account here, however,
     is the  overlapping  of the Primary HC/CO  and  the EGR  NOx  cate-
     gories.  That  is,  some vehicles  can be expected to  have  both a
     Primary  category  failure  mode  and an  inoperative  EGR  system,
     with  resulting high or medium  NOx levels.  This phenomenon  was
     observed in the X-Body  fleet and  is expected  to limit  the  size
     of the  Low  category.   This is especially  so since it can  logic-
     ally  be assumed  that  since tampering  contributes to both  the
     Primary  and EGR  categories, that  there  will  be  instances  in
     which both  the closed loop  fuel  system and the EGR  system  are
     tampered  concurrently.    Based   on  these  considerations   and
     accounting for the  overlap of the Low category with  the Misfuel-
     ing  category,   the  size   parameters of  the  Low  category  are
     assumed to be:

          Initial Size = 1.84%
          Growth Rate = 1.38%/10,000  miles

     d.  Secondary.

     As  with the   Secondary  category  for  the HC/CO  analysis,   the
     Secondary category  for  NOx  is  simply assumed  to represent  the
     remainder of the fleet.   Thus,  its size parameters are:

          Initial Size = 83.72%
          Growth Rate = -3.22%/10,000  miles

A.  Unit-of-Analysis Composite.

     a.  This section presents  the composite  emissions of  Closed  Loop
     vehicles designed to meet a  1.0 g/mi  NOx standard.  The  NOx  com-
     posite  values  are  presented in  units   of g/mi  at  10,000  mile
     intervals.

-------
    Table V-4
NOx Emissions for Closed Loop Vehicles
Miles
0
1
2
3
4
5
6
7
8
9
10


E(EGR)
1.58
1.71
1.84
1.97
2.10
2.23
2.36
2.49
2.62
2.75
2.88


S(EGR)
0.06
0.08
0.10
0.12
0.14
0.16
0.17
0.19
0.21
0.23
0.25


E(Low)
0.27
0.40
0.53
0.66
0.79
0.92
1.05
1.18
1.31
1.44
1.57


S(Low) E(Secondary) S(Secondary)
0.02 0.58 0.84
0.03 0.71 0.80
0.05 0.84 0.77
0.06 0.97 0.74
0.07 1.10 0.71
0.09 1.23 0.68
0.10 1.36 0.64
0.12 1.49 0.61
0.13 1.62 0.58
0.14 1.75 0.55
0.16 1.88 0.51
This table of composite emissions
NOx = 0.74
E(M/EGR)
4.69
4.82
4.95
5.08
5.21
5.34
5.47
5.60
5.73
5.86
5.99
yields
+ (0.15)
S(M/EGR)
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.02
E(M/Low)
0.80
0.93
1.06
1.19
1.32
1.45
1.58
1.71
1.84
1.97
2.10
S(M/Low)
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
E(M/2nd)
1.72
1.85
1.98
2.11
2.24
2.37
2.50
2.63
2.76
2.89
3.02
S(M/2nd)
0.07
0.07
0.07
0.06
0.06
0.06
0.06
0.05
0.05
0.05
0.04
Comp
0.74
0.88
1.03
1.18
1.33
1.48
1.62
1.77
1.92
2.07
2.21
a regression equation of:
(m)





(m = miles/10,000)

-------
                                   45
B.  Oxidation Catalyst vehicles designed to meet a 1.0 g/mi
NOx standard.

     1.  Discussion of the analysis.

     The  development of the  emission  rates  for  this  unit of  analysis
     relies on  the  same  methodology as that developed  in Sections IV.C.I.
     and IV.D.I. for Oxidation Catalyst  vehicles  meeting various HC and CO
     standards.   It  will  therefore  not  be  presented  in  detail.   In
     essence,  the  zero-mile  emission  levels  and deterioration rates for
     1980 Light  Duty Vehicles (as  developed in  the  1980 revision  of the
     Mobile  Source  Emission  Factors  Document  (EPA-400/ 9-78-005))  were
     adopted after taking into account the  effect of the Parameter Adjust-
     ment regulations and ratioing  the zero mile  emission level  to account
     for the different design standard.

     The  "percent  benefit"  figures  for  NOx attributable  to  the Parameter
     Adjustment  regulations are  presented in  Table  6.   They were derived
     in  the  same way as were the HC and  CO "percent benefit"  figures as
     described in Section IV.C.I.
                               Table V-5
       Percent  Benefit  from the Parameter Adjustment Regulations

                              1981                   1982+

        NOx                  -5,0%*                  -7.6%*

     These  figures of   "percent  benefit"  or,  more  precisely,   "percent
     penalty" were  then applied  to the  1980  Light Duty  Vehicle  zero  mile
     emission level (1.56  g/mi)  and deterioration rate  (0.10  g/mi/10,000
     mile).

     The final step consisted of applying a ratio  of (1.0/2.0)  to the  zero
     mile emission  level  to  account for  the  fact that  the  1980  fleet  is
     designed to meet a 2.0  g/mi  NOx standard whereas  the post-1980 fleet
     will be  designed  to meet a  1.0  g/mi  NOx  standard.   The  resultant
     zero mile  emission levels and deterioration rates are presented  in
     Table V-6.
     * These figures are  negative.   They reflect what was  observed  in  the
     data,  that  removing  those  vehicles  with maladjustments  which will
     most  likely be  prevented  by  the Parameter  Adjustment  regulations
     results  in an increase in average NOx  levels.

-------
                                       46
                                   Table V-6
              NOx Emission Factors for Oxidation Catalyst Vehicles

              1981                          1982+

      Zero Mile     Deterioration Rate     Zero Mile     Deterioration Rate

NOx      0.82              0.11               0.84              0.11

         2.  Unit-of-Analysis Composite.

         This section  presents  the composite NOx emissions  of Oxidation Cata-
         lyst vehicles  designed to  meet a  1.0  g/mi  NOx  standard.   The  NOx
         values  are presented  in units of  g/mi  at  10,000  mile  intervals.
         Table V-7 simply tabulates the equations shown in Table V-6.


                                   Table V-7
             NOx  Emission  Composites  for  Oxidation Catalyst Vehicles

         Miles               1981 NOx            1982+ NOx

           0                    0.82                 0.84
           1                    0.93                 0.95
           2                    1.04                 1.06
           3                    1.15                 1.17
           4                    1.26                 1.28
           5                    1.37                 1.39
           6                    1.48                 1.50
           7                    1.59                 1.61
           8                    1.70                 1.72
           9                    1.81                 1.83
           10                   1.92                 1.94

References for Section V.

1.   Internal  EPA memorandum; July,  1980;  from David  Brzezinski/David Hughes,
Inspection/Maintenance  Staff to  Tom  Cackette,   Chief,  Inspection/Maintenance
Staff.   "EGR  Tampering/Failure  Rates:   Comparison of 1978 MSED  and FY77 Emis-
sion Factors Surveys".

-------
                                        47
VI.  COMPOSITE EMISSION FACTORS FOR THE 1981 FEDERAL FLEET

    A.  Fraction of Total 1981 New Car Sales Contributed
    by Each Unit of Analysis"

    The  effect  of  the  CO  Waiver  decisions  of  1979 and  1980 for  Light  Duty
    vehicles was to allow certain  engine families for some of the manufactur-
    ers to  certify  to a 7.0 g/mi CO standard in  1981 and  1982.   (Not all  eng-
    ine  families which  were waived in  1981 were also  waived for  1982.)   The
    majority of  the waived families are  expected to employ  Closed Loop tech-
    nology in 1981  and  1982,  however,  several manufacturers  are  also expected
    to  certify  the  waived portions  of  their  Federal  fleets with  Oxidation
    Catalyst systems.

    The  model's  predicted  technology  mix  for  the Federal  fleet in 1981  and
    1982 is 93 percent Closed  Loop and  7 percent Oxidation Catalyst, based on
    confidential  EPA  Certification  sales  estimates  and  CO  Waiver   testi-
    mony. [1,2,3,4,5]*  It is assumed that  this technology  split is essentially
    unaffected by the waiver decisions.

    It is predicted  that  28 percent of the 1981  Federal fleet has been  waived
    to the  7.0  g/mi CO standard,  based on  EPA  Certification sales  estimates
    for  1980   and   the  CO   Waiver  decisions   for  specific   engine   fami-
    lies.[1,2,3,4,5]  The  emissions model  is  constructed  so  that the  waived
    segment of the  fleet  is apportioned among the different  units of analysis
    as predicted  from technologies employed  in   1981 Certification cars  cur-
    rently on record with EPA,  and CO Waiver testimony.[1,2,3,A,5]

    As was discussed in Section I.C.I,  the portion of the  Ford Motor  Co. fleet
    which may be produced as  an  open  loop system employing  a Three-Way cata-
    lyst has been subsumed into the Closed Loop vehicle  technology type.

    Table VI-1  presents  the percentage contribution  of  each  unit  of analysis
    to the 1981  fleet.
    * These references appear at the end of Section  VIII.

-------
                                      48
                                   Table VI-1
            Percentage Contribution"of the Various Uhits of Analysis
                           to the 1981 Federal Fleet

Technology         Percent of        CO Design       Percent of
   Type            Fleet Sales       Standard        Fleet Sales

Closed Loop             93               3.4              67.0
                                         7.0              26.0

Oxidation Catalyst      7                3.4              5.0
                                         7.0              2.0
    B.  Composite Emissions versus Mileage.

    In  the  emissions model,  the 1981  fleet composite  at  each mileage  point
    from zero to 100,000 miles, in intervals of 10,000 miles,  is  calculated as
    a  weighted  average  of  the  emission  levels  of  each  unit  of  analysis.
    Tables VI-2.a., VI-2.b. and  VI-2.C.  present  the sales  fractions  and  emis-
    sion  levels  of each unit  of analysis,  and the fleet composites,  at  each
    10,000 mile interval.

-------
                                                             Table VI-2.a.
ON




HC Emissions by
1981
Federal Fleet
Units of Analysis and Fleet Composites
Oxidation
Closed Loop

Mileage in
Miles/10,000

0
1
2
3
4
5
6
7
8
9
10
at 3.4 CO
Emission.
level in
g/mi
0.40
0.59
0.77
0.96
1.15
1.34
1.53
1.72
1.91
2.10
2.29

Sales
frac.

0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
Closed Loop
at 7.0 CO
Emission
level in
g/mi
0.40
0.59
0.77
0.96
1.15
1.34
1.53
1.72
1.91
2.10
2.29

Sales
frac.

0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
Catalyst
at 3.4
Emission
level in
g/mi
0.24
0.48
0.72
0.96
1.20
1.44
1.68
1.92
2.16
2.40
2.64
CO
Sales
frac.

0.05
0.05
0.05
0.05
0.05
0.05
0.05
•Q.Q.5
0.05
0.05
0.05
Oxidation
Catalyst
at 7.0
Emission
level in
g/mi
0.24
0.48
0.72
0.96
1.20
1.44
1.68
1.92
2.16
2.40
2.64
CO
Sales
frac.

0.02 -
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02


Fleet Composite

0.39
0.58
0.77
0.96
1.16
1.35
1.54
1.73
1.93
2.12
2.31

-------
                                                             Table VI-2.b.
o
u-i




CO Emissions by
1981
Federal Fleet
Units of Analysis and Fleet Composites
Oxidation
Closed Loop

Mileage in
Miles/10,000

0
1
2
3
4
5
6
7
8
9
10
at 3.4 CO
Emission
level in
g/mi
5.21
8.02
10.83
13 .65
16.46
19.27
22.08
24.90
27.71
30.52
33.33

Sales
frac.

0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
Closed Loop
at 7.0 CO
Emission
level in
g/mi
7.32
10.08
12.85
15.62
18.39
21.16
23.93
26.70
29.47
32.24
35.00

Sales
frac.

0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
0.26
Catalyst
at 3.4
Emission
level in
g/mi
2.24
4.39
6.54
8.69
10.84
12.99
15.14
17.29
19.44
21.59
23.74
CO
Sales
frac.

0.05
0.05
0.05
0.05
0.05
0.05 •
0.05
0.05
0.05
0.05
0.05
Oxidation
Catalyst
at 7.0. C
Emission
level in
g/mi
4.61
6.76
8.91
11.06
13.21
15.36
17.51
19.66
21.81
23.96
26.11
-u
Sales
frac.

0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02


Fleet Composite

5.60
8.35
11.11
13.86
16.62
19.37
22.12
24.88
27.63
30.39
33.14

-------
                     Table VI-2.C.
                   1981  Federal  Fleet
NOx Emissions by Units of Analysis and Fleet Composites
Closed Loop Vehicles
Mileage in
Miles/10, GOO

0
1
2
3
4
5
6
7
8
9
10
Emission
level in
g/mi
0.74
0.88
1.03
1.18
1.33
1.48
1.62
1.77
1.92
2.07
2.21
Sales
frac.

0.93
0.93
0,93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
Oxidation Catalyst Vehicles
Emission
level in
g/mi
0.82
0.93
1.04
1.15
1.26
1.37
1.48
1.59
1.70
1.81
1.92
Sales
frac.

0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07

Fleet Composite

0.75
0.89
1.04
1.18
1.32
1.47
1.62
1.76
1.91
2.06
2.20

-------
                                  52
C.  Regression Equations.

The  fleet  composites presented  in  Section VLB  need to be  summarized in
linear  form  for the  purposes  of emission  factors  programming.   However,
these emission  levels do not  constitute an  exact linear  function  of the
mileage  points  (although  they are  very close  to  a linear  function).
Therefore, these  composite emissions  have been  linearized by  regression
analysis.  Table VI-3 presents these linear equations by pollutant for the
1981  federal  fleet.   In all  the emission  equations, the  constant  equals
the Zero-mile emission level in  g/mi  and the slope  equals  the regression-
based composite deterioration rate in g/mi/10,000 miles.


                               Table VI-3
              Regression Equations  for 1981 Federal  Fleet

     Pollutant	Equation  (q/mi)

         HC                                     0.39 + (O.l9)(m)
         CO                                     5.60 + (2.75)(m)
         NOx                                    0.75 + (O.l5)(m)

     m = miles/10,000

-------
                                       53
VII.  COMPOSITE EMISSION FACTORS FOR THE 1982 FEDERAL FLEET

    A.  Fraction of Total 1982 New Car Sales Contributed
    by Each Unit of Analysis.

    Some engine families which received CO waivers in 1981 received 2 year
    waivers; these families amount to an estimated 10% of the 1982 Federal
    fleet.  No engine families received waivers for 1982 that had not previ-
    ously received them for 1981.  These engine families are accounted for in
    the model in the same way as are the waived families for 1981.  The pre-
    dicted technology split is the same (93% Closed Loop, 7% Oxidation Cata-
    lyst) and the waiver decisions are expected to have essentially no impact
    on this split.

    Table VII-1 presents the contribution of each unit of analysis to the 1982
    Federal fleet.
                                  Table VII-1
            Percentage Contribution  of the Various Units of Analysis
                           to the 1982 Federal Fleet

Technology          Percent of         CO Design       Percent of
   Type             Fleet Sales        Standard       Fleet Sales

Closed Loop             93                3.4             83.6
                                          7.0              9.4

Oxidation Catalyst       7                3.4              6.4
                                          7.0              0.6
    B.  Composite Emissions versus Mileage.

    Tables VII-2.a. through VII-2.C.  present composite emissions  and sales
    fractions for the various units.of analysis and the 1982  Federal fleet
    composite, as explained in Section VLB.

-------
                                                              Table  VII-2.a.

-------
                                                             Table VII-2.b.
I/I




CO Emissions by
1982
Federal Fleet
Units of Analysis and Fleet Composites
Oxidation
Closed Loop

Mileage in
Miles/10,000

0
1
2
3
4
5
6
7
8
9
10
at 3.4 CO
Emission
level in
g/mi
5.21
8.02
10.83
13.65
16.46
19.27
22.08
24.90
27.71
30.52
33.33

Sales
frac.

0.836
0.836
0.836
0.836
0.836
0.836
0.836
0.836
0.836
0.836
0.836
Closed Loop
at 7.0 CO
Emission
level in
g/mi
7.32
10.08
12.85
15.62
18.39
21.16
23.93
26.70
29.47
32.24
35.00

Sales
frac.

0.094
0.094
0.094
0.094
0.094
0.094
0.094
0.094
0.094
0.094
0.094
Catalyst
at 3.4
Emission
level in
g/mi
2.17
4.25
6.33
8.41
10.49
12.57
14.65
16.73
18.81
20.89
22.97
CO
Sales
frac.

0.064
0.064
0.064
0.064
0.064
0.064
0.064
0.064
0.064
0.064
0.064
Oxidation
Catalyst
at 7.0
Emission
level in
g/mi
4.47
6.55
8.63
10.71
12.79
14.87
16.95
19.03
21.11
23.19
25.27
CO
Sales
frac.

0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006
0.006


Fleet Composite

5.21
7.97
10.72
13.48
16.24
18.99
21.75
24.51
27.27
30.02
32.78

-------
                     Table VII-2.C.
                  1982 Federal Fleet
NOx Emissions by Units of Analysis and Fleet Composites
Mileage in
Miles/10,000
0
1
2
3
4
5
6
7
8
9
10
Closed Loop Vehicles
Emission
level in
g/mi
0.74
0.88
1.03
1.18
1.33
1.48
1.62
1.77
1.92
2.07
2.21
Sales
frac.
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
Oxidation Catalyst Vehicles
Emission
level in
g/mi
0.84
0.95
1.06
1.17
1.28
1.39
1.50
1.61
1.72
1.83
1.94
Sales
frac.
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
Fleet Composite

0.75
0.89
1.04
1.18
1.33
1.48
1.62
1.77
1.91
2.06
2.20

-------
                                  57
C.  Regression Equations.

Table VII-3 presents the regression equations for the 1982 Federal fleet,
as explained in Section VI.C.  The CO equation is different than 1981 due
to the effects of changing CO standards for certain engine families and
due to the impact of the second step of the Parameter Adjustment regula-
tions being implemented; the HC and NOx equations are different due only
to the impact of the second step of the Parameter Adjustment regulations
being implemented.
                              Table VII-3
             Regression  Equations  for  th~1982 Federal Fleet

Pollutant	Regression Equation (g/mi)
HC
CO
NOx
0.39 H
5.21 H
0.75 H
k (0.19)(m)
r (2.76)(m)
- (0.15)(m)
   m = miles/10,000

-------
                                      58
VIII.  COMPOSITE EMISSION FACTORS FOR THE 1983 AND BEYOND FEDERAL FLEET

    A.  Fraction of Total 1983 New Car Sales Contributed
    by Each Unit of Analysis.

    Starting in 1983 the entire Federal fleet will be required to certify to
    the 3.4 g/mi CO standard.  The predicted technology split is the same (93%
    Closed Loop, 7% Oxidation Catalyst).  Table VIII-1 presents the contribu-
    tion of each unit of analysis to the 1983 federal fleet.


                                  Table VIII-1
         Percentage Contribution of the Various Units of Analysis
         ~to  the  1983  Federal Fleet

Technology Type              CO Design Standard       Percent of Fleet Sales

Closed Loop                          3.4                        93
Oxidation Catalyst                   3.4                         7
    B.  Composite Emissions versus Mileage.
    Tables VIII-2.a. through VIII-2.C. present composite emissions and  sales
    fractions for the various units of analysis and the 1983 and Beyond Feder-
    al fleet composite.

-------
Table VIII-2.3.
1983 and Beyond Federal Fleet
HC Emissions by Units
Closed Loop at 3.4 CO
Mileage in
Miles/10,000

0
1
2
3
A
5
6
7
8
9
10
Emission
level in
g/mi
0.40
0.59
0.77
0.96
1.15
1.34
1.53
1.72
1.91
2.10
2.29
Sales
frac.

.0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
of Analysis and Fleet Composites
Oxidation Catalyst at 3.4 CO
Emission
level in
g/mi
0.24
0.48
0.72
0.96
1.20
1.44
1.68
1.92
2.16
2.40
2.64
Sales
frac.

0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07

Fleet Composite

0.39
0.58
0.77
0.96
1.16
1.35
1.54
1.73
1.93
2.12
2.31

-------
Table VIII-2.b.
1983 and Beyond Federal Fleet
CO Emissions by Uhits
Closed Loop at 3.4 CO
Mileage in
Miles/10,000

0
1
2
3
4
5
6
7
8
9
10
Emission
level in
g/mi
5.21
8.02
10.83
13.65
16.46
19.27
22.08
24.90
27.71
30.52
33.33
Sales
frac.

0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
of Analysis and Fleet Composites
Oxidation Catalyst at 3.4 CO
Emission
level in
g/mi
2.17
4.25
6.33
8.41
10.49
12.57
14.65
16.73
18.81
20.89
22.97
Sales
frac.

0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07

Fleet Composite

5.00
7.76
10.52
13.28
16.04
18.80
21.56
24.32
27.08
29.85
32.61

-------
Table VIII-2.C.
1983 and Beyond Federal Fleet
NOx Emissions by Units
Closed Loop at 3. A CO
Mileage in
Miles/10,000

0
1
2
3
A
5
6
7
8
9
10
Emission
level in
g/mi
0.7A
0.88
1.03
1.18
1.33
1.A8
1.62
1.77
1.92
2.07
2.21
Sales
frac.

0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
of Analysis and
Fleet Composites

Oxidation Catalyst at 3. A CO
Emission
level in
g/mi
.84
.95
1.06
1.17
1.28
1.39
1.50
1.61
1.72
1.83
1.9A
Sales
frac.

0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07

Fleet Composite

0.75
0.89
l.OA
1.18
1.33
1.A8
1.62
1.77
1.91
2.06
2.20

-------
                                      62
    C.  Regression Equations.

    Table VIII-3 presents the regression equations for the 1983 Federal fleet,
    as explained in Section VI.C.  The HC and NOx equations are unchanged from
    the 1982 equations; the CO equation is different due to the universality
    of the 3.4 g/mi CO standard for the 1983 and Beyond Federal fleet.
                                  Table VIII-3
                   Regression  Equations  for 1983 Federal Fleet

              Pollutant                          Regression Equation (g/mi)

                 HC                                  0.39  + (0.19)(m)
                 CO                                  5.00  + (2.76)(m)
                 NOx                                 0.75  + (O.l5)(m)

         m = miles/10,000


References for Chapters VI-VIII.

1)  Federal Register, Vol. AA, No.  179,  9/13/79,  pp.  53376-53A08.

2)  Federal Register. Vol. AA, No.  233,  12/3/79,  pp.  69A16-69A62.

3)  Federal Register, Vol. A5, No.  22,  1/31/80,  pp. 7122-7138.

A)  Federal Register, Vol. A5, No.  55,  3/19/80,  pp. 1791A-17922.

5)  Federal Register, Vol. A5, No.  107,  6/2/80,  pp. 37360-37371.

-------
                                      63
IX.  CONPARISON TO PREVIOUS EPA EMISSION FACTOR ESTIMATES

This section simply compares the emission factor equations obtained from this
analysis with the equations developed in Appendix E to the Mobile Source Emis-
sion Factors Document (EPA-400/9-78-005).  The attached figures (Figures
IX.a.-c.) compare only the 1983-and-Beyond equations from this analysis with
the Appendix E equations for HC, CO and NOx.

-------
                                  64
 3.0
 2.5
«2.0
to


i-l.S
z
cc
h-


jl.O
 0.5
 0.0
                             FIGURE  IX.fl,

                                  HC
New Analysis

flPPENOJX E
                 20000      UOOOO     60000

                                 MILES
                               80000    100000

-------
                                   65
 35.0
 30.0
 25.0
 15.0
£10.0
  5.0
  0.0
                             FIGURE IX.B
                                  CO •
	 New Analysis
	 flPPENDJX E
                  20000     UOOOO     60000
                                 MILES
                                  80000     100000

-------
                                   66
 3.0
 2.5
~2.0
x
N.
z
oc
                             FIGURE IX.C,

                                 NOX
  1.0
a
a.
 0.5
 0.0
New Analysis

flPPENDIX  E
                              I
                 20000     UOOOO      60000

                                MILES
                                 80000
100000

-------