EPA-AA-TEB-EF-86-03
      Emission Factor Testing
     Needs  in the Latter 1980s
                 by

        Thomas L. Darlington


              June  1986
     Test and Evaluation Branch
Emission Control Technology Division
      Office of Mobile Sources
U.S. Environmental Protection Agency

-------
                 Emission Factor Testing Needs
                       in the Latter  1980s

1.0    Summary

2.0    Background

3.0    General Testing Needs by Vehicle Type
3.1    Light-Duty Gasoline Vehicles
3.2    Light-Duty Gasoline Trucks
3.3    Heavy-Duty Trucks,  and Buses
3.4    Light-Duty Diesel Vehicles and Trucks
3.5    Off-Highway Vehicles

4.0    Specialized Needs
4.1    Evaporative Emissions
4.2    Exhaust Emissions at Different Temperatures
4.3    CO Emissions at Low Speeds
4.4    Effects of Repairs
4.5    Disablement Testing

-------
                Emission Factor Testing Needs in
                        the Latter 1980s
1.0    Summary
       Emission Factor testing is  important  because of  its  use
in   MOBILES,   which   is  in   turn  used   to  prepare   State
Implementation Plans   (SIPs),  Environmental  Impact  Statements
(EISs),  and  develop   real  world  benefits  for  EPA  regulatory
proposals.   Many  others  (auto  manufacturers,  GARB)  use  the
emission factor data directly for many different  purposes.

       There  are  a  variety of  current  and  future  emission
factor  testing  needs  which  are enumerated  below.   EPA  should
consider  increasing  its commitment  to emission factor  testing
so that these needs are adequately addressed.

           Summary of EF Testing Needs
       0   Carbureted Cars - at higher mileages (70+K)
       0   Fuel injected cars - at higher mileages (50+K)
       0   Continued or increased testing at alternate sites
       0   Older cars (10-15 years old)
       0   Future cars/trucks with onboard systems or systems
             certified with higher volatility fuels
           More emphasis on LDGTs
           Chassis testing of heavy duty trucks
           Continued testing of transit bus engines
           Continued testing of heavy duty gasoline engines, if
           feasible
       0   Some testing of LDDTs
           Continued testing of temperature effects on
             evaporative emissions
           Characterization of evaporative HC running losses
           High altitude evaporative HC testing
           Low temperature testing of carbureted and fuel
             injected cars with wintertime fuels
       0   Modal testing of new technology cars to aid in CO
             modeling at intersections
       *   Effectiveness of certain repairs
       4   FTP effects of pattern case fixes

-------
2.0    Background

       EPA's emission factor data base, which  contains  emission
tests  on  over  12,000  in-use  light-duty  vehicles  and  trucks
tested since  1971,  is EPA's best  source  of information  on how
cars and trucks actually perform on the road.   As a data  set on
in-use  car  performance  with  respect  to  emissions  and  fuel
economy,   it  is  second  to none  in size  and  completeness  of
information on  each  car  tested-  It  is  used  primarily  by the
Emission Control Technology Division of EPA's  Office of  Mobile
Sources  (QMS)  to update  its  in-use  emissions'  models such as
MOBILES,   and  to  estimate the  benefits of  certain  regulatory
proposals,  for  example,  volatility controls.   States and local
areas  are  relying on the data through the use  of MOBILE3  to
estimate emission inventories, determine the effects  of I/M and
antitampering   programs,   and   determine  the   effects   of
transportation  control   measures   and construction  projects.
Others are given free  access to  the data,  and  use  it on  a
periodic  or as-needed basis.   For example, most of the major
auto manufacturers have  requested  access  to the data,  and some
review the new data  on  a weekly or monthly basis  to determine
the in-use performance  of their cars.  The potential  benefits
of  their   concern   with  in-use  performance   should  not  be
underestimated.   Another  example  user is  the  California  Air
Resources  Board  (GARB),  which has  an  ongoing  in-use testing
program  of its own.  CARB is  currently using the  EPA  data to
compare  their  estimates  of light-duty vehicle emission factors
to  EPA's.   Other  QMS  divisions  have  also used the  emission
factor data base on numerous occasions.

       With all  of  its  current merits,  in order for  the data
base  to   remain  useful,  EPA must  have  a  high  commitment  to
continued   emission  factor   testing,   and   should   consider
increasing  that  commitment.    There  is,  and will  be,  a great
need  to  test new  technology  vehicles   at   a   wide  range  of
mileages  and  ages  (to  assess  durability); vehicles with  new
fuel delivery systems (fuel  injection), new emission standards,
and   perhaps    new    evaporative   control   hardware   (onboard
systems).  Additionally,  there are many special  emission factor
testing needs which have  not yet been addressed.  All  of these
needs, both present and future, will be discussed in this paper.

-------
3.0    General Testing Needs by Vehicle Type

3.1    Light-Duty Gasoline Vehicles (LDGVs)

       Carbureted  Cars  -  Although   fuel   injected   cars  are
rapidly becoming more  common,  carbureted closed loop cars  will
continue to  dominate the in-use  fleet in the  late 1980s.   The
emission factors for these  cars are based  on  quite  a few  cars
tested  at  low mileages  and  some cars  tested at  the  higher
mileages (see Figure 1).  The emission data  indicated  that  cars
could be  divided into three groups by emission  levels - normal
emitters, high emitters, and super emitters.  The  high emitting
cars  usually have  one  or  two  problems which  make  them  emit
above the  normal cars.  The super emitters have  extraordinary
problems,  either  in numbers  or  type, that make them emit  far
above the  high  emitters.   As the fleet  of  new  technology  cars
grows  older,  more   are expected to  migrate  from  the  normal
category  into the  high  and   super  emitter  categories.   The
emission rates of the  fleet  of carbureted  closed-loop cars are
very dependent on how  quickly  this migration occurs.   In turn,
the emission  rates  of  all  LDGVs are quite dependent on the high
mileage  cars,  since  50% of  the  LDGV VMT  is  from  cars  with
odometer values above  50,000 miles.   Therefore, it is important
to be able to predict  the rate of growth of the high and super
emitter  categories  with  confidence,  and  also  their  average
emission  levels.  To do this  effectively,  EPA needs  more data
on carbureted cars tested at mileages above  7OK miles.

       Another point  with respect to  the high mileage  data we
do have  is that it  is from cars that accumulated the  mileage
quickly,  since  most of the tests were  performed two or three
years ago  on  1981 cars when they  were  only about two to three
years old.   Most cars  accumulate  mileage slower than that,  for
example, a 1981  car sold in January  of  1981 would be expected
to have  about 55,000 miles on it  at this time.  Emissions from
normally accumulated mileage cars  could  be  somewhat higher than
advanced  accumulated  mileage  cars  since  they may  experience
more cold  starts and  more  severe stop-and-go driving than  the
advanced mileage cars.

       Fuel  Injected  Cars - Manufacturers  are  introducing fuel
injection on new  and existing engines  at  a rapid pace.   EPA's
emission factor  program has been  focusing on these cars in the
last  two years  for  evaporative  as  well  as  exhuast emission
reasons, but  we clearly have little  high  mileage data  on fuel
injected  cars  (see  Figure  2).   Generally, our experience  to
date  has  been  that   fuel  injected  cars  are  cleaner  than
carbureted cars.  We have not  found a fuel  injected car  in  the
super emitter  category  yet,  and  those  in  the high"  and normal
categories have  lower  average  emissions  than  their  carbureted
counterparts.   However, a  somewhat new development  with  fuel

-------
injected cars may change this outlook.  Many fuel  injected  cars
are  experiencing  injector  plugging  from  deposits  that   are
building up  in the nozzles  of the  injectors.   These  deposits
come  from   in-use  gasolines  with   little  or   no  detergent
additives.    (Higher  detergent  gasolines  supposedly can clear
away these  deposits.)   In a  multipoint  fuel injected  car,  the
deposits can  disturb  the distribution  of  gasoline  into  the
cylinders,   causing driveability problems  potentially  resulting
in  incomplete  combustion and higher  HC  emissions.  We plan to
monitor  this situation by inspecting and testing  injectors on
EF cars that have HC emissions in excess of about  0.6  g/mi,  and
have few or no  other  apparent  problems  (such  as  a  three-way
system  failure or faulty ignition  system).   In  the   last  few
months  we have seen  very few fuel injected cars above 0.6  g/mi
(most  are  0.2-0.3 g/mi HC); however,  these have been  mostly
lower mileage  (20K-30K) cars.

       All Cars  - For  the last  few  years  our  primary emission
factor  testing has been  conducted  in Ann Arbor (other smaller
and  specialized  programs  have  taken  place  in Denver  and  East
Liberty,  Ohio).    Resource   considerations   have  been   the
motivating factor  in consolidating  the testing at  the Ann Arbor
site.  We do not  fault the testing here in any way, but  we are
concerned about the  risks of basing nationwide vehicle emission
factor  estimates  on testing primarily  at one  site.   We  have
seen  differences  in   evaporative   emissions   between  samples
tested  in Ohio and Ann Arbor.  Most  of  the difference we  have
been able to attribute to minor test procedure differences, but
the  potential  exists  for   different   evaporative  emissions
between  different sites  because  of the  wide differences  in
in-use  fuels used.   It is possible  that some fuels cause  more
rapid deterioration in evaporative systems than other fuels.

       Another consideration  is  the  severity of weather  in the
Detroit  area.   Detroit cars  undoubtedly experience  more  low
temperature  cold  starts  than the national  average,  and these
conditions are well  known for producing engine wear.   There is
also  more  possibility of  choke malfunctions.    These factors
would make the Detroit sample a higher emitting  sample than the
national average.

       An   ideal   solution  would   involve  a   return  to   the
multicity EF programs  of the  past.   As an  alternative,  it is
imperative that we maintain  and  perhaps  increase our commitment
to  the  off-site testing  we  are  doing in  Ohio  and  possibly at
other   contractor   facilities.     In  addition   to   studying
specialized  concerns  at these sites,  we  can  continue to  make
comparisons  between  the  as-received  tests  on  cars  at   the
different sites.

       Future  Cars -  Due to  the  need  for ozone control  and
control  of  air toxics, it is almost a certainty  that EPA  will
require  auto   manufacturers  to  control   excess  evaporative

-------
emissions, refueling emissions, or both.  This  could occur with
the 1989  or  1990 model year.   The manufacturers  will  have  to
develop  new  hardware  and alter  purge  system  management  to
accomplish this.  Naturally, there  will be a need to test  how
these  systems  perform  in use.   A  continued  and  consistent
commitment to EF testing will  allow us to perform  this critical
testing when the time comes.

       Older  Cars  - The  data in Figure  3  illustrate that  age
0-5 cars  accumulate  50% of the fleet  VMT,  but only contribute
16-22  percent   of  the   fleet   emissions   because  of   their
relatively low  emission levels.   The  most  emissions come  from
the  age   11-15  group,  which  contributes 35-37  percent  of  the
fleet emissions.  These values take  into account the fact  that
many  of   these  11-15  year old  cars  have  been scrapped.   The
dominance  of  the  fleet emission factor  by these  cars  can  be
traced to  the deterioration rate,  or increase in emissions with
age/mileage.

       This concept  is  illustrated  further in  Figure  4  which
shows the yearly  emissions of a  1985 car  which lasts for 20
years.  Each  year's  emissions are  estimated  as the  product of
the  annual  VMT  and  the  average  emission  factor  (at  year
midpoint)  for  that  year.   For 1985  cars that  last 20  years,
they  produce  the  most  emissions  in  their twelfth year,  the
reason being  that  their  emission  rates  have  increased  much
faster than their yearly VMT has dropped.

       The implications  of these concepts  are  that  if EPA has
missed the mark at  all  in estimating  emission  rates  of  the
1970-75  cars  by  extrapolating  low  mileage  data  to   higher
mileages,   the  current   fleet   emission   rates   could   be
significantly  affected.    We  suggest  that  at  least 50  twelve
year old  cars be tested each year  to  check the  older  emission
rates.    If   the    average  emissions   of  these   cars   are
significantly  higher  than the  equations  predict,  then  some
adjustment could be made.                         ...     ...  .
       It  should be noted  that  it  will be  more expensive  to
test 50  older cars than  50 newer cars.  With newer cars,  our
current  rejection rate  is  about 12 percent.   These cars  are
rejected for  being too  expensive to repair prior to  testing  on
the dynamometer.   Where a  car  needs minor nonemission  control
repairs such  as  new brake linings or a new exhaust  pipe,  we  do
make those repairs  so the cars  can be  tested.   To  conserve
resources we  do  not,  however,  replace  expensive items  such  as
faulty transmissions unless they fail during testing.

       On older  cars there  is  a greater likelihood that we will
have to reject more cars  as too expensive to  repair, and that
cars will  fail  while we  are testing them.   This will  increase
the recruitment and repair costs over what we  now pay for  newer
cars.

-------
                                8

3.2    Light Duty Gasoline Trucks (LDGTs)

       Light  duty  gasoline  trucks,   which  are  all  gasoline
trucks under 8500  Ibs  gross weight,  make up roughly  20%  of  the
total VMT  of  all  vehicles  combined.   However,  because  their
emission  rates  are  higher  than   light-duty   vehicles,  they
account for about  32%  of  the HC and CO  emissions of the fleet
(see  Table  1),  and  therefore  it   is   important  that  their
emission factors are accurately characterized.

       Light duty truck manufacturers use very  similar  emission
control equipment  on their cars  and  light  trucks.   Many light
trucks  are  currently  equipped  with  closed  loop fuel  controls
and  3-way  catalysts,  even though the  current  NOx  standard is
2.3 gpm, which  in  many cases probably does not  require  closed
loop  fuel  control  to attain.   Because  of  the  similarities
between light  trucks and  cars,  the  emission factors of  trucks
are  in  part developed from the  emission  factors of cars,  for
which there are much more  data.  However, there are indications
that  the  resulting  emission factors  for trucks  are  too  low.
Recently we tested 50  1981 LDTs in Ohio, and found  that  the CO
emissions of  the  untampered trucks  were  20% higher than what
NOBILE3 would  predict  for  untampered  1981  trucks.   Therefore,
we have started  testing trucks  again in the current EF program,
and since the  emission standards of  trucks  are  still changing,
this  effort should  continue.   (The  1985  trucks  have  a  full
useful  life definition,   1987 trucks  will  have a  particulate
standard of  0.26 g/mile,  and 1988 and later trucks  will  have a
NOx  standard  of  1.2  and  1.7  g/mile for  0-6000  pounds  and
6000-8500  pound  trucks,   respectively).   Although  we do  not
think it is  necessary  or  feasible to devote equal  resources to
light  trucks  and  cars,  there  will  be  a  continuing  need to
support EF testing of trucks.

3.3    Heavy Duty Trucks (Gasoline and Diesel),  and Buses

       EPA's data  base on  in-use  emissions for gasoline  and
heavy   duty   trucks  has  been  greatly   augmented   through
cooperative testing with NVMA and EMA of in-use  truck  engines.
The  cooperative  testing  program is complete, and the emissions
of these  1979-82 engines  will  be used  in  a  future update of
model year 1979-1986 emission rates for heavy duty trucks.

       The  emission  rates  of trucks tested in  this program are
quite similar  to  MOBILE3  emissions  with  the  exception of HC
emissions  from  heavy-duty gasoline  vehicles,   which  from  the
test  data  are  around 8.7  g/bhp-hr,  where  MOBILES  predicts
around 3.7 g/bhp-hr.   The  HC emissions for  the test engines are
driven by one  truck with  HC  emissions of 43.8  g/bhp-hr, while
the  rest of the engines  ranged from 3-8 g/bhp-hr.   Many of the
gasoline engines were poorly maintained  and  had evidence of
being tampered.

-------
       The heavy-duty gasoline engine market is dominated  by GM
and Ford.  Only Ford  and Chrysler,  however, contributed engines
to the  in-use emissions project.   If  they could be  encouraged
to continue sending a few engines each year, we may  be able to
refine our emission estimates, particularly for HC.

       For  both  diesel  and  gasoline  engines,  it  would  make
sense  to encourage  ENA  and  MVMA  to  continue the  cooperative
in-use testing program on  a  more  limited  basis,  for  example,
each member agreeing  to  send  one or two engines per year.   This
way  we  could have  a continuing  program,   as with  light-duty
vehicles and  light-duty trucks.

       The emission  factors for trucks have also been improved
with  further  attention  to the  development of the   heavy duty
conversion factors which translate  engine emissions  in g/bhp-hr
to on-road emissions  in  g/mile.   While the emissions in g/mile
have been improved through this  analysis, there is still a need
to do further chassis testing of heavy  duty trucks  as  a  check
to  see  that  the  converted   engine emissions  are  realistic.
Several  trucks  from  each  truck  class  from  light-heavy  to
heavy-heavy   (both  gas  and  diesel)  would  be needed  in this
analysis.

       Looking ahead,  heavy duty truck HC and CO standards are
made  more  stringent  in  1987,   and  the  NOx and  particulate
standards are more stringent  in 1988.   The NOx  and particulate
standards  are further tightened  in 1991  and 1993.   The next
major cooperative  in-use program on heavy duty engines should
occur  when   the  1988   and   1989  model  year  engines   have
accumulated some mileage.

       EPA has recently  begun to test  some transit  bus engines
on bus  duty  cycles, which are significantly different than the
truck  transient   test,   and  yield  higher  levels   of   some
emissions.  Only one  bus engine has been tested so  far, and to
adequately  characterize  emissions  a.  few  more  are  probably
.needed.  Although  bus emissions  are probably  not a  large part
of  any urban inventory, they are important  from  an exposure
standpoint.                                            ;  :  •• - --

3.4    Light  Duty Diesel Vehicles and Trucks (LDDVs and LDDTs)

       Although  the  manufacturers  have  significantly  pulled
back  in their effort to  develop diesel cars,  there  has been
continued development and  demand for  diesel  engines  in  light
duty trucks.  In 1980 the percent  of  light truck VMT that was
attributed to diesel  trucks  was  less  than  1%.   It is expected
to grow to about 10% in 1990 and 28% by the year 2000.

       The zero mile levels of light diesel trucks  are based on
certification levels  and  the deterioration rates  are borrowed
from  light  diesel  vehicles.   This,  coupled with the fact that
the particulate  standard for diesel trucks was  lowered in 1985

-------
                                10

to  260  mg/mile, underscores  the need  to test  a  few  (perhaps
2Q-aQ) 1985 and 1986 light diesel trucks in the next  two years.

3.5    Off-Highway Vehicles

       There  are  four  main  types  of  off-highway  vehicles:
locomotives,   construction  equipment,   aircraft   and   ships
(vessels).   New data  are being compiled  for locomotives  and
aircraft,   and  the   current   AP-42   emissions   factors   for
construction equipment are based on a  recent  California study.
The  off-highway  source   that  needs  additional  attention  is
ships.  Although  there are some new data, the applicability is
limited.   For  example, the  new data  available are  for  large
diesels  used   in   commercial   ships.    However,   most  of  the
emissions  in a harbor area   come  from  the smaller  and  newer
diesel  powerplants  found  in  tugs  and  construction  barges.
There are  very little data  for this size  engine  (around  1500
hp).  Testing  of  a few of these engines  on  representative duty
cycles  is  needed  for  cities  like  San Diego,  Houston  and  New
York.

-------
                                11
4.0    Specialized Needs

4.1    Evaporative Emissions

       As it  has  become more apparent  that  many  areas  of  the
nation will not  attain the ozone  standard by the end of  1987,
EPA   has  searched   for   additional   HC   control   measures.
Evaporative emissions from mobile  sources  have  come  under  close
scrutiny because  the volatility of  in-use fuels has  increased
dramatically  over  the  last  decade,  thereby  leading to higher
vehicle evaporative HC emissions.

       EPA has concentrated its  testing efforts in  quantifying
the  "first order"  effects  on evaporative  emissions,   namely,
fuel  volatility  and  ambient  temperature.    EPA   now  has  a
substantial  volatility   data   base  with  which  to   estimate
volatility  effects  and  the benefits  of  volatility controls.
Characterizing the  effects  of temperature,  however,  has  thus
far  proved to  be  more elusive.   The  temperature  testing  is
resource intensive  on a per car basis,  therefore, only  20 fuel
injected  and  20  carbureted  cars  have  received  temperature
testing.  Furthermore, evaporative emissions  are very  sensitive
to temperature, leading to  high variability  in the  data.   This
underscores the need for continued  temperature  testing  so that
the temperature effects can be reasonably estimated.

       Quantifying  the  temperature  effects  is  also   a   very
important  issue   from  the  standpoint   of   controls.    Ozone
exceedences  are   known to  be  associated with  hotter  summer
weather.  The maximum diurnal  temperature  in EPA's Federal Test
Procedure  is  84°F.   EPA  needs  to  be  sure   that   whatever
evaporative controls are  put in place,  that  they  are  effective
for most situations.   Accurately  characterizing the effects of
temperature  on  evaporative emissions,  and  also  the  ambient
temperatures  associated  with  most  ozone   exceedences,   will
enable us to do that.   --..

       High ambient temperatures  and  extended  vehicle  driving
can  also produce  an evaporative  emission from cars  known  as
"running losses."   When  cars  are driven, fuel  in the  tank is
heated by the exhaust system,  and in fuel  injected cars  is also
heated  by  unused  fuel  recirculated  from the  injection  system.
The increase in tank  fuel temperature  produces  additional  vapor
which  is  normally  purged  into the engine  while  the  car  is
operating.   Under  some conditions,  however,   the rate of  vapor
generation can overwhelm the purge  system,  leading to  a  rapid
buildup  of pressure in  the  fuel  tank.  All   fuel  tanks  are
designed to release  that  pressure  through  the fuel cap at  about
1.5  psi  for  safety reasons,   and  that  release  is  a  "running
loss."  Running  losses can also occur during periods of engine
operation where  purge  is  not  taking  place,   such  as  extended
idle  operation.   Here the losses may not vent through  the cap,

-------
                                12

but  instead build-up  and overwhelm  the  canister,  since  the
canister is not being purged.

       At   this   time   there  are  very  little  data   on  the
conditions under which running losses occur.  EPA has  initiated
a  small  scale  test  program  on one  fuel  injected  and  one
carbureted  car to  characterize  the  extent  of  driving  which
would  trigger  running  losses.    However,   the  occurrence  of
running losses is probably also very system dependent, and so a
larger  scale effort on  more  cars  is needed  in the very near
future.

       A third area where EPA needs more evaporative data  is at
high altitude.   Starting with the 1985 model year  the high and
low  altitude  evaporative  standard is  the  same  (2g).    Since
atmospheric  pressure  is  lower  at  high  altitude,  more HC  is
produced during  a given diurnal  at  high altitude  than  at  low
altitude.   For the  same  reason,  vapor lock is more of  a  problem
at high altitude than at low altitude,  so  the oil  companies
typically market  lower volatility fuels  at  high altitude during
the  summer   than  are  marketed  at  low altitudes  with  similar
temperatures.

       In EPA's High Altitude Report  to  Congress,  it was stated
that the manufacturers might  need to  increase canister  capacity
to  meet  the  2g  standard  at  high  altitudes.   It  is  known,
however, that some companies  are  using exactly  the  same  systems
at high as  well as low altitude, either because they  feel that
their  low  altitude  systems  have  enough  capacity  for  high
altitude, or that there  is a  lack of EPA enforcement  effort at
high altitude, or both.

       The  current  MOBILE3 high altitude  evaporative  emission
factors  for  1985+  LDGVs  are  set  equal  to the   low  altitude
levels because the  standards  are the  same.   But if the  systems
are  not different, the  in-use high altitude evaporative  rates
may be  higher  than  at lower  altitudes.   Therefore, there  is a
clear need   for  in-use  evaporative  testing  of 1985 and  later
cars  at  high  altitude.    This testing   should  utilize  a
representative in-use high altitude fuel.

4.2    Exhaust Emissions at Different Temperatures

       Exhaust  HC  and  CO  emissions  are  very  sensitive  to
temperature.   During  a  cold  start at  cold  temperatures  choke
operation in carbureted  cars  produces very rich mixtures  at a
time when the catalyst is not operating, thereby producing high
HC  and  CO  emissions.    Fuel  injected cars  need rich mixtures
during  cold  starts  also, but  these  systems  achieve  tighter
control of  fuel/air ratios, thereby producing less HC and CO at
cold temperatures than carbureted cars.

-------
                                13

       The primary needs in the  area  of cold temperature  data
are  for   additional   testing  of  carbureted   cars   at   cold
temperatures on typical  wintertime fuels  of  13+  psi  (most tests
have been  done  on Indolene,  at  9 psi),  and  for fuel  injected
cars tested at  low  temperatures.   The  fuel injected  testing
should  also utilize  typical wintertime  fuels.    We  currently
have about  30 fuel  injected cars which have been tested  at 20°
and 50°, and we need a minimum of 30-40 more  cars to  accurately
characterize  the  behavior  of  these   cars separately  from
carbureted  cars.   Also, there  may be  substantial  differences
between throttle-body and multiport fuel  injection that need to
be identified.   The cold temperature testing  is  currently being
done by TEB,  drawing  on   the  pool  of  fuel  injected  emission
factor cars that we are recruiting.

       As previously discussed,  HC and CO emissions are  high at
very low  temperatures.   As the  temperature is increased,  these
emissions  decrease because  less choke  operation  is  needed.
Emissions  find  a  low  point  around 70°  to  80°.  This  is not
surprising because  it is the test temperature  range in  which
cars are certified, therefore,  manufacturers  have  designed cars
for low  emissions  in this  range.  At  temperatures   above  85°,
however,  HC and  CO  emissions  start to  increase  again.   This
increase  is due  to  a   change  in the  density  of air  as the
temperature  rises.   As  it  rises, the  air becomes less dense,
causing an  increase in  the  fuel/air  ratio since there  are  less
oxygen  molecules  per  unit  volume  of  air  than  at  a   lower
temperature.

       The  temperature  correction  factors   for  the  different
model  year  groups  for  HC  and  CO  are  shown in Figures   5-10.
There  is a  separate figure  for  each bag  of  the  FTP.   Figure 6-
shows  the  HC correction factors for  the stabilized  bag.   The
1975-80 cars display  significantly more  emission sensitivity to
higher  temperatures  than  the   other   cars.   (CO   correction
factors for  these cars  display  a similar sensitivity  in Figure
9.)  These correction  factors  are based on four 1977-78  cars
tested in the 1979 Gulf Study of emissions versus temperature.
Of  the four  cars, three  were  very  sensitive  to temperature,
i.e.,  the stabilized HC emissions of  a 1978 Buick  (V6>  were
1.93  g/mi  at   80F°,  and   26.01  g/mi  at  110°F with the air
conditioning on.

       The  high temperature  correction  factors for  1981  and
later  cars  are based on a few  prototype  and  California  cars
tested  in  the  same  (Gulf  Research)  test program.   Therefore,
there  appears to be a need for  more high temperature testing of
newer  fuel  injected and carbureted cars.   Some of this  testing
has already been done  as  a  part of  our evaporative emissions
versus  temperature and RVP  program  (exhaust  emissions  were
collected at 75,  85,  and 95°F).  This most recent testing also
includes  heat   builds   (diurnal  evap   emissions),   which  the
previous  Gulf  Research  testing  did  not.  These  data will  be

-------
                                14

analyzed, and  additional new  technology- emission  factoc  cars-
may be tested at high temperatures.

       The  effects  of  higher temperature  on fleet  HC and  CO
emissions of LDGVs  are  shown in Table 2.  NOBILE3 predicts  for
calendar year  1985 that LDGV CO emissions are  82% higher  at
100°  than   75°,  and  that HC  emissions are  20%  higher.    The
magnitude of these  differences  decrease  significantly in  the
year 2000,  when  all cars are assumed  to have closed loop fuel
control, and can better  compensate  for the change  in  fuel/air
ratio brought about by the less dense air.

4.3    CO Emissions at Low Speeds

       One  of  the most  persistent  problems facing  air'quality
modelers today  is how to estimate  low speed CO  emissions.  The
applications are numerous  since most states require  that  an
Environmental  Impact  Statement  (EIS)  be prepared  for  any  new
construction in  a  downtown  area.    The EIS  must  assess  the
impacts  of  changes  in  traffic  behavior on local or "hot  spot"
CO emissions.

       EPA  provided states with a low speed version  of MOBILE3
which  is capable of  predicting  CO emissions in  g/mi  down  to
about  2.5  mph.   These   are  transient emissions, which involve
stops,   idle  periods,   acceleration  and   deceleration  modes.
While  this  seems  to  have  met  the  needs  of some,  there  are
problems with  this approach.   The  emissions  in g/mi  are very
sensitive to low speeds,  as evidenced  by the emissions  at  3,  4,
and  5 mph  presented  in Table  3.    In an effort  to  study  CO
emission sensitivity  at  low speeds,  EPA  developed three  low
speed  cycles under  5 mph from the GM chase car data,  and will
soon  be  testing  all  emission  factor cars on  these and  other
speed cycles.                           ;;;               ;;•

       An alternative to transient modeling at  low speeds  is
modal  modeling,  where  a given  cycle is. broken  down  into  its
various  modes  (idle,  accel,  decel,  cruise)  and emissions  are
estimated for each mode.  These estimates are then placed  in an
indirect source  model such as HIGHWAY oc CALINE,  which predict
CO  concentrations  around  the   area   of   interest.    EPA  has
published modal  factors which can  be  used  to  convert transient
emissions  into  modal  emissions.   The  limitation  with  these
factors  is that   they  were  developed  eight   years   ago  on
1972-1976 cars,  and therefore do not account for the  behavior
of  new  technology closed   loop  fuel  control  cars.   Updating
these  factors  to  include  newer technology  cars is vital  to
indirect  source  modeling,   and  therefore  vital  in  order  to
accurately  prepare  EISs.   EPA  should  give  high  priority  to
developing  an intersection-type modal  testing  program for newer
technology cars.

-------
                                15


       Another  concern with  low speed  CO  emissions  is  the
interaction with temperature.  The speed  correction factors for
HOBILE3 were  developed from cycles which were run  in the  hot
stabilized  (as  opposed to  cold start)   mode at  about  75°F.
Alternately,  the  temperature  correction factors were  developed
with tests  at different temperatures using  the FTP cycle with
an average speed of 19.6 mph.   No other cycles were used.

       When a MOBILES user specifies that they want  emissions
at 7 mph and 25°, the  speed and  temperature correction factors
are  estimated separately  and applied  independently.    If  the
speed  correction factor is 3.0  and the  temperature correction
factor is 3.5, the resulting combined factor  is 10.5,  which can
result in very high emissions.

       There  are a  great  many   factors  which can affect  the
amount of CO  emissions produced  under  these low speed  and low
temperature conditions, such as  choke  action, extent  of  choke
pull-off,   idle   speed,  rate  of   engine   warm-up,   catalyst
light-off time,  etc.   Since CO emissions  are sensitive to both
speed  and temperature,  EPA should test some cars at low speeds
and temperatures to determine the interactions.

4.4    Effects of Repairs

       Cars   with   the  highest  as-received  emission  levels
generally receive  maintenance  (called  restorative  maintenance,
or  "RM")  and an  after maintenance FTP  test.  The  historical
objective   for   RM  is  to  confirm   our  diagnosis   of  the
malfunctioning  systems:   once  repaired,   the  car's   emission
levels should significantly drop.  If  they  do not, we  know we
have not yet made  the complete diagnosis  and  further  RM is
generally performed, depending on the  continued availability of
the car and/or other resource considerations.

       Recently,  however,  two  other   objectives   for  RM  have
emerged.  One has been the need to  characterize  the emission
reductions  associated with specific  three-way system repairs.
This is  in support  of  a   preliminary  effort  to   estimate  the
benefits of requiring on-board diagnostics  on all  cars.  These
systems are probably most  readily integrated with  the  three-way
system and computer.

       Emission benefits on cars  receiving 3-way system repairs
are  clearly  higher  than   on  those  receiving  only  non-3-way
system  repairs  as  the  data  in  Table  4  demonstrate.    The
limitations of these  data  are that many  of  the cars  receiving
3-way   repairs    also   received   some    non-3-way    repairs
simultaneously,  so  it  is  difficult to determine 'the emission
benefits  of  specific  3-way repairs.   For  this  reason we  are
implementing  a  step-wise  RM procedure  with cars which qualify
for RM.   Preference will  be given to repair 3-way items first.

-------
                                16

A  retest will  be performed,  and  then  additional  non-3-way
repairs  if  necessary.   We  are also  using a  short test  (the
restart idle test) as a "flash" estimate  for evaluating whether
repairs have been effective.

       The second  new objective in  KM is to determine effects
of field  fixes  on  "pattern case"  failures.   Pattern  case  cars
are those that  experience  an abnormally  high I/M failure rate,
usually because of a common problem.   These cars may or may not
have high FTP emissions.   In some cases, the manufacturers have
suggested field fixes for  these cars so they won't continue to
fail I/M tests.  Currently, we  do not know what the FTP effects
of these field fixes  are.   EPA should  do some testing  on  cars
which have received field fix repairs.

4.5    Disablement/Misfueling Testing

       EPA does disablement and misfueling testing to determine
the effects of tampering  on in-use emissions.   The  analysis of
the  effects  of  tampering  that  was  performed  prior  to  the
release of  MOBILES identified  areas  where  additional  testing
was  needed.    This   testing   included  disablement   testing
(primarily catalyst removal) at high altitude,  and  additional
misfueling studies.   That  testing is  nearly  complete  and the
results should be  summarized in the next few months.   This may
identify  additional disablement or  misfueling testing  that is
needed for cars.

       The  effects  of   tampering  on   truck   emissions  were
borrowed   from   cars  with   some   modifications.    However,
disablement testing should be  extended at least on  a limited
basis to  light-duty trucks, particularly in  light of  the  fact
that tampering rates are higher for trucks than for cars.

-------
                                17
                                   Table. L

            MOBILES  Fleet VMT and Emission Fractions:
                       1990 Calendar Year
Vehicle
Type
LDGVs
LDGTs
HDGVs
LDDVs
LDDTs
HDDVs
MCS

Fleet VMT
Fractions
0.635
0.201
0.041
0.046
0.021
0.049
0.007
1.000
Emission Factors
(g/mi)
HC CO
2.28
4.67
6.95
0.40
0.62
3.51
6.01

18.44
35.93
66.48
1.32
1.53
11.11
19.73

Emission Fractions
HC CO
0.50
0.32
0.09
0.01
0.0
0.06
0.02
1.00
0.52
0.32
0.12
0.0
0.0
0.03
0.01
1.00
Note:  Emission fractions  are  estimated by multiplying emission
       factors  by  fleet  VMT  fractions,  adding these products
       together to  get  the fleet emission  factor,  and dividing
       the weighted emission factor  of each vehicle type by the
       fleet emission factor.

-------
                                18
                            Table 2
                 LDGV Fleet HC and CO Emissions
                     at Varying Temperatures
Calendar
Year
1985

2000



Pollutant
HC
CO
HC
CO
(g/mi)
(g/mi)
(g/mi)
(g/mi)


50°
4
40
2
26
.32
.39
.31
.40



75°
3
27
1
13
.48
.52
.60
.66
4
50
1
15

100°
.21
.09
.79
.51
Source:  MOBILES, default operating mode percentages,
         speed =19.6 mph

-------
                   19
               Table 3

    CO Emissions  in g/mi of LDGVs
            at Low Speeds

                         CO,  g/mi
    Speed             75°F       30°F

  3 mph               159        305
  4 mph               121        231
  5 mph                96        181

Source:Low-speed MOBILES. Default operating
         mode percentages used.

-------
                                20
                                      Table  4

                           HC and CO Emission Benefits of
                             3-Way System and Non-3-Way
                                   System Repairs
    Category

Carbureted Cars
  No 3-way System Repair
  3-way System Repair

Fuel Injected Cars
  No 3-way System Repair
  3-Nay System Repair

NOTE: No 3-way  system repairs:  could have received multiple  repairs,  but none to
      3-way system.
      3-way system repairs: could include some non-3-way system repairs also.


N
144
48
26
21

Bef
Rep
1.
2.
0.
2.

ore
air
36
43
95
68

After
Repair
0.81
1.28
0.76
0.91


Reduc .
0.55
1.15
0.19
1.78

Be
Re;
18
47
11
51

fore
pair
.32
.78
.50
.86

After
Repair
10.49
23.76
8.80
10.16


Red
7
34
2
41


uct
.83
.02
.70
.70

-------
   20
i.
   10
                              Figure 1
                  Distribution of 1981+ Carbureted
                      Cars by Odometer Values

4.78
1
18.08
^
>>
y>
>>
>>
>>
>>
2x
13.04
1
9.27
!
14.2
i
13.04
I
13.91
i
9.42
!

|Corto.N»690|
3.78
W t19 U
         0-10  -20 -30  -40  -50  -60  -70 -80  -90  -100 100+
                       Mileage in Thousands of Miles
  35
  30-



  29
   10
                             Figure 2
                 Distribution of 1981+ Fuel Injected
                     Cars by Odometer Values
             3t41
        9.13

24.77


I
                               |nnj.N =
 14.8

I
                           9.08
                                4.93  4M
        0-10  -20  -30 -40  -50-60-70-60  -90  -100 100+
                      MilMfe in Thousand* of Mile.

-------
                         Figure 3
     Distribution of Travel fractions and HC and CO Emissions
               by Age Group, CY 1988 —  LDGVs
      w-
   •8
                                      CZHCCminiefM
                                      • COCmMm
      20
      10
      0-
               0-5     5-10     10-15
                         Age Group
15+
   180000-
X. 140000 H
 01
 a

1
 01 100000-
    60000-
    80000
                         FIGURE 4
             CO YEARLY EMISSIONS vs. AGE
           t  2  3  4 » • 1 » 9 W U U U 14 15 IS 17 18 19 20

-------
                       FIGURES
          TEMPERATURE CORRECTION FACTORS
       LIGHT DUTY GASOLINE POWERED
               BAG 1 - HYDROCARBONS
                     Low Altitude
   14
   12-
e  10-
§
OS
§
w
05.
os
w
Du
    8-
    2-
         198H
         1980
 1968-69

 1975-79

 1970-74

Pr«-1968
                                       1981+


                                       1970-74
                                       Pr«-1970
                                       1975-80
               10   30   50    70   90   110
               AMBIENT TEMPERATURE (F)
                                          •log
                                            i
                                          •8
                                          14
                                          12
  o
8 QS
  1
  OS
                                                •4
                                          •2

-------
                        FIGURE 6
           TEMPERATURE CORRECTION FACTORS
        LIGHT DUTY GASOLINE POWERED
                BAG 2 - HYDROCARBONS
                      Low Altitude
   2.6
2.4 A
2.2 H
HM
OS
§
Ed
B
§
05
Ed
Ed
         1980-f
        1975-79
1.4 A
    1968-69

    Pr«-1968

     1970-71
     1972-74
                                         197S-80
                                          1968-69
                                          1972-74
                                          Pr«-1968
                                          198H
                                          1970-71
 0.8-
                10    30   SO   70   90   110
                AMBIENT TEMPERATURE (F)
                                                2.6
                                                   2.4
                                                   2.2
                                                 2  K
                                                   I
                                                   6
                                                 1.8 Ed
                                                   o:
                                                hl.6 Ed
                                                hl.4
                                                    CU
                                                    S
                                                    Ed
                                                    0.8

-------
                 FIGURE?



          BAG 3 - HYDROCARBONS
               Low Altitude
UJk.
0.8
           10   30  50   70   90  110
           AMBIENT TEMPERATURE (F)

-------
                      FI6UBE8
   8
   7-
   
-------
                        FIGURE 9
           TEMPERATURE CORRECTION FACTORS
        LIGHT DUTY GASOLINE POWERED
               BAG 2 - CARBON MONOXIDE
                      Low Altitude
   5.5
    5-
   4.5-


-------
                       FIGURE 10
          TEMPERATURE CORRECTION FACTORS
        LIGHT DUTY GASOLINE POWERED VEHICLES
               BAG 3 - CARBON MONOXIDE
                     Low Altitude
   3.5-
    3-
   2.5-

CIS
§
OS
W  l'8
£

I
    1-
   0.5
        1972-74
1975-80




1970-71



1972-74

1968-69
                                        Pr«-1968

                                        198H
               10   30   50   70   90   110
               AMBIENT TEMPERATURE (F)
                                                 •2.5
          3.5
          ••  §
             o
             sa
             g
             O,
             2
             w
          -1
          0.5

-------