Final Regulatory Impact Analysis and
    Summary and Analysis of Comments
 Control of Vehicular Evaporative Emissions
              February 1993
   U.S. Environmental Protection Agency
         Office of Air and Radiation
          Office of Mobile Sources
Regulation Development and Support Division
   Engine and Vehicle Regulations Branch

-------
    Final Regulatory Impact Analysis and
    Summary and Analysis of Comments
 Control of Vehicular Evaporative Emissions
              February 1993
   U.S. Environmental Protection Agency
         Office of Air and Radiation
          Office of Mobile Sources
Regulation Development and Support Division
   Engine and Vehicle Regulations Branch

-------
                             Table of Contents
Chapter 1 Introduction	   1
      A.    Principles of Evaporative Emission Control  	   1
      B.    Overview of Proposals  	   3

Chapter 2 Test Procedures and Standard	   7
      A.    Sequence of Test Segments  	:	   8
      B.    Emission Standards	  15
      C.    Diurnal Emission Test	  20
      D.    Running Loss Test	  23
      E.    Preconditioning  	  32
      F.    Hot Soak Test	  37
      G.    Exhaust Emission Test  	  38
      H.    Heavy-Duty Vehicles and Engines	  39
      I.    Fuel Spitback	  39
      J.    Methanol-Fueled Vehicles  	  42
      K.    Other Issues  	  43
      L.    Comparison to CARB's Adopted Test	  45
      M.   Adjustments to Test Tolerances for EPA Testing	  53

Chapter 3 Technological Feasibility and Lead Time  	  55
      A.    Technological Feasibility	  55
      B.    Lead Time	  58

Chapter 4 EconoT"^ Tmpnrt	  63
      A.    Vehicle Hardware Costs	  64
      B.    Development and Production Costs  	  69
      C.    Overall Vehicle Lifetime Costs	  71
      D.    Fuel Dispensing Nozzles  	  72

Chapter 5 Environmental Impact	  75
      A.    Methodology	  75
      B.    Baseline Emissions  	  75
      C.    Emission Reductions 	  77
      D.    Projected Emission Factors	  87
      E.    Total Nationwide VOC Emission Reductions 	  89

Chapter 6 Cost-effectiveness	  91
Appendix A  Evaporative Modeling with In-Use Driving Patterns

Appendix B  MOBILES Input and Output Files

-------
                           Chapter 1  Introduction
      EPA's  concern  regarding  the  control of volatile organic compound  (VOC)
emissions has grown over the years as exceedances of the health-based ozone
standard have continued to be a problem in many areas.  On hot, sunny days VOC
emissions react  in the  air to form ground-level ozone, which causes respiratory
problems and is  associated with urban smog. Based on the most recently available
information for  1989  to 1991, there are 97 areas that fail to meet the National
Ambient Air Quality  Standard for ozone (0.12 parts per million).1 According to
ozone monitoring data, based on 1991 only, 70 million people continue to live in U.S.
counties exceeding the ozone standard.  Evaporative emissions from motor vehicles
are a significant source of VOC's and, as a result, EPA has initiated action aimed at
reducing these emissions.

      In addition, the 1990 amendments to the Clean Air Act direct EPA and the
states to carry out new programs to reduce levels of tropospheric ozone, especially in
urban areas. With respect to evaporative emissions, the amended Clean Air Act
states in §202(k):

      The Administrator  shall promulgate (and from time  to time revise)
      regulations applicable to evaporative emissions of hydrocarbons from all
      gasoline-fueled  motor vehicles—
            (1) during operation; and
            (2) over 2 or more days of nonuse;
      under ozone-prone summertime conditions (as determined by regulations
      of the Administrator). The regulations shall take effect as expeditiously
      as possible and shall require the greatest degree of emission reduction
      achievable by means reasonably expected to be available for production
      during any  model year  to  which the regulations apply,  giving
      appropriate consideration to fuel volatility and to cost, energy and safety
      factors associated with the application of the appropriate technology.
A.    Principles of Evaporative Emission Control

      In 1971  EPA began  testing motor  vehicles for  evaporative emissions by
subjecting  test vehicles to  typical drive and park conditions.  The evaporative
emission test procedure, which has changed little since then, measures diurnal and
hot soak emissions. These emission sources, as well as the more recently identified
sources of running losses  and  resting losses,  are  described  in the following
paragraphs.
    lMNational Air Quality and Emissions Trends Report, 1991," EPA, October 1992

-------
      Diurnal emissions occur during periods when a vehicle is not in operation.
They result from the heating of a vehicle's fuel tank in response to daily increases in
ambient temperature.  Emissions result when high ambient temperatures cause a
buildup of fuel vapors and eventually cause vapor venting out of the tank.  Current
vehicles are designed to prevent vapors from reaching the atmosphere by venting
them to an evaporative canister. The canister stores the vapors in a bed of activated
carbon until the vehicle is driven. At that time the engine draws ambient air through
the canister,  carrying the stored vapors to the engine to be burned as fuel.  If a
vehicle has enough storage capacity for fuel vapors generated in use, and can restore
that capacity with regular driving, it can maintain control of diurnal emissions.

      Hot soak emissions happen immediately after a vehicle has been driven, due
to residual fuel tank heating and the high  temperatures of the engine and fuel
system. As with diurnal emissions, hot soak emissions are controlled by storing
vapors in an evaporative canister and subsequently purging them to the engine.

      Running losses are evaporative emissions that occur while a vehicle operates.
They represent a greater in-use emission source than previously believed. Running
losses are caused by the generation of vapors from the fuel tank as the fuel is heated
during driving. This heating comes from several sources, including hot pavement, hot
surfaces of the engine and exhaust systems, recirculated fuel that has been heated
by the engine, and fuel pumps that are built into fuel tanks. Running losses can be
prevented by maintaining enough vapor purge to the engine so the vapors can be
consumed before they overfill the canister and reach  the atmosphere.  Techniques
that reduce vapor generation during driving also can contribute to running loss
control; for example, insulating or isolating fuel tanks from heat sources, increasing
underbody airflow, or using fuel systems that do not return fuel to the tank, can
contribute to reduced vapor generation.

      A fourth source of evaporative emissions, resting losses, has been more recently
identified.  Resting losses  are composed of fugitive  vapors that result from the
migration of fuel from the evaporative canister, and permeation through joints, seals.
and polymeric components of the  fuel system.  Resting losses can be controlled by
improved design of evaporative canisters and by material selection to prevent fuel
permeation.

      Control of evaporative emissions is fundamentally different  than control of
exhaust emissions.  Because exhaust emissions from vehicles are largely controlled
by engine and fuel management controls and their associated catalytic converters.
vehicles tend to provide proportional control of exhaust emissions over a wide variety
of conditions. Control is lost abruptly only if a control system component fails (such
as the catalyst) or the engine is operating  in an override mode (e.g., wide-open
throttle). A fully functioning evaporative system in normal operation can experience
a loss of control if the canister is simply overfilled with vapor.  Also, changes i n
exhaust emission standards result in incrementally reduced in-use tailpipe emissions
In contrast, the current standard for evaporative emission testing was intended t ..

-------
reflect the expectation that no fuel vapors will be emitted during the test.  The
nonzero standard for the evaporative emission test was intended as an allowance for
small nonfuel hydrocarbon emissions (e.g., from vinyl surfaces, paints, and other
polymers) and for test-to-test variability.

      Because changing the standard is not an effective means of improving the
performance of vehicle evaporative controls, EPA has focused on revising the test
procedure to expose the  vehicle to more challenging conditions typical of in-use
operation,  as required by  section 202(k) of the Clean Air Act.  When vehicles
temporarily lose evaporative emission control, emissions can be very high, several
grams per mile in some cases.  Evaporative emissions come disproportionately from
relatively infrequent experiences of temperatures and driving patterns when control
is most challenging. Such  conditions include high ambient temperatures, a small
amount of driving between  extended periods of nonuse, or several consecutive days
without driving.  Thus, increasing the stringency of evaporative emission testing
requires changing the test procedure to address these kinds of in-use conditions (see
Clean Air Act section 202(k)>.

      An effective test of evaporative emission control systems  needs to include a
sequence of three fundamental elements:  an initial loading of the  evaporative
canister, a period  of driving for opportunity to purge the canister,  followed by a
simulation of parking over a series of hot days. Sampling for emissions during the
final "day" ensures first that the vehicle can quickly regain storage capacity  during
driving, and also that the canister's total capacity is sufficient. A rigorous test with
this sequence of test segments provides assurance that all sources of evaporative
emissions are controlled.

B.    Overview of Proposals

      EPA initially proposed to change the evaporative emission test in August 1987
by adding a step to load the evaporative canister with vapors at'the beginning of the
test (52 FR 31274, August 19, 1987).  This change was prompted by emission data
showing that vehicles passing the current test with  an empty canister  during
certification were often unable to meet the standard with more varied in-use canister
conditions. EPA believed that initially loading the canister in the test would ensure
that  a  vehicle could,  with limited  driving,  purge vapors from the canister  in
preparation for the next loading experience.  Vehicles designed to meet those
requirements would be expected to maintain their level of control, regardless of how
much vapor was in the canister at any time.

      Based on subsequent analyses, EPA concluded that this change was insufficient
in itself to accomplish the goal of achieving acceptable in-use evaporative emission
control. Additional provisions were needed to increase the level of diurnal emission
control  and to ensure that running losses,  which  had since  been found to be
unexpectedly high, would be prevented in use. EPA held a public workshop in June

-------
1988 to discuss new information on these issues and published a new proposal on
January 19, 1990 (55 FR 1914).

      EPA proposed in the January 1990 Notice of Proposed Rulemaking (NPRM) to
keep the canister loading step at the beginning of the test procedure and to add two
high-temperature heat  builds after  the exhaust  emission  test.   Rather than
incorporate a running loss test, the January 1990 proposal had provisions that were
meant to ensure that running losses would rarely occur.  These provisions included
the initial canister loading, an engineering design review to ensure that vapors would
not be vented to the atmosphere during operation, and a "cap-off* requirement at the
beginning of the hot soak test to encourage low-pressure fuel tank designs.

      Shortly before EPA published the new proposal, General Motors (GM) proposed
a test concept it believed would be an improvement over the proposed test.   GM's
test differed  from that proposed in the NPRM, first by changing the method of
conducting  diurnal heat builds to a "real time" approach approximating outdoor
ambient cycles.  Real time diurnal heat builds were to be conducted by exposing the
whole vehicle to ambient temperature changes in 24-hour cycles, rather than by the
conventional  approach of repeatedly heating the freshly fueled tank with a local
heating element. Also, the GM proposal added a 70-minute, high-temperature
running loss test, placed between the exhaust and diurnal emission tests. Finally,
the hot  soak test was moved to follow the running loss test.

      In the January 1990 NPRM, EPA requested written comments on both EPA's
and GM's proposed test procedures.  At that time other vehicle manufacturers, citing
impacts that such a change would have on facility  requirements and the time
required for testing, did not support GM's proposed method of testing for diurnal
emissions.  Nearly all manufacturers, however, supported direct testing for running
losses instead of EPA's proposed design review and "cap-off* requirement.  Also, a
nearly universal request from the vehicle manufacturers was for EPA to work with
the California Air Resources Board (GARB), which proposed and eventually adopted
a test based on GM's proposed test sequence, to adopt a common test procedure.

      In response to the comments received following the January 1990 NPRM, EPA
published a notice requesting comment on several possible  modifications  to the
proposed test procedure (55 FR 49914, December 3,1990). In the modified procedure
the duration of the diurnal heat builds was extended from one to two hours, and a
third diurnal heat build was added. The design review and "cap-off* requirement
were replaced with a running loss test, very similar  to GM's, which would be
conducted after the series of diurnal heat builds. The preconditioning sequence was
rearranged so that the evaporative canister would be manually purged and  loaded
   2GM formalized its proposal in a letter from Lisa M. Fior to Tad Wysor, March
26, 1990 (Docket A-89-18, item IV-D-19).

                                     4

-------
with butane to IVa times its working capacity just before the exhaust emission test.3
Finally, a 4-hour period was added to the hot soak test, which was placed after the
new running loss test,  to measure resting losses.  EPA held a public workshop on
December 19, 1990 to discuss these changes and accepted written comments until
February 22, 1991.

      Finally, on December 17, 1991, EPA announced a public workshop to discuss
the analysis supporting its position on the previously proposed sequencing of test
segments (56 FR 65461). At that time the public also had opportunity to comment
on draft regulations containing the entire set of test procedures.  The test procedure
described in the draft regulations included diurnal emission testing by GM's real time
method, but was otherwise consistent with the modified procedure described in the
December 1990 notice.

      During the development of the final rule, EPA has incorporated most of the
substantive revisions to the proposed test suggested by GM and other commenters.
Most of the revisions,  however,  have been made to improve the simulation and
repeatability of testing, rather than changing the fundamental test requirements.
The expectations for basic hardware and vehicle configurations needed to meet the
test requirements have changed little since the January 1990 NPRM. EPA believes
the resulting test procedure will  ensure  that evaporative emission controls will be
designed to eliminate evaporative emissions for nearly all in-use events, including
those likely to occur under ozone-prone summertime conditions, as required by Clean
Air Act section 202(k).

      EPA's new test for evaporative emissions is based on the procedures proposed
by GM and adopted by GARB.  The EPA test  also contains a  supplemental test
requirement, which is necessary  to ensure adequate purge capacity, so that in-use
evaporative emissions are, in fact, controlled. GARB has similarly interpreted its
existing test as requiring a demonstration that control systems have adequate purge
capacity to control in-use evaporative emissions. EPA's final evaporative emission
test procedure is, therefore, basically an extension of CARB's current test procedure
(adopted in August 1990) to the rest of the nation.

      Chapter 2 provides a summary and analysis of comments related to the  test
procedure.  Chapter 3 evaluates  the technological feasibility of compliance and the
schedule for implementation. Chapters 4, 5, and 6 present EPA's detailed analyses
of the costs, emission  reductions,  and cost-effectiveness associated  with the new
procedure for testing evaporative emissions. A  description of a recently developed
computer model for evaluating evaporative emissions over a wide range of in-use
    For the purposes of the test procedure, the working capacity is the amount of
vapor that a canister, starting from a purged condition, would retain in loading to the
2-gram breakthrough point (that is, 2 grams of vapor emitted from the canister)

-------
driving conditions, as well as modeling  results relevant to this rulemaking, are
included as appendices.

-------
                   Chapter 2 Test Procedures and Standard
      Pursuing the goal of improved evaporative emission controls has significantly
broadened the understanding, both in the Agency and in industry, of the nature of
evaporative emissions and the means of their control. The Agency has benefited from
extensive public participation in this rulemaking.  Many aspects of the rule reflect
the input of outside participants.   Vehicle manufacturers, individually and in
conjunction  with  the  oil industry, have  made  valuable contributions to  the
understanding of evaporative emissions.  EPA has also worked very closely with
GARB,  which  has concurrently  developed and  adopted  its own  revisions to
evaporative emission testing requirements.5

      To better understand EPA's approach to resolving the individual issues raised
during the interaction with the public, it is important to identify some fundamental
differences between EPA's and GM's approaches to designing an evaporative emission
test procedure.  The following discussion expands on these basic differences.

      First, GM has promoted some of the changes to its proposed test procedure by
focusing on how they better represent the experience of an average vehicle in the in-
use fleet. The  amount  of driving  before the diurnal emission  test is the  most
important area where GM wanted to maintain a vehicle's "typical" experience in the
test. GM wanted to allow as much driving between the initial canister loading and
the diurnal emission test as an average car would experience in a full day.

      GM's focus on "typical" conditions is not consistent with the statutory mandate.
which is to control evaporative emissions to the greatest degree reasonably achievable
"under ozone-prone summertime conditions," including two or more days of nonuse.
Any test for evaporative emissions must be judged against this standard.  EPA's goal
in designing  a test is  therefore not to simulate a single, "representative" in-use
condition. Clearly, any specific procedure will only simulate one of a multitude of
actual  in-use patterns of operation.  The broader goal of EPA's test design is to
develop a test that  will result in  good  emission performance under nearly all
conditions that vehicles will experience in use (see Clean Air Act section 202(kn
Designing the test based on average conditions is inappropriate, because the resulting
vehicle designs would be incapable of performing well under the temperature and
driving conditions when high evaporative emissions are most  likely to occur and
control is meet  needed.  For example, EPA's key point of contention  with  GM's
   4The Air Quality Improvement and Research Program, undertaken jointly by the
auto and oil industries, has included extensive development of new methods to test
evaporative emissions.

   5In  August  1990  CARS  adopted  revised  test  regulations for controlling
evaporative emissions (Docket A-89-18, item IV-D-83).

-------
proposed procedure is that the "representative" amount of driving allowed between
the initial canister loading and the diurnal emission test, involving about  100
minutes  of driving over various patterns, would provide  an inadequate purge
requirement for many in-use driving scenarios, as discussed below.

      Second, GM has emphasized that in several ways its  test simulation would
more precisely duplicate the physical phenomena that a vehicle experiences in use.
GM claimed that these changes to the test would have the effect of improving the
accuracy and repeatability of the test results. For example, exposing a whole vehicle
to a diurnal heat build would allow the process of vapor generation in the tank and
adsorption of those vapors in the evaporative canister to occur at natural rates.

      Duplicating as completely  as possible the physical phenomena involved in a
vehicle's  in-use experience  is not, in itself, a primary goal  in EPA's approach to
designing a test procedure.  EPA would like to avoid expanding test requirements to
improve the test simulation (with consequent administrative burden and straining
of Agency resources) if there is no expected  positive  effect on vehicle design.
However, as  in  the example of the diurnal test  method discussed below,  other
considerations may be involved in EPA's choice of a final test procedure.

      In addition to these  general issues, participants had specific comments on
many aspects of the new test requirements. Following is a summary and analysis of
these comments, grouped by major topic area.

A.    Sequence of Test Segments

EPA Proposal

      The test sequence described in the December 1990 Federal Register notice
established the driving time between the canister loading and the diurnal emission
test  at about  30 minutes,  the amount driven during the exhaust emission test
(Figure 2-1).  This test sequence involved canister loading just before the  exhaust
emission test, which was in turn followed by the diurnal emission test.  A new
running loss test was added at the end of the test sequence, not affecting the driving
time before the diurnal emission  test.

      EPA's proposed test  sequence was different than that finalized by GARB
GARB's adopted procedure included a 70-minute running  loss test between  the
exhaust and diurnal emission tests, allowing a total of approximately 100 minutes of
driving for vehicles to purge their canisters.

Summary of Cornrnent3

      Manufacturers objected to EPA's proposal, arguing primarily that the proposed
test  sequence, compared to in-use driving patterns, represented  a rare and rather
extreme  scenario of vehicle operation.   Manufacturers claimed that because th**

                                     3

-------
                 Figure 2-1
         Basic Elements of GARB
    and EPA Proposed Sequences
Key Elements
  LOAD
  DRIVE/
  PURGE
 VERIFY
CAPACITY
                    CARB
Load/Condition
  Canister
 Exhaust Test
(30 min. drive)
                Running Loss Test
                 (70 min. drive)
                  EPA Proposal
Load/Condition
  Canister
 Exhaust Test
(30 min. drive)
                    Multiple
                   'Real time"
                    Diumals
                                 Running Loss Test
                                   (70 min. drive)

-------
specific drive-park sequence represented in the test would so rarely happen in real
driving, EPA's  approach was  invalid.   They reinforced their  position with the
observation that EPA's MOBILE model estimates that vehicles from the in-use fleet
average approximately 30 miles (48 km) of driving per day, much more than the 11
miles (18 km) of driving for the exhaust emission test.6  Most auto manufacturers
recommended that EPA adopt CARB's test procedure.

      After considering EPA's technical objections to its procedures, however, GARB
acknowledged that its procedure could lead to inadequate purge during short trips.
In a March 1992 letter, GARB thus stated,

      As written, the [GARB] procedure may not necessarily ensure adequate
      purge during short trips, and canister saturation is a possibility. This
      could  occur even on a vehicle which would pass the ARB procedure.
      ARB  and Environmental  Protection Agency (EPA)  analyses have
      confirmed significant in-use emissions benefits from requiring adequate
      purge during the exhaust testing portion of the current test/

To address these concerns, GARB suggested in the same letter  adding one of two
alternative methods to verify purge during the exhaust emission test. In the first
method, purge airflow would be measured and compared with a similar measurement
during the running loss test to verify a consistent purge rate. In addition, the change
in canister mass during the exhaust emission test would be measured to ensure that
approximately 70 percent of the canister's working capacity before breakthrough had
been made available. In the second method, GARB would conduct a special test with
two diurnal heat builds directly following the exhaust emission test. GARB proposed
that these additional test requirements would apply to certification and, potentially,
in-use testing.  In  addition, GARB  identified  the possibility of adopting EPA's
proposed test sequence if its other proposed changes were found not to be viable.

      Auto manufacturers had varying responses to  CARB's proposed approaches.
Some argued that current language in EPA rules that prohibits defeat devices would
be effective in ensuring sufficient purge  under  GARB's adopted  test.  These
manufacturers  suggested a requirement  to state at certification  that  they had
employed no defeat devices in designing their purge strategy.

      Manufacturers opposed CARB's suggestion of weighing canisters during a test
run.  They commented that such an operation could jeopardize the  repeatability.
reliability, and  validity of test results because of the need to remove and handle
components  of a vehicle's emission control system.
   6The series of MOBILE models is used to characterize the emission behavior of
the in-use fleet and to estimate the effectiveness of various control programs.

   7CARB Mail-Out #92-13, March 6, 1992 (Docket A-89-18, item IV-D-84).

                                     10

-------
      Commenters who did not object outright to the idea of a purge-verification
strategy  generally supported  the  concept  of measuring purge  airflow.   These
commenters noted that measuring purge airflow would be the least burdensome
strategy, and would give a direct measure of purge behavior.  Various formulas for
specifying a purge requirement were discussed.

      Ford and Chrysler  came  forward with nearly identical approaches  for a
potential compromise, consistent with CAKB's proposed option for  a special two-
diurnal test to ensure sufficient purge in short-trip driving patterns.8'9  Ford and
Chrysler recommended that EPA finalize CARB's adopted procedure, with minor
modifications, for certification  testing.   For recall  testing, they  suggested  an
abbreviated test, consisting of the preconditioning and exhaust emission test, followed
by a moderate-temperature hot soak test, and two diurnal heat builds.  Since Ford
and  Chrysler offered no explanation of the differences for recall  and certification
testing, EPA understands that they were merely responding to EPA's  desire to adopt
an enforceable in-use test that would ensure adequate purge rates.  The standard for
recall testing would be 2.5 grams for vehicles with fuel tank  capacity less than 30
gallons (110 liters), and would allow for exclusion of nonfuel emissions. Vehicles with
larger fuel tanks would be subject to a 3-gram standard. No explanation of the basis
for these relaxed standards was stated.

      GM  opposed the use  of  any alternate emission  measurement to  verify
purge.10  GM claimed that  the alternate procedures under consideration would
overburden the industry and increase the severity  of the  full  evaporative test
procedure.  GM claimed, though without explanation, that an alternate emission
measurement, with the existing 2-gram standard, would increase the overall  purge
requirement by 25 percent—with no air quality benefit.

      Several commenters recommended a streamlined version of CARB's adopted
procedure to facilitate EPA's in-use testing (e.g., see GM's March 23,1992 letter, page
11).  Commenters suggested driving through the running loss test without measuring
evaporative emissions to avoid installation of thermocouples and to prevent the need
for running loss measurement facilities.  Significant fuel heating (and thus  vapor
generation) would be prevented by holding ambient temperatures at  80° F (26.7° C >
and circulating air around the fuel tank.
   8Letter from Gordon E. Allardyce, Chrysler Corporation, to Docket A-89-18, March
23, 1992 (item IV-D-76).
          from Donald R. Buist, Ford Motor Company, to Richard D. Wilson, EPA.
March 27, 1992 (Docket A-89-18, item IV-D-77).

    10Letter from Samuel A. Leonard, General Motors, to Richard D. Wilson, EPA.
March 23, 1992 (Docket A-89-18, item IV-D-78).

                                     11

-------
Analysis
      After considering all of the comments, EPA still believes that CARB's adopted
test procedure, by allowing 100 minutes of driving time to purge the evaporative
canister, does not ensure effective emission control.  Most importantly, the majority
of the driving time, and therefore purging time, in CARB's test occurs when there is
no measurement of exhaust emissions. Vapors purged from the canister during the
running loss test could simply pass unburned out the vehicle's tailpipe as exhaust
emissions, without detection.  CARB's test sequence thus gives  manufacturers an
important incentive to minimize the amount of purge during the early part of the
test's  driving time, when exhaust emissions are measured. An inadequate purge
requirement would result in reduced evaporative control effectiveness for vehicles
experiencing mostly short trips, and could also cause increased exhaust emissions in
use, compared to today's vehicles.

      In addition, CARB's adopted procedure would be very difficult to use as the
exclusive test for in-use enforcement for three reasons.   First, CARB's adopted
procedure would  require that a full  running loss test be  conducted before every
diurnal emission  test.  EPA believes that the diurnal emission  test is of primary
importance in verifying the key parameters of canister purge and storage capacity.
EPA expects that the resource-intensive running loss test can be reserved for vehicle
designs with  higher  vapor  loads  to the  engine, such as those with high  fuel
temperatures during driving. CARB's adopted test would remove this flexibility, and
would require a greater investment in running loss facilities, significantly increasing
the cost and effort of testing.  Second, some of CARB's running loss test specifications
are very difficult to maintain, increasing the likelihood of invalid tests.  This would
also apply to certification confirmatory testing. Third, in-use vehicles would likely
need to have fuel tanks removed for installation of thermocouples for the running loss
test.  Thermocouple installation is a time-consuming procedure, and may call into
question the validity of test results if installation affects the integrity of the vehicle's
emission control system.

      EPA believes that its proposed test, with three diurnal heat builds following
the exhaust emission test,  is a feasible requirement that would achieve good in-use
control.  EPA has evaluated the  emission  benefits of its  proposed  test sequence
relative to CARB's.  This evaluation is described in a draft technical report and was
the subject of the January 1992 public workshop.11 The draft report concluded ( as
noted above) that CARB's test had so much driving time before the diurnal emission
test that manufacturers could substantially delay purging. Refinements made to the
analysis, described in Appendix A, only reinforce that concern. If vehicles designed
for CARB's adopted test delay purging, in-use emissions may actually increase from
current levels, contrary to the requirements of Clean Air Act section 202(k) (or section
    ^'Emission Evaluation of the GM Real Time Evaporative Test Procedure," drart
EPA report by Julie Hayden, September 25, 1991 (Docket A-89-18, item III-B-2)

                                     12

-------
202(a) for methanol-fueled vehicles).  The analysis shows that these vehicles would
perform poorly in use, because many in-use driving patterns involve short trips with
less driving time than is present in CARB's adopted test procedure. In comparison,
the analysis shows that vehicles designed to pass EPA's proposed test sequence with
three diurnal heat builds would almost completely control emissions for a wide range
of in-use driving patterns.

      EPA  has,  however,  made a  concerted  effort to  achieve  common  test
requirements for federal and California-only vehicles, within the constraints of its
legal  obligation  under section  202(k) of the Act.   EPA has considered possible
modifications to the GARB procedure to ensure effective in-use emission control,
while addressing manufacturers' expressed concerns about the relative stringency and
associated costs of test options, and the desirability of avoiding the expense and
administrative complication of maintaining different federal and California-only tests.
The following discussion evaluates the various proposed or suggested modifications
to CARB's test.

      Merely relying on existing requirements aimed at preventing defeat devices,
as suggested by some commenters, is insufficient to ensure adequate emission control.
Most  participants, including CARB (particularly in its March 6, 1992  letter), have
acknowledged that CARB's adopted test sequence allows manufacturers flexibility
that could result in poor in-use performance.  Defeat device regulations rely on a
subjective evaluation of designs to identify  possible defeat devices.   As much as
possible, the test itself should ensure effective in-use performance and so avoid the
need for such subjective inquiries.  Moreover, this is the  Agency's legal mandate
under section 202(k).

      The various suggested improvements to CARB's adopted test sequence are also
not satisfactory.  Measuring  a change in canister mass during the exhaust emission
test is an inappropriate way to verify purge during short trips. Any requirement for
a change in canister mass would effectively be a design standard,  because it would
dictate requirements for certain vehicle components rather than demonstrating the
vehicle's performance to an emission standard.  EPA strongly prefers  performance
standards over  design  standards because  design  standards  can unnecessarily
constrain manufacturers' design options, and may not be effective in improving in-use
performance in that they may not address possible unforeseen mechanisms by which
emissions occur. Also, the removal of a canister to determine its mass change would
involve an unnecessary intrusion  into the control system, both before and after the
exhaust emission test.

      Measurement of purge airflow is  also  an inappropriate way to verify purge.
Requiring some specified distribution of purge in different driving conditions would
effectively be a design standard,  and therefore not a preferred alternative  for the
reasons just noted.  Also, there is an enormous degree of latitude in defining the
criterion for acceptable purge distribution, so that setting such a criterion would
require a subjective evaluation of what constitutes an optimum strategy, to the

                                      13

-------
exclusion of other reasonable strategies. The nature of design standards virtually
ensures that any  such criterion would either be ineffective in ensuring in-use
emission control, or would unnecessarily restrict manufacturers' flexibility in vehicle
design, or both.  EPA believes the goals of establishing an effective, yet nonrestrictive
purge flow criterion are irreconcilable, as evidenced by the fact that GARB has been
unable to reach an agreement with manufacturers.  Measurement of purge airflow
may also require temporary, intrusive vehicle modifications that could impact vehicle
evaporative emissions and call into question the test results.

      Manufacturers' suggestions to  perform the running loss segment of the test
without measuring emissions, in order to increase testing capacity, does not address
EPA's primary concern: that manufacturers would minimize purge rates during the
exhaust emission test.  In fact, removing the vapor generation component from the
running loss test by holding the vehicle and its fuel at low nominal temperatures
would only increase the incentive for manufacturers to delay substantial purge until
the running loss test.

      A special  test measuring vehicle emissions from two diurnal  heat builds
immediately after the exhaust emission test is the only suggested  modification to
CARB's test procedure that addresses EPA's need for assurance of adequate purge.
This assurance comes from the fact that such a test measures emissions following a
relatively short amount of driving, as is common in use.  Measuring emissions is
necessary  to establish a performance standard,  and to  prevent the need for any
intrusive measurement of secondary variables such as canister mass or purge airflow.
A supplemental procedure could verify sufficient purge for short trips without being
more stringent overall than the full three-diurnal test. Such a procedure would only
change  the  overall test requirements for vehicles  that are indeed insufficiently
purging early in the test.

      In addition to verifying adequate purge, a supplemental test procedure is also
the best way of dealing with EPA's  other concerns regarding CARB's test.   The
simpler supplemental procedure measures the performance of vehicles' evaporative
emission controls with much lower resource requirements than the full sequence.
Also, the supplemental procedure can prevent the possibility of a significant increase
in exhaust emissions by ensuring that exhaust emissions are measured while the
canister is being purged.

      EPA thus considers the  fundamental elements of the alternate procedure
suggested by GARB, and developed further in the Ford and Chrysler comments, to
be effective and reasonable.  The approach taken in defining this procedure helps to
ensure that it does not introduce challenges to vehicle designers beyond those already
imposed by the three-diurnal test, except  for ensuring that vehicles can  purge
effectively to control evaporative emissions. For example, eliminating a diurnal heat
build, initially loading the evaporative canister only to breakthrough, measuring a
moderate-temperature hot soak, and increasing the standard from 2  to 2.5 grams all
contribute significantly to making the supplemental procedure effective in its limited

                                     14

-------
objective of ensuring proper purge without requiring additional design modifications
(such as increased canister size).  Figure 2-2 shows EPA's approach to designing an
evaporative test based on the Ford and Chrysler comments. Also, EPA believes that
the vehicle hardware that would be needed to meet the test requirements proposed
in EPA's January 1990 NPRM (e.g., canisters, purge valves) will be sufficient to meet
the requirements of the supplemental test.

      The supplemental test procedure would not in itself provide assurance that a
vehicle could meet all requirements of the longer three-diurnal test.  For example,
there is no measurement of running losses and the final diurnal heat build is omitted
in the supplemental test. Thus, the supplemental procedure is not a replacement for
the three-diurnal test. However, the opportunity for EPA to run the longer test, in
both confirmatory certification and in-use testing, provides the necessary assurance
that vehicle  designs will achieve optimum control.

      Because neither test sequence is sufficient in itself to demonstrate adequate
control of evaporative emissions, manufacturers would have to perform certification
testing using both sequences. Reserving the supplemental test only for EPA's testing
of in-use  vehicles,  as  suggested by  Ford and Chrysler, would  therefore  be
inappropriate. EPA recognizes that this adds some testing burden to the certification
process. However, the record established in the docket for this rulemaking makes it
amply clear  that the industry views consistency with CARB's requirements (with
potential implications for vehicle designs and costs) to be of more critical importance
than minimizing the test burden for federal testing.  A test based on the CARB
procedure, with  the addition  of the supplemental test,  deals with manufacturers'
concerns and, because it allows EPA to meet statutory requirements, is acceptable to
EPA. Moreover, CARB has expressed its willingness to recommend the adoption of
this approach to their Air Resources Board following action by EPA.12

B.    Emission Standards

      EPA  intends that further emission reductions from improved  evaporative
emission controls will be prompted by changes in the test procedure itself, rather
than by simply reducing the numerical value of the standard, as discussed in Chapter
1. The revisions to the test procedure do cause a change in the effective stringency
of the standard, since  the test requires the  same  performance  from a more
challenging sequence of events, and thus will lead to optimized control of evaporative
emissions under the situations contemplated  by Clean Air Act section 202(k).
    ^Letter from Thomas Cackette, CARB, to Charles L. Gray, EPA, September 15.
1992 (Docket A-89-18, item IV-D-88).

                                     15

-------
         Figure 2-2
Evaporative Emissions Test


1
1
f
Canister preconditioning
,

Cold start exhaust test


Hot start exhaust test
1

Hot soak test
68-86F
,
ambient

Vehicle soak
last 6 hours at 72 F
,
.
Diurnal emission test
2 heat builds in 48 hours
72-96 F excursion
d
^

_______________
fro minutes
Fuel drain & fill
9 RVP. 40% full
7.8 RVP for altitude
,

Vehicle soak
68-86 F
i

Preconditioning drive
one UDDS cycle
.

Fuel drain & fill
,

Canister preconditioning
i

Cold start exhaust test
,

Hot start exhaust test
_,
5 minutes MAX
I
1 hour

6-36 hours


Running loss test
95 F ambient
UDDS,NYCC,NYCC,UOOS
1 0" max tank pressure
i


1 2-36 hours



1 hour MAX
12-36 hours
110 minutes
[4 hours MAX

Hot soak test
95 +/-S F ambient


5 minutes MAX
1 hour
.
Vehicle soak
last 6 hours at 72 F
i

Diurnal emission test
3 heat builds in 72 hours
72-96 F excursion



»
6 - 36 hours


-------
EPA Proposal

      The existing standard is set at 2 grams of hydrocarbons for the combined
measurements of diurnal and hot soak emissions.  EPA proposed at the January 1992
workshop to adopt the same form of the evaporative standard that GARB adopted.
For light-duty vehicles (LDVs) and light-duty trucks (LDTs), this would require the
sum of the measured emissions from the diurnal and hot soak tests to be less than
2 grams, and the emissions from the running loss test to be less than 0.05 g/mi (0.03
g/km). Heavy-duty vehicles (HDVs) would be  subject to a similar standard based on
the existing levels of 3 and 4  grams for light and medium heavy-duty vehicles,
respectively, and a running loss standard of 0.05 g/mi (0.03 g/km). Heavy heavy-duty
vehicles would continue to be subject to an engineering evaluation to  demonstrate
compliance with the standards for medium heavy-duty vehicles.

      Unlike GARB, EPA proposed to include nonfuel emissions in the standards, not
allowing any subtraction or  exclusion of  nonfuel background  from  measured
emissions.   EPA revised the  wording of the  standard from "fuel evaporative
emissions" to "evaporative emissions" to reflect the intent to include all measured
emissions to determine compliance with standards. Again, this is in keeping with the
mandate of section 202(k) to optimize control of evaporative emissions.

      EPA also proposed at the January 1992  workshop to set a standard of 1.0 gram
for the spitback test during vehicle  refueling.

Summary of C
      Commenters  concentrated  on the  need to allow subtraction of nonfuel
background emissions. Manufacturers argued that they should not be responsible for
nonfuel emissions in evaporative testing.  This concern was  heightened by the
proposed change in the sampling period for the diurnal emission test from 1 hour to
24 hours, because nonfuel emissions would accumulate for a much longer time.

      For the supplemental procedure, manufacturers requested that the 2-gram
standard  be relaxed to avoid a net increase in  stringency.   Manufacturers
recommended a 2.5-gram standard for vehicles with fuel tanks smaller than 30
gallons (110 liters), and a 3.0-gram standard for vehicles with larger fuel tanks.

      Chrysler suggested introducing somewhat relaxed standards for intermediate
useful life testing.   Chrysler  argued that the lack  of experience with the new
technologies warranted protection from minor deterioration caused by unexpected
problems. For in-use testing, Chrysler proposed relaxing the 2-gram standard to '2 5
grams, and relaxing the 0.05-gram standard for the running loss test to 0.06 grams

      EPA received no other comments related to the running loss test standard <>r
the spitback test standard for light-duty vehicles and light-duty trucks, and recent
no comments related to any standards for heavy-duty vehicles.

                                     17

-------
Analysis of Comments

      EPA's  position on  the treatment  of nonfuel  background  emissions  was
established in the original action to implement the 2-gram standard for evaporative
testing (41 FR 35626, August 23, 1976).  The Agency stated at that time:

      Providing a factor to account for possible background emissions (e.g., 1
      g/test) would thus have the practical effect of easing the intended fuel
      evaporative emission standard by whatever the background allowance
      might be, since manufacturers would still be expected to provide test
      vehicles  with  minimized background  emissions....  If, however, a
      manufacturer uses paints or plastic materials that will continue for long
      periods of time to emit significant levels of background emissions, these
      emissions should properly  be charged  against  the  in-use vehicle
      inasmuch as such emissions are real hydrocarbons that contribute to air
      pollution.

EPA's position thus has been that the nonzero standard implies an expectation of full
control of evaporative emissions, with an allowance to account for nonfuel background
emissions and test variability. New understanding of resting losses prompts a small
readjustment of this view, since it is now  clear that evaporative  emissions can be
minimized, but  not completely eliminated.  The revised test and standard, however,
maintain the expectation that fuel emissions are near zero, even with the increased
performance requirements for control of diurnal, hot soak, running loss, and resting
loss emissions.

      Including nonfuel background emissions in  the reported values should not
jeopardize the validity of test results or the feasibility of compliance. EPA and GM
test data indicate that nonfuel emissions from stabilized vehicles are typically on the
order of 0.1 grams, and  occasionally as  high  as 0.5 grams,  in  a 24-hour
period.13'14 Even the worst expected background emission is considerably smaller
than the 2-gram standard.

      Furthermore, as a matter of practice, EPA typically excludes vehicles from
recall confirmatory testing for evaporative emissions if there is an indication that
nonfuel emissions may be unusually high.  For example, if a vehicle had received
rust-proofing, body repairs, or modifications to upholstery within the previous three
months, it would be excluded from testing.  Similarly, if a vehicle had experienced a
    13"Real Time Nonfuel Background Emissions," Harold Haskew, et al, General
Motors, October 1991, SAE 912373.

    14"Resting Loss and Background Emission Data," EPA note from Alan Stout. t<>
Charles Gray, June 20, 1991 (Docket A-89-18, item IV-B-5).

                                     18

-------
spill from any container of gasoline or kerosene within the previous three months, it
would be excluded from testing.

      In response to the comments requesting a relaxed standard for vehicles with
large fuel tanks,  EPA agrees that this is appropriate for certain vehicles.  EPA's
approach to  setting evaporative emission standards has  consistently been that,
although a good test procedure is the primary means by which control is assured,
setting  different  diurnal emission  standards based  roughly on vehicle size  is
appropriate.  This provides recognition of the fact that a larger vehicle, possessing
larger fuel system components, is likely to have higher emissions in diurnal testing
than a smaller vehicle, even when they are both achieving what would generally be
considered an equivalent degree of optimized control.  EPA's choice of standards for
heavy-duty vehicles in this  rule follows this pattern.  This is even more important
with the introduction of extended-time diurnal  emission testing, discussed  below.
Consistent with this approach, EPA believes that the larger light-duty trucks, those
with a gross vehicle weight rating (GVWR) between 6,000 and 8,500  pounds (2,700
to 3,900 kg), which have large fuel tanks (over 30 gallons (110 liters) nominal size)
comprise a category of vehicles for which it is appropriate to set a standard of 2.5
grams, halfway between the 2-gram light-duty standard and the 3-gram light  heavy-
duty standard.

      On the issue of the standard for the supplemental procedure, EPA is increasing
the level of the  standard by 0.5 grams for  all  vehicles in order to accommodate
manufacturers' concerns.   The additional 0.5  grams helps to  ensure  that the
supplemental procedure does not introduce challenges to vehicle designers beyond
those already imposed by the three-diurnal test, as described above, and the need to
provide assurance that vehicles can purge effectively within a relatively short driving
time. The standard for the  measured diurnal and hot soak emissions would then be
2.5 grams for light-duty vehicles and light-duty trucks, 3.0 grams  for light-duty
trucks with GVWR between 6,000 to 8,500 pounds (2,700 to 3,900 kg) that have fuel
tanks of nominal size of 30 gallons (110 liters) or more, and 3.5 and 4.5  grams for
light and medium heavy-duty vehicles, respectively. The slightly relaxed standard
for the  supplemental procedure will  not compromise in-use  control, because the
adequacy of control is primarily determined  by the test procedure (as explained m
Chapter 1), and because manufacturers must also design vehicles to meet the lower
standards for the full three-diurnal  test.   Even with the higher  standard,  the
supplemental procedure, in combination with the rest of the test,  will meet  the
statutory requirement to require the greatest achievable degree of control.

      The 0.05 g/mi (0.03 g/km) standard for the running loss test was first proposed
by GARB, and was widely supported by the auto industry. The standard allows about
1 gram  of emissions during the hour  of driving, comparable to current testing tor
diurnal  and hot soak emissions. EPA believes that CARB's adopted standard of 0 " ">
g/mi (0.03 g/km) will require the greatest achievable degree of control under
prone summertime conditions.
                                     19

-------
      Regarding Chrysler's suggestion for relaxed in-use testing standards, EPA does
not believe it necessary to relax these standards. Based on comments, EPA expects
that improvements in the evaporative control systems will be fairly straightforward
upgrades of existing technologies, as described in Chapter 3.  Commenters have
provided no clear rationale for a change in technology that would justify a relaxation
of the standards for in-use vehicles.

C.    Diurnal Emission Test

      The  generation  of vapors  from  diurnal  heating is  a straightforward
phenomenon that can be simulated in laboratory testing.  When a vehicle is not
driven but is exposed to outdoor conditions, its fuel temperatures rise and fall in
response to cycling ambient temperatures. The increase in fuel temperatures causes
evaporation  of liquid fuel and expansion of the vapor space so that fuel vapors are
driven from  the tank. The current method of simulation — rapidly heating the fuel
from an initial to a final temperature — is a simple procedure that simulates the
actual diurnal heat  build.  Over the course of the rulemaking EPA has considered
various changes that would not only make the test more challenging, but would also
increase the accuracy of the simulation.

EPA Proposalr-Test Method

      At the January 1992 workshop, EPA proposed the new diurnal test method
advocated by GM, in which the whole test vehicle  would be exposed to ambient
temperatures cycled over 24-hour periods. Emissions would be collected in a SHED
(Sealed Housing for Evaporative Determination).  EPA also requested comment on
a simplified method  of conducting the diurnal temperature cycling. The test vehicle
would be stabilized at the low diurnal temperature, then exposed to the high diurnal
temperature with fans positioned to ensure sufficient air circulation around the fuel
tank.  The air circulation around the fuel tank decreases the time required for the
fuel to experience the full temperature excursion. Because this procedure would not
be long enough to allow the measurement of resting losses separate from diurnal
emissions, a separate test for resting losses would be required.

          pf    rnent8
      GM championed the concept of extended-time diurnal emission testing; CARB
and almost all other manufacturers eventually expressed support for the approach.
GM identified several advantages of their method of testing.  First, cycling ambient
temperatures in 24-hour periods would improve the accuracy and repeatability of the
test.  Vapor loading rates to the canister  would be more representative, the fuel
heating would be close to an equilibrium process, as it is in use, and all emission
sources would be better measured and controlled.

      Second, by  allowing  the  fuel to cool without refueling,  the  test would
automatically account for fuel weathering.  Fuel weathering, which occurs as a

                                     20

-------
vehicle is driven or exposed to diurnal heat builds, decreases the volatility of the fuel
over time.

      Third, GM  claimed that the improved simulation of a  vehicle's  in-use
experience would  provide incentive for  manufacturers to  design their emission
controls to achieve maximum in-use emission reductions. The lengthy time in the
measurement enclosure would account for resting losses, providing an incentive for
manufacturers to modify vehicle designs to control these slowly emitted losses. Also,
allowing the fuel to cool naturally would encourage manufacturers to design their fuel
systems to allow the phenomenon of backpurge. A vehicle that can backpurge draws
air and vapors from the evaporative canister into the fuel tank when the liquid fuel
cools and causes condensation of fuel tank vapors.  Ambient temperature exposure
would also give full credit for insulating fuel tanks to minimize  fuel temperature
excursions (and thus vapor generation).

      Several manufacturers opposed the use of the accelerated method to measure
diurnal emissions.  These manufacturers supported the 24-hour testing method and
considered the flexibility provided by an alternate method unnecessary.

Analysis of CorniT>ent8

      While the Agency has no objection to the concept of extended-time testing, it
believes that an effective diurnal  emission test, with most of the advantages of
extended-time testing, can be realized with a much shorter test time. The procedure
in the December 1990 workshop notice already took many of these factors into
account.

      EPA believes that the alternate method for accelerated ambient heating of the
vehicle could achieve emission reductions comparable to those achieved by GM's test.
Such testing would require only minor modifications to existing equipment and
facilities and would greatly reduce the time involved  for each test.  Nevertheless.
because of the  broad support  for extended-time diurnal emission testing,  EPA is
finalizing regulations based on this approach.  Also, any possible emission  sources
that may not currently be identified would more likely be measured and controlled
with the real time  test.

EPA Proposal—Test Severity

      In the January 1990 NPRM, EPA proposed to conduct two consecutive diurnaJ
heat builds. EPA proposed to heat the fuel from 72° to 96° F (22.2° to 35.6° C), based
on a thorough analysis of summer ambient temperatures across much of the United
States.  At  the December 1990 and January 1992 workshops, EPA proposed  the
addition of a third  diurnal heat build.

      As part of the proposal at the January 1992 workshop to adopt GM's  method
of heating the  whole vehicle,  EPA included a provision to require an underbo
-------
circulation rate of at least 5 miles (8 km) per hour to maintain the transfer of heat
into and out of the fuel tank.
          of Cornrr>ents
      Manufacturers did not object to the specified temperature range.  However, to
deal with EPA's concerns for fuel temperatures lagging behind ambient temperatures,
manufacturers suggested the possibility of adjusting to an ambient temperature
range of 70° to 98° F (21.1° to 36.7° C). In later comments, manufacturers expressed
a preference to maintain the 72° to 96° F (22.2°  to 35.6° C) ambient temperature
range, but to specify a temperature tolerance for the air just below the fuel tank.

      Manufacturers questioned the basis for adding the third diurnal heat build,
and suggested conducting only two heat builds to  save test time and facility costs.

Analsis of
      Section 202(k) of the Clean Air Act specifically requires test conditions to
reflect "ozone-prone summertime conditions."  The temperatures chosen by EPA are
based on meteorological data for high ozone days. The high diurnal temperature of
96° F (35.6° C)  is the average of the daily high temperatures for the hottest ten
percent of the days in this data.   Since daily low temperatures in this data subset
were, on average, 24° F (4.4° C) less than the associated high temperatures, 72° F
(22.2° C) was chosen as  the low diurnal temperature.  Thus, to meet statutory
requirements, EPA is  retaining the  72° to 96° F (22.2° to 35.6° C)  temperature
specification.

      Fuel temperatures are expected to experience temperature cycling during the
diurnal emission test comparable to what would occur outdoors under ozone-prone
summertime conditions.  Since test data indicate that peak fuel temperatures are
typically at least as  high as peak ambient temperatures, EPA would expect fuel
temperatures in uninsulated tanks to experience the full temperature cycling from
72° to 96° F.   Specifying a minimum 5 mile (8 km) per hour circulation under the
vehicle is important  in achieving this goal.  Testing by EPA shows that, with air
circulation  around the  fuel tank and gradually cycled ambient temperatures, fuel
    15"Procedure for Determining Daily Maximum and Diurnal Temperatures," EPA
memo from Mark Wolcott to John Anderson, September 12, 1988 (Docket A-89-1*
item II-B-2).

    16"Peak Fuel Temperatures in Parked Vehicles," EPA memo from Alan Stout to
Joanne I. Goldhand, January 13, 1993 (item IV-B-9).

                                     22

-------
temperatures in uninsulated tanks  stay  very  close to ambient  temperatures.17
Specifying a temperature tolerance for the space below the fuel tank, as suggested
by commenters, would be inappropriate, because the temperature gradients involved
are very small (especially relative to  the ±3°  F (±1.7° C) ambient temperature
tolerance).  Also, even a very thin layer of still  air around the fuel tank would be
effective in insulating the tank.  Because adequate air circulation plays a key role in
the diurnal  heat transfer process, EPA will retain the opportunity to determine
appropriate fan configurations when performing evaporative emission tests.

      Moreover, EPA may compare  a vehicle's  fuel temperatures under outdoor,
summertime conditions with test fuel temperatures and take measures to correct any
demonstrated discrepancy. EPA recognizes that it is not feasible for a test procedure
to ensure a  SHED environment that simulates the wide range of environmental
factors affecting fuel  tank heating in  use, particularly those induced by direct solar
heating of the vehicle and nearby surfaces. EPA may adjust ambient temperatures
as necessary to induce the temperature swing in the fuel representative of in-use
conditions. For vehicles with fuel tanks isolated from underbody airflows, by physical
barriers or by location,  EPA may adjust fans to ensure sufficient air circulation
around the fuel tank.

      Regarding the comments objecting to the addition of a third diurnal heat build
to the test procedure, EPA believes that the third heat build is appropriate.  GARB
first introduced a third  diurnal heat build to evaporative testing, arguing that it
would further reduce in-use evaporative emissions. The addition of the third heat
build will prompt more  effective vehicle designs (in terms of canister capacity and
canister purge), which will cause reduced emissions for vehicles that are parked for
three consecutive days, as well as other driving scenarios. For example, if a vehicle
is driven only a short amount between one-day parking episodes, the additional
effectiveness of the canister purge could enable the vehicle to maintain control. In
addition, the statute clearly provides discretion to take  this action. Section 202(k)
states that an evaporative emission standard should provide optimized control "over
two or more days of nonuse."

D.    Running LOSS Test

      When federal  testing for evaporative  emissions was  first developed,  the
possibility of running losses was acknowledged,  but  the understanding  of  the
phenomenon and the development of test equipment was insufficient at that time to
prescribe a test that would prevent running losses under typical summer conditions.
The  original test for running losses, which involved sampling for emissions with
carbon traps during the exhaust emission test, is ineffective because of the  low
temperatures and the inadequacy of the collection method.
    17"Testing in Support of Short Diurnal Test", EPA memo from Bryan Manning to
Chester J. France, December 16, 1991, (Docket A-89-18, item IV-B-4).

                                     23

-------
      In recent years EPA's understanding of the extent of in-use running losses has
grown as testing of in-use vehicles revealed unexpectedly high levels of running
losses.  EPA therefore included in the proposed revisions to the test procedure a
means of ensuring running loss control. Furthermore, the amended Clean Air Act
mandates that EPA regulate running losses from in-use vehicles.

      The fundamental question in regulating running losses is whether vehicles
should be directly tested for emissions during driving, or whether the arrangement
of the other test segments — loading the canister, executing a prescribed drive, and
conducting consecutive  diurnal heat  builds— can ensure sufficient running  loss
control. The following presentation first deals with this fundamental question, then
pursues the whole range of issues that arise if a running loss test is required. In the
January 1990 NPRM, EPA published GM's running loss  test for comment,  and
proposed a similar test at the December 1990 and January 1992 workshops.

EPA Proposal — Adoption of a Running Loss Test

      EPA proposed a set  of test provisions in the January 1990 NPRM to prevent
the occurrence of in-use running losses. These provisions were intended to guarantee
that all fuel vapors generated in the  tank would eventually be consumed by the
engine.  First,  the initial  loading of the evaporative canister would ensure that
sufficient purge occurs during the exhaust emission test so that in-use vehicles would
be able  to  create canister  capacity  with limited driving.   Second, designs of
evaporative emission  controls would  be  subject to an  engineering  review for
certification to ensure that  all hydrocarbon vapors would be routed to the evaporative
canister or directly to the engine.  Finally, fuel cap removal at the beginning of the
hot soak test (after the exhaust emission test) would encourage  the use of low-
pressure fuel tank designs during driving. Vehicles with low-pressure designs are
preferable, because they are less likely to emit high rates of hydrocarbons if a fuel
system loses the ability to  hold pressure.

Summary
      Almost all commenters supported adoption of a running loss test and objected
to the other  provisions that were intended to ensure running  loss control.   In
particular, manufacturers objected to an undefined design review and opposed  any
limitation of fuel tank pressures. GM and Toyota submitted very  similar proposals
for conducting a high-temperature running loss test for a long drive. The New York
State Department of Conservation supported running loss testing because it included
an incentive to minimize vapor generation.

Analysis of Comments

      The main  factor in controlling running losses is a vehicle's ability to purge
vapors to  the engine as they are generated.  The sequence of events in the  test
procedure (with or without a running loss test) should generally ensure that vehicles

                                     24

-------
passing the diurnal emission test will have sufficient purge to pass a running loss
test. A direct test for running losses, broadly supported by industry, would provide
positive assurance that this is the case. EPA has therefore decided to include it as
part of the test procedure. A new running loss test would also have the benefit of
giving  manufacturers the incentive to pursue vehicle designs that  reduce fuel
temperature increases (and vapor generation) during driving.

      EPA is omitting the cap-removal step from the hot soak test. EPA is instead
adopting CARB's requirement to prevent fuel tank pressures higher than 10 inches
of water (2.5 kPa), unless the vapors, other than refueling emissions, are vented to
the evaporative canister when the fuel cap is removed.  EPA  is adopting this
requirement to maintain consistency with CARB's test  requirements wherever this
can be done without sacrificing emission control benefits.  Although EPA does  not
believe that this requirement will fully prevent high fuel tank pressures, because of
the vent-to-canister provision, CARB's requirement should be effective in preventing
gasoline emissions whenever fuel caps are removed after a pressure buildup in  the
fuel tank. In the future EPA expects to pursue other regulatory action as necessary
to deal with the problems associated with fuel systems  that fail to hold pressure.

      EPA is omitting the design review as a routine part of the certification process,
but is including in the regulations the requirement to route all vapors to the engine
or the evaporative canister. This provides the manufacturers with a straightforward
requirement aimed at ensuring that vehicles do not routinely lose control. There is
no justification for maintaining a valve that would release vapors under expected
driving conditions.  The  regulations allow for an exception in emergency situations
and are not intended to include refueling emissions.

EPA Proposal—Test Parameters

      EPA proposed that a vehicle be exposed to 95° F (35° C) ambient conditions
with humidity held at 100 grains per pound (14.3 grams per kilogram) of dry air
during the running loss test.

      EPA  also recommended  that the driving cycle consist of  one  Urban
Dynamometer Driving Schedule (UDDS), two New York  City Cycles (NYCC), and
another UDDS. After the first  UDDS and the second NYCC the vehicle would
experience a two-minute idle. The  two NYCCs of this driving cycle were a revision
to GM's proposed driving cycle of three UDDSs, which CARB adopted.

      EPA specified a vehicle refueling event before the running loss test to prevent
excessive weathering of  the test fuel.  EPA also specified an overall air circulation
requirement of 5.5±0.5 cfxn per cubic foot (fpm per liter) of enclosure volume.
                                     25

-------
Summary of Comments

      Manufacturers endorsed the 95° F (35° C) ambient temperature specification.

      Most manufacturers preferred a driving schedule with three UDDS cycles, but
they acknowledged that either driving schedule would be acceptable, especially if it
would be adopted by both EPA and GARB. Most commenters did not expect the
different driving schedules to result in product differences or different in-use control
of running losses.  The New York City Department of Environmental Protection
emphasized that vehicles should be tested with low-speed driving and short trip
cycles, such as is reflected in the NYCC, to ensure control in these conditions.

      Manufacturers requested  lowering or removing the humidity specification to
avoid uncomfortable conditions for the test  drivers,  adding that the humidity
specification is not a significant  test requirement.

      Manufacturers requested that the refueling event before the running loss test
be omitted, to avoid having to test to an unlikely in-use scenario and to prevent
unnecessary test effort.

      Honda asked that EPA change the air circulation requirement, because of the
large volume of their enclosure,  and because of CARB's different specification for a
high-volume road-speed modulated fan.

Analysis of Co*"rnents

      The selection of 95° F (35° C) as the ambient temperature for the running loss
test is based on the control scenario of a vehicle starting a long trip near the point
of its maximum fuel temperature  during the day, clearly a reasonable possibility.
With no comments to the  contrary, EPA is adopting this specification.

      EPA analysis supports the driving schedule proposed at the December 1990
workshop.  EPA analysis of test  data indicates that  the New  York City Cycle,
representative of approximately 15 percent of urban driving,  can increase the
effectiveness of the running loss test.18 The analysis showed that vehicles with
adequate purge for higher-speed UDDS driving can have inadequate purge at lower
speeds.
    18"Running Loss Emission Control: LA-4 vs. NYCC Test," EPA memo from Rick
Rykowski to Charles Gray, April 17, 1990 (Docket A-89-18, item IV-B-6).

                                     26

-------
      Controlling the humidity of the test environment is desirable, because humidity
has  some effect on the vehicle's ability to purge its canister.19  The  proposed
humidity specification of 100 grains per pound (14.3 grams per kilogram) of dry air.
equivalent to a dew point of 67° F (19.4° C.) and a relative humidity of 40 percent at
95° F (35° C), represents typical summer conditions in the United States.   EPA
acknowledges, however, that driver discomfort at the high temperatures  is a
reasonable concern. Given the concern for driver discomfort, the proposed humidity
specification, as a test parameter of secondary importance, has not been retained.

      EPA has revised the specified circulation to  a minimum 2 cfm per cubic foot
(Cpm per  liter) of enclosure volume.   This minimum overall circulation rate is
expected to be sufficient to prevent temperature and hydrocarbon stratification in the
enclosure. Omitting the maximum circulation rate allows full use of a fan that varies
with dynamometer roll speed.

EPA Proposal—Fuel Temperature Profile

      EPA proposed that the manufacturer establish fuel tank liquid temperature
profiles for driving in summer conditions. The fuel temperature profile would be used
as a target during the running loss test to simulate the heating of the vehicle's fuel
tank during driving.  EPA proposed to limit testing to certain ambient conditions.
including steady or increasing ambient temperatures over 80° F (26.7° C), a sustained
wind speed under 10 miles (16 km) per hour, and a road surface temperature at  least
30° F  (16.7° C)   above ambient  temperature,  on  average,  during  the  test
Manufacturers would generate the profile by obtaining a trace of fuel temperature
versus time while driving the vehicle over the established driving  schedule.

      The measured temperature profile would be adjusted to an initial temperature
of 95° F (35° C).  Fuel temperatures during the running loss test would be controlled
to match the profile temperature within 3° F (1.7° C); this tolerance would be reduced
to 2° F (1.1° C) during the last two minutes of the test.

Summary of Comments

      CARB  adopted a procedure with different requirements for developing fuel
temperature profiles. CARB specified ambient temperatures at least 95° F (35' <' .
wind speeds under 15 miles (24 km) per hour, road surface temperature at  least
20° F (11.1° C) above ambient  temperature, and a maximum cloud cover  of 2^
percent.
    19"Effect of Humid Purge Air on the Performance  of Commercial Activ.u.
Carbons Used for Evaporative Emission Control," J. Urbanic, et al, September l'«-
SAE 892039.

                                     27

-------
      Commenters acknowledged the need to develop fuel temperature profiles in
summer conditions. Manufacturers recommended that  EPA and GARB adopt  the
same ambient condition requirements to avoid the need for two  profiles.  GM and
Ford proposed ambient temperature specifications of 87°  to 100° F (30.6° to 37.8° C)
and 90° to 100° F (32.2° to 37.8° C), respectively.  Honda recommended specifying an
ambient temperature of at least 95° F (35° C), a sustained wind speed of less than
10 mph (16 km/hr), and road surface temperatures at least 20° F (11.1° C) above  the
ambient temperature, measured  at the beginning and  end of the drive.  Toyota
commented that the proposed ambient requirements would seriously limit, the ability
of Japanese and European manufacturers to establish their fuel temperature profiles.

      Honda requested that EPA allow manufacturers to adjust the measured fuel
temperature profiles to an initial temperature less than 95° F (35° C). The lower
initial temperature would be based on a demonstration that there  is adequate
insulation of the fuel system to ensure lower-than-ambient fuel temperatures at  the
beginning of any drive.

      Honda also  requested that EPA add a requirement to control the vapor
temperatures in the fuel tank during the running loss test.  Manufacturers would
generate a vapor temperature profile for the last two minutes of the driving cycle;
vapor temperatures during the last  two minutes of the  running loss test would be
required to match this profile to within 3° F (1.7° C).

Analysis of Comments

      Through the development of fuel temperature profiles for in-use vehicles, EPA
has learned much about the factors that affect the heat transfer to the fuel tank
during vehicle operation.  The temperature of the pavement relative to ambient
temperature correlates most strongly with fuel temperatures.  Test data shows that
pavement temperatures are often 30° F (16.7° C) and sometimes 40° F (22.2° C) or
more  above ambient temperatures.   EPA is therefore specifying  a  minimum
temperature difference of 30° F (16.7° C) to provide adequate assurance of conditions
representative of ozone-prone summertime conditions.

      EPA is requiring measurement of pavement  temperatures throughout  the
drive.  Pavement temperature is very sensitive to instantaneous solar loading
Pavement temperatures measured only at the beginning and end of the drive could.
therefore,  give a poor indication of conditions on a partly cloudy day.

      EPA does  not believe that very high ambient temperatures are needed to
develop fuel temperature profiles.  As long as pavement temperatures are high
   20"Determination of Tank Fuel Temperature Excursions", Final Report by ATL
Inc. (EPA contract 68-C9-0027, Work assignment 2-1), November 19, 1991 (Docket
A-89-18, item IV-A-4).

                                    28

-------
enough  above  ambient  temperatures,  fuel  temperature  behavior should  be
representative of high-temperature conditions.  In fact, with all other conditions
constant, higher ambient temperatures should coincide somewhat with lower fuel
temperature increases. This relationship can be explained by the dependence of heat
transfer on temperature gradients.  When the fuel in the tank starts at a lower
temperature, there is a greater temperature gradient relative to the underbody air
heated by  the  engine, and  also  a  greater temperature difference  between the
recirculated fuel and the fuel in the tank. However, this effect does not appear to be
primary, and,  considering manufacturers'  desire for consistency with CARB's
requirements, EPA is specifying a minimum ambient temperature of 95° F (35° C>,
with no maximum temperature.

      Because pavement temperature serves as a good indicator of the degree of solar
loading, and is  fundamental in affecting fuel temperatures,  EPA believes that an
extra specification for maximum  cloud cover is not  necessary.   However, to be
consistent with  CARB's requirements, EPA is including CARB's specified maximum
cloud cover of 25 percent.   Considering the subjectivity involved in cloud cover
assessments  and ambient temperature  measurements, EPA  is  also  requiring
submission of meteorological data from the nearest weather station.  EPA  expects
manufacturers to justify any significant discrepancy between the reported figures.

      EPA is not allowing manufacturers to use lower initial temperatures for the
running loss  test for several reasons.  First, in the common in-use occurrence in
which a vehicle parks for a short time between trips,  fuel temperatures at the
beginning of the second trip could be much higher than the ambient temperature
because of the previous drive.  In such a scenario, tank insulation would actually
increase the likelihood of higher fuel temperatures. Second, EPA test data indicate
that peak  fuel temperatures  are typically at least as high  as peak  ambient
temperatures,  and  sometimes as  much as 5°  or 6° F   above  peak  ambient
temperatures.   Third, 95° F  (35° C) is not intended to be a maximum possible
ambient temperature. Therefore, even with substantial insulation of the fuel  tank.
fuel  temperatures  could easily reach 95° F (35° C)  on a warmer day.  Finally
specifying different temperatures for fuel and  ambient at the beginning of the test
introduces a technical difficulty.  Because dissimilar temperatures in the  lab  are
inherently unstable, maintaining different temperatures until the start of the  te*t
as well as controlling the fuel temperatures during the early stages of the profile
would be difficult.

      EPA strongly opposes  the use of vapor temperature profiles for the running
loss test. If the vapor temperature during testing were to begin to depart from t h-
profile, technicians, using airflow to control liquid fuel temperatures, would have :
additional control variable to independently manage vapor temperature.  The  • I •.
   21"Peak Fuel Temperatures in Parked Vehicles," EPA memo from Alan St.
Joanne I. Goldhand, January 13, 1993 (item IV-B-9).

                                     29

-------
temperature  requirements would  greatly increase the difficulty  of  successfully
completing a running loss test. In addition, because the fuel temperature is the key
parameter linking the running loss test with on-road summertime driving, matching
an additional parameter may result in running loss tests that simulate an  overly
narrow range of in-use conditions.

EPA Proposal—Testing Method

      EPA believes that vehicles should be subject to running loss testing either in
a sealed enclosure or by the point-source method, in which emissions are sampled
from the several expected sources on the vehicle. The opportunity to test in a sealed
enclosure would ensure that emissions from unexpected sources would be accounted
for.

Summary of Comments

      Auto manufacturers opposed the testing for running losses by the enclosure
method. First, manufacturers wanted to be responsible for passing only a single test
for all enforcement testing.  Manufacturers claimed that having two alternative test
methods would make them subject to dual test requirements. Because they would
have to pass a test by either method, they would have to  develop facilities and
operations to  demonstrate  compliance  with  both  methods.    Furthermore,
manufacturers believed that emissions from unexpected sources would be detected
in the diurnal or the hot soak tests.

      Second, commenters were concerned about the  repeatability, accuracy, and
technological feasibility of the enclosure method. GM questioned the mass resolution
and hydrocarbon retention capabilities of the larger  SHED with  the higher  air
circulation, and expected the vehicle's demand for engine and supplemental air to
distort measurements.  In comparison, GM claimed  much  better than 0.1-gram
precision  and over  98 percent propane  recovery for the  point-source method.
Manufacturers  claimed that nonfuel  emissions  in an enclosure would  pose  an
unmanageable problem. Manufacturers questioned the feasibility of maintaining the
specified ambient temperatures in an enclosure and of routing  air to the induction
system and from the tailpipe.

      Third, manufacturers claimed that  an enclosed dynamometer would  be  too
costly to justify. GM stated that the  enclosure facility cost  would be at least ten
times the cost of the point-source sampling system, since the point-source test would
utilize existing equipment configurations.  Similarly, Nissan claimed that the
enclosure  method would cost from $459,000-$473,000 per enclosure, compared to
$87,000 per cell for point-source testing. Ford suggested that the enclosure test may
require robotic control, which could cost an additional $250,000 per unit.

      Finally, manufacturers urged EPA to eliminate the enclosure method for safety
reasons.   Manufacturers claimed  that driving the vehicle  in the enclosure was

                                     30

-------
unacceptable because of the potentially high rates of vapor generation  and the
presence of high-pressure fuel-injection systems without ventilation in the enclosure.
These conditions  could result in high levels of hydrocarbons  or other gases in the
enclosure, which  could place the driver at risk of high exposure or fire if he or she
had difficulty escaping the enclosure.

      The Auto/Oil Air Quality Improvement Research Program also indicated a
preference for point-source testing, based on a much higher observed frequency of
void tests with the enclosure method.   The joint test effort involved difficulties
primarily in maintaining control of fuel  and ambient temperatures during running
loss tests. The difficulties were presumably caused by  retaining the mass of air in
the enclosure, and by the limited access  of technicians to make adjustments to fans
and other equipment.

Analysis of Comments

      As in-use vehicles age, they can emit running losses from many unpredictable
sources.  Testing by the point-source method would therefore be appropriate only for
vehicles having readily identifiable and measurable sources of emissions.  Therefore.
if a running loss test is to be effective in achieving the emission reductions mandated
by the Clean Air Act, the Agency believes it must be able to test for running losses
in a sealed enclosure.

      Enclosure testing for running losses is technically feasible.  A standard of 0.05
grams per mile (0.03 g/km) for the running loss test would translate to 0.9 grams of
hydrocarbons or,  for a 4000 ft3 (113  m3) enclosure, about 12 ppm carbon.  Such
concentrations are well within the measurement range of existing equipment. The
large  running  loss enclosure should present  no new problems  of hydrocarbon
retention, even with the higher air circulation. Since air is  routed from outside the
enclosure directly to the engine's air intake, EPA  does not expect  the vehicle's  air
consumption to distort test results. EPA's experience with running loss testing has
clearly demonstrated the feasibility of supplying engine air in an enclosure and of
controlling ambient and fuel temperatures.

      Nonfuel emissions should not compromise test results.  New dynamometers a re
designed to emit no hydrocarbons at test temperatures, and the rate  of nonfuel
emissions from the vehicle should be comparable to that during the diurnal emission
test (much less than 0.1 gram per hour).

      Retaining the option to test for running losses in a sealed enclosure should
pose no new safety problems for testing.  First, EPA believes there is no technic.il
difference in test safety between enclosure and point-source testing, or between
running loss testing and conventional testing for exhaust emissions. In each type ••(
testing, the vehicle is run on a dynamometer in a closed room with air supplied to t hr
engine  and exhaust gases routed out;  sensors  are in place to detect  for unsnN*
conditions; and a door is accessible for personnel to exit.  Second, to prevent ha«i!.

                                      31

-------
emitting vehicles from  producing dangerous hydrocarbon  levels, a  test may be
terminated at the point  of failure and remain valid.  For those manufacturers that
do not want to invest in the enclosure method, contract testing facilities should be
available to perform any desired enclosure testing.

      The experience gained from the Auto/Oil testing does not warrant a conclusion
that the point-source method is superior to the enclosure method.  Maintaining fuel
temperatures within  the  prescribed tolerance  should not be significantly  more
difficult in a sealed enclosure. EPA expects that fuel temperature control will require
some learning, but that technicians will be able to meet test tolerances equally with
either method. As indicated in the comments, programmable automatic controls may
be able to eliminate void  tests caused by fuel temperature  control problems.
Maintaining ambient  temperatures within test  tolerances  is a matter of correctly
sizing, configuring, and operating the heating, ventilating and air conditioning system
and is not expected  to be problematic, either for point-source or enclosure testing.

      In summary,  EPA continues to believe it necessary to retain the option  to
conduct running loss testing in a SHED.  Manufacturers clearly want to retain the
option to conduct their own routine testing using the point-source method, and EPA
has no objection to this method. Furthermore, GARB also allows testing by either the
point-source or the  enclosure method.  Finally,  manufacturers may choose to rely
solely on the point-source  method; manufacturers would have to gain enough
confidence that testing  by the point-source  method would be adequate to ensure
compliance with running loss testing by  the enclosure method.  EPA is therefore
retaining both methods in its test procedure. Thus, the rule makes clear that EPA
may conduct testing using the enclosure method; manufacturers using the point-
source method would have to design vehicles able to pass EPA's enclosure testing to
be in compliance.

E.    Preconditioning

      The preconditioning of the vehicle includes all parts of the test from the
beginning of the test  sequence up to the beginning of the exhaust emission test.
Though no sampling or testing occurs during the preconditioning period, this portion
of the test is important in stabilizing the test vehicle to a known initial condition and
preparing it for the  series of test segments that  follow.

EPA Proposal—Method  of Canister Loading

      In the January 1990 NPRM, EPA proposed to change the existing procedure
by adding a step initially to load the evaporative canister to breakthrough.  The
evaporative canister would be loaded beginning from its as-received condition h\
placing the vehicle in a SHED and repeatedly heating the fuel tank until 2 grams < >f
vapor was  detected in the SHED.  In the December  1990 workshop notice, EPA
revised its proposal to allow as an option CARB's more convenient method to load th-
evaporative canister. In that option the canister would be manually purged and then

                                     32

-------
loaded by sending to the canister an amount of butane equivalent to \Vz times the
canister's working capacity at a rate of 15 grams butane per hour. Of course, since
this would involve loading the canister past breakthrough, not all of this butane
would  be  retained.  In the December 1991 workshop notice, EPA proposed to load
canisters, without purging, to 2-gram breakthrough using a mixture of butane and
nitrogen at a rate of 40 grams butane per hour.

Summary of Comments

      After the December 1990 workshop, auto manufacturers contested the severity
of the revised canister loading procedure.  Since vehicles could pass the test only by
keeping the canister from reaching breakthrough, commenters felt that any load
beyond breakthrough was unjustified. Several recommended that canisters be loaded
with an amount of butane equivalent to the 2-gram breakthrough point.  However,
following the January 1992 workshop, auto manufacturers generally recommended
adopting CARB's procedure, including the canister preconditioning.

      Auto manufacturers agreed with the specification to load canisters with butane
instead of with gasoline, but most commenters objected to the specification of loading
with pure butane. They claimed that the butane should be mixed with equal parts
of nitrogen to better approximate the mixture of gasoline and air generated from the
fuel tank. Manufacturers' data indicated that canisters loaded to breakthrough with
pure butane contained 10  to 60 percent more butane than canisters loaded with a
mixture of butane and nitrogen.  >23

      The American Petroleum Institute (API) pointed out that changes to in-use fuel
volatilities continue to decrease the amount of butane in summer gasolines.  API
advocated a vapor composition for canister loading that includes pentane and hexane,
as well as butane.

      Manufacturers requested that the procedure include an  initial purge of the
vehicle's evaporative canister to eliminate the variability involved in  testing as-
received vehicles. GARB preferred to purge canisters because it enabled them to
specify a  fixed quantity for  subsequent  vapor loading, which reduces the labor
requirement for monitoring the procedure.

      EPA received several suggestions for canister loading rates. Toyota and Ford
recommended 30 and 50 grains per hour, respectively, to avoid a prolonged canister
loading procedure. CARB specified a rate of 15 grams per hour for  their test. GM
   22Letter from Satoshi Nishibori, Nissan, to Docket A-89-18, February 22. 1991
(Docket A-89-18, item IV-D-42).

   23Letter from Samuel A. Leonard, GM, to Richard D. Wilson, EPA February 2'2
1991 (docket A-89-18, item IV-D-45).

                                     33

-------
suggested loading canisters at whatever rate would result in about a 6-hour loading
period, since in-use canisters of various sizes experience different loading rates over
a constant time period.

Analysis of Comments

      The Agency believes that initially loading to the 2-gram breakthrough point
represents a sufficient amount of vapor to test vehicle purge rates.  Loading the
evaporative canister to less than breakthrough would not require emission control
throughout the range of expected evaporative system behavior.  A load represented
by IVfe times the working capacity of the canister is expected by EPA to be a very
unusual in-use occurrence, because properly functioning vehicles that can meet the
new test requirements should rarely experience canister loading beyond breakthrough
in use.  However, to be consistent with CARB's test requirements for the three-
diurnal  test sequence  (as supported  by the manufacturers), EPA would find  it
acceptable to load the canister with an amount of vapor equal to IVz times its working
capacity.  This degree of preconditioning  will contribute  to optimized control  of
evaporative  emissions by  ensuring  control even under  the  relatively  extreme
conditions simulated by loading the canister beyond breakthrough.

      For the supplemental two-diurnal test  sequence, initially loading beyond
breakthrough would be excessive, because of the short amount of driving before the
diurnal emission test.  EPA is therefore specifying an initial canister load to 2-gram
breakthrough for the two-diurnal test.

      Data submitted by manufacturers show that, for loading evaporative canisters,
butane is a reasonable substitute for gasoline.  The  data indicate that  canister
behavior is nearly the same when loaded with the two different types of vapors.24
This is not surprising, since butane comprises about half of the fuel vapor over a
typical fuel with Reid vapor pressure (RVP) equal to 9 psi (62 kPa); the rest of the
vapor is approximately 30 percent pentane,  10 percent hexane,  and 10 percent
heavier  hydrocarbons.    The use of butane would greatly simplify the loading
procedure by  reducing the time  required and making possible  the  use of less
expensive, more readily available equipment.

      At the time of the December  1990  workshop there had been  no  technical
justification for the added test complexity of adding nitrogen to butane for loading the
evaporative canister.  However, manufacturers' data submitted since,  while widely
varying, clearly indicate that the presence of nitrogen in the loading stream decreases
   24Letter from Noboru Fujii, Nissan, to Alan Stout, EPA, June 21, 1990 (Docket
A-89-18, item IV-D-13a).

      Composition  of  Vapor Emitted From a Vehicle Gasoline Tank  During
Refueling," Robert L. Furey and Bernard E. Nagel, GM, February 1986, SAE 860086

                                     34

-------
the amount of hydrocarbon retained by the canister.  Moreover, a review of the
literature revealed a potential theoretical basis for the claimed effect.  The amount
of vapor that can be adsorbed onto activated carbon increases as the partial pressure
of the vapor increases. Increasing the concentration of butane in the vapor stream
from 50 to 100 percent doubles the partial pressure of butane and seems to change
the adsorption from a single-layer to a multiple-layer phenomenon.   This would
explain the more effective retention of pure butane in the canister.  Because of this
new data, and the recognition that nitrogen is a major component of fuel tank vapors,
EPA considers this suggested change to be justified.

      In spite of the advantages of loading canisters with butane, EPA believes it is
important to retain the option of better simulating a canister's in-use experience. The
two-diurnal test therefore allows, as an alternate procedure, canister loading with
fuel tank  vapors.   Maintaining such an alternate procedure  would  avoid the
complication of routinely adding pentane and hexane to butane for canister loading.

      To  better simulate vehicles'  in-use  experience,  EPA is also specifying no
canister purging prior to loading to breakthrough for the supplemental two-diurnal
test. Initially purging a canister from an in-use vehicle would remove the system
from its as-received condition, which may cause a subsequent test to underestimate
that vehicle's actual emission potential. Removing initially resident vapors from the
canister,  and replacing them with butane or  other  gasoline vapors,  may be
detrimental for several reasons. First, as vehicles age, their canisters gradually lose
capacity as larger molecules occupy available sites; these harder-to-purge molecules
should not be removed from the canister for the sake of test convenience, or for the
sake of getting the vehicle into a repeatable condition. Second, if in-use vehicles have
overloaded canisters, then EPA would want to test them in that condition. Third,
butane has not been demonstrated as a good substitute for either oxygenated fuels
or methanol.  On the contrary, manufacturers have acknowledged that, for existing
canister designs, purging methanol vapors is more difficult than purging gasoline or
butane vapors.  Furthermore, the initial 23-minute preconditioning drive should
enable a properly functioning vehicle to almost completely purge its evaporative
canister.  In that case, an additional purge would be redundant, and would require
unnecessary handling of the evaporative control  system.  A test would more easily
identify inadequate  and malfunctioning  systems, on the other  hand, by initially
loading the canisters from the as-received (unpurged) condition.
   2*Hydrogenous Catalysis in Practice, C.N. Satterfield, McGraw Hill Co. 1980. p.
28.
   eyrj
     For example, with a fuel vapor pressure of 9 psi (62 kPa) (e.g. 9 psi (62 kPa'
RVP fuel at 100° F (38° C)) and a barometric pressure of 14.7 psi (101 kPa), the vapor
space would be composed of approximately 60 percent fuel vapors and 40 percent air
(which is primarily nitrogen).

                                      35

-------
      Regarding the rate of vapor loading, it is not possible simply to set the loading
rate to simulate in-use experience. Test data from Automotive Testing Laboratories
(ATL) indicate that loading rates can vary from 2 grams per hour for  a moderate
diurnal heat build to 300 grams per hour or more for a long, high-temperature
drive.28  Also, as GM pointed out, vehicles with larger fuel tanks and canisters
generally experience higher  vapor  loading rates.  For its test, GARB  specified a
loading rate of 15 grams butane per hour. EPA agrees that this specification is a fair
representation of a canister's in-use experience,  and is also specifying 15 grams
butane per  hour to load canisters in the three-diurnal procedure.  Because the
canister is loaded with a known quantity of vapor at a low, constant rate, the process
is slow, but requires a technician only to set up the equipment and then turn it  off
after a certain amount of elapsed time.

      In contrast, for canister loading in the two-diurnal procedure, the apparatus
must be  monitored continuously until  breakthrough  is observed.  It is therefore
advantageous to accelerate the canister  loading rate and minimize the total loading
time.  As the loading rate is increased,  however, the canister becomes  less able to
retain  the  hydrocarbon vapors,  reducing the  canister  load at  the  point  of
breakthrough. EPA testing indicates that a loading rate of 40 grams butane per hour
is the best balance between these opposing factors.29  The data show that purged
canisters, loaded at a rate of 40 grams butane per hour to breakthrough, may hold
10 to 20 percent less vapors than the same canisters loaded at 5 grams butane per
hour.  The data also shows that canisters are  much less effective at storing vapors
at higher rates.  EPA is therefore specifying a loading rate of 40 grams butane per
hour to load canisters for the two-diurnal procedure.

      EPA recognizes GM's concern for avoiding a prolonged canister loading step for
larger vehicles.  If the above specifications require canister loading for longer than
12 hours (the minimum  soak time), manufacturers may submit data supporting a
faster rate.   The faster rate would be based on completing the canister loading
procedure in 12 hours.

Summary of Comments—Miscellaneous Issues

      Manufacturers suggested various means for detecting breakthrough with the
butane-loading  apparatus, including use of a trap canister that would indicate
   28A11 test data generated by ATL for EPA is stored in publicly available files in
a Micro database (account name "SMAJ") on the Michigan Terminal System (MTS)
at Wayne State University.

   29"E£Fect of Load Rate on Canister Load at Two Gram Breakthrough," EPA memo
from Dan Barba to Joanne I. Goldhand, December 10, 1992 (Docket A-89-18, item
IV-B-10).

                                     36

-------
attainment of the 2-gram breakthrough point when its mass had increased by  2
grams, or use of a mini-SHED to contain emitted vapors.

      Mercedes Benz and Ford requested that they be allowed to remove canisters
and replace them with stock canisters already loaded according to specifications.

      Manufacturers requested that the initial preconditioning drive be initiated as
a cold start  for  purposes of adaptive memory.  GM added that an oil sump
temperature could be specified to ensure that the test vehicle would experience a cold
start for the initial drive.

Analysis of Comments

      EPA believes that there is more than one acceptable detection method to
accomplish loading to the 2-gram breakthrough point.  The baseline method of
detecting this is to place the vehicle in a SHED to load the canisters) until 2 grams
of hydrocarbons are measured in the SHED. An auxiliary evaporative canister may
also be used to collect emitted vapors from the vehicle's canisters); vapors would be
loaded until the mass of the auxiliary canister increases by 2 grams.

      Evaporative canisters may not be removed from test vehicles.  A vapor hose
will typically have to be disconnected to load the canister,  but EPA wants to avoid
unnecessary  handling of vehicle components.   Such temporary,  intrusive vehicle
modifications may call into question the validity of test results if they affect the
integrity of the vehicle emission control system.  As a result, canisters cannot be
loaded with butane separately in a mini-SHED, and loaded stock canisters cannot be
used to substitute for a vehicle's existing canister.

      The test vehicle soaks  for a minimum of 12 hours (6 hours at EPA's option >
before the preconditioning drive to ensure that the vehicle is stabilized at the  test
temperature. EPA believes this stabilization period is sufficient to ensure that all
vehicles will have a cold start for the initial preconditioning drive.

F.    Hot Soak Test

EPA Proposal

      EPA proposed at the December 1990 workshop to conduct the hot soak test for
one hour at 95° F (35° C) ambient following the running loss test.  Based on current
regulations, seven minutes would be allowed between completion of the running lo-«
test and the  start of the hot  soak test; EPA proposed to decrease that time to fr-••*
minutes.  EPA proposed an overall air circulation requirement of 0.4±0.2 cfm p--r
cubic foot (0pm per liter) of enclosure volume.
                                      37

-------
Summary of Comments

      Honda requested that EPA allow the full seven minutes to transition from the
running loss test to the beginning of the hot soak test.  Commenters also requested
that the air circulation requirement be changed to 0.6±0.2 cfm per cubic foot (fpm per
liter) of enclosure volume to match the specification of the diurnal emission test.

Analysis of Comments

      Because the hot soak emission rate is highest shortly after a drive, it is critical
for the hot soak measurement to make the transition as quickly as possible.  EPA
expects five minutes to be enough time to move the vehicle from the driving cell for
the running loss  test to the  hot  soak enclosure. EPA agrees that  the  specified
circulation rate should match that of the diurnal emission test.

G.    Exhaust Emission Test

EPA Proposal

      EPA proposed at the December 1990 workshop to match fuel temperatures
during the exhaust emission  test to  a target  profile that the manufacturer would
develop on the road for the same driving cycle and roughly the same  conditions.
Current practice frequently allows the addition of a supplemental fan that directs air
at the fuel tank,  ostensibly to simulate underbody airflow, but often resulting in
overcooling of the fuel. This unrepresentative situation prevents the expected heating
of, and vapor generation from, the fuel tank  during the exhaust emission test, so
vehicles that have supplemental cooling during the test may not be able to maintain
control of exhaust emissions when vapors are  generated in use.

          of Comments
      CARB joined the auto manufacturers in questioning the benefits of requiring
fuel temperature control during the exhaust emission test, especially considering the
costs involved in the generation of the target profiles and the task of controlling fuel
temperatures during the test.
        of  orr>Tnents
      EPA agrees that the burden involved in requiring manufacturers to generate
fuel temperature profiles, and to control fuel temperatures according to that profile.
is too great compared to the modest benefits of that improvement to the test. EPA
is therefore not requiring fuel temperature control during the exhaust emission test
as part of this rulemaking.

      The Agency may later evaluate the need to revise the current procedure tor
justifying supplemental cooling in an effort to improve the exhaust emission test

                                     38

-------
Revisions in this area may include new requirements for more rigorous justification
for additional cooling, as well as changes to the method of supplying cooling during
the test.

H.    Heavy-Duty Vehicles and Engines

EPA Proposal

      EPA proposed in the August 1987  NPRM to revise the test procedure for
heavy-duty engines to require that the engine start the exhaust emission test with
a loaded canister. This change would ensure that engines could purge hydrocarbons
from the evaporative canister without  causing an increase in exhaust emissions.

      The Agency proposed in the January 1990 NPRM to make changes to  the
evaporative  emission test for heavy-duty gasoline- and methanol-fueled  vehicles
similar  to the changes to the test for  light-duty vehicles, i.e., initially loading  the
evaporative canister with vapor, and conducting consecutive diurnal heat builds after
the dynamometer run.

          f Cornrr|ent3
      Ford and Chrysler had serious reservations about the feasibility of passing the
proposed heavy-duty test, because of the magnitude of vapor generated from the fuel
tank and purged from the evaporative canister.  They pointed out that heavy-duty
vehicles have larger fuel tanks and different driving patterns that make it difficult
to purge the evaporative canister.
Analysis of
      Manufacturers have provided insufficient basis to warrant any change in the
Agency's position.  Although larger fuel tanks result in a need for a larger canister
capacity, the proportionately greater fuel consumption of these engines allows for the
purging of the additional stored vapor. EPA believes that compliance will be possible
through  proper  selection of existing hardware and technology.30  The different
driving patterns for heavy-duty vehicles are already reflected in the specified drm re-
cycle  for those vehicles.   EPA acknowledges that the  test  specifications tor
temperatures, fill level, and driving schedule may be especially challenging for hen \-v
duty vehicles; however, the fact that these challenging specifications represent rn.il
conditions reinforces the importance of finalizing the proposed changes to the t— t
procedure.
   30"0nboard and Evaporative Control System Cost Estimates for the SNPRM
EPA memo from Jean Schwendeman, to the Record, December 22,  1988 ' D-  k
A-89-18, item H-B-6).

                                      39

-------
I.     Fuel Spitback

      This rulemaking includes provisions to prevent fuel spitback during refueling.
Fuel spitback can be a problem when the design of the fuel filler neck is inadequate
to accommodate in-use fuel fill rates. The result can be fuel spillage, which is both
an environmental and a safety hazard.

EPA Proposal

      EPA proposed in  1987 and again in the January 1990 NPRM to  limit
commercial dispensing rates to 10 gallons (37.9 liters) per  minute in an attempt to
prevent spitback during refueling events.  At  the  January 1992 workshop, EPA
proposed also to test vehicles for spitback by collecting liquid fuel emissions during
a refueling event, either by collecting emissions in a bag, or by a visual observation
of spitback.

          of Comments
      Auto manufacturers acknowledged spitback as a legitimate emission source
that warrants control, but expressed concern about the ability of the proposed test
to measure spitback accurately, to distinguish between vapor and liquid emissions,
and to distinguish between spitback and dragout losses (fuel dripping from nozzle
after dispensing).  Manufacturers were opposed to the proposed reliance on visual
observation of spitback. Manufacturers wanted either to treat spitback as a customer
satisfaction  issue,  or  to deal with it  outside  the scope  of this rulemaking.
Commenters suggested that the issues be resolved either in a future workshop or by
the involvement of the Society of Automotive Engineers (SAE).

      The oil industry argued that spitback is independent of fill rate, that  auto
manufacturers  should bear the responsibility to  design their vehicles to  prevent
spitback, and that current dispensing rates have increased and are significantly
higher than  the proposed limit. They also argued that any limitation on dispensing
rates should include an exemption for fuel pumps devoted to refueling heavy-duty
vehicles, which have much larger fuel tanks.

        of Comments
      EPA is taking action to limit spitback during fuel dispensing, since spitback
is a known contributor to air pollution that may endanger public health and welfare
within the meaning of section 202(a) of the Clean Air Act.  EPA has conservatively
estimated a fleet average value of 0.15 grains per gallon (0.04 g/liter) for spitback
                                     40

-------
emissions.31  Nationwide, spitback emissions were projected to be nearly 4 million
gallons (15 million liters) per year.  Reducing spitback emissions would thus result
in an air quality benefit, a substantial fuel savings, and health and safety benefits
for in-use refueling events.

      Rather than relying only on a limitation of fuel dispensing rates, EPA is
depending on direct testing of vehicles to prevent spitback. Because the underlying
goal is to prevent in-use emissions, the Agency is implementing a test that simulates
the experience of concern.

      The spitback test thus consists of draining the vehicle's fuel tank, filling the
tank to 10 percent of its nominal capacity, operating the vehicle over one UDDS, then
promptly refueling the vehicle with at least 85 percent of the tank's nominal capacity
at 10 gallons (37.9 liters) per minute. Compliance would be determined by catching
liquid  emissions in a plastic bag secured around the dispensing nozzle and then
weighing the collected fuel.

      Heavy-duty vehicles over 14,000 pounds (6,400 kg) GVWR will not be tested
for spitback. These vehicles are typically designed with filler necks so short that fuel
can be dispensed directly into the fuel tank. These vehicles would therefore not be
expected to experience spitback.  A small number of these heavy-duty vehicles may
have filler necks long enough to  make spitback possible; however, the overall air
quality benefit of extending the spitback test to these vehicles is negligible.

      EPA has modified the proposed spitback test to accommodate manufacturers
concerns.  First, the test now clarifies that the vapor in the spitback collection bag
must be expelled before weighing. Second, the test specifies a means of handling the
nozzle to  prevent any dragout losses from affecting measured spitback emissions.
Also, the  final  rule establishes a standard of 1 gram  per test to  ensure that the
accuracy of the procedure was sufficient to determine compliance.  Finally, the test
specifies the use of a commercially available  dispensing nozzle.  Any issues of
nozzle/filler neck compatibility are beyond the scope of this test procedure.

      EPA conducted testing to develop the spitback collection procedure. Emissions
were collected in a rectangular tedlar bag,  approximately 15" x 20", with two small
openings on opposite ends of the bag to allow passage of the dispensing nozzle. Each
opening was fitted with a tedlar insert  for the bag to be clamped onto the filler pipe
and the dispensing nozzle, ensuring that  liquid emissions are trapped in the  bag
One side  of the bag  was left open to allow displaced fuel vapors to escape easily
Upon completion of the dispensing operation, the nozzle was removed from the
vehicle and the bag, with the opening of the nozzle held up to prevent any dripping
    o -i
      Investigation of the Need for In-Use Dispensing Rate Limits and Fuel Nozzle
Geometry Standardization," EPA Technical Report, May 1987 (Docket A-89-18, it^m
IV-A-2).

                                      41

-------
into the bag.  Then the bag was folded several times to retain the trapped liquid fuel,
to eliminate  any vapor space,  and  to facilitate weighing the bag.   Because the
procedure was effective in collecting all (and only) liquid fuel emissions, it served as
the basis for  drafting the regulatory language for the spitback test.

      In addition, the final rule limits in-use dispensing rates to 10 gallons (37.9
liters) per minute. With a vehicle test for spitback in place, the limitation on in-use
dispensing rates ensures that the test specifications will reflect actual dispensing
conditions. Because of the minimal cost of complying with the rate limitation, as
described  in  Chapter 4, EPA does not believe that a trend toward higher in-use
dispensing rates, if it exists, would be an obstacle to meeting the new requirements.

      Insufficient basis was provided for the comment that spitback is independent
of fill rate. EPA test data indicate that higher flow rates are associated with a more
frequent occurrence of spitback.32 Furthermore, a consideration of the mechanics
involved in the spitback phenomenon indicates that, although the configuration of the
filler neck/nozzle interface plays  a major role, it would be difficult to envision a
situation in which the rate of fuel crossing the interface is not also important.

      EPA agrees that the limitation on dispensing rates should not extend to pumps
that service only heavy-duty vehicles. Such dedicated pumps would be expected to
service  primarily  the heavy-duty vehicles that are  exempt from spitback  test
requirements, as described above.  All other gasoline- or methanol-dispensing pumps
belonging to retailers or wholesale purchaser-consumers are subject to the 10 gal/min
(37.9 liters/mm) limit.
J.    Methanol-Fueled Vehicles

EPA Proposal

      EPA proposed that the regulations apply to both gasoline- and methanol-fueled
vehicles.

Summar  pf
      For flexible-fueled vehicles, auto manufacturers objected to the use of low-level
blends of methanol for testing.  A mixture of 10 percent methanol in gasoline has a
volatility of about 12.5 psi (86.2 kPa) RVP, which causes a much greater amount of
vapor  generation than most  other  compositions of  methanol  and  gasoline
    32"Application of Onboard Refueling Emission Control System to a 1988 Ford
Taurus Vehicle," EPA technical report,  EPA-AA-SDSB-91-06, page  36 ff. (Docket
A-89-18, item IV-A-6).

                                      42

-------
Manufacturers requested a different fuel composition for flexible-fueled vehicles, or
at least an extended time before requiring use of the worst-case fuel.

      Several manufacturers requested additional lead time to comply with the new
test requirements for flexible-fueled vehicles. The purge requirements at slow speeds
for varying fuel vapor composition, and the need to prevent permeation were cited as
the most challenging aspects of the proposed test that warranted additional time for
development.

      Manufacturers identified the need for flexible-fueled vehicles to have additional
preconditioning whenever the test fuel would be changed to prevent a test run from
being affected by previously used fuels.

Analysis of Comments

      In a separate rulemaking, EPA set out the requirements for composition of test
fuels for flexible-fueled vehicles (54 FR 14426, April 11,1989). EPA concluded at that
time that, to maintain control in all expected in-use conditions, vehicles should be
tested with the worst-case fuel mixture. EPA therefore specified a test fuel mixture
of 90 percent gasoline and 10 percent methanol (M10) for evaporative testing.

      As described in Chapter 3, EPA expects that flexible-fueled vehicles will not
need to be designed to meet the new evaporative emission test requirements until the
last  year of  the  phase-in schedule.    The  approximately 6V6 years  between
promulgation of EPA's final rule and the first sales of flexible-fueled vehicles subject
to the new evaporative test requirements provides much more lead time than
provided by past EPA actions.  Moreover, manufacturers were aware of CARB's
similar requirements adopted in August 1990 and so have had some time to prepare.
Concerning permeation, the industry is currently making great progress in improving
the permeation-resistance of materials. Since even gasoline-fueled vehicles need to
be able to prevent  permeation of methanol and other  oxygenated compounds, much
of the materials development and selection  for flexible-fueled vehicles  is  well
underway.  EPA therefore believes that the specified lead time is sufficient to design
vehicles to meet test requirements, within the meaning of section 202(a)(2) of the Act.

      Testing with different fuel mixtures does require additional preconditioning
when the test fuel is changed.  EPA has drafted a procedure of repeated drives and
refuelings, based on the procedures of the Auto/Oil research effort, for inclusion in the
final  regulations.33    The  procedure consists  of  purging and   reloading  the
evaporative canister, draining and refilling the fuel tank, starting the vehicle several
times, and driving through one UDDS.
   33Telefax from  Dave  Brooks, Auto/Oil Air Quality  Improvement Research
Program, March 17, 1992 (Docket A-89-18, item IV-D-81).
                                     43

-------
K.    Other Issues

      In addition to the areas covered above, comments were received on a number
of other issues. These comments are discussed below.

      EPA proposed various methods to minimize the impact of invalid tests.  These
included an accelerated procedure to bypass a portion of the test that had been
successfully completed before a test error, and a means of accepting test data after
a test error, depending on how the error affected the test results.  Manufacturers
objected, arguing that the proposed provisions might modify test requirements, and
would encourage sloppy testing.  In response, EPA does not intend to pursue these
methods to deal with invalid tests. Instead, EPA has  attempted to minimize the
potential for test errors in the development of the test procedure.

      Some commenters requested that EPA change some of the tolerances for test
parameters.   Commenters wanted to change the  tolerance  on the  ambient
temperature  of the  diurnal emission test  from ±1°  to  ±3° F  (0.6°  to 1.7° C).
Commenters also  wanted to relax the proposed specification to maintain enclosure
pressures within  0.2 inches (0.5 cm) of water of barometric pressure,  allowing a
difference of up to 2 inches (5 cm)  of water.  EPA  agrees that  both  of these
specifications should be changed to make the test procedure easier to execute.  To
avoid too much variability in diurnal emission test, EPA is retaining the specification
that, on average for the whole test, ambient temperatures need to be within 2° F of
the target profile. Neither of these changes should affect test stringency.

      Honda requested that EPA change its specification for the volatility of test fuel
from 9  to  7.8 psi (62  to 53.8 kPa) RVP for all testing, because new volatility
requirements specify 7.8 psi (53.8 kPa) RVP for nonattainment areas. However, EPA
is retaining the specification of 9 psi (62 kPa) RVP test fuel. Decreasing the volatility
of test fuel is inappropriate because 9 psi (62 kPa) RVP fuel will still be widely
available in much of the country. EPA's volatility requirements limit summer fuel
volatilities throughout the continental U.S. to a maximum of 9 psi (62  kPa) RVP;
nonattainment areas in some warm climates or at high altitudes have a
fuel volatility of 7.8 psi (53.8 kPa) RVP (55 FR 23658, June 11, 1990).  States with
the lower volatility requirement may, however, justify changing to a maximum of 9
psi (62 kPa) RVP.  Because 9 psi (62 kPa) RVP fuel will be widely available in the
U.S.  for  the foreseeable future, EPA is requiring manufacturers to demonstrate
sufficient control on that fuel.

      API requested that EPA specify 10.5 psi (72.4 kPa) RVP test fuel to allow for
a higher in-use fuel volatility without foregoing  emission  control.  Increasing the
volatility of the test fuel is also inappropriate; EPA's rulemaMng to establish
maximum volatility levels demonstrated that it was cost-effective to reduce in-use
volatilities to the current levels. Moreover, the Clean Air Act forbids the sale of any
summer fuel with a volatility higher than 9 psi (62 kPa) (section 211(h)).
                                     44

-------
      Manufacturers requested specification of 7.8 psi (53.8 kPa) RVP test fuel for
high-altitude testing, and wanted confirmation that EPA would not change its policies
by requiring vehicle testing for high-altitude. EPA agrees that 7.8 psi (53.8 kPa) RVP
fuel is appropriate for high-altitude testing, since EPA's regulations limiting summer
fuel volatilities establish 7.8 psi (53.8 kPa) RVP as the maximum volatility level for
high-altitude areas (55 FR 23658, June 11, 1990). For this rulemaking, EPA is not
changing  its  policy  on  the  need  for high-altitude  testing  for  certification.
Manufacturers may continue to submit test data indicating compliance  with
standards, or, in lieu of testing, may submit a statement that they can meet the test
requirements, based on sound engineering judgment.  EPA may in the future require
high-altitude testing if there is an indication that  testing is needed to  improve
emission control in these areas.

      GM objected to EPA's proposed regulatory text that required all fuel vapors to
be routed to the evaporative  canister or to the engine.   GM claimed  that the
requirement was not clear, and that EPA had communicated no plan to enforce the
provision.   EPA believes  that  the  provision is  straightforward and involves no
ambiguities  that unduly jeopardize  a  manufacturer's certification.   Enforcement
would be based  on an engineering evaluation of the fuel system for each vehicle
design submitted for certification. EPA would expect vapors  to be routed to the
evaporative canister without any restriction or release valve that could cause vapor
venting under normal operating conditions, other than vehicle refueling. This would
be similar to EPA's successfully implemented requirement for eliminating engine
crankcase emissions.

      Ford requested  that EPA not require the new test procedures for Selective
Enforcement Audits (SEAs) at assembly facilities.  Ford wanted to avoid evaporative
testing on newly assembled vehicles because of the lack of measurement facilities and
the concern  for high levels of nonfuel emissions  from new polymeric components.
Ford also wanted to avoid the canister  preconditioning for exhaust-only testing,
because of the facility requirements, and because  of the time constraints in the
assembly process. EPA will continue the practice of omitting evaporative testing from
the SEA process because  of the  concern for high nonfuel emissions from newly
assembled vehicles. EPA may, however, conduct the new canister preconditioning for
SEA exhaust testing.  Purging canisters, and then loading them with butane and
nitrogen, requires very little equipment and adds  little time to the procedure to test
for exhaust  emissions.  Moreover, the ability to control exhaust emissions while
purging a loaded canister is central to any new requirements for evaporative emission
control.

      GM requested that EPA hold a public workshop after it publishes the final rule
to discuss detailed improvements to the test procedure. GM expected the  industry
to gain experience in the early stages of complying with test requirements, making
possible a set of improvements to the test that would make it easier to run. EPA is
resolving as  many technical issues as possible before publishing this final  rule. After
publication,  EPA expects to work with manufacturers in setting up test facilities and

                                     45

-------
initiating testing. EPA will consider the need for any amendments as part of that
process.

L.    Comparison to CARB's Adopted Test

      In addition to the differences in test fuel RVP and  diurnal/running loss
temperatures, and the inclusion  of the supplemental two-diurnal test, the test
procedure being adopted by EPA differs from the one adopted by CARB on a number
of lesser points. These are listed below. Many of these differences are also discussed
in the previous sections of this chapter but are repeated here in order to present a
complete comparison. Because of the expressed desire of many commenters for
maximum consistency between the EPA and CARB procedures, the reasons for the
differences are also provided.   Note that some  of these deviations from the CARB
procedure are desired by manufacturers.

Running loss test

o     Delete requirement for vapor temperature control; instead specify narrowed
      fuel temperature tolerance at the end of the running loss test
            Vapor temperatures are very difficult to control independent of liquid
            fuel temperatures, considering that a single control mechanism, ambient
            heating of the vehicle, is used.  To better control the vapor generation
            during the running loss test, the temperature tolerance  for the liquid
            fuel is instead narrowed from ±3°  F (1.6° C)  to ±2° F (1.1° C) for the
            last two  minutes.

o     Delete requirement for proportional-speed fan for vehicle's full frontal area
            EPA believes that specifying a proportional-speed fan for the vehicle's
            radiator provides very little improvement to  the test  simulation; the
            large expense for the fan is therefore not justified. EPA is maintaining
            the  current  specification for radiator  cooling  during  the exhaust
            emission test  (40  CFR 86.135-90(b)).   Also, EPA  is  adapting
            specifications for underbody circulation, and direct blowing on the fuel
            tank, to  provide adequate fuel temperature control.

o     Replace the second UDDS in the driving  cycle with two NYCCs
            Thia provides a broader range of driving patterns, including more low-
            speed and idle operation, thus ensuring a more robust running loss test,
            as described  in Chapter 2.

o     Delete allowance for initial fuel temperature below 95°  F  (35° C)
            Two-temperature systems, such as  would be required at the start of the
            test under this provision, are inherently unstable, thus making the test
            prone to voiding.  Furthermore, lower temperatures are unwarranted.
            since they  would  less accurately reflect  ozone-prone  summertime
            conditions, as required by statute.

                                     46

-------
o     Allow driver to put car in neutral between cycles
            This change allows the  driver to rest or stand twice  for up to two
            minntea during the drive. This should make it easier for the driver to
            continue accurate driving through the whole 70-minute drive. No effect
            on the vehicle's emission performance is expected.

o     Increase initial stabilization period from one to four hours maximum
            EPA and manufacturers believe that, in some cases, more time is
            needed to stabilize fuel temperatures to 95° F (35° C) to start  the
            running loss test.

o     Specify maximum fuel tank pressure allowance of 10 inches of water (2.5 kPa)
      for both SHED and point-source methods
            Maintain consistency between optional test methods.

o     Define provision to deal with high coolant temperatures
            The GARB test monitors coolant temperatures but provides for no action
            based on this information. EPA clarifies the meaning by specifying that
            a dashboard warning of high engine temperatures should prompt the
            termination of the test run.

o     Add provision to allow continuous sampling
            With either the point-source or the enclosure method, emissions could
            be measured continuously by a flame ionization detector, or collected in
            bags for subsequent measurement.

o     Add minimum overall circulation requirement (1.0 cfin/ft3) (fpm/liter) for the
      enclosure method
            CARB's specification of fans for the vehicle's radiator, underbody, and
            fuel tank provide no specification to prevent temperature or hydrocarbon
            stratification in the enclosure.

o     Delete allowance to correct suspected  equipment errors
            This provision leaves unwarranted discretion to testing personnel
            Clearly, if equipment is  shown to be in error, test results may be
            invalidated.

o     Clarify that cap removal is not allowed during the test
            Modify language to avoid the  possible interpretation of the provision
            limiting pressures to allow cap  removal during the running loss test

o     Change test speed resolution during the running loss test to ±0.1 mph < 0 N
      km/hr), once per second
            Speed measurement frequency  reflects  current test specifications
            Resolution  is  relaxed  somewhat from  current  EPA specification-:
                                     47

-------
            because fuel temperatures, and thus emissions, are not sensitive to
            small variations in vehicle speed.

      Lower minimum surface temperature for running loss test from 70° to 40° F
      (21° to 4.4° C)
            Specifying 70° F (21° C) as a minimum surface temperature makes it
            very difficult to maintain ambient temperatures within EPA's specified
            range of 95±5° F (35.0±2.8° C). Allowing surface temperatures as low
            as 40° F (4.4° C) does not affect the accuracy of testing.
Diurnal test
      Do not allow subtraction of nonfuel background emissions
            EPA applies the evaporative standards to measured fuel and nonfuel
            emissions.

      Set underbody  circulation to 5 mph  (8 km/hr); add discretion to  ensure
      temperature swing representative of that experienced in use
            Circulation is required to ensure that fuel in an uninsulated fuel tank
            experiences appropriate heating; discretion is required to ensure that
            airflow is not defeated by vehicle designs that provide more effective
            insulation during the test than actually occurs in use.

      Define average tolerance (±2° F (1.6° O) for cycled temperatures
            Average tolerance is defined as the average of the absolute values of the
            differences between measured and target temperatures. Not considering
            absolute values would allow a significantly smaller temperature range.

      Allow 6 to 36 hours between the hot soak test and the diurnal emission test,
      with the last 6 hours held at the initial ambient temperature for the diurnal
      emission test
            The stabilization of the test vehicle at  the initial test temperature
            makes it easier to start the diurnal emission test.
Preconditioning
      Define working capacity in terms of 2-gram breakthrough
            Working capacity has various definitions in industry, and GARB left the
            term undefined.   EPA chose the 2-gram breakthrough point to  be
            consistent with its own procedures, and with the understanding of most
            participants, including GARB.

      Add option to conduct additional preconditioning driving on a dynamometer
      (not just outdoors); delete 20-mile (32 km) minimum for extra driving.
                                     48

-------
            EPA does not drive test vehicles outdoors, but would like to maintain
            the flexibility of additional vehicle operation to prepare vehicles for
            testing.

      Allow canister loading in three-diurnal test at faster rates if more than 12
      hours is required
            EPA believes that testing personnel should be able to start the exhaust
            emission test within 12 hours after the end of the preconditioning drive.
            For some very large vehicles, canister capacity may require more than
            12 hours of canister loading at the specified rate of 15 grams butane per
            hour. For those vehicles, manufacturers may alter loading rates and
            times to accomplish the full canister  loading in  a 12-hour period.
            Allowing the accelerated 12-hour loading time is consistent with the
            experience of a parked vehicle undergoing a diurnal heat build.

      Specify a maximum period of one hour to  refuel the test vehicle after the
      preconditioning drive
            CARB's test specified no time requirement for the refueling event

      Change humidity specification during canister purge to 50±25 grains per pound
      (7.1±3.6 grams per kilogram) of dry air
            Lowering the humidity specification  and widening the tolerance will
            make testing easier, with no loss in the effectiveness of the test.
Hot soak test
      Specify ambient temperatures:  95±10° F (35.0±5.6° C) for first five minutes.
      95±5° F (35.0±2.8° C) for remainder of hot soak test
            CARB's regulations included inconsistent specifications, requiring either
            55 or 60 minutes of temperature control during the hot soak test.  EPA
            believes that the temperature tolerance should be relaxed for the first
            five minutes of the test, because of the difficulty of sufficiently cooling
            the SHED at the beginning of the hot soak test. This is consistent with
            the vehicle's experience outdoors, since a  newly parked vehicle  cools
            gradually, rather than being exposed to high rates of circulation with
            cool air.

      Require the hot soak test to start no more than five minutes after completion
      of the running loss test, and no more than two minutes after engine shutdow n
            Most of the vapor generation (and thus emissions) from the hot soak test
            come  in the first few minutes of the  test period.   It is  therefor^
            important to minimize the time between the moping loss test and th«-
            beginning of the hot soak test.  EPA's testing experience indicates th.it
            five minutes is a sufficient allowance.  The 2-minute specification <.~
            consistent with established test practices.
                                      49

-------
o     The initial hydrocarbon measurement is taken before the vehicle enters the
      enclosure
            With this change from current practice, the test would account for all
            hot  soak emissions from the instant the enclosure is sealed.

Fuel temperature profile

o     Change "representative" profile to "highest expected" profile
            Evaporative families, based on vehicle characteristics such as canister
            capacity and tank size, include no means of grouping vehicle models by
            their fuel  temperature increase  during driving.  Nevertheless  fuel
            temperatures  are the most basic variable involved in conducting the
            running loss test.  Therefore, only a test vehicle with the highest fuel
            temperatures  would  ensure adequate control for all the vehicles it
            represents. Manufacturers are allowed to develop fuel temperature
            profiles for subsets of an evaporative family (e.g., for an engine family,
            or for an individual model), at their discretion.

o     Require use of any vehicle options that limit underbody airflow
            Because features such as air dams and running boards can dramatically
            influence fuel temperatures during driving, they must be in place during
            development of a vehicle's fuel temperature profile. Not requiring these
            optional features  would   allow many  in-use  vehicles to  routinely
            experience conditions more severe than test conditions.

o     Relax driving tolerances to ±4 mph (6.4 km/hr)
            EPA's experience  with fuel temperature profiles indicates  that  fuel
            temperatures  are  not sensitive to small variations in vehicle speed.
            Relaxing the driving tolerance would make it easier to complete test
            runs without sacrificing the accuracy of the data collection.

o     Allow temporary deviation from the driving schedule—up to three times for a
      maximum of 15 seconds each
            Outdoor driving according  to a prescribed schedule poses some risk of
            collisions with other vehicles.  Departing from the driving schedule for
            up to 45 seconds during the 70-minute drive should minimize that risk
            without affecting the validity of the data.

o     Increase the required  pavement temperature  to 30° F above  ambient.
      measured  throughout the drive
            The higher pavement  temperatures  are critical  in developing  fuel
            temperature profiles that represent the vehicle's behavior in ozone-prone
            summertime conditions.  In addition, because pavement temperatures
            are very sensitive to instantaneous solar loading, temperatures must be
            measured  throughout the  driving period, rather than only at the
            beginning  and end of the drive.

                                     50

-------
      Measure ambient temperature and wind speed throughout the drive
           Measurement of ambient conditions throughout the drive ensures that
           the desired conditions are met during the whole drive and minimizes the
           potential for measurements that are not characteristic of the drive
           period.

      Specify standard procedure for measuring ambient temperature and wind
      speed, and require submission of weather station data
           EPA has observed that temperatures reported by testing personnel are
           sometimes much different than those reported by local airports.   To
           resolve  differences  in these measurements, EPA has incorporated
           requirements  to  meet federal  standards for  making  ambient
           measurements.  These procedures should make the test requirements
           clear, objective, and uniform for all involved. Obtaining data from the
           nearest weather station provides a valuable but inexpensive check on
           the data collected at the test site, particularly for the very subjective
           cloud cover assessment.

      Require pavement temperature  measurement to be  made on a  surface
      representative of the driving surface
           Measurement on a surface with a different color, texture, or composition
           could cause erroneous results.

      Calculate fuel profile independent of ambient temperatures
           CARB's procedure unnecessarily ties  fuel temperatures to measured
           ambient temperatures.  With EPA's approach, fuel temperatures must
           be reported as a relative profile (e.g., a 25° F (13.9° C) temperature rise).
           independent of  ambient temperatures.   The relative profile  is then
           adjusted to the initial fuel temperature of 95° F (35° C) to establish the
           absolute profile for the running loss test.

      Require submission of results from  all valid tests and create a composite
      profile
           Consistent with data submission  requirements for  manufacturers
           emission testing, EPA is requiring the submission of all test data for
           fuel temperature profiles. Because fuel temperature profiles can vary
           significantly, even under constant ambient conditions, manufacturers
           must create (based on simple averaging) a composite fuel temperature
           profile from all valid tests for a vehicle model. This composite profile is
           to be used in running loss testing.
Equipment
      Define the  nominal  volume  of  variable-volume  SHED  based  on rrud-
      temperature of diurnal range
                                     51

-------
            Because the  volume in  a variable-volume  SHED  depends  on the
            temperature, an average temperature must be used to determine SHED
            volume.

o     Delete maximum surface temperatures in running loss enclosures
            EPA does not believe that this provision  is necessary.  Furthermore,
            engine and exhaust surface temperatures, as well as the specified upper
            tolerance for tank temperature control, can easily exceed the "global"
            maximum temperature specified in the GARB procedure.

o     Increase  diurnal test and hot  soak test air circulation to 0.8±0.2 cfm/ft3
      (Cpm/liter)
            CARB's specification of 0.3 to 0.6 cfm/ft3 (Cpm/liter) is increased to
            provide additional circulation for adequate heating and cooling during
            the diurnal emission test. The circulation rates for  the hot soak test
            match those for the diurnal emission test.

o     Specify enclosure ambient pressure of 0 to -2 inches of water (0 to -0.5 kPa) for
      fixed-volume enclosures
            Maintaining a slight  negative  pressure  would simplify the  needed
            controls, without affecting the integrity of the enclosure or the test
            results.

o     Adapt the enclosure calibrations to EPA's test specifications
            Ambient temperatures and other detailed specifications have been
            changed to adapt to EPA's test requirements.  If manufacturers request
            that EPA accept a demonstration of calibration according to CARB's
            procedures, EPA would consider such a request under the provisions for
            alternate equipment configurations (40 CFR 86.106-9
-------
            EPA has separately established M10 as the appropriate fuel for testing
            flexible-fueled vehicles (54 FR 14426, April 11, 1989).
Miscellaneous
      Omit specified procedures for generating deterioration factors
            Because there were no proposed changes to the requirements for
            manufacturers to submit  durability  data, EPA  will not change its
            current  policy of allowing manufacturers discretion in establishing
            evaporative deterioration factors.

      Revise format, adjust constants for evaporative calculations
            Changes are necessary to maintain consistency with existing language
            in the Code of Federal Regulations.

      Add  language for testing with methanol, as appropriate
M.    Adjustments to Test Tolerances for EPA Testing

      EPA has identified the following specifications from CAKB's test that can be
relaxed, without increasing the stringency of the test, to minimize the number of
invalid tests  and avoid unnecessary facility modifications.  In order to maximize
consistency with the CARB procedure, EPA will retain the CARB specifications in its
regulations (that is, without these adjustments) and add provisions for conducting
testing with each of these adjustments at EPA's option.

Running loss test

o     Delete need for heated pumps and sample lines (point-source)
            EPA believes that heated pumps and sample lines are not necessary to
            prevent condensation of collected emissions during running loss testing
            by the point-source  method.   The relatively  low concentrations of
            hydrocarbon or methanol vapors in the dilution stream ensure that the
            dew point of the vapor at  any time during the test stays well below
            95° F (35° C).   Furthermore, any error caused by such  condensation
            would directionally decrease the stringency of the test.

o     Omit measurement of fuel tank pressure
            Omitting  the  pressure measurement will clearly  not  affect  test
            stringency. EPA may  choose not to measure fuel tank pressures in
            order to simplify testing.
                                     53

-------
Preconditioning
      Change initial cold soak to 6 hour minimum (from 12 hours)
            EPA believes that 6 hours is adequate to stabilize the vehicle prior to
            the preconditioning drive.   Since no evaporative  measurement is
            involved in this portion of the procedure, the shorter soak time should
            have no effect on any subsequent evaporative measurement.
                                     54

-------
               Chapter 3 Technological Feasibility and Lead Time
A.    Technological Feasibility

      Because the new evaporative test procedure implements a performance-based
standard, manufacturers have substantial flexibility in deciding how to upgrade
vehicle designs in response.  However, several likely modifications are evident from
comments and discussions with industry:  larger canisters, better canister purge,
permeation-resistant fuel and vapor lines, anti-spitback features, and some means of
limiting fuel tank temperatures. It is not expected that all of these changes would
be required on all models.

      No manufacturer  has indicated  that the new test procedures represent
requirements that are technologically infeasible.  Even for EPA's last proposed test
sequence, manufacturers challenged the cost-effectiveness, but not the feasibility, of
compliance.  Industry commenters almost universally requested that EPA finalize a
test procedure based on CARB's adopted procedure.  Therefore, because the test
procedure being  finalized is based  on  CARB's  procedure,  there is even less
uncertainty concerning technological feasibility. Because manufacturers have been
aware of CARB's test requirements for over two years, and must soon certify vehicles
with  it, EPA  believes  that  industry  support  of  CARB's  procedure  implies
manufacturer acknowledgement of the technological feasibility of the test.

      Larger canisters involve no major technological challenges. The required size
increase is modest, a liter or two for light-duty vehicles.  Furthermore, more efficient
carbon media are available that would minimize the size increase, if vehicle designers
desired to do so.  A straightforward redesign of the internal structure of the canister
may be necessary to ensure a good distribution of airflow as well. Chapter 4 provides
additional discussion on the likely hardware changes involved in producing larger
canisters.

      The primary need for demonstration of feasibility in the revised test procedure
relates  to the requirement  of purging  a loaded evaporative  canister during the
exhaust emission test. There are two related requirements: the vehicle must be able
to purge enough vapor from its canister to create capacity for the subsequent diurnal
heat builds, and the vapors must be purged from the canister without  simply being
passed unburned through the engine as exhaust emissions.

      Manufacturers have various means to create sufficient vapor storage capacity
First, increasing the rate of purge would cause the vehicle to more quickly  remove
vapors from the canister. Since the existing evaporative emission test has very little
purge requirement, the purge design of current vehicles varies greatly. For example.
GM has quoted 1V2-13 fl?/LA-4 (42-368 liters/LA-4) as the range for purge volume for
                                      55

-------
their current vehicles.34'35   EPA believes that increased purge rates will be one
of the most important results of revising the test procedure. Second, manufacturers
could incorporate methods to reduce the magnitude of vapor generation, thereby
reducing the amount of vapor that would have to be removed from the canister to
create the needed storage capacity. Finally, if the canister volume is insufficient, in
spite of aggressive purging, to create the required capacity, a larger canister would
make it possible  to remove enough  vapor with lower  purge rates.   Different
combinations  of  canister size  and  purge  capability could  achieve  the same
performance.

      GM, in modeling for CARB's adopted test procedure, assumed a purge rate of
10 ft3/LA-4 (283 liters/LA-4) for a vehicle with a 20-gallon (76-liter) fuel tank.36
This falls within the range of values EPA would expect from vehicles designed to
meet the new test requirements. GM estimated further that passing a supplemental
two-diurnal test would require a 25 percent increase in purge rates.  EPA believes
that the test parameters of its supplemental two-diurnal procedure have been relaxed
enough  that the  additional test requirements verify sufficient purge, without
increasing test stringency. Nevertheless, GM's estimates make clear that the purge
rates required by EPA's test are within the range of current technology, and are
therefore feasible for all vehicles.

      Test data indicate that  some recent-model  vehicles are already capable or
nearly capable of managing the required amount of vapor without exceeding exhaust
emission standards. During  the exhaust emission test, a typical vehicle would  have
to consume about 75 grams  of vapor, roughly six percent of the vehicle's total fuel
requirement, to restore canister storage capacity for two diurnal heat builds.37
Methods to reduce vapor generation from the fuel tank would reduce this amount and
make it easier to meet the performance requirements of the test, as described above.
Testing by Ford showed that  a current model vehicle was able to purge up to at  least
116 grams during an exhaust emission test with no more than a ten percent increase
    4The rate of purge is quantified in terms of volume of ambient air (in ft ) that
is drawn across the evaporative canister during the UDDS (or LA-4) driving cycle.

   35Transcript from January 28,1992 public workshop, p. 102 (Docket A-89-18, item
IV-F-4).

   36Letter from Samuel A. Leonard, GM,  to Richard D.  Wilson, EPA, March 23,
1992 (Docket A-89-18, item IV-D-78).

   37With  a fuel economy of 25 mpg (9.5 J/100 km) and a  fuel with specific gravity
of 0.74, a vehicle would need 1,246 grams of fuel to drive 11.1 miles.

                                     56

-------
in exhaust hydrocarbons.38  Testing by GARB on 1990 and  1991 Buick LeSabres
showed a negligible effect of increased purge vapors on exhaust emissions, even when
the vehicles were re-equipped with large, heavily loaded canisters.39  Based on this
information, EPA believes that the technology already exists for vehicles to meet the
purge requirements of the revised test procedure, and to achieve all other emission
standards as well.

      The current evaporative test procedure provides little or no control of resting
losses. The new evaporative test procedure measures emissions in a SHED over long
periods of time and will, therefore, control resting losses.  One technology that may
be employed in response to this requirement is the use of permeation-resistant fuel
lines and fuel  vapor lines. Steel lines provide good permeation resistance and are
already in  use in these applications.  Although the  use of steel lines could be
maximized, it  is  expected that the need for some  flexible segments will remain.
Newly developed teflon-coated nylon tubing provides very good permeation resistance
and could be used in these applications.   Chapter 4 provides additional discussion
of the costs associated with these design changes.

      Plastic fuel tanks, by being more susceptible to permeation, also face a new
challenge in limiting resting losses.  One manufacturer, however, by developing a
technology to limit permeation, has committed to supplying plastic fuel tanks that
emit less than 0.1 gram in 24 hours.41  EPA expects that  auto manufacturers can
initially, or in  the near future, meet test requirements with plastic fuel tanks.

      Anti-spitback valves are already used in some vehicles.  EPA test data show
that some vehicle designs are capable of meeting the spitback  requirements at a
dispensing rate of 10  gallons (37.9 liters) per minute. There are no technological
barriers to more widespread use of these valves if manufacturers determine that the
valves are needed to comply with the anti-spitback test requirements.

      To control running losses,  manufacturers  are  expected to  manage fuel
temperatures to ensure that fuel boiling does not occur during the running loss test.
   38"EPA Meeting with Ford," EPA memo from Alan Stout to Docket A-89-18, July
17, 1991 (item IV-E-13).

   39"Effects of Vapor Purging on Exhaust Emissions," GARB memo from Michael
Carter to Bob Cross, February 4, 1992 (Docket A-89-18, item IV-D-82).
         example, Pilot Industries has available its "P-CAP" line of permeation
resistant fuel lines and fuel vapor lines.  Letter from Edward K. Krause, Pilot
Industries, Inc., to Alan Stout, EPA, August 25, 1992 (Docket A-89-18, item IV-D-92

   41Letter from Norman W. Johnston, Solvay Automotive, to Richard D. Wilson
EPA, February 3, 1992 (Docket A-89-18, item IV-D-67).

                                     57

-------
This can be accomplished by several methods, including rerouting fuel lines, adding
insulation or heat shields, and locating fuel pumps outside of fuel tanks (described
in more detail in Chapter 4). None of these methods of control pose any technological
difficulties.  EPA expects  such methods to be adequate to meet requirements for
running loss control.

      The limit of 10 gallons (37.9 liters) per minute on fuel dispensing rates poses
no new technological challenge. Fuel nozzles that can accomplish this purpose are
already in use in the U.S., at little or no  additional cost, as discussed further in
Chapter 4.  The limited flow rate is  achieved with  a simple spring-loaded device
installed in the fuel nozzle.

B.    Lead Time

EPA Proposal

      The implementation date originally  proposed in August 1987 was the 1990
model year.  Most recently, however, EPA proposed at the January 1992 workshop
to adopt the following phase-in schedule, which was also adopted by CARB:


           Model Year                  Percent of Production

              1995                              10

              1996                              30
              1997                              50

              1998                             100
      At the January 1992 workshop, EPA proposed to implement the 10 gallon (37.9
liter) per minute dispensing limit June 1, 1993.

Summary
      After  the January  1992 workshop,  auto  manufacturers agreed with  the
proposed phase-in schedule, provided there were no substantial differences between
CARB's and EPA's test procedures, and provided that the nilemakLng was concluded
promptly.  In September  1992 EPA released draft regulations for the final test
procedure.42  In response,  manufacturers indicated  that EPA's final rule had been
delayed enough that compliance in the 1995 model year was no longer possible, sine?
   42Docket A-89-18, item IV-C-10.

                                     58

-------
there was insufficient time to make significant design changes or to complete federal
certification testing for 1995 model year vehicles.

      There were additional requests for delayed implementation for various special
concerns, including small manufacturers, and manufacturers of methanol-fueled and
heavy-duty vehicles. Daihatsu requested that the new test be implemented for all
vehicles in  the 1997  model year, arguing that it  would  be most  difficult  for
manufacturers with only a small number  of families to meet test requirements.

      EPA received no comments regarding the lead time for the limit on in-use fuel
dispensing rates.

Analysis of Comments

      Three different sections of the Clean Air Act provide guidance for determining
lead time  for  test  requirements.   These are sections 202(k),  202(a)(2),  and
202(a)(3)(C).

Section 2Q2(k)

      Section 202(k), in directing EPA to promulgate new regulations to control
evaporative  emissions  from all gasoline-fueled motor vehicles, provides that "the
regulations shall take effect as expeditiously as possible."  This applies  to  the
evaporative emission test requirements for gasoline-fueled light-duty vehicles, light-
duty trucks, and heavy-duty vehicles.

      EPA now believes that maintaining  a phased approach  as proposed,  but
reducing manufacturers' compliance requirements by ten percent each year, is  the
most expeditious implementation schedule that is realistically feasible. This schedule
effectively delays the start of the phase-in until model year 1996, according to  the
following schedule:

           Model Year                 Percent of Production

              1996                             20

              1997                             40

              1998                             90

              1999                             100
         , for example, Letter from Kelly M. Brown, Ford Motor Company, to EPA AJ r
Docket, November 5, 1992 (Docket A-89-18, item IV-D-90).

                                     59

-------
      This schedule, which decreases manufacturers' compliance requirements by ten
percent each  year,  is  designed to respond  to  manufacturers' concerns while
maintaining  consistency  with  the proposed  implementation  schedule.   EPA
acknowledges that manufacturers can no longer be expected to make changes and
complete certification for the 1995 model year on a national scale due to the date of
promulgation of this rule.

      The revised implementation schedule, while responding to the manufacturers
concerns, reduces the phase-in requirement by only ten percent in each year of the
phase-in, and so minimizes the effect on the air quality benefit provided by improved
evaporative emission controls. Furthermore, in California (and in any states adopting
the California evaporative standards), a larger percentage of vehicles will be  sold in
these model years, and in the  1995  model year, with greatly improved evaporative
emission controls compared to current  controls. Some of these additional vehicles
may be sold in  states  that have  not adopted California  standards  as well, if
manufacturers choose not to market  separate California and non-California versions
of certain models. Thus, the effect of the revised schedule on the air quality benefit
will be minimized.

      The phased in approach to implementation, starting as it does with a modest
20 percent requirement, recognizes that current vehicle designs vary with respect to
degree of modification needed to comply with the new requirements. The phase-in
allows earlier implementation  of the easier-to-fix designs in a large manufacturer's
model line, while allowing additional lead time for more challenging redesigns.

      For heavy-duty vehicles and engines, manufacturers presented no evidence that
an additional delay in the implementation schedule was necessary, and EPA knows
of none; therefore, gasoline-fueled heavy-duty vehicles and heavy-duty engines are
subject to the same implementation  schedule as for light-duty vehicles.

      It is reasonable to expect manufacturers designated as small entities to comply
only with the 1999 model year requirement, primarily because of the large impact the
revised test  will have on  capital requirements  for  facility modification  and
construction. Also, the advantage of a phased in implementation schedule is  clearly
reduced where a manufacturer's domestic offering is limited.

Section 202(aX2)

      Requirements finalized  pursuant to  section 202(aXD of the Act are to  be
implemented according to the provisions of section 202(aX2).  This includes the new
evaporative  test procedure for methanol-fueled light-duty vehicles and light-duty
trucks, the fuel spitback test, and the dispensing limit for fuel pumps.  The 202( k >
provision for implementation "as expeditiously as possible" therefore does not apply
Section 202(a)(2)  states:
                                     60

-------
      Any regulation prescribed under paragraph (1) of this subsection (and
      any revision  thereof) shall  take effect  after  such period  as  the
      Administrator  finds  necessary  to  permit  the  development  and
      application of the requisite technology, giving appropriate consideration
      to the cost of compliance within such period.
      Although the requirement for implementation "as expeditiously as possible"
does not apply to methanol-fueled light-duty vehicles and light-duty trucks,  EPA
believes that the same phase-in schedule for gasoline-fueled vehicles should apply to
these vehicles. EPA does not expect methanol-fueled vehicles, including flexible-
fueled vehicles, to represent a large percentage of any manufacturer's sales volume
in the early years of the phase-in.  Therefore, manufacturers will be able to delay
certifying these vehicles to the new requirements until later in the phase-in schedule,
if desired, without incurring undue costs of compliance.

      EPA is phasing in the fuel spitback test with the  new  evaporative test
requirements. As described in Chapter 4, the requisite technology already exists for
preventing fuel spitback during refueling. Since designs for spitback control may be
affected  by  a manufacturer's  approach  to meeting  the  new evaporative test
requirements, implementing the fuel spitback test on a different schedule may cause
an unnecessary increase in the cost of compliance.

      The limitation on fuel dispensing rates will be delayed until January 1, 1996
for large fuel handlers. The technology for compliance is currently being used by a
portion of the industry. Implementing the dispensing limit in January 1996 provides
ample time for the replacement of the remaining dispensing nozzles and will ensure
that the first vehicles subject to fuel spitback  control will generally be fueled at
pumps subject to the dispensing limit-  Moreover, EPA intends to minimize the cost
of compliance by allowing industry to meet  the requirements through the natural
turnover of equipment, rather than requiring retrofit or replacement of equipment to
meet test requirements.  Therefore, entities that handle less than 10,000 gallons
(38,000 liters) per month are allowed an additional two years to meet the limitation
on dispensing rates.

Section 202(aX3XC)

      Clean  Air Act section 202(a)(3XC) applies to heavy-duty methanol-fueled
vehicles. That provision states:

      Any standard promulgated  or revised  under  this  paragraph  and
      applicable to classes or categories of heavy-duty vehicles or engines shall
      apply for a period of no less  than 3 model years beginning no earlier
      than the model year commencing 4 years after such revised standard is
      promulgated.
                                      61

-------
      To comply with the requirements of section 202(a)(3)(C), implementation of the
new requirements for methanol-fueled heavy-duty vehicles and engines should begin
in model year 1998. Implementation for these vehicles is then phased in on the same
schedule as for other vehicles, namely 90 percent in model year 1998 and 100 percent
in following years.
                                     62

-------
                         Chapter 4  Economic Impact
      The  revised evaporative  test procedure will result  in  costs for various
components on the vehicle, as well as costs for new facilities and the effort spent for
research and testing.

      The January 1990 NPRM depended on a detailed assessment of the technology
and costs  that  would be necessary to meet test requirements.44  The expected
changes included  larger evaporative canisters and vapor lines, more sophisticated
purge valves, and new rollover valves. EPA estimated at that time a retail price
equivalent (RPE) total cost of $9.65,  $13.40, and $11.25 for light-duty vehicles, light-
duty trucks, and heavy-duty vehicles, respectively.

      In response to the January 1990 NPRM, several commenters claimed that EPA
underestimated vehicles costs.  EPA received cost estimates ranging from $30 to
$110.   However, no commenter responded to any of the individual  assumptions
comprising EPA's cost analysis, and no commenter attempted to justify  any of the
new estimates.

      At the  January  1992 public  workshop, GM offered new  cost figures.  GM
estimated a total cost of $100 to meet the  requirements of the GARB procedure, and
a total of $200 to meet the requirements of the procedure discussed at EPA's January
1992 workshop.  For the GARB  procedure, GM identified the need for "a larger
canister, running loss control by thermal management techniques, and a 'smart
purge' system."    For EPA's proposed procedure, GM said it  would need,  in
addition, a hydrocarbon sensor to maintain control of exhaust emissions, and a heated
canister system to more rapidly remove vapors from  the canister.

      Industry's cost estimates have been insufficiently supported for EPA to change
its analysis. Commenters have identified no hardware that EPA's analysis had not
accounted for, and have provided no direct challenge to the assumptions that went
into EPA's cost estimate.  Moreover, GARB estimated the cost of compliance for its
evaporative test procedure to be $18 per vehicle, very close to EPA's estimate.46
   ^"Onboard and Evaporative Control System Cost Estimates for the Supplemental
Notice of Proposed Rulemakmg," EPA memo from Jean Schwendeman to the Record.
December 22, 1988 (Docket A-89-18, item II-B-6).

   ^Letter from Samuel A. Leonard, General Motors, to Richard D. Wilson, EPA.
March 23, 1992 (Docket A-89-18, item IV-D-78).

      Technical Support Document for a Proposal to Amend Regulations Regarding
Evaporative Emissions Standards, Test Procedures, and Durability Requirements.  .
California Air Resources Board, August 9, 1990 (Docket A-89-18, item IV-D-87).

                                    63

-------
CARB's estimate assumed the evaporative control system would be composed of a 5-
liter canister, some additional vapor lines, and a fuel cooling system.

      The highest figure, GM's $200 estimate, was  associated with EPA's older
proposed test requirements, which are not being finalized. The test requirements
that EPA is finalizing are based on the GARB test procedure, to which GM applied
its $100 cost estimate. Nevertheless, GM's $200 estimate, involving the hydrocarbon
sensor and heated canister, was not substantiated with cost information. EPA is
unconvinced that manufacturers could not meet the EPA's proposed test requirements
by increasing the size of the evaporative canister and improving the aggressiveness
and control of the purge system. Also, the technology for hydrocarbon sensors seems
to be unavailable in the near term. Moreover, GM relied on a complex and expensive
method of heating canisters externally, ignoring evidence in the technical literature
that canisters could be most effectively, and very cheaply, heated by purging with
heated air.

      EPA has revised its cost estimates, depending largely on the assumptions used
in the NPRM.   Updated cost estimates  are described below.  Unless otherwise
indicated, all prices in the following discussion represent the original equipment
manufacturer (OEM) cost that auto manufacturers  pay,  in 1992 dollars.   Cost
estimates are rounded to the nearest $0.05 in intermediate calculations. Previously
published costs expressed in 1988 dollars, where used as a  basis for the present
calculations, are increased by 15 percent to account for inflation.

A.    Vehicle Hardware Costs

      The following paragraphs detail EPA's revised cost estimates associated with
vehicle changes expected to result from the new test procedures. The estimates are
summarized in Table 4-1.
   47"Vapor Canister Heater for Evaporative Emissions Systems," Robert P. Bishop
and Peter G. Berg, February 1987, SAE 870123.
                                     64

-------
                                  Table 4-1
                           Hardware Cost Estimates
Cost Item
Canister
Purge valve
Fuel and vapor line
Fuel temperature
management
Spitback valve
Total Cost
Light-Duty
Vehicles
3.50
1.10
0.80
0.30
0.30
6.00
Light-Duty
Trucks
6.05
1.10
1.10
0.00
0.35
8.60
Heavy-Duty
Vehicles
2.95
1.10
1.45
0.00
0.35
5.85
      1.  Evaporative Canisters

      Evaporative canisters are expected to increase in size to accommodate the need
to store larger quantities of fuel vapor. The NPRM cost analysis, using an estimated
volume increase from 1.3 to 2.7 liters for light-duty vehicles, relied on an established,
detailed  methodology to estimate the increased canister cost.  EPA received  no
comments on the methodology of estimating canister cost.

      Current modeling efforts indicate the need for canisters of approximately the
same size as used in the NPRM analysis.  GM has indicated that it expects to meet
CARB's requirements with a 2-liter canister, indicating that the NPRM estimate may
be conservative.  As a  result, EPA has  revised its cost estimate for evaporative
canisters  only  to  account  for inflation,  resulting in an increased cost  of  $3.50.
Similarly, light-duty trucks and heavy-duty vehicles are expected to incur increased
costs of $6.05 and  $2.95, respectively.

      2.  Purge Valves

      Purge valves are among the most important components to be improved as a
result of the revised evaporative test requirements. EPA expects purge valves to be
upgraded to manage higher vapor flow rates and maintain improved control  of low
flow rates.  This  would involve  changes to the valve assembly, as well as the
electronic control of the valve.

      The NPRM cost analysis characterized various in-use designs for purge valves
and calculated a  cost increase for a frequency-modulated solenoid valve.   EPA
received  no comments on the methodology of estimating the  cost of purge valves
EPA therefore  maintains the earlier estimate for a fleet-average increased cost.
                                      65

-------
adjusted for inflation, of $1.10 per vehicle.  The same cost was assumed for light-duty
trucks and heavy-duty vehicles.

      Additional costs will be incurred in the programming of the electronics that
control the valve. The programming is considered to be part of the effort for research
and development, described below.

      3.  Rollover Valves

      In the NPRM cost analysis, EPA recognized the fact that many vehicles already
had some form of vapor vent or rollover valve, but conservatively estimated that all
vehicles would need a new rollover valve to allow full venting of fuel vapors from the
tank to the canister, without allowing liquid fuel to enter the  vapor line.  Rollover
valves are now installed on all vehicles to meet safety requirements preventing liquid
fuel leaks in an accident.

      EPA expects the new test requirements to require no change in size or function
for existing rollover  valves.   The  diurnal  emission  test involves more  vapor
generation, but with the very gradual heating in 24-hour cycles, vapor flow rates
through the rollover valve during the test will decrease.  The test procedure also does
not increase vapor flow rates during driving.  For example,  manufacturers are not
required to avoid a reliance on pressurized systems to reduce vapor generation. Also,
EPA expects that fuel temperatures during driving,  and therefore vapor generation,
will decrease for many vehicles, as described below.

      EPA therefore expects no cost for new or improved rollover valves.

      4.  Vapor Lines

      The  NPRM  cost analysis allocated a cost  for vapor lines, assuming that
increased vapor flow rates would require  larger tubes to route vapor from the fuel
tank to the canister.  This assumption is no longer valid, because the test procedure
no longer involves an increased vapor flow rate, as described above.

      Manufacturers  commented that both fuel and vapor lines were suspected
sources of emissions from permeation of fuel. Comments fell short of stating a need
to change materials, and included no estimate of any cost.  For this analysis, EPA
conservatively assumes that fuel and vapor  lines, currently composed of steel and
nylon tubes, will have to be modified to reduce fuel permeation.

      EPA assumes that manufacturers will maximize the use of impermeable steel
in fuel and vapor lines, leaving several  short  sections that  require a somewhat
flexible material. A  new technology, in which nylon tubes  are lined with a teflon
material,  seems to provide adequate resistance to  permeation.  The teflon-coated
nylon tubes cost roughly $0.30 per foot ($0.98 per  meter), for either fuel or vapor
lines, compared to a cost of about $0.14 per foot ($0.46 per meter) for either steel of

                                     66

-------
plain nylon tubes.48  EPA estimates that a total of 5, 7, and 9 feet (1.5, 2.1, 2.7
meters) of the new nylon tubes will be required for the whole set of fuel and vapor
lines on light-duty vehicles, light-duty trucks, and heavy-duty vehicles, respectively,
resulting in per-vehicle cost increases of $0.80, $1.10, and $1.45.

      EPA  understands that concerns  for  chemical  resistance and electrostatic
dissipation  are prompting some changes in materials selection.   EPA has  not
considered the possible positive or negative effects of such changes in estimating the
cost of meeting evaporative requirements.

      5.  Fuel Temperature Management

      Auto  manufacturers are expected to  design  vehicles with lower  fuel tank
temperatures during driving to avoid fuel boiling during the running loss test.  As
described in Chapter 5, EPA estimates that a total of 58 percent of light-duty vehicles
will need  moderate design changes to avoid fuel overheating.  However,  EPA  has
confirmed that the two percent of vehicles with the highest fuel temperatures are
scheduled to go out of production before implementation of the new evaporative test
requirements. EPA estimates that the remaining 56 percent of light-duty vehicles
would need to reduce fuel temperature profiles by as much as 15° F.

      Manufacturers have several possible options for reducing fuel temperatures
during driving. For example, fuel lines are often exposed to very hot engine surfaces,
allowing the recirculated fuel to be heated before it returns to the fuel tank.  Isolating
the fuel lines, either with a simple heat shield, or  through more careful routing.
would greatly decrease  the amount of heat absorbed by the fuel.   Similarly  hot
exhaust pipes are sometimes located very close to fuel tank surfaces, resulting in
localized heating  of the fuel in the tank; a heat shield or other insulation, or more
careful routing, would again  greatly reduce fuel heating.  Adding  heat shields or
rerouting lines should involve a very small amount of additional parts and  material
EPA estimates an average cost of $0.50 for the vehicles that need any of these
changes. The cost of $0.50 per vehicle, spread over 56 percent of the fleet, results in
an average cost of $0.30 for each light-duty vehicle.

      More expensive means  are available to deal with high fuel temperatures, but
manufacturers have not indicated that they intend to change to the more expensive
methods.  For example, installation of a variable-flow fuel pump would eliminate the
recirculation of fuel, eliminating the heating involved  as the unused fuel  goes past
the engine and back to the fuel tank. The variable-flow fuel pump, installed inside
the fuel tank, would also deliver much less fuel and give off much less heat. Also
some vehicles  are  currently  designed  to  route the recirculating fuel  past an
conditioning components to prevent the fuel from heating.  The current use of such
   ^Letter from Edward K. Krause, Pilot Industries,  Inc., to Alan  Stout. KPX
August 25, 1992 (Docket A-89-18, item IV-D-92).

                                     67

-------
a system supports EPA's expectation that limiting heat transfer to the fuel line is
effective in limiting fuel temperatures in the fuel tank.

      In the future manufacturers can probably design new models to sufficiently
limit fuel temperatures with no direct costs.  Factoring fuel temperatures into early
engineering efforts could prevent the need for additional hardware or other features
to ensure adequately low fuel temperatures.  For example, the fuel lines  could be
routed to minimize exposure to hot surfaces. The fuel tank could be isolated from the
exhaust system and from hot underbody air. The fuel pump could be located outside
the fuel tank.  Total airflow underneath the vehicle could be increased as necessary.
These design changes would involve significant costs for existing models, but should
be fully available at very little or no cost for future models.

      Light-duty trucks and heavy-duty vehicles are designed differently than light-
duty vehicles in ways that prevent overly high fuel temperatures. For example, light-
duty trucks and heavy-duty  vehicles typically have relatively high underbody
clearance, resulting in a greater cooling effect from underbody airflow. Because there
is more space underneath the vehicle, fuel tanks can also be isolated from exhaust
systems. In addition, light-duty trucks and especially heavy-duty vehicles are more
likely to have carbureted fuel systems, which do not circulate hot fuel back to the fuel
tank.  EPA therefore expects that light-duty trucks and heavy-duty vehicles do not
need to  be modified to reduce  fuel  temperatures  in order to  meet new  test
requirements.

      6.  A"ti-spitback yalva

      Manufacturers may need a valve installed in the vehicle's filler neck to prevent
spitback during refueling. The NPRM cost analysis quoted a vendor's estimated cost
of $0.25 for  the valve.  Manufacturers provided no comment on the need for or the
cost of an anti-spitback valve.  Some current vehicles are already designed  with the
valves, but EPA conservatively assumes that all vehicles need the valve, adjusted for
inflation to a current price of $0.30. Light-duty trucks and heavy-duty vehicles have
a slightly higher estimated cost to account for vehicles with multiple tanks.

      7.  Seals

      Kautez of Canada was  the only participant to comment on the need for a
change in seals, such as O-rings and gaskets.  Kautex claimed that current seals
would be inadequate to pass EPA's proposed 4-hour resting loss test, but gave no
indication that  the proposed 24-hour diurnal testing would require new  seal
materials.  Because EPA is adopting only the 24-hour diurnal emission test, no cost
was estimated for more expensive seals.
                                      68

-------
B.    Development and Production Costs
      The following paragraphs describe EPA's revised estimates of the overall costs
associated with the new test procedures. The estimates are summarized in Table 4-2.

                                  Table 4-2
                         Total Vehicle Cost Estimates
Cost Item
Vehicle Components
Packaging
RD&T
Certification
Facilities
Total OEM Cost
Markup (26 %)
Total RPE Cost
Light-Duty
Vehicles
6.00
0.75
0.20
0.15
0.60
7.70
2.00
9.70
Light-Duty
Trucks
8.60
0.90
0.35
0.15
0.60
10.60
2.75
13.35
Heavy- Duty
Vehicles
5.85
1.00
0.90
0.15
0.60
8.50
2.20
10.70
      1.  Packaging

      The NPRM cost analysis included estimated costs of $0.75, $0.90, and $1.00
associated with the need to accommodate the larger evaporative canister for light -
duty  vehicles,  light-duty trucks,  and  heavy-duty vehicles, respectively.   EPA
demonstrated in its 1987 Draft Regulatory Impact Analysis ("1987-DRIA") that such
modifications can almost always be  avoided.    EPA found that most vehicles at
that time could easily accommodate the expected increase in canister, volume   If a
vehicle cannot accommodate the anticipated larger canister, other design options
would be available. A higher grade of activated carbon would reduce canister volume
with  little or no  increase in  total cost.   Also,  converting from cylindrical t«>
rectangular or other shapes could allow more efficient packaging of the canister  EPA
expects a negligible cost for  vehicle body  modifications to accommodate larger
evaporative canisters. However, EPA will maintain its conservative estimate from
the NPRM cost analysis.
   49"Draft  Regulatory Impact  Analysis,  Control of  Gasoline Volatility  m
Evaporative Hydrocarbon Emissions from New Motor Vehicles," EPA, pp. 4-34  M
July 1987 (Docket A-85-21, item II-A-45).
                                     69

-------
      2.  Research. Development, and Testing

      EPA's  estimate of  the  cost  for  research,  development, and testing is
documented in the 1987 DRIA.  EPA's cost estimate, which received no comment,
factored  in engineering time, technician time and the cost  of testing  for each
evaporative family. The new cost estimate, therefore, is based on the 1987 estimate
adjusted only for inflation, and is $0.20, $0.35, and $0.90 for light-duty vehicles, light-
duty trucks, and heavy-duty vehicles, respectively.

      3.  Certification

      The NPRM cost analysis estimated a cost of $0.05 per vehicle for the effort
involved  in certifying  vehicles to  the  new test procedures.   While receiving no
comments on this estimate, EPA believes that the previous  estimate should be
adjusted to reflect changes to the test procedure that  increase the burden of the
certification process. The method of cycling temperatures in 24-hour periods is the
primary test change that increases the certification burden. Estimating a per-vehicle
cost for certification testing is difficult; EPA therefore makes a simple estimate of a
tripled overall cost for certification testing, to $0.15 per vehicle.

      4.  Facilities

      GM has estimated that the automotive industry will  need 72  variable-
temperature  enclosures to comply with CARB's test  procedure requirements.50
EPA has not conducted an independent assessment, but this estimate appears to be
reasonable based on recent information obtained from GM, and based on the number
of manufacturers that currently perform their own certification testing.51  EPA
assumes that manufacturers performing their own  certification tests will need to
build one running loss test site as well. Thus, EPA estimates that the industry will
require 72 variable-temperature enclosures and 30  running loss sites for the new
evaporative test procedure.

      Based on GM's comments and EPA's investigation of facility costs, the cost of
each variable-temperature enclosure and running loss site is estimated to be $225,000
and $700,000, respectively. The estimated cost to the industry  for these enclosures
is then 37.2 million dollars (Table 4-3).  Amortizing this cost over ten years at ten
percent per year, and conservatively assuming an annual production of approximately
ten million vehicles per year, results in a per-vehicle cost of $0.60.
   50Letter from Samuel A. Leonard, GM, to Richard D. Wilson, June 5, 1990, page
39 (Docket A-89-18, item IV-D-30).

   5^'Telephone Conversation with General Motors Staff," EPA memo from Dan
Barba to Docket A-89-18, October 19, 1992 (item FV-E-24).

                                     70

-------
                                  Table 4-3
                     Facility Costs for Automotive Industry
Equipment Requirements
Variable-Temperature SHED
Running Loss SHED
Number
72
30
Unit Cost
$225,000
$700,000
Total
Cost
$16,200,000 !
$21,000,000
$37,200,000
C.    Overall Vehicle Lifetime Costs

      1.     Fuel Savings

      The vehicle costs discussed above are offset by a savings in fuel consumption
over the life of the vehicle. This is due to the capture of fuel vapors that would have
otherwise been lost to the atmosphere, and the subsequent burning of these vapors
in the engine. Two factors affect the determination of the fuel savings.

      First, a cost element is added to reflect the additional fuel consumption caused
by the extra weight of the  improved evaporative control system.   In 1988 EPA
prepared a detailed analysis to quantify the effect of the added weight, concluding
that light-duty vehicles  would add a discounted lifetime  cost of approximately
$0.50.    The analysis assumed, among other  things, a 1.8 pound (0.8 kg) increase
in weight for canister, vapor  line and valve changes, and a crude oil price of $20 per
barrel. Although adjustments to some of the assumptions could be made to  reflect
updated information, such modifications would be minor and would not significantly
affect the $0.50 estimate.   Similarly, weight penalties  of $1.15 and $0.60 were
calculated for light-duty trucks and heavy-duty vehicles, respectively.

      Second, a cost savings is credited to account for the retention and combustion
of fuel vapors that  would otherwise be lost to the  atmosphere.  To estimate the
benefit of burning the vapors, EPA assumes that all the vapor is butane, and that all
the vapor purged from the canister is burned in the engine to power the vehicle. The
heat of combustion (BTU/pound) of butane is six percent higher than that of 9  psi * H2
kPa) RVP gasoline, so a pound (kg) of butane is considered equivalent to 1.06 pounds
(kg) of gasoline. With a gasoline density of 6.18 pounds per gallon (0.74 kg/liter), and
a conservatively assumed cost of $1 per gallon ($0.26 per liter), the resulting cost
credit is $0.38 per kilogram of butane.
    52"Onboard and Evaporative Control System Cost Estimates for the Supplemental
Notice of Proposed Rulemaking," EPA memo from Jean Schwendeman to the Record
December 22,  1988 (Docket A-89-18, item II-B-6).
                                     71

-------
      The amount of butane  recovered  over the lifetime of the vehicle  can be
estimated by applying per-vehicle emission reductions (in g/mi) over the vehicle's life.
Vehicle life is assumed to be ten years and 100,000 miles for all vehicle classifications
in order to simplify the analysis. This results in an estimated fuel recovery of 40, 25,
and 105 kg for each light-duty vehicle, light-duty truck,  and  heavy-duty vehicle,
respectively.  The associated cost credits, discounted at a rate often percent per year
over ten years, are $9.35, $5.85, and $24.20 for each light-duty vehicle, light-duty
truck,  and heavy-duty vehicle, respectively.  Combining these credits with the
offsetting weight penalties and rounding to the nearest whole dollar yields net fuel
savings of approximately $9, $5, and $24 for each light-duty vehicle, light-duty truck,
and heavy-duty vehicle, respectively.

      2.    Overall Vehicle Costs

      Overall, OEM costs for light-duty vehicles are estimated to increase by $7.70.
A 26 percent markup results in a RPE cost of $9.70.  However, this initial cost for
purchasing a vehicle is offset over the life of the vehicle by the net fuel savings.  A
summary of costs for each type of vehicle is provided in Table 4-4.

      Assuming that  10 to 15 million vehicles  requiring  improved evaporative
controls are sold per year, and conservatively using the light-duty truck costs of $13
per vehicle without considering the net fuel savings, EPA estimates an annual cost
of $130 to 200 million.   This cost would be largely or completely  offset by the
associated fuel savings.

                                   Table 4-4
                                Cost Summary

Cost to manufacturer
Cost to consumer
Net Fuel savings
Net cost to consumer
LDV
$8
$10
$9
$1
LDT
$11
$13
$5
$8
HDV
$9
$11
$24
-$13 i
D.    Fuel Dispensing Nozzles

      Fuel dispensing nozzles must be designed to limit fuel flow to a maximum rate
of  10  gallons  (37.9 liters)  per  minute.    Husky  Corporation, representing
approximately 30 percent of the market, already installs a simple flow-limiting device
                                      72

-------
                                        CO
in all its nozzles at no cost to the purchaser.'   EPA therefore assumes that the cost
of such a device is negligible if a nozzle needs to be replaced before the effective date
of the dispensing rate limitation. Nozzle turnover rates vary widely, but on average
nozzles  are estimated to last from one to three years.54 That replacement rate is
expected to  increase because  of new safety  requirements for  fuel  dispensing
equipment.    Moreover, stations that must install stage II vapor recovery systems
are already required to limit fuel flow to 10 gallons (37.9 liters) per minute (57 FR
13498, April  16, 1992).  With a lead time of three years for large stations and five
years for small stations, EPA  expects that the cost of implementation  of this
requirement will be negligible.
   53"Phone Contact with Husky Corporation," EPA memo from Alan Stout to Docket
A-89-18, November 2, 1992 (item IV-E-27).

   ^"Investigation of the Need for In-Use Dispensing Rate Limits and Fuel Nozzle
Geometry  Standardization," EPA  technical report prepared  by the  Standards
Development and Support Branch, May 1987 (Docket A-89-18, item IV-A-2).

     For example, the National Fire Protection Association set a standard for fuel
dispensing nozzles to prevent the possibility of an operator returning a nozzle to it*
stored position without first deactivating the fuel flow (ANSI NFPA 30-A, effecti.-
August 17, 1990).

                                      73

-------
                       Chapter 5  Environmental Impact
A.    Methodology
      Emission reductions resulting from the improved evaporative test procedure
are estimated using EPA's MOBILE emission factor model, version 5.0 (MOBILES).
MOBILES is EPA's model for calculating fleet average motor vehicle emission factors.
Each evaporative emission factor (diurnal, hot soak, running loss, and resting loss)
calculated by MOBILES is a composite of emission factors from vehicles with properly
functioning emission controls and vehicles with failed controls.56

      The projections used here are made for the year 2020 in order to provide
benefit predictions for a fully turned-over fleet and to factor in other known trends,
such as the effects of other  new  Clean Air Act programs.  Emission inventory
reductions are estimated  by determining  the difference between  post-control
emissions (2020 projection with the new evaporative test procedure in place) and
baseline  emissions (2020 projection without the  new  procedure).   Post-control
emission factors are calculated in MOBILES using the percent  reduction factors
discussed  in Section C below.  To simplify the discussion, this chapter focuses the
methodology description on light-duty vehicles. The calculations for light-duty trucks
and heavy-duty vehicles use essentially the same approach.

      EPA has developed a supplemental evaporative emissions model, separate from
MOBILES, to evaluate the expected reductions in VOC emissions associated with the
improved evaporative emission test procedure.  The supplemental model, described
in Appendix A, calculates canister  emissions during diurnal, hot soak, and driving
episodes and makes use  of actual in-use trip patterns to track canister condition.
Results of this modeling effort are presented in Appendix A and summarized in this
chapter.

B.    Baseline Emissions

      A summary of the MOBILES input parameters used to determine the baseline
emissions is provided in  Appendix B, along with the MOBILES output files.  The
primary inputs include the use  of 9 psi (62 kPa) RVP gasoline, a daily diurnal
temperature swing of 72° to 96° F (22.2° to 35.6° C), full implementation  of EPA's
high-technology inspection and maintenance (I/M) program, and the implementation
of a reformulated gasoline program as required in the Clean Air Act.  The RVP was
chosen to represent designated Class C areas in Phase U of EPA's volatility controls
   56These  vehicle categories are described in  "Draft MOBILES Hot Soak and
Diurnal Emissions," handout from EPA MOBILES Workshop, July 8, 1992 (Docket
A-89-18, item IV-B-8).

                                     75

-------
(55 FR 23658, June 11, 1990).  The temperature swing is based on meteorological
data from these Class C areas, as discussed in Section C of Chapter 2.

      The  inspection  and maintenance program reflects EPA's  proposed rule
requiring   high-technology  I/M  programs in   serious,  severe  and   extreme
nonattainment areas (57 FR 52950, November 5, 1992).  To simplify the estimate of
emissions, EPA conservatively assumed that vehicles in areas not covered under the
high-technology I/M program will experience the same emission reductions due to the
improved evaporative test procedure as vehicles in high-technology I/M areas.  In
fact, emission reductions resulting from the improved evaporative test procedure will
be somewhat greater in areas not covered under the I/M program.  However, the
effect of I/M on emission reduction estimates for the new test  procedure is small,
because the I/M  program  reduces  evaporative emissions primarily from failing
vehicles,  whereas the improved evaporative  test procedure  reduces emissions
primarily from properly functioning vehicles.

      The use of reformulated gasoline in certain areas, as required in the Clean Air
Act, will also have some effect on the MOBILES predictions. Although the details of
the reformulated gasoline program have not yet been finalized, MOBILES is capable
of modeling the effects of the program, assuming a 25 percent overall VOC emission
reduction standard.  Unlike the I/M program,  the use of reformulated gasoline is
expected to significantly reduce evaporative emissions from properly functioning
vehicles.  Thus, in the areas  of the country using reformulated gasoline,  baseline
evaporative emissions would be lower, resulting in correspondingly smaller emission
reductions from the improved evaporative test procedure.

      To account for the use of reformulated gasoline in affected areas, emission
factors are calculated  using  MOBILES  for vehicles  operating with  and without
reformulated  gasoline.  The two emission factors are then weighted, based on the
expected use of reformulated gasoline across  the nation, to  produce an overall
nationwide emission factor. Currently, it is anticipated that reformulated gasoline
will be used in the nine cities specified in the Clean Air Act, all of California, several
areas that  are likely to opt in to the Clean Air Act program, and some additional
areas that  will be included due to the effects of fuel distribution system  spillover.
Based on this, it is estimated  that about 40 percent of the nation would use
reformulated  gasoline.57

      Table 5-1 presents the projected baseline  evaporative emissions calculated
using MOBILES  and weighted  to account for the expected use of reformulated
gasoline in 40 percent of the nation. Appendix  B contains the MOBILES input and
output files used in this analysis.
   57"Evaluation of the  Use of Ethanol and  MTBE  in Reformulated Gasoline,"
prepared by Sobotka & Co, Inc.  for the U.S.  EPA, September 30, 1992 (Docket
A-89-18, item IV-A-5).

                                     76

-------
                                  Table 5-1
                        Baseline LDV Emission Levels
                     for Calendar Year 2020 in g/mi (g/km)
Category
Running loss
Hot soak
Diurnal
Resting loss
TOTAL
Problem-free
0.32 (0.20)
0.05 (0.03)
0.10 (0.06)
0.07 (0.04)
0.54 (0.34)
Purge-fail
2.76(1.72)
0.68 (0.42)
0.18(0.11)
0.07 (0.04)
3.69 (2.35)
Pressure-fail
2.76(1.72)
0.69 (0.43)
0.42 (0.26)
0.07 (0.04)
3.94 (2.45)
Composite
0.39(0.24)
0.06 (0.04)
0.11(0.07)
0.07 (0.04)
0.63 (0.39)
C.    Emission Reductions

      Percent reductions resulting from improved evaporative emission controls are
estimated for each of the four categories of evaporative emissions: diurnal, hot soak,
running losses, and resting losses.  Percent reduction factors are then input into
MOBILES to  estimate per-vehicle reductions in g/mi and total VOC  inventory
reductions in metric tons. The new evaporative test procedure's effect on emissions
from vehicles  with inoperative evaporative control systems (referred to as failed
vehicles) will be much different than its effect on emissions from properly functioning
vehicles (pass vehicles).  Therefore, failed vehicle percent reductions are discussed
separately near the end of this section.

Diurnal Emissions

      Diurnal emissions are classified in MOBILES as partial, full, or multiple-day
events.  A partial diurnal occurs when a diurnal is interrupted by a trip during the
period in the day when ambient temperatures are rising. Multiple-day diurnals occu r
when a vehicle has not been driven for two or  more consecutive days.

      Vehicles designed to the existing evaporative test procedure, which simulate*
a 1-day diurnal event, control a large percentage of partial and full-day diurnal
emissions. It is expected that vehicles designed to the new test procedure will further
reduce emissions from partial and  full-day diurnals, for several reasons.  FITM
current vehicles are designed to pass the test  based on a 60° to 84° F (15.6  t<
28.9° C) diurnal heat build. The improved evaporative test procedure will requn«
control during a higher temperature heat build of 72° to 96° F (22.2° to 35 6  <
Based on vapor generation estimates using the Wade Model, modified based  on w< r k
by  Reddy, the higher temperature diurnal  results  in  almost twice the  vap. •
                                      77

-------
generation.58'59  Thus, the higher temperature diurnal requirement will result in
larger canisters capable of controlling emissions on these hotter days.

      Second, the improved evaporative test procedure will require control of three
consecutive diurnals  rather  than only one  under the current evaporative test
procedure.  The increased canister capacity required for the extra diurnals in the test
procedure provides added assurance that a single full-day diurnal will be controlled.

      Furthermore, the improved evaporative test procedure is expected to result in
higher canister purge rates. The improved test procedure will require vehicles to be
capable of purging a loaded canister during the exhaust emission test in preparation
for a multiple-day diurnal heat build.  This provides an advantage for partial and
full-day diurnal control because the rapid purge helps to ensure that storage capacity
is restored, even after the short drives typical of many in-use driving patterns.

      However, available data show that in-use vehicles with properly functioning
evaporative systems consistently emit  a small quantity of hydrocarbons during a
diurnal heat build, even with a well purged canister. Figure 5-1 shows a distribution
of diurnal emission data collected from EPA's emission factor program.60 The data
exclude failed and tampered vehicles and is limited to fuel-injected vehicles tested
according to the existing test procedure. The data show that vehicles most frequently
emit around 0.5 to 0.6 grams during a 60° to 84° F (15.6° to 28.9° C) temperature
diurnal and that a  majority of vehicles emit less than 1.8  grams per test.  EPA
conservatively assumes that these emissions will continue to exist even with the new
test procedure in place and that as a result, the improved evaporative test procedure
will control only some fraction of current full-day diurnal emissions.

      A conservative method of estimating the emission reduction is to assume that
all vehicles in the test data base that emit more than 1.8 grams would be brought
below this level via increased canister capacity and better purge. Vehicles emitting
more  than  1.8 grams are assumed to  be  experiencing the canister breakthrough
phenomenon associated with  insufficient canister capacity, which the new test
procedure is specifically intended to address. As mentioned above, vehicles emitting
at lower levels are conservatively assumed to be unaffected by the new procedure in
this analysis. The average diurnal emissions in this data set (Figure 5-1), weighted
   58"Factors Influencing Vehicle Evaporative Emissions," D.T. Wade, January 1967
SAE 670126.

   59EPA applied a correction factor of 0.78 to Wade Model predictions based on the
work of S.R. Reddy, "Prediction of Fuel Vapor Generation From a Vehicle Fuel Tank
as a Function of Fuel RVP and Temperature," September 1989, SAE 892089.

   ^"I/M Costs, Benefits, and Impacts," EPA, November 1992, Appendix A, pages TJ
& 33 (Docket A-91-75, item V-B).

                                     78

-------
CO

O
C
O
O
O
                                Figure 5-1.
             Distribution of  Diurnal  Emissions Data
    120
    100
     80
     60
     40
     20
     0
         0.2 0.4 0.6 0.8  I  L2 1.4 1.6 1.8  2  2.2 2.4 2.6 2.8  3  4  5  6 7  8  9  10  11 11 +

                              Diurnal Emissions (grams)

 <• I.' pii 'I'K-iii tii-i- vehicles
  • i > !•'\ I' I'll .si I iriii|>ri.iiiiir use lucl inji'i Icil vehicles only

-------
according to fuel system type using MOBILES parameters, is 1.31 grams per test.
Eliminating the emissions greater than 1.8 grams per test from the data set (thus
effectively assuming that the vehicles emitting at these higher levels are redesigned
to emit at the weighted average emission level for vehicles in the data base that emit
at below 1.8 grams per test) results in an average emission level for the full data set
of 0.65 grams per test.  Thus, a 50 percent reduction in full-day diurnal emissions is
a reasonable, albeit  conservative, estimate of the effect of the improved test
procedure.

      This estimate should  also reasonably approximate the percent reduction in
partial diurnal emissions, due to the similarity between partial diurnals and full
diurnals in terms  of  the degree of control provided  by the current and new
procedures.  Thus, a reduction of 50 percent is expected to occur in both partial and
full-day diurnal situations as a result of the improved evaporative test procedure.
Reductions in multiple-day diurnal emissions are expected to be somewhat higher
than for partial and full-day  diurnals.  Because EPA's existing evaporative test
procedure does not address multiple-day diurnal events,  unlike the one-day diurnal
scenario examined above, larger reductions are possible in multiple-diurnal emissions
than in one-day diurnal emissions.

      EPA emission factor data also provides some basis for estimating the reduction
in two- and  three-day diurnal emissions.   The data  identifies  how well in-use
evaporative systems controlled diurnals as a result of the existing test procedure
requirements. It is expected that future reductions in second- and third-day diurnal
emissions will be similar to the reductions in full-day diurnal emissions that resulted
from adoption of the current evaporative test procedure.  Based on the EPA-modified
Wade Model, EPA estimates that, without evaporative controls, vehicles would emit
roughly 17 grams during a 60° to 84° F (15.6°  to 28.9° C)  diurnal, assuming  a
16-gallon (61-liter) tank, 9.0 psi (62 kPa) RVP fuel, 40 percent tank fill level, and a
permanent fuel tank vapor space of 2.4 gallons  (9 liters). Considering the average
1.31 grams per test for controlled diurnal emissions discussed above, a reduction of
over 90 percent results.

      However, due to the  uncertainty  associated with the  various temperature
conditions and driving patterns surrounding  multiple-day diurnal events,  EPA
believes that a more conservative estimate is appropriate for two- and three-day
diurnal emission reductions.  Therefore, EPA estimates that adoption of the new
evaporative test procedure will result in a reduction in second- and third-day diurnal
emissions of 75 percent.

      Though not specifically addressed by the  new test procedure, some control of
emissions on the fourth day of an extended multiple-day diurnal episode is expected
This control is due to the ability of the  canister  to collect some vapor even after the
canister has been loaded beyond its normal working capacity.  Consistent with past
EPA analysis of this phenomena, it is expected that roughly half of the emissions on
the fourth day of a multiple-diurnal event would be captured by a canister sized to

                                      80

-------
                                  f\ 1
control emissions from three diurnals.   The same canister would also control some
smaller fraction of diurnal emissions from a fifth day, probably around 25 percent.
Due to the occurrence of backpurge between successive diurnal events, EPA estimates
that the evaporative system will  continue to control roughly  25  percent of all
multiple-day diurnal emissions beyond five days.

      Due to the infrequency of diurnal events lasting four or more days, MOBILES
calculations are simplified by applying a single reduction factor for these events. In
this case, the 50 percent figure for fourth-day control was adjusted downward to
reflect the smaller benefits expected in multiple-diurnal episodes of five days or more.
Based on the fraction of the fleet experiencing multiple day diurnals of four days or
more, EPA estimates a  40 percent reduction in emissions during diurnal events
lasting four or more days.62

Hot Soak Emissions

      Faster purge and larger canisters will ensure that canisters also have capacity
at the end of a  drive to collect generated  hot soak vapors.  An estimate of the
reduction  in hot soak emissions can be derived from the data collected in EPA's
emission factor testing.63 The distribution of the hot soak data from EPA testing
is shown in Figure 5-2.  The average hot soak emissions in the data set, weighted
according to fuel system type using MOBILE4.1 parameters, is 0.48 g/test.

      Using an approach similar to the one taken in the above discussion on diurnal
emissions, it is assumed that the new test procedure will produce little incentive to
eliminate the small hot soak emissions evident in the test results for a large number
of vehicles. These small emissions  (most often in the 0.2 to 0.3 gram range) can be
expected to  continue to occur in future vehicles designed to meet  the improved
evaporative test procedure.

      In the hot soak data, most of the vehicles emit less than 0.7 grams per test.
Eliminating all hot soak emissions above 0.7 grams per  test from the distribution
(thus effectively assuming that vehicles emitting at these higher levels are redesigned
to emit at the weighted average emission level for vehicles in the data base that emit
at below 0.7 grains per test) results in an average hot soak emission value of 0 25
   61"Multiple Diurnal Emissions," EPA memo from David Bartus to Celia Shih.
December 19, 1989 (Docket A-89-18, item H-B-4).
   CO                          _
      Reductions in Evaporative Emissions and Running Losses from Enhanced
Vehicle-Based Control," EPA memo from Alan Stout to Charles Gray, December 19
1989 (Docket A-89-18, item II-B-5).

   63"I/M Costs, Benefits, and Impacts," EPA, November 1992, Appendix A, pages (j
& 33 (Docket A-91-75, item V-B).

                                     81

-------
    200
    150
    100
u
o
    50
     0
                              Figure 5-2.

           Distribution of Hot Soak  Emissions  Data
         0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9  1 1.1 1.2 1.3  1.4 1.5 1.6 1.7 1.8 1.9 2 4  6  8 10  12 12+


                            Hot Soak Emissions (grams)
 <>\? piultlfm-lrcc vehicles

 •' 1 1 K VI1 hUl- k in|x i jlua I in I ii i )«.•». li-il vcliK Ic.s only

-------
g/test.  Thus, it is estimated that hot soak emissions will be reduced by approximately
50 percent (0.48 to 0.25 g/test, on average) as a result of the improved evaporative
test procedure.

Running Losses

      Because  the scope of in-use running losses has been identified only  fairly
recently, it is difficult to estimate the reductions in running losses that are likely to
result from the new evaporative test procedure. However, a reasonable estimate can
be produced using running loss emission data contained in MOBILES.

      MOBILES estimates that the running losses from current problem-free vehicles
average 0.32 g/mi (0.20 g/km).  EPA sees no reason to expect that the new procedure
will not reduce these emissions to a level near the standard of 0.05 g/mi (0.03 g/km),
a reduction of 84 percent. However, considering the uncertainties in these estimates
and the fact that extreme in-use conditions may result in some loss of control, it is
reasonable to decrease the full assumed benefit somewhat to a percent reduction level
of 80 percent.

Resting Losses

      Reductions in resting losses can be estimated by comparing resting losses from
current vehicles to an  estimate of the control level required to pass the improved
evaporative  test procedure. Because resting losses are measured concurrently with
diurnal emissions, a separate standard has not been specified for resting losses in the
new test procedure.  However, an upper limit for resting losses expected, following
implementation of the improved test procedure, can be estimated by subtracting the
emissions expected from the hot soak and diurnal portions of the  test from the 2
gram per test overall evaporative standard.

      Manufacturers will be required to design a system that emits no more than 2
grams during the hot soak test and the highest-emitting 24-hour cycle of the diurnal
emission test (which includes resting losses).  Based on data from EPA's emission
factor testing as discussed above, it  is estimated that diurnal emissions normally
contribute approximately 0.5  to 0.6  grams  per test and  that a hot  soak can be
expected to add another 0.2 to 0.3 g/test. Providing a 0.4 g/test margin of safety to
pass the test, 0.7 to 0.9 grams per 24-hour period (or about 0.03 g/hr) appears to be
a reasonable estimate of the upper limit for future resting losses.  This estimate is
consistent with GM's testimony at the January 1992 workshop that its vehicles would
need to control resting losses to levels under 0.7 g per 24 hours to comply with test
requirements.

      Resting loss emission data contained in MOBILE4.1 indicate that current
vehicles emit an average of 0.11 grams of hydrocarbons per hour.  Comparing this
value to the  expected post-control resting loss rate of approximately 0.03 g/hr result-
                                      83

-------
in approximately a 75 percent reduction in resting losses due to the new evaporative
test procedure.

Failed Vehicles

      EPA expects that some vehicle designs will need to be revised to ensure lower
fuel temperatures during driving in order to prevent fuel boiling during the running
loss test.  Lower fuel tank temperatures during driving  would result in reduced
emissions, even for vehicles having inoperative evaporative controls, because of the
reductions in vapor generation rates at the lower fuel temperatures. EPA analyzed
ATL's running loss data on a representative sample of failed vehicles to estimate the
effect of the new test for running losses on in-use emission levels.   The 85 light-
duty vehicles tested all had  fuel systems that did not hold pressure.  Testing was
conducted at 5 different combinations of ambient temperature and fuel volatility.

      EPA expects manufacturers to design vehicles that limit fuel temperature rises
to 29° F (16.1° C) during the running loss test.  This is based on the expectation that
vehicles designed to pass the running loss test (which begins with fuel temperatures
at 95° F (35° C)) would limit maximum fuel temperatures  to no higher than 124° F
(51° C), in order to provide an 8° F (4.4° C) margin of safety below the boiling point
of certification fuel. Table 5-2 shows a breakdown of the fuel temperature profiles for
the test vehicles. The data indicate that 41 percent of the vehicles require no change
to prevent fuel boiling during the running loss test; 56 percent of the vehicles require
a reduction in fuel temperatures of up to 15° F (8° C); 2 percent of the vehicles need
more than a 15° F (8° C) reduction.
   ^"MOBILES Inputs for Evaporative Emission Reductions from Fail Vehicles.
EPA memo from Alan Stout to Joanne I. Goldhand, January 5,1993 (Docket A-89-18.
item IV-B-11).

                                     84

-------
                                  Table 5-2
                   Distribution of Fuel Temperature Profiles
Percent of
Vehicles
41
28
20
8
0
2
Range of Current Profiles
°F (°C)
0-29 (0-16.1)
30-34 (16.7-18.9)
35-39 (19.4-21.7)
40-44 (22.2-24.4)
45-49 (25.0-27.2)
50-54 (27.8-30.0)
Expected Reduction
°F (°C)
0(0)
1-5 (0.6-2.8)
6-10 (3.3-5.6)
11-15 (6.1-8.3)
16-20 (8.9-11.1)
21-24 (11.7-13.3)
      To  estimate emission  reductions,  EPA assumed  that vehicle families
represented by ATL test vehicles with fuel temperature profiles greater than 29° F
(16.1° C) would be redesigned such that their distribution of fuel temperatures match
that of the lower-temperature vehicles in the ATL data.  It is possible that some of
these vehicle  designs may be modified only to the point of achieving the  29° F
(16.1° C) profile; however, EPA expects that the incentive to reduce the magnitude
of vapor generation during driving  will prompt manufacturers  to  reduce fuel
temperatures  beyond the  minimum  required  to  avoid  fuel  boiling.   Thus the
distribution of ATL running loss test data for vehicles that maintained fuel at 29° F
(16.1° C) or lower is assumed to be the distribution for all vehicles meeting the new
evaporative emission test requirements. The estimated emission reductions are then
the calculated differences between the average running loss emissions (measured by
ATL) for the distribution of all tested vehicles and the modified distribution.

      The calculated emission reductions for each test condition are presented  in
Table 5-3.  The data indicate a reduction of approximately 30  percent for ambient
temperatures of at least 95° F (35° C). The testing at 80° F (27° C) involved only 19
vehicles, but clearly indicated  less of an effect, with a  17.8 percent reduction.  At
lower ambient temperatures, fuel boiling is less likely, even for vehicles with large
fuel temperature increases during driving. The data do not indicate any dependence
of emission reductions on fuel volatility.  EPA therefore estimates that failed vehicle
running loss  reductions resulting  from the  new evaporative emission control
requirements  will be  as indicated  in Table 5-4.  Reductions for intermediate
temperatures are calculated in MOBILES by linear interpolation.
                                     85

-------
                                  Table 5-3
           Emission Reduction Results from Testing of Failed Vehicles
Ambient
Temperature
80° F (27° C)
95° F (35° C)
105° F (41° C)
Fuel Volatility (RVP)
7.0
-
28.6%
27.0%
9.0
-
29.6%
28.7%
10.4
17.8%
-
-
      EPA believes that these reductions are appropriate for hot soak emissions as
well, due to the dependence of hot soak emissions on fuel heating  at the end of
vehicle operation. EPA expects no reduction in diurnal emissions for failed vehicles,
because diurnal vapor generation is a function of ambient temperatures rather than
heat produced by vehicle operation.

                                  Table 5-4
                    Running Loss and Hot Soak Emissions:
                     Percent Reductions for Failed Vehicles
Temperature
65°F(18°C)
80° F (27° C)
95° F (35° C)
105° F (41° C)
Reduction
0%
15%
30%
30%
      For resting losses, EPA expects that failed vehicles will reduce resting losses
as effectively as pass vehicles (75 percent reduction).  Because resting losses are
controlled primarily by material selection, resting losses should be unaffected by a
vehicle's ability to purge or hold pressure.

      EPA has  not  quantified  any expected  change in the frequency of failing
evaporative  systems  attributable  to  the  new  evaporative  test  procedure.
Manufacturers have  argued that the test procedure will cause a revamping of fuel
system designs, which will improve system durability.  It is possible that changes to
reduce permeation will result in more durable  components.  However, the new test
procedure does not require the use of more durable systems, which makes it difficult
to estimate the degree of improvement in durability.  Therefore, EPA is unable to
quantify this effect at this time. It should be noted  that the MOBILES analysis
                                     86

-------
accounts for the effects of the enhanced I/M and onboard diagnostics programs in
determining the number of failed vehicles in the fleet.

Other Reductions

      The  calculated  reduction in VOC  emissions also includes a  substantial
reduction in benzene,  a known carcinogen.   Benzene  reductions will have  an
important societal benefit.  EPA has, however, not factored benzene reductions into
the calculation of benefits, because they are difficult to quantify in cost terms, and
because the new test procedure is clearly justified without this additional calculation.

      Table 5-5 summarizes EPA's estimate of the percent reductions resulting from
the improved evaporative test procedure.  The percent reduction factors are applied
to MOBILES to estimate fleet-average emission reductions and final emission levels,
in grams per mile, resulting from the improved evaporative test procedure.

                                  Table 5-5
                       Percent Reductions in Emissions
Category
Running loss
Hot soak
Diurnal:
Partial
Full
2-3 Days
4+ Days
Resting Loss
Problem-free
Reduction
80%
50%

50%
50%
75%
40%
75%
Pressure-fail
Reduction
£30%*
£30%*

0
0
0
0
75%
Purge-fail
Reduction
£30%*
<30%*

0
0
0
0
75%
          See Table 5-4.
D.    Projected Emission Factors

      The emission reduction factors discussed in Section C above are based on light
duty vehicle technology, but also apply to light-duty trucks and heavy-duty vehicle*
It is considered appropriate to make this simplifying assumption for the light-dutv
truck classification because light-duty trucks resemble light-duty vehicles in term.*
of both vehicle technology and in-use driving patterns. The technology used to con t r< i
                                      87

-------
evaporative emissions from heavy-duty vehicles may be somewhat different than that
used for light-duty vehicles,  potentially affecting emission reductions.  However,
because gasoline-fueled heavy-duty vehicles do not make up a large percentage of
total vehicle miles traveled,  assuming the same level of control for heavy-duty
vehicles should not significantly affect overall emission estimates.

      Table 5-6 summarizes the predicted post-control evaporative emission levels
by applying the percent reduction factors summarized in Tables 5-4 and 5-5 to
baseline emission levels from MOBILES.  Table 5-7 summarizes the corresponding
emission reductions for light-duty vehicles.  The MOBILES projection for evaporative
emissions for light-duty  vehicles designed to meet  the new  evaporative control
requirements is 0.23 g/mi (0.14 g/km), a reduction of 0.40 g/mi (0.25 g/km).

                                  Table 5-6
           Post-Control  LDV Emission Levels for Calendar Year 2020
                                in g/mi (g/km)
Category
Running loss
Hot soak
Diurnal
Resting loss
TOTAL
Problem-free
0.06 (0.04)
0.02 (0.01)
0.04 (0.02)
0.02 (0.01)
0.14 (0.09)
Purge-fail
2.01 (1.25)
0.50 (0.31)
0.19 (0.12)
0.02 (0.01)
2.72 (1.69)
Pressure-fail
2.01 (1.25)
0.51 (0.32)
0.42 (0.26)
0.02 (0.01)
2.96 (1.84)
Composite
0.12 (0.07)
0.04 (0.02)
0.05 (0.03)
0.02 (0.01)
0.23 (0.14)
                                  Table 5-7
         Summary of LDV Emission Reductions for Calendar Year 2020
                                in g/mi (g/km)
Category
Running loss
Hot soak
Diurnal
Resting loss
TOTAL
Baseline
0.39 (0.24)
0.06 (0.04)
0.11 (0.07)
0.07 (0.04)
0.63 (0.39)
Post-control
0.12 (0.07)
0.04 (0.02)
0.05 (0.03)
0.02(0.01)
0.23 (0.14)
Reduction
0.27 (0.17)
0.02 (0.01)
0.06 (0.04)
0.05 (0.03)
0.40 (0.25)
                                     88

-------
      The  supplemental modeling discussed in Appendix A,  which uses a very
different approach than MOBILES to determine emission factors, predicts similar
emission reductions resulting from the new evaporative test procedure. The results
of this modeling support EPA's position that vehicles designed to CARB's adopted test
could have very high in-use emissions if manufacturers substantially delayed canister
purging at  the beginning of a trip. The current results reinforce the findings of the
modeling presented at the January 1992 workshop. The modeling indicates that the
addition  of the supplemental test sequence provides assurance that vehicles will be
designed to perform well under in-use driving conditions.  In fact, the results show
that the test procedure being finalized, by protecting against excessive purge delays,
will provide air quality benefits very near those sought in the last EPA proposal.

      Applying the percent reduction factors in Tables  5-4 and 5-5 to baseline
emissions for light-duty trucks, MOBILES estimates that emissions will drop from
0.45 to 0.20 g/mi (0.28 to 0.12 g/km) as a result of the improved  evaporative test
procedure,  a reduction of 0.25 g/mi (0.16 g/km). For heavy-duty vehicles, MOBILES
estimates that emissions will drop from 2.98 to 1.94 g/mi (1.85 to 1.21 g/km), a
reduction of 1.04 g/mi (0.65 g/km). Appendix B contains the MOBILES output files
from  which these numbers  were calculated.   Baseline and  post-control emission
factors from the  MOBILES  output files were weighted  to account for the use of
reformulated gasoline in 40 percent of the country. Overall, MOBILES estimates that
total motor vehicle VOC emissions will be reduced from 1.67 g/mi to 1.32 g/mi (1.04
to 0.82 g/km) as a result of the new  evaporative test procedure, a reduction of 20
percent.

E.    Total Nationwide VOC Emission Reductions

      EPA estimates nationwide evaporative emission reductions by applying the
MOBILES  gram/mile emission reduction estimates to projections of future VMT
(vehicle miles traveled).  The year 2020 was selected for this  purpose  in order to
remain consistent with the environmental benefits analysis discussed above.

      Based on EPA's Fuel  Consumption Model, VMT projections for light-duty
vehicles, light-duty trucks and heavy-duty vehicles for the year 2020 are projected to
be 1780,  970, and 160 billion miles (2860, 1560, and 260 billion km), respectively *5
Applying these VMT projections, the total VOC reductions in the year 2020 resulting
from  the improved evaporative test procedure for light-duty vehicles, light-duty
trucks, and heavy-duty vehicles are projected to be 710,000, 240,000, and  170.000
metric tons of VOC, respectively; this is a total of 1,120,000 metric tons annually As
discussed in the previous section, this results in a total motor vehicle VOC inventory
reduction of 20 percent.
   ^"Draft MOBILE4 Fuel Consumption Model," U.S. EPA and Computer Science
Corporation, April 1991 (Docket A-89-18, item IV-A-3).

                                      89

-------
90

-------
                         Chapter 6 Cost-effectiveness
      Comparing  benefits and costs makes possible an  estimate of the  cost-
effectiveness of emission reductions for the new test requirements.  The estimated
per-vehicle emission reductions (in g/mi) discussed in Chapter 5 are projected over
the vehicle's life, then discounted to quantify the vehicle's lifetime reductions in
present terms.  Vehicle life is assumed  to be ten years and 100,000 miles for all
vehicle classifications in order to  simplify the analysis.  The analysis uses a ten
percent discounting rate. This is the rate commonly used by EPA in performing cost-
effectiveness analyses.

      The resulting discounted lifetime total emission reductions are 26, 16, and 68
kg for light-duty vehicles, light-duty trucks, and heavy-duty vehicles, respectively.
Dividing the costs discussed in Chapter 4 by benefits gives cost-effectiveness figures
of $380, $810, and $160 per metric ton for light-duty vehicles, light-duty trucks, and
heavy-duty vehicles, respectively, and a weighted average cost-effectiveness (based
on projected vehicle registrations in the year 2020) of  $500 per metric  ton.
Table 6-1 summarizes the cost-effectiveness results.

      Although these cost-effectiveness calculations have used a ten percent discount
factor, EPA believes that three percent may be a more realistic rate.    Using this
rate, the resulting cost-effectiveness figures would be $290, $620, and $130 per metric
ton for light-duty vehicles, light-duty trucks,  and heavy-duty vehicles, respectively.
and $380 per metric ton overall.

      These figures are conservative in that they do not factor in the cost savings
over the lifetime of the vehicle caused by improved fuel economy. Applying these fuel
consumption credits (Table 4-5) results in an overall cost-effectiveness  of $170 per
metric ton.

      Even  considering GM's  cost  estimate of $100 per  vehicle,  which  was
insufficiently substantiated, the cost-effectiveness would be $3800 per metric ton for
light-duty vehicles.
    ^"Draft MOBILE4 Fuel Consumption Model," U.S. EPA and Computer Science
Corporation, April 1991 (Docket A-89-18, item IV-A-3).

      "Supplemental Guidelines on Discounting in the Preparation of Regulator-.
Impact Analysis," U.S. EPA, Office of Policy, Planning, and Evaluation, 1989

                                      91

-------
               Table 6-1
           Cost-Effectiveness
of Improved Evaporative Emission Control

Emission reduction
Discounted lifetime total
Vehicle cost
Discounted cost per metric ton
Light-Duty
Vehicles
0.40 g/mi
(0.25 g/km)
26kg
$10
$380
Light-Duty
Trucks
0.25 g/mi
(0.16 g/km)
16kg
$13
$810
Heavy-Duty
Vehicles
1.04 g/mi
(0.65 g/km)
68kg
$11
$160
                   92

-------
                  Appendix A
Evaporative Modeling with In-Use Driving Patterns

-------
A.    Overview

      In September 1991, EPA completed a draft report describing an analysis that
compared the effects of using various evaporative emission test procedures for vehicle
certification.1 This analysis concluded that the proposed EPA test provided the best
assurance for control of evaporative emissions.  An opportunity to critique the
analysis was given in the January 1992 workshop. The  analysis at that time used
a model that was designed to predict how often a vehicle's canister would be loaded
past breakthrough based on the vapor  generation from diurnal heat  builds and
canister purge  during driving.   Since  the January  1992 workshop, EPA has
considered comments on the original model and has developed an upgraded model
that better simulates real world evaporative emissions.

      The objective of the upgraded evaporative emission model is to calculate
canister emissions during diurnal, hot soak, and driving episodes by maintaining a
continuous accounting of canister condition (storage capacity).  This is achieved by
calculating the mass of hydrocarbons (HO going to the canister due to (1) ambient
heating, (2) hot soak following drives, and (3) heating of the fuel in the fuel tank
when the vehicle is driven.  Also, the model calculates the mass of HC purged from
the canister during vehicle operation and backpurge as a result of ambient cooling.
Resting losses are not modeled. The model simulates vehicle  driving patterns by
tracking the start and end times of each drive and park event and the distance
traveled in each drive event.  These driving patterns were taken from a database
developed  for General Motors  by  National Purchasing  Diary  (NPD) as described
below.

B.    NPD Details

      The vehicle drive characteristics in the September 1991  analysis and in this
model are based on an automotive usage database developed for General Motors in
1979  by NPD.   This database spanned one week  and included 2,870  vehicles.
Participants in the NPD survey were asked to keep a travel diary for each of their
vehicles for one week.  Entries into the  diary included:  number of trips per day,
   ^'Emission Evaluation of the GM Real Time Evaporative Test Procedure," Draft
EPA technical  report by Julie Hayden,  September  1991 (Docket A-89-18,  item
III-B-2).

   2"Automobile Usage Data Base," William M. Spreitzer, General Motors Research
Labs, 1979.

                                    A-l

-------
distance traveled per trip, and time of the beginning and end of every trip taken.  A
trip was defined as the time between engine start-up and engine shut-down.

      For the model, the trip data were screened in order to identify and remove any
obviously mistaken trip entries. After the screening process, 1,787 vehicles remained.
Because the canister condition and evaporative control of a given vehicle depends on
each trip, vehicles  with trips having  any of the  following  characteristics were
completely removed from the database:

      1.    longer than 1440 minutes (24 hours),
      2.    greater than 1560 miles (2510 km) (65 mph (105 km)  for 24 hours),
      3.    average speed less than 1 mph (1.6 km/h) or greater than 75 mph-(120
           km/h),
      4.    average speed less than 15 mph (24  km/h) for over 600 minutes,
      5.    overlapping trips.
      Compared to U.S. census data, NPD underestimates the average number of
secondary and tertiary vehicles per household. To correct for this inconsistency, EPA
added additional secondary and tertiary vehicles to the database. Driving patterns
for  the  additional vehicles were created by copying the characteristics of NPD
secondary and tertiary trip records. The resulting input file used for the analysis had
2,067 vehicles.

      In addition to the revised NPD data, the model provides the option of using an
input file with drive characteristics defined by the user.  This was necessary for
designing a vehicle to pass a given test sequence.

C.    Canister Loading

      An evaporative canister stores fuel vapors generated from the fuel tank until
those vapors are purged from the carbon matrix. When the load exceeds the capacity
of the canister, breakthrough occurs. For the purpose of this analysis, breakthrough
is defined as the point at which the first significant amount of HC escapes from the
canister. For each vehicle, the model tracks the canister condition through the week,
calculating how often breakthrough occurs and how much vapor is generated.
Figure A-l  shows how  the  model calculates canister loading efficiency beyond
breakthrough. Once  vapors begin escaping, the efficiency gradually decreases until
the canister reaches saturation (zero percent collection efficiency). The breakthrough
point on the curve is defined by a canister load equal to two-thirds of the canister
load at saturation. This curve is based on data collected by EPA.
                                    A-2

-------
s
 uoofc-	
<->   V
a   \
•rt     X Empty
                                          - Breakthrough
   50
 CO
 ^•^
u
    o
                  2/3
                                            Saturation
              Canister Condition (mass)

-------
D.    Purge Modeling

      The model uses a purge curve,  developed  from General Motors data, to
represent the effectiveness of purge from the canister as a function of the volume of
air drawn through the canister (Figure A-2). Each point on the curve defines the
purge behavior of a canister at a certain fill level. For example, the origin represents
a saturated canister.  The model uses the purge curve and the load curve to keep
track of the canister condition at any time. The purge curve shows that vapor
removal is most effective when the canister is  full; as the canister empties, an
increasing amount of purge air is required to remove the same mass of HC. The
model can be run with various purge curves, but the analysis described here relies
solely on the curve in Figure A-2.

E.    Diurnal

      A diurnal heat build is simulated by the model by adding a mass of HC to the
canister for every day that the vehicle experiences diurnal loading. This amount is
a function of the fuel volatility, ambient temperatures, operation characteristics, and
the fill level of the fuel tank.  To compare the GARB and EPA test procedures, EPA
analysis uses both 7 psi (48 kPa) and 9 psi (62 kPa) RVP fuel, 55 percent fuel tank
fill level,  and diurnal loading  based on  both  test temperatures  and  recorded
temperatures. The 55 percent fill level  represents an average in-use gasoline fuel
tank fill level determined by analysis of vehicle refueling data.3

      The amount of vapor generated each day in a full diurnal event depends on the
daily temperature swing. For the modeling discussed in this report, maximum daily
temperature data from an actual hot  week in Chicago in 1988 were used.   These
temperatures are not  atypical of hot  weather conditions in Class C areas of the
country.  The average  of the maximum daily temperatures for the week is equal to
96° F, which matches the  maximum daily temperature  used in  the MOBILES
modeling discussed in Chapter 5  of the Final RIA for this rulemaking. Modeling the
day-to-day variations  in ma-rimnm temperatures in  this  way provides canister
loadings  more typical of real-life conditions than would  a hypothetical week of
nonvarying hot  days.   The measured lower temperatures were conservatively
increased by 2° F (1.1° C) so that the average temperature swing would match the
EPA test temperature swing of 24° F (13.3°  C) (i.e. 72° to 96° F (22.2° to 35.6° C».
The temperature data and resulting diurnal canister loadings are shown in Table A-1.
The  diurnal  canister  loadings  in Table A-l were determined  using a  method
   3"Use of RADIAN Fuel Weathering Model to Generate  a MOBILE4.0  Fuel
Weathering Correlation," EPA memo from Gina Shreve to the Record, June 1,  19 By
(Docket A-89-18, item IV-B-7)

                                    A-4

-------
100
Figure A-2: Canister Purge Curve
     (Normalized to 1  Liter)
          breakthrough
            Purge Air (cubic feet)

-------
developed by EPA based on work by Reddy.4
                                 Table A-l
               HC Grams Loaded During Full Diurnal (Per Day)
Day
One
Two
Three
Four
Five
Six
Seven
Temperature Range
in °F (°C)
75-93 (23.9-33.9)
72 - 97 (22.2 - 36.1)
70 - 103 (21.1 - 39.4)
78 - 103 (25.6 - 39.4)
76 - 101 (24.4 - 38.3)
67-92 (19.4-33.3)
64-80 (17.8-26.7)
Vapor Generated
in grams
16.9
21.0
31.1
28.8
26.8
17.2
14.3
      The model assumes that the canister only experiences diurnal loading when
the vehicle is parked and the fuel temperature is less than the ambient temperature.
In order to determine whether a vehicle experiences diurnal vapor generation, the
model first checks to see if the vehicle is parked during the day's period of increasing
ambient temperature.  If the vehicle is parked long enough for the fuel to cool  to
ambient temperature, then the appropriate mass of HC is added to the canister.  If
the vehicle is never parked and cooled during the increasing temperature period, then
no HC is added to the canister.

      The period of daily temperature increases and the length of time for the fuel
in the tank to cool to ambient temperature are input by the user.  For this analysis,
EPA assumes  a four-hour  cool down time and an ambient heating window from
10:00 a.m. to 4:00 p.m.  Under these conditions, if a vehicle's last drive were to end
at 9:00 a.m., the fuel would be cooled at 1:00 p.m. and the vehicle would receive a
diurnal loading for the remainder of the ambient heating time.  However, if this
vehicle were to be driven again from, say 12:00 p.m. to 12:30 p.m., the fuel would not
cool until 4:30 p.m., and the vehicle would receive no diurnal loading for that day.

      The relationship between the vapor generated and the period that the vehicle
is parked is modeled using the following equation:
   4"PT Evaporative Emissions Model, Description and Users Guide, Release 2.02
David B. Bartus, U.S. EPA, June 1990, (Docket A-89-18, item IV-B-2).
                                    A-6

-------
      Load = a x [(Tf)b - (T/]
(1)
where:

      Tj = time when ambient temperature begins heating fuel
      Tf = time when ambient temperature stops heating fuel
      a =   c/[(16)b - (10)b]  (a is constant for each day; 16 and 10 correspond to
            4 p.m. and 10 a.m., the limits of the full diurnal heating window)
      b = volatility factor =  1.5 for this analysis
      c a full diurnal loads from Table A-l

F.    Hot Soak

      The model simulates a hot soak by estimating vapor generation from the fuel
tank, which depends on fuel volatility, trip mileage before the hot soak, and park
duration (Table A-2).   The model adds HC  to the  canister following each trip
depending on the distance traveled during the trip and the duration of the park time
after the trip. Data from CRC-Radian show that about 60 percent of the hot soak
loading occurs in the first ten minutes following a drive; about 80 percent of hot soak
loading occurs in the first 30 minutes.5

                                 Table A-2
                      Hot Soak Canister Loading (grams)
Trip Distance
in miles (km)
0-3 (0-4.8)
3-7 (4.8-11.3)
7-11 (11.3-17.7)
>11 (17.7)
Park Duration (minutes)
0-10
2.5
3.9
5.1
5.6
10-30
3.6
5.5
7.2
7.9
30+
4.2
6.5
8.5
9.3
   5"CRC-Radian Evaporative Emissions Model:  Evaluation of Time and Driving
Effects,"  1990 Annual Report, Radian Corporation, ABRAC VE-4 Project Group.
Coordinating Research Council, May 4, 1992.
                                    •A-7

-------
G.    Vehicle Operation

      1. Purge Design

      For periods of vehicle operation, the model calculates the mass of hydrocarbons
purged from the canister to the engine. Vehicles are typically designed so that purge
does not occur continuously while the engine is running.  A purge delay of two
minutes is assumed to allow the engine to warm up and begin closed-loop operation
before accepting purged hydrocarbons from the canister. The amount of delay is a
variable in the model.

      The model assumes that no purge will occur during decelerations or at idle.
To calculate the amount of purge from the canister during a trip, the model breaks
the trip into purge and no-purge intervals that better simulate real driving.  The
length of the purge delay is subtracted from the trip duration.  The remaining trip
is divided into seven equal alternating intervals of purge and no-purge (Figure A-3).
In addition, the last no-purge period simulates deceleration of the vehicle at the end
of a trip.  This alternation better simulates actual drive patterns than assuming that
the vehicle would purge continuously.
                                 Figure A-3
                  Representation of Simulated Purge Strategy
   purge
   delay             PNPPNP          PNPPNP
                     *•      * *      *     *™          *     *• *    *    ^*
time -»

       P = period of purge          NP a period of no-purge


      2. Vapor Load

      Hydrocarbons generated during drive periods are also added to the canister
through the use of a vapor generation curve included in the model (Figure A-4). The
figure shows vapor generation as a function of driving time. The model assumes that
all vapor generated is sent to the engine during periods of canister purging and that
all vapor generated  during no-purge periods is sent to the canister.

      The model accounts for the fact that fuel temperatures are not the same at the
beginning of each trip.  Fuel tank temperature rises are calculated using a heating
curve and estimates of cooling during non-drive periods are calculated using one of

                                    A-8

-------

-------
two cooling curves. These temperature curves were developed from test data on in-
use vehicles.6  Slightly different cool down curves are used to model daytime and
nighttime ambient temperature conditions.

      The heating and cooling curves contained in the model relate temperature to
driving time. The fuel tank temperature curves are used in conjunction with the
vapor generation curve to provide estimates of vapor generation as a function of both
fuel temperature and driving time.

H.    Backpurge

      In order to represent backpurge due to the cooling of the fuel at night, an
amount of HC is removed from the canister for vehicles not driven when ambient
temperatures decrease. The model uses a similar method for calculating backpurge
as it does for diurnal loading by removing a prescribed mass (0.35 ft3 (9.9 liters) for
this  analysis) of HC  from the canister  of those vehicles that are  parked in an
appropriate time window.  This window was assumed to  be from 4:00 a.m. to
8:00 a.m.

I.     Cases Modeled and Results

      Modeling was performed for four cases, using canister sizes and purge rates
designed to pass the following test procedures:

      1.     EPA's current procedure,
      2.     CARB's procedure (three-diurnal test following exhaust test and running
            loss test), assuming 9 psi (62 kPa) RVP fuel and 72° to 96° F (22.2° to
            35.6° C) diurnal temperatures, and assuming purge is delayed for 20
            minutes to avoid exhaust interactions,
      3.     EPA's proposed procedure for the January 1992 workshop (running loss
            test following three-diurnal test following exhaust test),  and
      4.     Final rulemaking (FRM) procedure (three-diurnal test following exhaust
            test and running loss test and supplemental test involving two-diurnal
            test following exhaust test).
      These cases were run using the 2,067 NPD vehicles described above in order
to calculate evaporative emissions.  The results of this analysis are presented in
Table A-3.
   6EPA Contract 68-C-90027, Automotive Testing Laboratories, 1988 through 1991

                                    A-10

-------
      The results confirm the conclusion reached by the simpler modeling presented
at the January 1992 workshop: if design responses to the GARB test procedure were
to include large purge delays in order to avoid exhaust interactions during  the
exhaust emission test, very large in-use emissions would result.  The results also
show that the FRM test procedure, by protecting against excessive purge delays, will
provide air quality benefits very near those sought in the last EPA proposal.

                                  Table A-3
                Test Procedure Effects on Evaporative Emissions
Test Procedure
Current
GARB with large purge delay
January 1992 workshop proposal
EPA Final Rule
Emissions
in g/mi (g/km)
0.5 (0.3)
1.0 (0.6)
0.04 (0.02)
0.05 (0.03)
                                    A-ll

-------
         Appendix B
MOBILES Input and Output Files

-------
                                                         INPUT TIL*
1Enhanced  I/M without Reformulated Gasoline  & without New Evap  Test  Procedure  (No Phase-In)
 MOBILES  (4-Dec-92)
OEvaporative Teat Procedure Phase-in Years and Percentages:
 Model Year: 2020
 Percentage:   0.
OI/M program fl selected!                           I/M program  |2 selected:
OStart year  (Jan 1):  1983
 Pre-1981 stringency: 20%
 First MYR covered:   1968
 Last MYR covered:    2020
 Haiver  (pre-1981):    1.%
 Waiver  (1981+):       1.%
 Compliance Rate:     98.%
 Inspection type:
 Teat Only
 Inspection frequency: Annual
 I/M program |1 vehicle types
      LDGV - Yes
     LOGT1 - Yes
     LOGT2 - Yes
      HDGV - No
 1981 t  later MYR test type:
      Idle
 Outpoints, HC:   220.000
 Outpoints, CO:     1.200
 Cutpoints, NOx:  999.000

OFunctional Check Program Description:
OCheck Start   Model Yrs  Vehicle Classes Covered
       (Janl)  Covered    LDGV   LDGT1  LOGT2  HDGV
                                     Start  year  (Jan  1):   1990
                                     Pre-1981  stringency:  20%
                                     First  MYR covered:    1984
                                     Last MYR  covered:     2020
                                     Naiver (pre-1981):     1.%
                                     Haiver (1981+):        1.%
                                     Compliance  Rate:      98.%
                                     Inspection  type:
                                     Test Only
                                     Inspection  frequency:
                                     I/M program |2 vehicl
                                         LDGV - Yes
                                        LDGT1 - Yea
                                        LDGT2 - Yes
                                         HDGV - No
                                     1981 £ later MYR teat
                                         IM240  test
                                     Cutpoints,  HC:     0.
                                     Cutpoints,  CO:     IS.
                                     Cutpoints,  NOx:  999.
 Annual
e types
 type:

800
000
000
                                           Inspection
                                        Type      Freq
 Press 1990
 Purge 1990
 ATP   1983
OAir pump syst<
1971-2020  Yes    Yes    Yes    No  Test Only     Annual
1984-2020  Ye*    Yes    Yes    No  Test Only     Annual
1975-2020  Yes    Yes    Yes    No  Teat Only     Annual
m disablements:       Yes  Catalyst removals:
 Fuel inlet restrictor disablements: Yes
 EGR disablement:                    No
 PCV system disablements:            Yes
                           Tailpipe lead deposit test:
                           Evaporative system disablements:
                           Missing gas caps:
                                Minimum Temp:  72.  (F)
                       Period 1 RVP: 11.5      Period 2 RVP:  9.0
OVOC HC emission factors include evaporative HC emission factors.
0
   Comp
   Rate

   98.0%
   98.0%
   98.0%
      Yea
      Yea
      Yea
      Yea
      Maximum Temp:  96.
 Period 2 Start Yr: 1992
                                                                                                  (F)
                                                       B-l

-------
                                                       OUTPUT 7H.KS
Paaai.no Vahi.cJ.aa
OEmiasion factors are as of July 1st  of the indicated calendar year.
OCal. Year:  2020
0  Veh. Type:
                                                                    90.5
                                                                    27.3
                        I/M Program:  Yes
                  Anti-tarn. Program:  Yes
                   Reformulated Gas:  No
                  User-Supplied Pass,  Purge Fail,  and Pressure Fail Rates
                                               Ambient Temp:  90.5
                                             Operating Mode:  20.6
                                                      / 90.5  (F) Region: Low
                                                      / 20.6   Altitude:   500.
                   LDGV
 Veh. Speeds:     19.6
     VMT Mix:      O.S75
OComposite Emission Factors
          HC:
          HC:
 VOC
 Exhaust
 Evaporat HC:
 Refuel L HC:
 Runing L HC:
 Rating L HC:
 Exhaust  CO:
 Exhaust NOX:
1.53
0.71
0.17
0.20
0.38
0.07
9.76
1.27
LDGT1

19.6
 0.207
(Gm/Mile)
 1.40
 0.77
 0.18
 0.26
 0.11
 0.07
10.58
 1.41
                                      LOGT2
OEvaporative Emissions by Component
 (All Components in Grams per Mile)
 Hot Soak
 Diurnal
 Multiple
 Crankcase
 8am-11am
 10am-3pm
 8am-2pm
                   0.07
                   0.13
                   0.39
                   0.00
                   0.00
                   0.01
                   0.02
          0.06
          0.13
          0.46
          0.00
          0.00
          0.01
          0.02
                                      13.02
                                       2.00
                                       0.06
                                       0.12
                                       0.42
                                       0.00
                                       0.00
                                       0.01
                                       0.02
                              LDGT
                                                 1.46
                                                 0.84
                                                 0.18
                                                 0.26
                                                 0.11
                                                 0.07
                                                11.31
                                                 1.59
                                                   06
                                                   13
                                                   45
                                                 0.00
                                                 0.00
                                                 0.01
                                                 0.02
                                                           HDGV

                                                          TiTTe
                                                           0.034
                                                            .03
                                                            .96
                                                            .19
                                                            .42
                                                            .38
                                                           0.08
                                                          24.38
                                                           3.73
4.
1.
1.
0.
0.
 LOOV

1976
 0.002

 0.52
 0.52
          1.44
          1.09
                                                                          (*) : 100.,
                                                                               LOOT
19.6
 0.005

 0.73
 0.73
           1.61
           1.24
0.,   0.
  HDDV

 T976
  0.084

  2.10
  2.10
                                                                                                   Ft
                                                                                                   MC
19.6
 0.004

 6.38
 1.89
 4.03
                                                        Weathered RVP:  8.5
                               0.72
                               0.55
                               1.15
                               0.01
                     0.46
          11.18     26.11
           6.58      0.76
            Hot Soak Temp:
        Running Loss Temp:
        Resting Loaa Temp:
                     2.16
                     1.87

                     0.00
                                                                                                           All Veh
                                                                                          1
                                                                                       91.8
                                                                                       92.6
                                                                                       85.2
 1.655
 0.913
 0.206
 0.205
 0.262
 0.069
10.847
   890
   (F)
   (F)
   (F)

-------
Pure-Fail
0Emission factors are aa of July lat of the indicated calendar year.
OCal. Year: 2020
   Veh. Type:
  Ambient Temp: 90.5 /
Operating Mode: 20.6 /
90.5
27.3
      I/M Program: Yea
Anti-tarn. Program: Yea
 Reformulated Gaa: No
User-Supplied Paaa, Purge Fail, and Pressure Fail Rates
 LDGV  .   LOGT1     LOGT2      LDGT      HDGV      LOOV
/ 90.5 (F) Region: Low
/ 20.6   Altitude:  500. Ft.
                             (*):   0., 100.,   0.
                                  LOOT      HDDV
                                                      MC
                                                             All  Veh
Veh . Speeds :
VMT Mix:
"i575
19.6
0.207
OComposite Emission Faotors (Gm/Milej
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
5.45
0.71
1.05
0.20
3.41
0.07
9.76
1.27
OEvaporative Emissions by
(All Component a

Hot Soak
Diurnal
Multiple
Crankcaae
8 am- 11 am
10am- 3pm
8 am- 2pm
in Grama

1.11
0.41
0.41
0.00
0.02
0.12
0.17
5.61
0.77
1.09
0.26
3.41
0.07
10.58
1.41
Component
per Mile)

1.09
0.50
0.50
0.00
0.04
0.17
0.22
19.6
0.089
1
5.73
1.00
0.98
0.27
3.41
0.07
13.02
2.00


5
0
1
0
3
0
11
1


.65
.84
.05
.26
.41
.07
.31
.59
19
0

9
1
3
0
3
0
24
3
.6 19.6
.034 0.002

.85 0.52
.96 0.52
.97
.42
.41
.08
.38 1.44
.73 1.09
Weathered RVP: 8.5


0.99
0.46
0.46
0.00
0.04
0.15
0.20


1
0
0
0
0
0
0


.07
.49
.49
.00
.04
.16
.22


3
1
1
0





.20
.15
.15
.01



19.6 19.6 19.6
0.005 0.084 0.004

0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting








2.35
1.89
0.00


0.46
26.11
0.76
Soak Temp :
Loss Temp:
Loss Temp:
0.00
0.00

0.00




5
0
1
0
3
0
10
1
91 .8
92.6
85.2








.325
.913
.049
.205
.089
.069
.847
.890
(F)
(F)
(F)







                                                       B-3

-------
Preaaure-Fail Vehicles

OEmiasion factors are as of July 1st  of the indicated calendar year.
OCal. Year: 2020
      I/M Program:  Yes
Anti-tarn. Program:  Yes
 Reformulated Gas:  No
                                             Ambient Temp:  90.5 /  90.5 /
                                            Operating Mode:  20.6 /  27.3 /
90.5 (F) Region: Low
20.6   Altitude:  500. Ft.
User-Supplied Pass,
0 Veh. Type: LDGV .
Veh. Speeds: 19.6
VMT Mix: 0.575
LDGT1
19.6
0.207
Purge Fail, and Pressure Fail Rates (%) :
LDGT2
19.6
0.089
LDGT


HDGV
19.6
0.034
LDDV
19
0
.6
.002
0-,
LOOT
19
0
.6
.005
0., 100
HODV
19.6
0.084
.
MC
19.6
0.004

All Veh


OComposite Emission factors (Cm/Mile)
VOC HC: 5.69
Exhaust HC: 0.71
Evaporat HC: 1.29
Refuel L HC: 0.20
Runing L HC: 3.41
Rsting L HC: 0.07
Exhaust CO: 9.76
Exhaust NOX: 1.27
OEvaporative Emissions by
5.89
0.77
1.36
0.26
3.41
0.07
10.58
1.41
Component
5.99
1.00
1.23
0.27
3.41
0.07
13.02
2.00

5.92
0.84
1.32
0.26
3.41
0.07
11.31
1.59

10.18
1.96
4.30
0.42
3.41
0.08
24.38
3.73
Weathered
0
0




1
1
RVP:
.52
.52




.44
.09
8.5
0
0




1
1

.73
.73




.61
.24

2.10
2.10




11.18
6.58
Hot
2.35
1.89
0.00


0.46
26.11
0.76
Soak Temp:
5.555
0.913
1.278
0.205
3.089
0.069
10.847
1.890
91.8 (F)
(All Components  in  Crams  per Mile)
 Hot Soak
 Diurnal
 Multiple
 Crankcase
 8am-11am
 10am-3pm
 8am-2pm
1.12
0.62
0.62
0.00
0.31
0.40
0.44
1.11
0.74
0.74
0.00
0.39
0.48
0.53
1.01
0.68
0.68
0.00
0.36
0.45
0.48
1.08
0.73
0.73
0.00
0.38
0.47
0.51
3.24
1.58
1.58
0.01



                                                                    Running Loss Temp:  92.6
                                                                    Resting Loss Temp:  85.2
                                                                                 0.00
                                                                                 0.00

                                                                                 0.00
                                                                                                               (F)
                                                       B--1

-------
Compos 1t e
OEmiasion factors are as of July 1st of the  indicated  calendar  year.
OCal. Year: 2020
I/M Program: Yes
Anti-tarn. Program: Yea
Ambient
Operating
Tamp:
Mode:
90.5 / 90.5 / 90.5 (F) Region: Low
20.6 / 27.3 / 20.6 Altitude: 500. Ft.
Reformulated Gas : No
0 Vah. Type:
•f
Veh . Speeds :
VMT Mix:
LDGV

TsTfi
0.575
LDGT1

T576
0.207
LOGT2

19
0

.6
.089
LDGT



HDGV LDDV

19
0

.6 19.6
.034 0.002
LDDT HDDV

19.6 19.6
0.005 0.084
MC

19. b
0.004
Ail



Veh



OComposite Emission Factor* (On/Mile)
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Honing L HC:
Rating L HC:
Exhauat CO:
Exhaust NOX:
1.65
0.71
0.20
0.20
0.47
0.07
9.76
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8 am- 11 am
10am- 3pm
8 am- 2pm
in Cram*

0.10
0.14
0.41
0.00
0.01
0.02
0.03
1.54
0.77
0.22
0.26
0.21
0.07
10.58
1.41
Component
per Mile)

0.10
0.15
0.47
0.00
0.01
0.02
0.03
1
1
0
0
0
0
13
2
.77
.00
.20
.27
.23
.07
.02
.00
1.61
0.84
0.21
0.26
0.22
0.07
11.31
1.59
5
1
2
0
1
0
24
3
.85 0.52
.96 0.52
.12
.42
.27
.08
.38 1.44
.73 1.09
Weathered RVP: 8.5


0
0
0
0
0
0
0


.09
.14
.43
.00
.01
.02
.03


0.10
0.15
0.46
0.00
0.01
0.02
0.03


1
0
1
0





.45
.80
.57
.01



0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting







6.38
1 .89
4.03


0.46
26.11
0.76
Soak Temp:
Loss Temp:
Loss Temp :
2.16
1.87

0.00



1
0
0
0
0
0
10
1
91 .8
92.6
85.2







.829
.913
.267
.205
.375
.069
.847
.890

-------
                                           POST-CONTROL  INPUT FZUE
INew Evap Test Procedure with Enhanced I/M & without Reformulated Gasoline
 MOBILES <4-Dec-92)
OI/M program |1 selected:                          I/M program |2 selected:
OStart year (Jan 1):  1983
 Pre-1981 stringency: 20%
 First MYR covered:   1968
 Last MYR covered:    2020
 Waiver (pre-1981):    1.%
 Waiver (1981+):       1.%
 Compliance Rate:     98.%
 Inspection type:
 Test Only
 Inspection frequency: Annual
 I/M program 11  vehicle types
      LOGV - Yes
     LDGT1 - Yes
     LDGT2 - Yes
      HDGV - No
 1981 4 later MYR test type:
      Idle
 Outpoint3,  HC:    220.000
 Cutpoints,  CO:      1.200
 Outpoints,  NO:i:  999.000
Start year (Jan 1):  1990
Pre-1981 stringency: 20%
First MYR covered:   1984
Last MYR covered:    2020
Waiver (pre-1981):     1.%
Waiver (1981+):        l.ft
Compliance Rate:     98.%
Inspection type:
Test Only
Inspection frequency:  Annual
I/M program |2 vehicle types
     LOGV - Yes
    LOGT1 - Yes
    LOGT2 - Yes
     HDGV - No
1981 £ later MYR test  type:
     IM240 test
Cutpoints, HC:     0.800
Cutpoints, CO:     15.000
Cutpoints, ttOx:  999.000
OFunctional Check Program Description:
OCheck Start   Model Yrs  Vehicle Classes Covered
       (Janl)   Covered    LDGV   LDGT1   LDGT2  HDGV
      Inspection
    Type      Freq
 Press 1990    1971-2020  Yes
 Purge 1990    1984-2020  Yes
 ATP   1983    1975-2020  Yes
OAir pump system disablements:
 Fuel inlet restrictor disablements;  Yes
 EGR disablement:                    Mo
 PCV system disablements:            Yes
0	
                       Period 1 RVP:  11.5
Yes    Yes    No  Test Only     Annual
Yes    Yes    No  Test Only     Annual
Yes    Yes    No  Test Only     Annual
    Yes  Catalyst removals:
         Tailpipe lead deposit test:
         Evaporative system disablements:
         Missing gas caps:
              Minimum Temp:  72. (F)
              Period 2 RVP:  9.0
Comp
Rate

98.0%
98.0%
98.0%
   Yes
   Yes
   Yes
   Yes
   Maximum Temp:
                                          96.  (F)
                      Period 2 Start Yr: 1992
OVOC HC emission factors include evaporative HC emission factors.
0                                       	
                                                       B-6

-------
                                          POST-CONTROL OUTPUT
Paaaini Vehiclaa
OEmiaaion factors are  aa  of July  lat of the indicated calendar year.
OCal. Year:  2020
      I/M Program:  Yea
Anti-tarn. Program:  Yea
 Reformulated Gaa:  No
  Ambient Temp: 90.5 /  90.5  /
Operating Mode: 20.6 /  27.3  /
90.5 (F)  Region:  Low
20.6   Altitude:   500.  Ft.
0 Veh . Type :
Veh . Speeda :
VMT Mix:
Uaer-Supplied Pass, Purge Fail, and Pressure Fail Rates (%)
LDGV LDGT1 LOGT2 LDGT HDGV LDDV
19.6
0.575
19.6
0.207
19
0
.6
.089

19
0
.6 19.6
.034 0.002
: 100., 0., 0.
LDDT HDDV MC
All
Veh
19.6 19.6 19.6
0.005 0.084 0.004
OCompoaite Emission Factors (Gm/Mile)
VOC HC:
Exhauat HC :
Evaporat HC:
Refuel L HC:
Run ing L HC:
Rating L HC:
Exhauat CO:
Exhauat NOX:
1.08
0.71
0.08
0.20
0.08
0.02
9.76
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcaae
8 am- 11 am
10am- 3pm
8am-2pm
in Grama

0.03
0.06
0.14
0.00
0.00
0.01
0.01
1.16
0.77
0.08
0.26
0.02
0.02
10.58
1.41
Component
per Mile)

0.03
0.07
0.17
0.00
0.00
0.01
0.01
1
1
0
0
0
0
13
2
.39
.00
.07
.27
.02
.02
.02
.00
1.23
0.84
0.08
0.26
0.02
0.02
11.31
1.59
3
1
0
0
0
0
24
3
.09 0.52
.96 0.52
.60
.42
.09
.03
.38 1.44
.73 1.09
Weathered RVF: 8.5


0
0
0
0
0
0
0


.03
.06
.15
.00
.00
.01
.01


0.03
0.07
0.16
0.00
0.00
0.01
0.01


0
0
0
0





.37
.28
.43
.01



0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting







6.38
1.89
4.03


0.46
26.11
0.76
Soak Temp :
Loss Temp :
Los a Temp:
2.16
1.87

0.00



1
0
0
0
0
0
10
1
91.8
92.6
85.2







.296
.913
. 104
.205
.054
.020
.847
.890
(F)
(F)
(F)







                                                      B-7

-------
Purge—Fall Vehicles
0Emission factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020
0  Veh. Type:
                                               Ambient Temp: 90.5
                                             Operating Mode: 20.6
/ 90.5 / 90.5
/ 27.3 / 20.6
      I/M Program: Yes
Anti-tarn. Program: Yea
 Reformulated Gas: No
User-Supplied Pass,  Purge Fail, and Pressure Fail Rates
 LDGV     LDGT1     LDGT2      LDGT      HOGV      LDDV
(F)  Region: Low
  Altitude:  500. Ft.
                                                                          (%):   0.,  100.,    0.
                                                                               LOOT      HDDV
                                                                                                    MC
                                                                                                           All Veh
Veh . Speeds :
VMT Mix:
19.6
0.575
1576-
0.207
OComposite Emission Factors (Cm/Mile]
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Riming L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
4.23
0.71
0.83
0.20
2.48
0.02
9.76
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8am-llam
10am- 3pm
8an-2pm
in Grams

0.81
0.41
0.41
0.00
0.02
0.12
0.17
4.41
0.77
0.88
0.26
2.48
0.02
10.58
1.41
Component
per Mile)

0.81
0.50
0.50
0.00
0.04
0.17
0.22
19
0
1
4
1
0
0
2
0
13
2
.6
.089

.57
.00
.79
.27
.48
.02
.02
.00


4.46
0.84
0.85
0.26
2.48
0.02
11.31
1.59
19
0

8
1
3
0
2
0
24
3
.6 19.6
.034 0.002

.06 0.52
.96 0.52
.15
.42
.51
.03
.38 1.44
.73 1.09
Weathered RVP: 8.5


0
0
0
0
0
0
0


.73
.46
.46
.00
.04
.15
.20


0.79
0.49
0.49
0.00
0.04
0.16
0.22


2
1
1
0





.38
.15
.15
.01



19.6 19.6 19.6
0.005 0.084 0.004

0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting








2.35
1.89
0.00


0.46
26. 11
0.76
Soak Temp :
Loss Temp:
Loss Temp :
0.00
0.00

0.00




4
0
0
0
2
0
10
1
91.8
92.6
85.2








.216
.913
.836
.205
.242
.020
.847
.890
(F)
(F)
(F)







                                                       B-8

-------
Preasure-Fail Vehicles

OEmission factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020
      I/M Program: Yes
Anti-tarn. Program: Yes
 Reformulated Gas: No
  Ambient Temp: 90.5 / 90.5 / 90.5
Operating Mode: 20.6 / 27.3 / 20.6
                                                                                (F) Region: Low
                                                                                  Altitude:  500. Ft.
0 Veh . Type :
Veh . Speeds :
VMT Mix:
User-Supplied Pass, Purge Fail, and Pressure Fail Rates (*)
LDGV LOGT1 LDGT2 LDGT HDGV LDDV
19.6
0.575
19.6
0.207
OCompoaite Emission Factors (Gm/Mile]
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Runing L HC:
Rsting L HC:
Exhaust CO :
Exhaust NOX:
4.47
0.71
1.07
0.20
2.48
0.02
9.76
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8 am- 11 am
10am- 3pm
8 am- 2pm
in Grams

0.82
0.62
0.62
0.00
0.31
0.40
0.44
4.69
0.77
1.15
0.26
2.48
0.02
10.58
1.41
Component
per Mile)

0.82
0.74
0.74
0.00
0.39
0.48
0.53
19
0
1
4
1
1
0
2
0
13
2
.6
.089

.82
.00
.05
.27
.48
.02
.02
.00


4.73
0.84
1.12
0.26
2.48
0.02
11.31
1.59
19
0

8
1
3
0
2
0
24
3
.6 19.6
.034 0.002

.38 0.52
.96 0.52
.46
.42
.51
.03
.38 1.44
.73 1.09
Weathered RVP: 8.5


0
0
0
0
0
0
0


.74
.68
.68
.00
.36
.45
.48


0.79
0.73
0.73
0.00
0.38
0.47
0.51


2
1
1
0





.41
.58
.58
.01



: 0., 0., 100.
LOOT HODV HC
All
Veh
19.6 19.6 19.6
0.005 0.084 0.004

0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting








2.35
1 .89
0.00


0.46
26.11
0.76
Soak Temp :
[.033 Temp :
Loss Temp :
0.00
0.00

0.00




4
0
1
0
2
0
10
1
91 .8
92.6
85.2








. 443
.913
.063
.205
.242
.020
.847
.890
(F)
(F)
(F)







                                                       B-'J

-------
OEmission factors are as of July 1st
OCal. Year: 2020

I/M Program:
Anti-tai
B. Program:
Reformulated Gas:
0 Veh. Type:
•f
Veh . Speeds :
VMT Mix:
LOGV

IJTfi
0.575
LDGT1

19.6
0.207
of the indicated calendar year.
Yes
Yes
No
LOGT2

19.6
0.089
Ambient
Operating

LDGT



Temp:
Mode:

90.5
20.6

HDGV

19
0

.6
.034
/ 90
/ 27

.5/90
.3 / 20

LDOV

19
0

.6
.002
.5 (F) Region: Low
.6 Altitude: 500.

LOOT HOOV

19.6 19.6
0.005 0.084

Ft.

MC

19.6
0.004



All Veh



OComposite Emission Factors (Gin/Mile)
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
1.18
0.71
0.10
0.20
0.15
0.02
9.76
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcasa
8am-llam
10 am- 3pm
8am-2pm
in Grams

0.06
0.08
0.15
0.00
0.01
0.01
0.02
1.27
0.77
0.11
0.26
0.10
0.02
10.58
1.41
Component
per Mile)

0.06
0.09
0.18
0.00
0.01
0.02
0.02
1.51
1.00
0.11
0.27
0.11
0.02
13.02
2.00
1.34
0.84
0.11
0.26
0.10
0.02
11.31
1.59
4
1
1
0
0
0
24
3
.64
.96
.43
.42
.81
.03
.38
.73
Weathered


0.06
0.08
0.17
0.00
0.01
0.02
0.02


0.06
0.09
0.18
0.00
0.01
0.02
0.02


0
0
0
0





.97
.61
.85
.01



0
0




1
1
RVP:









.52
.52




.44
.09
8.5









0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot Soak
Running Loss
Resting Loss







6.38
1.89
4.03


0.46
26.11
0.76
Tamp:
Temp:
Temp:
2.16
1.87

0.00



1.437
0.913
0.157
0.205
0.142
0.020
10.847
1.890
91.8 (F)
92.6 (F)
85.2 (F)







b- lu

-------
                             BASKLXHE  IHPOT FXLB -  WITH  RKrOKMDZJlTBD GASOLXHX
lEnhancad I/M with Reformulated Gasoline 6 without New Cvap Test Procedure (No Phase-In)
 MOBILES (4-Dec-92)
OEvaporative Test Procedure Phase-in Years and Percentages:
 Model Year: 2020
 Percentage:   0.
OI/M program |1 selected;                          I/M program |2 selected:
OStart year (Jan 1):  1903
 Pre-1981 stringency: 20%
 First MYR covered:   1968
 Last MYR covered:    2020
 Waiver (pre-1981):    1.%
 Waiver (1981+):       1.%
 Compliance Rate:     96.%
 Inspection type:
 Test Only
 Inspection frequency: Annual
 I/M program |1 vehicle types
      LOGV - Yes
     LDGT1 - Yes
     LDGT2 - Yes
      HDGV - No
 1981 & later MYR test type:
      Idle
 Outpoints,  HC:   220.000
 Outpoints,  CO:     1.200
 Cutpoints,  NO:<:  999.000
                                                   Start year (Jan 1) :   1990
                                                   Pre-1981 stringency: 20%
                                                   First MYR covered:    1984
                                                   Last MYR covered:    2020
                                                   Waiver (pre-1981):     1.%
                                                   Waiver (1981+):       1.%
                                                   Compliance Rate:     98.%
                                                   Inspection type:
                                                   Test Only
                                                   Inspection frequency: Annual
                                                   I/M program |2 vehicle types
                                                        LDGV - Yes
                                                       LOGT1 - Yes
                                                       LDGT2 - Yes
                                                        HDGV - No
                                                   1981 t later MYR test type:
                                                        IM240 test
                                                   Cutpoints, HC:     0.800
                                                   Cutpoints, CO:    15.000
                                                   Cutpoints, NOx:  999.000
OFunctional Check Program Description:
OCheck Start   Model Yra  Vehicle Classes Covered
       (Janl)  Covered    LDGV   LDGT1  LDGT2  HDGV
                                                         Inspection
                                                       Type      Freq
                                 Yea
                                 Yea
                                 Yea
 Preaa 1990    1971-2020  Yea
 Purge 1990    1984-2020  Yea
 ATP   1983    1975-2020  Yea
OAir pump aystem disablements:
 Fuel inlet restrictor disablements: Yea
 EGR disablement:
 PCV system disablements:
                                        Yea    No  Test Only     Annual
                                        Yea    No  Teat Only     Annual
                                        Yea    No  Teat Only     Annual
                                     Yea  Catalyat removals :
                                          Tailpipe lead deposit test:
                                          Evaporative system disablements :
                                          Missing gas caps:
                                               Minimum Temp:  72. (F)
                                     No
                                     Yea
Comp
Rate

98.0%
98.0%
98.0%
   Yes
   Yes
   Yes
                       Period 1 RVP: 11.5      Period 2 RVP:  9.0
OVOC HC emission factors include evaporative HC emission factors.
0
   Yes
   Maximum Temp:
                                                                         Period 2 Start Yr:
                                                                                             96.
                                                                                            1992
                                                                                                 (F)
                                                       B-ll

-------
y.TCT OUTVl!)T
                                                     -  WITH

      a Vahi.cJ.aa
0Emission factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020        I/M Program: Yes       Ambient Temp: 90.5 /
                  Anti-taM. Program: Yes     Operating Mode: 20.6 /
                   Reformulated Gas: Yes         ASTM Class: C
                                     90.5 / 90.5 (F) Region: Low
                                     27.3 / 20.6   Altitude:  500. Ft.
0 Veb. Type:
+
Veh. Speeds:
VMT Mix:
User-Supplied Pass, Purge Fail, and Pressure Fail Rates (%)
UXSV LDCT1 LDGT2 LDGT HDGV LDDV
19.6
0.575
19.6
0.207
19.6
0.089


19
0
.6 19.6
.034 0.002
: 100., 0., 0.
LOOT HDDV MC
All
Veh
19.6 19.6 19.6
0.005 0.084 0.004
OComposite Emission Factors (Cm/Mile)
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Runing L HC:
Rsting L HC:
Exhaust CO:
Exhaust NOX:
1.12
0.57
0.10
0.17
0.22
0.07
7.73
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8 am- 11 am
10 am- 3pm
8a»-2pm
in Grams

0.05
0.04
0.23
0.00
0.00
0.00
0.01
1.14
0.61
0.10
0.22
0.13
0.07
8.35
1.40
Component
per Mile)

0.05
0.05
0.27
0.00
0.00
0.00
0.01
1.31
0.79
0.10
0.23
0.13
0.07
10.15
2.00
1
0
0
0
0
0
8
1
.19
.67
.10
.22
.13
.07
.89
.58
3
1
1
0
0
0
17
3
.29 0.52
.63 0.52
.00
.36
.22
.08
.27 1.44
.77 1.09
Weathered RVP: 7.1


0.05
0.04
0.24
0.00
0.00
0.00
0.01


0
c 0
0
0
0
0
0


.05
.04
.26
.00
.00
.00
.01


0
0
0
0





.58
.55
.81
.01



0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting







7.14
1.80
4.88


0.46
22.45
0.76
Soak Temp:
Loss Temp:
Loaa Temp :
1.89
2.99

0.00



1
0
0
0
0
0
8
1
91.8
92.6
85.2







.319
.766
.140
.174
.170
.069
.706
.886
(F)
(F)
(F)







                                                       B- U

-------
Purge-Fail Vehicles

OEmission factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020
      I/M Program: Yes
Anti-tarn. Program: Yea     Operating Mode: 20.6
 Reformulated Gas: Yes         ASTM Class: C
Ambient Temp: 90.5 / 90.5 / 90.5 (F) Region: Low
                                                                  / 27.3 / 20.6   Altitude:  500. Ft.
0 Veh . Type :
Veh . Speeds :
VMT Mix:
User-Supplied Pass, Purge Fail, and Pressure Fail Rates (%)
LDGV LOGT1 LDGT2 LDGT HOGV LDOV
19.6
0.575
19.6
0.207
19.6
0.089


19
0
.6 19.6
.034 0.002
: 0., 100., 0.
LOOT HODV MC
All
Veh
19.6 19.6 19.6
0.005 0.084 0.004
OCompoaite Emission Factors (Cm/Mile)
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Run ing L HC :
Rating L HC:
Exhaust CO :
Exhaust NOX:
3.19
0.57
0.59
0.17
1.79
0.07
7.73
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8 am- 11 am
10am- 3pm
8 am- 2pm
in Grams

0.62
0.24
0.24
0.00
0.01
0.05
0.08
3.32
0.61
0.62
0.22
1.79
0.07
8.35
1.40
Component
per Mile)

0.61
0.30
0.30
0.00
0.02
0.08
0.12
3.43
0.79
0.56
0.23
1.79
0.07
10.15
2.00
3
0
0
0
1
0
8
1
.35
.67
.60
.22
.79
.07
.89
.58
6
1
2
0
1
0
17
3
.19 0.52
.63 0.52
.33
.36
.79
.08
.27 1.44
.77 1.09
Weathered RVP : 7.1


0.55
0.28
0.28
0.00
0.02
0.07
0.11


0
0
0
0
0
0
0


.59
.30
.30
.00
.02
.08
.12


1
0
0
0





.78
.81
.81
.01



0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting







2.26
1.80
0.00


0.46
22.45
0.76
Soak Temp :
Loss Temp :
Loss Temp :
0.00
0.00

0.00



3
0
0
0
1
0
8
1
91.8
92.6
85.2







.225
.766
.595
.174
.621
.069
.706
.886
(F)
(F)
(F)







                                                       b-l i

-------
Presaijf q-Fail Vehicles

OEmiaaion factors are as of July 1st
OCal. Year: 2020        I/M Program:
                  Anti-tarn. Program:
                   Reformulated Gaa:
                  User-Supplied Pass,
of the indicated calendar year.
Yea       Ambient Temp: 90.5 / 90.5  /  90.
Yea     Operating Mode: 20.6 / 27.3  /  20.
Yea         ASTM Class: C
 Purge Fail, and Preaaure Fail Rates (%):
5 (F)  Region: Low
6   Altitude:  500. Ft.
0 Veh. Type:
•f
Veh . Speeda :
VMT Mix:
LDGV
19.6
0.575
LDGT1
19.6
0.207
LOGT2
19.6
0.089
LOGT


HDGV LDDV
19
0
.6 19.6
.034 0.002
LOOT HOOV
MC
All
Veh
19.6 19.6 19.6
0.005 0.084 0.004
OCompo»ite Emission Faotora (Gm/Mile)
VOC HC:
Exhaust HC :
Evaporat HC:
Refuel L HC:
Run ing L HC:
Rating L HC:
Exhauat CO:
Exhaust NOX:
3.44
0.57
0.84
0.17
1.79
0.07
7.73
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
Sam- 11am
10am- 3pm
8 am- 2pm
in Grama

0.62
0.49
0.49
0.00
0.29
0.34
0.37
3.61
0.61
0.91
0.22
1.79
0.07
8.35
1.40
Component
per Mile)

0.61
0.59
0.59
0.00
0.36
0.42
0.45
3.70
0.79
0.83
0.23
1.79
0.07
10.15
2.00
3
0
0
0
1
0
8
1
.64
.67
.89
.22
.79
.07
.89
.58
6
1
2
0
1
0
17
3
.55 0.52
.63 0.52
.69
.36
.79
.08
.27 1.44
.77 1.09
Weathered RVP: 7.1


0.56
0.54
0.54
0.00
0.33
0.39
0.41


0
0
0
0
0
0
0


.60
.58
.58
.00
.35
.41
.44


1
1
1
0





.81
.32
.32
.01



0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting







2.26
1.80
0.00


0.46
22.45
0.76
Soak Temp:
Loss Temp :
Loss Temp :
0.00
0.00

0.00



3
0
0
0
1
0
8
1
91.8
92.6
85.2







.466
.766
.837
.174
.621
.069
.706
.886
(F)
(F)
(F)







                                                       B-14

-------
Composite

OEmission factors are as of July lat of the indicated calendar year.
OCal. Year: 2020        I/M Program: Yea       Ambient Temp: 90.5 /
                  Anti-tarn. Program: Yea
                   Reformulated Gaa: Yea
Operating Mode: 20.6
    ASTM Class: C
90.5 / 90.5 (F) Region: Low
27.3 / 20.6   Altitude:  500. Ft.
0 Veh. Type: LDGV
-t- 	
Veh. Speeds: 19.6
VMT Mix: O.S75
OCompoaite Emission
VOC HC :
Exhaust HC:
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
LDGT1
19.6
0.207
Factor* (Gm/Mile)
1.18 1.22
0.57
0.12
0.17
0.26
0.07
7.73
1.27
OEvaporative Emissions by
(All Component a in Grama
Hot Soak 0.07
Diurnal
Multiple
Crankcaae
Bam-llam
10 am- 3pm
8am-2pm
0.06
0.24
0.00
0.01
0.01
0.01
0.61
0.13
0.22
0.18
0.07
8.35
1.40
Component
per Mile)
0.07
0.06
0.27
0.00
0.01
0.01
0.02
LOGT2
19
0
1
0
0
0
0
0
10
2
0
0
0
0
0
0
0
.6
.069
.39
.79
.12
.23
.19
.07
.15
.00
.07
.06
.25
.00
.01
.01
.02
LOGT

1.27
0.67
0.13
0.22
0.18
0.07
8.89
1.58
0.07
0.06
0.27
0.00
0.01
0.01
0.02
HDGV LDDV
19
0
4
1
1
0
0
0
17
3
.6 19.6
.034 0.002
.24 0.52
.63 0.52
.49
.36
.68
.08
.27 1.44
.77 1.09
Weathered RVP : 7.1
0.94
0
1
0


.72
.14
.01


LDDT HDDV
MC
Ali
Von
19.6 19.6 19.6
0.005 0.084 0.004
0.73 2.10
0.73 2.10

1.61 ' 11.18
1.24 6.58
Hot
Running
Resting




7. 14
1.80
4.88
0.46
22.45
0.76
Soak Temp :
Loss Temp:
Loss Temp:
1.89
2.99
0.00


1
0
0
0
0
0
8
1
91.8
92.6
85.2




.413
.766
. 176
.174
.228
.069
.706
.886
(F)
(F)
(F)




                                                       b-li

-------
                           POST-CONTROL INPUT VXLE  - WITH RKTORMULATU) GASOI.INB
INew Evap Teat Procedure with Enhanced I/M and Reformulated Gasoline
 MOBILES (4-Dec-92)
OI/M program |1 selected:                          I/M program |2 selected:
OStart year (Jan 1):  1983
 Pre-1981 stringency! 20%
 First MYR covered!   1968
 Last MYR covered<    2020
 Waiver  (pre-1981)i    1.%
 Waiver  (1981+):       1.%
 Compliance Rate:     98.%
 Inspection type:
 Test Only
 Inspection frequency: Annual
 I/M program fl vehicle types
      LOGV - Yea
     LDGT1 - Yes
     LDGT2 - Yes
      HDGV - Ho
 1981 4 later MYR test type:
      Idle
 Cutpointa,  HC:   220.000
 Outpoints,  CO:     1.200
 Outpoints,  NOx:  999.000
                                                   Start year (Jan 1):  1990
                                                   Pre-1981 stringency: 20%
                                                   First MYR covered:   1984
                                                   Laat MYR covered:    2020
                                                   Waiver (pre-1981):    1.%
                                                   Waiver (1981+):       l.t
                                                   Compliance Rate:     98.%
                                                   Inspection type:
                                                   Test Only
                                                   Inspection frequency: Annual
                                                   I/M program 12  vehicle types
                                                        LDGV - Yes
                                                       LDGT1 - Yes
                                                       LOGT2 - Yes
                                                        HDGV - No
                                                   1981 4 later MYR teat type:
                                                        IM240 test
                                                   Cutpointa,  HC:      0.800
                                                   Cutpoints,  CO:     15.000
                                                   Cutpoints,  NOz:  999.000
OFunctional Check Program Description:
OCheck Start   Model Yrs  Vehicle Classes Covered
       (Janl)   Covered    LDGV   LDGT1   LDGT2  HDGV
                                                         Inspection
                                                       Type      Freq
 Press 1990    1971-2020  Yes    Yes
 Purge 1990    1984-2020  Yes    Yes
 ATP   1983    1975-2020  Yes    Yes
OAir pump system disablements:
                                        Yes    No  Teat  Only     Annual
                                        Yes    No  Test  Only     Annual
                                        Yes    No  Test  Only     Annual
                                     Yes  Catalyst  removals:
 Fuel inlet restrictor disablements:  Ye*  Tailpipe lead deposit teat:
 EGR disablement:                    No   Evaporative system disablements:
 PCV system disablements:             Yes  Missing  gas caps:
0	                               Minimum Temp:   72.  (F)
                       Period 1 RVP:  11.5      Period 2  RVP:   9.0
OVOC HC emission factors include evaporative  HC emission factors.
0
Comp
Rate

98.0*
98.0%
98.0%
   Yes
   Yes
   Yes
   Yes
   Maximum Temp:
                                                                                             96. (F)
                                                                         Period 2 Start Yr: 1992
                                                       B-16

-------
                       POST-CONTROL OUTPUT FXLKS - WITH RKTOBMDLATXD GASOLINB
       Vehicles
OEmiasion factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020
0 Veh. Type:
Veh. Speeds:
VMT Mix:
I/M Program: Yea
Anti-taat. Program: Yea
Reformulated Gas: Yea
Uaer-Supplied Pass,
LDGV LDGT1
19.6
0.575
0.207
OCompoaite Emission Factor* (G»/Mile)
VOC HC: 0.84 0.93
Exhaust HC :
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
0.57
0.04
0.17
0.04
0.02
7.73
1.27
OEvaporativa Emissions by
(All Components in Grama
Hot Soak 0.03
Diurnal
Multiple
Crankcase
8 am- 11 am
10 am- 3pm
8a»-2pm
0.02
0.08
0.00
0.00
0.00
0.00
0.61
0.05
0.22
0.03
0.02
8.35
1.40
Component
per Mile)
0.03
0.02
0.10
0.00
0.00
0.00
0.00
Ambient Temp:
Operating Mode:
ASTM Class:
90.5 / 90.5 / 90.5 (F) Region: Low
20.6 / 27.3 / 20.6 Altitude: 500. Ft.
C
Purge Fail, and Pressure Fail Rates (*) : 100., 0., 0
LDGT2 LDGT HDGV LODV LOOT HODV
19
0
1
0
0
0
0
0
10
2
0
0
0
0
0
0
0
.6
.089
.10
.79
.04
.23
.03
.02
.15
.00
.03
.02
.09
.00
.00
.00
.00

0.98
0.67
0.05
0.22
0.03
0.02
8.89
1.58
0.03
0.02
0.09
0.00
0.00
0.00
0.00
19
0
2
1
0
0
0
0
17
3
.6 19.6
.034 0.002
.57 0.52
.63 0.52
.51
.36
.05
.03
.27 1.44
.77 1.09
Weathered RVP: 7.1
0.30
0
0
0


.29
.30
.01


19.6 19.6
0.005 0.084
0.73 2.10
0.73 2.10

1.61 11.18
1.24 6.58
Hot
Running
Resting




MC
19.6
0.004
7.14
1.80
4.88
0.46
22.45
0.76
Soak Temp :
Loss Temp :
Loss Temp :
1.89
2.99
0.00


All

1
0
0
0
0
0
8
1
91.8
92.6
85.2




Veh

.071
.766
.077
.174
.035
.020
.706
.886
(F)
(F)




                                                   b-17

-------
Pure-Fail Vehiclaa
OEmission factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020 I/M Program: Yes Ambient Temp: 90.5 / 90.5 / 90
Anti-tarn. Program: Yes Operating Mode: 20.6 / 27.3 / 20
Reformulated Gas: Yes ASTM Class: C
User-Supplied Pass, Purge Fail, and Pressure Fail Rates (%)
0 Veh. Type: LDGV LDGT1 LDGT2 LDGT HDGV LDDV
Veh. Speeds:
VMT Mix:
19.6
0.575
19.6
0.207
OComposite Emission Factors (Ga/Mile)
VOC HC:
Exhaust HC:
Evaporat HC :
Refuel L HC:
Runing L HC:
Rsting L HC:
Exhaust CO:
Exhaust NOX:
2.52
0.57
0.47
0.17
1.30
0.02
7.73
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8 am- 11 am
10am- 3pm
8am-2pm
in Grams

0.45
0.24
0.24
0.00
0.01
0.05
0.08
2.66
0.61
0.50
0.22
1.30
0.02
8.35
1.40
Component
per Mile)

0.45
0.30
0.30
0.00
0.02
0.08
0.12
19
0
1
2
0
0
0
1
0
10
2
.6
.089

.79
.79
.45
.23
.30
.02
.15
.00


2.70
0.67
0.49
0.22
1.30
0.02
8.89
1.58
19.
0.

5.
1.
1.
0.
1.
0.
17.
3.
6 19.6
034 0.002

20 0.52
63 0.52
87
36
32
03
27 1.44
77 1.09
Weathered RVP: 7.1


0
0
0
0
0
0
0


.41
.28
.28
.00
.02
.07
.11


0.44
0.30
0.30
0.00
0.02
0.08
0.12


1.
0.
0.
0.





32
81
81
01



.5 (F) Region: Low
.6 Altitude: 500. Ft.
: 0., 100., 0.
LDDT HDDV MC
All
Veh
19.6 19.6 19.6
0.005 0.084 0.004

0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting








2.26
1.80
0.00


0.46
22.45
0.76
Soak Temp:
Loss Temp :
Loaa Temp :
0.00
0.00

0.00




2
0
0
0
1
0
8
1
91.8
92.6
85.2








.613
.766
.477
.174
.177
.020
.706
.886
(F)
(F)
(F)







                                                       B-ia

-------
OEmission factors are as of July 1st of the indicated calendar year.
OCal. Year: 2020 I/M Program: Yes Ambient Temp: 90.5 / 90.5 / 90.5 (F) Region: Low
Anti-tarn. Program: Yea Operating Mode: 20.6 / 27.3 / 20.6 Altitude: 500. Ft.
Reformulated Gas: Yes ASTM Class: C
User-Supplied Pass, Purge Fail, and Pressure Fail Rates (%) : 0., 0., 100.
0 Veh . Type :
*
Veh . Speeds :
VMT Mix:
LDGV
19.6
0.57S
LDGT1
1976
0.207
LDGT2
19.6
0.089
LOOT




HDGV LDOV
19
0
.6 19.6
.034 0.002
LDDT HDDV
19.6 19.6
0.005 0.084
MC
19.6
0.004
All


Veh


OComposite Emission Factor* (Gm/Mile)
VOC HC:
Exhaust HC :
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating L HC:
Exhaust CO :
Exhaust MOX:
2.77
0.57
0.72
0.17
1.30
0.02
7.73
1.27
OEvaporative Emissions by
(All Components

Hot Soak
Diurnal
Multiple
Crankcase
8am-llam
10am- 3pm
8am-2pm
in Grama

0.46
0.49
0.49
0.00
0.29
0.34
0.37
2.95
0.61
0.79
0.22
1.30
0.02
8.35
1.40
Component
per Mile)

0.45
0.59
0.59
0.00
0.36
0.42
0.45
3.06
0.79
0.72
0.23
1.30
0.02
10.15
2.00
2
0
0
0
1
0
8
1
.98
.67
.77
.22
.30
.02
.89
.58
5
1
2
0
1
0
17
3
.55 0.52
.63 0.52
.22
.36
.32
.03
.27 1.44
.77 1.09
Weathered RVP: 7.1


0.41
0.54
0.54
0.00
0.33
0.39
0.41


0
0
0
0
0
0
0


.44
.58
.58
.00
.35
.41
.44


1
1
1
0





.34
.32
.32
.01



0.73 2.10
0.73 2.10




1.61 11.18
1.24 6.58
Hot
Running
Resting







2.26
1 .80
0.00


0.46
22.45
0.76
Soak Temp :
Loss Temp:
Loss Temp:
0.00
0.00

0.00



2
0
0
0
1
0
8
1
91.8
92.6
85.2







.853
.766
.717
.174
.177
.020
.706
.886
(F)
(F)
(F)







B-19

-------
OEmiaaion factora are aa of July 1st of the
OCal. Year: 2020 I/M Program: Yea
Anti-tarn. Program: Yea
Reformulated Caa: Yea
0 Veb. Type: LDGV
-f
Veh. Speeda; 1578
VMT Mix: 0.97S
LDGT1
TO"
0.207
OCompoaite Emiaaion raotoxa < Cm/Mile)
VOC HCs 0.99 0.99
Cxhauat HCs 0.57
Cvaporat HCs 0.06
Refuel L HCs 0.17
Runing L HCs 0.08
Rating L HCs 0.02
Exhaust COs 7.73
Cxhauat NOX: 1.27
OCvaporative Emiaaiona by
(All Componenta in Grama
Hot Soak 0.04
Diurnal 0.03
Multiple 0.09
Crankcaae 0.00
8am- 11 am 0.01
10am- 3pm 0.01
8am-2pm 0.01
0.61
0.07
0.22
0.07
0.02
8.35
1.40
Component
per Mil*)
0.04
0.04
0.11
0.00
0.01
0.01
0.01
LDGT2
19.6
0.089
1.17
0.79
0.06
0.23
0.07
0.02
10.15
2.00

0.04
0.04
0.10
0.00
0.01
0.01
0.01
indicated calendar year.
   Ambient Temp:  90.5
 Operating Mode:  20.6
     ASTM Claaa:  C
     LOOT
       04
       67
       07
       22
     0.07
     0.02
     8.89
     1.58
     0.04
     0.04
     0.10
     0.00
     0.01
     0.01
     0.01
   HDGV

  1976
   0.034

   3.43
   1.63
   0.99
    .36
    .43
            90.5 /
            27.3 /

             LDDV
                90.5  (F) Region: Low
                20.6   Altitude:  500. Ft.
0.
0.
   0
  17
  .03
  ,27
   3.77
Weathered
   0.61
   0.53
   0.63
   0.01
         19.
          0.

          0.
          0.
          1.
          1.
       RVP:
.6
.002

.52
.52
 44
 09
  7.1
 LDDT

T976
 0.005

 0.73
 0.73
 1.61
 1.24
                                             HDDV

                                            1976
                                             0.084
     10
     10
 MC

19.6
 0.004

 7.14
 1 .80
 4.88
                                               All  Veh
             0.46
  11.18     22.45
   6.58      0.76
    Hot Soak Temp:
Running Loaa Temp:
Reating Loaa Temp:
             1 .89
             2.99

             0.00
   1 . 150
   0.766
   0. 109
   0. 174
   0.081
   0.020
   8.706
   1.886
91.8 (F)
92.6 (F)
85.2 (F)
           U- J

-------