EPA-420-R-74-107
Mobile Source Evaporative Emissions
by
C. Do'n Paulsell
June 1974
Environmental Protection Agency
Office of Air and Waste Management Programs
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Regulations Development and Support Branch
Ann Arbor, Michigan
-------
Executive Summary of Evaporative Emissions Document
The paragraphs below summarize the important points contained
in each of the three parts of the attached document on mobile source
evaporative emissions.
Position Paper: Evaporative emissions have been studied and measured
since 1958. Federal regulations became effective in 1971 to reduce
evaporative emissions.
The Federal canister test procedure which has been used with
those regulations indicates that vehicles are emitting about 0.13
grams/mile from fuel evaporative processes. Tests have shown that
this procedure does not adequately cover all potential sources and
does not accurately measure the true magnitude of all evaporative
emissions. An improved procedure (SAE J171) involves placing the
vehicle inside an enclosure where all emissions can be measured on
a mass basis by monitoring the concentration of hydrocarbons in a
known volume. The SAE J171 test procedure indicates that 1972 model
vehicles (controlled) are actually emitting about 1.87 grams/mile.
If evaporative regulations are notfcchanged prior to implementation
of the statutory exhaust emission hydrocarbon standard of 0.41 grams/
mile, evaporative emissions will constitute the dominant hydrocarbon
emission. The total hydrocarbon emissions would then exceed the
established reductions necessary to meet the air quality standards
for oxidant levels.
Adoption of the SAE 0171 enclosure procedure will require that
necessary changes be made in evaporative control systems to reduce
evaporative emissions. However, to simply replace the canister test
with the enclosure test would be very costly. Data indicate this
implementation plan is unnecessary. By developing selection criteria
for evaporative system families, the number of required tests could be
reduced to less than 1000 per year, and the enclosure procedure could
be implemented in a cost effective manner.
A program plan has been outlined to achieve the implementation of
the enclosure procedure by December, 1975 (Start of 1977 MY testing).
Program Plan: A data base of 202 enclosure tests exists for a vehicle
population covering 1957-72 model year vehicles. The program plan
provides for supplementing that data with tests that specifically look
for the cause of the emissions. A contract will be negotiated to
assess evaporative sources and to develop the cost/effectiveness
analysis for various evaporative control approaches.
Communications will be established with the automobile industry to
maximize their lead time for hardware development and to solicit their
comments and data on our approach to revised regulations.
-------
-2-
This program will require the assignment of high priority in
order to accomplish the critical deadlines scheduled in FY75. Costs,
manpower, and tasks have been itemized and summarized.
Technical Appendix: The various specifications used in the enclosure
procedure are discussed. Simulation of real world conditions related
to time, temperature, fuel volumes, and types of fuel are documented.
The four phases of the evaporative process, (diurnal losses,
running losses, hot soak losses, and refueling losses) have been
discussed. The preferred measurement and recommended calculation
procedures are covered.
A complete list of references on the subject of mobile source
evaporative emissions has been compiled..
-------
TABLE OF CONTENTS
A. A Position Paper on Evaporative Emissions
Page
I. Introduction
II. Background on Evaporative Emissions Regulations
III. Evaporative Emissions Test Procedures
9
A. Canister Test Procedure
B. Enclosure Test Procedure
IV. Implementation Alternatives
V. Problems of Implementation
VI. Conclusions
VII. Recommendations
10
14
15
16
16
17-19
VIII. Closure
IX. References
B. A Program Plan for Evaporative Emissions Regulation Development
I. Title 1
II. Responsibility 1
III. Problem Assessment 1
IV. Purpose 1
V. Objectives 1
VI. Approach/Scope of Work 1
VII. Milestones/Accomplishments 1
VIII. Current Status 2
IX. Technical Support 2
X". Coordination and Manpower 2
XI. Program Description and Timetable 2
C. A Technical Appendix on the Measurement of Evaporative Emissions
I. Introduction 1
II. Summary Statement and Background 1
III. Diurnal Emissions 2
IV. Running Losses 8
V. Hot Soak Emissions 9
VI. Measurements and Calculations 10
VII. Closure 11
VIII. Charts and Graphs 12-19
IX. References 20-22
-------
A POSITION PAPER ON EVAPORATIVE EMISSIONS
I. Introduction: This document summarizes the background and the
present state of knowledge regarding evaporative emissions from light
duty motor vehicles. It discusses the effectiveness of our present
evaporative emissions regulations and explores the revision of those
regulations to facilitate more accurate measurement, assessment, and
control of evaporative emissions.
The purpose of this document is: to present the facts that define
the problem, to discuss the test techniques available as regulatory
procedures, to weigh the pros and cons of actions which may effect a
solution to the problem, and to outline the conclusions and recommenda-
tions that will lead to the revision of those regulations.
II. Background on Evaporative Emissions Regulations: Section 202(a)
of the Clean Air Act requires the Administrator to prescribe regula-
tions applicable to motor vehicle emissions which contribute to air
pollution and endanger public health or welfare.
Hydrocarbons emitted to the atmosphere from motor vehicles have
been judged as a pollutant requiring regulation. Hydrocarbons are
generated from three processes: gaseous,, products from combustion,
engine crankcase blowby, and valors from fuel evaporation. These
emissions contribute to the formation of smog. References (4,19,20,23)
at the end of this paper discuss this process and the interactions of
the components. Basically, the conclusions have been that it is neces-
sary to reduce both hydrocarbons and oxides of nitrogen to reduce the
oxidant levels produced by the photochemical reactions; furthermore,
a reduction in hydrocarbons has been shown more effective than an
equal reduction in oxides of nitrogen. This fact underscores.the
importance of hydrocarbon regulation.
The preceding paragraph illustrates the importance of controlling
al1 hydrocarbon emissions, but this document has been written to assess
one source of hydrocarbons, evaporative emissions. Studies and measure-
ments on evaporative losses have been documented since 1958. One of
the first studies (1 ) which deals with carburetor losses v/as prompted
by interest in idling performance, fuel economy, and air pollution.
Throughout the mid-sixties, many studies were conducted to quantify
the magnitude, mechanism, and sources of fuel vapor emissions. At the
time of this research vehicles were uncontrolled; fuel tanks had vented
caps, carburetors had externally vented float bowls, and some engines
had vented crankcases. Basically, the emissions have been differentiated
into three categories: diurnal breathing losses, running losses, and
hot soak losses. Diurnal breathing, losses are caused by the daily
temperature rise and resultant expulsion of vapors from the fuel tank
vent. Running losses are similar to the diurnal losses, except that the
temperature rise occurs during vehicle operation from engine waste heat
flowing over the fuel tank and carburetor. Hot soak losses are caused
-------
-2-
by boiling in the carburetor bowl after engine shutdown. A comprehen-
sive discussion of the mechanisms and parameters related to these
emissions has been written as a technical appendix to this paper.
Several test techniques for measuring these specific source loca-
tions were developed (16) and used to assess the amount, composition,
and influencing factors on the fuel evaporative emissions. The tech-
nique most widely used was vapor condensation in a cold trap immersed
in a slurry of dry ice, alcohol, and acetone to maintain a temperature
of -80°F.
One AMA (Automobile Manufacturer's Association) program (15)
extensively tested 5 vehicles and reported carburetor losses as high
as 50 grams/test. Vehicle design was the largest single factor affect-
ing carburetor losses. The program also reported tank emissions as
high as 125 grams/day, with 69% of the losses occurring during vehicle
operation (running loss). Fuel volatility and maximum temperature had
the major impact on these tank emissions.
After the influence of fuel volatility and temperature had been
assessed, a test procedure was developed that standardized the fuel
and fuel temperature variations,. Additional data were collected using
this procedure and interim standards were proposed by the California
Motor Vehicle Pollution Control Board in November, 1966. Their stan-
dards v/ere based on an 80% reduction of baseline emissions from un-
controlled vehicles. The baseline emissions were consolidated from
several references (25,2Ł,21) and v/sre set at average values of 30 grams
per day from the diurnal and 10 grams per test from the hot soak.
Hence, California standards were 6 grams and 2 grams for the respec-
tive diurnal and hot soak tests.
One system to control evaporative emissions had been in the develop-
ment phase since 1960. It was the forerunner of the present control
system and basically incorporated a canister of activated charcoal that
was plumbed to the tank and carburetor vents and operated on a controlled
adsorption-desorption principle. A report (8 ) on the effectiveness of
this system concluded that vehicles controlled by this system achieved
90-100% reduction in emissions. This reduction has been shown dependent
on the test procedure used to measure the emissions,
The California standards were never finalized as they had been
proposed. Three months later, February 4, 1967, the Federal Government
(HEW) issued a notice of proposed standards for evaporative emissions
(27). The standard proposed, 2 grams per test, was a composite sample
representing the total diurnal and hot soak emissions. The test pro- _
cedure in this proposal involved running an exhaust emissions test on
a dynamometer, fueling the vehicle with cold fuel, moving the vehicle
into a sealable enclosure, and monitoring the hydrocarbon concentration
within the enclosure to determine the mass of fuel vapors emitted, This
proposal was technically correct and realistically practical because it
-------
-3-
required no modification of the vehicle and measured the total diurnal
and hot soak emissions from aj_]_ sources, Unfortunately, the proposed
concept had not been substantiated by quantitative data, was not sup-
ported by the automotive industry, and was consequently replaced by
an alternate technique. This technique involved trapping the vapors
from specific sources by attaching charcoal canisters; this was basi-
cally the proposal California had made. A test procedure employing
the vapor trapping technique was finalized and published as a federal
regulation on June 4, 1968. The evaporative emission standard of 6
grams per test was applicable to 1971 model year vehicles sold nation-
wide; the MVPCB of California adopted the procedure and standard to
apply to 1970 model year vehicles sold in California.
The 1971 HEW certification results showed that 107 of 131 (82%)
vehicles tested emitted less than 1.0 gram. Consequently, the standard
was lowered to 2 grams per test for 1972 vehicles and evaporative
emissions regulations have not changed since.
This brief summary on the regulation of evaporative emissions has
hopefully provided the reader with the necessary background to place
the remainder of this paper in the proper perspective.
The enclosure procedure th^t had bŁen proposed in 1967 by HEW was
technically correct and had many merits, as mentioned before. A few
days after that proposal was made, one company of the automotive industry
built an enclosure and conducted extensive tests to study the potential
problem areas and to assess the value of the proposal. Two published
reports (5, 6 ) on these evaluations concluded that the technique was
accurate, simple, and repeatable, and represented "a superior technique
and versatile tool" for evaporative emissions measurements. Subsequent
tests by HEW engineers and various other organizations substantiated
those initial results and lead to refinements which were published by
the Society of Automotive Engineers (12) in a formal recommended test
procedure. This procedure, SAE J171 a, has the same specifications as
the presently used canister procedure except that the enclosure-FID
setup replaced the gravimetric determination as the measurement method.
During the last three years of certification testing, the official
results shown in Table A were obtained from the canister evaporative
emissions test.
-------
-4-
Table A
1971-1974 Certification Results
Grams/Test
No. % Tests % Tests Max. Avg.
MY Tests < 0.1gm < 1.Oqm Value Value Std.
71
72
73
74
131
370
351
399
32%
45
56
45
82%
91
94
98
3.65
1.90
1.90
1.90
.545
.307
.251
.258
6
2
2
2
Note: 60% of tests < .1 gm were actually 0.0 gms.
The emissions shown in Table A seem to indicate that the evaporative
control systems are very effective. However, for this analysis, all the
data gathered should be considered. The data collected during the 1974
model year program was readily accessible.
Actually, there were a total of 114*1 evaporative emissions tests
performed in that program. Durability vehicles accounted for the 399
published results but the emission data vehicles tested totaled 742,
There were 20 failures in the emission data population. These failures
occurred on .12 vehicles. Most failures were caused by test procedure
errors; the vehicles passed subsequent retests. Three.of the vehicles
required hardware changes to pass. There were no failures on any dura-
bility vehicles and the deterioration factors for evaporative emissions
were less than .2 (additive D.F.) for 93% of the vehicle families.
Another observation can be made concerning the tests listed as 0.0
emissions. Roughly 60% of those values less than 0.1 grams were actually
0.0. Investigation disclosed that values listed as 0.0 were almost
always negative canister weight changes. This aspect of the canister
procedure is the most indicative evidence of the canister technique's
questionable validity.
As a followup to the certification process, the mobile source program
conducts surveillance testing to assess the net effect of the regulatory
process. Vehicles are procured from private owners and emissions tests
are performed to quantify the emission levels generated from production
vehicles that have been maintained by the consumer.
-------
-5-
The surveillance program in 1972 measured evaporative emissions from
both controlled and uncontrolled vehicles. The questionable results
which had been obtained from the canister procedure prompted the sur-
veillance personnel to consider using the enclosure technique which was
more simply applied to both controlled and uncontrolled vehicles, The
enclosure technique accounts for emissions from areas that are diffi-
cult to "trap", i.e. gaskets and throttle shafts, whereas the canister
procedure requires many connections and lengthy preparation for uncon-
trolled vehicles. Consequently, in order to provide a simple and common
basis for comparison of data, the surveillance program adopted the
enclosure procedure, SAE J171. The measurements made in Los Angeles
encompassed a distribution of 136 California vehicles from model years
1957 to 1971. In addition, twenty-two (22) 1971 model vehicles were
tested in Denver to assess the effect of high altitude. The next year,
the surveillance program concentrated on 1972 model vehicles only, test-
ing twenty-two (22) each at Los Angeles and Denver. The results of all
these tests are summarized in Table B.
Table B
Surveillance Results - Encjosure Procedure
—- — / "
L.A. Data (qms)
Denver Data (qms)
Weighted
Values**
Model
No.
Year.
Tests
Diurnal
Hot Soak
Diurnal Hot Soak
qms/nile
'57-'
69 102
26.08
14.67
-
2,72
'70
13
17.75
10.70
-
1.95
'71
21
14.87
10.89
-
1,88
'71
22
-
-
47.2* 34.8*
6.02*
'72
22
12.40
11.80
-
1.87
'72
22
-
-
17.4 14.2
2.40
Note:
*Winter grade fuel
(11.7 RVP)
used on all tests,
L.A. Data up
to 1971 used all types of fuel (7.8 - 12.0 RVP).
**Emission = Diurnal + 4.7(Hot Soak)/35 miles/day.
Before the comparison of Table A and Table B is made, a few qualifying
remarks are necessary. The first contrast one observes in Table B is the
disparity between the 1971 and 1972 Denver values. This large difference
has been explained by the fact that the 1971 vehicles were tested using
commercial winter grade fuel of high volatility (Reid Vapor Pressure (RVP)
= 11.7 psi), and the 1972 vehicles used the standard test fuel of summer
grade volatility, (RVP = 8.8 psi). Several reports (1,2,3,7,9,11) have quanti-
fied that the emissions from fuel having an RVP of 11.7 will be double to
triple those from fuel with an RVP of 8.8. This correction factor will
-------
-6-
tend to faring the Denver results for the two model years (both controlled
vehicles) into closer agreement. The same situation occurred during the
tests in Los Angeles, but the effect on the overall average is less pro-
nounced because a wide variety of vehicles was tested at random times with
both types of fuel. However, the standard test fuei was used during the
Los Angeles tests on 1972 model vehicles and a valid comparison to the
Denver data can be made. The consistent difference in this comparison
can be attributed to the effect on the evaporative process of the lower
barometric pressure at Denver (24.5"Hg) as compared to that at L.A.
{30.
A comparison ef the values for the diurnal and fiat soak emissions
from controlled ar.ci uncontrolled vehicles does not 'indicate a significant
improvement. Diurnal emission control shows the highest reduction (.547^1,
but the hot soak control techniques; produced a much snailer reduction
(27SJ. It is reiterated here tftat trie c&rractiOT? factor will
lower the uncontrolled vehicle emissions, a change which demonstrates
even lower reductions than those indicated.
The very perplexing aspect of the diurnal data has been the magnitude
of the emission from a supposedly closed system. Diurnal tests have been
coid^ctsd on leak-tight coTtrolJsd vehicles at tfie PISAPC laboratory
usirp the enclosure technique. The results confirmed that a control
system can limit diurnal emissions to less than \ gram/"test. Ore must
conclude that a leak exists in the control system-, speculation as to the
sources would point to the canister, lines, tank fittings, or tank cap.
Investigation of tfifs latter speculation disclosed that 1Q-20S of the
pressure "leak tests performed during the 1974 certification program
failed because of leaking caps. Generally, the manufacturers were
allowed to replace the cap arid the evaporative test proceeded. The
surveillance program did net require a leak test because the purpose
of the test was to measure "as received1' emissions. However, testing
experience indicates a cap leak would be the likely cause of these high
diurnal emissions. The data indicate that hardvjare improvements are
needed in this area.
The control of hot soak losses relies primarily on plumbing the
c&rburetor bowl snd air cleaner to the control canister. In most sys-
tems the internal carburetor vents continue to fill the air cleaner witfi
vapors, and the diffusion-expulsion process follows the path of least
resistance, the air cleaner inlet. Tn the canister test procedure, these
vapors must make thejr my through small tubing before they can be ad-
sorbed in the charcoal canister. The data indicate this is not happening.
A test was reported (5 ) that investigated this problem inside an .
enclosure. TEie author concluded that the air cleener merely stores the
vapors during a canister test when its inlet is plugged. However, sub-
sequent removal of the plug produced a sudden increase in the enclosure
tydrocarbon concentration. This is an example of how the trod ifications
required by the canister procedure actually fnfn'6it tfte emfssfc/i of
hydrocarbons.
-------
-7-
The comparison of values from Table A and Table B indicates that the
canister test procedure does not accurately assess evaporative emissions.
There are several reasons to conclude that the enclosure procedure more
accurately assesses the true magnitude..
During the surveillance program, replicate tests were performed on
14 of the 188 vehicles tested. After considering the variations caused
by the fuel volatility problems cited earlier, the repeatability of the
enclosure procedure is on the order of + 1 OX"for all replicates. Simi-
lar repetitive testing using the canister procedure has produced stan-
dard deviations as high as 200% of the mean for a group of tests.
Another reason to accept the enclosure values is that the enclosure
can be calibrated. Calibration tests are typically repeatable within +2%
for all tests. Enclosure background emissions (if any exist) can be
measured and subtracted to yield net vehicle emissions. Finally, the
enclosure measures all vehicle emissions without modification to the
control system.
The discussion to this point has attempted to present the background
and to illustrate the quantitative data^available for evaluation. If
the most valid block of data, the L.A. and Denver tests on 1972 vehicles,
is representative of true emissions, the impact of evaporative emissions
on air quality can be estimated. The weighting factors that have been
used for exhaust emissions will be applied to evaporative emissions.
The model assumes that a vehicle experiences 4,7 starts per day and
accumulates 35 miles in average daily operation. In the analysis of
total evaporative emissions, the assumption of one diurnal and 4.7 hot
soaks per day is made. When the average values from the 1972 vehicles
are used in the model for total emissions, the final value is:
(Diurnal qms/test x test/day)+(Hot soak gms/test x tests/day)
gms/mile Miles/day ~
Evap. gms/mi le = .02-4 x 1 -3 x 4-7)
35
65.5 gms/day
35
rSSSw™ - 1.87 gms/mile
Emissions 3 '
These 1972 vehicles, by similar analysis using the 2 gram standard
test results, were certified at evaporative emission levels less than
0.13 gms/mile. To place this evaporative emission in perspective, the-
tailpipe tmissions have also been considered.
-------
-8-
The certification HC standard in California for 1972 vehicles was
3.2 grams/mile as opposed to the national standard of 3.4 grams/mile.
The surveillance data that was measured on 35-1972 vehicles showed an
average HC emission of 4.07 grams/mile.
One additional source of HC vapor should be discussed to complete
this analysis of total emissions. Two reports (10,24) have been
written on the fuel vapors emitted during fueling operations. For the
purpose of this analysis, the rates of 5.0 gms/gal. and 13.4 miles/gal.
will be used as the average values for fuel transfer emissions and fuel
consumption. These values translate to approximately 0.4 grams/mile.
If the emission discussed represent the total mobile source contri-
bution to HC pollution, the following table shows the relative contribu-
tion of each specific source for in-use 1972 vehicles.
Table C
Total HC Emissions From 1972 Vehicles
Exhaust
/qms/mile
4.07
64.2%
Evaporative
1.87
29.5%
Refueling
.40
6.3%
Total HC
6.34
100.0%
This conservative analysis has shown that the- minimum impact on HC
emissions due to evaporative losses is presently about 30% from vehicles
with evaporative control. The analysis can be developed one step further
to show the impact of evaporative emissions relative to 1975 model year
vehicles. Evaporative emissions or refueling losses are not expected to
change for 1975 or 1976 vehicles because the evaporative and fueling control
systems will be essentially the same as those used in 1972. Assuming the
1972 (in-use/cert. std.) factor (4.07/3.2 = 1.27) will be the same in 1975
and 1976, the in-use tailpipe emissions from 1975 and 1976 vehicles can be
estimated and compared as shown in Table D.
Table D
Predicted HC Emissions From 1975/76 Vehicles
HC Standard
1975 (1.
gms/mile
5 gms'/mile)
1976 (.41 gms/mile)
gms/mile
Exhaust
1.91
45.7%
.52
18.5%
Evaporative
1.87
44.7%
1.87
67.0%
Refueling
.40
9.5%
.40
14.4%
Total HC
4.18
100.0%
2.79
100.0%
-------
-9-
This analysis illustrates that evaporative emissions will represent
the primary source of hydrocarbons from mobile sources if the present
regulatory procedures and standards remain unchanged.
The cause of this problem is simply the inability of the evaporative
emissions test procedure to measure the true total emissions from the
test vehicle. The presence of this weakness has caused an erroneous
assessment of the true magnitude of the emissions and has resulted in
few improvements in control technology. The two effects are interdependent.
One may be tempted to jump at the obvious solution to this problem,
which is to change the test procedure. While this is certainly an im-
portant aspect of the solution, it alone will not automatically result
in the control of emissions.
III. Evaporative Emissions Test Procedures: The two principle techniques
which have been used for measuring evaporative emissions, the canister
and enclosure procedures, have been discussed in general terms previously
in this paper. This section will discuss the advantages, limitations,
costs, and problems associated with each procedure.
A. Canister Evaporative Test Procedure
Advantages:
1. Test equipment is simple and relatively low in cost.
2. Required test area is small and test setup is flexible.
Limitations:
1. Inability to measure all losses.
2. Poor repeatability of test data.
3. Use of tubing and connectors may actually inhibit
real emissions.
4. Low percent of total emissions collected.
5. Time consuming and tedious procedure to perform.
Costs:
1. Lengthy preparation phase results in high manpower costs.
(Preconditioning study indicates preparation could be
shortened)
-------
-10-
Problems:
1. High void test rate caused by numerous and complex steps.
2. Accurate calibration of test procedure is not possible.
B. Enclosure Evaporative Test Procedure
Advantages:
1. Accurate measurement of all emission sources.
2. Minimum vehicle modifications and preparation.
3. Good repeatability (+10%) on replicate tests.
4. Precise calibration and verification of equipment is possible.
5. Familiar instrumentation and calculation routines.
Limitations:
1. Area required to house ar^ enclosure is large.
Costs:
1. Initial investment costs are very high (est. $25K per
enclosure)
Problems:
1. Vehicle background (tires, seals, paint, undercoat, etc.)
emissions from new cars are high. Emissions decrease to
low levels after 90 days of ageing.
2. Abnomal high ambient temperatures during hot soak tests
must be avoided.
Other questions related to simulation of realistic environmental
parameters have been addressed in a technical appendix to this report.
IV. Implementation Alternatives: Previous position papers and program
plans on the subject of evaporative measurement by the enclosure pro-
cedure have taken the basic position that the enclosure technique re-
places the canister in a test applied to each vehicle subject to certi-
fication. The cost benefit relationship, as well as the physical space
requirements of such an implementation plan cannot be justified in light
of the data available on evaporative emissions.
-------
-11-
As discussed earlier, evaporative emissions occur from two primary
sources, the fuel tank and the carburetor. The test procedure specifies
the measurement of separate diurnal and hot soak losses. This differen-
tiation is an important requirement because of the weighting factors which
must be applied to translate test values to a grams per mile basis. Pro-
cedures that combine the diurnal and hot soak tests to reduce test time
surrender a necessary part of the data.
A reduction of test time should be attainable without sacrifice of
important data. An alternative method to reduce the manpower and test
time required to obtain the necessary data is to reduce the number of
vehicles required for testing. This proposal is more than an alterna-
tive; it is a technique that is justified by the character of the emis-
sions and the data meas.ured on many systems. For example, the factors
that govern the diurnal emission are tank vapor volume, fuel volatility,
and maximum fuel liquid temperature. These parameters have been regressed
against diurnal emissions (11) which produced a correlation coefficient
of 0.93. Similarly, carburetor hot soak losses are primarily dependent
on fuel distillation characteristics (% @ 160°F), carburetor maximum fuel
temperature, and carburetor bowl liquid volume. Hot soak emissions have
been regressed against these three variables with a correlation coefficient
of 0.89.
These characteristics can be applied to vehicle evaporative control
system configurations to determine the appropriate vehicle selection criteria
and sampling plan. This method of vehicle selection would predictably
reduce the number of vehicles to be tested for evaporative emissions. It
is then possible to assess the capability of all control system configurations
without testing every vehicle and without expenditures for numerous en-
closures and instruments. This implementation plan is justified by the
data, and has a very low cost: benefit ratio.
In 1973, the Mobile Source Laboratory tested approximately 399 durabi-
lity vehicles which represented 114 engine families and 35 manufacturers.
In addition 742 emission data vehicles were tested. If sample selection
criteria had been applied to fuel tank configurations and carburetor sizes,
the number of evaporative emissions tests would have been reduced.
For example, if a manufacturer uses the same fuel tank system with
several different carburetors, it is not necessary to perform a diurnal
tank test for each carburetor hot soak test. The emission value could
be assigned to each evaporative configuration on the basis of fuel tank
and carburetor combinations.
It has been estimated that a testing scheme based on selection criteria
would result in approximately 50 fuel tank families and 150 carburetor
families. If each selected system is tested three times, i.e. 5000, 25,000
and 50,000 miles, the test load would be reduced from an estimated 4500
tests to 600 for upcoming model years.
-------
-12-
The discussion above is contrary to the established certification
process. However, evaporative emissions are generated from physical
mechanisms that are hardware oriented rather than chemical reactions
as in the case of exhaust emissions. Furthermore, control systems
are prone to deteriorate with time rather than with the mileage ac-
cumulated in the accelerated durability testing. For these reasons,
it does not appear to be necessary to test each durability vehicle
at 5 mileage points.
In general, the evaporative emissions test should be separated
from the exhaust emissions test to allow more flexibility in the
testing sequence. However, when a hot soak test is required on a
vehicle, the hot soak test should obviously follow a phase of hot
vehicle operation. The diurnal test could be conducted as an in-
dependent test. Tests ( 22 ) have been conducted that showed that
the inclusion or exclusion of the diurnal test prior to the exhaust
emissions test has no detectable effect on the exhaust emissions.
Similarly, vehicle operating temperatures are generally at stable
conditions during hot operation, and the hot soak temperature profile
tends to achieve the same maximum temperature regardless of the previous
operation. Hot soak testing obviously (toes not influence the exhaust
test phase.
The evaporative emission that has not been specifically measured
heretofore is the "running loss" emission. At present, the assumption
is made that the only possible source for a running loss would be at
the tank cap or recognized vents. In most cases, no canister measure-
ment is made at the tank cap because the caps are considered leak tight
and void of emissions. The fallacy of this assumption is that, running
losses could occur from many unrecognized sources.
A scheme has been developed at the MSAPC laboratory to measure the
total running losses by essentially operating the dynamometer and vehicle
inside an enclosure. A flow through the enclosure is maintained and
sampled during the exhaust emissions test. In practice, this has been
achieved by sampling the inlet and outlet of the test cell air handling
system. Exhaust emissions were measured by the normal CVS.while the air
handler acted as a running loss CVS. This concept is a CVS within a CVS
which measures and differentiates all vehicle hydrocarbon emissions. A
description of this concept and some data collected using it have been
presented in the technical appendix.
This method of measuring running losses has been discussed here as
a potential part of the implementation plan because it is compatible with
the enclosure principle, CVS concept, and vehicle sample selection plan".
This total concept could be implemented by constructing one enclosure that
incorporated all the phases of a total emissions test. One such enclosure
would have the capability of handling about 2 total emissions tests/day
or about 40 vehicles per month; this capacity will not accommodate the
number of tests that might be required under the sample selection criteria.
-------
-13-
However, if running losses are not measured this way, one enclosure could
handle 120 diurnal and hot soak tests per month if performed independently.
A fourth evaporative emission, refueling losses, can also be measured
by the enclosure technique. Although this is presently unregulated, the
effectiveness of a standard fueling connector which might be adopted to
eliminate refueling losses could be assessed by this method. Again, the
flexibility of the enclosure method has been illustrated.
Most of this discussion has been about an implementation) plan that
is based on vehicle sample selection criteria and testing with the en-
closure - hydorcarbon analyzer combination. Several alternative plans
have been previously proposed and are listed below with a brief dis-
cussion of their limitations and problems.
A. SAE Procedure applied to all vehicles.
1) Description: Replace the federal canister procedure with
the enclosure technique and apply the test to every vehicle
tested under the present vehicle selection plan.
2) Problems: The cosl;, space*- and time involved to perform
the test sequence on each vehicle are not justified by the
data. It would be technically and politically unpopular to
propose such a plan, and very costly to execute such a scheme.
B. SAE modified procedure applied to all vehicles.
1) Description: The diurnal and hot soak phases have been
combined to shorten test time.
2) Problems: Separate emission data are lost and weighting
factors cannot be properly applied. Diagnostic aspect is
also sacrificed. Same problems of (a) apply here also.
Cost: benefit ratio is still high. Safety regulations
may not permit the fueling operations required.
C. MSAPC/ECTD procedure.
1) Description: A procedure that employed radiant heat lamps
instead of the electric blanket pads for the diurnal test,
which was also combined with the hot soak.
2) Problems: The critical temperature profile, fuel tank
liquid temperature, was' not attained by this method. Ve-
hicle color requirements (all black) were too restrictive.
Enclosure ambient temperatures would be too high, separate
emission data would be sacrificed, and energy consumption
per test would be very high.
-------
-14-
These alternatives were proposed to improve, simplify, and shorten
the test. Unfortunately, each had an associated aspect that rendered
the proposal less beneficial and more costly than the one recommended
in this paper. That proposal has been summarized again for review.
D. MSAPC evaporative emission procedure.
1) Vehicle sample selection criteria are used to define
evaporative diurnal and hot soak families and to reduce
the number of vehicles and tests required for certifica-
tion of evaporative control systems.
2) SAE J171 procedure used for testing.
3) Independent diurnal and hot soak tests are performed on
selected vehicles; each vehicle may not get both tests.
4) Calculations are performed to yield weighted emission
rates (g/mi) for each system defined.
V. Problems of Implementation: 'The proposal that has been made is not
competely without problems. Lead time, control technology, vehicle sample
selection criteria, and evaporative emission standards are all problems
which must be addressed prior to a final implementation.
The lead time problem affects both MSAPC and industry laboratories.
MSAPC has two enclosure that are operational, but a formal data col-
lection and analysis program is required prior to conducting large
numbers of tests. These problems can be resolved in the anticipated
lead time. The industry has tv:o lead time problems. Some automotive
laboratories will require construction of enclosures, and all companies
will require lead time to develop the appropriate evaporative emission
control hardware. Communications should be established with the in-
dustry to inform them of our intentions so as to maximize their lead time.
The proposed program plan, Part II of this document, estimates that 1977
MY vehicles could be tested by the enclosure procedure. These tests might
begin as early as December, 1975.
As mentioned, the control technology represents the biggest problem
to the industry. The evaporative emissions standards that would be set
weigh heavily on the control technology changes that would be required.
As the program plan outlines, the potential for control by various tech-
nques and devices will be determined. The cost of extensive changes will
be weighed against their impact on overall air quality. This is an area^
where additional information and data are needed. The program plan provides
for obtaining this information by a control technology assessment contract.
The final problem associated with adopting a new procedure is to
determine the appropriate standard of compliance. The data indicate
that the average evaporative emissions as measured by the enclosure are
much higher than the 2 gram/test (.13 g/mi) standard presently in effect.
The existing data base from enclosure tests can be combined with data
-------
-15-
generated during the control technology assessment, development, and
demonstration programs to provide a concise determination of the ap-
propriate standard and timetable for compliance.
The control systems now in use appear to have an emission rate
somewhere betv/een 1.5-2.0 g/mi as calculated form the weighted en-
closure emission values and average daily mileage. Had the evaporative
standards been established using the enclosure procedure and weighted
calculations, the 1975 standard would have been 0.27 g/mi. Assuming
the diurnal emission can be reduced to zero, the allowable one hour
hot soak emission translates to a value of 2.0 grams per test. There
have been some hot soak enclosure results obtained at the 2.0 gram
per test level, but the overall vehicle population average of 11.0
grams per test illustrates the amount of reduction still required.
It is unlikely that this reduction can be achieved in one year, al-
though the fact that the problem is a single component, hardware
oriented phenomenon makes its solution much simpler than the multi-
component, chemical reactions associated with vehicle exhaust emis-
sions. The design work being done to meet impact/fuel spillage safety
regulations may improve evaporative emissions.
Other questions related to simulation of realistic environmental
parameters have been addressed in a technical appendix to this report.
VI. Conclusions;
A. The canister test procedure is so limited in its accuracy and
integrity that its continued use has little value towards con-
tinued improvement in air quality.
B. The enclosure test procedure as specified in SAE J171 will
accurately and repeatedly measure true total evaporative emis-
sions. The parameters of the procedure represent a realistic
simulation of the critical environmental conditions. Its use
will correctly assess the performance of control techniques
and their impact on air quality.
C. Data indicate that evaporative emissions control can be assessed
on a vehicle sample selected by use of characteristic criteria.
This is an evaporative family approach which is similar to the
engine family scheme used for exhaust emissions. This type of
regulatory program would be highly cost effective.
D. Evaporative emissions will become the dominant hydrocarbon emis-
sion by 1976 under present control techniques and will result in
reductions that are lower than those needed to meet ambient air
quality requirements. Actions to regulate evaporative emissions
are required by the Clean Air Act.
-------
-16-
E. The problems of changing the procedure and standard can be solved.
Some automotive companies will require a larger effort to im-
plement the change and to develop the control techniques.
F. A program plan can be executed to provide a regulatory program
and evaporative emissions standard by September, 1975. A
milestone chart has been included at the end of this paper to
show the tasks required.
VII. Recommendations:
A. The use of the canister procedure should be minimized as soon
as possible for the following reasons:
1) High costs are incurred that produce no benefits.
2) Evaporative control technology has not changed; therefore,
evaporative emission values should remain the same. From
the data shown in Table A, the probability of a vehicle
failing the evaporative standard is very low.
3) Tests ( 22 ) have shown that the vehicle preparation and
diurnal test have no detectable effect on exahust emissions
and could therefore be deleted or separated.
4) Test time could be reduced to facilitate highway fuel economy
measurements and enclosure testing for evaporative emissions.
B. The enclosure procedure should be adopted as the official test
procedure for measuring evaporative emissions. Communications
with the automobile industry should be established as soon as
possible to express the intended changes in evaporative regulation,
to maximize lead time, and to solicit their comments on our program.
C. The program plan to achieve the development and implementation of
an appropriate standard should be approved and given high priority.
The revised regulation could possibly be implemented by December,
1975 if high priorities are assigned.
VIII. Closure: This document has attempted to reveal the problems
associated with mobile source evaporative emissions. These environ-
mental, technical, and political problems pose a challenge to the
OMSAPC which requires action. In the consideration of alternative
actions, "no action" is not a responsible alternative.
-------
IX. References for Evaporative Losses
SAE Vehicle Emissions - Part I Vol. 6, 1955-1962
1. Wentworth, J. T., "Carburetor Evaporation Losses", SAE Transactions
1955.
SAE Vehicle Emissions - Part II Vol. 12, 1963-1966
2. Muller, H. L., Kay, R. E., and Wagner, T. 0., "Determining the
Amount and Composition of Evaporative Losses from Automotive Fuel
Systems". SAE Transactions, Vol. 75, 1967.
3. Ebersole, G. D., and McReynolds, L. A., "An Evaluation of Automobile
Total Hydrocarbon Emissions", SAE Transactions, Vol. 75, 1967.
4. Caplan, J. D., "Smog Chemistry Points and the Way to Rational
Vehicle Emissions Controls". SAE Transactions, Vol. 74, 1966.
SAE Vehicle Emissions - Part III, Vol. 14, 1967-1970
5. Martens, S. W., and Thurstor), K. W.* "Measurement of Total Vehicle
Evaporative Emissions" SAE Transactions 680125, 1968, p. 191.
6. Martens, S. W., "Evaporative Emissions Measurements with the Shed -
A-Second Progress Report" SAE Transactions 690502, 1969, p. 202.
7. Wade, D. T., "Factors Influencing Vehicle Evaporative Emissions",
SAE Transactions, 670126, 1967, p. 743.
8. Wade, Clarke, Gerrard, Skarstrom, Vardi, "An Adsorption-Regeneration
Approach to the Problem of Evaporative Control", SAE Preprint, 670126
1967, p. 756.
9. Koehl, W. J. Jr., "Mathematical Models for Prediction of Fuel Tank
and Carburetor Evaporative Emissions", SAE Preprint, 690506, May,
1969.
10. Nelson, E. E., "HC Control for Los Angeles by Reducing Gasoline
Volatility", SAE Transactions, 690087, 1969, p. 775.
11. Jackson, M. W., and Everett, R. L., "Effect of Fuel Composition on
Amount and Reactivity of Evaporative Emissions", SAE Transactions,
690088, 1969, p. 802.
12. Society of Automotive Engineers, SAE J171a, "Measurement of Fuel
Evaporative Emissions Using the Enclosure Technique", July , 1972.
13. Deeter, W. F., et. al. "An Approach for Controlling Vehicle Emissions
SAE Preprint 680400, (1968).
-------
-18-
14. Deeter, W. F., and Jewell, R. G., "Factors Affecting Carburetor
Vapor Losses".
15. Automobile Manufacturer's Association, AT-1A Engineering Notes 6/6,
"Fuel Systems Evaporation Losses", September, 1961.
16. Coordinating Research Council, CRC Report No. 400, "1966 CRC Motor
Vehicle Evaporation Loss Measurement Technique Evaluation", January,
1967.
17. U. S. Dept. of Commerce, NOAA, "Climatography of the United States
No. 84, Daily Normals 1941-70", September, 1973.
18. National Academy of Sciences, "Semiannual Report to EPA by CMVE of
NAS", January 1, 1972.
19. National Academy of Sciences, "A Critique of the 1975-76 Federal
Automobile Emissions Standards for Hydrocarbons and Oxides of
Nitrogen", Prepared for EPA, May 22, 1973.
20. Korth, M. W., Rose, A. H., and Stahman, R. C., "Effect of Hydrocarbon
to Oxides of Nitrogen Ratios'on Irradiated Auto Exhaust", APCA Annual
Meeting, June, 1963, PHS Document.
21. Rose, A. H. Jr., "Summary Report of Vehicular Emissions and Their
Control", PHS Report, Unpublished, 1966.
22. Kruse, R. E., "Evaluation of the Effect of Evaporative Emission
Testing Upon Exhaust Emission Test Results", January, 1974. EPA
Memo to M. W. Korth.
23. Patterson, D. J., and Henein, N. A., "Emissions from Combustion
Engines and Their Control", (Ann Arbor: Science Publishers, Inc., 1972)
Chapter 6, p. 181.
24. Society of Automotive Engineers, SAE J1045, "Instrumentation and
Techniques for Vehicle Refueling Emissions Measurement", August,
1973.
25. Letter from D. A. Jensen to Technical Advisory Group on Evaporative
Emissions, October 18, 1965.
26. Letter from M. L. Brubacker to the Evaporative Emissions Advisory
Group, January 10, 1966.
27. Federal Register, Volume 32, Number 24, Part 85, "Control of Air
Pollution from New Motor Vehicles and New MOtor Vehicle Engines",
February 4, 1967, p. 2448.
-------
-19-
28. Federal Register, Volume 33, Number 108, Part II, "Standards for
Exhaust Emissions, Fuel Evaporative Emissions and Smoke Emissions,
Applicable to 1970 and Later Vehicles and Engines". June 4, 1968,
p. 8304.
29. Federal Register, Volume 35, Number 136, July 15, 1970.
30. Federal Register, Volume 36, Number 70, "Certification Test Results
for 1971 Model Year", April 10, 1971.
31. Federal Register, Volume 37, Number 114, "Certification Test Results
for 1972 Model Year", June 13, 1972.
32. Federal Register, Volume 38, Number 84, "Certification Rest Results
for 1973 Model Year", May 2, 1973.
33. Federal Register, Volume 38, Number 212, "Certification Test Results
for 1972 Model Year", November 5, 1973.
34. Miscellaneous reports on file in EPA Procedures Development File
on Evaporative Emissions. MSAPC, Ann Arbor, Michigan, 1973
35. Coordinating Research Council, CRC-Scott Project No. 2602, CAPE-5-68,
"Time-Temperature Histories of Specified Fuel Systems". Vol. 1,
October, 1969.
36. Coordinating Research Council, CRC-SDC, CAPE-10, "Summary Report
on Vehicle Driving Patterns", January 1971.
------- |