EVAP 76-4
Technical Support Report for Regulatory Action
A Study of Methods for Reducing Evaporative
Background Hydrocarbon Emissions
from New Vehicles
October 1976
Thomas Rarick
Lou Donate
Notice
Technical support reports for regulatory action do not necessarily
represent the final EPA decision on regulatory issues. They are intended
to present a technical analysis of an issue and recommendations resulting
from the assumptions and constraints of that analysis. Agency policy
considerations or data received subsequent to the date of release of
this report may alter the recommendations reached. Readers are cautioned
to seek the latest analysis from EPA before using the information contained
herein. •
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air and Waste Management
U.S. Environmental Protection Agency
-------
Contents
Page
1. Introduction 1
2. Summary and Conclusions 1
3. Technical Discussion 2
3.1 Program Objective 4
3.2 Program Design 4
3.3 Facilities and Equipment 4
3.3.1 Facilities 4
3.3.2 Test Equipment 5
3.3.3 Test Vehicles 5
3.3.4 Test Fuel 5
3.4 Test Procedures 5
4. Test Results 8
4.1 Total Evaporative Emission Tests 8
4.2 Background Emission Tests 8
5. Discussion 14
6. References 15
-------
1. In t roduc t ion
Evaporative hydrocarbon emissions from gasoline fueled vehicles
come from two major sources: 1) fuel evaporation; and 2) background
emissions which are given off from hydrocarbon sources such as paints,
solvents, plastics, vinyl, etc. Background emissions are high from
brand new vehicles (1-10 days old) but decay rapidly to some stabilized
level (.5 g or less). Mobile Source transient background emissions
(background emissions above the stabilized emission level) probably do
not have a significant impact on air quality since they last for a
relatively short time (3 to 6 months as opposed to 10 years for fuel and
stabilized background evaporative emissions).
A problem arises when evaporative emissions testing is done on new
vehicles which still have high transient background emission levels.
The proposed method of evaporative emission measurement is the evapora-
tive enclosure method. This test involves putting the test vehicle in a
sealed enclosure and measuring the change in hydrocarbon levels within
the enclosure over a period of one hour. A test such as this cannot
distinguish between fuel evaporative emissions and background emissions.
A similar test caja be used to measure background emissions onlyj by
removing the fuel; system during the test, but again such a test cannot
distinguish between transient and stabilized background emissions.
In order to solve the problem of finding a test which will only
measure fuel and stabilized background evaporative emissions, testing of
very new (2-5 weeks old) vehicles was conducted. The purpose of this
study was two fold: 1) determine if the magnitude of transient back-
ground levels could be accurately predicted at some point in time by
performing total evaporative emission tests; and 2) determine if tran-
sient background levels could be artificially lowered to the point of
insignificance by elevated temperatures (baking) and/or driving.
2. Summary and Conclusions
It was the objective of this study to determine if transient evapora-
tive background emissions can be predicted by mathematically generating
a curve using total evaporative emission data. Also, the feasibility of
lowering background emission levels by baking and driving test vehicles
was studied.
Tests were conducted on four test vehicles for which total emis-
sions were measured intermittently over a 3-week period. Between emis-
sion measurements certain vehicles were baked and/or driven on a dynamo-
meter to attempt to reduce their background emissions. At the end of
this testing at least two repeat background emissions tests were conducted
on each vehicle.
The results of the curve fitting of total evaporative emission
measurements indicated that background emission levels could not be
accurately predicted using such a technique.
-------
-2-
The results of the background tests on three identical vehicles
indicated that baking and driving did result in lower background emis-
sion levels than for the control vehicles. The background levels of all
tests cars were below .5 grams 40 days after they were built. It is
recommended that vehicles tested for evaporative emissions be required
to have low background levels as it appears this is feasible and much
more reliable than any acceptable prediction techniques now available.
3. Technical Discussion
Recent information resulting from comments to the proposed Evapora-
tive Emission Regulations for light duty vehicles and light duty trucks
have increased the understanding of background emission decay. A study
done by Ford Motor Company using a fleet of 10 cars showed that back-
ground decay follows the following equation form
B = aTb
where: B = background emission level
T = Time
a & b = constants.
If the constants a and b are known, then the background level at
any time can be calculated. This calculated value (minus the level
considered to represent stabilized emission levels) could then be used
to adjust the emission levels measured during an evaporative emission
test. Thus, a means of determining the constants a and b is needed.
The equation form for background emissions can be transformed (by
a logarithmic transformation) to a linear equation form, as shown:
log B = b log T + log a
Using this equation form, one can determine the coefficients a and
b by using a least squares curve fitting technique. Data can be generated
by making background emission measurements at several times. (Preferably
the tests should be separated by at least 10 days)* Background measure-
ments can be made using the SAE J171a (Appendix C) procedure. This
procedure calls for total removal of the fuel system. The vehicle then
undergoes a cold soak in the SHED. The vehicle is then driven over an
Urban Dynamometer Driving Cycle (UDDS), followed by a hot soak in the
SHED. The carburetor must be put on prior to the drive portion of the
test and then quickly removed prior to the hot soak.
This procedure generating the constants a and b of the background
equation could be done prior to testing a vehicle for evaporative emis-
sions. However, this procedure has two major problems: 1) the integrity
of the fuel system may be destroyed in the process of making background
measurements, such that the resulting evaporative emission measurements
would be affected; and 2) the background measurement technique is complex
and time consuming.
-------
-3-
An easier method of predicting background emissions would be to
measure total evaporative emissions Concluding background), and then
somehow separate out the background contribution mathematically. The
following equation could be used to describe total (fuel and background)
evaporative emissions:
E = aTb + F
where E = Total evaporative emissions (Fuel and background)
T = Time
F = Fuel evaporative emissions
a & b = constants
This equation assumes that fuel evaporative emissions are constant
as a function of time. This is a reasonable assumption from the stand-
point that a new fuel system would not deteriorate over the relatively
short time _p_eriod such measurements could be made. However, the vari-
ability, S/X, of fuel evaporative emissions is roughly 10%. Thus, a
large number of test points would be required to generate values of a
and b with an acceptable level of confidence. Also, the equation form
for total evaporative emissions cannot be transformed into a linear form
such that a least squares curve fit technique could be used.
Data could be fit to the equation for total evaporative emissions
by non-linear methods. There are several methods which could be employed
to accomplish this. One way would be to select a value for b, solve for
the coefficients a and F using a least squares method and calculate the
regression coefficient and standard error. Several values of b could be
selected and the one resulting in the best regression coefficient and
standard error would be used.
The question that must be answered is how accurate are the back-
ground values predicted by an equation arrived at in the way described.
The accuracy would depend on the number of observations made, the close-
ness of the observations (the more time separation the better the accuracy),
and the accuracy of each individual measurement. There are, however,
practical limits on the number and closeness of test points due to the
cost and manpower of conducting tests and the need to determine during a
6 month period, whether or not a given car can meet certification require-
ments . The question of accuracy will be looked at in this report by
subjecting total evaporative emission data to a curve fitting technique.
Another way of pursuing the problem of transient background emissions
would be to try and eliminate them. Transient background emissions are
known to decay with time. If the decay can be accelerated, or if some
or all background sources can be eliminated, then the prediction techniques
discussed would not be needed. Background emissions are believed to
result from paints, solvents, plastics, etc. An earlier study by Ford
Motor Company indicated that subjecting a vehicle to elevated temperatures,
-------
-4-
and driving a vehicle resulted in lower background emissions than would
have normally been expected. This report will analyze the results of a
similar study.
3.1 Program Objective
It is the objective of this study to determine if transient evapora-
tive background emissions levels can be predicted by generating total
fuel emission data and then curve fitting those data. Another objective
of this study is to determine whether or not background levels can be
artificially lowered by exposing the vehicle to elevated temperatures
and/or by accumulating mileage.
3.2 Program Design
The test program involved the repeat testing of 4 brand new test
vehicles over a 2-3 week period. Total evaporative emissions were
measured on the four test vehicles five times. These data were then
used to determine the accuracy of curve fitting techniques used in
predicting transient background levels.
Three of the test vehicles were identical (3 Chrysler Volare's).
These three vehicles were used to determine the relative effectiveness
of baking and driving a vehicle in reducing background emission levels.
One vehicle was used as a control. It was allowed to soak at a rela-
tively constant temperature in the soak area. A second vehicle was
repeatedly baked in a bake oven at roughly 160°F for 12 hours. The
third vehicle was baked similarly to the second vehicle and was also
driven over 6 Urban Dynamometer Driving Schedules (UDDS's) over a 6 hour
period after each baking. Actual background emission measurements were
made on all vehicles at the end of the test program to determine the
relative magnitude of the background levels.
A fourth vehicle (Chevrolet Nova) was baked and driven similarly to
the one Volare'. Tests were conducted on this vehicle to get a further
indication of how far background emissions could be lowered.
3.3 Facilities and Equipment
3.3.1 Facilities
All evaporative emission tests took place at the Motor Vehicle
Emission Laboratory in Ann Arbor, Michigan. The Light Duty Vehicle
evaporative enclosure was used during all evaporative emission tests.
The enclosure is nominally 8 feet high x 10 feet wide x 20 feet long and
has a measured volume of 1540 ft . Calculation of the enclosure volume
with a propane injection and recovery test compared within + 2% of the
measured volume. Propane retention tests performed periodically indicated
a leakage rate of less than 0.5% per hour.
-------
-5-
Mileage accumulation and vehicle operation prior to the hot soak
tests were done on Clayton DDVIF dynamometers.
Baking was done in a local paint shop oven. The oven used forced
hot air and the oven temperature was kept between 155°F and 165°F. The
windows and trunks of the vehicles were opened during baking. The fuel
tank was vented to the outside of the oven by use of a specially fitted
gas cap and a length of nylon tubing. This was done as a safety pre-
caution and to prevent abnormally loading the evaporative control
system.
3.3.2 Test Equipment
Evaporative emission enclosure hydrocarbon measurements were taken
using a Flame lonization Detector (FID). A continuous sample was taken.
3.3.3 Test Vehicles
Four 1976 MY vehicles were used during testing. The criteria for
vehicle selection were: 1) the vehicles had to be less than 3 weeks
old; and 2) the vehicles had to have low mileage (less than 250 miles).
All vehicles were assembled during the week of April 11, 1976. The four
vehicles were 3 Plymouth Volare's and a Chevrolet Nova. Table 3-1 gives
specific information about these vehicles.
3.3.4 Test Fuel
Indolene type HO lead-free test fuel was used throughout the program.
The RVP of the fuel was between 8.7 and 9.2 psi.
3.4 Test Procedures
The test procedures used during the total evaporative emission
measurement portion of the test program are shown in Figure 3-1.
Baseline tests were conducted and then the vehicles went through the
remaining sequence four times. The bake portion of the test involved
baking the vehicles for 12 hours (overnight) at approximately 160°F.
The driving portion of the test consisted of driving a UDDS cycle once
each hour for 6 hours. In this way the engine could be expected to stay
hot for a period of several hours. The evaporative emission test consis-
ted of a cold soak in the SHED for one hour, followed by a UDDS driving
cycle; followed immediately (less than 5 minutes) by a one hour hot soak
in the SHED.
Background emission tests were taken in a similar fashion as the
total evaporative emissions tests except the fuel system components were
either removed or plugged during the tests. The fuel tank, carburetor,
and evaporative canister were removed from each of the vehicles. Fuel
lines, lines connected to the canister, the opening to the intake mani-
fold, dipstick hole, lines to and from the fuel pump, and the exhaust pipe
-------
-6-
Age at Baseline,
days
Age at final
test , days
Mileage at
baseline
Mileage at
final test
Type of
testing
Vinyl Top
Vinyl Seats
Under coated
Nova
18
38
227 miles
520 miles
Baked
and driven
NO
NO
NO
White
Volare'
20
42
30 miles
105 miles
Control
(soaked only)
YES
YES
NO
Copper
Volare'
21
43
28 miles
154 miles
Baked
YES
YES
NO
. Blue
Volare'
18
39
28 miles
320 miles
Baked and
driven
YES
YES
, NO
Table 3-1 Vehicle Information
-------
Repeat 3 times
Nova & Blue Volare'
Baseline
Test
(hot &
cold soak)
Copper Volare' _
1
Bake
<§ 160°F
for 12 hours
i
Bake
@ 160 °F
for 12 hours
Repeal
Drive .
6 UDDS's over
6 hour period
: 3 times
12 hour ^
soak
12 hour
soak ~~
Repeat 3
White Volare'
Retest
Cold
hot s
and
oak
Retest
Cold
and
hot soak
times
1
Retest
Cold &
hot soak
i
i
Figure 3-1 Test Procedure Flowchart for Evaluation
of Total Evaporative Emissions.
-------
-8-
were all sealed during these tests. The UDDS cycle was driven with the
carburetor in place fueled with an auxiliary fuel tank outside the
vehicle. After vehicle operation, the carburetor was removed and the
intake and exhaust pipe were sealed. This procedure was essentially the
same as prescribed in Appendix C of the SAE J171a recommended test
procedure .
4. Test Results
4.1 Total Evaporative Emission Tests
The emission test results for the four vehicles are given in Table
4-1, for cold soak, hot soak and total emissions. The hot and cold soak
test results for the four vehicles are shown graphically in figures 4-1
and 4-2 respectively. Figure 4-3 shows the total emission results for
each of the four vehicles as a function of time and also shows the best
fit curve of the data in the equation form E = aT + F. The equations
shown were derived by trying several values of b, solving for a and F,
and then picking the set of coefficients which gave the best correlation
coefficient.
Figure 4-3 shows that the curves generated fit the data reasonably
well but the large scatter in data for certain vehicles resulted in
equations which do not fit the physical situation. The equation for the
Copper Volare'predicts a negative fuel emission term. The equations for
the blue and white Volare's predict negative background emissions for any
time. The equation which describes the data for the Nova is the only
one which fits the actual situation (fuel and background emissions are
positive and the function is continuously decreasing). The predicted
background emission line is also given in the figure for the Nova. It
shows that a background emission level of 3.28 g and 3.18 g would be
expected at 37 days and 38 days old, respectively. Actual background
emission measurements were made at these times arid showed actual levels
of .41 g and .35 g.
4.2 Background Emission Tests
The results of the hot .plus cold background evaporative emission
tests are given in Table 4-2 for the four test vehicles. The results
are also shown graphically in Figure 4-4. The most striking result of
these tests is that the background levels are quite low for the age of
the vehicles. Even the control Volare' (which was not baked or driven)
had less than .4g total background after 40 days. Baking and driving
the vehicles appeared to do some good in lowering background emissions
as the blue Volare' (baked and driven) had lower background levels than
the copper Volare' (baked only). The control vehicle had the highest
final background emission levels which were almost double the levels of
the blue Volare'. As stated earlier the Nova showed low background
levels of .41 and .35 g at 37 and 38 days respectively.
-------
Vehicle
Nova
(baked and
driven)
White
Volare'
(Control)
Copper
Volaie'
(baked)
Blue
Volare'
(baked
and
driven)
Type
of
Test
Cold
Hot
Total
Cold
Hot
Total
Cold
Hot
Total
Cold
Hot
Total
Baseline
Age,
davs
18
18
18
20
20
20
21
21
21
18
18
18
HC loss,
grams
0.39
9.62
10.01
0.68
4.62
5.30
0.54
5.23
5.77
0.50
3.58
4.08
Retest #1
Age.
days
22
22
22
23
23
23
23
23
23
21
21
21
HC loss,
grams
0.40
9.72'
10.12
0.58
4.58
5.16
0.35
7.61
7.96
0.32
4.31
4.63
Retest #2
Age,
rtaya
25
25
25
27
27
27
27
27
27
25
25
25
HC loss,
grama
0.29
8.17
8.46
0.51
3.15
3.66
0.35
3.38
3.73
0.22
2.45
2.67
r
Retest #3
Age,
rtaya
30
30
30
30
30
30
31
31
31
28
28
28
HC loss,
grams
.0.25
7.87
8.12
0.57
3.07
3.64
0.49
1.14
1.63
0.31
2.52
2.83
Retest #4
Age,
davs
35
35
35
34
34
34
36
36
36
33
HC loss,
grams
0.20
7.36
7.56
0.66
2.09
2.75
0.38
3.57
3.95
0.29
*
*
*Hot soak test voided due to loose carburetor fitting.
Table 4-1 Evaporative Emission Test Results
-------
12
-10-
10
CO
c
o
•H
CO
CO
•H
CB
O
C/3
8 .
—i—
30
—r—
10
20
40
12 .
Vehicle Age, days
Figure 4-1 Hot soak emissions vs. time
for 4 test vehicles.
LEGEND
&—6 Nova (baked & driven)
y.—x White Volare' (control)
O—o Copper Volare'(baked)
a—o Blue Volare'(baked
& driven)
10 .
co
C
o
•rl
CO
co
ffl
O
to
o
o
8 _
6 .
4 .
2 .
10
20 30
Vehicle age, days
Figure 4-2 Cold Soak Emissions vs Time
For 4 Test Vehicles
40
-------
15 -,
10 J
3
o
•H
to
a 5
o
m
30 _
20 J
10 J
Blue Volare'
(baked and driven)
E = -0.1201 X 10~3T 3>0+5.084
(r - 0.8528)
Vehicle Age, days
Nova
i\ (baked and driven)
E - 174.04T"1'1 + 4.126
(r = 0.965)
Predicted
background
10
20 30
Vehicle age, days
40
15
to
a 10
0)
m
15 _
10 J
a
Copper Volare'
(baked)
E - 89.84T"0'6 -8.087
(r - 0.678)
10 20 30
Vehicle Age, days
White Volare'
(control)
E - -2.724T°'40+14.06
(r •» 0.940)
10
20
30
. Vehicle Age, days
Figure 4-3 Evaporative HC Emissions vs. Age for 4 Test Vehicles
40
40
-------
-12-
.
Vehicle
Nova
(Baked and
driven)
White
Volare'
(Control)
Copper
Volare'
(baked)
Blue
Volare'
(baked and
fl T"! VPn}
Test
M*-*
WO •
i
2
1
2
1
2
3
4
1
2
..J
Age,
days
37
38
41
42
38
41
42
44
36
39
Background
Cold Soak
0.132
0.062
0.126
0.132
0.115
0.064
0.074
0.060
0.117
0.042
Emissions,
Hot Soak
0.275
0.290
0.272
0.257
0.506
0.203
0.217
0.150
0.275
0.178
grams
! Total
1
| 0.407
i
0.352
0.398
0.389
0.621
0.267
0.291
0.210
0.392
0.220
Table 4-2 Background Emission Test Results
-------
-13-
0.6
co
C
o
S 0.4
13
C
2 0.2
oo
^!
o
tfl
oa
,_ Copper Volare'
White Volare'
Blue Volare'
36
44
44
48
Vehicle Age, Days
Figure 4-4 Background HC Emissions vs Age
For 5 Test Vehicles
-------
-14-
5. Discussion
Two possible ways of dealing with transient background emissions
have been explored. The first method was to measure total vehicle
emissions and use a mathematical curve fitting technique to generate
predicted background emission levels at some future time. The test
results generated indicated that this technique does not generate
accurate predicted values. Of the four sets of test data,, three resulted
in equations which did not correspond to real world conditions, and the
fourth predicted emission levels eight times higher than actual levels.
This result was most likely due to the expected variability of fuel
evaporative emissions. If future improvements in the test procedure can
reduce test variability significantly, this method of accounting for
transient background emissions could work. However, on the basis of
these tests, it is recommended that other ways of handling background
emissions be investigated.
The other way of handling background emissions was to attempt to
artificially lower the background levels of new vehicles by baking '
and/or driving vehicles. The results of this study indicated that
background levels were lowered when vehicles were driven and/or baked.
More significantly, however, was the fact that a vehicle which was
neither baked nor driven had low background levels (less than .4g) at 40
days old. The methods for lowering the background levels used are only
a couple of many methods that can be used. Vehicles can be built
without plastic or vinyl components, without paint, without sound
deadening, etc. The use of several of these techniques together should
produce vehicles with background levels much below expected stabilized
background levels. It is recommended, therefore, that low background
vehicles be required for testing rather than using some analytical test
method for predicting background emission levels.
-------
-15-
6. References
1. Ford Motor Company's Submission of Comments to the January 13, 1976
Notice of Proposed Rulemaking for Evaporative Emission Regulations,
February 27, 1976.
2. "Measurement of Fuel Evaporative Emissions from Gasoline Powered
Passenger Cars and Light Trucks using the Enclosure Technique," SAE
Recommended Practice, SAEJ171a, SAE Handbook.
3. EPA Technical Memorandum, "Background Test Evaluation at Ford Motor
Co.," Memo from Gary Wilson, EPA, to Charles Gray, EPA, April 9,
1976.
------- |