United States Air and Radiation EPA420-P-99-003
Environmental Protection M6.EVP.003
Agency February 1999
&EPA Evaluating Multiple Day
Diurnal Emissions Using
RTD Tests
DRAFT
> Printed on Recycled Paper
-------
EPA420-P-99-003
February 1999
M6.EVP.003
Assessment and Modeling Division
Office of Mobile Sources
U.S. Environmental Protection Agency
NOTICE
This technical report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data which are currently available.
The purpose in the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical developments which
may form the basis for a final EPA decision, position, or regulatory action.
-------
- Draft -
Evaluating Multiple Day Diurnal
Evaporative Emissions Using RTD Tests
Report Number M6.EVP.003
January 1999
Phil Enns
U.S. EPA Assessment and Modeling Division
1.0 INTRODUCTION and BACKGROUND
This report documents an analysis of diurnal evaporative emissions from light-duty
vehicles (LDVs) and light-duty trucks (LDTs) occurring over periods of more than one day.
Results of this study will be used in MOBILE6 in conjunction with estimates of vehicle and truck
activity and estimates of evaporative emissions for shorter periods to obtain total diurnal
emission values.
The underlying causes of diurnal evaporative emissions are discussed at length in several
reports1'2'3. By definition, diurnals are those emissions associated with daily temperature change
and its effect on vaporization of a vehicle's fuel and the expansion of fuel vapor. The evolution
of technology and regulations is assumed to influence diurnal emission rates. These trends also
are discussed in the references cited above. In the modeling of multiple day diurnals presented
here, several categories of vehicles are considered, based on model year, fuel metering and
purge/pressure test1. These are chosen to achieve consistency with groupings employed in the
MOBILE emissions inventory model.
2.0 DATA SOURCES
In this analysis, EPA considered real-time diurnal (RTD) test data from testing programs
(i.e., work assignments) performed under contract for EPA. The data consist of hourly values of
HC emissions (in grams) measured under varying conditions of fuel Reid vapor pressure (RVP)
and ambient temperature. Daily totals are obtained directly from these hourly values.
Sandman, L. "Evaluating Resting Loss and Diurnal Evaporative Emissions Using RTD
Tests," Report No. M6.RTD.001.
2Heirigs, P.L. and R.G. Dulla, "Analysis of Real-Time Evaporative Emissions Data,"
Sierra Research, Report No. SR97-12-01, December, 1997.
3Haskew, H.H. and T.F. Liberty, "Diurnal Emissions from In-Use Vehicles,"
Coordinating Research Council, CRC E-9, January, 1998.
Last Revised 1/13/99
-------
The RTD testing performed for EPA was done by its testing contractor (Automotive
Testing Laboratories) over the course of five (5) work assignments from 1994 through 1996
(performed under three different EPA contracts). A total of 119 light-duty vehicles (LDVs) and
light-duty trucks (LDTs) were tested in these programs. Table 1 displays the distribution of
vehicles and individual tests by several characteristics. Of special interest is the length of the
tests, ranging from 33 to 72 hours. More complete descriptions of these data are found in the
reports cited earlier.
Other reports on diurnal emissions utilize data from a testing program performed for the
Coordinating Research Council (CRC). However, because all these tests were run for 24 hours
only, and yield no information on multiple day emissions, they are not employed in the current
study.
In addition, the two EPA vehicles identified as "gross liquid leakers" are omitted from
these analyses. The emissions of these vehicles are large, tending to skew estimates for non-
leakers, while the mechanisms by which emissions are produced are quite different from the two
groups. EPA proposes to treat multiple day emissions from gross liquid leakers as constant.
3.0 METHODOLOGY
This work involves estimating the change in diurnal evaporative emissions from the first
day to later days. In the MOBILE model these estimates can be used to determine emissions for
full Days 2 and 3 given total emissions for Day 1. These in turn can be subdivided into hourly
values as needed.
When modeling RTD emissions, potential explanatory factors include fuel metering
technology, model year, and outcome of purge and pressure tests performed on the vehicle.
Ambient temperature and fuel volatility also are known to play a central role.
3.1 Model Form
The percent change in emissions from one day to the next can be modeled by expressing
the natural logarithm of emissions as a linear function of potential explanatory factors:
ln(Emissions) = b0 + bjXj + b2X2 + ... + bkXk
where the bj coefficients are constants and the Xj 's are factors related to emissions. In this model,
the coefficient bj is interpreted as an approximate measure of the percentage change in emissions
per unit change in Xj when the other factors are unchanged (see Appendix).
Consider the following representation of multiple day diurnal evaporative emissions:
ln(HC) = b0 + bŁ> + b2P,*D + b3P2*D + b^D + b5Y2*D + b6F*D + b7R + bgT
where dummy variables are used to switch on or off the categorical factors of day, purge/pressure
test status, model year, and fuel metering:
Last Revised 1/13/99
-------
D = 0ifdayi, 1 ifdayi+1;
P! = 1 if vehicle fails the purge test and passes the pressure test (F/P), 0 otherwise;
P2 = 1 if vehicle fails the pressure test (P/F or F/F: these outcomes are combined
due to lack of data), 0 otherwise;
Yj = 1 for pre-1980 model years; 0 otherwise;
Y2 = 1 for 1980-85 model years; 0 otherwise;
F = 1 for carbureted vehicles; 0 for fuel injected;
and
R = Reid vapor pressure (pounds per square inch);
T = temperature (degrees Fahrenheit).
The nominal factors are chosen because they represent the categories to be used in MOBILE6. As
shown in the Appendix, for a given combination of purge/pressure status, model year range, and
fuel metering, the percent change in HC from Day i to Day i+1 is given by:
t>! + b2Pi + b3P2 + b^ + b5Y2 + b6F (1)
For example, with 1986-96 fuel injected vehicles that fail the pressure test, the percent change
over a one day period is
b!+b3P2 (2)
For the continuous variables of Reid vapor pressure and temperature, the coefficients (b7 and b8)
represent the percent change in emissions per unit change in the given variable.
3.2 Model Estimation
The above model can be fitted using ordinary least squares regression. In order to account
for additional variation, a vehicle factor was included. This effectively fits a different intercept
term to each vehicle and helps produce sharper estimates of the coefficients shown above. The
goal of the analysis is to obtain point estimates of the linear combinations of the type shown in
equation (2). Given the categories of fuel metering, model year and purge/pressure test status, a
total of 18 different values can be estimated for each day (and vapor pressure/temperature
combination). This number can be reduced if there is insufficient evidence to justify separating
categories. That approach is adopted in the analysis reported below.
Because the available data include tests of varying length, it is difficult to compare
emission values from all tests for the purpose of estimating full day changes. In particular,
complete 72-hour tests are available in only six of the technology, model year and pressure purge
test status categories. However, as seen in Table 1, there are a large number of EPA 33-hour and
38-hour tests, and these provide more complete coverage of the categories. These tests give some
Last Revised 1/13/99
-------
indication of change in evaporative emissions from the first day to the second. One way to use
these data is to consider only the first nine hours of each day, since the 33-hour tests give only
that number of hours in Day 2. If it is assumed that the total emissions in the first nine hours are
comparable across days then the effective data set numbers 564 tests.
4.0 Initial Results
Two models were fitted to the 9-hour data described above, one for Days 1 and 2, and the
other for Days 2 and 3. Regression coefficient estimates, computed using the SAS GLM
procedure, are found in Tables 2(a) and 2(b), respectively.
4.1 Effect of Fuel Metering, Pressure/Purge Test Status, and Model Year
In both models, neither of the model year terms is statistically significant. Therefore, as a
first step toward simplification, the model year factor was removed from the analysis. Refitting
the models gives estimates shown in Tables 3(a) and 3(b). In the Day 1-to-Day 2 equation, all
terms are significant. For Day 2-to-Day 3, the purge/pressure test terms are not significant,
possibly because 23 of the 26 vehicles tested for three days were from the single pass/pass
purge/pressure group.
Actual percentage effects from the various combinations of fuel metering and
purge/pressure test status can be estimated using the ESTIMATE feature of the SAS GLM
procedure. This is applied to the linear functions illustrated by equation (2). Tables 3(a) and 3(b)
display the results. Across the two models, the only percent change that is clearly different than
zero is for the class of fuel injected vehicles that pass both purge and pressure tests. For the Day
1-to-Day 2 model, two other categories are significant at the five percent level: fuel injected
vehicles that fail the purge test and carbureted pass/fail vehicles. However, when the three
categories of pressure/purge test result are compared on a pairwise basis, it is seen that the F/P
and (F/F or P/F) groups do not differ significantly (p=0 . 5423). Therefore, a further
simplification is proposed in which a vehicle is classed as "PASS" (pass both tests) or "FAIL"
(fail one or both tests). Table 4 gives results for the model using this classification.
Table 5 shows estimates for the Day 1-to-Day 2 changes when the sample includes only
the vehicles for which 72-hour data was collected. This is the same subsample that applies to the
Day 2-to-Day 3 estimates. Therefore, the values are more directly comparable for the two sets of
estimates. For this reduced set, only the percentage effect for the fuel injected P/P group is
significant, and its value is substantially larger than for the full sample (49.6% vs. 36.5%).
Using the values from Tables 3 to 5, the following recommendations are made for multiple day
percent changes to be used in MOBILE6:
1. Fuel-injected vehicles passing both purge and pressure tests. Point estimate percentage
increases are 36.5% for Day 1-to-Day 2; and 43.8% for Day 2-to-Day 3. However, when only the
72-hour data are considered (Table 5), the Day 1-to-Day 2 value is 51.2%. We can argue that the
first figure is more precise (because it is based on the larger sample) and should be used, but the
Last Revised 1/13/99
-------
ratio of the two daily changes ought to reflect estimation which is based on the same data, i.e.,
the 72-hour data. That ratio, 43.8/51.2 or 0.856, applied to the first day percentage gives an
estimate for Day 2-to-Day 3 of 31.2% for the larger sample. This suggests that while the daily
emissions are continuing to increase into the third day, they appear to be leveling off. EPA
proposes to use the 36.5% value as the percent increase in diurnal emissions from the first to the
second day, and 31.2% for the second to third day increase.
2. Fuel-injected vehicles failing one or both of the purge and pressure tests. When the P/F and
F/P-F/F groups are estimated separately, the significance tests give mixed results: for Day 1-to-
Day 2, the P/F group percent increase is not significant, while the F/P-F/F group is significant
(p=0.010). These two groups show similar percentages. After combining these groups, the
estimated percentage for Day 1-to-Day 2 is 13.3%, and is statistically significant (Table 5a). The
Day 2-to-Day 3 value is not significant and EPA proposes to set it to zero.
3. For the carbureted vehicles, the Day 1-to-Day 2 P/F class has a marginally significant value.
Because it is negative, and the other classes are not significant, EPA proposes setting all
carbureted vehicle percent changes to zero.
These results are summarized in the following table.
Pressure/
Puree
Fail One or Both
Fail One or Both
Pass Both
Pass Both
Fuel
Meteri n e
Carbureted
Fuel Injected
Carbureted
Fuel Injected
Dav 1 to Dav 2
0.0%
13.3%
0.0%
36.5%
Dav 2 to Dav 3
0.0%
0.0%
0.0%
31.2%
It is further proposed that all changes be assumed to stabilize at zero following Day 3. This
appears reasonable for the first three cases in the table, where none of the Day 2-to-Day 3 percent
changes is statistically significant. For the most common situation, fuel injected vehicles that
pass both pressure and purge tests, an argument could be made for modeling continued positive
but decreasing changes in diurnal evaporative emissions for succeeding days. That is not
proposed here since we lack data with which to form estimates.
Otherwise, the numbers in the above table do not seem unreasonable. In passing fuel-
injected vehicles, the evaporative emission control system is assumed to be functioning properly.
For these vehicles, the Day 1 base level of evaporative emissions is comparatively small. Over
time, the canister fills and excess evaporative emissions escape from the vehicle. The daily
increase in these emissions is estimated to be greatest on the second day, somewhat smaller the
third day, and constant thereafter. Thus, absolute emissions are larger on the second day than on
the first, still larger on the third day, and unchanged beyond that time.
In the other categories, base emissions are higher so that canister overloading is a smaller
component of overall emissions. This along with smaller sample sizes would account for
estimates of zero multiple day change for all but the failing fuel-injected Day 1-to-day 2 class.
Last Revised 1/13/99
-------
The large percentage changes estimated for fuel-injected vehicles that pass both pressure
and purge tests are derived from base Day 1 emissions that are considerably smaller than those of
the failing fuel-injected vehicles. However, in practice a straightforward application of these
numbers can lead to projected multiple day evaporative emissions for passing vehicles which
exceed those of vehicles that fail at least one test. For example, suppose Day 1 emissions for a
model year class of fuel injected vehicles passing both tests is modeled as 4 grams. For vehicles
failing at least one test let this value be 6 grams. Applying the growth factors gives the following
estimates for:
Pass Both Tests:
Day 1-to-Day 2: 4 grams * (1+.365) = 5.46 grams
Day 2-to-Day 3: 5.46 grams * (1+.312) = 7.16 grams
Fail One or Both Tests:
Day 1-to-Day 2: 6 grams * (1+.133) = 6.80 grams
Day 2-to-Day 3: 6.80 grams * (1+0) = 6.80 grams
Thus, by Day 3, the passing vehicle class is projected to have higher emissions than the failing
group.
To avoid this anomaly, it is proposed that within a given model year range, the pass/pass
fuel-injected vehicle projections be capped by the projection for failing fuel-injected vehicles in
that group. The cap proposal would set the passing vehicle estimate equal to that of the vehicles
that fail at least one test. Thus, in the illustration above, both passing and failing fuel-injected
vehicle emissions in the third day would be assigned the value 6.80 grams.
For the data employed in this analysis, the problem does not exist if judged using sample
means. The following table shows mean daily HC emissions for the (nine-hour) data in the
categories proposed.
Carbureted
Fuel -Injected
Fail
Day 1
22.12
10.41
Day 2
19.09
10.004
Day 3
19.44
5.94
Pass
Day 1
11.80
3.93
Day 2
12.69
4.78
Day 3
10.71
5.52
4The apparent decrease in the failing fuel-injected mean from Day 1 to Day 2 appears
inconsistent with the finding of a 13.3% rate of increase. This is explained by the fact that the
percent change is derived from the logarithms of individual emission levels, which has a
disproportionate effect larger emission values. For these two subsamples, the means of the
logarithms increase (from 1.59 to 1.76) as expected.
Last Revised 1/13/99
-------
The mean values do not refute the hypothesis that passing vehicles perform better than failing
vehicles, and that fuel-injected vehicles outperform carbureted. While these data suggest that the
differences between categories are more extreme than in the example, they are subject to
sampling variation.
4.2 Temperature and RVP
The models considered above include the covariates Reid vapor pressure and
temperature. The RVP variable was controlled at nominal values of 6.3, 6.8, and 9.0 psi.
Temperature varied over a 24-hour period according to three cycles: 60 to 84 degrees F; 72 to 96;
and 82 to 106. For this analysis, the midpoint values, 72, 84 and 94 degrees, were used. As seen
in Tables 3 to 5, RVP and temperature have high statistical significance. This implies that
emissions are sensitive to the values of these variables. Thus, it is appropriate to include them in
models of real-time diurnal emissions.
For the purpose of MOBILE6 modeling, the somewhat different question has been raised
of whether the adjustments for fuel metering and pressure/purge status are related to RVP and
temperature. To answer this, terms were added to the linear models to account for possible
interactions between the categorical variables and the continuous. A significant interaction
between, say fuel metering and temperature, would suggest that the factor used to adjust for fuel
metering should vary with temperature. In those models, which are not reported here, none of
these interaction terms was found to be statistically significant. Therefore, EPA recommends that
the adjustment scheme described in Section 4.1 be applied without regard for temperature and
RVP.
5.0 CONCLUSION
Day-to-day diurnal evaporative emissions are found to change over the first three days for
several combinations of a vehicle's fuel metering and pressure/purge test status. Temperature and
fuel vapor pressure effects also are evident. Estimates of these changes are proposed for
application in MOBILE6
The MOBILE model distinguishes between resting loss and diurnal evaporative
emissions. The analysis presented here takes a simplified approach, treating resting losses as
constant so that any change from one day to the next is entirely due to the diurnal. The estimated
multiple day diurnal effect is greatest when applied only to 1986-95 fuel injection vehicles that
pass both purge and pressure tests.
Last Revised 1/13/99
-------
Appendix
The results presented in this report hinge on the interpretation of regression coefficients
as measures of percent change in emissions. The mathematics supporting this assumption
follows.
For the emissions function
ln(HC) = b0 + bjD + b^D + b3P2*D + b^D + b5Y2*D + b6F*D + b7R + bgT ,
if we invert the log transformation we get:
HC = exp(b0 + bjD + b^D + b3P2*D + b^^D + b5Y2*D + b6F*D + b7R + bgT )
The change in HC with respect to D is found by differentiating:
dHC/dD = (bj + b2Pj + b3P2 + b4Yx + b5Y2 + b6F) *exp(-)
As a percentage of HC, this is simply the ratio of the last two expressions,
[dHC/dD]/HC = [(bj + b2Pj + b3P2 + b4Yx + b5Y2 + b6F) *exp(-)]/exp(-)
= bj + b2Pj + b3P2 + b4Yj + b5Y2 + B6f
Last Revised 1/13/99
-------
Table 1
Distribution of EPA Vehicles and Tests
MODEL
YEAR
Pre-80
80-85
86-95
FUEL
METERING
CARB
CARB
FI
CARB
FI
PURGE/
PRESSURE
F/P
P/F
P/P
F/P
P/F
P/P
F/P
P/F
P/P
F/P
P/F
P/P
F/P
P/F
P/P
ALL
HOURS
33
VEHS
1
2
1
5
5
4
2
3
1
3
2
17
19
20
85
TESTS
6
12
6
24
19
21
12
12
4
12
6
96
96
88
414
38
VEHS
1
2
1
1
2
7
TESTS
4
8
4
4
8
28
72
VEHS
1
6
1
1
1
16
26
TESTS
4
27
1
6
4
80
122
ALL
VEHS
1
4
1
5
5
8
4
2
3
1
3
3
19
21
38
267
TESTS
6
20
6
24
19
35
21
12
12
4
12
7
106
104
176
736
Last Revised 1/13/99
-------
Table 2(a)
Dependent Variable: LHC
Source
Model
Error
Corrected Total
Day 1 to Day 2 - Full Model
Parameter
Day
FI-F/P
FI-F/F or P/F
CARB-P/P
Pre-80-P/P
80-85-P/P
TEMP
RVP
DF
123
971
1094
R-Square
0.888760
Estimate
Sum of Mean
Squares Square F Value Pr > F
1869.22443 15.19695 63.07 0.0001
233.95714 0.24094
2103.18157
C.V. Root MSB LHC Mean
33.19699 0.49086 1.47863
T for HO: Pr > T
Parameter=0
D Std Error of
Estimate
0.37355527
-0.24896979
-0.21158987
-0.23482963
-0.01467024
-0.08646772
0.06577755
0.31847117
7 .
-3.
-2.
-2.
-0.
-0.
36.
23.
.77
.35
.94
.32
.09
.92
.78
.07
0.
0.
0.
0.
0.
0.
0.
0.
.0001
.0008
.0033
.0203
.9301
.3563
.0001
.0001
0.
0.
0.
0.
0.
0.
0.
0.
. 04810178
.07427266
.07189721
. 10100772
.16723868
.09369706
.00178864
.01380463
Table 2(b): Day 2 to Day 3 - Full Model
Dependent Variable: LHC
Source DF
Model 120
Error 535
Corrected Total 655 1082.122745
R-Square
0.907006
Sum of
Squares
981.491408
100.631337
C.V.
Parameter
Estimate
Day
FI-F/P
FI-F/F or P/F
CARB-P/P
80-85-P/P
TEMP
RVP
0.
-0.
-0.
-0.
0.
0.
0.
.43752226
.44914073
.41435187
.44407966
.06286394
.07473522
.41597033
28.63212
T for HO:
Parameter=0
Mean
Square F Value
8.179095 43.48
0.188096
Root MSB
0.43370
Pr > T
Pr > F
0.0001
LHC Mean
1.51473
Std Error of
Estimate
6.
-1.
-1 .
-0.
0.
36.
26.
.38
.73
.32
.72
.10
.62
.35
0.
0.
0.
0.
0.
0.
0.
.0001
.0842
.1879
.4721
.9199
.0001
.0001
0.
0.
0.
0.
0.
0.
0.
.06857404
.25961712
.31424573
.61716636
.62459983
.00204069
.01578645
Last Revised 1/13/99
10
-------
Table 3(a): Day 1 to Day 2 - Reduced Model
Dependent Variable: LHC
Sum of
Source DF Squares
Model
121
1868.98843
Mean
Square F Value
15 .44619
64.17
Pr > F
0.0001
Error
973
234.19314
0.24069
Corrected Total
1094
2103.18157
Parameter
R-Square
0.888648
Estimate
C.V.
33.17957
T for HO:
Parameter=0
Root MSB
0.49060
Pr > T
LHC Mean
1.47863
Std Error of
Estimate
Day
FI-F/P
FI-F/F or P/F
CARB-P/P
TEMP
RVP
Selected Linear
Combinations
FI-P/P
FI-F/P
FI-F/F or P/F
CARB-P/P
CARB-F/P
CARB-F/F or P/F
0.36562242
-0.25756963
-0.20883320
-0.28232652
0 . 06577721
0.31847302
0.36562242
0.10805279
0.15678922
0.08329590
-0.17427373
-0.12553730
Parameter
F/P vs F/F or P/F
P/P vs F/F or P/F
P/P vs F/P
Estimate
-0.04873643
0.57445562
0.62319205
7 .71
-3.50
-2.93
-3.99
36.79
23.08
7.71
1.79
.69
.15
.20
2.
1.
-2.
-1.67
0.0001
0.0005
0.0035
0.0001
0.0001
0.0001
0.0001
0.0740
0.0073
0.2493
0.0281
0.0951
T for HO:
Parameter=0
-0.61
5.41
5.77
Pr >
0 . 04740431
0.07356632
0.07131000
0.07080813
0.00178770
0.01379739
0.04740431
0.06041319
0.05831582
0.07225962
0.07924244
0.07514976
Std Error of
Estimate
0.5423 0.07995367
0.0001 0.10613121
0.0001 0.10802126
Last Revised 1/13/99
11
-------
Table 3(b): Day 2 to Day 3 - Reduced Model
General Linear Models Procedure
Dependent Variable: LHC
Sum of
Source DF Squares
Model
Error
Corrected Total
119
536
655
R-Square
0.907004
981.489503
100.633242
1082.122745
C.V.
28.60566
Mean
Square F Value
8.247811
0 .187749
Root MSE
0.43330
43.93
Pr > F
0.0001
LHC Mean
1.51473
Parameter
Day
FI-F/P
FI-F/F or P/F
CARB-P/P
TEMP
RVP
Selected Linear
Combinations
FI-P/P
FI-F/P
FI-F/F or P/F
CARB-P/P
CARB-F/P
CARB-F/F or P/F
Estimate
0.43752226
-0.44914073
-0.41435187
-0.38346086
0.07473522
0.41597033
0.43752226
-0
01161846
0.02317039
0.05406140
0.39507932
0.36029047
T for HO:
Parameter=0
6.39
Pr >
Parameter
F/P vs F/F or P/F
P/P vs F/F or P/F
P/P vs F/P
Estimate
-0.03478886
0.85187413
0.88666299
.73
.32
.85
36.66
26.37
-1.
-1.
-2.
6.39
-0.05
0.08
0.47
-1.39
-1.08
0.0001
0.0839
0.1875
0.0045
0.0001
0.0001
0.0001
0.9630
0.9397
0.6408
0.1648
0.2821
T for HO:
Parameter=0
-0.09
2.54
3.11
Pr >
Std Error of
Estimate
0.06851069
0.25937729
0.31395542
0.13455232
0.00203881
0.01577186
0.06851069
0.25016567
0.30638912
0.11580420
0.28405491
0.33463207
Std Error of
Estimate
0.9299 0.39554665
0.0114 0.33563247
0.0020 0.28523275
Last Revised 1/13/99
12
-------
Table 4(a): Collapsed P/F and F/F - Day 1 to Day 2
Dependent Variable: LHC
Sum of
Mean
Source
Model
Error
Corrected Total
Selected Linear
Parameter
FI-PASS
FI-FAIL
GARB -PASS
GARB -FAIL
TEMP
RVP
DF
120
974
1094
R-Square
0.888606
Combinations
Estimate
0.36532558
0.13337025
0.08445850
-0 . 14749683
0.06577568
0.31848139
Squares Square F Value Pr > F
1868.89900 15.57416 64.75 0.0001
234.28257 0.24054
2103.18157
C.V. Root MSB LHC Mean
33.16887 0.49045 1.47863
T for HO: Pr > T
Parameter=0
D Std Error of
Estimate
7.71 0.0001 0.04738652
3.04 0.0024 0.04385692
1.17 0.2424 0.07221113
-2.24 0.0255 0.06593070
36.81 0.0001 0.00178712
23.09 0.0001 0.01379293
PASS vs FAIL
0.59728092
6.02
0.0001
0.09927384
Table 4(b): Collapsed P/F and F/F - Day 2 to Day 3
Dependent Variable: LHC
Sum of
Mean
Source
Model
Error
Corrected Total
Selected Linear
Parameter
FI-PASS
FI-FAIL
GARB -PASS
GARB -FAIL
TEMP
RVP
DF
118
537
655
R-Square
0.907003
Combinations
Estimate
0.43752226
0.00229708
0.05406140
-0.38116378
0 . 07473522
0.41597033
Squares Square F Value Pr > F
981.488051 8.317695 44.38 0.0001
100.634694 0.187402
1082.122745
C.V. Root MSB LHC Mean
28.57922 0.43290 1.51473
T for HO: Pr > r
Parameter=0
r Std Error of
Estimate
6.39 0.0001 0.06844736
0.01 0.9905 0.19359838
0.47 0.6405 0.11569716
-1.62 0.1064 0.23569304
36.69 0.0001 0.00203692
26.40 0.0001 0.01575720
PASS vs FAIL
0 . 87274745
3.68
0.0003
0.23710862
Last Revised 1/13/99
13
-------
Table 5(a): Day 1
Dependent Variable: LHC
to Day 2 using 72 hour data only
Source
Model
Error
Corrected
Parameter
FI-PASS
FI-FAIL
GARB -PASS
GARB -FAIL
TEMP
RVP
Dependent
Source
Model
Error
Corrected
Parameter
FI-PASS
FI-FAIL
GARB -PASS
GARB -FAIL
TEMP
RVP
DF
29
206
Total 235
R-Square
0.876085
Estimate
0.51180929
-0.14496360
0.09150151
-0.56527139
0.07496020
0.36832894
Table 5 (b) : Day 2
Variable: LHC
DF
29
206
Total 235
R-Square
0.918035
Estimate
0.43752226
0.00229708
0.05406140
-0.38116378
0.08776948
0.51871100
Sum of Mean
Squares Square F Value Pr > F
392.266280 13.526423 50.22 0.0001
55.482658 0.269333
447 .748938
C.V. Root MSB LHC Mean
66.04906 0.51897 0.78574
T for HO: Pr > r
Parameter=0
r Std Error of
Estimate
6.24 0.0001 0.08205688
-0.62 0.5329 0.23209192
0.66 0.5102 0.13870145
-2.00 0.0468 0.28255634
18.63 0.0001 0.00402374
11.70 0.0001 0.03148806
to Day 3 using 72 hour data only
Sum of Mean
Squares Square F Value Pr > F
372.666174 12.850558 79.56 0.0001
33.272886 0.161519
405.939060
C.V. Root MSB LHC Mean
35.92281 0.40189 1.11877
T for HO: Pr > r
Parameter=0
r Std Error of
Estimate
6.89 0.0001 0.06354504
0.01 0.9898 0.17973250
0.50 0.6153 0.10741072
-1.74 0.0830 0.21881226
28.17 0.0001 0.00311599
21.27 0.0001 0.02438442
Last Revised 1/13/99
-------
February ??, 1999
Document Released for Stakeholder Review and Comment
Evaluating Multiple Day Diurnal Evaporative Emissions Using RTD
Tests
Report Number EPA420-99-003
The office of Mobile Sources, Assessment and Modeling Division
announces the release of
"Evaluating Multiple Day Diurnal Evaporative Emissions Using RTD
Tests" for stakeholder review and comment. This document EPA420-
99-003 also known as document M6.EVP.003.is available at the
MOBILE6 section of the QMS Web Site
(http://www.epa.gov/oms/m6.htm).
This draft report presents an analysis of diurnal evaporative
emissions from light-duty vehicles (LDVs) and light-duty trucks
(LDTs) occurring over periods of more than one day, using real-
time diurnal (RTD) test data from testing programs performed
under contract for EPA.
The data consists of hourly values of HC emissions (in
grams)measured under varying conditions of fuel Reid vapor
pressure (RVP) and ambient temperature. Daily totals are
obtained directly from these hourly values.
Comments on this report and its proposed use in MOBILE6 should be
sent to the attention of Phil Enns. Comments may be submitted
electronically to mobile@epa.gov, by fax to (734) 214-4821, or by
mail to MOBILE6 Review Comments, US EPA Assessment and Modeling
Division, 2000 Traverwood Drive, Ann Arbor, MI 48105. Electronic
submission of comments is preferred. In your comments please
note clearly the document that you are commenting on including
the report title and the code number listed. Please be sure to
include your name, address, affiliation and any other pertinent
information.
This document is being released and posted on February ??, 1999.
Comments will be accepted for sixty (60) days ending April ??,
1999. EPA will then review and consider all comments received
and will provide a summary of those comments and how we are
responding to them in the form of a follow-up document.
Thank you for your continuing interest in the development of
MOBILE6.
Sincerely,
Emission Inventory Group, Assessment and Modeling Division, US
EPA, Office of Mobile Sources
------- |