EPA420-P-98-018 June 1998 Hot Soak Emissions as a Function of Soak Time Edward L. Glover Assessment and Modeling Division Office of Mobile Sources U.S. Enviromental Protection Agency NOTICE These reports do not necessarily represent final EPA decisions or positions. They are intended to present technical analysis of issues using data which are currently available. The purpose of releasing these 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. ------- 1.0 INTRODUCTION MOBILE6 will compute and report (as a user option) hourly emission factors for start, running, and evaporative emissions. These outputs will be in addition to the standard daily emission estimates which are currently calculated by MOBILES. The hourly emission factors will allow the MOBILE6 model to provide more precise output that accounts for the time of day that vehicle emissions occur. Hot soaks are one type of evaporative emissions which will be reported on an hourly basis. They are, by definition, the evaporative hydrocarbon (HC) emissions which escape from a vehicle during the first hour after the engine is stopped. The limited data in this analysis suggests that the emissions are not distributed evenly throughout the hour, but decline as the hour passes. This is likely due to the cooling of the vehicle and its evaporative system. However, the exact mechanism may include back purge from the canister to the fuel tank, leaks and permeation effects in the lines, and canister breakthrough considerations. The reasons why the emissions occurred and their resulting distribution formed were not investigated in this analysis. This document (M6.EVP.007) presents an analysis of the rate hot soak emissions decline over the one hour time period. This information will be used in MOBILE6 with the hot soak activity information from the document "Soak Length Activity Factors for Hot Soak Emissions" - EPA Report Number M6.FLT.004, and the hot soak emission information from the document "Hot Soak Emissions" (M6.EVP.004). Structurally, this document is divided into three sections. The first section briefly describes the data which were analyzed. The second section discusses the analysis performed on the data. The third section provides the results of the analysis and shows how they will be applied in the MOBILE6 model. 2.0 DATA All of the data used to determine the distribution of hot soak emissions as a function of soak time were collected as part of an EPA study. This study was designed to be similar to a previous Auto / Oil test program (See SAE Paper 951007 "Real World Hot Soak Emissions - A Pilot Study"). In the EPA study, 250 vehicles were recruited and given the standard hot soak test. Only 240 vehicles were used in this analysis since ten of them received emission control system repairs, and were not representative of the general fleet. During the test, the hot soak emissions were measured at 10 minute intervals. Thus, hot soak emission measurements are available at 10, 20, 30, 40, 50 and 60 minute intervals. By definition, the hot soak emissions at time = 0 were assumed to be 0 grams. ------- Other information was collected during the study in addition to the hot soak values. This included vehicle identification information, canister type (open bottom, closed bottom, and unknown), preconditioning prior to the hot soak, and fuel RVP. For more details on the EPA study please refer to the Final Contractor's (ATL) Report of Work Assignment 0-2 of EPA Contract 68-C5-0006 "Real World Hot Soak Evaporative Emissions". 3.0 ANALYSIS 3.1 Hot Soak Fraction All of the hot soak data (hydrocarbon measurements made in a SFLED - sealed housing for emission detection) were collected at 10 minute intervals during the 60 minute hot soak test. For this analysis, these measurements were transformed into emission fractions based on the 60 minute test result. This was done by dividing the individual vehicle hot soak emissions at interval X (i.e., 10 minutes) by the hot soak emissions at the 60 minute test point (the end of the test). By definition, the zero point (t=0) was assumed to have a hot soak fraction of 0, and the end of the test (t=60) was assumed to have a hot soak fraction of 1.0. Figure 1 shows the hot soak fraction data points versus test time duration. The figure indicates that many vehicles quickly reach their maximum hot soak emission value, and emit very little thereafter. Considerable scatter is evident in the figure, particularly during the early portions of the hot soak test. The reduced scatter at the end of the test as evidenced in Figure 1, and the slightly concave pattern of the mean hot soak fractions in Figure 2, suggest that typically hot soak emissions do not occur at or near the end of the 60 minute test. 3.2 Linear Regressions A linear regression was performed on the data to determine if test duration, canister type, preconditioning prior to the hot soak, and fuel RVP significantly affect the results. The regression results in Appendix A show that test duration and canister type are statistically significant variables. Preconditioning and fuel RVP were not found to be significant at a 95% confidence level; although fuel RVP was significant at a 90% confidence level. Despite its significance, canister type was eliminated from the analysis because it is not a variable which will be present in MOBILE6. Thus, analysis by canister type would prevent the functional relationship from being used in the model. Fleet canister type information is difficult to obtain, making an accurate default or user defined MOBILE6 input impractical. ------- 3.3 Quadratic Fit Based on Figures 1 and 2, it was decided that a quadratic fit of the hot soak fraction versus soak time would produce a model which would adequately fit the data, and be simple to implement in the MOBILE6 model. The quadratic fit was obtained from a least squares regression of hot soak fraction versus soak time with the regression intercept fixed at zero. Fixing the zero point was done because the hot soak fraction is defined to be zero when the soak time is zero. The regression statistics are shown in Appendix A. The regression equation is: HSFract = 0.0258 * Soaktime + 0.000156 * Soaktime Eqn 1 4.0 Using the Hot Soak Fraction in MOBILE6 The hot soak fractions developed in this analysis will be used in conjunction with basic hot soak emission values and hot soak activities to predict hourly hot soak emission rates in MOBILE6. 4.1 Basic Hot Soak Emission Value from MOBILE6 For this illustration the basic hot soak emission value is assumed to be X. It is based on testing and is an average result which reflects the entire 60 minute hot soak test. The actual hot soak value will be calculated by the model based on the specific characteristics of the vehicle class and model year. This calculation will be described in EPA document M6.EVP.004. 4.2 Hot Soak Fraction The variable HS Fraction shown in Equation 1 is the cumulative hot soak fraction at a given time. For example using Equation 1, at a soaktime of 30 minutes, 59.8% of the hot soak emissions have been emitted. 4.3 Hot Soak Activity Equation 2 is the general equation for hot soak activity. It can be found in the EPA document "Soak Length Activity Factors for Hot Soak Emissions" - EPA Report Number M6.FLT.004. This equation calculates the cumulative hot soak activity at a give time during the test. ------- Activity(soaktime) = bl - b2 * exp( -b3 * soaktime**b4 ) Eqn 2 For example, substituting the coefficients for the 9 to 10 AM Weekday curve estimates that 44.1 percent of all hot soaks have a soak time of 30 minutes or less. Activity = 3212.9 - 32.712*exp[4.589 * 30 **-0.001003] However, a cumulative hot soak activity is not used in the MOBILE6 model. Instead, the cumulative distribution is broken into one minute intervals, and the amount of activity for each interval is calculated. This is done by subtracting the previous activity value from the current value. Mathematically this is: Interval Activity(t) = Activity(t) - Activity(t-1) Eqn 3 Where t is the time from 1 to 60, and "Interval Activity" represents the fraction of hot soaks with a soak time between t and t-1. 4.4 Calculating the Hourly Hot Soak Emissions In the MOBILE6 model, the hot soak emission and activity distributions will be calculated in one minute intervals ranging from t=0 to t=60 minutes. The overall hot soak emission (X), the cumulative HS Fraction and the Interval Activity parameters will be calculated by weighting the three pieces together and summing the product to produce an overall hourly hot soak emission result. The hot soak emission value for an individual interval at time = t is calculated by multiplying the cumulative HS Fraction at time = t, and the Interval Activity at time = t, and the mean hot soak 60 minute test emission value (X) together using Equation 5. Interval Hot Soak© = Cumulative HS Fraction(t) * Interval Activity(t)*X Eqn 5 Where X is the average hot soak emissions for a 60 minute test. The hourly hot soak emissions are calculated by summing the 60 individual Interval Hot Soak(t) values. Hourly Hot Soak = SUM[Interval Hot Soak(t)] Eqn 6 where t ranges from 0 to 60 minutes. A sample calculation spreadsheet illustrating the hot soak calculation for the 9 to 10 5 ------- AM weekday case is shown in Appendix B. In the example calculation, the one hour average hot soak (X) is assumed to be 100 grams. This is not the actual value which will be used in MOBILE6. It is shown as a round number (100) for illustration only. For this example when 100 grams is used for the overall average hot soak after 60 minutes, the resulting hourly soak with emission fraction and activity fraction weighting is 63.179 grams. ------- 1.0' .8' o .6. o 03 LL .4' ^ 05 <£ .2< 0 ^ 0.0, c Figure 1 Fraction of HS Emissions vs Test Length ) 10 20 30 40 50 60 TIME Figure 2 Mean HS Emission Fraction 1 .6' '*= O 03 iV 4' ^: 03 0 „ C/) 2' 0 I 00, =•= =!= 0. 10. 20. 30. 40. 50. 60. TIME Mean T Upper 95% CL 1 Low95%CL a MEAN ------- Appendix A Statistical Results -> USE ALL. -> COMPUTE filter_$=(test_seq = 17). -> VARIABLE LABEL filter_$ 'test_seq = 17 (FILTER)1. -> VALUE LABELS filter_$ 0 'Not Selected' 1 'Selected' -> FORMAT filter_$ (fl.O). -> FILTER BY filter_$. -> EXECUTE . -> EXAMINE -> VARIABLES=hs_perc BY time -> /PLOT NONE -> /STATISTICS DESCRIPTIVES -> /CINTERVAL 95 -> /MISSING LISTWISE - > /NOTOTAL. HS_PERC By TIME Valid cases: 10 240 .0 Missing cases: Percent missing: .0 Mean .2659 Std Err .0120 Min Median .2100 Variance .0343 Max 5% Trim .2441 Std Dev .1853 Range 95% CI for Mean (.2423, .2894) IQR .0000 Skewness 2.0959 1.0200 S E Skew .1571 1.0200 Kurtosis 4.5601 .1300 S E Kurt .3130 HS_PERC By TIME Valid cases: 20 240 .0 Missing cases: Percent missing: .0 Mean .4687 Std Err .0104 Min Median .4300 Variance .0260 Max 5% Trim .4549 Std Dev .1613 Range 95% CI for Mean (.4482, .4892) IQR .1800 Skewness 1.4434 1.0000 S E Skew .1571 .8200 Kurtosis 1.9371 .1400 S E Kurt .3130 HS_PERC By TIME Valid cases: 30 240 .0 Missing cases: Percent missing: .0 Mean .6276 Std Err .0083 Min Median .5950 Variance .0164 Max 5% Trim .6198 Std Dev .1282 Range 95% CI for Mean (.6113, .6439) IQR .3100 Skewness 1.0080 1.0600 S E Skew .1571 .7500 Kurtosis 1.0257 .1300 S E Kurt .3130 HS_PERC By TIME Valid cases: 40 240 .0 Missing cases: Percent missing: .0 ------- Mean .7686 Std Err .0059 Min Median .7500 Variance .0085 Max 5% Trim .7649 Std Dev .0919 Range 95% CI for Mean (.7569, .7803) IQR .5300 Skewness .8471 1.1600 S E Skew .1571 .6300 Kurtosis 1.2231 .1075 S E Kurt .3130 HS_PERC By TIME Valid cases: 50 240 .0 Missing cases: Percent missing: .0 Mean .8853 Std Err .0033 Min Median .8800 Variance .0026 Max 5% Trim .8849 Std Dev .0514 Range 95% CI for Mean (.8788, .8919) IQR .6700 Skewness .0098 1.0000 S E Skew .1571 .3300 Kurtosis 1.5568 .0650 S E Kurt .3130 HS_PERC By TIME Valid cases: 60 240 .0 Missing cases: .0 Percent missing: .0 >Note # 17570. Command name: EXAMINE >The number of unique data values for this cell is equal to one. The cell >will be included in any boxplots produced but other output will be omitted. ------- -> USE ALL. -> COMPUTE filter_$=(test_seq = 17). -> VARIABLE LABEL filter_$ 'test_seq = 17 (FILTER)'. -> VALUE LABELS filter_$ 0 'Not Selected' -> FORMAT filter_$ (fl.O). -> FILTER BY filter_$. -> EXECUTE . -> REGRESSION -> /MISSING LISTWISE -> /STATISTICS COEFF OUTS CI R ANOVA -> /CRITERIA=PIN(.05) POUT(.10) -> /NOORIGIN -> /DEPENDENT hs_perc -> /METHOD=ENTER time can fuel_rvp precond -> /RESIDUALS HIST(ZRESID) NORM(ZRESID) . 'Selected' * * * * MULTIPLE REGRESSION * * * * Listwise Deletion of Missing Data Equation Number 1 Dependent Variable.. HS_PERC Block Number 1. Method: Enter TIME CAN FUEL RVP PRECOND Variable(s) Entered on Step Number 1.. PRECOND 2 . . TIME 3.. CAN 4 . . FUEL RVP Multiple R .89825 R Square .80686 Adjusted R Square .80632 Standard Error .12165 Analysis of Variance DF Regression 4 Residual 1435 Sum of Squares 88.71651 21 .23650 Mean Square 22.17913 .01480 F = 1498 .69575 Signif F = .0000 Variable TIME CAN FUEL_RVP PRECOND (Constant) Variables in the Equation 3 SE B 95% Confdnce Intrvl B .014461 - .020136 -3 .19716E-04 - .004923 .233692 1.8771E-04 .014093 .002674 -.025381 1.7195E-04 -6.57019E-04 .008094 -.020800 .014555 .205140 .014830 - .014891 1.75872E-05 .010954 .262244 Beta .893783 -.087556 - .021644 -.007070 in Variable T Sig T 10 ------- TIME 77.041 .0000 CAN -7.531 .0000 FUEL_RVP -1.859 .0632 PRECOND -.608 .5431 (Constant) 16.055 .0000 **** MULTIPLE REGRESSION **** Equation Number 1 Dependent Variable.. HS_PERC End Block Number 1 All requested variables entered. Residuals Statistics: Min Max Mean Std Dev N *PRED .2560 1.0746 .6694 .2483 1440 *RESID -.3512 .7089 .0000 .1215 1440 *ZPRED -1.6649 1.6319 .0000 1.0000 1440 *ZRESID -2.8872 5.8271 .0000 .9986 1440 Total Cases = 1440 Hi-Res Chart # 13:Histogram of *zresid Hi-Res Chart # 14:Normal p-p plot of *zresid 11 ------- -> GET -> FILE='D:\MOBILE6\EVAP\EMIS\AOH95_2.SAV -> EXECUTE . -> * Curve Estimation. -> TSET NEWVAR=NONE . -> CURVEFIT /VARIABLES=hs_perc WITH time -> /NOCONSTANT -> /MODEL=QUADRATIC -> /PRINT ANOVA -> /PLOT FIT. MODEL: MOD 8. Dependent variable.. HS_PERC Listwise Deletion of Missing Data Multiple R .98589 R Square .97197 Adjusted R Square .97194 Standard Error .12122 Analysis of Variance: DF Sum of Squares Regression Residuals 2 1498 763 .37082 22.01098 Method.. QUADRATI Mean Square 381 .68541 .01469 F = 25976.34126 Variable TIME TIME**2 Signif F = .0000 Variables in the Equation B SE B Beta .025772 .000326 -.000156 6.5297E-06 T Sig T 1.387091 78.938 .0000 -.419536 -23.875 .0000 Notes: * Equation was estimated without the constant term; Rsq is redefined. Notes: * Equation was estimated without the constant term; Rsq is redefined. 12 ------- Appendix B - Sample Calculation Soak time 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Activity % 0.0000 4.4196 10.3965 14.6290 17.9073 20.5828 22.8427 24.7988 26.5230 28.0644 29.4580 30.7295 31.8987 32.9808 33.9877 34.9293 35.8135 36.6468 37.4348 38.1822 38.8929 39.5704 40.2176 40.8370 41.4311 42.0017 42.5507 43.0796 43.5898 44.0826 44.5592 45.0205 45.4676 45.9013 46.3223 46.7314 47.1292 47.5164 47.8934 48.2609 48.6192 48.9689 49.3102 49.6437 49.9696 50.2884 50.6002 50.9054 51.2044 51.4972 51.7842 52.0656 52.3416 52.6124 52.8783 53.1393 53.3957 53.6475 53.8951 54.1385 100.0000 Activity Fract 0.0000 0.0442 0.1040 0.1463 0.1791 0.2058 0.2284 0.2480 0.2652 0.2806 0.2946 0.3073 0.3190 0.3298 0.3399 0.3493 0.3581 0.3665 0.3743 0.3818 0.3889 0.3957 0.4022 0.4084 0.4143 0.4200 0.4255 0.4308 0.4359 0.4408 0.4456 0.4502 0.4547 0.4590 0.4632 0.4673 0.4713 0.4752 0.4789 0.4826 0.4862 0.4897 0.4931 0.4964 0.4997 0.5029 0.5060 0.5091 0.5120 0.5150 0.5178 0.5207 0.5234 0.5261 0.5288 0.5314 0.5340 0.5365 0.5390 0.5414 1.0000 Delta Act Fract 0.0442 0.0598 0.0423 0.0328 0.0268 0.0226 0.0196 0.0172 0.0154 0.0139 0.0127 0.0117 0.0108 0.0101 0.0094 0.0088 0.0083 0.0079 0.0075 0.0071 0.0068 0.0065 0.0062 0.0059 0.0057 0.0055 0.0053 0.0051 0.0049 0.0048 0.0046 0.0045 0.0043 0.0042 0.0041 0.0040 0.0039 0.0038 0.0037 0.0036 0.0035 0.0034 0.0033 0.0033 0.0032 0.0031 0.0031 0.0030 0.0029 0.0029 0.0028 0.0028 0.0027 0.0027 0.0026 0.0026 0.0025 0.0025 0.0024 0.4586 Emission 0.0256 0.0510 0.0760 0.1007 0.1251 0.1492 0.1730 0.1964 0.2196 0.2424 0.2649 0.2871 0.3090 0.3306 0.3519 0.3729 0.3935 0.4139 0.4339 0.4536 0.4730 0.4921 0.5109 0.5293 0.5475 0.5653 0.5829 0.6001 0.6170 0.6336 0.6499 0.6659 0.6815 0.6969 0.7119 0.7266 0.7410 0.7551 0.7689 0.7824 0.7956 0.8084 0.8210 0.8332 0.8451 0.8567 0.8680 0.8790 0.8896 0.9000 0.9100 0.9198 0.9292 0.9383 0.9471 0.9556 0.9638 0.9716 0.9792 1.0000 HS Averaae 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Wt Emiss 0.1133 0.3047 0.3217 0.3301 0.3347 0.3372 0.3383 0.3387 0.3384 0.3378 0.3369 0.3357 0.3344 0.3329 0.3313 0.3297 0.3279 0.3261 0.3243 0.3224 0.3204 0.3185 0.3165 0.3145 0.3124 0.3104 0.3083 0.3062 0.3041 0.3020 0.2998 0.2977 0.2955 0.2934 0.2912 0.2891 0.2869 0.2847 0.2825 0.2804 0.2782 0.2760 0.2738 0.2716 0.2694 0.2672 0.2649 0.2627 0.2605 0.2583 0.2561 0.2539 0.2516 0.2494 0.2472 0.2450 0.2428 0.2405 0.2383 45.8615 63.1797 13 ------- |