United States Air and Radiation EPA420-P-99-029 Environmental Protection November 1999 Agency M6.EXH.006 vvEPA Determination of Methane Offsets as a Function of Mileage for Light-Duty Cars and Trucks > Printed on Recycled Paper ------- EPA420-P-99-029 November 1999 of as a of for Phil Enns 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 - Determination of Methane Offsets as a Function of Mileage for Light-Duty Cars and Trucks Report Number M6.EXH.006 September 1999 Phil Enns U.S. EPA Assessment and Modeling Division 1.0 Introduction Previous reports document how the MOBILE6 emission factor model will allocate vehicle exhaust emissions between engine start (start emissions) and travel (running emissions)1'2. This split allows the separate characterization of start and running emissions for correction factors such as fuel effects and ambient temperature. It also enables a more precise weighting of these two aspects of exhaust emissions for particular situations such as morning commute, parking lot and freeway driving. Because methane does not contribute significantly to ozone formation, MOBILE attempts to separate it from total hydrocarbons in estimating vehicle emissions. This document describes methodologies for calculating this methane "offset" based on the separation of start and running emissions proposed for MOBILE6. The procedure follows the scheme of modeling the in-use deterioration of emissions as a function of accumulated mileage. For model year 1981-1993 light-duty cars and trucks, stratified into carbureted versus fuel-injected, the method parallels that used for total hydrocarbons, carbon monoxide and oxides of nitrogen. This analysis utilizes actual methane data from emissions tests conducted on vehicles from those model years. For pre-1981 model years, data of the type used in the newer vehicles is not available. Therefore, procedures are described for estimating methane as a function of JCarey, P., P. Enns, E. Glover, and M. Sklar, "Determination of Running Emissions as a Function of Mileage for 1981-1993 Model Year Light-Duty Cars and Trucks," Report Number M6.EXH.001, October 1998. 2Carey, P. and E. Glover, "Determination of Start Emissions as a Function of Mileage and Soak Time for 1981-1993 Model Year Light-Duty Vehicles," Report No. M6.STE.003, October 1998. Methane -8- 11/8/99 ------- mileage using existing data in combination with methods that are applied in MOBILE version 5. 2.0 Data The data underlying the analysis of 1981-93 light-duty vehicles are drawn from a subset of the Federal Test Procedure (FTP) tests described in the reports cited above. These tests were conducted by EPA, the American Automobile Manufacturers Association (AAMA), and the American Petroleum Institute (API). Most, but not all, of these tests produced measurements of methane. In particular, the sample sizes from the full data set and the reduced sets for which methane is recorded are compared in the table below, subdivided by vehicle type and the model year/technology groups used to determine basic emission rates of total HC, CO and NOx. CARS GROUP 81-82 GARB 81-82 FI 83-85 GARB 83-87 FI 86-93 GARB 88-93 PFI 88-93 TBI DATA REDUCED N 580 88 203 688 93 1361 437 SET FULL N 1166 126 253 726 96 1605 444 GROUP 81-83 GARB 81-87 FI 84-93 GARB 88-93 PFI 88-93 TBI TRUCKS DATA REDUCED N 72 92 125 199 458 SET FULL N 180 94 134 330 467 Another key data set is that based on a sample of FTP tests to which were appended a 505-second cycle without an engine start. This cycle is identical to that of Bags 1 and 3, but contains no emissions associated with the cold start of Bag 1 or warm start of Bag 3. It is referred to as the Hot Running 505 (HR505). Pure cold and warm start emissions are estimated by deducting HR505 emissions from the 505 bags that include a start.3 The data from this test program were used to estimate the relation between the 3Brzezinski, D. and P. Enns, "The Determination of Hot Running Emissions from FTP Bag Emissions", Report No. M6.STE.002, December, 1997. Methane 11/8/99 ------- HR505 and Bags 1 to 3 of an FTP. From this function, the portions of FTP emissions attributed to start and running are computed. This calculation then was applied to the large FTP data set described above for which Bags 1 to 3, but not the FIR505, are measured. 3.0 Modeling Methane Deterioration in MOBILE6 - 1981-93 Vehicles To understand the current analysis, it is helpful to review how emissions deterioration is modeled in MOBILE6. The basic method involves separating start and running emissions. 3.1 Running Emissions For the running component, simple linear functions of emissions versus mileage are fitted by the method of least squares regression. To improve the fit at low mileage, the mean of emissions is used. This produced piecewise continuous functions in which emission rates (in grams per mile) are constant at low mileage and increase when mileage accumulation exceeds approximately 20,000 miles. The higher mileage portion of the function has constant slope in most cases, but under certain conditions the slope may change, adding a second "corner" point to the graph of emissions as a function of mileage. There was a concern that the FTP data suffer from sample bias due to the self-selecting nature of the data collection. In order to account for such bias, adjustments were made to these running emission lines using data from a large sample of inspection and maintenance tests conducted in Dayton, Ohio. (See document (1) for details.) However, methane was not recorded in these tests, so it was not possible to compute this adjustment for the analysis described in this report. Instead, we propose using a proportional adjustment for sample bias corresponding to that used for total HC. Table 1 reports the deterioration coefficients derived for methane using the reduced data set. The first slope is always zero, reflecting the use of the low mileage mean (ZML Emissions). The first corner occurs at the mileage where the emission rate begins to slope positively. For total HC, only the 1983-87 fuel-injected car category has a second corner. With methane, several categories have two corners, while the 1983-85 carbureted car group has zero deterioration at all mileages. The high emitter adjustment was determined as proportionately the same at a given mileage as for total HC. These adjustment factors are applied additively to the initial THC emissions. They raise or lower THC by an amount that is a linear function of mileage and is zero at mileage zero. In order to apply this factor to methane it is necessary to first compute unadjusted and adjusted THC at a given mileage. Their ratio is then multiplied by unadjusted methane to obtain adjusted methane. Figure 1 illustrates the unadjusted and adjusted methane deterioration lines for the case of 1988 to 1993 model year port fuel-injected cars. As described, the adjustment in methane is proportional to the corresponding HC adjustment, which also is illustrated. Figure 2 shows the final adjusted methane lines by model year-technology group for cars and trucks. Methane -8- -3- 11/8/99 ------- 3.2 Start Emissions Deterioration of start emissions (measured in grams) was modeled somewhat differently. First, the tested vehicles were stratified into categories of "normal" and "high" emitters depending on how their FTP values compare to applicable standards. For HC and CO, the normal emitter emissions were regressed against mileage, while high emitter emissions were fitted by the mean. At a given mileage, the normal and high emitter estimates were combined in a weighted average, where the weight reflects the proportion of high emitters at that mileage. Reference (2) gives a complete description of this procedure. Table 2 presents coefficients used to compute the start portion of emission rates at a given mileage. The majority of the groups actually show negative slope estimates for the normal emitter regression lines. However, in all but one of these, the slope is not statistically significant when tested against zero. Since negative deterioration is intuitively unexpected, it was decided to use the normal emitter mean value when the regression estimate is negative. It also is possible for estimated cold start emissions to be negative. This occurs when the fitted Hot Running 505 value exceeds the observed Bag 1 value. In the case of 1988-93 TBI cars, this produced an anomaly in which the mean methane of the high emitters is negative and the start emission deterioration line has negative slope. For this case, the deterioration line was assigned a constant value equal to its zero mile level. Table 3 gives the fraction of high emitters estimated for THC which is used to compute the weighted average described above. These are the same values used for the full sample total HC calculations. Graphs of start emissions versus mileage for cars and trucks appear in Figure 3. 3.3 FTP Emissions Final FTP deterioration functions are obtained by combining the start and running estimates according to their relative importance in that test. The weights used coincide with those used for HC, CO and NOx. They form the equation FTP=(7.5*Run + .43*CS + .57*HS)/7.5 where Run is emissions in grams per mile from the running LA4 portion; CS is emissions in grams from cold start; and HS is hot start, computed as simply 0.16*CS. The factor 7.5 is the driving mileage corresponding to Bags 1 and 2 (the LA4 cycle) in the FTP. 4.0 Other Model Years 4.1 Pre-1981 Model Years In MOBILES, start and running emissions are not separated. MOBILES contains deterioration functions for methane for each of the three FTP bags. This enables the Methane -8- -4- 11/8/99 ------- calculation of bag values at selected vehicle mileages. These values can in turn be used to compute running and start emissions when combined with the regression coefficients derived from the 77-car test data. The validity of this approach rests on the assumption that start and running emissions in older vehicles occur in a manner similar to that of late model vehicles. For carbureted cars, graphs of running and start emissions are shown in Figure 4. The graphs include the proposed MOBILE6 functions for 1981-82 and for 1983-85. It is apparent that the deterioration functions for the two versions of the model are reasonably compatible. Similar graphs for fuel injected cars and for trucks yield the same impression. Therefore, in the absence of new data, EPA proposes to separate start and running methane emissions in pre-1981 light-duty vehicles and trucks using the methodology of MOBILE6 applied to the deterioration functions employed in MOBILES. For diesel vehicles and motorcycles, FTP bag values are not provided in MOBILES. Thus, it is not possible to compute running and start emissions by the method outlined above. In these cases, EPA proposes to use fractions described in a separate report4 that are applied to total hydrocarbon deterioration functions to determine methane function coefficients. 4.2 Model Years 1994 and Later For model year 1994 and later vehicles subject to Tier I standards, MOBILE6 predicts nonmethane hydrocarbon (NMHC) emissions directly. In the case of LEVs, the model predicts nonmethane organic gas (NMOG). No adequate methane data exist for estimating methane deterioration functions for these model years. Therefore, EPA proposes developing multiplicative methane offsets from these projections in a manner consistent with the treatment of NMHC and NMOG. In general, emission rates for those pollutants are obtained using ratios of standards that apply to the newer vehicles to those of 1993 model year vehicles. For a given class of vehicles, there exists a ratio between NMHC (or NMOG) in the two model year ranges. The proposed method would apply the same ratio to compute methane for newer vehicles. 5.0 Summary EPA proposes combining available methane data with methods employed for methane in MOBILES and total hydrocarbons in MOBILE6 to develop basic emission rates for methane in MOBILE6. This results in an eclectic approach that attempts to fully utilize the data while making reasonable assumptions where insufficient data exists. In general, the assumptions are consistent with methods applied to total and nonmethane HC. 4Brzezinski, D. and J. Gilmore, "Basic Exhaust Emission Rates of Open Loop Vehicles forMOBILE6," Report No. M6.EXH.005, May 1999. Methane -8- 11/8/99 ------- Table 1: Methane Running Emission Deterioration Model Coefficients for Light-Duty Vehicles Cars Model Year/ Technology 88-93 PFI 88-93 TBI 83-87 FI 86-93 CARB 83-85 CARB 81-82 FI 81-82 CARB ZML Emission (gr/m) 0.0167 0.0240 0.0365 0.0405 0.0721 0.0271 0.0845 First Slope (gr/m/lOOOm) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 First Corner (1000 miles) 15.47 32 .18 14.12 15.19 N/A 13 .92 22.11 Second Slope (gr/m/lOOOm) 0.0003 0.0002 0.0006 0.0002 N/A 0.0005 0.0005 Second Corner (1000 miles) 67.89 N/A 81.29 71.91 N/A 265.40 N/A Third Slope (gr/m/lOOOm) 0.0003 N/A 0.0004 0.0001 N/A 0.0005 N/A Trucks Model Year/ Technology 88-93 PFI 88-93 TBI 84-93 CARB 81-87 FI 81-83 CARB ZML Emission (gr/m) 0.0291 0.0253 0.1118 0.0594 0.1033 First Slope (gr/m/lOOOm) 0.0000 0.0000 0.0000 0.0000 0.0000 First Corner (1000 miles) 19.18 16 .25 36.51 29.76 12.35 Second Slope (gr/m/lOOOm) 0.0005 0.0004 0.0008 0.0006 0.0002 Second Corner (1000 miles) N/A 54 .46 N/A N/A 80.35 Third Slope (gr/m/lOOOm) N/A 0.0003 N/A N/A 0.0001 Methane 11/8/99 ------- Table 2: Methane Start Emission Deterioration Model Coefficients for Light-Duty Vehicles Cars GROUP 88-93 PFI 88-93 TBI 86-93 CARB 83-87 FI 83-85 CARB 81-82 FI 81-82 CARB ZML (gr/m) 0.102 0.084 0.108 0.115 0.174 0.077 0.177 SLOPE (gr/m/lOOOm) -0.0002 -0.0003 -0.0001 -0.0004 -0.0001 0.0008 0.0007 MEAN METHANE HIGHS (gr/m) 0.178 -0.073 0.242 0.151 0.601 0.335 0.551 MEAN METHANE NORMALS (gr/m) 0.095 0.071 0.105 0.097 0.172 0.116 0.211 Trucks GROUP 88-93 PFI 88-93 TBI 84-93 CARB 81-87 FI 81-83 CARB ZML (gr/m) 0.151 0.148 0.302 0.098 0.423 SLOPE (gr/m/lOOOm) -0.0022 -0.0002 -0.0001 -0.0000 -0.0001 MEAN METHANE HIGHS (gr/m) 0.556 0.183 0.686 MEAN METHANE NORMALS (gr/m) 0.050 0.140 0.297 0.098 0.417 Methane 11/8/99 ------- Table 3: Fraction of High Emitters for Light-Duty Cars • MILES (xlOOO) 2 .142 12.823 29.335 50 60.006 74.239 87.786 100.01 112 .948 124.625 135.738 146 .315 156 .38 165.96 175.077 183 .753 192 .01 199.869 207.349 214 .466 221.241 227.688 233.823 239.663 245.22 250.509 81-82 CARB 0.0282 0.0543 0.1580 0.2906 0.3560 0.4503 0.5416 0.6253 0.7152 0.7976 0.8772 0.9539 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 81-82 FI • 0.0203 0.0654 0.1613 0.2861 0.3485 0.4393 0.5275 0.6094 0.6986 0.7812 0.8620 0.9407 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 83-85 CARB • 0.0232 0.0158 0.0047 0.0917 0.1348 0.1972 0.2578 0.3135 0.3737 0.4290 0.4826 0.5345 0.5847 0.6332 0.6801 0.7253 0.7690 0.8111 0.8516 0.8907 0.9284 0.9646 1.0000 1.0000 1.0000 1.0000 GROUP 83-87 FI • 0.0223 0.0157 0.0406 0.1003 0.1298 0.1723 0.2078 0.2346 0.2634 0.2898 0.3153 0.3400 0.3638 0.3868 0.4089 0.4303 0.4508 0.4706 0.4896 0.5079 0.5255 0.5425 0.5587 0.5743 0.5893 0.6036 86-93 CARB • 0.0052 0.0197 0.0526 0.1042 0.1296 0.1661 0.2012 0.2334 0.2678 0.2992 0.3295 0.3586 0.3866 0.4135 0.4393 0.4641 0.4879 0.5108 0.5327 0.5537 0.5738 0.5931 0.6116 0.6293 0.6462 0.6624 88-93 PFI • 0.0184 0.0227 0.0422 0.0800 0.0987 0.1260 0.1525 0.1770 0.2036 0.2280 0.2518 0.2748 0.2972 0.3189 0.3398 0.3601 0.3798 0.3988 0.4171 0.4348 0.4519 0.4683 0.4842 0.4994 0.5141 0.5283 88-93 TBI • 0.0239 0.0251 0.0270 0.0386 0.0458 0.0561 0.0661 0.0753 0.0851 0.0940 0.1026 0.1110 0.1190 0.1267 0.1341 0.1412 0.1480 0.1546 0.1609 0.1669 0.1727 0.1782 0.1836 0.1887 0.1936 0.1982 (Cont.) Methane -10- 11/8/99 ------- Table 3: Fraction of High Emitters for Light-Duty Trucks MILES (xlOOO) 2 .142 12.823 29.335 45.05 60.006 74.239 87.786 100.678 112 .948 124.625 135.738 146 .315 156 .38 165.96 175.077 183 .753 192 .01 199.869 207.349 214 .466 221.241 227.688 233.823 239.663 245.22 250.509 GROUP 81-83 CARB| 81-87 FI | 84-93 CARB | 88-93 PFI 0.0500 0.0405 0.0733 0.1181 0.1637 0.2101 0.2573 0.3052 0.3538 0.4030 0.4528 0.5032 0.5540 0.6052 0.6568 0.7087 0.7608 0.8132 0.8656 0.9181 0.9706 1.0000 1.0000 1.0000 1 . 0000 0.0659 0.0515 0.0283 0.0370 0.0883 0.1393 0.1900 0.2405 0.2905 0.3401 0.3891 0.4376 0.4855 0.5328 0.5793 0.6251 0.6702 0.7144 0.7578 0.8004 0.8420 0.8827 0.9225 0.9614 0.9993 0.0000 0.0000 0.0000 0.0000 0.0164 0.0697 0.1230 0.1761 0.2291 0.2819 0.3343 0.3864 0.4381 0.4892 0.5399 0.5899 0.6393 0.6881 0.7361 0.7833 0.8297 0.8753 0.9200 0.9638 1.0000 0.0301 0.0294 0.0356 0.0546 0.0734 0.0919 0.1102 0.1281 0.1458 0.1631 0.1801 0.1968 0.2131 0.2291 0.2446 0.2599 0.2747 0.2891 0.3032 0.3169 0.3302 0.3431 0.3556 0.3677 0.3795 88-93 TBI 0.0094 0.0066 0.0202 0.0381 0.0555 0.0721 0.0884 0.1043 0.1198 0.1349 0.1497 0.1641 0.1781 0.1918 0.2050 0.2178 0.2303 0.2424 0.2541 0.2655 0.2764 0.2870 0.2973 0.3072 0.3167 1.0000 1.0000 1.0000 0.3909 0.3259 Methane -10- 11/8/99 ------- (g/mfl 0.75 0.50- 025 0.00 figure 1: METHANE and THC vs. MILEAGE, RUNNING LA4, PFI CARS 50 100 MILES (xlOOO) 150 Methane UNADJUSTED THC UNADJUSTED METHANE -10- ADJUSTED THC ADJUSTED METHANE 200 11/8/99 ------- 0.0 Figure 2(a): METHANE vs. MILEAGE, RUNNING LA4 CARS MILES (xlOOO) Methane 81-82 GARB 86-93 GARB -10- 81-82 Fl 88-93 PR 83-85 GARB 1/8/99 83-87 Fl ------- Figure 2(b): METHANE vs. MILEAGE, RUNNING LA* TRUCKS (g/mi) 025 020 0.15 0.10 0.05 0.00 \ ' ' ' ' ' ^ 0 81 -83 GARB i 50 81-87 Fl 100 MILES (X1000) 84-93 GARB Methane -10- 150 -93 PR 11/8/99 200 1-93 TBI ------- grams 0.6 Rgure 3(a): METHANE vs. MILEAGE, START EMISSIONS, CARS 0.4 0.0 50 100 MILEAGE 150 200 81 -82 GARB 86-93 GARB Methane 81-82 R 88-93 PR -10- 83-85 GARB 88-93 TH • 83-87 R 11/8/99 ------- grams 0.75 0.50 0.25 Rgure 3(b): METHANE vs. MILEAGE, START EMISSIONS, TRUCKS 0.00 81-83 GARB Methane i 50 81-87 Fl 100 MILEAGE 84-93 GARB -10- 150 88-93 PR 11/8/99 200 88-93 TBI ------- 0.4 Figure 4(a): MOBILES and MOBILES METHANE DETERIORATION CARBURETED CAR RUNNING EMISSIONS (g/mQ 0.3 0.2 MY Methane MILES 0(1000) M51975-79 M5PRE-1975 M51980 M61981-82 10- M5:198l-82 M61983-85 M51983+ 11/8/99 ------- 0.2 0.0 MY Methane Figure 4(b): MOBILES and MOBILES METHANE DETERIORATION CARBURETED CAR START EMISSIONS (grams) 50 M51975-79 M5PRE-1975 100 MILES (XIOOO) 150 M5:1980 M6:1981-82 10- M5:19B1-82 M6:1983-85 200 M5:1983+ 11/8/99 ------- Methane -10- 11/8/99 ------- |