EPA-AA-CORR 76-2 Ford-EPA Emission Laboratory Correlation Study April 1976 Environmental Protection Agency Office of Air and Waste Management Office of Mobile Source Air Pollution Control Emission Control Technology Division ' Standards Development and Support Branch Ann Arbor, Michigan ------- Abstract A specific emissions correlation program between the EPA Motor Vehicle Emissions Laboratory and the Ford Motor Company AEO facility has been completed. This report summarizes emission and cross check results for this program. Examination of the Ford mass simulator results, gas cross check results, and emission and fuel economy comparisons do not indicate a serious correlation problem exists between laboratories. ------- Introduction Analysis of paired 1977 model year durability emission test data from January through March have produced significant differences between Ford and EPA results. The problem has been most severe in the measurement of CO where EPA consistently reported cold start 1977 FTP values 20% higher than Ford results. Several pairs of test data have produced HC and CO differences in excess of 50%, although these large percent dif- ferences are most often observed at or below statutory standards. Ford and EPA have conducted gas cross checks, mass emission simulator tests, and a vehicle correlation program. The results of these tests are part of a larger program to locate and correct or account for the discrepancies in emission results between manufacturers facilities and the EPA-MVEL based on paired emission differences. 1. Test Design Three checks were made between facilities: a gas cross check analysis, tests with the Ford mass emission simulator, and a vehicle cross check tising a 1977 Ford durability vehicle. ' Mass Emission Simulator; On March 3, 1976, the Ford emissions simulator was tested using EPA CVS's 21 and 22 (analytical train #9). A CO analyzer failure prevented additional tests on CVS's 23 and 24 (analy- tical train #19). ------- Gas Cross Check: On March 4, 1976, a gas cross check was made at EPA to verify analyses of Ford supplied reference gases. The gases were read on EPA certification analysis systems 9 and 19, with the exception of the high concentration CO sample which was named on analysis train 9 only. Vehicle Cross Check; A series of vehicle correlation tests were arranged using a 1977 Ford durability vehicle. A vehicle with a known history of emissions repeatability, REPCA III, supplied to EPA by General Motors, was used as a confirmatory test vehicle. i The Ford vehicle cross check consisted of cold start 1977 type tests, highway fuel economy tests, and hot start 1974 type tests. Each test day included one cold start test, 2 HFET's, and 2 hot start tests. This series of tests was completed on two successive days at each facility. The vehicle was then exchanged between laboratories until each laboratory had completed 6 cold starts, 12 HFET's, and 12 hot starts. All tests were scheduled for a single dynamometer with each facility supplying its test driver. Testing with the Ford durability vehicle began on March 25 and ended on April 14. / The GM REPCA III vehicle was tested two times at each facility usi,ng the same dynamometer and analysis equipment that were used for testing the Ford vehicle. Emission and torque data from the REPCA III vehicle were collected and statistically analyzed. ------- 2. Test Vehicles The Ford vehicle, I.D. No. 7A1-400-5AINP, had completed 50,000 miles durability testing in March. The vehicle was tested three times at EPA during its mileage accumulation and a summary of the emissions history of this vehicle is presented below: Composite Emission Results 'est Site Ford EPA Ford EPA Ford EPA Ford •EPA Mileage 22,866 i 46,052 51,033 Average of 5 tests HC 0.405 0.408 0.333 0.498 0.518 0.454 0.460 0.377 CO 5.88 7.75 3.64 7.60 6.26 7.98 6.20 .5.57 co2 710 697 695 702 700 722 694 686 NOx 1.72 1.73 1.57 1.65 1.72 1.83 1.66 1.77 F.E. 12.31 12.49 12.64 12.40 12.47 12.06 12.56 12.78 The last set of data summarize the FTP results of this study. The vehicle is in the 5000 pound inertia class and has a 400 CID engine with automatic transmission. Emission controls include: engine modifications, an oxidation catalyst with secondary air injection, exhaust recirculation, and canister evaporative control. The vehicle was equipped to measure carburetor inlet temperature, air cleaner inlet temperature, engine speed, change in engine speed, and manifold vacuum. ------- The repeatable emissions vehicle, REPCA III, is a 350 CID, 4500 pound inertia weight vehicle which has been modified to produce stable hot start emissions at approximately 1974 federal standards. REPCA III is equipped with torque wheels and a fifth wheel speed pickup. The digital recording system is designed to measure and display positive and negative torque and calculate and display positive and negative horsepower. The Ford-EPA correlation program summed torque and horsepower over the 1372 second LA-4 cycle. 2.1 Preconditioning: Preconditioning for the first cold start at each test facility consisted of AMA, LA-4 dynamometer preconditioning, a 12 to 20 hour soak, and a one hour heat build. Evaporative emissions were not measured for the emission tests. Emission and fuel economy tests served as preconditioning for the second cold start test at each facility. Dynamometer preconditioning consisting of steady state warmup was necessary to* achieve a stable engine temperature before sampling emissions from REPCA III. r 2.2 Facilities; 2.2-1 Equipment: Gas cross checks at EPA were conducted using analytical trains #9 and 19. Mass simulator tests were made on CVS's 21 and 22 (analytical train #9). Vehicle emission and fuel economy tests ------- were run using test cells 2, 4, and 5 at Ford, while all tests were run using cell 5 and analytical train #33 at EPA. t 2.2-2 Calibration; Dynamometers, CVS's, and analytical systems were determined to be operating properly when tests were conducted at Ford and EPA. Equipment checks and calibrations at EPA and Ford are performed at least as often as specified in the Federal Register. 3. Test Results 3.1 Mass Emissions Simulator Results Mass emission results were obtained on only two sites due to a CO chopper motor failure on train #19 and a malfunctioning solenoid valve on the Ford emissions simulator. Results of the emissions simulator tests are shown in Tables 1 and 2. 3.2 Gas Cross Check Analyses Ford reference gases were named within 2.6 percent on analytical systems 9 and 19. These results are presented in Table 3. 3.3 Exhaust Emission Results The results of the exhaust emissions and fuel economy test are presented in Appendix A. HC, CO, C02 results are plotted in Figures 1- 3. All emission and fuel economy results are composite values with ------- 8 units of gm/mile. NOx emissions were not plotted because the composite results are approximately equal for cold start, fuel economy, and hot start tests. Changes in engine speed were measured and examined to determine if notable differences among test drivers at Ford and EPA could be detected. No significant variations could be observed. 3.4 REPCA III Results > The emission results from REPCA III exhibited much more variability than the baseline data supplied by GM and the vehicle was not judged useful for comparisons between laboratories. A baseline of twelve tests were run at GM to verify its repeatibility before its delivery to EPA. HC and CO emissions from the second sample bag of the LA-4 cycle have coefficients of variation of 1.5 and 2.7%, while 30 tests at EPA have variability of 8.5 and 16.0% for HC and CO respectively. The vehicle is now at the GM Proving Grounds to correct several mechanical problems and determine why the repeatability is poor. Torque and horsepower data from the vehicle are valid and comparisons of two dynamometers are summarized in Table 4. 4. Analysis of Test Results HC, CO, and CO- mass values from the Ford emission simulator are within + 5% of the expected results. NOx values measured by EPA are ------- slightly higher than expected theoretical limits. The reason for higher NOx values is not known, but the gas cross check and vehicle cross checks do not show similar NOx correlation problems. Six gas cylinders were read on analysis trains //9 and #19: 1 HC, 2 CO, 1 C0«, and 2 NOx. The high concentration CO cylinder was not read on train 19 because of an instrument failure. All gases were determined to be within 2.6% of their named concentrations. Ford and EPA exhaust emission and fuel economy results are compared by using a statistical "t" test. As shown in Tables 5, 6, and 7 for cold start, fuel economy and hot start emission tests, statistically significant differences were proven at high confidence levels. A statistical difference between Ford and EPA HC results exists at the 90% C.L. for cold start, fuel economy, and hot start tests. The differences are not believed to be important, however, because the mag- nitude of the differences are small; 0.08, 0.02, and 0.04 gm/mi for cold start, fuel "economy, and hot start tests, respectively. Ford measured higher average HC values for all three types of tests and generally had higher variability for all emission and fuel economy tests. For CO, cold start tests showed differences between facilities of over 11%, with the Ford results higher. No significant CO differences are proven at the 90% C.L. for fuel economy or hot start tests. Statistical differ- ences are not proven for NOx results from cold start, fuel economy, or hot start tests. A statistical difference between Ford and EPA measurements of C02 (8 gm/mi) is evident from an analysis of cold start tests at the ------- 10 90% C.L. No other differences between measurements of CO,, are observed, indicating the CO- differences between facilities are not serious. A difference of 0.2 mi/gal for urban fuel economy is statistically signifi- cant at the 90% C.L. This reflects the slight measurement differences in C02. The effects of ambient conditions have not been accounted for in the analysis, although differences between facilities are significant. The notable differences are: x Ford x EPA PB , 29.35 29.02 TD 82.3 73.0 *H 20.6 41.0 KH 0.841 0.901 Corrections for barometric pressure would be expected to produce a larger difference between Ford and EPA HC and CO measurements but the correlation between ambient conditions and HC and CO results is very poor. The significance of dry bulb temperature differences cannot be assessed from this program. Slightly lower Ford NOx measurements are probably related to their low levels of relative amd absolute humidity. j An analysis of the torque and horsepower results show statistical differences between facilities at the 90% C.L. but the importance r of these differences can be assessed in terms of the differences between Ford and EPA values of NOx, C02 and fuel economy. Only slight differences in cold start measurements of NOx and calculated urban fuel economy were t proven. The differences between torque measurements at Ford and EPA are ------- 11 less than previous measurements from the 1975 MVMA-EPA correlation program. The trends in dynamometer differences for torque and horsepower are more important than the statistical differences between Ford and EPA. Conclusions 1) Based on this program the correlation between Ford and EPA test sites is acceptable. Emission and fuel economy tests and checks of dynamometers and analysis equipment do not indicate serious correlation problems. 2) The effect of ambient test conditions for this particular correlation vehicle are relatively unknown. Ford ambient test conditions are sig- nificantly different from average EPA ambient conditions but emissions measurements and calculated fuel economy results between facilities correlate well. 3) This program was unable to determine the causes for the poor correlation between Ford-EPA paired test results, or the reasons why the Ford-EPA / correlation has now improved. r Recommendations 1) An MVMA-EPA correlation program would have identified correlation problems among the participating manufacturers much more efficiently. A ------- 12 mutually acceptable correlation program between MVMA and EPA test labora- tories would have been useful in determining the extent of suspected emissions correlation problems. Such a program should be conducted in the future. 2) Humidity control at Ford and EPA should be improved, particularly at Ford. ------- Table 1 AEO VEHICLE EXHAUST SIMULATOR RESULTS ON EPA CVS SYSTEM #1 •#1 #2 #3 March 3, 1976 - One Test Orifice Bank #1 & #2 #2 #1 Constituent IIC CO 2 NOx HC CO 2 NOx HC CO o C02 2 NOx Measured Gran Value 3.01 827 5.U9 1.81 28.3 832 1.2k 13.7 839 ' 2.21 Time Adjusted Gram Value 2.U6 33-6 677 l.tfS 23.2 681 2.90* 1.01 11.2 686 1.80* AEO 5$ , Upper Limit 2.63 • 3^-9 726 If . *5k 1 . 5^- 23.6 727 2.80 1.11 . H-5 729 1 1.78 AEO 5$ Lover Limit 2.38* 32.2. 656 U.10 i.ko 21.3 657 2.53 1.00 10. k 659 1. Multiplied by 180 sec../220 sec for 180 sec. test. 2. Corrected NOx values have "been divided by the humidity , correction factor of 0.88^ to yeild uncorrected results. of 5$ Limits. ------- //I Table 2 AEO VEHICLE EXHAUST SIMULATOR RESULTS ON EPA CVS SYSTEM #2 March 3, 1976 - One Test Orifice Bank *'* #2 #1 Constituent HC CO C02 2 NOx HC CO C02 2 NOx HC CO C02 2 NOx Measured Gram Valve 3-04 41.0 822 5-59 1.83 29.0 832 3.56 1.27 14.0 . 849 2.27 Time Adjusted1 Gram Value 2.^9 33-5 672 1.50 23. 7* 681 2. 91*' 11.4 695 1.85* AEO 5$ Upper Limit 2.63 34.9 726 Jt tr^i 1 ^IL 23.6 727 2.80 1.11 11.5 729 1.78 AEO 5$ Lower Limit 2.38 32.2 656 4.10 1.40 21.3 657 2.53 1.00 10.4 659 1.61 1. Multiplied "by^ l8o sec/220 sec for 180 sec. test. 2. Corrected NOx' values have been divided "by the humidity correction factor of 0.877 to yeild uncorrected results. *0ut of 5$ limits. ------- 15 Ford Results A 3470 CO 675 A 1563 CO 1275 A 10817 HC 50.82 A 14870 C02 1.93 A 5482 NO 44.7 x A 4969 NO 81.2 J*L * % Difference = Table 3 Ford-EPA Gas Cross Check EPA Results Train 19 684.0 ** 51.30 1.92 44.6 80.7 Ford - EPA Train 9 692.7 1250.4 50.24 1.90 44.9 83.4 % Difference* Train 19 -1.3 -0.9 +0.5 +0.2 +0.6 Train 9 -2.5 +2.0 +1.2 +1.6 -0.4 -2.6 EPA ** High CO Range on Train 19 Inoperative ------- Figure 1 FCCDAVE .c EPA AVE HC 0.3 O.I EPA jF.z ,F,2 EPA 1/5J FOfiO EPA •H •H • H.2. •H -H •H •H C s COLO H= HOT F V/7,? FOfcO EPA ------- Figure 2 CO 3 FORD FPA •H •H 'H -H •H • F • H •H.2 -H -H • H 3/2$- 3/36 FOCO 3/30 EPA 3/3/ EPA FO/JO EPA ------- 00 6(50 3.0O £PA £DpO EPA . -HU 3/2? FO^D Figure 3 ' ' 3/30 £PA FO£D EPA ------- 19 FORD Table 4 Dynamometer Correlation Site n = EPA Site n = V *F' X t - 4 3_ X a CV 5 6 X a CV ^ X x 100% T? £i statistic +Torque ft-lbf-sec 208,750 1,129 0.5 ,„ -KTorque 206,535 1,576 , 0.8 2,215 1.1 2.14 -Torque 114,837 1,126 1.0 -Torque 110,856 1,113 1.0 3,981 3.6 5.04 Roll Ft. 38,619 117 0.3 Roll Ft. 38,281 111 0.3 338 0.9 4.24 +Horsepower horsepower-sec 10,203 64 0.6 +Horsepower 9,994 95 1.0 209 2.1 3.38 -Horsepower 3,806 20 0.5 -Horsepower 3,645 25 0.7 161 4.4 9.54 t - values 99% C.L. 3.50 90% C.L. 1.90 ND D D D ND D D ------- 20 Table 5 Cold Start Correlation Ford n = 5 X a CV Min. Max. EPA n = 5 x t a CV Min. Max V*. -IT i r\na/ X J-UU/o t = statistic t - valuer 99% C.L. 3.36 90% C.L. ' 1.86 HC 0.460 0.051 11.0 0.419 0.556 0.377 0.015 4.0 0.356 0.390 0.083 +22.0 3.26 ND* D** CO 6.201 1.041 16.8 5.23 7.64 5.570 0.497 8.9 4.97 6.15 0.631 +11.3 0.92 ND _ ND NO X 1.664 0.107 6.4 1.54 1.79 1.766 0.090 5.1 1.68 1.86 -0.102 -5.8 -1.63 ND ND C0_ FE 694.2 12.56 8.9 0.15 1.3 1.2 685 12.4 706 12.8 686.4 12.78 7.5 0.13 1.1 1.0 678 12.6 694 12.9 7.8 -0.22 +1.1 -1.7 1.88 -1.99 ND ND *No difference exists **Difference exists ------- 21 Table 6 Fuel Economy Correlation FORD n = 12 X o CV Min. Max. EPA n = 12 X o • CV Min. Max. HC 0.082 0.006 7.3 0.072 0.093 0.066 0.007 10.3 0.059 0.08 CO 0.298 0.273 91.5 0.077 1.03 0.309 0.26 85.0 0.042 0.76 NO X 1.610 0.106 6.6 1.46 1.86 1.645 0.162 9.8 1.47 1.86 co2 502.2 17.1 3.4 483 534 512.6 10.1 2.0 502 531 FE 17.67 0.58 3.4 16.6 18.3 17.33 0.36 2.1 16.7 17.9 X., - X., 0.016 -0.011 -0.035 -10.4 0.3 .t £i —— -xlOO% +24.2 -3.6 -2.1 -2.0 +1.7 t - statistic 6.18 -0.10 -0.63 - 1.68 +1.72 t - values 99% C.L. D ND ND ND ND 90% C.L. D ND ND ND ND 1.72 ------- 22 Table 7 Hot Start Correlation FORD n = 12 X a CV Min. Max. EPA n = 12 X 0 CV Min. Max. XF ~ XE ir i no"7 t - statistic t - values 99% C.L. 2.82 r 90% C.L. 1.72 HC 0.215 0.041 19.0 0.159 0.333 0.174 0.023 13.0 0.155 0.243 0.041 +23.6 3.05 D D CO 2.656 0.692 26.0 1.70 3.82 2.470 0.421 17.1 1.56 3.08 0.186 +7.5 0.79 ND ND NO X 1.538 0.080 5.2 1.43 1.67 1.583 0.067 4.2 1.49 1.70 -0.045 -2.8 -1.47 ND ND co2 654.3 24.3 3.7 583 678 648.8 6.7 1.0 641 659 5.5 +0.8 0.76 ND ND ------- APPENDIX A Composite Emission Results from Correlation Vehicle 7A1-400-5A1NP Date Ford 3/25 3/26 EPA 3/29 3/30 Test Type. CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H Site HC CO co2 7 NOx F.E. T./T d w PB *H NOx K. 5 VOID 0 0 0 0 2 0 0 0 0 0 5 0 0 0 0 0 5 ' 0 0 0 0 0 .089 .084 .204 .204 .464 .085 .082 .213 .224 .376 .072 .074 .172 .176 .390 .064 .066 .18 .18 0.117 0.175 2.54 3.24 5.94 0.512 0.315 2.83 3.82 5.55 0.575 0.533 2.67 2,46 5.21 0.253 0.216 2.35 3.02 534' 532 583 666 706 509 501 664 672 ' 694 ,531 524 655 649 686 516 509 649 642 1.65 1.72 1.59 1.49 1.54 1.63 1.86 1.64 1.47 1.86 1.86 1.91 1.57 1.60 1.86 1.79 1.80 1.70 1.65 16.6 16.7 15.1 13.2 12.4 17~4 17.7 13.3 13.1 12.6 16.7 16.9 13.4 13.6 12.8 17.2 17.4 13.6 13.7 80.0/57.0 80.0/57.0 80.0/57.0 80.0/57.0 80.0/60.0 84.0/62.0 88.0/65.0 86.0/64.0 82.0/64.0 75.0/60.5 72.0/60.0 73.5/60.5 74.0/61.5 74.0/61.2 76.0/63.5 74,5/62.0 73.0/62.0 74.5/63.0 75.5/62.5 29.30 29.30 29.31 29.33 29.31 29.28 29.27 29.23 29.15 29.05 29.06 29.06 29.02 29.01 28.74 28,72 28.73 28.73 28.73 21 21 21 21 29 27 28 29 * 37 43 49 47 49 48 50 ' 49 54 53 48 0.833 0.833 0.833 0.833 0.878 0.887 0.917 0.911 0.939 0.921 0.932 0.930 0.947 0.941 0.977 0.957 0.968 0.978 0.960 N5 LO ------- Page 2 cont. Appendix A Date Ford 3/31 4/1 EPA 4/5 4/6 Test Type CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H Site HC 2 0. 0. v o. 0. 0. 2 0. 0. 0. 0. • 0. 5 0. 0. 0. 0. 0. 5 0. 0. 0. 0. 0. 419 085 093 220 211 449 084 082 333 201 356 07 08 169 237 386 06 06 156 157 CO 5.38 0.378 1.03 1.97 2.49 6.20 0.351 0.354 3.39 1.74 6.15 0.76 0.72 2.89 3.08 5.97 0.15 0.12 2.23 2.22 co2 685 489 483 650 648 697 514 503 666 658 694 522 516 655 653 680 506 505 641 643 NOx 1.60 1.46 1.51 1.43 1.50 1.63 1.50 1.61 1.57 1.44 1.75 1.68 1.67 1.67 1.74 1.68 1.53 1.51 1.53 1.50 F.E. 12.8 18.1 18.3 13.6 13.6 12.5 17.2 17.6 13.2 13.4 12.6 16.9 17.1 13.4 13.4 12.8 17.5 17.6 13.7 13.7 TjT d' w 80.0/57.0 82.0/57.0 83.0/57.0 84.0/58.0 80.0/57.0 83.0/58.0 83.0/58.0 83.0/58.0 83.0/58.0 83.0/58.0 73.5/58.5 72.0/57.5 72.0/57.0 72.5/56.5 72.5/57.0 70.5/58.5 72.0/58.0 74.0/57.5 73.0/56.5 72.5/56.5 PB 29.18 29.17 29.17 29.10 29.31 29.12 29.14 29.15 29.14 29.16 29.02 29.00 29.00 28.96 28.97 28.94 28.94 28.94 28.94 28.93 KH 21 18 16 17 21 19 19 19 19 19 40 40 39 36 38 49 42 35 34 36 NOx 1C 0.834 0.824 0.819 0.827 0.833 0.832 0.832 0.832 0.832 0.832 0.896 0.889 0.882 0.871 0.879 0.916 0.898 0.878 0.869 0.872 NJ ------- Page 3 cont. Appendix A Date Ford 4/7 4/8 4/9 EPA 4/13 4/14 Test Type CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H CVS-C HFET HFET CVS-H CVS-H Site HC 4 0. 0. 0. 0. 0. 4 0. 0. 0. 0. 0. 556 076 076 215 185 420 077 072 159 216 CO 7.64 0.086 0.086 3.36 2.16 5.23 0.100 0.077 2.63 1.70 co2 690 491 490 657 678 702 493 487 648 661 • NOx 1.76 1.58 1.61 1.48 1.58 1.79 1.60 1.59 1.60 1.67 F.E. 12.6 18.1 18.1 13.4 13.0 12.5 18/0 18.2 13.6 13.4 f/T d w 81/57 81/57 81/57 83/57 83/58 84/57 85/60 85/58 80/56 81/56 PB 29.53 29.54 29.54 29.61 29.62 29.68 29.66 29.64 29.61 29.60 *H 19 19 19 16 18 14 20 15. 18 17 toxK, 0.826 0.826 0.826 0.815 0.829 0.810 0.846 0.818 - 0.818 0.813 5 VOID 0. 0. 0. 0. 5 0. 0. 0, 0, 0. 061 066 163 155 379 059 059 165 169 0.109 0.160 2.56 1.56 4.97 0.068 0.042 2.43 2.17 507 496 659 653 678 508 502 642 641 1.49 1.47 1.54 1.59 1.68 1.56 1.47 1.49 1.54 17.5 17.9 13.4 13.5 12.9 17.5 17.7' 13.7 13.7 73.0/55.5 72.5/55.0 72.5/55.0 72.0/55.0 72.0/56.5 72.0/57.0 73.0/57.0 72.0/56.5 72.0/57.0 29.31 29.31 29.31 29.31 29.18 29.17 29.17 29.17 29.17 31 30 30 32 37 39 36 37 39 0.851 0.847 0.847 0.850 . 0.872 0.880 0.874 0.872 0.880 to Ul ------- |