Report No. SR00-03-01 Failure Rate Analyses and Development of Fast-Pass, Retest, and CPP Algorithms for IM147 Max CO Cutpoints prepared for: U.S. Environmental Protection Agency March 2000 prepared by: Sierra Research, Inc. 1801 J Street Sacramento, California 95814 (916) 444-6666 ------- Failure Rate Analyses and Development of Fast-Pass, Retest, and CPP Algorithms for IM147 Max CO Cutpoints prepared for: U.S. Environmental Protection Agency Regional and State Programs Division Office of Mobile Sources Under Contract No. 68-C7-0051 Work Assignment No. 1-08 March 2000 prepared by: Richard W. Joy Garrett D; Torgerson Thomas R. Carlson Sierra Research, Inc. 1801 J Street Sacramento, CA 95814 (916) 444-6666 ------- DISCLAIMER Although the information described in this report has been funded wholly or in part by the United States Environmental Protection Agency under Contract No. 68-C7-0051, it has not been subjected to the Agency's peer and administrative review and is being released for information purposes only. It therefore may not necessarily reflect the views of the Agency and no official endorsement should be inferred. ------- Failure Rate Analyses and Development of Fast-Pass, Retest, and CPP Algorithms for IM147 Max CO Cutpoints Table of Contents page 1. Summary 1 Scope of Work .2 Projected IM147 Failure Rates 3 Modal IM 147 Fast-Pass Standards 6 Predictive Retest Algorithms 6 Modal Fast-Fail Criteria 7 Integrated Fast-Pass, Retest, and Fast-Fail Algorithm Results 9 Segment 2 Revised Integrated Algorithms Results 9 Development of IM147 Driver Variation Standards 12 Integration of CPP Variation Limits 12 SIP Credit Analysis . 12 Need for Follow-Up Analysis ; 14 2. Introduction 15 Scope IB Organization of the Report 20 3. Test Data 21 4. IM147 Cutpoint Analysis 26 IM147 Test Length ... ..;.... ;... 26 Cutpoint Analysis 27 Comparison of Failure Rates 30 5. Optimized IM147 Test Criteria 36 Modal Fast-Pass Standards 36 Predictive Retest Algorithms 40 Modal Fast-Fail Criteria 43 Integration of Fast-Pass, Retest, and Fast-Fail Algorithms 50 Segment 2 Revised Integrated Algorithms Results 53 SIP Credit Analysis 53 Need for Follow-Up Analysis 54 ------- Table of Contents (continued') page 6. Development of IM147 Variation Limits 55 Existing Tolerance Limits 55 Previous Analysis of Speed Variation Limits 57 Development of Positive Power-Based Variation Limits 58 Evaluation of IM147 Variation Limits on Test Time 66 7. Integration of CPP Variation Limits 68 8. References 71 Appendix A - Startup, Intermediate and Final IM240 and IM147 Cutpoints Appendix B - Max CO, Startup, Intermediate and Final IM147 Failure Rates Appendix C - IM 147 Regression Coefficients Appendix D - Excess Emissions Identification for Max CO Cutpoints Appendix E - Second-by-Second CPP Variation Limits Appendix F - Regression Summaries Appendix G - Fleet Distribution Data -ii- ------- List of Tables page 1-1 Failure Rates, Third IM147 3 1-2 Predicted Arizona IM147 Failure Rates with MaxCO Outpoints and Integrated Algorithms 5 1-3 Modeled Fast-Pass Results Excess Emissions vs. Average Test Time 7 1-4 Retest Algorithm Results 8 1-5 Fast-Fail Algorithm 8 1-6 Comparison of Integrated Algorithms vs. Standard IM147 Impact on Test Time and Excess IM240 Emissions (MaxCO Outpoints) Lost 11 1-7 Average Test Time by Test Time Reduction Methodology and CPP 13 1-8 Comparison of IM147 Max CO Outpoints to IM240 Final Standards Impact on Excess Emissions Lost 13 4-1 IM240 to IM147 Composite Regression Equation Coefficients 28 4-2 IM147 Composite to Phase 2 Regression Equation Coefficients 29 4-3 Revised IM147 Max CO Outpoints (Composite/Phase 2) 30 4-4 Failure Rates, Third IM147 31 5-1 Revised IM147 Segments . . 38 5-2 Modeled Fast-Pass Results Excess Emissions vs. Average Test Time 40 5-3 Excess IM240 (Max CO Outpoints) Emissions Identified . (With and Without Fast-Pass) 41 5-4 Retest Algorithm Results 48 5-5 Fast-Fail Algorithm 50 5-6 Comparison of Integrated Algorithms (No CPP) vs. Standard IM147 Impact on Test Time and Excess IM240 Emissions (Max CO Outpoints) Lost ... 51 5-7 Comparison of IM147 Max CO Cutpoints to IM240 Final Standards Impact on Excess Emissions Lost 54 6-1 Preliminary CPP Cutpoints Based on Top-50% Drivers 63 6-2 Centered CPP Cutpoints and Resulting Test Alport Rates 64 7-1 End Test Decisions Affected by CPP Errors 68 -iii- ------- List of Figures page 1-1 Modal IM147 Fast-Pass Standards 4 1-2 Integrated Test Results Maximum CO Outpoints Without CPP Limits 10 3-1 Test Sequence Used to Investigate Triplicate IM147 Tests in Arizona Test Lanes 21 3-2 Model Year Distribution of Sample Fleet and AZ Overall Fleet 24 4-1 IM147 Trace Phase 1 and Phase 2 27 4-2 Failure Rate by Consecutive IM147 Max CO Cutpoints 32 4-3 Failure Rate by Consecutive IM147 Startup Cutpoints 33 4-4 Failure.Rate by Consecutive IM147 Intermediate Cutpoints 34 4-5 Failure Rate by Consecutive IM147 Final Cutpoints .35 5-1 EM147 Test Segments Used for Fast-Pass Cutpoint Development 39 5-2 Retest Algorithm During First IM147-LDGV 44 5-3 Retest Algorithm During Second LM147-LDGV 45 5-4 Retest Algorithm During First IM147-LDGT1, LDGT2 46 5-5 Retest Algorithm During Second IM147-LDGT1, LDGT2 47 5-6 Integrated Test Results Max CO Cutpoints Without CPP Limits 52 6-1 Comparison of IM147 Variation Limit Metrics , 60 6-2 Distribution of Arizona IM147 Positive Power Differences 62 6-3 Illustration of Second-by-Second IM147 Cumulative Positive Power Variation Limits 65 6-4 Effect of Initial CPP Limits Multiplier on Effective IM147 Abort Rate 66 7-1 Process for Integrating CPP Variation Limits 70 -iv- ------- 1. SUMMARY Under the Clean Air Act Amendments of 1990, metropolitan areas with the most serious air quality problems are required to implement so-called "enhanced" I/M programs. Two different test procedures for exhaust emissions testing in enhanced programs have been approved by EPA: the "IM240" test, and the "Acceleration Simulation Mode" (ASM) test. With either procedure, the efficiency of the testing process depends on how quickly accurate decisions can be made as to whether a vehicle should pass or fail. Inadequate vehicle preconditioning has previously been identified as a cause of false failures in I/M programs. In fact, previous EPA and DEQ analyses estimate that 25% of the vehicles failing the final IM240 standards would pass with further preconditioning, and that these vehicles can be identified through modal analysis. To address this problem, Sierra suggested, in another study, that Phase 1 of the IM240 test be eliminated, instead using only the second hill of the IM240 (i.e., the "IM147") up to three times in succession to ensure adequate preconditioning. Based on this recommendation, and to address the issue of inadequate preconditioning without compromising test throughput excessively, Arizona therefore decided to change its test procedure to the IM147 beginning January 1, 2000. This program upgrade will also include implementation of "Max CO" cutpoints previously developed by Sierra, which are designed to maximize the CO benefits of the program. Regarding this proposed alternative to the existing IM240 test, several concerns needed to be addressed prior to implementation in Arizona. First of all, how would emissions identification and credit change with the new procedure. Secondly, because the length of the proposed test may be equivalent to that of three IM147 tests, it can use considerably more dynamometer time than the other aforementioned tests, thus increasing the cost of the program. Prior to this study, two other studies have already been conducted to start addressing these issues. The first study addressed, among other things, reducing test time and projected emission credit levels for the IM147 test. Part of the data analyzed in this study consisted of 101 tests where vehicles were given three back-to-back IM147 tests. These data were used to create "Phase 2b" cutpoints, which are analogous to phase 2 cutpoints for the IM240 test, thus giving vehicles two ways to pass. In addition, fast-pass cutpoints for both the entire IM147 as well as Phase 2b were also developed. The other data used in this study consisted of 2% random sample IM240 test data collected in the Arizona IM240 program. These data, in combination with the 101 vehicle data, were used to project emission credits. Unfortunately, since this study did not include back-to-back IM147 to IM240 testing, excess emissions identification rates and SIP credit could not be conclusively -1- ------- established. In a follow-up work assignment in 1998,1* EPA asked Sierra to evaluate 304 triplicate IM147 tests followed by an IM240 test. These data were analyzed to verify the preliminary excess emission identification rates and average test time estimates projected for the Arizona IM program from the Phase 2 and IM240 data sets collected in previous studies. In addition, improved fast-pass and retest algorithms were developed for the IM147 test using the same approach used previously in developing similar IM240 algorithms; however, the cutpoints scenarios evaluated in the study did not include the Max CO cutpoints previously developed for DEQ. PKE Speed Variation Criteria - In the 1997 IM240-related evaluation for EPA (SR98-02- 01),2 Sierra developed improved speed variation criteria based on the total Positive Kinetic Energy (PKE) change per mile traveled during the IM240 cycle. These criteria were designed to minimize the variation in emissions while still being feasible for use by minimally trained drivers with a reasonable aptitude for dynamometer driving. However, only IM240 drive cycle criteria were developed in the 1997 study. Therefore, further analysis was needed to develop similar speed variation criteria for the IM147. Scope of Work To aid in the Arizona IM147 implementation effort, EPA issued a work assignment (#1-08) to Sierra to complete the following tasks: 1. Develop projected IM147 failure rates for the Arizona I/M program; 2. Develop modal IM147 fast-pass standards for the Max CO cutpoints; 3. Develop modal predictive IM147 retest algorithms for the Max CO cutpoints; 4. Develop modal IM147 fast-fail criteria; and 5. Develop fast-pass and full-duration PKE criteria for the IM147 test. Three distinct data sets were used in this study. The first two sets were the 304 vehicle study, collected for SR99-10-02, and a 543-vehicle sample collected for this study. For both of these sets, randomly selected vehicles were given triplicate IM147 tests followed by an IM240 test. The third data set comprised 2,518 vehicles given triplicate IM147 tests and, if they failed the third test, an IM240 test. After removing invalid tests, the test data sets used for this study consisted of 300, 535, and 2,512 vehicles, respectively (i.e., 3,347 vehicles total). All of the data were collected by Gordon-Darby I/M lanes in Phoenix, Arizona. Superscripts denote references listed in Section 8. -2- ------- Projected IM147 Failure Rates Prior to projecting failure rates for the IM147 test using each of the four set of emissions standards (Startup, Intermediate, Final, Max CO), Max CO cutpoints developed as part of SR99-10-02 were revised using the combined 300- and 535-vehicle data sets. To this end, IM147 scores were regressed against IM240 scores. The resulting regression equations were then used to derive IM147 cutpoints from the IM240 cutpoints. IM147 phase 2 cutpoints were developed similarly by regressing IM147 composite scores against IM147 phase 2 scores. A scaling factor of 0.9 was multiplied against the predicted phase 2 cutpoint to make the phase 2 cutpoint slightly more stringent than the composite cutpoint, as it is with the IM240 test. Table 4-4 in Section 4 shows the revised Max CO cutpoints. After revising the Max CO cutpoints, failure rates for the IM147 test using the Startup, Intermediate, final, and Max CO cutpoints were determined using the 3,347-vehicle data set. These failure rates, which are shown in Table 1-1, were based upon the results of the third IM147 test. Table 1-1 Failure Rates, Third IM147 HC CO NOx OVERALL Vehicle Type Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Max CO Cutpoints LDGV 114 1666 6.4% 250 1530 14.0% 155 1625 8.7% 390 1390 21.9% LDGT1 49 915 5.1% 156 808 16.2% 70 894 7.3% 219 745 22.7% LDGT2 25 578 4.1% 63 540 10.4% 30 573 5.0% 93 510 15.4% All 188 3159 5.6% 469 2878 14.0% 255 3092 7.6% 702 2645 21.0% Startup Cutpoints LDGV 106 1674 6.0% 130 1650 7.3% 141 1639 7.9% 281 1499 15.8% LDGT1 52 912 5.4% 33 931 3.4% 48 916 5.0% 108 856 11.2% LDGT2 33 570 5.5% 30 573 5.0% 28 575 4.6% 67 536 11.1% All 191 3156 5.7% 193 3154 5.8% 217 3130 6.5% 456 2891 13.6% Intermediate Cutpoints LDGV 157 1623 8.8% 161 1619 9.0% 196 1584 11.0% 367 1413 20.6% LDGT1 80 884 8.3% 53 911 5.5% 73 891 7.6% 165 799 17.1% LDGT2 40 563 6.6% 42 561 7.0% 42 561 7.0% 93 510 15.4% All 277 3070 8.3% 256 3091 7.6% 311 3036 9.3% 625 2722 18.7% Final Cutpoints LDGV 280 1500 15.7% 240 1540 13.5% 277 1503 15.6% 507 1273 28.5% LDGT1 120 844 12.4% 85 879 8.8% 127 837 13.2% 239 725 24.8% LDGT2 73 530 12.1% 56 547 9.3% 82 521 13.6% 149 454 24.7% All 473 2874 14.1% 381 2966 11.4% 486 2861 14.5% 895 2452 26.7% -3- ------- This table helps clarify the relationship between the four sets of outpoints. The Startup, Intermediate, and Final outpoints result in nearly equal HC, CO, and NOx failure rates, by vehicle category, which increase as the stringency of the cutpoints increase. The Max CO standards, when compared to the final standards, result in lower HC and NOx failure rates, and higher CO failure rates, especially for light-duty trucks. .Figure 1-1 shows the failure rates for each of the three IM147s, based on the Max CO standards. As the figure shows, most of the decrease in failure rates due to the use of multiple test cycles (i.e., to address the lack of adequate preconditioning) occurs between the first and second IM147s. As a result, it is reasonable to expect that algorithms designed to shorten the test would often end the test prior to the third IM147. This is, in fact, the case, as is shown later in this report. Figure 1-1 Overall Failure Rate by Vehicle Type Max CO Standards LDT1 LDT2 Vehicle Type ~ 1st IM147 02nd IM147 03rdIM147 Later in this report, implementation of fast-pass, fast-fail, and retest algorithms will be discussed. The addition of the algorithms has a significant impact on failure rate. In predicting which vehicles will ultimately pass the test, minor predictive errors associated with the fast-pass algorithm lead to a small fraction of false passes and in turn a reduction in the failure rate. The fast-fail and retest algorithms, which are used to predict failing vehicles, will err on the side of false failures, hence acting to increase the failure rate. One of the goals in developing the revised Max CO cutpoints was to minimize the projected false failure rate; however, false passes were treated somewhat differently. Rather than attempting to keep the number of false passes to a minimum, the study instead focused on maximizing excess emissions identification while also minimizing test time. As a result, the fast-pass algorithms allow a number of marginal vehicles that have little impact on the overall excess emissions to falsely pass. While relatively insignificant -4- ------- from an emissions perspective, this results in a fairly substantial decrease in the projected failure rate. There is also a second factor affecting the projected failure rates shown in Table 1-1. Review of the individual results for each of the three sequential IM147s shows that roughly 80 vehicles (i.e., about 2.5% of the total sample) failed the third IM147 after having passed the first test cycle. The reason for this degradation in emissions over the three IM147s is unknown, but appears to be an artifact of the test protocol. Since these vehicles would have passed out as a result of the first EM 147, they are considered false failures that would not occur with the addition of the integrated algorithms. Table 1-2 shows the effect on projected failure rates of applying the algorithms to the 3,347-vehicle sample. The failure rates shown in the table have also been adjusted by normalizing the results to the vehicle fleet distribution contained in 2% random sample data collected in the Arizona program from July 1997 through March 1998. (These were the most recent vehicle type-specific data readily available to Sierra.) This fleet distribution is detailed in Appendix G. Table 1-2 Predicted Arizona IM147 Failure Rates with MaxCO Cutpoints and Integrated Algorithms Vehicle Class Predicted Failure Rate* LDGV 16.7% LDGT1 11.4% LDGT2 10.4% Overall 14.8% Normalized to Arizona fleet distribution obtained from 2% random sample data for the period July 1997 - March 1998. Historical Arizona IM240 test results also show a significant seasonal effect. This is believed to be caused by two primary factors: the impact of changes in ambient temperatures on the purging of fuel vapors to the canister (in some older models these vapors are emitted directly to the engine), and the wintertime use of oxygenated gasoline. Both factors will act to reduce wintertime failure rates. At the high ambient temperatures typically experienced in Phoenix in the summer and fall, high purge rates from vehicles idling in the queue lead to canister overloading and breakthrough. This in turn results in higher emissions and an increase in projected failure rates. Older-technology vehicles are particularly susceptible to this phenomenon, due to the vehicles' poorer fuel delivery capabilities. This effect will not occur nearly as often during colder temperatures, thus -5- ------- contributing to lower wintertime failure rates. The use of oxygenated gasoline reduces CO emissions, leading to further decreases in failure rates. A review of historic failure rate data for the Arizona program shows that this seasonal effect appears to have a 2%-2.5% impact on failure rates. For example, the IM240 failure rate in January 1999 was 14.0% versus a failure rate of roughly 16%-16.5% .during the July-September 1999 period. Since most of the data analyzed in this study were collected during late summer, the wintertime failure rate would be expected to be significantly lower than the projections shown in Table 1-2. Modal IM147 Fast-Pass Standards Using the same methodology employed in SR99-10-02, this study developed fast-pass regression standards for the Max CO cutpoints. This methodology requires dividing the IM147 test into a series of short segments over which emission mass is accumulated. By performing multivariate linear regressions of these incremental segments, modal fast-pass coefficients were developed to predict when vehicles would pass without having to complete the entire test. Unlike the previous study, which divided the IM147 test into 14 segments, this study divided the test into 20 segments, thus increasing the frequency of opportunities for fast-pass. While a fast-pass procedure of this nature has the ability to greatly reduce test time, this reduction has to be balanced against false passes. False passes occur when vehicles that would otherwise fail an inspection are fast-passed out because their emissions over the drive cycle are not appropriately characterized by the regression. For this study, false passes are quantified by measuring excess emissions, which are defined as emissions collected during an IM240 test in excess of the applicable standard for a given vehicle. The IM147 test receives credit for identifying excess emissions if it fails a vehicle that had excess IM240 emissions when using the same emission cutpoints (e.g. Max CO). In this study, both the regression of the IM147 segments and the analysis of that regression were performed using the combined data sets (3,347 vehicles) and the Max CO standards. Table 1-3 details the results of this analysis. Predictive Retest Algorithms The workplan for the study called for Sierra to refine algorithms originally developed for SR99-10-02 and then to apply them to the total vehicle sample (3,347 vehicles) to determine their net effect on test time. In contrast to the fast-pass algorithm, where misidentification results in false-passing vehicles and a loss in excess emissions identification, the retest algorithm errors result in false failures, which can lead to consumer complaints. Like the fast-pass algorithm, decreases in test time need to be weighed against false failures to determine a reasonable compromise. -6- ------- Table 1-3 Modeled Fast-Pass Results Excess IM240 Emissions3 (w/ MaxCO Cutpoints) vs. Average Test Time No Fast-Passb Fast-Pass Enabled Excess HC Identified 97.8% 95.7% Excess CO Identified 96.3% 92.7% Excess NOx Identified 95.4% 82.4% Average Test Time0 217 seconds 125 seconds a Excess emissions normalized to Arizona 2% random sample fleet distribution data, July 1997 to March 1998. b While vehicles cannot terminate in the middle of an IM147 without the fast-pass algorithm, the test may end prior to completing three IM147s if the emissions measured at the end of any one of the IM147s meet the applicable standards. c Test time refers to the time actually required to operate the vehicle on the dynamometer. While the original retest procedure described in SR99-10-02 used a combination of mass and concentration emissions measurements to anticipate whether a vehicle would benefit from additional testing, concerns expressed by Gordon-Darby regarding the complexity of this algorithm led to a different approach for this analysis. Instead, a variation of the fast-pass regression calculation was developed and used to predict emissions improvement over an IM147 test. In short, the emissions result predicted after Segment 7 is compared to the emissions result predicted after Segment 19 to determine whether the vehicle emissions are converging on the applicable standard. Using this procedure, test time was reduced from 125 seconds (with fast-pass enabled) to 96 seconds. Table 1-4 details further results of the retest analysis. Modal Fast-Fail Criteria One of the requirements of this work assignment was to develop modal fast-fail criteria. Unlike the retest procedure, which can terminate the test at the ends of the individual IM147s, the fast-fail algorithm can terminate tests during an IM147 test. Like the retest algorithm, errors committed by the fast-fail criteria result in false failures. As a result, decreases in test time have to be weighed against false failures. With this in mind, fast-failures cannot be made during the first of the three IM147s. Analysis showed that the results of the first IM147 were too unpredictable relative to the final result to risk false-failing vehicles. The final two IM147s, however, would serve reasonably well for this purpose. -7- ------- Table 1-4 Retest Algorithm Results LDGV LDGT1, LDGT2 Total Number of Complete Tests 1567 1780 # of Failures Without Retest Algorithm 327 (20.9% of 1567) 273 (15.3% of 1780) # of Correctly Identified Failures3 245 (74.9% of 327)b 180 (65.9% of 273)b # Failing After 1 IM147 94 (38.4% of 245)c 97 (53.9% of 180)c # Failing After 2 IM147s 151 (61.6% of 245)c 83 (46.1% of 180)° # of Passing Vehicles Falsely Failed by Retest 0(0% of 1567) 0(0% of 1780) a "Correctly identified failures" refers to those vehicles that were still failing at the end of the third IM147. bThe number shown in parentheses is the number of failures without the retest algorithm. cThe number shown in parentheses is the total number of IM147 Cycle 2 and 3 failures. There are two different fast-fail algorithms, one for each of the final two IM147 tests. The fast-fail algorithm for the second IM147 fails vehicles with excessively high predicted emissions after segment 7 of the second IM147. The fast-fail algorithm for the third IM147 test uses a variation of the fast-pass algorithm to predict failing vehicles throughout the test. Both of these algorithms are described more completely in the body of this report. Table 1-5 shows the results of the fast-fail algorithms when applied to the 3,347-vehicle sample. The fast-fail algorithm reduces average test time an additional 2 seconds, from 96 seconds (with fast-pass and retest enabled) to 94 (with fast-pass, retest, and fast-fail enabled). Table 1-5 Fast-Fail Algorithm . Vehicle Class Second IM147 Fast- Failures Third IM147 Fast- Failures False Failures LDV 124 211 3 LDT 99 182 3 Total 223 393 6 -8- ------- Integrated Fast-Pass. Retest. and Fast-Fail Algorithm Results This portion of the study combined all of the optimized algorithms to determine their net effect on test time, false failures, and excess emissions while using the Max CO standards with the 3,347-vehicle sample. The flow chart contained in Figure 1-2 shows the point at which vehicles concluded the test and the reason they passed or failed. It also indicates the average dynamometer test time for each category of vehicles. Following the flow chart, Table 1-6 shows the net effect of the procedures on test time and excess emissions. The table shows that excess emission identification using the Max CO cutpoints with the fast-pass, retest, or fail-fail algorithms enabled is 95.9%, 93.1%, and 97.0% for HC, CO, and NOx, respectively. (For comparison, the respective identification rates are 97.8%, 96.3%, and 95.4% without the various test criteria enabled.) This is down from the identification rates developed in SR99-10-02,1 which identified 99.6% of the HC, 98.2% of the CO, and 99.9% of the NOx with the fast-pass and retest algorithms enabled (fast- fail was not considered). However, direct comparison of the two sets of results may not be relevant for several reasons. First, the previous study measured excess emissions captured against the Final Cutpoints rather than the Max CO cutpoints use for this study. Second, because fast-fail was created for this study, it was not included in the previous study results. Third, the retest algorithm has been modified as part of this study and will therefore have a different effect on the results. Finally, the majority of the data used in this study were collected with the newer model year exemptions in place and normalized to the vehicle inspection fleet distribution for the period July 1997 to March 1998. As a result, the vehicle distribution was skewed toward older vehicles relative to that in the previous study. Less rigorous test criteria were also evaluated as part of this latest study. However, their use yielded relatively little improvement in identification rate at the cost of a large increase in test time. It was therefore decided not to pursue this latter option. Segment 2 Revised Integrated Algorithms Results The original integrated algorithm results were initially determined assuming fast-pass and fast-fail results could not be rendered prior to the fourth segment (i.e., no earlier than Test Time = 28 seconds). This was consistent with the procedure established in SR98-02-01. Gordon-Darby, wishing to further minimize test time, requested that Sierra explore the feasibility of rendering fast-pass and third IM147 fast-fail decisions after earlier segments without degrading excess emission identification. Further investigation found that decisions could be made as early as the end of segment 2 (i.e., at Test Time = 16 seconds) if the error multiplier used in the fast-pass decision was increased during segments 2 and 3. For segment 2, the error multiplier was 3, while it was 2.5 for segment 3. Using these criteria, average test time was reduced to 91 seconds without sacrificing any excess emissions identification. -9- ------- Figure 1-2 Integrated Test Results Maximum CO Cutpoints Without CPP Limits TOTAL PASS = 2,802 TOTAL FAIL = 545 TOTAL FALSE FAILURES = 3 -10- ------- Table 1-6 Comparison of Integrated Algorithms vs. Standard IM147 Impact on Test Time and Excess IM240 Emissions (Max CO Cutpoints) Lost2 Class Model Year Group Sample Sizeb Mean Test Time Standardb Mean Test Time w/ Algorithmsb % Excess Emissions Identified' HC CO NOx LDGV 1981-82 105 286.4 140.1 - 87.6% 100.0% 1983-85 228 311.2 169.4 100.0% 97.6% 100.0% 1986-89 425 248.4 112.0 99.2% 97.1% 100.0% 1990-95 952 184.8 78.3 96.8% 85.3% 91.2% 1996+ 70 154.3 37.1 - - - All 1780 221.0 100.0 98.2% 93.3% 96.8% LDGT1 1981-85 260 306.6 158.7 78.5% 97.3% 99.4% 1986-89 222 230.8 101.6 93.8% 90.8% 99.6% 1990-95 450 173.3 59.2 0.0% . 0.0% 100.0% 1996+ 32 155.1 31.6 - - 100.0% All 964 221.9 94.9 83.4% 92.6% 99.5% LDGT2 1981-85 94 307.5 158.8 100.0% 100.0% 77.2% 1986-87 64 253.2 101.4 100.0% 100.0% 100.0% 1988-95 427 166.5 54.2 - 0% 44.4% 1996+ 18 146.0 28.0 - - - All 603 197.1 74.7 100.0% 93.8% 83.1% Weighted Average 3,347 217.0 94.0 95.9% 93.1% 97.0% 3 Test time results do not include impact of driver variation limits. b Mean test time standard refers to the average dynamometer test time without the algorithms enabled. This was determined using the 3,347-vehicle sample. c Percent of IM240 (Max CO) excess emissions identified with the integrated algorithms enabled. This was determined using the 835-vehicle sample and normalized to the Arizona 2% random sample fleet distribution data, July 1997 to March 1998. -11- ------- Development of IM147 Driver Variation Standards As will be shown in the body of this report, previous work performed on driver variation limits utilized Positive Kinetic Energy (PKE) limits to evaluate driver performance. Analysis conducted for this study, however, revealed limitations with this metric, which can result in inappropriate driver errors. As a result, a new statistic, Cumulative Positive Power (CPP), was developed to remedy this problem. Designed to be used in conjunction with EPA-specified absolute speed variation limits, CPP produces an improved, more predictable, driver evaluation criteria than PKE. When applied to the total vehicle sample for this study (minus vehicles with absolute speed violations greater than ą 2 mph), the new CPP criteria produced a total abort rate of 3.4%. Since the IM147 test, even without fast-pass, fast-fail, or retest, allows for vehicles to pass the test after a single passing IM147, many of the driver errors in the total sample would not be experienced since they occurred after the vehicle had already passed. Taking this into account, the effective abort rate was 2%. Integration of CPP Variation Limits The CPP analysis was conducted on a subset of the 3,347-vehicle population, with absolute speed excursion violations (as defined in EPA's IM240 guidance) removed. This resulted in an overall data set of 3,006 vehicles. Using the 3,006-vehicle sample, the average test time, with the fast-pass, retest, and fast-fail criteria enabled but without the CPP criteria applied, was 89 seconds. Once the CPP criteria were enabled, 87 tests (of the 3,006 vehicles) were extended, increasing the average test time by 1 second to 90 seconds. This resulting increase of 1.1% is less than the 2% increase projected at the end of Section 6, which makes sense given that the 2% projection was made without the fast- pass/fail and retest algorithms in place. The overall test time reductions caused by the fast-pass/fail and retest algorithms would mean that fewer errors would be committed. Table 1-7 summarizes the change in dynamometer test time with each succeeding set of enabled criteria. The overall impact of all the criteria is to reduce the test time by 58%, from 217 to 90 seconds. SIP Credit Analysis The above comparison of excess emissions identification between the IM240 and IM147 is based on the use of CO Max standards for both test cycles. To develop an estimate of the allowable SIP credit that should be allocated to the revised IM147 CO Max standards, it is also necessary to compare excess emissions identification between this scenario and the IM240 with EPA-recommended final cutpoints in place. This is due to the need to establish a link to using MOBILE for SIP modeling purposes. Configuring MOBILE with CO Max standards is not feasible; therefore, a better approach is to run the model with final EPA standards in place and use the excess emissions identification rates developed in this study to adjust the resulting model outputs. -12- ------- Table 1-7 Average Test Time (seconds) by Test Time Reduction Methodology and CPP Scenario Dyno Test Time Test Time Reduction (%) Cutpoint only, two possible retests 217 Added fast-pass 125 42 Added retest 96 56 Added fast-fail 94 57 Allowed fast-pass at end of Segment 2 91 58 Removed speed excursion violations 89 59 Added CPP limits 90 58 Table 1-8 shows the excess emission identification rates when the IM147 Max CO cutpoints are compared to the IM240 final standards. Pollutant-specific identification rates are shown both without and with the fast-pass, retest, or fail-fail algorithms enabled. (The latter scenario includes fast-passing vehicles as early as at the end of segment 2.) Since Arizona will be implementing the IM147 test procedure with the algorithms enabled, the identification rates for this scenario are the ones that should be used to adjust the MOBILE modeling results (based on final IM240 standards) for SEP credit purposes. Table 1-8 Comparison of IM147 Max CO Cutpoints to IM240 Final Standards Impact on Excess Emissions Lost3 Class % Excess Emission Identified (Without Fast-Pass, Retest, Fast-Fail) % Excess Emissions Identified (With Integrated Algorithms) HC CO NOx HC CO NOx LDGV 95.2% 96.2% 84.8% 91.9% 95.1% 85.7% LDGT1 80.5% 100.0% 70.6% 68.4% 98.1% 81.2% LDGT2 87.9% 100.0% 46.6% 98.5% 97.7% 73.0% Weighted Average 91.3% 97.3% 79.9% 86.7% 95.9% 81.6% a Percent of IM240 (Final Standards) excess emissions identified was determined using the 835 vehicle sample. Data normalized to Arizona 2% random sample fleet distribution data, July 1997 to March 1998. -13- ------- As expected, the table shows that HC and NOx identification rates are significantly lower with the IM147 Max CO cutpoints relative to final IM240 standards. This is due to the fact that the Max CO cutpoints are designed to maximize the CO benefits of the program at the expense of HC and NOx benefits, while keeping maximum failure rates in each cutpoint category to acceptable levels. The CO identification rate of 95.9% (with the algorithms enabled) shows that the Arizona program will achieve nearly all of the modeled benefit of the final IM240 standards. Note that this will be substantially more effective than the current phase-in IM240 standards. The table also shows that the addition of the integrated algorithms results in little more than a 1% reduction in the excess emissions identification rate for CO. (As noted above, the addition of the algorithms reduces dynamometer test time from 217 to 90 seconds.). Need for Follow-Up Analysis As discussed above, the analysis results presented in the report are based on a relatively small sample of IM147 and IM240 data. While the available data are significantly more robust than the previous sample of 300 vehicles, it is clear that these results should be revisited with a much larger sample once IM147 testing is initiated in Arizona. We therefore recommend that as soon as one to two months of IM147 data are collected in the program, they should be used to verify the validity of the cutpoints and algorithms developed in this study. This follow-up analysis would allow for any required fine- tuning of the cutpoints and algorithms. IT It it -14- ------- 2. INTRODUCTION Under the Clean Air Act Amendments of 1990, metropolitan areas with the most serious air quality problems are required to implement so-called "enhanced" I/M programs. One element of an enhanced program is a more effective test procedure, than the simple idle tests used in "basic" I/M programs. Two different test procedures for exhaust emissions testing in enhanced programs have been approved by EPA: the "IM240" test, and the "Acceleration Simulation Mode" (ASM) test. Both of these procedures have been shown to be capable of separating vehicles with excessive exhaust emissions from other vehicles; however, the accuracy of the test depends on whether tested vehicles have been adequately preconditioned and whether the speed-time profile associated with each test procedure is closely followed. With either procedure, the efficiency of the testing process depends on how quickly accurate decisions can be made as to whether a vehicle should pass or fail. Inadequate preconditioning of vehicles prior to testing is a potential cause of inaccurate or inconsistent test results because exhaust emission levels depend on how thoroughly a vehicle has been warmed up. Before the vehicle is thoroughly warmed up, high emissions can be caused by air-fuel ratio enrichment or an inactive catalytic converter. In addition, increased emissions due to purging of loaded canisters may also be an issue associated with inadequate preconditioning prior to I/M testing. Inadequate vehicle preconditioning has previously been identified as a cause of false failures in I/M programs. Under current EPA guidance, IM240 preconditioning procedures are woven into the "two-ways-to-pass" standards. Vehicles that exceed the emissions standards established for the entire 239-second test are passed or failed based on emissions occurring during the last 147 seconds of the test (also called Phase 2 or the IM147). The separate set of standards that applies to Phase 2 is slightly more stringent. For vehicles that initially demonstrate high emissions, the first 93 seconds (Phase 1) of the test are used to precondition the vehicle for the second phase of the test. In addition, EPA calls for a "second-chance" test whenever a vehicle fails the initial test by less than 50% of the standard and was in a queue for more than 20 minutes before being tested. Previous EPA and DEO Analyses - Considerable data have already been collected regarding the preconditioning requirements for IM240 testing. During 1996 and 1997, Sierra conducted evaluations of this issue using data obtained from samples of vehicles recruited from IM240 lanes in Phoenix, Arizona, and a laboratory test program at Sierra's facilities in Sacramento. The results of the 1996 analysis were reported in SAE Paper No. 962091.3 The 1997 evaluation also included an analysis of the effect on test duration ------- of adopting EPA-recommended "final" IM240 outpoints. Preliminary conclusions from the two evaluations are summarized below. 1. Using the current IM240 test procedures, it is estimated that 25% of the vehicles failing the final IM240 standards would pass with further preconditioning. 2. Vehicles that would benefit from further preconditioning can be identified through modal analysis of the emissions recorded during the IM240 test. 3. Two possible approaches to modifying the current preconditioning procedures would be to: a. retain existing IM240 test procedure and two-ways-to-pass standards, with the entire IM240 to be repeated if the Phase 2 emissions failure is marginal, emissions near the end of Phase 2 are relatively low, or emissions during Phase 2 are significantly lower than during Phase 1; or b. eliminate Phase 1 and make the initial pass/fail decision based on.running only the IM147, with a second-chance test (another IM147) for all vehicles that initially fail, and a third-chance IM147 test if emissions during the second-chance test are significantly lower than emissions during the initial test. 4. Adoption of final cutpoints and more effective preconditioning procedures involving a second full-IM240 (Option 3.a. above) will increase the portion of the test involving dynamometer operation by more than 100%. The 1997 evaluation also involved the development of improved IM240 fast-pass cutpoints using a modal regression approach originally pioneered by the New York Department of Environmental Conservation (NYDEC).4 This study also involved the development of modal predictive IM240 retest algorithms designed to minimize the fraction of vehicles either (1) identified as needing a retest when they would still fail, or (2) not identified as needing a retest when they would have passed if retested. As a follow-up to the 1997 evaluation for EPA, Sierra subsequently conducted an analysis (SR98-05-01)5 for the Arizona Department of Environmental Quality (DEQ) of the effect on failure rates, I/M program benefits, and test duration of the following changes to the current IM240 procedure: (1) implementation of the Option 3.b preconditioning procedures summarized above; (2) adoption of interim ("Max CO") cutpoints designed to maximize the carbon monoxide emission reduction benefits being achieved by the program; and (3) the exemption of either the first four or first five model years from program requirements. To analyze the IM147-only preconditioning option, Sierra used a combination of data from the 2% random test sample (consisting entirely of full-duration tests) that is -16- ------- routinely collected in the Arizona IM240 program and a limited number (101 tests) of triplicate (back-to-back-to-back) IM147 tests that were conducted as part of the 1997 EPA evaluation. A key element of the analysis methodology involved the development of "Phase 2b" cutpoints to complement the full IM147-only cutpoints. (The Phase 2b cutpoints were applied in a manner similar to the current IM240 procedure in which vehicles passing in Phase 2 are considered passing for the entire test.) Fast-pass cutpoints for both the entire IM147 and Phase 2b were also developed. As noted above, while the 1997 EPA study involved the analysis of a considerable amount of IM240 data, only a small subset (101 vehicles) was of use in projecting credit levels and test times for an IM147 test program. An additional concern is that the 101- vehicle study was not specifically designed to determine excess emission identification rates and SIP credit levels. As a result, EPA issued a follow-up work assignment to Sierra in 1998 that involved the collection of test data from triplicate IM147 tests followed immediately by a full-duration IM240. Data were collected from 304 randomly selected light-duty cars and trucks arriving at the test lane during normal queuing conditions. These data were then analyzed to verify the preliminary excess emission identification rates and average test time estimates projected for the Arizona IM program from the Phase 2 and IM240 data sets collected in previous studies. As part of the 1998 study for EPA, improved fast-pass and retest algorithms for the IM147 were developed using the same approach used previously in developing similar IM240 algorithms; however, the cutpoint scenarios evaluated in the study did not include the Max CO cutpoints previously developed for DEQ. Given Arizona's need for the maximum feasible CO reductions from its I/M program, DEQ has decided to implement this set of cutpoints. A follow-up study is therefore needed to develop improved fast-pass and retest algorithms for the Max CO IM147 cutpoints. Gordon-Darby collected additional test data from roughly 3,000 vehicles in the Phoenix area that can be used in this analysis. Of these vehicles, approximately 2,500 received triplicate IM147s only; the remaining 500 vehicles will receive triplicate IM147s followed by a single full-duration IM240. PKE Speed Variation Criteria - An additional IM147 implementation issue was the lack of allowable speed variation criteria for the shortened drive trace. In addition to the false failures caused by inadequate preconditioning, inadeqjuate control over vehicle operation during the IM240 test procedure can contribute to inaccurate results. The ability of a driver to follow the IM240 speed-time trace has a significant effect on the emissions recorded during the test. To limit this variation in test results, tolerances are applied to driver performance. In the 1997 IM240-related evaluation for EPA (SR98-02-01),2 Sierra developed improved speed variation criteria based on the total Positive Kinetic Energy (PKE) change per mile traveled during the IM240 cycle. These criteria were designed to minimize the variation in emissions while still being feasible for use by minimally trained drivers with a reasonable aptitude for dynamometer driving. However, only IM240 drive cycle criteria were developed in the 1997 study. Therefore, further analysis was needed to develop similar speed variation criteria for the IM147. -17- ------- Scope To address the issue of inadequate preconditioning without compromising test throughput excessively, Arizona decided to change its test procedure to the IM147 beginning January 1, 2000. This program upgrade will also include implementation of the Max CO cutpoints previously developed by Sierra, which are designed to maximize the CO benefits of the program. To aid in this implementation effort, EPA issued a work assignment (#1-08) to Sierra to develop the necessary test criteria. Under Work Assignment 1-08, Sierra is to complete the following tasks: 1. Develop projected IM147 failure rates for the Arizona I/M program using the 3,000-vehicle data set currently being collected by Gordon-Darby. This evaluation is to include start-up, midpoint, Max CO, and final IM147 cutpoint scenarios. 2. Develop modal IM147 fast-pass standards for the Max CO cutpoints using (a) the modal regression technique used in the 1998 EPA study to develop IM147 fast-pass standards, and (b) the 3,000-vehicle data set and the 304-vehicle data set collected in 1998. 3. Develop modal predictive IM147 retest algorithms for the Max CO cutpoints using (a) the same technique used in the 1998 EPA study to develop IM147 retest algorithms, and (b) the 3,000-vehicle data set and the 304-vehicle data set collected in 1998. 4. Develop modal IM147 fast-fail criteria that can be used to terminate retests if emissions performance is not improving during the retest. 5. Develop fast-pass and full-duration PKE criteria for the IM147 start-up, midpoint, Max CO, and final cutpoints using the 3,000-vehicle and 304-vehicle data sets, as well as the 16,581-vehicle data set from the 1997 study of IM240 PKE limits for EPA. Seven different tasks were proposed to accomplish these objectives. Task 1. Test Plan Development and Data Collection Assistance - This task covered working with Gordon-Darby in its efforts to collect the test data needed to complete the remaining tasks. Data collection, driver participation incentives, and other program- related details were performed under the guidance of DEQ and Gordon-Darby and were not Sierra's responsibility. Sierra provided assistance on an as-needed basis to resolve any problems or questions (e.g., regarding test protocols, data record format, etc.) that developed during the data collection process in Arizona. -18- ------- Task 2. Failure Rates - After completion of the vehicle testing described in Task 1, Sierra analyzed the resulting data. Using the same approach as utilized in the previous EPA and DEQ studies, projected IM147 failure rates were developed for both passenger cars and light-duty trucks. Separate projections were generated for the start-up, midpoint, Max CO, and final IM147 cutpoints developed under the previous analyses. Task 3. Fast-Pass Cutpoints - Data obtained in Task 1 as well as the 304-vehicle IM147 data set collected in 1998 were analyzed to develop fast-pass cutpoints and algorithms associated with the Max CO cutpoints. The same approach previously used to develop fast-pass cutpoints for the other cutpoint scenarios (i.e., start-up, midpoint, and final) was followed in this analysis. Sierra developed fast-pass cutpoints and algorithms for the Max CO cutpoints, with minor adjustments to the model year groups when the data indicated that such a change improved accurate emission identification. Separate sets of fast-pass cutpoints were developed for these new model year groups and vehicle classes as contained in EPA's IM240 test guidance. The impact of the resulting fast-pass cutpoints on average dynamometer test time and excess emissions identified was evaluated using the same techniques as in the previous analyses. Excess emissions identified will be expressed as the percent of excess emissions that are identified relative to those identified on the IM240 test. Task 4. Retest Algorithms - The same data used in Task 3 were analyzed to develop retest algorithms associated with the Max CO cutpoints. While this task originally charged Sierra with utilizing the same approach previously used to develop retest algorithms for the other cutpoint scenarios (i.e., start-up, midpoint, and final), subsequent comments from Gordon-Darby resulted in alternate retest algorithms. The impact of the resulting retest algorithms on average dynamometer test time was evaluated using the same technique as in the previous analyses. Task 5. Fast-Fail Criteria - The same data used in Task 3 were analyzed to develop criteria for evaluating mid-test emissions during IM147s in order to determine whether emissions performance is improving during the retest. The resulting criteria were structured to "fast fail" vehicles that are not benefitting from such retesting. The impact of the resulting fast-fail criteria on average dynamometer test time was evaluated using the same technique as in the previous analyses. Task 6. Driver Variation Criteria - The same data as used in Task 3 were analyzed to develop fast-pass and full duration driver variation limits for the IM147 start-up, midpoint, Max CO and final cutpoints.* The analytical approach used in the 1997 The original workplan called for this latest analysis to use the 16,581-vehicle data set from Sierra's 1997 PKE study for EPA (SR98-02-01); however, as explained later in this report, a "time realignment" (of emissions versus vehicle speed) was incorporated into the analysis, which made use of the previous data problematic. It was determined that the effort required to adjust the data could not be justified in terms of a significant increase in the accuracy of the results; thus, it was decided not to use these previous data. -19- ------- analysis was initially followed; however, subsequent results led to the development of an improved variation metric, Cumulative Positive Power (CPP). Consistent with the approach used in the previous analysis, the evaluation was structured to develop CPP limits designed to keep the effective abort rate due to drive trace violations to less than 3%. The impact of the resulting criteria on average dynamometer test time was also evaluated. Organization of the Report Following this introduction, Section 3 describes data collection and the data sets used throughout this study. Section 4 explains the cutpoint analysis including the revision of the Max CO cutpoints using the larger data sample; it also details failure rates for the Startup, Intermediate, Final, and Max CO standards. Section 5 describes the optimized fast-pass and rete'st criteria. In addition, it details the new fast-fail algorithm and criteria as well as the net results of optimized criteria when hin simultaneously. Section 6 describes the new CPP driver variation limits and Section 7 integrates the CPP limits with the optimized IM147 to show the net effect on test time. Section 8 lists the references cited in the report. ### -20- ------- 3. TEST DATA Data used in this study are divided into three distinct groups consisting of 304, 543, and 2,518 vehicles. Vehicles in each of the groups were given three consecutive IM147 tests regardless of the result. The 304-vehicle sample was collected by Gordon Darby for Task 1 of Work Assignment SR99-10-021 at the Gordon Darby I/M lanes in Phoenix, Arizona, during March 1998. The data included 193 cars and 111 light-duty trucks tested over triplicate IM147 tests followed by a full IM240 test as illustrated in Figure 3-1. Figure 3-1 Test Sequence Used to Investigate Triplicate IM147 Tests in Arizona Test Lanes 5-15 minutes First Second Third in queue IM147 test IM147test IM147 test Time Hot IM240 Test -21- ------- For this set, the study inspector would select vehicles by scanning the queue for the closest white vehicle waiting in the lanes. If there were no white vehicles in the queue, the inspector would look for the palest, closest vehicle waiting in the lanes. The inspector approached the first 1981 or newer vehicle following that vehicle, checked to make sure the vehicle had at least half a tank of gas, and asked the vehicle owner if he or she was interested in participating in a study that would take approximately 30 minutes, for a payment of $50. The selection process resulted in vehicles waiting in a queue for approximately 5 to 15 minutes prior to testing. Most of the vehicles participating in the program were receiving their initial test; however, 12 vehicles in the database were being re-tested after an initial failing score. As discussed in SR99-10-02, four vehicles were pulled out of the original 304 vehicle data set due to anomalous results. For this study, while Sierra did find some vehicles with anomalous results in the newer data samples (e.g., passing the initial IM147 yet grossly failing the final one), these vehicles were not removed from the sample, with the data set instead being viewed as representative of the in-use fleet. To remain consistent regarding the treatment of data from the older sample, however, the same four vehicles were removed for this analysis. The four vehicles are described below. Record 14, a 1988 Pontiac Bonneville, had relatively low CO emissions during the first and second IM147 test (1.35 and 3.42 g/mi, respectively). However, CO emissions during the third IM147 increased substantially (to 55.72 g/mi) and were higher still during the IM240 following the IM147 testing. It is interesting to note that CO was emitted in measurable quantities throughout the test, and the large increases are not attributable to a specific section of the trace. It thus appears that the gradual emissions increase could be attributable to excessive purge as the vehicle warmed up or to some kind of catalyst protection scheme. Record 15, a 1989 Dodge Dynasty, had moderate CO emissions during the first three IM147 tests (14 to 18 g/mi), but emissions during the IM240 test were excessive, particularly during the end portion of that test (106 g/mi). Reviewing the modal CO emissions in Figure 3-5, one observes that the vehicle appears to go into open-loop operation at the start of the large hill of the end portion of the test (i.e., beginning at about second 160 of the IM240). Although CO emissions accrue throughout this test, the period from 160 to 230 comprises the bulk of the emissions. Record 23, a 1993 Ford Ranger, shows a very similar emissions response throughout the three IM147 tests. As seen in Figure 3-6, most of the CO emissions occur during seconds 62 to 75 of the IM147. During the end portion of the IM240, a similar pattern is observed. In that test, however, substantial CO is also emitted during the high-speed portion of the trace. It is not entirely clear what has caused this, but it appears that the vehicle did not follow the speed-time trace as smoothly during the end portion of the IM240 as it did during the first three IM147 tests. -22- ------- Record 24, a 1995 Toyota 4Runner, had decreasing emissions throughout the first three IM147 tests, emitting only 0.72 g/mi CO during the third IM147. After these vehicles were removed, the remaining sample from the original data set consists of 191 cars and 109 trucks, for a total of 300 vehicles. The 2,518- and 543-vehicle data sets were collected for this work assignment by Gordon Darby at the ten I/M lanes in Phoenix, Arizona during June, July, and August 1999. Like the original data set, these tests were conducted at the I/M lanes in Phoenix, Arizona. Unlike the original data set, however, all motorists were asked to participate in the study, rather than simply those following a white-colored vehicle in the queue. The exception to this occurred toward the end of the testing when Gordon Darby staff, based on direction from Sierra, targeted certain vehicle model years and vehicle types to ensure that these groups were adequately represented in the test data. The main difference between the 2,518- and 543-vehicle samples was administration of the IM240 test at the conclusion of the IM147 test. For the 2,518-vehicle sample, only failing vehicles were given the IM240 test. In the 543-vehicle sample, all vehicles were given the IM240 test regardless of their IM147 result. Motorists were not required to participate in this testing. To encourage motorists to participate, inspection fees were waived for these tests. Inspection fees amount to $25 per inspection. Statistics detailing the number of refusals were not kept. As previously mentioned, anomalous vehicle test data were not thrown out of the latter two samples. There was, however, one vehicle identification number (VIN), "123456," that appeared multiple times with different vehicles. Gordon-Darby staff confirmed that this was a test VIN and should be excluded from analysis, which was done. Once this VIN was removed, the larger sample comprised 2,512 vehicles (1,360 cars and 1,152 trucks) while the smaller sample comprised 535 vehicles (229 cars and 306 trucks). The model year distribution for each of the samples is shown in Figure 3-2. The YTD (year-to-date) October 1999 line represents the initial test for 2% random sample vehicles tested in 1999 through October. Anomalous vehicles have been removed. For the most part, the model year distribution of the data samples mirrors the random sample distribution reasonably well. One notable exception can be seen with newer model year vehicles for the 300-vehicle sample. When that data set was collected, model year exemptions for the five newest model years were not in place. As a result, this sample has greater representation throughout these years. The newer data sets were collected with the model year exemptions in place; therefore, they follow the 1999 YTD random sample more closely. One important difference in this study versus previous studies was how time-alignment was handled. For the previous studies, individual channels (HC, CO, etc.) were aligned -23- ------- Figure 3-2 Model Year Distribution of Sample Fleet and AZ Overall Fleet at a> T I 00 o> o> c 0) o d) a. 20% 15% 10% 5% 0% 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 Model Year 2512 Vehicle Sample Y//A 535 Vehirle Samplp [HI 300 Vehicle Sample YTD October 1999 according to their T90 response times. The T90 response time is the length of time necessary for the analyzer to see 90% of a positive step change in the gas concentration that was introduced at the exhaust collection cone. By correcting for response time, emission events can be linked to the corresponding drive trace event. The justification for T90 time-alignment centered on the response curve of the gas bench. Typically, the response curve for a bench will appear somewhat asymptotic; as the measured gas value closes in on the actual value, the absolute rate at which the measured value approaches the target value decreases. As a result, the T90 time, which is relatively short, becomes a good approximation to time an event. / Unfortunately, the response time measurement of the analyzer is composed of two elements, gas bench response time and transport time,, which is the amount of time necessary for the CVS (Constant Volume Sample) blower to transport the sample from the collection cone to the gas analyzer bench. This second response time element, transport time, obscures the gas bench response curve through gas mixing that occurs during transport. -24- ------- Noting this effect, Gordon-Darby staff, after reviewing the raw data, suggested that the T90 time alignment overcompensated and thus caused short-duration emission events during a transient cycle to appear to precede the triggering speed event. Given the interaction of the transport time with the bench response time, Gordon Darby suggested that a more appropriate alignment measure is T50 response time, which would be more conservative and alleviate the aforementioned problem. To accommodate this change, all the existing data, which were previously aligned using T90 alignment by Gordon Darby, had to be realigned to T50 response times. This included data collected specifically for this study. Per Gordon-Darby staff, this change required shifting data for each of the channels (HC, CO, NOx, and C02) four seconds later relative to the speed signal. This was accomplished by adding four seconds of data to the front of each test, in which it was assumed that the modal emissions for the entire period were identical to those measured during the "previous first second" (i.e., now second 5) of the test. This assumption is considered reasonable since the vehicle is at idle during this entire period. I! I! it -25- ------- 4. IM147 CUTPOINT ANALYSIS This section of the report discusses the application of the IM147 Max CO cutpoints to the 3,347-vehicle data set. Divided into three parts, the section first discusses changes in overall IM147 test time. The second part of the section addresses revising the Max CO cutpoints originally developed in SR99-10-021 based on the large data set now available. The third and final part of the section details the failure rate when these revised cutpoints are applied to the" 3,347-vehicle data set. IM147 Test Length Sierra originally developed the IM147 test cycle based on the last 147 seconds of data from the IM240 drive trace. (Figure 4-1 shows the speed/time profile of the IM147 drive trace.) Following the precedence EPA established with the IM240 drive trace, this meant that the modal results of the IM147 test would have 147 seconds of data. In short, there were no constraints regarding an odd versus even number of seconds in the overall test cycle, nor in Phase 2 of the test cycle. To simplify implementation of the IM147 test in Arizona, Gordon Darby, and DEQ agreed that modal data would be recorded once every two seconds instead of second-by- second as is done with the IM240. This allows the IM147 cycle data to fit into the same size record format as is currently used for full duration IM240 tests. As a result, having an odd number of seconds in the drive trace creates a problem of what to do with the odd second. To alleviate the problem, Gordon Darby suggested that the speed/time trace define the boundaries for the test time instead of the actual number of data points reported. In other words, assuming the first speed/time point is labeled zero seconds, the first modal data result would be recorded for second 1 and the last for second 146. Phase 2 of the test would also be revised to start at second 66 (first data reported for second 67) and extend to the end of the test. The net effect of both of these changes is that both the composite results and the Phase 2 results will contain an even number of seconds. Since this addresses the issue of Gordon-Darby's two-second average data collection with no apparent negative consequences, Sierra revised the test length accordingly. Additional information on the drive trace is presented in Section 5. -26- ------- Figure 4-1 IM147 Trace Phase 1 and Phase 2 Time (sec) Cutpoint Analysis Task 1 of the work assignment required projecting failure rates for the Arizona I/M program using the combined (3,347 vehicles) data set collected by Gordon Darby. The projection includes four sets of emission cutpoints: start-up, midpoint, Max CO, and final for the IM147 test. Since the original Max CO cutpoints were developed using the 304- vehicle sample compiled for SR99-10-02, Sierra first revisited these cutpoints and model year groupings to ensure their accuracy against the larger data set. Accordingly, the model year groupings were modified to avoid anomalously high or low failure rates for any individual model year/vehicle type combination. A second objective in establishing the endpoints of the model year groupings was to ensure that changes in emissions control technology were properly reflected in the various model year groupings. (There is an obvious and direct relationship between the control technology installed on a vehicle and its ability to comply with a given set of cutpoints.) -27- ------- Once the model year groupings were revised, the IM147 Max CO outpoints were developed by regressing the final emissions results from the IM240 test against the final emission results from the third IM147 test for each of the three exhaust constituents. For this regression, Sierra used only vehicles from the two of the three data sets where vehicles were automatically given an IM240 test regardless of their IM147 test results (the 300- and 535-vehicle samples). Because the third data set contained IM240 data only for vehicles failing the IM147, this data set would have created a regression bias and was therefore excluded from this part of the analysis. The resulting regression equations were then used to extrapolate IM147 composite cutpoints from the IM240 composite cutpoints developed previously in SR99-10-02. Three linear regression equations were developed for each of three model year groupings. Equation 4-1 illustrates the regression equation. IM147-Cutpoint = slope * (IM240 Cutpoint) + Intercept [4-1] Table 4-1 details the appropriate model year/emission constituent regression coefficients to be used in the above equation. Coefficients of correlation (r2 values) are also shown for each of the regression equations. As expected, the r2 values shown in the table demonstrate good correlation between the IM240 and IM147 cutpoints. (More detailed regression results, including graphical plots, are shown in Appendix F.) The model year groupings identified in this table are not the same as the model year groupings used in the emission cutpoint tables, which vary depending upon vehicle type. IM240 to IM147 Com Table 4-1 josite Regression Equation Coefficients Model Year Group Emission Constituent Slope Intercept Correlation Coefficient (r2 value) 1981-1985 HC 0.896629 0.110694 0.963 CO 1.020463 0.858255 0.979 NOx 1.065128 0.085613 0.978 1986-1989 HC 0.933646 0.056509 0.976 CO 0.939067 1.679632 0.969 NOx 1.077932 0.058971 0.956 1990+ HC 0.963839 0.026672 0.949 CO 1.037836 0.392486 0.840 NOx 1.102698 0.048771 0.918 -28- ------- Phase 2 cutpoints were developed by regressing the IM147 composite test scores against the IM147 Phase 2 scores. Unlike the composite score regressions, however, the equations had to be adjusted to preserve the relationship between composite versus Phase 2 scores existing in the IM240 test. EM240 Phase 2 cutpoints are more rigorous than the composite cutpoints, presumably to minimize falsely passing vehicles. After studying the IM147 data, a multiplier of 0.9 was used with the regression since it provided additional defense against false failures while maintaining the possibility of a Phase 2 pass. The following Phase 2 regression equation (4-2) was used to extrapolate IM147 Phase 2 cutpoints from the IM147 composite cutpoints. IM147Phase 2 = 0.9 * (Slope * (IM147 Composite) + Intercept) [4-2] Table 4-2 details-the appropriate regression coefficients to be used with the above equation, as well as the resulting r2 values. The r2 values contained in the table demonstrate excellent correlation between the composite and Phase 2 cutpoints. Unlike the composite regression coefficients shown in Table 4-1, IM147 composite to Phase 2 coefficients were held constant across model years. Table 4-2 IM147 Composite to Phase 2 Regression Equation Coefficients Emission Constituent Slope Intercept Correlation Coefficient (r2 value) HC 0.807408 0.012886 0.973 CO 0.881965 -0.569281 0.975 NOx 0.989412 -0.083696 0.981 Table 4-3 shows the revised IM147 Max CO cutpoints developed using both sets of regression equations. In two places, the revised NOx cutpoints seem anomalous because they are actually less stringent for newer vehicles (1989 to 1990 LDGV and 1987 to 1988 LDGT2). As it turns out, these anomalies occurred at or near the regression equation breakpoints.* The slight discontinuity caused by this change created the apparent anomaly. To apply the regression equations to the model year groupings, which vary depending upon the vehicle type, cutpoints for some model years were determined using the adjacent regression equation. -29- ------- Table 4-3 Revised IM147 Max CO Cutpoints (Composite/Phase 2) Vehicle Class Model Years HC CO NOx LDGV 1981-82 2.80/2.05 26.37/20.42 3.28/2.85 1983-85 2.08/1.53 17.19/13.13 3.28/2.85 1986-89 1.46/1.07 15.77/12.00 2.75/2.38 1990-95 0.99/0.73 12.85/9.68 2.81/2.42 1996+ 0.80/0.59 12.85/9.68 2.25/1.93 LDGT1 - 1981-85 3.70/2.70 31.47/24.47 5.41/4.74 1986-89 2.86/2.09 25.16/19.46 4.91/4.30 1990-95 1.95/1.43 21.15/16.28 4.46/3.90 1996+ 1.57/1.15 21.15/16.28 3.36/2.91 LDGT2 1981-85 4.06/2.96 51.88/40.67 6.48/5.69 1986-87 3.79/2.77 39.24/30.64 5.99/5.26 1988-95 2.92/2.13 26.34/20.39 6.11/5.37 1996+ 2.34/1.71 26.34/20.39 4.46/3.90 Comparison of Failure Rates After revising the Max CO outpoints using the 835-vehicle sample, the failure rates were evaluated using each of the four sets of IM147 outpoints (Startup, Intermediate, Final, and Max CO) and the combined vehicle data set (3,347 vehicles). While the IM147 Max CO cutpoints were revised for this study, the IM147 Startup, Intermediate, and Final cutpoints were developed as part of SR99-10-02 and are shown in Appendix A. Table 4-4 shows how the failure rate changed with the different cutpoints for the third and final IM147. The overall failure rate will be slightly less when actually implemented since some vehicles will pass and therefore complete the test after an earlier IM147 even though they would go on to fail the third one if the test was continued. More detailed information on the failure rates is provided in Appendix B. This table helps to clarify the relationship between the four sets of cutpoints. The Startup, Intermediate, and Final cutpoints result in nearly equal HC, CO, and NOx pollutant-specific failure rates for each cutpoint category (e.g., the pollutant-specific failure rates for the startup cutpoints range from 5.7% to 6.5%). Both the pollutant- specific and overall failure rates increase with increasing cutpoint stringency. The Max CO standards, when compared to the final standards, reduce the HC and NOx failure rates, and instead increase the CO failures, especially for light-duty trucks. -30- ------- Table 4-4 Failure Rates, Third IM147 Vehicle Type HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Max CO Cutpoints LDGV 114 1666 6.4% 250 1530 14.0% 155 1625 8.7% 390 1390 21.9% LDGT1 49 915 5.1% 156 808 16.2% 70 894 7.3% 219 745 22.7% LDGT2 25 578 4.1% 63 540 10.4% 30 573 5.0% 93 510 15.4% All 188 3159 5.6% 469 2878 14.0% 255 3092 7.6% 702 2645 21.0% Startup Cutpoints LDGV 106 1674 6.0% 130 1650 7.3% 141 1639 7.9% 281 1499 15.8% LDGT1 52 912 5.4% 33 931 3.4% . 48 916 5.0% 108 856 11.2% LDGT2 33 ~ 570 5.5% 30 573 5.0% 28 575 4.6% 67 536 11.1% All 191 3156 5.7% 193 3154 5.8% 217 3130 6.5% 456 2891 13.6% Intermediate Cutpoints LDGV 157 1623 8.8% 161 1619 9.0% 196 1584 11.0% 367 1413 20.6% LDGT1 80 884 8.3% 53 911 5.5% 73 891 7.6% 165 799 17.1% LDGT2 40 563 6.6% 42 561 7.0% 42 561 7.0% 93 510 15.4% All 277 3070 8.3% 256 3091 7.6% 311 3036 9.3% 625 2722 18.7% Final Cutpoints LDGV 280 1500 15.7% 240 1540 13.5% 277 1503 15.6% 507 1273 28.5% LDGT1 120 844 12.4% 85 879 8.8% 127 837 13.2% 239 725 24.8% LDGT2 73 530 12.1% 56 547 9.3% 82 521 13.6% 149 454 24.7% All 473 2874 14.1% 381 2966 11.4% 486 2861 14.5% 895 2452 26.7% Figures 4-2 through 4-5 show how the failure rate changes as vehicles progress through the three IM147 tests. Since most of the increase in failures occurs between the first and second IM147 tests, it is reasonable to expect that algorithms designed to shorten the test would end the test prior to the third IM147. As will be shown later in the report, this is indeed the case. Once the fast-pass, retest, and fast-fail criteria are applied, only 162 vehicles out of 3.347.(4.8%) are tested beyond the second EM147. ------- Figure 4-2 Failure Rate by Consecutive IM147 Max CO Outpoints HC Failure Rate by Vehicle Type u i ~ 1st 1M 147 Ś 2nd IM 147 03rd IM 147 PC LDT1 LDT2 Vehicle T ype All NOxFailure Rate by Vehicle Type O z ED 1st IM 147 Ś 2nd IM 147 ~ 3rd IM 147 PC LDTI LDT2 Vehicle Type All CO Failure Rate by Vehicle Type 0 1st IM 147 Ś 2nd IM 147 ~ 3rd IM 147 PC LDT 1 LDT2 All Vehicle T ype Overall Failure Rate by Vehicle Type 0 1st IM 147 Ś 2nd IM 147 03rd IM 147 PC LDT1 LDT2 Vehicle Type All ------- I u> u> Figure 4-3 Failure Rate by Consecutive IM147 Startup Outpoints HC Failure Rate by Vehicle Type u X [0 I st IM 147 Ś 2nd IM 147 ~ 3rd IM 147 LDT1 LDT 2 Vehicle Type NOx Failure Rate by Vehicle Type o z 0 1st IM 147 Ś 2nd IM 147 03rd IM 147 PC LDT1 LDT2 Vehicle Type All CO Failure Rate by Vehicle Type o u O 1st IM 147 Ś 2nd IM 147 03rd IM 147 PC LDT I LDT2 Vehicle Type All Overall Failure Rate by Vehicle Type 25.0% Ł 20.0% a S 15.0% ~ 1st IM 147 Ś 2nd IM 147 Q 3rd IM 147 PC LDT1 LDT2 Vehicle Type All ------- Figure 4-4 Failure Rate by Consecutive IM147 Intermediate Cutpoints HC Failure Rate by Vehicle Type ~ '1st IM 147 Ś 2ndIM 147 ~ 3rd IM 147 PC LDTI LDT2 All Vehicle Type NOxFailure Rate by Vehicle Type ~ 1st IM 147 Ś 2nd IM 147 ~ 3rd IM 147 PC LDT I LDT2 All Vehicle Type CO Failure Rate by Vehicle Type o CJ ED 1st IM 147 |2nd IM147 EJ3rd IM 147 PC LDT1 LDT2 Vehicle Type All Overall Failure Rate by Vehicle Type ~ 1st IM 147 Ś 2nd IM 147 ~ 3rd IM 147 PC LDTI LDT2 Vehicle Type All ------- Figure 4-5 Failure Rate by Consecutive IM147 Final Cutpoints HC Failure Rate by Vehicle Type El 1st 1M 147 Ś 2nd IM I 47 ~ 3rd IM 147 PC LDTI LDT2 Vehicle T ype All NOxFailure Rate by Vehicle Type (011st IM 147 Ś 2nd IM 147 123rd IM 147 0.0% PC LDTI LDT2 Vehicle T ype All CO Failure Rate by Vehicle Type o o EH 1st IM 147 Ś 2nd IM 147 ~ 3rd IM 147 PC LDTI LDT2 Vehicle Type All Overall Failure Rate by Vehicle Type 0 1st IM 147 Ś 2nd IM 147 03rd IM 147 LDTI LDT2 Vehicle Type ------- 5. OPTIMIZED IM147 TEST CRITERIA This section presents the analysis methodology and results used to develop optimized IM147 test criteria. Criteria developed in the study include fast-pass standards, retest algorithms, and fast-fail criteria. As described below, each set of criteria was evaluated in turn to determine its overall effect on test time and excess emissions identification. Modal Fast-Pass Standards IM147 fast-pass standards were originally developed for startup, midpoint, and final standards as part of SR99-10-02.1 The current study furthered that work by developing fast-pass standards for the Max CO cutpoints. The fast-pass regression coefficients determined for both SR99-10-02 and the current study were developed using the methodology described in Sierra Report No. SR98-02-01,2 "Additional Study of Preconditioning Effects and Other IM240 Testing Issues." As detailed in that report, the selected drive trace is divided into segments over which mass emissions for HC, CO, and NOx are summed. By performing a multivariate linear regression of the modal mass emissions against the composite emissions result, we can determine the coefficients needed to predict final emissions at mode ends throughout the test. Equation 5-1 illustrates how these coefficients are used to predict emissions. P240* = Cn + Ł {Snm X Xnm} + {Me X En) t5*1] m=1 Where: P240n = Predicted emissions after completing n segments Cn = Regression intercept for equation n Snm = Regression coefficient for segment m in equation n Xnm = Total emissions over a given segment m in equation n Me = Error multiplier (usually 2 unless otherwise specified) En = Error in regression equation n n = Equation number (corresponds to the number of modal segments completed) m = Segment number After reviewing the proposed fast-pass and retest procedures described in SR99-10-02, Gordon-Darby staff expressed several concerns regarding actual in-use implementation of that procedure. In response, the workplan developed for the current project indicated that -36- ------- Sierra would initiate further discussions with Gordon-Darby and evaluate possible changes to the previously developed procedures to address these concerns. Specific concerns that were voiced by Gordon-Darby staff include the following: 1. Insufficient number of fast-pass segments - SR99-10-02 divides the IM147 drive trace into 14 segments. While the segments in the first half of the test usually comprise 8 to 10 seconds, several segments in Phase 2 of the drive trace are considerably longer (up to 19 seconds in duration). Gordon-Darby expressed concern that extended segments may force vehicles to be tested longer than necessary prior to a fast-pass. 2. Fast-pass segments containing an odd number of seconds - To allow the resulting modal test data up to three possible IM147 cycles to fit into the same size record format as currently used for full duration IM240 tests, Gordon- Darby -plans to record two-second averages, rather than one-second recordings as is presently done. To simplify its lane software, Gordon-Darby requested that Sierra realign the segments to agree with the planned frequency of data storage. 3. Nonalignment of fast-pass segments with proposed retest modes - To further simplify its programming process, Gordon-Darby asked that Sierra try to coincide retest mode breaks with segment breaks. In addition to simply developing updated fast-pass regression coefficients and retest algorithms based on an expanded data set, this study addresses the above concerns. In response to the first two concerns as well as the test length issue (odd vs. even number or seconds), the fast-pass segments have been revised as shown in Table 5-1. As the table shows, there are now 20 segments ranging from 4 to 10 seconds in duration. Each segment contains an even number of seconds. Phase 2 of the IM147 now begins at segment 11 and extends through the end of the test. The test includes 146 seconds of data (collected starting at second 0 and ending at second 146). When the IM240 segments were originally created as part of SR98-02-01, they were divided in such a way that segments characterized different modes of vehicle operation. Some segments were composed of hard accelerations, while others characterize cruises, and everything between, including decelerations. While this study re-aligned the segments with deference to ensuring 4- to 10-second segment lengths divided along even increments, care was taken to preserve the original intent of the segmentation. Figure 5-1 shows the re-aligned segments positioned against the drive trace. -37- ------- Table 5-1 Revised IM147 Segments Segment Initial Second Data Final Second Data 1 1 4 2 5 16 3 17 22 4 23 28 5 29 34 6 35 42 7 43 48 8 49 54 9 55 60 10 61 66 11 67 76 12 77 82 13 83 92 14 93 98 15 99 108 16 109 112 17 113 116 18 117 122 19 123 132 20 133 146 -38- ------- Figure 5-1 IM147 Test Segments Used for Fast-Pass Cutpoint Development Time (sec) Using these revised test modes, Sierra applied the aforementioned regression method to the 3,347-vehicle sample supplied by Gordon-Darby. Thirty-nine distinct regression models, shown in Appendix C, were developed from these data. Models were created for each vehicle type (LDGV, LDGT1, and LDGT2), emission constituent (HC, CO, and NOx), and phase (composite and Phase 2 only). There are 20 equations for the composite models and 10 equations for the Phase 2 models. After revising the regression models, they were applied to the 835-vehicle sample (unbiased IM240 sample) to determine how the models would affect excess emissions identified by the IM147 test. Excess emissions are defined as emissions collected during an IM240 test in excess of the applicable standard for a given vehicle. The IM147 test receives credit for identifying excess emissions if it fails a vehicle that had excess IM240 emissions. Table 5-2 shows the excess emissions versus test time results when the Fast-Pass algorithm was enabled. Max CO standards were used for both the IM147 and the IM240. As Table 5-2 shows, the IM147 test with the fast-pass criteria enabled still identifies over 91% of the excess emissions for each of the three exhaust constituents and still -39- ------- Table 5-2 Modeled Fast-Pass Results Excess Emissions vs. Average Test Time No Fast-Pass Fast-Pass Enabled Excess HC Identified 97.8% 95.7% Excess CO Identified 96.3% 92.7% Excess NOx Identified 95.4% 82.4% Average Test Time 217 seconds 125 seconds significantly reduces average test time.* Without fast-pass enabled, the pass/fail evaluation occurs only after each IM147 is complete. If the vehicle passes the IM147, the test would be complete at that point. If the vehicle fails, another IM147 would be run, with up to three IM147s conducted on any one vehicle. If the vehicle is still failing at the end of the third IM147, it fails the overall test. Since each IM147 lasts 146 seconds, the maximum time a vehicle could be tested is 438 seconds. Without fast-pass, the average test time for the 3,347-vehicle sample was 217 seconds. Once the fast-pass algorithm was applied, the average test time dropped to 125 seconds, for a reduction of 92 seconds (42%). This is similar to the test time estimate of 121 seconds shown in SR99-10-021 for the final standards with fast-pass. Table 5-3 details excess emissions identification by vehicle type and model year groups. Appendix D provides additional excess emission analysis results. As will be shown later in the report, excess emissions identified increase with the addition of retest and fast-fail algorithms. Predictive Retest Algorithms As with the fast-pass standards, the predictive retest algorithms were developed as part of SR99-10-02. Retest algorithms are intended to predict whether a vehicle would benefit from additional testing. If the algorithm determines that a vehicle that was failing at the end of the first or second of three IM147 tests would benefit from additional testing, then the next IM147 would begin. If, on the other hand, the algorithm determines that the vehicle would not benefit from additional testing, the test would be terminated at that point and the vehicle would fail the inspection. * Average test time refers to the estimated time for that portion of the test in which the vehicle is driven on the dynamometer. -40- ------- Table 5-3 Excess IM240 (Max CO Cutpoints) Emissions Identified3 (With and Without Fast-Pass) Vehicle Class Model Year No Fast-Pass Fast-Pass Enabled Excess HC Identified Excess CO Identified Excess NOx Identified Excess HC Identified Excess CO Identified Excess NOx Identified LDGV 81-82 N/A 87.6 100.0 N/A 87.6% 100.0% 83-85 100.0% 97.6 100.0 100.0% 97.6% 100.0% 86-89 100.0% 99.8 100.0 99.2% 97.1% 100.0% 90-95 99.0% 8.6.3 91.2 96.7% 85.3% 46.0% 96+ N/A N/A N/A N/A N/A N/A ALL 99.6% 94.8 96.8 98.2% 93.3% 80.6% LDGT1 81-85 78.5 99.6 99.4 78.5% 92.0% 99.4% 88-89 100.0 99.0 71.3 93.8% 90.8% 71.3% 90-95 100.0 0.0 100.0 0.0% 0.0% 100.0% 96+ N/A N/A 100.0 N/A N/A 0.0% ALL 90.0 99.0 94.6 83.4% 91.0% 92.6% LDGT2 81-85 92.3 100.0 79.7 92.3% 100.0% 79.7% 86-87 100.0 100.0 53.1 100.0% 100.0% 53.1% 88-95 100.0 100.0 65.7 100.0% 36.2% 56.6% 96+ N/A N/A N/A N/A N/A N/A ALL 96.1 100.0 65.8 96.1% 95.2% 64.5% Total ALL 97.8 96.3 95,4 95.7% 92.7% 82.4% ° N/A means that no vehicles failed the applicable IM240 cutpoint within this vehicle class/model year grouping. In the current study, the workplan called for Sierra to refine the algorithms using the 3,347-vehicle sample and the Max CO cutpoints. The original algorithms predicted test outcomes using a combination of mass and concentration readings during specific modes of the IM147 test. In conversation with Gordon-Darby staff, however, Sierra learned that the suggested retest logic would be difficult to implement. While gas concentrations could be determined during the test, it would be far less work if all of the retest criteria referred only to mass. Another concern, as previously mentioned, was that the modes did not align with the fast-pass segments. With these issues in mind, Gordon-Darby asked if Sierra could find a more user-friendly retest logic. While Sierra agreed to explore alternatives to concentration measurement in the retest procedure, it was not clear that any viable alternatives existed since the justification for using both concentration and mass made sense technically. While mass emission -41- ------- measurements are useful to gauge emission performance relative to the actual cutpoint, the concentration measurements provide a measure of emission performance less affected by engine load than mass emissions. As a result, emissions concentration measurements seemed especially relevant for measuring improvements in engine performance during a transient test. After some experimentation, Sierra settled on a procedure that utilizes the previously determined fast-pass regression coefficients to predict whether a vehicle would benefit from additional testing. In short, if the vehicle has not fast-passed the inspection by the end of the 19th segment and the predicted emissions fall outside a tolerance level allowing automatic retest, the emissions result predicted after segment 7 is compared to the emissions result predicted after segment 19 to determine whether the vehicle emissions are converging on the applicable cutpoint. Equation (5-2), named the convergence ratio for this study, illustrates how emission convergence on the applicable standard is determined. Segment! - Comp Std Convergence._Ra,io - SegmentIg_ Comp_Std [5-2] Where: Segment7 = Predicted emissions after segment 7 (Composite Regression) Segmentl9 = Predicted emissions after segment 19 (Composite Regression) Comp_Std = Composite Cutpoint for the specific emission Common sense would dictate that the convergence ratio would be greater than one if the emissions are converging on the cutpoint for vehicles where the emissions are above the cutpoint. A convergence ratio less than 1, on the other hand, might suggest that the emissions are actually diverging from the cutpoint. With this in mind, it makes sense that a conservative decision threshold for the retest algorithm would utilize a convergence ratio greater than one for the HC and CO channels. In general, emissions of both of these pollutants reduce as a vehicle warms up. NOx, on the other hand, may actually increase the longer a vehicle operates, so the appropriate convergence ratio threshold may be greater than one. Iterative solutions to this equation using actual vehicle data suggest, however, that both of these assumptions may result in overly stringent application of the retest algorithm, thus creating unacceptable false failure rates. In short, some vehicles failing for either HC or CO during one of the early IM147 tests may go on to pass a later IM147 in spite of the fact that their emissions appeared to be diverging from the cutpoint during the earlier IM147. Regarding NOx emissions, modal emissions are even more difficult to predict. As a result, the convergence ratio was not applied to the NOx channel. Instead, the retest algorithm predicts NOx failures by simply comparing segment 7 and 19 NOx readings to the previously mentioned tolerance level. If the readings are greater than the prescribed multiple of the cutpoint, then the vehicle does not receive a retest. -42- ------- Separate algorithms were developed for LDGVs and LDGTs based upon whether one or two IM147 tests had been completed. In all cases, a predicted emissions score of less than the standard (but that does not trigger a fast-pass) is treated as cause for a retest. This approach avoids the need to deal with negative convergence ratios, while also providing maximum potential for vehicles to pass the test. The flow charts shown in Figures 5-2 through 5-5 show the specifics regarding how these algorithms work. Unlike the fast-pass algorithms, the error term in the fast-pass regression equations is not included when determining predicted emissions for the Segment 7 to Segment 19 comparison. If the error term (which varies between regression models) were included, the retest procedure would need to be refined for specific regression models in addition to vehicle types. Without including the error term, LDGVs can simply be separate from LDGTs while still maintaining acceptable accuracy. Table 5-4 presents the results of the retest algorithm independent of the fast-pass algorithm. As shown in the table, the retest criteria eliminate more vehicles after the second test than after the first test. This is in contrast to the fast-pass criteria, which pass a disproportionate number of vehicles after the first IM147 when compared to the second IM147. Given the difference between the two criteria, this makes sense. While vehicles can pass the test after early IM147 tests without the fast-pass enabled, the only way a failing vehicle can end the test without the retest algorithm enabled is to run the full duration of the test. As a result, the retest criteria on the first IM147 are judged by their ability to predict emissions on the third IM147 whereas the fast-pass criteria are judged by their ability to predict emissions on the current IM147, an easier task. The retest criteria must therefore be more conservative on the first IM147 than the fast-pass criteria. In addition to listing the point at which vehicles were denied a retest, Table 5-4 shows the number of false fails as a result of the retest procedure. The criterion for false failing is quite simple: a false failure occurs any time a vehicle that would go on to pass one of the subsequent IM147 tests is denied a retest. As shown, there were no false failures. Despite the conservatism evidenced by this result, average test time dropped from 125 seconds to 96 seconds when the retest algorithm was added to the fast-pass criteria. Modal Fast-Fail Criteria One of the requirements of this work assignment was to develop modal fast-fail criteria. Unlike the retest procedure, which can terminate the test at the ends of the individual IM147 tests, the fast-fail algorithm can terminate tests during an IM147 test. Like the retest algorithms, the fast-fail algorithm has the potential to falsely fail vehicles that would otherwise pass the inspection in the algorithm's absence. After looking at the test data, it was apparent that the data from the first IM147 were too unpredictable to fast-fail any significant number of vehicles during the first IM147 without also significantly increasing false failure levels. While the retest algorithm utilizes almost all of the first IM147 data before it makes a decision not to retest a vehicle, the fast-fail algorithm must make essentially the same decision in a smaller amount of time. For this reason, it was decided that the fast-fail algorithm would not function before the second IM147. -43- ------- Figure 5-2 Retest Algorithm During First IM147 - LDGV Max CO Outpoints Note: STD = Composite Standard for the specific pollutant. Mode 7 Predicted = Predicted emissions after 7 segments using the composite regression equations without error term included. Mode 19 Predicted = Predicted emissions after 19 segments using the composite regression equations without error term included. -44- ------- Figure 5-3 Retest Algorithm During Second IM147 - LDGV Max CO Outpoints Note: STD = Composite Standard for the specific pollutant. Mode 7 Predicted = Predicted emissions after 7 segments using the composite regression equations without error term included. Mode 19 Predicted = Predicted emissions after 19 segments using the composite regression equations without error term included. -45- ------- Figure 5-4 Retest Algorithm During First IM147 - LDGT1, LDGT2 Max CO Cutpoints Note: STD = Composite Standard for the specific pollutant. Mode 7 Predicted = Predicted emissions after 7 segments using the composite regression equations without error term included. Mode 19 Predicted = Predicted emissions after 19 segments using the composite regression equations without error term included. -46- ------- Figure 5-5 Retest Algorithm During Second IM147 - LDGT1, LDGT2 Max CO Cutpoints Note: STD = Composite Standard for the specific pollutant. Mode 7 Predicted = Predicted emissions after 7 segments using the composite regression equations without error term included. Mode 19 Predicted = Predicted emissions after 19 segments using the composite regression equations without error term included. -47- ------- Table 5-4 Retest Algorithm Results LDGV LDGT1, LDGT2 Total Number of Complete Tests 1567 1780 # of Failures Without Retest Algorithm 327 (20.9% of 1567) 273 (15.3% of 1780) # of Correctly Identified Failures3 245 (74.9% of 327)b .180 (65.9% of 273)b # Failing After 1IM147 94 (38.4% of 245)c 97 (53.9% of 180)° # Failing After 2 IM147s 151 (61.6% of 245)c 83 (46.1% of 180)c # of Passing-Vehicles Falsely Failed by Retest 0(0% of 1567) 0(0% of 1780) a "Correctly identified failures" refers to those vehicles that were still failing at the end of the third IM147. b The number shown in parentheses is the number of failures without the retest algorithm. c The number shown in parentheses is the total number of IM147 Cycle 2 and 3 failures. In order to build on work already completed for the retest procedure, the fast-fail algorithm for the second IM147 trace evaluates predicted emissions after segment 7 of the drive trace.* To maintain uniformity for lane software programmers, the error term was not included in the prediction of the mode 7 emissions since it was not in the retest algorithm. The vehicle fast-fails the second IM147 if any of the following are true: For LDVs: Predicted HC after 7 segments > (1.5 x Composite HC standard) Predicted CO after 7 segments > (2.2 x Composite CO standard) Predicted NOx after 7 segments > (1.4 x Composite NOx standard) ~ It is theoretically possible to identify additional fast-fails by using the same or different predictive algorithms at the end of subsequent segments. While Gordon-Darby has expressed interest in this enhancement in order to further reduce average test time, the timing of the study did not allow this issue to be evaluated. -48- ------- ForLDTs: Predicted HC after 7 segments > (1.1 x Composite HC standard) Predicted CO after 7 segments > (1.5 x Composite CO standard) Predicted NOx after 7 segments > (1.5 x Composite NOx standard) Using these criteria, 99 trucks and 124 cars are fast-failed after segment 7 of the second IM147. Different fast-fail criteria were used during the third IM147, since the potential for falsely failing vehicles in subsequent IM147 tests is eliminated. As a result, the modal fast-pass algorithm, with a minor modification, can be implemented to predict failing vehicles. Instead of the error term being added to the predicted score, as is done with the fast-pass algorithm, the error term is subtracted. This adjustment ensures conservative emission estimates, which will help to minimize false failures when using the algorithm. Equation 5-3 illustrates the third IM147 fast-fail algorithm. n P240n = Cn + J] {iSnrn X Xnm} ( Me X En) o-i m= 1 P"-5! Where: P240n = Predicted emissions after completing n segments Cn = Regression intercept for equation n Snm = Regression coefficient for segment m in equation n Xnm = Total emissions over a given segment m in equation n Me = Error multiplier (usually 2 unless otherwise specified) En = Error in regression equation n n = Equation number (corresponds to the number of modal segments completed) m = Segment number Table 5-5 details the results of the fast-fail algorithm for both the second and third IM147 tests. As the number of false failures indicates, the fast-fail criteria were developed with the intention of minimizing false failures. The results shown in the table are also deceptive, since they are based on an analysis of the impact of the fast-fail algorithm in the absence of the other test criteria. While over 600 fast-failures are shown in the table, many of these are also subject to the retest criteria, resulting in a much smaller fast-fail impact when the criteria are combined. In this latter case, the number of fast-fails falls to roughly 200 vehicles. This effect, combined with the fact that such fails occur relatively late in the 3-IM147 test cycle, leads to a fairly small impact on average test time. Once the fast-fail algorithm was enabled with the fast-pass and retest algorithm, average test time was reduced from 96 seconds to 94 seconds. -49- ------- Table 5-5 Fast-Fail Algorithm Second IM147 Third IM147 False Vehicle Class Fast-Failures Fast-Failures Failures LDV 124 211 3 LDT 99 182 3 Total 223 393 6 Integration of Fast-Pass. Retest. and Fast-Fail Algorithms The next step in the analysis was to integrate the Fast-Pass, Retest, and Fast-Fail criteria to determine their net effect. Table 5-6 shows how average test times and excess emissions identified vary by model year range and vehicle class using the integrated criteria. As you can see, once the retest and fast-fail algorithms were added to the fast-pass algorithm, excess emissions identification improved. There are two reasons for this: (1) vehicles that would be falsely passed later in the test are now failed prior to that decision being made; and (2) some vehicles that passed according to the IM147 criteria yet would have failed the IM140 are now falsely failed on the IM147, thus increasing IM240 excess emissions identification. The excess emissions identification rate shown in Table 5-6 is down from the identification of SR99-10-02,1 which identified 99.6% of the HC emissions, 98.2% of the CO, and 99.9% of the NOx with the fast-pass and retest algorithms enabled. However, direct comparison of these results may not be relevant for several reasons. First, the previous study measured excess emissions captured against the Final Outpoints rather than the Max CO cutpoints developed for this study. Second, because fast-fail was created for this study, it was not included in the previous study results. Third, since the retest algorithm has been modified as part of this study, it will have a different effect on the results. In the previous study, the retest algorithm improved excess emission identification by 2.7% versus 1.1% for this study. Lastly, most of the vehicle data used in this study were collected with the model year exemptions in place and then normalized to the inspection fleet, based on 2% random sample data collected from July 1997 to March 1998. This skews the model year distribution older when compared to SR99-10-02, which used data unbiased by model year exemptions. Figure 5-6 illustrates how the integrated criteria combined to produce emission results and final test times. Note that only 7 of the 3,347 vehicles were tested over the full duration of all three IM147 tests. -50- ------- Table 5-6 Comparison of Integrated Algorithms (No CPP) vs. Standard IM147 Impact on Test Time and Excess IM240 Emissions (Max CO Cutpoints) Lost Class Model Year Group Sample Size" Mean Test Time Standard3 Mean Test Time w/ Algorithms" % Excess Emissions Identified1" HC CO NOx LDGV 1981-82 105 286.4 140.1 - 87.6% 100.0% 1983-85 228 311.2 169.4 100.0% 97.6% 100.0% 1986-89 425 248.4 112.0 99.2% 97.1% 100.0% 1990-95 952 184.8 78.3 Ś 96.8% 85.3% 91.2% 1996+ 70 154.3 37.1 - - - AH" 1780 221.0 100.0 98.2% 93.3% 96.8% LDGT1 1981-85 260 306.6 158.7 78.5% 97.3% 99.4% 1986-89 222 230.8 101.6 93.8% 90.8% 99.6% 1990-95 450 173.3 59.2 0.0% 0.0% 100.0% 1996+ 32 155.1 31.6 - - 100.0% All 964 221.9 94.9 83.4% 92.6% 99.5% LDGT2 1981-85 94 307.5 158.8 100.0% 100.0% 77.2% 1986-87 64 253.2 101.4 100.0% 100.0% 100.0% 1988-95 427 166.5 54.2 - 0% 44.4% 1996+ 18 146.0 28.0 - - - All 603 197.1 74.7 100.0% 93.8% 83.1% Weighted Average 3,347 217.0 94.0 95.9% 93.1% 97.0% a Mean test time standard refers to the average dynamometer test time without the algorithms enabled. This was determined using the 3,347-vehicle sample. b Percent of IM240 (Max CO) excess emissions identified with the integrated algorithms enabled. This was determined using the 835-vehicle sample and normalized to the Arizona 2% random sample fleet distribution data, July 1997 to March 1998. -51- ------- Figure 5-6 Integrated Test Results Maximum CO Outpoints Without CPP Limits TOTAL PASS = 2,802 TOTAL FAIL = 545 TOTAL FALSE FAILURES = 3 -52- ------- Segment 2 Revised Integrated Algorithms Results The original integrated algorithm results were determined assuming fast-pass and fast fail results could not be rendered prior to the fourth segment (i.e., no earlier than Test Time = 28 seconds). This was consistent with the procedure established in SR98-02-01. Gordon-Darby, wishing to further minimize test time, requested that Sierra explore the feasibility of rendering fast-pass and third IM147 fast-fail decisions after earlier segments without degrading excess emission identification. Unfortunately, simply moving the decision forward with the existing algorithms, while shortening test time, did degrade excess emission identification. To adjust for this change, the error multiplier used in the fast-pass decision needed to be increased to 3 during segment 2 and to 2.5 during segment 3. While the fast-fail algorithm used during the third EM 147 also uses an error multiplier, it can be left at 2 during segments 2 and 3 without increasing false-failure incidence. Using these criteria, the earliest possible fast-passes were moved to the end of segment 2 (i.e., Test Time = 16 seconds) and average test time was reduced to 90.5 seconds without sacrificing any excess emissions identification. Excess emissions identification remained at 94.4% for HC, 95.4% for CO, and 95.7% for NOx. SIP Credit Analysis The comparison of excess emissions identification between the IM240 and IM147 that is presented above is based on the use of CO Max standards for both test cycles. However, to develop an estimate of the allowable SIP credit that should be allocated to the revised IM147 CO Max standards, it is also necessary to compare excess emissions identification between this scenario and the IM240 with EPA-recommended final cutpoints in place. This is due to the need to establish a link to using MOBILE for SIP modeling purposes. Configuring MOBILE with CO Max standards is not feasible; therefore, a better approach is to run the model with final EPA standards in place and use the excess emissions identification rates developed in this study to adjust the resulting outputs. Table 5-7 shows the excess emission identification rates when the IM147 Max CO cutpoints are compared to the IM240 final standards. Pollutant-specific identification rates are shown both without and with the fast-pass, retest, or fail-fail algorithms enabled. (The latter scenario includes fast-passing vehicles as early as at the end of segment 2.) Since Arizona will be implementing the IM147 test procedure with the algorithms enabled, the identification rates for this scenario are the ones that should be used to adjust the MOBILE modeling results (based on final IM240 standards) for SIP credit purposes. As expected, the table shows that HC and NOx identification rates are significantly lower with the IM147 Max CO cutpoints relative to final IM240 standards. This is due to the fact that the Max CO cutpoints are designed to maximize the CO benefits of the program at the expense of HC and NOx benefits, while keeping maximum failure rates in each cutpoint category to acceptable levels. The CO identification rate of 97.9% (with the algorithms enabled) shows that the Arizona program will achieve nearly all of the modeled benefit of the final IM240 standards. Note that this will be substantially more -53- ------- Table 5-7 Comparison of IM147 Max CO Cutpoints to IM240 Final Standards Impact on Excess Emissions Lost* Class % Excess Emission Identified (Without Fast-Pass, Retest, Fast-Fail) % Excess Emissions Identified (With Integrated Algorithms) HC CO NOx HC CO NOx LDGV 95.2% 96.2% 84.8% 91.9% 95.1% 85.7% LDGT1 80.5% 100.0% 70.6% 68.4% 98.1% 81.2% LDGT2 87.9% 100.0% 46.6% 98.5% 97.7% 73.0% Weighted Average 91.3% 97.3% 79.9% 86.7% 95.9% 81.6% * Percent of IM240 (Final Standards) Excess Emissions Identified was determined using the 835-vehicle sample. effective than the current phase-in IM240 standards. The table also shows that the addition of the integrated algorithms results in less than a 1% reduction in the excess emissions identification rate for CO. Need for Follow-Up Analysis As discussed above, the analysis results presented in the report are based on a relatively small sample of IM147 and IM240 data. While the available data are significantly more robust than the previous sample of 300 vehicles, it is clear that these results should be revisited with a much larger sample once IM147 testing is initiated in Arizona. We therefore recommend that as soon as one to two months of IM147 data are collected in the program, they should be used to verify the validity of the cutpoints and algorithms developed in this study. This follow-up analysis would allow for any required fine- tuning of the cutpoints and algorithms. ### -54- ------- 6. DEVELOPMENT OF IM147 VARIATION LIMITS In addition to fast-pass/fail cutpoints and retest criteria, trace variation limits were also developed for the IM147 test. Under Task 6 of the Work Assignment, an analysis methodology that was used to develop IM240 variation limits under an earlier EPA study2 was applied to the pilot IM147 data to develop similar limits for the IM147 test. During the course of the effort, an alternate statistical metric, Cumulative Positive Specific Power (CPP), was identified that resulted in better second-by-second variation limits than the Positive Kinetic Energy (PKE) metric employed in the previous EPA study. This section of the report describes why the new statistic was selected, and how IM147 CPP variation limits were developed and evaluated to ensure they do not produce excessive test abort rates. It also assesses their individual impact on average dynamometer test time. (The combined effect of fast-pass/fail cutpoints, retest criteria and IM147 trace variation limits is discussed in Section 7.) Before describing the effort performed under the current Work Assignment, a review of the existing IM240 tolerance limits and a summary of the previous evaluation of those limits are presented. Existing Tolerance Limits The prescribed driving cycles for the transient IM240 and IM147 tests consist of varying second-by-second speeds ranging from zero (i.e., idle) to 56.7 mph, with maximum speed changes of ą3.3 mph/sec. During actual I/M testing, the driver watches a graphical display of the prescribed or "reference" speed/time trace overlayed with the actual second-by-second trace as it is being driven as an aid to following the reference trace and anticipating upcoming speed changes. (The visual display also indicates prescribed shift points for manual transmission vehicles along the trace.) Since each vehicle has different performance characteristics, it is impossible, even for highly skilled drivers, to precisely follow the second-by-second reference trace speeds during actual "one-time-only" testing. As a result, EPA originally developed a set of speed-based tolerance limits for the IM240 test that defined the leeway allowed to the driver in trying to follow the reference trace for the test to be considered valid. Those tolerance limits consisted of two components: (1) speed excursion limits; and (2) speed variation limits. Each of these criteria is described below.6 -55- ------- Speed Excursion Limits f85.2221 (e) (4)1 - Speed excursion limits shall apply as follows : (i). The upper limit is 2 mph higher than the highest point on the trace within 1 second of the given time. (ii). The lower limit is 2 mph lower than the lowest point on the trace within 1 second of the given time. (iii). Speed variations greater than the tolerances (e.g., during gear changes) are acceptable provided they occur for no more than 2 seconds on any occasion. (iv). Speeds lower than those prescribed during accelerations are acceptable provided they occur for no more than 2 seconds on any occasion. Speed Variation Limits [85.2221 Ye) (5)1 (i). A linear regression of feedback value on reference value shall be performed on each transient driving cycle for each speed using the method of least squares, with the best fit equation having the form: y = mx + b, where: (A), y = the feedback (actual) value of speed; (B). m = the slope of the regression line; (C). x = the reference value; and (D). b = the y-intercept of the regression line. (ii). The standard error of estimate (SE) ofy on x shall be calculated for each regression line. A transient driving cycle lasting the full 240 seconds that exceeds the following criteria shall be void and the test shall be repeated: (A). SE = 2.0 mph maximum. (B). m= 0.96 -1.01. (C). r2 = 0.97 minimum. (D). b = ą2.0 mph. Simply stated, the speed excursion limits require that vehicles be driven within ą2 mph of the reference trace, accommodating for gear changes and other momentary excursions. -56- ------- The speed variation limits, though a bit more difficult to comprehend, were intended to ensure that speed differences from the reference trace within the ą2 mph excursion limits "envelope" would not bias the resulting measured emissions. However, EPA suspended the use of the speed variation limits on November 23, 1993, pending further evaluation. Previous Analysis of Speed Variation Limits Inadequacy of Speed Variation Limits - In 1998, Sierra completed a study for EPA2 that evaluated the ability of the linear regression-based speed variation limits to identify high emissions-producing speed variations. It was found that the linear regression criteria were inadequate in flagging high-emissions speed variations. The reason for this finding was that the standard error (SE) statistic does not give "appropriate" higher weighting to speed deviations occurring at the critical high-emission points along the IM240 trace. Instead, it gives equal weighting to all speed variations and thus (along with the other regression statistics used in the speed variation criteria) is ill-suited to identifying those speed deviations that substantially affect IM240 emissions. Evaluation of Alternative Statistics - During this study, two alternative statistical measures were evaluated for their ability to better identify IM240 speed variations that significantly affect measured emissions: 1. DPWRSUM7 - the sum of absolute changes in specific power; and 2. Positive Kinetic Energy fPKE) - the sum of positive differences in kinetic energy per unit distance. It was found that the PKE statistic provided a better measure than DPWRSUM for identifying those speed variations from the reference trace that produce high emissions. This finding was supported by analysis of modal (i.e., second-by-second) speed and emissions data from a random sample of 16,581 full* IM240 tests from the Arizona I/M program. It was determined that the high-emission portions of the IM240 test closely corresponded with periods of acceleration. A examination of both the DPWRSUM and PKE statistics found that DPWRSUM is increased during both decelerations and accelerations. PKE, on the other hand, is increased only during acceleration periods. Thus, the DPWRSUM statistic is "diluted" with speed variations during decelerations that have little effect on emissions. As a result, the PKE statistic was reasoned to provide a better measure of significant emissions-producing speed variations. Development of PKE-Based IM240 Variation Limits - From this finding, PKE variation limits were then developed as potential replacements to the original regression-based "Full" tests refer to those run over the entire test duration, regardless of fast-pass or fast-fail status. -57- ------- speed variation limits. The basic approach used to develop PKE-based speed variation criteria for the IM240 consisted of the following elements: 1. Establishing upper and lower "composite" PKE limits for full 240-second tests from the Arizona data sample; and 2. Scaling these composite limits based on the cumulative PKE at each second of the IM240 reference trace to produce second-bv-second PKE variation limits. Second-by-second PKE variation limits were established to ensure that compliance with the reference trace was maintained throughout the test to minimize emissions bias during fast-pass and fast-fail determinations. From analysis of-the Arizona data, the PKE variation limits were established over a range expected to produce no more than a 3% increase in the test abort rate when applied as a replacement for the speed variation criteria, in conjunction with the existing speed excursion (i.e.. ą2 mph) criteria- Development of Positive Power-Based Variation Limits Under the current study, the randomly collected triplicate Arizona IM147 tests described in Section 3 were also analyzed to develop IM147 trace variation limits. Similar to the earlier IM240 study, upper and lower composite limits were first established over the full duration of the IM147 test. The composite variation limits were then scaled at each second of the reference trace to produce second-by-second variation limits. Before the details of how these IM240-developed methodologies were adapted for the IM147 test are discussed, an explanation of why positive power was used as a replacement for PKE as the variation limits metric is provided. Use of Positive Power Instead of PKE - During the course of developing the second-by- second IM147 variation limits, a number of tests were identified for which PKE-based variation limits were being exceeded under periods of deceleration. This was clearly a problem. As discussed earlier, statistical metrics used to establish variation criteria were selected based on their ability to identify high emissions-producing variations that coincide with acceleration events. Although the composite PKE statistic does this well on a cumulative basis over the entire test, it is less-suited when applied on a second-by- second basis. The reason for this can be seen by first considering how PKE is calculated. Over a traveled driving cycle of distance x, cumulative PKE per unit distance is defined as follows: ------- Em CPKE(x) = x [6-1] jcbc The term PP, is referred to as the positive specific power at time t and is given by the following equation: iV2 - (V<-i)2 whenV, >Vtml f 0 whenV, ŁV,_, PP' = i [6-2] By definition, positive (specific) power* is non-zero during acceleration and zero during cruise and deceleration events. Cumulative positive power (CPP) is defined as the sum of positive power at each second t over a transient driving cycle of T seconds, or [6-3] CPP(T) = Y,PP, (=0 Thus, CPP increases during accelerations over a transient driving cycle and remains constant during cruise and deceleration. Conversely, cumulative PKE decreases during cruise and deceleration because the denominator in Equation [6-1], which represents the summed distance driven, still increases while a vehicle is cruising or decelerating. Cumulative PKE becomes constant only during periods of idle (i.e., zero speed). Figure 6-1 illustrates the different behavior of each metric over the IM147 test. It shows second-by-second speed, CPP, and cumulative PKE over the reference trace. Speed (in mph) is plotted against the left axis. Cumulative positive power and cumulative PKE are plotted against the right axis (in mph2/sec and miles/hr2, respectively). As seen in Figure 6-1, CPP either increases or remains constant over the entire duration of the transient IM147 test. On the other hand, cumulative PKE decreases during both deceleration and cruise events, by definition. Furthermore, cumulative PKE can actually decrease during modest accelerations after the initial portion of the IM147 test. For Power is literally defined as the rate of change in kinetic energy (or work). Strictly speaking, specific power (i.e., power per unit mass) would be calculated by the velocity times the acceleration at time t. When the change in kinetic energy is evaluated on a second-by-second basis as defined in Equation 6-2, instead of as a net change from the beginning to the end of the cycle, PP, as defined in that equation approaches the strict definition of positive power. -59- ------- Figure 6-1 Comparison of IM147 Variation Limit Metrics Time (seconds) example, this phenomenon can be seen at Seconds 11-14 in Figure 6-1, where the reference trace exhibits acceleration from Second 5 through Second 14 but cumulative PKE begins dropping beyond Second 11. Because of this behavior, it was difficult to establish reasonable second-by-second variation limits based upon cumulative PKE differences between the reference and driven traces using a scaling approach similar to that developed under the 1998 IM240 study. That approach basically consisted of scaling the composite PKE interval limits established at the end of the test by the percentage difference of these limits from the composite reference value. Since the magnitude of critical emissions-affecting deviations during accelerations is difficult to distinguish from the magnitude of deviations that do not substantially affect emissions during deceleration and cruise, this scaling approach fails when based upon cumulative PKE. The result is that cumulative PKE-based second-by-second variation limits of a specific scaled interval width will either falsely flag less important deviations during decelerations or not identify deviations during accelerations that do affect measured emissions. Thus, two alternate approaches were considered for establishing reasonable second-by- second IM147 variation limits: 1. Use of a better-behaved statistic, such as positive power, in conjunction with the basic scaling approach; and 2. Development of separate variation limits for each second using deviation distributions (reference vs. actual) at each second compiled from a full modal analysis of the test data sample. -60- ------- Although conceptually appealing since variation limits are established independently for each individual second, the latter approach would require a rigorous modal analysis just to develop initial variation limits for each second. A complex iterative process would then be required to evaluate these limits over the entire trace and "tune" them in a manner that yields an acceptable overall abort rate. This timing step would also consider the relative impact of trace variations at each second on measured emissions. For example, it would be desirable to apply tighter trace variation limits during acceleration segments that produce high emissions than during less emissions-significant idling segments if the resulting overall abort rates (across all segments) can be kept at acceptable levels. In addition, it is believed that imposing tighter, independently established second-by- second trace variation limits would result in a much greater degree of "re-leaming" by the lane inspectors as they adjust to the impact of these limits. This re-learning process and any necessary re-tuning of the variation limits could best be evaluated through an initial pilot study before being implemented on a program-wide basis. During this pilot study, it is also envisioned that different approaches to providing visual feedback to drivers as they attempt to follow the trace could be evaluated. For example, this could involve providing dynamically updated forward-looking "trace envelopes" or speeds that guide the driver back toward the reference trace when a nominal excursion begins in a manner that complies with the second-by-second variation limits. As a result of the scope of the latter approach, the first approach was selected because it was believed to substantially overcome the shortcoming of the PKE-based metric while being less resource-intensive to apply and test than limits developed from a full modal analysis. It should also be noted that using CPP as a replacement for PKE assumes that the existing ą2 mph criteria developed by EPA will be applied in conjunction with CPP- based variation limits. This assumption is dictated by the use of a cumulative metric in specifying second-by-second variation limits. Development of Composite Variation Limits - Composite IM147 variation limits were generated using a similar methodology to that employed in the 1998 IM240 study, except cumulative positive power (CPP), rather than cumulative PKE, was used as the statistical metric. Figure 6-2 shows the distribution of composite (i.e., 147-second) CPP calculated from the second-by-second actual speeds in the Arizona data sample, expressed as the percent difference between actual and reference IM147 CPP. As the figure shows, actual CPP appears normally distributed, although the median CPP is approximately 1% higher than the reference value. To determine how far to go along the "tails" of the CPP differences distribution to set composite limits, CPP differences among the test lane drivers in the Arizona data were examined. The basic concept applied in setting the CPP limits was to identify a significant fraction of drivers who, historically, could always (or nearly always) run IM147 tests within the selected CPP limits. Given a mixture of ability among individual drivers to follow the reference trace, Sierra sought to identify the fraction of "competent" drivers who could follow the trace ------- Figure 6-2 Distribution of Arizona IM147 Positive Power Differences % Difference Between Reference and Actual Positive Power (Sample Size = 9,306) 14% g 12% o 5 10% O B/o O g. 6% 2% 0% Ś 111 11 I III I I I I I 111 I I I I I III III 1111111 -10%-9% -8% -7% -6% -5% -4% -3% -2% -1% 0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11% 12% 13% 14% 15% Percent Difference from Reference Value more consistently than others when conducting IM147 tests for a range of vehicles. This subset of competent drivers and tests was used to establish composite CPP limits. From the Arizona data sample of over 10,000 valid triplicate IM147 tests, records in . which speed excursions (over ą2 mph from the reference trace) occurred were discarded, leaving a remaining sample of 9,306 tests. For the purpose of establishing composite CPP variation limits based on the capabilities of "good" drivers, speed excursion tests were removed from this portion of the analysis. The remaining data sample was then grouped by individual drivers, one for each of the 231 drivers that were found in the sample. Driver groups containing fewer than 25 test records were then discarded; this left a total of 112 driver groups, which encompassed 86% of the tests in the total sample (i.e., before discarding small-sample driver groups). Composite CPP was then calculated for each test in the remaining driver groups. The mean and standard deviation of CPP from the tests within each driver group were also computed. The driver groups were then ranked by increasing CPP standard deviation and the top 50% of the drivers (based on lowest CPP standard deviation) were used to compute possible composite CPP outpoints. Table 6-1 lists a series of possible CPP cutpoints computed from the percentile CPP variance among the top 50% drivers. For example, the CPP cutpoints shown for the 2% row under the "Top 50% Percentile" column (the third column in Table 6-1) indicate that 96% (100% - 2 x 2%) of the tests from the top 50% drivers had composite CPP within 4,437 and 4,949 mph2/sec. -62- ------- Table 6-1 Preliminary CPP Outpoints (mph2/sec) Based on Top-50% Drivers (Sample Size = 7,962 Tests) Top-50% Driver Lower Tail Top-50% Driver Upper Tail Top-50% Driver Percentile Low-End CPP (mphVsec) High-End CPP (mphVsec.) Interval Width (mphVsec) 0.25% 99.75% ą0.25% 4,348 5,077 728 0.5% 99.5% ą0.5% 4,368 5,039 671 0.75% 99.25% ą0.75% 4,391 5,013 622 1.0% 99.0% ą1.0% 4,403 4,987 584 2.0% 98.0% ą2.0% 4,437 4,949 512 3.0% 97.0% ą3.0% 4,459 4,920 462 4.0% 96.0% ą4.0% 4,475 4,903 428 5.0% 95.0% ą5.0% 4,483 4,882 398 Note that these preliminary outpoints are not centered about the IM147 reference CPP value of 4,617 mph2/sec (as evidenced by the shifted CPP distribution shown earlier in Figure 6-2). To generate a series of "final" composite limits for evaluation, Sierra applied the interval widths shown in Table 6-1 to the reference CPP value to produce "centered" limits about the reference value. Incremental abort test rates for each set of centered cutpoints were then calculated based on both the entire 9,306 test Arizona data sample and the top-50% driver subset. The results are presented in Table 6-2, which lists both "simple" and "effective" abort rates. Simple abort rates represent the fraction of tests in the sample for which the composite CPP cutpoints would have been exceeded. Effective abort rates were calculated from simple rates by subtracting the fractions of emission-pass tests that exceeded the upper CPP cutpoint and emission-fail tests that exceeded the lower CPP cutpoint. The idea is that tests on vehicles that had passing emission scores but were driven with high CPP should not be aborted. Similarly, tests that failed on emissions despite being driven below the lower CPP cutpoint should also be considered valid tests and not aborted. Thus, tests should be aborted only when the upper CPP variation limit is exceeded for an emissions failure or the lower CPP variation limit is exceeded during a passing test. The emissions pass/fail determinations used to calculate the effective abort rates shown in Table 6-2 are based upon the "Max CO" cutpoints developed earlier in the study. In addition to the calculated test abort rates, Table 6-2 also shows the percentage of drivers who are always within the limits of each set of CPP cutpoints. -63- ------- Table 6-2 Centered CPP Cutpoints and Resulting Test Abort Rates Top-50% Driver Percentile Centered CPP Limits (mph2/sec) Abort Rates (%) from All Tests Abort Rates (%) from Top-50% Driver Tests Percentage of All Drivers Within Limits Lower CPP Limit Upper CPP Limit Interval Half- Width Simple Abort Rate Effective Abort Rate Simple Abort Rate Effective Abort Rate > Lower Limit < Upper Limit ą0.25% 4,253 4,982 364.2 5.8% 1.6% 1.2% 0.2% 95.7% 53.9% ą0.5% .4,282 ; | 4,953_ \ 335,6 ' JW] ' , 2:1%* ^ " 1.9%; -0.4%. ATA% ą0.75% 4,307 4,928 310.8 8.7% 2.6% 2.7% 0.6% 91.8% 40.5% ą1.0% 4,325 4,909 292.1 10.5% 3.1% 3.7% 0.9% 88.4% 37.1% ą2.0% 4,361 4,873 255.9 14.1% 4.2% 6.3% 1.6% 81.5% 28.9% ą3.0% 4,386 4,848 230.9 17.3% 5.3% 8.9% 2.6% 77.6% 26.3% ą4.0% 4,403 4,831 214.2 19.9% 6.2% 10.9% 3.1% 74.1% 24.6% ą5.0% 4,418 4,816 199.1 22.6% 7.2% 13.2% 3.9% 67.2% 22.4% Sample Size - - - 9,306 Tests 7,962 Tests 231 Drivers Note: Shading indicates the CPP limits proposed for use by Sierra, and corresponding data. Based on the results given in Table 6-2, Sierra proposes the use of composite lower and upper CPP variation limits of 4,282 and 4,953 mph2/sec, respectively (shown in the shaded row in Table 6-2). As indicated in the table, these composite CPP limits are expected to increase the (effective) abort test rate by 2.1% relative to that resulting from EPA's recommended ą 2 mph limits based on available test data. Since the goal of the analysis was to specify variation limits that kept the abort rate due to these variation limit violations to 3%, composite CPP limits resulting in only a 2.1% incremental abort rate were selected. This left some "room" below the 3% target for the impact of also imposing second-bv-second CPP variation limits. If drivers are selected based on their ability to perform as well as the best 50% of the current drivers, then the abort rate would drop to just 0.4%. In practice, it is expected that the abort rate will increase by less than this amount as drivers "adjust" to the new limits. Development of Second-bv-Second Variation Limits - Using the recommended composite CPP limits of 4,282 and 4,953 mph2/sec, second-by second CPP limits were generated by scaling the percentage difference of these limits from the composite reference value (7.3%) to the CPP calculated at each second from the reference trace. This approach was further modified as described below. 1. To provide drivers with a short period to "learn" to drive each test vehicle, second-by-second CPP limits were not imposed until t=30 seconds; and -64- ------- 2. To further accommodate wider allowable variations (on a percentage basis) in second-by-second CPP at the beginning of the transient IM147 test, a "CPP Multiplier" factor was applied that widened the allowed CPP limits progressively from the end of the test to where limits begin at t=30 seconds. From its maximum value beginning at t=30 seconds, the CPP multiplier factor was linearly decreased to a value of unity (i.e., 1.0) at t=146. In other words, at the end of the test, the CPP limits were set equal to the composite CPP limits. Furthermore, this linear narrowing was applied only over the acceleration sections of the IM147 trace, during which the reference CPP is increasing. During cruise and deceleration periods, the limit widths were held constant (as the reference CPP also remains constant). This latter improvement to the methodology employed in the 1998 IM240 study, in conjunction with the use of CPP instead of CPKE, enabled second-by-second IM147 CPP limits to be specified so variation limit aborts were not falsely triggered during deceleration and cruise portions of the transient test. Note that second-by-second CPP variation limits developed in this manner can still be exceeded during cruise and deceleration events, signaling tests that should be aborted. However, falsely triggered "anomalous" exceedances that occur from the use of a PKE-based metric are eliminated under this modified approach. Figure 6-3 illustrates this modified second-by-second CPP-based variation limit concept. The thick solid line shows the reference CPP over the IM147 test; the thinner solid lines represent the lower and upper CPP limits established as described above. These CPP traces are plotted against the left axis. Speed, indicated by the dashed line, is plotted against the right axis. Note that the CPP limits can be held constant during cruise and deceleration periods. Thus, variation limit exceedances in these intervals are real, rather than anomalous artifacts of the statistical metric. Figure 6-3 Illustration of Second-by-Second IM147 Cumulative Positive Power Variation Limits Time (sec) -65- ------- To establish second-by-second CPP variation limits that produced expected test aborts near the 3% target rate, a range of initial CPP multipliers (from 2.0 to 6.0) were evaluated. These initial multipliers specify the width of the variation limits at the starting point (t=30 seconds) relative to the composite interval width at the end of the test. For example, an initial CPP multiplier of 2.0 means that the starting interval width was 14.6% (2.0 x 7.3% composite CPP interval width) of the reference CPP trace at that point. Figure 6-4 shows the increase in effective abort rate as a function of varying initial CPP multipliers. The diagonally striped region represents lower CPP variation limit aborts, the shaded region above shows upper limit aborts. Since these abort events are mutually exclusive, their sum represents the total expected effective abort rate from implementation of the CPP variation limits. Figure 6-4 Effect of Initial CPP Limits Multiplier on Effective IM147 Abort Rate (Sample Size = 9,306) 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Initial CPP Limits Multiplier Based on the analysis results, an initial CPP multiplier of 3.5 is recommended for implementation in Arizona. Second-by-second CPP variation limits based on the use of this multiplier are shown in Appendix E. Evaluation of IM147 Variation Limits on Test Time Using the second-by-second CPP variation limits described in the preceding section, an analysis was conducted of the impact of these variation limits on average dynamometer test time. This was a simplistic analysis since it addressed only the singular impact of the variation limits. (A more exhaustive analysis of the combined test time impact of CPP variation limits in conjunction with fast-pass, fast-fail and retest decisions is presented in Section 7.) -66- ------- In this simple analysis, the "without limits" or base average test time was assumed to be 146 seconds, the length of a full IM147 test. This assumption was necessitated by the Arizona data sample, which contained only full tests. The recommended CPP limits were found to produce a total effective abort rate of 3.4% based on the Arizona IM147 data. The average time at which the aborts occurred under these limits was determined to be 101.6 seconds. Thus, the "with limits" average test time was then calculated as follows: With Limits Test Time = Base Test Time + (Abort Rate x Average Abort Time) = 146 + (3.4% x 101.6 seconds) = 149.5 seconds It should be noted that this approach also assumes that an aborted test is performed successfully on the subsequent re-test. Under these assumed conditions, the CPP variation limits will increase average test times by approximately 2% [(149.5-146) + 146], II II II mm -67- ------- 7. INTEGRATION OF CPP VARIATION LIMITS The final phase of this analysis involved integrating the CPP criteria with the other algorithms included in this study to determine net test time. As shown in Section 6, there are both high-end CPP errors and low-end CPP errors. High-end CPP errors occur when the vehicle is driven too aggressively, whereas low-end CPP errors occur when the vehicle is driven too smoothly, essentially minimizing the peaks and valleys of the trace. Because high-end CPP errors will create additional emissions and in turn make it more difficult for a vehicle to pass the test, a test where a vehicle passed in spite of high-end CPP errors is considered a valid test. If the same vehicle failed because of high emissions, the cause is assumed to be the high-end CPP error; therefore, the test would need to be extended to ensure fairness. Low-end CPP errors, however, would result in lower mass emissions and make it easier for vehicles to pass falsely. In those cases where the vehicle passes with a low-end CPP error, the result would be invalid and the test would need to be extended. If a vehicle fails with a low-end CPP error, the result is valid and the test would terminate. Table 7-1 -details which CPP violations affect which decisions. Table 7-1 End Test Decisions Affected by CPP Errors Decision Type Prohibited by: Fast-Pass Low-end CPP error Retest (initiate another IM147 cycle) High-end CPP error Fast-Fail High-end CPP error End-of-Test Pass Low-end CPP error End-of-Test Fail High-end CPP error While the above decision types can be prohibited by the corresponding CPP errors, the error must occur while data for that decision were being produced in order to affect the decision. In other words, not all high-end CPP errors occurring during an IM147 test -68- ------- would necessarily invalidate failing results, nor would all low-end CPP errors invalidate passing results. If an applicable power error occurred while data required to make a particular decision were being collected, then the decision would be invalidated and the test would continue. Specifically, this would apply to fast-pass and fast-fail decisions. For example, if the vehicle's emission data were clean enough to permit a fast-pass at second 40, but there was a low power violation at second 34, then the fast-pass decision would be invalidated and the test would continue. On the other hand, if the first power violation occurred after second 40 (e.g., at second 41), then the vehicle could be fast- passed at second 40. In a related issue, the CPP error is reset at the conclusion of each IM147. As a result, low-end CPP errors occurring during the first IM147 do not prohibit pass-oriented decisions in subsequent IM147 tests. The same is true for high-end CPP errors and fail- oriented decisions. Using this logical framework, the CPP variation limits were applied to the 3,347-vehicle sample to determine how this algorithm would affect test time. Since the CPP variation limits are designed to be imposed in concert with the ą2 mph speed limits detailed in the IM240 guidance, vehicles already failing the ą2 mph speed limits were eliminated from the sample since they would be aborted regardless of the CPP outcome. Of the original 3,347 vehicles, 3,006 remained after these vehicles were eliminated from the sample. The flow chart detailed in Figure 7-1 shows how the CPP decision integrates with the fast-pass/fail and retest algorithms previously discussed in this report. Note that in cases where a CPP violation prevented a decision during the third IM147, the vehicle would, after completing the third IM147, restart the third IM147 again. For the average test time computation, it was assumed that CPP errors during the third IM147 extended the test time 146 seconds. Given the 3,006-vehicle sample, the average test time, with the fast-pass, retest, and fast- fail criteria enabled but without the CPP criteria applied, was 89.07 seconds. Once the CPP criteria were enabled, 87 vehicles' tests (of the 3,006 vehicles) were extended, increasing the average test time by 0.98 seconds (or 1.1%) to 90.05 seconds. This is less than the 2% increase projected at the end of Section 6, which makes sense given that the 2% projection was made without the fast-pass/fail and retest algorithms in place. The overall test time reductions caused by the fast-pass/fail and retest algorithms would mean that fewer errors would be committed. ------- Figure 7-1 Process for Integrating CPP Variation Limits ### -70- ------- 8. REFERENCES "Determination of Emissions Credit and Average Test Times for IM147 Testing," Prepared by Sierra Research for the U.S. Environmental Protection Agency, Report No. SR99-10-02, October 10, 1999. "Additional Study of Preconditioning Effects and Other IM240 Testing Issues," Prepared by Sierra Research for the U.S. environmental Protection Agency, Report No. SR98-02-01, February 2, 1998. Heirigs, P.L. and J. Gordon, "Preconditioning Effects on I/M Test Results Using IM240 and ASM Procedures," SAE Paper No. 962091, 1996. Whitby, R., C. Shih, W. Webster, and R. Card, "Enhanced Inspection and Maintenance Program: An Investigation of Real Time Data Utility and Dynamic Quality Assurance Instrumentation - Task 1: A Predictive Modal Regression Methodology for IM240 Fast-Pass/Fail Decisions," New York State Department of Environmental Conservation, EPA Cooperative Agreement #X992008-01-0, November 29, 1996. "Analysis of Alternate IM240 Cutpoints, Phase 2 Testing, and Exempting New Vehicle Models on Test Duration and Projected I/M Benefits," Prepared by Sierra Research for the Arizona Department of Environmental Quality, Report No. SR98-05-01, May 12, 1998. "High-Tech I/M Test Procedures, Emission Standards, Quality Control Requirements, and Equipment Specifications: IM240 and Functional Evaporative System Tests," (Revised Draft), U.S. Environmental Protection Agency, EPA-AA-RSPD-IM-96-1, June 1996. W. J. Webster and C. Shih, "A Statistically Derived Metric to Monitor Time- Speed Variability in Practical Emissions Testing," New York State Department of Environmental Conservation, presented at the 6th CRC On- Road Vehicle Emissions Workshop, March 18-20, 1996. ### -71- ------- Appendix A Startup, Intermediate, and Final IM240 and IM147 Cutpoints ------- Startup IM240 Outpoints and IM147 Cutpoints* (Composite/Phase 2 Cutpoints in g/mi, IM240 - IM147) Model Years HC CO NOx LDGV 1981-82 2.00/1.25 -2.00/1.20 60.0/48.0 - 58.0/30.0 3.0-3.3/1.2 1983-85 2.00/1.25 -2.00/1.20 30.0/24.0-30.0/15.0 3.0-3.3/1.2 1986-90 2.00/1.25 -2.00/1.00 30.0/24.0 - 30.0/10.0 3.0-3.0/1.2 1991-93 1.20/0.75 - 1.30/0.60 20.0/16.0-21.0/10.0 2.5-2.9/1.0 1994-95 1.20/0.75 - 1.20/0.60 20.0/16.0-21.0/10.0 2.5-2.7/1.0 1996+ (Tier 1) 0.80/0.50 - 0.80/0.50 15.0/12.0-15.0/7.0 2.0-2.1/0.9 LDGT1 1981-83 7.50/5.00 - 6.70/4.70 100.0/80.0 - 95.0/50.0 7.0 - 7.6/2.9 1984-85 3.20/2.00 - 2.90/2.00 80.0/64.0 - 76.0/40.0 7.0 - 7.6/2.9 1986-87 3.20/2.00-2.90/1.60 80.0/64.0-76.0/31.0 7.0 - 7.0/2.7 1988-90 3.20/2.00-2.90/1.60 80.0/64.0-76.0/31.0 3.5-3.6/1.3 1991-93 2.40/1.50-2.60/1.20 60.0/48.0-61.0/31.0 3.0-3.4/1.1 1994-95 2.40/1.50-2.40/1.20 60.0/48.0-61.0/29.0 3.0-3.2/1.1 1996+(Tier 1) 1.00/0.63 - 1.0/0.60 20.0/16.0-21.0/10.0 2.5-2.7/1.1 LDGT2 1981-83 7.50/5.00 - 6.70/4.70 100.0/80.0-95.0/50.0 7.0 - 7.6/2.9 1984-86 3.20/2.00 - 2.90/2.00 80.0/64.0 - 76.0/40.0 7.0 - 7.6/2.9 1987 3.20/2.00-2.90/1.60 80.0/64.0-76.0/31.0 7.0 - 7.6/2.7 1988-90 3.20/2.00-2.90/1.60 80.0/64.0-76.0/31.0 5.0-5.1/1.9 1991-93 2.40/1.50-2.60/1.20 60.0/48.0-61.0/31.0 4.5-5.1/1.9 1994-95 2.40/1.50-2.40/1.20 60.0/48.0-61.0/29.0 4.5-4.8/1.9 1996+ (Tier 1) 2.40/1.50-2.40/1.20 60.0/48.0-61.0/29.0 4.0-4.3/1.7 Developed for SR99-10-02 A-l ------- Intermediate IM240 Outpoints and IM147 Cutpoints Developed in This Study (Composite/Phase 2 Cutpoints in g/mi, IM240 - IM147) Model Years HC CO NOx LDGV 1981-82 1.40/0.88- 1.40/0.90 45.0/36.0 - 44.0/23.0 2.3-2.8/1.0 1983-85 1.40/0.88 - 1.40/0.90 23.0/18.0-23.0/12.0 2.3-2.8/1.0 1986-90 1.40/0.88 - 1.40/0.70 23.0/18.0-23.0/9.0 2.3-2.6/1.0 1991-93 1.00/0.63 - 1.10/0.50 18.0/14.0-18.0/9.0 2.3 - 2.6/0.9 1994-95 1.00/0.63 - 1.00/0.50 18.0/14.0- 18.0/9.0 2.3 - 2.5/0.9 1996+ (Tier 1) 0.70/0.45 - 0.80/0.40 13.0/10.0 - 15.0/6.0 1.8-2.2/0.8 LDGT1 1981-83 5.50/3.50-4.90/3.40 85.0/68.0-81.0/43.0 . 5.8 - 6.3/2.4 1984-85 2.40/1.50-2.30/1.50 60.0/48.0 - 60.0/30.0 5.8 - 6.3/2.4 1986-87 2.40/1.50-2.30/1.20 60.0/48.0 - 60.0/26.0 5.8 - 5.8/2.2 1988-89 2.40/1.50-2.30/1.20 60.0/48.0 - 60.0/26.0 3.0-3.3/1.2 1990 2.40/1.50-2.30/1.20 60.0/48.0 - 59.0/26.0 3.0-3.3/1.2 1991-93 2.00/1.25 - 2.10/1.00 50.0/40.0 -51.0/26.0 2.8-3.2/1.1 1994-95 2.00/1.25-2.00/1.00 50.0/40.0-51.0/25.0 2.8-3.0/1.1 1996+(Tier 1) 0.90/0.57- 1.30/0.60 17.0/13.0-31.0/8.0 2.2-2.7/1.0 LDGT2 1981-83 5.50/3.50 - 4.90/3.40 85.0/68.0-81.0/43.0 5.8 - 6.3/2.4 1984-86 2.40/1.50-2.30/1.50 60.0/48.0 - 60.0/30.0 5.8 - 6.3/2.4 1987 2.40/1.50-2.30/1.20 60.0/48.0 - 60.0/26.0 5.8 - 6.3/2.2 1988-90 2.00/1.50 - 2.30/1.20 60.0/48.0-60.0/26.0 4.3-4.6/1.6 1991-93 2.00/1.25 -2.20/1.00 50.0/40.0-51.0/26.0 4.0-4.6/1.6 1994-95 2.00/1.25 -2.00/1.00 50.0/40.0-51.0/25.0 4.0-4.3/1.6 1996+(Tier 1) 1.60/1.00-2.00/0.90 38.0/30.0-51.0/18.0 3.0-4.1/1.3 A-2 ------- Final IM240 Cutpoints and IM147 Cutpoints Developed in This Study (Composite/Phase 2 Cutpoints in g/mi, IM240 - IM147) Model Years HC CO NOx LDGV 1981-82 0.80/0.50 - 0.80/0.50 30.0/24.0-30.0/15.0 2.0 - 2.3/0.8 1983-85 0.80/0.50-0.80/0.50 15.0/12.0-16.0/8.0 2.0 - 2.3/0.8 1986-89 0.80/0.50 - 0.80/0.50 15.0/12.0 - 16.0/8.0 2.0 - 2.2/0.8 1990-93 0.80/0.50 - 0.80/0.50 15.0/12.0- 15.0/8.0 2.0 - 2.2/0.7 1994-95 0.80/0.50 - 0.80/0.50 15.0/12.0 - 15.0/7.0 2.0 - 2.2/0.7 1996+ (Tier 1) 0.60/0.40 -0.80/0.30 10.0/8.0- 15.0/5.0 1.5-2.2/0.6 LDGT1 1981-83 3.40/2.00-3.10/2.10 70.0/56.0 - 67.0/35.0 4.5-4.9/1.8 1984-85 1.60/1.00- 1.70/1.00 40.0/32.0-43.0/20.0 4.5-4.9/1.8 1986-87 1.60/1.00- 1.70/0.80 40.0/32.0-43.0/20.0 4.5-4.6/1.7 1988-89 1.60/1.00- 1.70/0.80 40.0/32.0-43.0/20.0 2.5-2.9/1.0 1990-93 1.60/1.00- 1.60/0.80 40.0/32.0-41.0/20.0 2.5-2.9/1.0 1994-95 1.60/1.00- 1.60/0.80 40.0/32.0-41.0/20.0 2.5-2.7/1.0 1996+ (Tier 1) 0.80/0.50- 1.60/0.50 13.0/10.0-41.0/6.0 1.8 - 2.7/0.8 LDGT2 1981-83 3.40/2.00-3.10/2.10 70.0/56.0 - 67.0/35.0 4.5-4.9/1.8 1984-86 1.60/1.00- 1.70/1.00 40.0/32.0-43.0/20.0 4.5-4.9/1.8 1987 1.60/1.00- 1.70/0.80 40.0/32.0 - 43.0/20.0 4.5-4.9/1.7 1988-91 1.60/1.00- 1.70/0.80 40.0/32.0 -43.0/20.0 3.5-4.0/1.3 1992-93 1.60/1.00- 1.70/0.80 40.0/32.0-41.0/20.0 3.5-4.0/1.3 1994-95 1.60/1.00- 1.70/0.80 40.0/32.0-41.0/20.0 3.5-3.8/1.3 1996+ (Tier 1) 0.80/0.50- 1.60/0.50 15.0/12.0-41.0/7.0 2.0-3.8/0.9 A-3 ------- Appendix B Max CO, Startup, Intermediate, and Final IM147 Failure Rates ------- Failure Rate Max CO Cutpoints HC CO NOx OVERALL V Type Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail First IM147 LDGV 207 1573 11.6% 352 1428 19.8% 236 1544 13.3% 543 1237 30.5% LDGT1 76 888 7.9% 204 760 21.2% 83 881 8.6% 281 683 29.1% LDGT2 31 572 5.1% 87 516 14.4% 44 559 7.3% 127 476 21.1% All 314 3033 9.4% 643 2704 19.2% 363 2984 10.8% 951 2396 28.4% Second IM147 LDGV 126 1654 7.1% 256 1524 14.4% 166 1614 9.3% 406 1374 22.8% LDGT1 59 905 6.1% 167 797 17.3% 70 894 7.3% 231 733 . 24.0% LDGT2 24 579 4.0% 61 542 10.1% 31 572 5.1% 91 512 15.1% All 209 3138 6.2% "484 2863 14.5% 267 3080 8.0% 728 2619 21.8% Third IM147 LDGV 114 1666 6.4% 250 1530 14.0% 155 1625 8.7% 390 1390 21.9% LDGT1 49 915 5.1% 156 808 16.2% 70 894 7.3% 219 745 22.7% LDGT2 25 578 4.1% 63 540 10.4% 30 573 5.0% 93 510 15.4% All 188 3159 5.6% 469 2878 14.0% 255 3092 7.6% 702 2645 21.0% Failure Rate By Model Year Grouping - First IM147 Max CO Cutpoints HC CO NOx OVERALL V Type Year Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 19 86 18.1% 32 73 30.5% 24 81 22.9% 55 50 52.4% 83-85 48 180 21.1% 108 120 47.4% 53 175 23.2% 144 84 63.2% 86-89 85 340 20.0% 96 329 22.6% 87 338 20.5% 169 256 39.8% 90-95 55 897 5.8% 112 840 11.8% 72 880 7.6% 171 781 18.0% 96+ 0 70 0.0% 4 66 5.7% 0 70 0.0% 4 66 5.7% LDGT1 81-85 38 222 14.6% 109 151 41.9% 54 206 20.8% 152 108 58.5% 86-89 22 200 9.9% 57 165 25.7% 17 205 7.7% 73 149 32.9% 90-95 16 434 3.6% 38 412 8.4% 11 439 2.4% 55 395 12.2% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 18 76 19.1% 44 50 46.8% 16 78 17.0% 55 39 58.5% 86-89 8 56 12.5% 13 51 20.3% 13 51 20.3% 27 37 42.2% 88-95 5 422 1.2% 30 397 7.0% 15 412 3.5% 45 382 10.5% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 314 3033 9.4% 643 2704 19.2% 363 2984 10.8% 951 2396 28.4% B-l ------- Failure Rate By Model Year Grouping - Second IM147 Max CO Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 13 92 12.4% 29 76 27.6% 18 87 17.1% 49 '56 46.7% 83-85 32 196 14.0% 81 147 35.5% 43 185 18.9% ' 116 112 50.9% 86-89 57 368 13.4% 82 343 19.3% 67 358 15.8% 141 284 33.2% 90-95 24 928 2.5% 64 888 6.7% 38 914 4.0% 100 852 10.5% 96+ 0 70 0.0% 0 70 0.0% 0 70 0.0% 0 70 0.0% LDGT1 81-85 33 227 12.7% 99 161 38.1% 48 212 18.5% 139 121 53.5% 86-89 19 203 8.6% 48 174 21.6% 10 212 4.5% 59 163 26.6% 90-95 7 443 1.6% 20 430 4.4% 11 439 2.4% 32 418 7.1% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 16 78 17.0% 40 54 42.6% 14 80 14.9% 53 41 56.4% 86-89 5 59 7.8% 11 53 17.2% 11 53 17.2% 22 42 34.4% 88-95 3 424 0.7% 10 417 2.3% 6 421 1.4% 16 411 3.7% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 209 3138 6.2% 484 2863 14.5% 267 3080 8.0% 728 2619 21.8% Failure Rate By Model Year Grouping - Third IM147 Max CO Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 13 92 12.4% 29 76 27.6% 16 89 15.2% 46 59 43.8% 83-85 29 199 12.7% 79 149 34.6% 39 189 17.1% 110 118 48.2% 86-89 48 377 11.3% 77 348 18.1% 65 360 15.3% 137 288 32.2% 90-95 24 928 2.5% 64 888 6.7% 35 917 3.7% 96 856 10.1% 96+ 0 70 0.0% 1 69 1.4% 0 70 0.0% 1 69 1.4% LDGT1 81-85 31 229 11.9% 95 165 36.5% 47 213 18.1% 134 126 51.5% 86-89 15 207 6.8% 43 179 19.4% 12 210 5.4% 56 166 25.2% 90-95 3 447 0.7% 18 432 4.0% 10 440 2.2% 28 422 6.2% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 16 78 17.0% 40 54 42.6% 13 81 13.8% 53 41 56.4% 86-89 6 58 9.4% 13 51 20.3% 10 54 15.6% 23 41 35.9% 88-95 3 424 0.7% 10 417 2.3% 7 420 1.6% 17 410 4.0% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 188 3159 5.6% 469 2878 14.0% 255 3092 7.6% 702 2645 21.0% B-2 ------- Failure Rate Startup Cutpoints V Type HC CO NOx OVERALL Fail Pass % Fail Fail | Pass | % Fail Fail | Pass % Fail Fail Pass % Fail First 1M147 LDGV 180 1600 10.1% 179 1601 10.1% 216 1564 12.1% 416 1364 23.4% LDGT1 75 889 7.8% 37 927 3.8% 84 880 8.7% 166 798 17.2% LDGT2 56 547 9.3% 37 566 6.1% 45 558 7.5% 103 500 17.1% All 311 3036 9.3% 253 3094 7.6% 345 3002 10.3% 685 2662 20.5% Second IM147 LDGV 109 1671 6.1% 133 1647 7.5% 152 1628 8.5% 291 1489 16.3% LDGT1 51 913 5.3% 40 924 4.1% 51 913 5.3% 115 849 11.9% LDGT2 32 571 5.3% 31 572 5.1% 34 569 5.6% 74 529 12.3% All 192 3155 5.7% 204 3143 6.1% 237 3110 7.1% 480 2867 14.3% Third IM147 LDGV 106 1674 6.0% 130 1650 7.3% 141 1639 7.9% 281 1499 15.8% LDGT1 52 912 5.4% 33 931 3.4% 48 916 5.0% 108 856 11.2% LDGT2 33 570 5.5% 30 573 5.0% 28 575 4.6% 67 536 11.1% All 191 3156 5.7% 193 3154 5.8% 217 3130 6.5% 456 2891 13.6% Failure Rate By Model Year Grouping - First IM147 Startup Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 30 75 28.6% 13 92 12.4% 24 81 22.9% 48 57 45.7% 83-85 56 172 24.6% 75 153 32.9% 49 179 21.5% 122 106 53.5% 86-89 57 368 13.4% 50 375 11.8% 72 353 16.9% 128 297 30.1% 90-95 36 916 3.8% 37 915 3.9% 71 881 7.5% 114 838 12.0% 96+ 1 69 1.4% 4 66 5.7% 0 70 0.0% 4 66 5.7% LDGT1 81-85 44 216 16.9% 24 236 9.2% 21 239 8.1% 74 186 28.5% 86-89 24 198 10.8% 11 211 5.0% 29 193 13.1% 53 169 23.9% 90-95 7 443 1.6% 2 448 0.4% 32 418 7.1% 37 413 8.2% 96+ 0 32 0.0% 0 32 0.0% 2 30 6.3% 2 30 6.3% LDGT2 81-85 29 65 30.9% 26 68 27.7% 9 85 9.6% 41 53 43.6% 86-89 14 50 21.9% 6 58 9.4% 7 57 10.9% 20 44 31.3% 88-95 13 414 3.0% 5 422 1.2% 29 398 6.8% 42 385 9.8% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 311 3036 9.3% 253 3094 7.6% 345 3002 10.3% 685 2662 20.5% B-3 ------- Failure Rate By Model Year Grouping - Second IM147 Startup Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 24 81 22.9% 13 92 12.4% 18 87 17.1% 38 67 36.2% 83-85 35 193 15.4% 53 175 23.2% 43 185 18.9% 92 136 40.4% 86-89 36 389 8.5% 42 383 9.9% 55 370 12.9% 101 324 23.8% 90-95 14 938 1.5% 25 927 2.6% 36 916 3.8% 60 892 6.3% 96+ 0 70 0.0% 0 70 0.0% 0 70 0.0% 0 70 0.0% LDGT1 81-85 31 229 11.9% 23 237 8.8% 18 242 6.9% 58 202 22.3% 86-89 18 204 8.1% 14 208 6.3% 18 204 8.1% 39 183 17.6% 90-95 2 448 0.4% 3 447 0.7% 14 436 3.1% 17 433 3.8% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 20 74 21.3% 23 71 24.5% 9 85 9.6% 35 59 37.2% 86-89 8 56 12.5% 5 59 7.8% 8 56 12.5% 16 48 25.0% 88-95 4 423 0.9% 3 424 0.7% 17 410 4.0% 23 404 5.4% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 192 3155 5.7% 204 3143 6.1% 237 3110 7.1% 480 2867 14.3% Failure Rate By Model Year Grouping - Third IM147 Startup Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 23 82 21.9% 13 92 12.4% 14 91 13.3% 35 70 33.3% 83-85 33 195 14.5% 56 172 24.6% 39 189 17.1% 90 138 39.5% 86-89 35 390 8.2% 36 389 8.5% 56 369 13.2% 96 329 22.6% 90-95 15 937 1.6% 24 928 2.5% 32 920 3.4% 59 893 6.2% 96+ 0 70 0.0% 1 69 1.4% 0 70 0.0% 1 69 1.4% LDGT1 81-85 34 226 13.1% 20 240 7.7% 17 243 6.5% 56 204 21.5% 86-89 17 205 7.7% 12 210 5.4% 16 206 7.2% 35 187 15.8% 90-95 1 449 0.2% 1 449 0.2% 14 436 3.1% 16 434 3.6% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 21 73 22.3% 22 72 23.4% 7 87 7.4% 33 61 35.1% 86-89 9 55 14.1% 6 58 9.4% 7 57 10.9% 16 48 25.0% 88-95 3 424 0.7% 2 425 0.5% 14 413 3.3% 18 409 4.2% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 191 3156 5.7% 193 3154 5.8% 217 3130 6.5% 456 2891 13.6% B-4 ------- Failure Rate Intermediate Cut >oints V Type HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail First IM147 LDGV 286 1494 16.1% 238 1542 13.4% 281 1499 15.8% 533 1247 29.9% LDGT1 121 843 12.6% 62 902 6.4% 122 842 12.7% 247 717 25.6% LDGT2 84 519 13.9% 47 556 7.8% 71 532 11.8% 142 461 23.5% All 491 2856 14.7% 347 3000 10.4% 474 2873 14.2% 922 2425 27.5% Second EM147 LDGV 170 1610 9.6% 161 1619 9.0% 214 1566 12.0% 386 1394 21.7% LDGT1 87 877 9.0% 57 907 5.9% 78 886 8.1% 176 788 18.3% LDGT2 46 557 7.6% 40 563 6.6% 48 555 8.0% 100 503 16.6% All 303 3044 9.1% 258 3089 7.7% 340 3007 10.2% 662 2685 19.8% - Third IM147 LDGV 157 1623 8.8% 161 1619 9.0% 196 1584 11.0% 367 1413 20.6% LDGT1 80 884 8.3% 53 911 5.5% 73 891 7.6% 165 799 17.1% LDGT2 40 563 6.6% 42 561 7.0% 42 561 7.0% 93 510 15.4% All 277 3070 8.3% 256 3091 7.6% 311 3036 9.3% 625 2722 18.7% Failure Rate By Model Year Grouping - First IM147 Intermediate Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 52 53 49.5%. 20 85 19.0% 27 78 25.7% 64 41 61.0% 83-85 88 140 38.6% 91 137 39.9% 64 164 28.1% 146 82 64.0% 86-89 94 331 22.1% 65 360 15.3% 100 325 23.5% 169 256 39.8% 90-95 51 901 5.4% 58 894 6.1% 90 862 9.5% 150 802 15.8% 96+ 1 69 1.4% 4 66 5.7% 0 70 0.0% 4 66 5.7% LDGT1 81-85 71 189 27.3% 42 218 16.2% 37 223 14.2% 117 143 45.0% 86-89 36 186 16.2% 14 208 6.3% 39 183 17.6% 71 151 32.0% 90-95 14 436 3.1% 6 444 1.3% 44 406 9.8% 57 393 12.7% 96+ 0 32 0.0% 0 32 0.0% 2 30 6.3% 2 30 6.3% LDGT2 81-85 37 57 39.4% 32 62 34.0% 17 77 18.1% 53 41 56.4% 86-89 19 45 29.7% 9 55 14.1% 10 54 15.6% 24 40 37.5% 88-95 28 399 6.6% 6 421 1.4% 44 383 10.3% 65 362 15.2% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 491 2856 14.7% 347 3000 10.4% 474 2873 14:2% 922 2425 27.5% B-5 ------- Failure Rate By Model Year Grouping - Second IM147 Intermediate Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 36 69 34.3% 15 90 14.3% 23 82 21.9% 49 56 46.7% 83-85 51 177 22.4% 66 162 28.9% 60 168 26.3% 118 110 51.8% 86-89 61 364 14.4% 50 375 11.8% 76 349 17.9% 134 291 31.5% 90-95 22 930 2.3% 30 922 3.2% 55 897 5.8% 85 867 8.9% 96+ 0 70 0.0% 0 70 0.0% 0 70 0.0% 0 70 0.0% LDGT1 81-85 57 203 21.9% 34 226 13.1% 31 229 11.9% 95 165 36.5% 86-89 24 198 10.8% 18 204 8.1% 25 197 11.3% 52 170 23.4% 90-95 6 444 1.3% 5 445 1.1% 21 429 4.7% 28 422 6.2% 96+ Q 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 29 65 30.9% 32 62 34.0% 15 79 16.0% 50 44 53.2% 86-89 10 54 15.6% 5 59 7.8% 11 53 17.2% 19 45 29.7% 88-95 7 420 1.6% 3 424 0.7% 22 405 5.2% 31 396 7.3% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 303 3044 9.1% 258 3089 7.7% 340 3007 10.2% 662 2685 19.8% Failure Rate By Model Year Grouping - Third IM147 Intermediate Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 34 71 32.4% 16 89 15.2% 20 85 19.0% 48 57 45.7% 83-85 48 180 21.1% 66 162 28.9% 58 170 25.4% 113 115 49.6% 86-89 56 369 13.2% 49 376 11.5% 72 353 16.9% 128 297 30.1% 90-95 19 933 2.0% 29 923 3.0% 46 906 4.8% 77 875 8.1% 96+ 0 70 0.0% 1 69 1.4% 0 70 0.0% 1 69 1.4% LDGT1 81-85 56 204 21.5% 34 226 13.1% 29 231 11.2% 92 168 35.4% 86-89 21 201 9.5% 16 206 7.2% 25 197 11.3% 50 172 22.5% 90-95 3 447 0.7% 3 447 0.7% 18 432 4.0% 22 428 4.9% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 25 69 26.6% 31 63 33.0% 14 80 14.9% 47 47 50.0% 86-89 10 54 15.6% 8 56 12.5% 8 56 12.5% 19 45 29.7% 88-95 5 422 1.2% 3 424 0.7% 20 407 4.7% 27 400 6.3% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 277 3070 8.3% 256 3091 7.6% 311 3036 9.3% 625 2722 18.7% B-6 ------- Failure Rate Final Cutpoints V Type HC CO NOx OVERALL Fail Pass % Fail Fail Pass | % Fail Fail | Pass % Fail Fail Pass % Fail First IM147 LDGV 499 1281 28.0% 346 1434 19.4% 396 1384 22.2% 733 1047 41.2% LDGT1 210 754 21.8% 109 855 11.3% 184 780 19.1% 348 616 36.1% LDGT2 135 468 22.4% 66 537 10.9% 120 483 19.9% 219 384 36.3% All 844 2503 25.2% 521 2826 15.6% 700 2647 20.9% 1300 2047 38.8% Second IM147 LDGV 309 1471 17.4% 247 1533 13.9% 296 1484 16.6% 539 1241 30.3% LDGT1 137 827 14.2% 91 873 9.4% 134 830 13.9% 256 708 26.6% LDGT2 77 526 12.8% 56 547 9.3% 88 515 14.6% 157 446 26.0% All 523 2824 15.6% 394 2953 11.8% 518 2829 15.5% 952 2395 28.4% Third IM147 LDGV 280 1500 15.7% 240 1540 13.5% 277 1503 15.6% 507 1273 28.5% LDGT1 120 844 12.4% 85 879 8.8% 127 837 13.2% 239 725 24.8% LDGT2 73 530 12.1% 56 547 9.3% 82 521 13.6% , 149 454 24.7% All 473 2874 14.1% 381 2966 11.4% 486 2861 14.5% 895 2452 26.7% Failure Rate By Model Year Grouping - First IM147 Final Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 70 35 66.7% 30 75 28.6% 39 66 37.1% 78 27 74.3% 83-85 148 80 64.9% 120 108 52.6% 81 147 35.5% 183 45 80.3% 86-89 179 246 42.1% 102 323 24.0% 144 281 33.9% 247 178 58.1% 90-95 100 852 10.5% 90 862 9.5% 132 820 13.9% 220 732 23.1% 96+ 2 68 2.9% 4 66 5.7% 0 70 0.0% 5 65 7.1% LDGT1 81-85 116 144 44.6% 69 191 26.5% 70 190 26.9% 165 95 63.5% 86-89 71 151 32.0% 29 193 13.1% 53 169 23.9% 104 118 46.8% 90-95 23 427 5.1% 11 439 2.4% 59 391 13.1% 77 373 17.1% 96+ 0 32 0.0% 0 32 0.0% 2 30 6.3% 2 30 6.3% LDGT2 81-85 50 44 53.2% 41 53 43.6% 32 62 34.0% 73 21 77.7% 86-89 29 35 45.3% 14 50 21.9% 20 44 31.3% 42 22 65.6% 88-95 56 371 13.1% 11 416 2.6% 68 359 15.9% 104 323 24.4% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 844 2503 25.2% 521 2826 15.6% 700 2647 20.9% 1300 2047 38.8% B-7 ------- Failure Rate By Model Year Grouping - Second IM147 Final Outpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 56 49 53.3% 26 79 24.8% 37 68 35.2% 74 . 31 70.5% 83-85 104 124 45.6% 91 137 39.9% 79 149 34.6% 157 71 68.9% 86-89 110 315 25.9% 83 342 19.5% 101 324 23.8% 181 244 42.6% 90-95 39 913 4.1% 47 905 4.9% 79 873 8.3% 127 825 13.3% 96+ 0 70 0.0% 0 70 0.0% 0 70 0.0% 0 70 0.0% LDGT1 81-85 90 170 34.6% 63 197 24.2% 65 195 25.0% 145 115 55.8% 86-89 36 186 16.2% 22 200 9.9% 39 183 17.6% 69 153 31.1% 90-95 11 439 2.4% 6 444 1.3% 29 421 6.4% 41 409 9.1% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 46 48 48.9% 41 53 43.6% 31 63 33.0% 72 22 76.6% 86-89 14 50 21.9% 12 52 18.8% 18 46 28.1% 33 31 51.6% 88-95 17 410 4.0% 3 424 0.7% 39 388 9.1% 52 375 12.2% 96+ 0 18 0.0% 0 18- 0.0% 0 18 0.0% 0 18 0.0% ALL 523 2824 15.6% 394 2953 11.8% 518 2829 15.5% 952 2395 28.4% Failure Rate By Model Year Grouping - Third IM147 Final Cutpoints V Type Year HC CO NOx OVERALL Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail Fail Pass % Fail LDGV 81-82 53 52 50.5% 25 80 23.8% 32 73 30.5% 67 38 63.8% 83-85 91 137 39.9% 88 140 38.6% 81 147 35.5% 156 72 68.4% 86-89 100 325 23.5% 76 349 17.9% 96 329 22.6% 171 254 40.2% 90-95 36 916 3.8% 50 902 5.3% 68 884 7.1% 112 840 11.8% 96+ 0 70 0.0% 1 69 1.4% 0 70 0.0% 1 69 1.4% LDGT1 81-85 83 177 31.9% 57 203 21.9% 61 199 23.5% 137 123 52.7% 86-89 31 191 14.0% 24 198 10.8% 36 186 16.2% 65 157 29.3% 90-95 6 444 1.3% 4 446 0.9% 29 421 6.4% 36 414 8.0% 96+ 0 32 0.0% 0 32 0.0% 1 31 3.1% 1 31 3.1% LDGT2 81-85 44 50 46.8% 41 53 43.6% 31 63 33.0% 73 21 77.7% 86-89 16 48 25.0% 11 53 17.2% 17 47 26.6% 32 32 50.0% 88-95 13 414 3.0% 4 423 0.9% 34 393 8.0% 44 383 10.3% 96+ 0 18 0.0% 0 18 0.0% 0 18 0.0% 0 18 0.0% ALL 473 2874 14.1% 381 2966 11.4% 486 2861 14.5% 895 2452 26.7% B-8 ------- Appendix C IM147 Regression Coefficients ------- APPENDIX C IM147 Composite Regression Coefficients, HC, 1981 to 1985 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT1 81-85 0.73806 0.95039 13.5112 92 LDGT1 81-85 0.52814 0.34501 2.7265 5.5143 P3 LDGT1 81-85 0.40889 0.23152 0.5547 3.2006 6.2124 P4 LDGT1 81-85 0.38717 0.22535 1.0106 3.4095 3.2694 3.5979 P5 LDGT1 81-85 0.38872 0.23976 -0.8175 2.5639 3.1218 2.0148 3.4755 P6 LDGT1 81-85 0.34978 0.25562 -2.8842 2.7609 1.6633 1.5365 2.3446 3.2626 P7 LDGT1 81-85 0.33929 0.24395 -2.5969 2.4215 2.0512 1.644 0.5818 2.2896 2.9199 P8 LDGT1 81-85 0.30707 0.21404 -0.9821 1.8007 1.4784 1.745 0.0989 0.8955 2.4381 4.82 P9 LDGT1 81-85 0.30661 0.21485 0.9303 1.7873 1.3395 1.7754 0.0281 0.6238 2.4791 4.5115 0.8987 P10 U0GT1 81-85 0.30246 0.21468 1.189 1.8114 1.1577 2.1188 0.3265 0.5235 2.6637 4.255 -0.5878 2.1284 P11 LDGT1 81-85 0.25475 0.05761 -0.2293 0.7411 1.0904 1.5866 0.0582 1.3727 1.9974 2.4699 0.822 1.4541 1.8008 P12 LDGT1 81-85 0.24251 0.05503 -0.1605 0.7753 0.9307 1.635 0.0664 1.336 2.0956 2.1337 0.2754 2.1379 1.2063 1.1106 P13 LDGT1 81-85 0.19345 0.05879 1.3617 0.2166 0.9329 1.5508 -0.1316 1.2355 2.3977 1.6874 0.0101 0.9701 0.9488 0.4982 3.2351 P14 LDGT1 81-85 0.1631 0.04219 1.279 0.1384 1.1759 1.4082 0.0948 1.1156 2.0131 1.2938 0.729 1.1342 0.4477 -0.3274 1.562 4.0745 P15 . LDGT1 81-85 0.13928 0.03346 1.8241 0.0799 1.2164 1.2299 0.2827 0.8749 2.2275 0.6681 0.8755 0.7239 0.5659 0.1723 0.7424 1.1987 2.7671 P16 LDGT1 81-85 0.13703 0.03021 1.8786 0.023 1.2823 1.1806 0.305 0.8514 2,2773 0.5758 0.8491 0.7433 0.5775 0.2633 0.6459 1.2728 2.0331 2.1897 P17 LDGT1 81-65 0.09132 0.00475 1.4179 0.434 0.6904 0.7838 1.0571 0.616 1.5938 0.7518 0.2064 0.8618 0.5896 0.4328 0.7951 1.3837 1.3066 0.6626 2.7633 P18 LDGT1 81-85 0.07314 -0.00235 1.5086 0.5696 0.7494 0.8068 0.8363 0.5804 1.3972 0.5907 0.4057 0.6221 0.5658 0.584 0.7038 0.8587 0.988 0.3277 1.9729 1.7883 P19 LDGT1 81-85 0.06979 -0.00621 1.375 0.6281 0.6996 0.852 0.9084 0.5029 1.4414 0.5441 0.3942 0.5659 0.6012 0.6081 0.7093 0.7313 0.6296 0.2443 1.7266 1.5869 0.729 IM147 Phase 2 Regression Coefficients, HC, 1981 to 1985 Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg- Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 81-85 0.5023 0.12381 4.4447 B12 L0GT1 81-85 0.48187 0.11013 3.3244 1.9633 B13 LDGT1 81-85 0.35264 0.07757 1.7904 -1.1796 5.9021 B14 LDGT1 81-85 0.32195 0.06184 0.9978 -1.0185 3.8122 5.5161 B15 LDGT1 81-85 0.2826 0.04064 1.0928 0.0499 1.7408 0.7784 4.7049 B16 LDGT1 81-85 0.27957 0.0348 1.0816 0.2105 1.5379 0.9523 3.4207 3.6479 B17 LDGT1 81-85 0.14692 -0.01627 1.0035 0.5208 1.1766 2.2402 1.4703 0.4909 4.6919 B18 L0GT1 81-85 0.11692 -0.02022 0.9177 0.8016 0.9222 1.5324 0.8911 0.0868 3.252 2.7658 B19 LDGT1 81-85 0.11286 -0.02299 0.9799 0.8216 0.9206 1.4305 0.4043 0.0294 2.8156 2.4898 0.9698 ------- IM147 Composite Regression Coefficients, HC, 1986 to 1989 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 PI LDGT1 86-69 0.49429 0.5423 17.2823 P2 L0GT1 86-89 0.37436 0.2461 2.2738 4.5192 P3 LDGT1 86-89 0.33313 0.22224 0.5252 2.5885 4.6599 P4 LDGT1 86-69 0.29595 0.20426 0.5943 2.3293 2.3375 3.7634 P5 LDGT1 86-69 0.2936 0.19895 0.6763 2.0377 2.3025 2.9749 1.8038 P6 LOGT1 86-69 0.28058 0.20644 -0.6191 2.2969 1.2069 2.4549 -0.1084 4.2386 P7 LDGT1 86-69 0.27458 0.19491 -0.0117 2.0167 1.3562 2.7382 -1.5023 2.6456 2.8231 P8 LDGT1 86-89 0.26273 0.17858 0.9768 1.9562 1.0027 2.2073 -1.5778 1.5977 2.003 3.2606 P9 LOGT1 86-69 0.25683 0.17484 0.544 2.0132 0.9204 2.1628 -2.2132 0.7045 2.0543 2.3838 3.4294 P10 LDGT1 86-89 0.25153 0.17214 0.4457 2.0253 0.8623 1.9864 -2.2523 0.4119 2.2232 1.6726 1.6078 3.6006 P11 LDGT1 86-69 0.19659 0.07478 2.1606 0.6298 1.0463 1.2006 -1.6683 1.7941 1.4349 0.7533 1.4174 2.3957 1.957 P12 LDGT1 66-89 0.17271 0.05609 2.4645 0.4889 0.5378 1.0905 0.6636 1.2017 1.3698 1.1634 0.8772 3.332 1.2131 1.5195 P13 LDGT1 66-89 0.13816 0.05847 2.1195 0.4529 0.7286 1.2692 -0.7285 1.427 1.284 0.6442 0.4274 2.4657 0.9117 0.2322 2.6432 P14 LDGT1 86-69 0.12661 0.04656 2.2369 0.3811 0.6514 1.4033 -0.4687 1.3427 1.1456 0.9582 0.0864 2.2192 0.7164 0.4248 1.479 2.6071 P15 LDGT1 86-89 0.11362 0.03727 2.1223 0.3411 0.6537 1.4601 -0.202 1.1529 1.232 0.6887 -0.2172 2.1843 0.7658 0.6133 0.9282 0.7457 2.1404 P16 LDGT1 86-69 0.10769 0.03525 1.9852 0.3076 0.7867 1.4374 -0.022 1.1859 1.1329 0.515 -0.1541 2.0989 0.7572 0.7681 0.784 0.499 0.7808 4.7323 P17 LDGT1 86-89 0.05356 0.01318 1.0985 0.6236 0.7621 0.7653 0.5658 1.0327 0.7036 0.4129 0.7975 1.2149 0.7601 0.6019 0.8454 0.937 0.7894 0.9132 3.009 P18 LDGT1 86-89 0.04127 0.00269 1.2483 0.5985 0.676 0.9508 0.3871 1.095 0.7732 0.6052 0.8039 0.761 0.6892 0.7308 0.7205 0.82 0.9126 0.0726 2.0163 1.4844 P19 LDGT1 86-89 0.03395 0.00397 1.0933 0.6168 0.695 0.9026 0.4507 1.0319 0.9511 0.5369 0.7983 0.6339 0.6758 0.6909 0.7183 0.7503 0.7572 -0.3355 1.6677 1.2005 0.8458 IM147 Phase 2 Regression Coefficients, HC, 1986 to 1989 Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant. C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 ;> C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 86-69 0.29834 0.1158 3.3457 B12 LDGT1 86-69 0.26979 0.09053 2.3987 1.9564 B13 LDGT1 86-89 0.20617 0.08073 1.4658 0.0492 4.1671 B14 LDGT1 86-89 0.19052 0.06515 1.1799 0.3012 2.4794 3.6519 B15 LDGT1 86-89 0.16862 0.04542 1.1829 0.6531 1.4949 0.6859 3.2681 B16 LDGT1 86-69 0.16083 0.04255 1.1483 0.9014 1.2632 0.3783 1.3376 6.662 B17 LDGT1 86-89 0.07626 0.01566 1.0548 0.7801 1.1751 1.2709 1.2392 0.9683 4.2486 B18 L0GT1 86-89 0.05999 0.00012 0.9466 0.9733 1.0114 1.038 1.3674 -0.1144 2.9584 1.987 B19 LDGT1 86-89 0.04869 0.00402 0.9313 0.9308 0.9901 0.9768 1.1033 -0.6703 2.4005 1.5697 1.23 ------- IM147 Composite Regression Coefficients, HC, 1990 to 1995 Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C16 C19 PI LDGT1 90-95 0.37406 0.22818 17.1854 P2 LDGT1 90-95 0.23507 0.05902 -2.145 5.1262 P3 LDGT1 90-95 0.21189 0.05784 -1.1729 3.3619 3.4597 P4 LDGT1 90-95 0.20276 0.05212 -2.7245 3.2374 1.6019 3.7987 P5 LDGT1 90-95 0.19505 0.05057 -3.0392 2.7597 1.1351 2.24 3.8993 P6 LDGT1 90-95 0.18478 0.05162 -3.3501 2,6388 0.6397 1.4629 1.9249 4.9984 P7 LDGT1 90-95 0.17282 0.04206 -2.9083 2.3805 0.6434 1.9124 -0.3621 3.0178 4.1577 P8 LDGT1 90-95 0.16217 0.03882 -1.4065 1.7566 1.0327 1.3221 -0.6046 1.7837 2.9574 4.5516 P9 LDGT1 90-95 0.15997 0.03954 -1.5284 1.8223 0.7356 1.6157 -0.8447 0.9682 2.5193 3.7565 3.0158 P10 LDGT1 90-95 0.15899 0.03962 -1.3222 1.7605 0.8841 1.526 -0.7219 0.6441 2.5385 3.3964 1.3921 2.2962 P11 LDGT1 30-95 0.1242 0.00888 0.4575 0.8942 1.2216 1.078 -0.2219 0.7626 1.9585 1.2513 1.7074 1.5915 1.5461 P12 LDGT1 90-95 0.108 0.00853 0.5566 0.7543 0.7122 1.489 -0.1897 0.6244 1.9355 0.7067 0.8076 2.4795 0.9978 1.8755 P13 LDGT1 90-95 0.08846 0.01744 0.8836 0.5089 0.6095 1.7266 0.0115 0.5537 1.4386 0.4008 0.5776 1.904 0.7549 0.8219 2.3703 P14 LDGT1 90-95 0.07835 0.01772 0.858 0.474 0.6951 1.5244 0.3476 0.7542 1.4394 0.4383 0.7477 2.2305 0.5943 0.9555 0.9132 2.1616 P15 LDGT1 90-95 0.06853 0.01326 0.7115 0.5394 0.7616 1.4283 0.4159 0.4769 1.3363 0.4071 0.6623 1.647 0.6826 0.8356 0.5762 0.7447 2.2358 P16 LDGT1 90-95 0.0593 0.01319 0.6956 0.6235 0.517 1.2912 0.6027 0.5332 1.4561 0.3365 0.4783 1.6692 0.6904 0.6659 0.7576 0.7049 0.1973 5.5813 P17 LDGT1 90-95 0.04054 0.0066 0.721 0.7143 0.508 0.9341 0.7735 0.3397 1.3289 0.8072 0.4656 0.9685 0.6693 0.7801 0.6168 0.8507 0.941 2.2139 2.3062 P18 LDGT1 90-95 0.03216 0.00497 0.6809 0.6934 0.5834 0.9437 0.759 0.6708 0.9442 0.9559 0.356 0.6599 0.6671 0.8103 0.5884 1.0269 0.3629 1.3443 1.5874 1.5359 P19 LDGT1 90-95 0.02955 0.00468 0.7103 0.6964 0.6123 0.9353 0.834 0.5442 0.9229 0.8944 0.4112 0.8891 0.6698 0.7038 0.6909 0.7838 0.3123 0.6669 1.5046 1.3036 0.7468 IM147 Phase 2 Regression Coefficients, HC, 1990 to 1995 Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 811 LDGT1 90-95 0.19826 0.0335 3.1539 B12 LDGT1 90-95 0.16235 0.0226 1.7885 3.1368 B13 LDGT1 90-95 0.12683 0.02866 1.1305 1.1678 3.5627 B14 LDGT1 90-95 0.11708 0.02986 0.9821 1.3215 2.042 2.4823 B15 LOGT1 90-95 0.09769 0.0205 1.024 1.124 1.0503 0.5065 3.5529 B16 LDGT1 90-95 0.08453 0.02072 1.0421 1.0995 1.3341 0:4695 0.6556 7.7639 B17 LDGT1 90-95 0.05719 0.01154 1.0027 0.9977 1.0045 0.9826 1.4994 2.7725 3.1834 B18 LDGT1 90-95 0.04461 0.008 0.9558 1.0711 0.8811 1.2953 0.5862 1.7403 2.1672 2.144 B19 LDGT1 90-95 0.04108 0.00785 0.9823 0.9356 1.0185 0.9731 0.499 0.8303 2.0505 1.8394 1.0152 ------- IM147 Composite Regression Coefficients, HC, 1996 and Newer Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 CIS C19 P1 LDGT1 96+ 0.05316 0.03402 18.1995 P2 LDGT1 96* 0.04307 0.02411 0.2121 2.417 P3 LDGT1 96+ 0.03308 0.02193 2.6279 1.1433 2.9158 P4 LDGT1 96+ 0.0321 0.02215 3.4627 1.1835 1.5689 1.9266 P5 LDGT1 96+ 0.0267 0.01928 1.088 0.5294 2.0247 -3.8687 8.4714 P6 LDGT1 96+ 0.02641 0.01933 0.5547 0.4887 1.8979 -4,0225 7,8031 1,8996 P7 LDGT1 98+ 0.02534 0.01654 0.953 0.3268 2.2496 -3.2672 4.7342 1.2515 3.3211 P8 LDGT1 96+ 0.02524 0.01597 -1.5963 0.6417 2.0178 -4.3447 3.8465 1.9882 2.8209 1.9699 P9 LDGT1 96+ 0.02522 0.0165 -0.9944 0.5289 2.1342 4.3968 4.6568 1.7412 2.8435 2.4683 -2.6396 P10 LDGT1 96+ 0.02529 0.01642 -1.4673 0.5347 2.3503 -5.1056 4.7009 2.0105 2.7607 2.6563 -1.6108 -0.7547 P11 LDGT1 96+ 0.01755 0.00739 3.4599 0.1081 1.652 -1.1947 3.4359 0.4622 0.6672 0.794 -0.5304 -0.2246 1.7144 P12 LDGT1 96+ 0.0151 0.00659 3.8459 0.4104 0.2567 1.1211 0.4848 0.8471 1.2853 0.5367 1.1122 0.0179 1.1163 1.9287 P13 LDGT1 96+ Ś 0.01241 0.00389 1.7514 0.4378 1.7539 -0.5187 -0.8979 0.5987 2.4378 1.2817 1.9041 -1.2482 0.9844 0.1784 2.6171 P14 LDGT1 96+ 0.01176 0.00367 1.875 0.4721 1.4005 0.0285 -0.4781 0.4911 2.4771 0.9339 1.5111 -0.5207 0.7769 0.1877 2.2359 1.1134 P15 LDGT1 96+ 0.00769 0.00241 1.0482 0.6074 0.9552 -0.3978 1.4364 1.1394 0.7747 1.4416 0.0533 -0.1206 0.7233 0.4499 1.3519 0.2003 2.2561 P16 LDGT1 96+ 0.00774 0.0024 1.0165 0.6108 0.9582 -0.4078 1.4359 1.1436 0.7653 1.4501 0.0568 -0.1213 0.7218 0:4484 1.3528 0.1992 2.2428 0.0735 P17 LDGT1 96+ 0.00585 0.00306 1.3687 0.5499 1.0977 0.7636 0.4468 0.6346 0.7879 0.6314 0.3617 0.6995 0.7613 0.3632 1.0963 0.709 0.7912 1.2412 3.4916 P18 LDGT1 96+ 0.00432 0.00121 1.569 0.5806 0.9368 0.6842 0.9482 0.5032 0.7086 0.6239 0.5212 0.4646 0.6817 0.6338 0.8277 0.6983 0.744 0.4028 2.2628 1.7838 P19 LDGT1 96+ 0.00416 0.00114 1.2715 0.6216 0.8437 1.0272 0.6835 0.4345 0.8099 0.3982 0.7078 0.75 0.6649 0.7239 0.6761 0.8112 0.5347 0.2273 2.2233 1.3046 0.5973 IM147 Phase 2 Regression Coefficients, HC, 1996 and Newer Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5- C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 96+ 0.02393 0.00865 2.4226 B12 L0GT1 96+ 0.02095 0.00863 1.6528 1.9763 B13 LDGT1 96+ 0.01824 0.00691 1.444 0.667 2.5512 B14 LDGT1 96+ 0.01653 0.00607 1.2007 0.3963 2.3614 1.9099 B15 LDGT1 96+ 0.01077 0.00256 1.0854 0.6091 1.8127 0.4665 3.018 B16 LDGT1 96+ 0.01076 0.00241 1.0549 0.6405 1.7899 0.4709 2.8141 0.9908 B17 LDGT1 96+ 0.00794 0.00414 0.8756 0.7439 1.2805 1.0073 1.458 1.4709 3.9228 B18 LDGT1 96+ 0.00581 0.00161 0.8235 0.9534 1.0124 1.0609 1.0031 0.6433 3.018 2.4016 B19 LDGT1 96+ 0.00549 0.00161 0.823 1.0394 0.8747 1.0551 0.9342 0.2958 2.5227 1.8686 0.7675 ------- 0 1 U\ 1 M147 Composite Repression Coefficients, HC, 1981 to 1985 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 CIS C16 C17 C18 C19 P1 LDGT2 81-85 0.92059 1.08952 14.1316 92 LDGT2 81-65 0.49433 0.24032 2.1062 6.4379 P3 LDGT2 81*85 0.47471 0.22581 0.921 5.3132 2.7842 P4 L0GT2 81-85 0.43631 0.17782 -1.7872 5.4763 0.3261 3.8518 P5 LDGT2 81-85 0.42908 0.2114 1.5166 4.7372 0.0105 2.9709 2,5245 P6 LDGT2 81-85 0.41687 0.23473 -3.8569 4.5338 -0.9009 2.7375 1.9498 3.2209 P7 L0GT2 81-85 0.41066 0.23366 -3.2863 4.2499 -1.1231 2.3552 1.2812 2.4743 2.1991 P8 LDGT2 81-85 0.3672 0.16688 -1.6179 3.5835 -0.8816 1.6341 -0.5481 3.1967 0.3748 5.7221 P9 LDGT2 81-85 0.3566 0.18117 -1.9127 3.5055 -0.9418 2.6495 -0.9804 1.9891 0.0338 4.0901 3.7721 P10 L0GT2 81-65 0.34151 0.16655 -1.8705 3.3075 -0.9637 3.3159 -1.7891 1.43 -0.0942 4.2235 1.3603 3.899 P11 LDGT2 81-85 0.2931 -0.04548 1.0746 1.8029 -1.2122 2.864 -0.6499 0.7309 0.2679 3.2607 -0.113 4.0698 1.9325 P12 L0GT2 81-85 0.2758 -0.01823 0.7863 1.8538 -0.984 2.6077 -0.7763 0.9299 0.773 2.1956 -0.3985 4.4575 0.873 2.4503 P13 LDGT2 81-85 0.21405 -0.01142 0.4126 1.5075 0.6583 2.6141 -0.5084 0.917 1.449 -0.2757 1.0903 2.7846 0.8509 0.03 3.3522 P14 LDGT2 81-85 0.17574 -0.01255 0.2502 1.2847 -0.0742 1.733 0.4325 1.3417 1.2187 0.2611 1.1197 1.7719 0.3758 0.7006 0.8312 4.2084 P15 LDGT2 81-85 0.15062 -0.02221 0.8141 1.325 -0.0587 1.4355 0.2426 1.2996 1.3459 0.694 0.5531 1.836 0.507 0.8254 -0.1245 2.1165 2.3277 P16 LDGT2 81-85 0.14738 -0.02755 1.0121 1.3194 -0.0133 1.4206 0.2735 1.0742 1.3679 0.895 0.1188 2.0783 0.5591 0.9525 -0.3444 1.9718 1.5317 2.7401 P17 LDGT2 81-85 0.11644 -0.04735 1.4605 1.1844 0.1561 0.8181 0.2509 0.6122 1.2664 1.4097 -0.3806 2.1461 0.6223 0.997 -0.2513 1.5706 1.3565 1.6797 2.2361 P18 LDGT2 81-85 0.10411 -0.05875 1.9421 1.2153 0.2284 0.6248 0.1763 0.9731 0.9762 1.0664 0.3199 1.6679 0.7965 0.7901 0.01 1.0203 0.9553 1.6582 1.5344 1.4185 P19 LDGT2 81-85 0.08606 -0.02861 1.9188 1.1185 0.3074 0.5037 0.6515 0.9559 0.5634 1.1722 0.4972 1.0016 0.7179 0.584 0.4013 1.317 0.5551 -0.9186 1.2833 0.9697 1.3485 IM147 Phase 2 Regression Coefficients, HC, 1981 to 1985 Model Year LDGT2's Repression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 81-85 0.53604 -0.06735 4.7635 B12 LDGT2 81-85 0.50867 -0.03629 2.9814 3.9143 813 LDGT2 81-85 0.37516 -0.03573 2.2833 -1.1494 5.7782 B14 LDGT2 81-85 0.31715 -0.05881 1.5613 0.0866 1.7691 6.5887 B15 LDGT2 81-85 0.29115 -0.05873 1.6798 0.2156 0.484 3.6579 3.2814 B16 LDGT2 81-85 0.28504 0.05551 1.6963 0.4373 0.1942 3.4658 1.7593 4.9957 B17 LDGT2 81-85 0.19206 -0.07062 1.2418 0.7375 0.4405 2.7184 1.6887 1.1505 4.1633 B18 LDGT2 81-65 0.17741 -0.07525 1.3222 0.5566 0.4852 2.1651 0.9953 1.6325 3.2854 1.6783 B19 LDGT2 81-65 0.15121 0.03936 1.1587 0.4814 1.1411 2.1958 0.3955 -1.7449 2.7898 1.0076 1.9981 ------- 0 1 Os 1 M147 Composite Repression Coefficients, HC, 1986 to 1987 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 0 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT2 86-67 0.76767 0.79061 14.62 P2 LDGT2 86-67 0.53753 0.28237 1.2627 6.2363 . !l P3 LDGT2 86-67 0.36485 0.18393 -0.7945 3.7354 5.7608 P4 LDGT2 66-67 0.34875 0.16586 -2.6149 3.7629 2.8334 4.1762 P5 LDGT2 86-67 0.34617 0.17953 -3.6873 3.212 2.761 3.4285 2.4018 P6 L0GT2 86-67 0.29325 0.2Q542 -3.5626 3.7025 1.7413 0.4872 1.0679 4.3236 . " P7 LDGT2 86-87 0.2934 0.19684 -3.5264 3.6876 1.7296 0.6143 0.2317 4.2624 0.8683 pa LDGT2 86-87 0.26213 0.17396 -1.8292 3.3926 1.1562 0.6702 0.0208 3.449 0.3557 2.6082 P9 L0GT2 86-67 0.27163 0.1983 -1.1117 3.3764 1.0296 0.699 0.2722 2.4556 0.6353 0.6 2.1369 P10 LDGT2 66-67 0.26655 0.20643 0.6291 3.3422 0.5757 0.6074 0.6213 2.1855 0.9202 0.0129 1.3566 1.7368 P11 LDGT2 86-87 0.23982 0.0819 0.4024 2.3049 0.6701 0.0523 1.2179 2.4136 0.3361 0.0779 0.9104 1.6901 1.4014 P12 LDGT2 86-87 0.21585 0.04628 0.8765 1.6118 0.7539 0.3407 0.9193 1.8565 0.6585 0.4701 0.4832 1.6797 1.0229 2.1489 P13 LDGT2 86-67 0.17953 0.04669 1.6471 0.9949 0.665 1.0029 0.4324 1.1786 0.7455 0.9956 -0.3607 1.9373 1.0646 0.024 2.6969 P14 LDGT2 66-67 0.16923 0.01715 2.1778 0.6591 1.0222 0.7528 0.0545 1.7038 0.8256 0.5799 0.1446 1.6426 1.1863 0.1661 1.154 2.4636 P15 LDGT2 86-67 0.16212 0.02841 2.045 0.4649 0.7435 .1.0759 0.3389 1.3816 1.0395 0.9294 0.0477 1.5028 1.0456 0.1781 0.9342 1.5234 1.2274 P16 LDGT2 86-67 0.15721 0.03244 2.1199 0.4169 0.672 0.9335 0.3563 1.582 1.3979 0.0804 0.2926 1.6162 1.1907 0.4448 0.6092 0.8887 0.3523 3.6537 P17 LDGT2 86-87 0.08118 0.01059 0.6472 0.7516 0.3931 0.9394 1.5168 0.882 0.3132 0.5162 0.8422 0.3101 0.8207 0.6645 0.7572 1.2216 0.5723 1.2293 3.38 P10 LDGT2 86-67 0.06876 -0.02245 0.747 0.8866 0.3561 0.7928 1.5497 0.9638 -0.0777 0.7405 0.6591 0.4782 0.7233 0.7721 0.8208 0.8129 0.6412 0.1007 2.9199 1.2143 P19 LDGT2 86-87 0.05744 -0.02534 0.8277 0.9324 0.4772 0.6834 1.4638 0.7072 0.6021 0.5295 0.8642 0.1132 0.7642 0.7717 0.5765 0.3706 0.4851 0.0561 2.1533 0.9384 1.4322 IM147 Phase 2 Regression Coefficients, HC, 1986 to 1987 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 "i C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 86-67 0.42555 0.04924 4.1247 B12 LOGT2 86-87 0.34262 0.01535 2.3444 4.5554 B13 LDGT2 86-87 0.26401 0.06727 1.8518 0.0098 4.3001 B14 LDGT2 86-67 0.26019 0.04305 1.8614 0.0618 3.3888 1.611 B15 LDGT2 86-67 0.24666 0.06334 1.6413 0.0316 2.594 0.5597 1.9278 816 LDGT2 86-67 0.24422 0.06216 1.731 0.2275 2.4543 ' -0.1869 1.1753 3.2006 B17 LDGT2 86-87 0.11773 0.01838 1.1424 0.827 1.1111 1.9999 0.6793 1.223 4.3846 B18 LDGT2 86-67 0.10087 -0.03741 1.0229 1.0633 1.2587 1.1759 0.6269 0.1785 3.8251 1.6179 B19 LDGT2 86-87 0.09638 0.03963 1.0707 1.0296 1.0107 1.2016 0.7089 -0.201 3.2092 1.6709 0.8058 ------- o ij I M147 Composite Repression Coefficients, HC, 1988 to 1995 Model Year LDGT2's Reqression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 Cfi C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT2 88-95 0.48034 0.5235 11.3071 P2 LDGT2 88-95 0.33713 0.21906 -2.96 4.9422 P3 LDGT2 88-95 0.29651 0.19834 -1.7489 2.8353 4.9489 P4 L0GT2 88-95 0.28747 0.18089 -1.9477 2.7951 2.9733 3.2647 P5 LDGT2 88-95 0.27016 0.17705 2.2098 1.8312 2.2325 1.406 5.8661 P6 LDGT2 88-95 0.25421 0.19481 -2.5612 2.2441 0.5267 0.4177 3.0039 5.8248 P7 LDGT2 88-95 0.24477 0.17308 -2.1598 1.867 0.6968 0.5785 0.8635 4.352 4.2662 P8 LDGT2 88-95 0.22604 0.14292 -1.1587 1.5845 0.5456 0.1762 0.5776 1.9469 3.7472 4.4401 P9 LDGT2 88-95 0.22226 0.14551 -1.2733 1.703 0.284 0.1479 0.9179 0.6334 3.6501 3.448 2.3522 P10 LDGT2 88-95 0.21442 0.13617 -0.7091 1.5866 0.3214 0.3077 0.4365 0.1063 3.9046 2.8806 -0.9464 5.4516 P11 LDGT2 88-95 0.1654 . 0.04125 0.6538 0.773 0.5132 0.6898 0.6807 0.4255 2.0621 1.1813 0.4048 2.5588 2.1807 P12 LDGT2 88-95 0.14608 0.03783 1.0502 0.5966 0.307 0.8399 0.3903 0.6138 1.7604 1.3774 -0.14 2.8081 1.4104 2.3284 P13 L0GT2 88-95 0.08638 0.03352 1.3017 0.5004 0.8055 0.8896 0.2765 1.4656 0.5695 0.5182 0.3106 1.2433 0.9247 -0.2803 4.1519 P14 LDGT2 88-95 0.07556 0.02491 1.2979 0.524 0.7653 0.8982 0.3475 1.5181 0.5104 0.7567 0.2981 1.4094 0.6811 0.0851 2.7444 1.9985 P15 LDGT2 88-95 0.06718 0.02356 1.1938 0.5183 0.8188 0.8118 0.379 1.4304 0.7007 0.8512 0.4013 1.2797 0.6663 0.2996 1.6768 0.8504 1.7815 P16 LDGT2 88-95 0.06237 0.02258 1.1224 0.5458 0.8938 0.7295 0.4333 1.4798 0.6019 0.7965 0.4454 1.2725 0.6721 0.4389 1.2102 0.7846 1.1152 3.1105 P17 LDGT2 88-95 0.04421 0.01056 0.9076 0.7137 0.6003 0.6133 0.8089 0.9512 0.7459 0.7414 0.5887 0.9827 0.6061 0.6414 0.972 0.8458 1.1687 1.7368 1.9774 P18 LDGT2 88-95 0.03718 0.00462 0.9099 0.6818 0.6872 0.6822 0.7561 0.9525 0.6907 0.7852 0.6052 1.0177 0.6169 0.7309 0.6146 0.7739 1.0312 0.9695 1.4043 1.372 P19 LDGT2 88-95 0.03217 0.00297 0.8825 0.7085 0.7059 0.6951 0.7723 0.8271 0.7424 0.7667 0.703 0.8462 0.6346 0.6886 0.6361 0.6291 0.7448 0.7476 1.2353 1.0716 0.9076 IM147 Phase 2 Reqression Coefficients, HC, 1988 to 1995 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 88-95 0.23751 0.05786 3.8759 B12 LDGT2 86-95 0.2067 .0.04911 2.4274 3.3621 B13 LOGT2 88-95 0.1199 0.03897 1.2412 -0.4078 5.5892 B14 LDGT2 88-95 0.10757 0.02785 1.0283 0.0365 3.9664 2.4645 B15 LDGT2 88-95 0.09668 0.02657 1.0392 0.3282 2.5887 0.8805 2.4135 B16 LDGT2 88-95 0.09084 0.02471 1.0542 0.5363 1.9743 0.7835 1.5325 4.1018 B17 LDGT2 88-95 0.06106 0.01248 0.8883 0.8408 1.4659 1.0499 1.5851 2.3095 2.7427 B16 LDGT2 88-95 0.05212 0.00425 0.9142 0.9654 1.0256 0.9252 1.4081 1.2094 2.0957 1.7973 B19 LDGT2 88-95 0.0449 0.00374 0.9299 0.9243 1.0294 0.748 0.9948 0.9323 1.7813 1.4026 1.2714 ------- o t oo IM147 Composite Regression Coefficients, HC, 1996 and Newer Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 015 C16 C17 C18 C19 P1 LDGT2 96+ 0.07243 0.06002 10.2 P2 LDGT2 96+ 0.04564 0.0237 -7.2812 4.6718 P3 LDGT2 96+ 0.04313 0.03223 -2.4032 1.3094 6.1404 P4 LDGT2 96+ 0.03729 0.03292 -1.4437 2.1679 -2.2327 9.7739 P5 LDGT2 96+ 0.02867 0.03311 -7.1815 1.5592 3.8084 -6.997 8.5612 P6 LOGT2 96+ 0.02468 0.03246 -5.2125 2.8958 -0.7992 -3.162 2.3757 7.8985 P7 LDGT2 96+ 0.02471 0.03212 -5.6903 2.6269 -0.3486 -3.879 1.8571 6.769 2.1598 P8 LDGT2 96+ 0.02458 0.03221 -5.7203 2.2642 1.1436 -7.0745 3.5654 5.3631 0.5718 2.5776 P9 LDGT2 96+ 0.02153 0.02755 -5.6174 3.2239 -0.1186 -6.6505 2.7679 1.8663 -1.0747 0.5327 8.5733 P10 LDGT2 96+ 0.02175 0.0276 -5.5557 3.1793 -0.113 6.663 2.9877 1.5294 -1.4181 0.9223 7.6307 0.9647 P11 LDGT2 96+ 0.01561 0.01512 -0.1593 1.3225 0.8979 -3.5644 1.4176 3.2321 -2.661 2.4201 3.7658 1.2056 1.069 * P12 LDGT2 96+ 0.01535 0.01155 0.1184 1.3108 0.2317 1.9686 1.0103 3.2791 -1.9369 2.0526 2.5266 1.346 0.9439 0.9442 P13 LDGT2 96+ 0.0116 0.00591 0.5316 0.6014 1.3453 -1.5835 1.0716 4.5702 -0.6276 1.6068 1.2052 -0.739 0.9742 0.8137 3.1462 P14 LDGT2 96+ 0.00894 0.00333 0.9294 0.6687 0.8295 0.2036 0.5645 4.0094 0.2628 1.1846 -0.1226 0.2024 0.7931 0.6864 2.3879 2.3089 P15 LDGT2 96+ 0.00519 0.00298 0.5728 0.3247 2.2863 -2.5268 2.2831 1.4541 -0.658 2.6037 -0.4889 1.5224 0.8251 -0.0309 1.4354 0.4248 1.7293 P16 LDGT2 96+ 0.00341 0.00091 0.6938 0.6433 0.9494 -0.2435 1.2108 1.9848 0.1134 1.3832 0.4148 0.7129 0.7604 0.2037 1.1446 0.8148 0.7145 4.0429 P17 LDGT2 96+ 0.003 0.0006 0.5483 0.8035 0.6446 0.0208 1.0382 2.0883 0.1055 1.0991 0.7076 0.6796 0.7787 0.2756 1.0048 0.7528 0.7994 3.2034 1.1992 P18 LDGT2 96+ 0.00229 0.00034 0.8423 0.7701 0.5405 0.6178 0.7621 1.54 0.2461 0.9463 0.785 0.7135 0.7393 0.5553 0.7347 0.5376 0.9234 2.5586 0.3526 1.0986 P19 LDGT2 96+ 0.00148 -0.00015 0.5489 0.7737 0.6253 0.5369 0.7319 1.0865 0.7396 0.777 0.8586 0.6472 0.714 0.5043 0.8109 0.8951 0.552 1.6026 0.688 0.6722 0.8733 IM147 Phase 2 Regression Coefficients, HC, 1996 and Newer Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 96+ 0.02667 0.01375 1.9495 B12 LDGT2 96+ 0.02266 0.00456 1.3011 3.138 B13 LDGT2 96+ 0.01837 0.00086 1.3623 0.2518 4.0937 B14 LDGT2 96+ 0.015 -0.00298 1.1501 0.0589 2.9165 3.3772 B15 LDGT2 96+ 0.01018 -0.00336 1.1443 0.9415 1.7337 0.5943 2.5841 B16 LDGT2 96+ 0.00753 0.00463 1.0828 1.0417 1.2129 0.903 1.0508 6.8124 B17 LDGT2 96+ 0.00666 -0.00476 1.1331 1.0977 1.0456 0.7683 1.3074 4.7654 2.3941 B18 LDGT2 96+ 0.00371 0.00109 0.9732 1.1893 0.7137 - 0.4116 1.4972 2.8064 0.0005 2.4412 B19 LDGT2 96+ 0.00243 -0.00121 0.9785 0.9384 0.9374 1.0534 0.8374 1.4572 0.714 1.4339 1.3705 ------- M147 Composite Regression Coefficients, HC, 1981 to 1982 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 011 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGV 81-82 0.49218 0.57705 15.1354 P2 LDGV 81-82 0.33532 0.27587 2.7958 4.3107 P3 LDGV 81-82 0.28385 0.2387 3.237 1.8573 5.3824 P4 LDGV 81-82 0.26651 0.22287 1.6681 2.0142 3.3342 3.0522 P5 LDGV 81-82 0.25039 0.21722 -0.3542 1.7343 3.0584 -0.4158 5.9309 P6 LDGV 81-82 0.23492 0.22612 -1.1998 1.7931 2.0663 -0.0624 2.4987 4.8232 P7 LDGV 81-82 0.22127 0.21033 -1.4452 1.5589 1.4953 0.1839 1.2039 3.8754 3.5496 PS LDGV 81-82 0.21552 0.19445 0.9213 1.4891 1.3035 -0.0672 1.0624 2.967 2.9559 2.5679 P9 LDGV 81-82 0.2067 0.18041 -1.0513 1.5796 1.3012 -0.0519 1.564 -0.0917 1.2979 2.0321 6.4454 PtO LDGV 61-82 0.18913 0.16356 -1.3284 1.5978 1.0833 0.3527 0.1824 0.3708 1.1991 2.1788 0.884 7.9922 P11 LDGV 61-82 0.15452 0.0563 0.0087 0.5393 1.7119 0.1012 0.6423 0.9748 1.0961 0.177 1.9283 6.7153 1.4557 P12 LDGV 81-82 0.13848 0.03382 0.3059 0.4616 1.2496 0.2523 0.8742 0.2381 0.6512 1.0074 1.8257 7.1 0.6496 1.7116 P13 LDGV 81-82 0.11063 0.03458 1.1018 0.2871 1.8035 0.9765 -0.5314 0.4907 1.1534 0.6704 1.3477 4.1068 0.6637 0.0626 3.0307 P14 LDGV 81-82 0.10379 0.02583 0.937 0.3096 1.715 1.0249 -0.5161 0.4784 1.2933 0.6542 1.7892 3.4911 0.4746 0.3833 2.0844 1.8075 P15 LDGV 81-82 0.09549 0.02733 1.3271 0.2571 1.7659 0.7304 0.2147 0.1823 1.1956 0.9924 1.0719 2.8131 0.4955 0.6194 1.2124 0.4439 2.0986 P16 LDGV 61-82 0.07932 0.0312 1.0118 0.4766 1.3033 0.8794 0.0472 0.402 1.6192 1.449 -0.2368 2.1792 0.4806 0.6063 1.0279 0.7352 0.5135 5.3144 P17 LDGV 81-82 0.05521 0.00814 1.2935 0.6849 0.6167 0.5737 0.616 0.8743 0.8448 1.724 0.1427 0.6903 0.6243 0.6678 0.6972 1.4033 0.6874 2.3373 2.2858 P18 LDGV 81-82 0.04359 0.00235 1.4765 0.7063 0.5683 0.9993 -0.1434 1.1276 0.8748 1.3533 -0.3586 1.3373 0.6347 0.6998 0.796 0.716 0.6921 1.817 1.4148 1.4078 P19 LDGV 81-82 0.03591 0.00148 1.2991 0.7134 0.6447 0.8407 0.2792 0.9872 0.8646 0.9518 0.1384 1.1208 0.6929 0.6275 0.8021 0.5739 0.7875 0.2237 1.2967 0.9834 1.1392 IM14 7 Phase 2 Regression Coefficients, HC, 1981 to 1982 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 . C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 81-82 0.32813 0.13912 3.2894 B12 LDGV 81-82 0.29163 0.08753 1.8437 3.4587 B13 LDGV 81-82 0.18796 0.06684 0.9136 -0.314 6.0599 B14 LDGV 81-82 0.1799 0.05326 0.6172 0.1991 4.6274 2.6027 B15 LDGV 81-82 0.15358 0.05277 0.6903 0.7618 2.1672 -0.1274 4.2479 B16 LDGV 81-82 0.12174 0.04997 0.8196 0.7862 1.6353 0.667 1.0053 8.5142 B17 LDGV 81-82 0.08162 0.00913 1.0095 0.7175 1.2863 1.8724 0.9087 3.3981 3.148 B16 LDGV 81-82 0.06421 0.00073 0.9515 0.8623 1.1518 0.998 0.9777 2.5859 1.9343 1.9181 B19 LDGV 81-82 0.05142 0.00017 0.9854 0.8032 1.1865 0.7374 1.1 -0.3037 1.7543 1.3113 1.6987 ------- 0 1 M147 Composite Repression Coefficients, HC, 1983 to 1985 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGV 83-65 0.40712 0.48122 11.2061 P2 LDGV 83-85 0.28564 0.23584 0.3793 4.0312 P3 LDGV 83-65 0.2251 0.18194 0.1504 1.6625 5.6839 P4 LDGV 63-85 0.20949 0.1636 -0.6566 1.89 3.4063 3.0941 P5 LDGV 83-85 0.19447 0.16094 -1.0863 1.2885 2.6869 2.1291 3.7766 P6 LDGV 83-65 0.19153 0.16531 -1.4624 1.3044 2.1907 1.8106 3.1299 2.0652 P7 LDGV 83-85 0.16355 0.1542 -1.1191 1.2382 2.1956 1.3759 2.2879 1.3053 2.1756 P8 LDGV 83-65 0.1742 0.14505 -1.1029 1.1144 1.7862 0.8889 2.4267 1.2368 1.6895 1.9949 P9 LDGV 83-85 0.16499 0.14615 -0.7229 1.073 1.8176 0.7769 1.4795 0.6194 1.4189 0.9084 4.175 P10 LDGV 83-65 0.16024 0.14347 -0.7243 1.1373 1.5682 0.8427 0.8657 0.8685 1.1993 0.834 2.645 2.9155 P11 LDGV 83-85 0.12238 0.06297 0.38 0.2293 1.7054 0.1936 1.2879 1.6893 0.5856 0.5369 2.6104 1.7875 1.5441 P12 LDGV 83-85 0.10785 0.05022 0.4878 0.4094 0.8846 0.5066 1.0774 1.8031 0.7369 0.5491 1.8603 2.2819 1.0196 1.6327 P13 LDGV 83-65 0.07961 0.0331 0.7767 0.4332 0.8179 0.9806 1.0015 1.2108 0.561 0.4915 1.5593 1.6272 0.8246 0.3997 2.5228 P14 LDGV 83-85 0.068 0.02445 0.861 0.3782 0.9405 0.9692 0.9471 1.0556 0.6664 0.4702 1.5692 1.559 0.6216 0.5139 1.483 2.5758 P15 LDGV 83-85 0.05617 0.01633 0.6641 0.3847 0.9494 0.9339 0.8878 1.1718 0.6879 0.3691 1.2553 1.3016 0.7499 0.5579 0.6795 1.1332 2.2871 P16 LDGV 83-85 0.05545 0.01675 0.7233 0.393 0.9564 0.9028 0.9463 1.1024 0.7194 0.3887 1.2701 1.1977 0.7436 0.5865 0.627 1.1212 1.6334 1.417 P17 LDGV 63-85 0.03679 0.00841 0.893 0.5126 0.7936 0.6136 1.0495 0.8218 0.8047 0.6079 0.984 0.9734 0.7209 0.7276 0.5005 1.2506 1.4269 0.4297 2.417 P18 LDGV 83-65 0.02939 -0.0015 0.9063 0.6276 0.7336 0.799 0.8401 0.7909 0.6547 0.6256 1.0592 0.5782 0.7013 0.749 0.5944 0.8615 1.0441 0.6308 1.7952 1.3705 P19 LDGV 83-65 0.0284 -0.00036 0.6248 0.6354 0.7403 0.7849 0.8685 0.8016 0.6772 0.6005 1.0662 0.4753 0.7239 0.7376 0.6183 0.7249 0.9182 0.3388 1.7224 1.1415 0.5047 \MU 7 Phase 2 Regression Coefficients, HC, 1983 to 1985 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C.1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 83-65 0.2057 0.10209 2.648 B12 LDGV 83-85 0.17708 0.0755 1.6316 2.7713 Q13 LDGV 63-65 0.12398 0.04472 1.2782 0.5298 4.0757 B14 LDGV 83-65 0.10965 0.03254 0.9666 0.713 2.5673 3.6248 B15 LDGV 83-85 0.08609 0.01872 1.0596 0.7937 1.0694 1.2338 3.6814 916 LDGV 83-85 0.08613 0.02003 1.0528 0.6555 0.9553 1.2083 2.7852 2.7867 B17 LDGV 83-65 0.05674 0.00853 0.9868 0.9946 0.7356 1.5594 2.1184 0.6029 3.6309 B18 LDGV 83-85 0.04106 -0.0038 0.9461 1.0589 0.7763 1.1008 1.4725 0.8392 2.602 1.9484 B19 LDGV 6345 0.03964 -0.00206 0.9706 1.0465 0.8189 0.9395 1.277 0.41 2.4945 1.6256 0.6907 ------- o M147 Composite Repression Coefficients, HC, 1986 to 1989 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 . C14 C15 C16 C17 C18 C19 P1 LDGV S6-89 0.31427 0.27681 15.5205 P2 LDGV 66-89 0.23056 0.12883 1.6181 4.2149 P3 LDGV 86-69 0.19626 0.11002 2.0358 2.3971 3.9903 P4 LDGV 86-69 0.18598 0.09989 0.6425 2.546 1.7035 3.3548 P5 LDGV 86-69 0.17584 0.09767 0.5819 1.852 1.5615 1.555 3.8429 P6 LDGV 86-89 0.16772 0.10204 -0.8376 1.9082 1.0801 1.0965 1.9459 3.8412 P7 LDGV 86-69 0.16444 0.09594 -0.6742 1.6725 0.9995 1.3082 0.7861 3.2631 2.3581 P8 LDGV 86-89 0.15806 0.08909 0.1422 1.5556 0.6844 0.9331 1.0208 2.2001 1.5379 2.5805 P9 LDGV 86-69 0.1519 0.08696 0.1948 1.5757 0.7441 1.0414 0.6635 1.0744 1.3289 1.468 3.5915 P10 LDGV 86-69 0.14467 0.0797 -0.4894 1.5393 0.9657 0.7863 0.5873 1.2938 1.3345 1.1773 0.2863 4.8853 P11 LDGV 86-69 0.11095 0.02972 0.2469 0.721 1.1728 0.8987 0.4133 1.9423 0.6973 0.4156 0.9936 2.4194 1.6263 P12 LDGV 86-89 0.09318 0.02564 0.8246 0.6275 0.7263 0.7966 0.542 1.6116 0.6663 0.672 0.4503 2.6883 1.1012 1.7602 P13 LDGV 66-69 0.06902 0.02184 1.3196 0.4061 0.8931 0.8959 0.5829 1.0567 0.7867 0.7859 0.5864 1.9088 0.9881 0.291 2.4099 P14 LDGV 66-89 0.06241 0.01809 1.4552 0.31 1.0344 0.9038 0.6073 1.022 0.7521 0.848 0.6425 1.6254 0.813 0.5647 1.1453 2.4731 P15 LDGV 66-89 0.05136 0.01585 1.2694 0.4095 0.9607 0.9108 0.6871 0.8398 0.7304 0.8585 0.5526 1.7815 0.6924 0.6284 0.7316 0.9325 2.1051 P16 LDGV 66-69 0.04861 0.01432 1.1846 0.4732 0.8917 0.9259 0.7055 0.8742 0.7619 0.7564 0.5578 1.5489 0.7021 0.6507 0.7862 0.9721 1.2367 2.5815 P17 LDGV 66-89 0.03332 0.00699 1.0682 0.673 0.7585 0.7037 0.8742 0.6284 0.7246 0.6208 0.7319 1.1727 0.7152 0.6517 0.7253 1.0692 1.1351 1.0816 2.2316 P18 LDGV 86-69 0.02685 0.0032 1.1443 0.6139 0.8508 0.7766 0.7017 0.7447 0.6277 0.623 0.6745 1.0469 0.7238 0.6588 0.7335 0.8575 0.8976 0.668 1.5797 1.3415 P19 LDGV 86-69 0.02176 0.00312 1.0887 0.6218 0.766 0.6703 0.7888 0.6802 0.7991 0.5738 0.7502 0.9406 0.7447 0.674 0.6947 0.6605 0.8184 0.1544 1.3072 0.9142 0.9739 IM14 ^7 Phase 2 Regression Coefficients, HC, 1986 to 1989 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 811 LDGV 86-89 0.17572 0.05913 2.9828 B12 LDGV 86-89 0.14079 0.04187 1.8474 2.837 B13 LDGV 86-89 0.10263 0.03308 1.4761 0.4901 3.5664 B14 LDGV 66-89 0.09666 0.02888 1.2557 0.6541 2.0611 2.9206 B15 LDGV 86-89 0.07943 0.02543 1.0638 0.9286 1.331 0.6468 3.2335 B16 LDGV 86-89 0.07243 0.02198 1.0486 0.9587 1.3534 0.8765 1.5227 4.8276 B17 LDGV 86-89 0.04689 0.01029 1.0457 0.8812 1.097 1.3615 1.4856 1.7603 3.2281 B18 LDGV 86-89 0.03789 0.00418 0.9974 0.9025 1.0724 1.0277 1.1846 1.1391 2.3711 1.6028 B19 LDGV 86-89 0.03036 0.00369 1.0225 0.6952 0.9808 0.8231 1.1 0.3076 1.8483 1.2733 1.3385 ------- 0 1 tť K) M147 Composite Repression Coefficients, HC, 1990 to 1995 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 05 . C6 C7 C6 C9 C10 C11 C12 C13 . C14 C15 C16 C17 C18 C19 P1 LDGV 90-95 0.2971 0.16705 15.7948 . if 92 LOGV 90-95 0.19863 0.05505 -1.9351 4.6837 P3 LDGV 90-95 0.16628 0.03974 -1.5889 2.5546 5.6414 P4 LDGV 90-95 0.1584 0.03534 -1.9531 2.652 3.0556 3.5632 P5 LDGV 90-95 0.14861 0.0377 -1.7859 2.1399 2.3054 1.5953 3.8796 P6 LDGV 90-95 0.1428 0.04157 -2.1978 2.3528 1.3397 1.1611 1.6243 4.1795 . i P7 LDGV 90-95 0.13017 0:03906 -1.447 1.8386 1.3743 1.2125 -0.2829 2.1347 4.9626 L P8 LDGV 90-95 0.12213 0.03722 -0.5405 1.6099 1.2031 0.3856 -0.1326 1.4892 3.3678 3.4765 P9 LOGV 90-95 0.11733 0.03754 -0.7341 1.7237 1.0051 0.4924 -0.9573 0.6324 3.2462 2.3004 3.9191 P10 LDGV 90-95 0.11526 0.0368 -0.6148 1.6847 1.071 0.425 1.176 0.676 2.845 2.502 1.4335 2.9572 P11 LDGV 90-95 0.08741 0.00996 0.6446 0.8311 1.0111 0.5067 -0.4113 1.2288 2.2201 1.2696 1.6991 1.3219 1.5051 P12 LDGV 90-95 0.07472 0.0105 0.6979 0.7429 0.706 0.5223 -0.0564 1.3101 1.9432 1.288 1.2908 1.3897 0.6836 1.8454 P13 LDGV 90-95 0.05916 0.0139 1.3045 0.4375 0.9242 0.6499 0.0689 1.2742 1.493 1.04 1.3582 0.4913 0.8544 0.3112 2.5569 P14 LDGV 90-95 0.05362 0.01549 1.386 0.3479 1.0903 0.5466 0.252 1.3522 1.424 0.6349 1.5658 0.679 0.6813 0.3943 1.6017 2.0376 P15 LDGV 90-95 0.04337 0.01025 1.2116 0.4485 0.9328 0.7036 0.5038 1.3664 1.1003 0.7903 1.3103 0.2473 0.6803 0.7135 0.7193 0.5976 2.3136 P16 LDGV 90-95 0.04152 0.00959 1.1338 0.4558 0.9892 0.8217 0.5423 1.2542 0.9416 0.7779 1.3289 0.1879 0.6808 0.7232 0.692 0.6981 1.451 2.8503 P17 LDGV 90-95 0.0313 0.00573 0.9123 0.6301 0.5621 0.8487 0.6701 1.1772 0.8709 0.7111 1.0362 0.425 0.686 0.7574 0.6663 0.8447 1.2224 1.3549 2.4338 P18 LDGV 90-95 0.02304 0.004 1.0116 0.6089 0.6873 0.6706 0.6772 1.1438 0.7472 0.6904 0.8929 0.5431 0.6656 0.7093 0.6728 0.682 1.042 0.6819 1.3735 1.591 P19 LDGV 90-95 0.01929 0.00299 0.874 0.6682 0.6242 0.6993 0.6995 1.2226 0.7716 0.6058 0.6359 0.606 0.6915 0.724 0.6677 0.6623 0.8334 0.1215 1.1562 1.238 0.9297 IM1A 7 Phase 2 Repression Coefficients, HC, 1990 to 1995 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 90-95 0.13911 0.0203 2.9778 B12 LDGV 90-95 0.11712 0.01783 1.7491 2.9654 B13 LDGV 90-95 0.08624 0.01945 1.2881 0.3333 3.9332 B14 LDGV 90-95 0.08064 0.02126 1.0669 0.4262 2.8321 2.3928 B15 LDGV 90-95 0.06433 0.0142 1.0083 0.9588 1.2508 0.4227 3.4379 B16 LDGV 90-95 0.06126 0.01375 0.9905 0.9989 1.1536 0.6257 2.0703 4.3157 B17 LDGV 90-95 0.04442 0.00749 0.9692 0.997 1.0495 1.0206 1.8198 2.0719 3.5584 B18 LOGV 90-95 0.03245 0.00457 0.9306 0.9602 0.9645 0.8007 1.3774 1.0699 1.9911 2.2266 B19 LDGV 90-95 0.02708 0.00366 0.9309 0.9723 0.9484 0.8169 1.0622 0.3307 1.6177 1.7198 1.2796 ------- IM147 Composite Regression Coefficients, HC. 1996 and Newer Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGV 96+ 0.10038 0.04026 29.2462 P2 LDGV 96+ 0.07371 0.01788 5.9162 4.0658 P3 LDGV 96+ 0.06844 0.01874 2.9638 1.9564 5.5186 P4 LDGV 96+ 0.06132 0.01973 2.7332 1.6433 3.0543 4.4161 P5 LDGV 96+ 0.06046 0.02162 2.8327 1.3424 2.314 3.7839 1.694 P6 LDGV 96+ 0.05435 0.02387 1.2378 1.2522 0.457 3.4931 -1.2143 7.4758 P7 LDGV 96+ 0.05447 0.0241 1.0091 1.2788 0.4429 3.4792 -1.2783 7.8471 0.2285 P8 LDGV 96+ 0.05359 0.02337 0.9941 0.7729 0.7794 3.2658 -2.1201 7.6602 -0.5933 3.1409 P9 LDGV 96+ 0.05279 0.02482 2.6229 0.6273 0.6458 3.3195 -1.1123 4.6464 -0.1966 0.788 4.1995 P10 LDGV 96+ 0.05143 0.02476 1.0359 0.4556 1.1796 3.0156 -1.3707 3.8218 0.108 0.823 0.5809 6.8862 P11 LDGV 96+ 0.03867 0.00936 5.4115 -0.1934 1.9861 0.9553 -0.8572 0.8886 0.8704 0.1062 0.029 6.6498 1.6503 P12 LDGV 96+ 0.02114 0.009 3.1864 0.1934 1.3914 1.1882 0.032 -0.1178 0.7255 1.0187 0.0701 4.5106 0.6495 2.1154 P13 LDGV 96+ 0.01901 0.00675 3.9129 0.3208 1.0647 0.8952 0.4296 0.6086 0.9736 0.1471 -0.7025 3.5491 0.6163 1.6571 1.3153 P14 LDGV 96+ 0.01746 0.00454 4.3463 0.3067 1.0782 1.0439 0.3355 1.3018 1.0886 -0.2129 -0.7694 2.9793 0.5207 1.3494 1.2261 1.5726 P15 LDGV 96+ 0.01604 0.00595 4.6859 0.2669 0.8615 0.9562 0.5265 1.8859 0.9556 -0.0501 -1.2322 2.431 0.5884 1.2496 1.3707 -0.6015 1.5225 P16 LDGV 96+ 0.01579 0.00648 4.6016 0.2309 0.8473 0.9826 0.3155 2.3189 0.8167 0.4521 -1.291 2.0816 0.5345 1.4982 1.0515 -0.1967 2.1338 -2.9193 P17 LDGV 96+ 0.01236 0.00351 2.9817 0.5075 0.6011 1.1216 0.3319 1.6187 0.6518 0.5096 0.0325 1.425 0.5667 1.2205 0.6224 1.0067 1.5045 -2.3679 2.3031 P18 LDGV 96+ 0.00892 0.00122 2.9957 0.5568 0.7792 0.788 0.2152 1.8089 0.6335 0.4301 0.2697 0.9495 0.5389 1.1609 0.7993 0.5202 1.4602 -1.9165 1.033 1.7225 P19 LDGV 96+ 0.00643 0.00092 1.469 0.8076 0.4864 0.7865 0.6192 1.1938 0.534 0.2013 0.8084 1.3759 0.6416 0.824 0.4897 0.7037 0.7724 0.0797 0.3611 1.5812 0.8653 I M147 Phase 2 Regression Coefficients, HC, 1996 and Newer Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 96+ 0.06231 0.01367 2.7246 B12 LDGV 96+ 0.03471 0.01289 1.0956 3.2189 B13 LDGV 96+ 0.02909 0.00972 0.7958 2.4462 2.013 B14 LDGV 96+ 0.02639 0.00805 0.7353 2.2615 1,8431 1.2834 B15 LDGV 96+ 0.02692 0.00988 0.8024 2.2022 1.7552 -1.4225 1.8689 B16 LDGV 96+ 0.02666 0.01042 0.7459 2.4894 1.4031 -0.8725 2.5744 -3.5714 B17 LDGV 96+ 0.0192 0.00442 0.8885 1.711 0.9711 0.9804 1.8655 -2.5604 3.883 B18 LDGV 96+ 0.01449 0.00165 0.7232 1.6635 1.0171 0.5616 1.7084 -1.9289 2.0901 2.3916 B19 LDGV 96+ 0.01043 0.00074 0.92 1.1468 0.7967 0.6882 1.2025 -0.0572 0.6503 2.302 1.1734 ------- 0 1 ~* -U 1 M147 Composite Repression Coefficients, CO, 1981 to 1985 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11' C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT1 81-85 10.1754 14.5276 12.5442 P2 LDGT1 81-65 7.53286 8.22078 5.3429 3.5563 P3 LDGT1 81-85 6.63268 6.60978 2.7905 2.1525 5.3312 P4 LDGT1 81-85 6.52317 6.63083 1.5125 2.2786 3.2782 3.8618 K P5 LDGT1 81-85 6.508 6.69361 1.2349 2.1748 2.865 3.4811 1.0432 . P6 LDGT1 81-85 6.43582 6.71221 0.1813 2.3646 2.9243 1.5307 0.1722 2.7427 P7 LDGT1 81-85 6.19445 6.37324 0.5635 1.9693 2.7172 0.6361 -1.561 3.2843 2.9955 P8 LDGT1 81-85 5.94027 5.98064 1.2119 1.6393 2.7167 0.5356 -1.4442 1.3759 1.7876 3.6004 . .H P9 LDGT1 81-65 5.9353 5.97383 1.0059 1.6737 2.5314 0.6549 -1.7102 0.8508 1.7967 3.4358 1.1942 P10 LDGT1 81-85 5.85738 6.02848 1.177 1.7285 2.0533 0.7067 -2.0802 1.0202 1.2427 3.6953 -1.2444 3.1472 P11 LDGT1 81-85 4.17486 2.67355 0.0251 1.2531 0.5349 0.6814 -0.4647 2.1824 0.0334 2.27 1.2267 2.243 1.0397 P12 LDGT1 81-85 3.65635 1.97698 -0.5472 1.2512 -0.2635 2.0162 -0.5276 2.225 0.3841 1.8143 0.5359 2.6593 0.7892 1.1045 P13 LDGT1 81-85 2.30258 0.92158 0.7798 0.5868 0.3445 2.2218 0.3028 1.3536 0.523 1.1833 0.7632 1.2554 0.7979 0.6638 2.4441 P14 LDGT1 81-65 1.68789 0.51912 1.7468 0.429 0.6515 1.5816 0.7182 0.8416 0.48 1.2506 1.093 1.1172 0.6993 0.5859 1.7776 1.5319 P15 L0GT1 81-85 1.13559 0.36037 1.4986 0.481 0.5203 1.5061 0.696 0.646 0.8751 0.9462 0.7302 1.215 0.6963 0.7448 0.985 0.9478 1.4391 P16 LDGT1 81-85 1.03437 0.37532 1.4149 0.5205 0.4715 1.4258 0.7753 0.6646 0.8775 0.7646 0.5543 1.4051 0.6989 0.7734 0.9487 0.9689 0.8955 1.6125 P17 LDGT1 81-85 0.96568 0.20115 1.2423 0.5923 0.2634 1.3236 0.9825 0.7034 0.7693 0.7682 0.1979 1.3783 0.7136 0.7407 0.9773 0.9693 0.8911 1.2362 1.2653 P18 L0GT1 81-85 0.75949 0.06889 1.1529 0.5685 0.4385 1.1423 0.7942 0.9943 0.7711 0.6846 0.5815 1.031 0.7017 0.7499 0.8098 0.8097 0.647 0.846 1.2494 0.9829 P19 LDGT1 81-85 0.34973 -0.06051 1.1508 0.6559 0.4115 1.1472 0.7599 0.9619 0.799 0.6744 0.5327 0.985 0.7106 0.7294 0.7378 0.6891 0.6385 0.7063 1.0232 0.6907 0.8306 IM147 Phase 2 Regression Coefficients, CO, 1981 to 1985 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C16 C19 B11 LDGT1 81-85 9.2129 9.11082 1.5105 B12 LDGT1 81-85 5.96613 4.72873 1.1535 1.7914 B13 LDGT1 81-85 3.55873 1.89325 1.0491 0.955 3.6598 B14 LDGT1 81-85 2.92833 1.4436 0.9077 0.9194 2.7661 1.929 B15 L0GT1 81-65 1.93926 0.86159 0.9453 1.0943 1.32 1.0292 2.3486 B16 LDGT1 81-85 1.76086 0.83361 0.9517 1.1151 1.2896 1.1023 1.3595 2.6847 B17 LDGT1 81-65 1.44658 0.21699 0.9768 0.9803 1.3491 1.2314 1.27 1.5265 2.526 B18 LDGT1 81-65 1.172 0.01644 0.9407 1.0108 1.0591 1.0119 1.2102 1.0895 2.58 1.3347 B19 LOGT1 81-85 0.69787 -0.12536 0.9656 0.9797 0.9936 0.8578 0.9082 0.6798 2.1293 0.9645 1.1513 ------- o ~I 1 M147 Composite Regression Coefficients, CO, 1986 to 1989 Model Year LDGTVs Repression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGTl 86-89 7.18532 10.8086 9.9102 P2 LDGT1 86-89 5.68789 7.0459 1.8741 2.6779 P3 LDGT1 86-89 5.04319 5.5558 1.7455 1.371 5.3393 P4 LDGT1 86-69 4.88921 5.13459 1.4696 1.5629 2.9026 5.3333 P5 LDGT1 86-89 4.78915 5.08892 1.6656 1.4035 2.0371 4.1723 2.0022 P6 LDGTl 86-89 4.74332 5.11226 1.0229 1.4755 1.9866 3.8059 1.0898 2.4375 P7 LDGT1 86-89 4.65225 4.91955 1.7462 1.3554 1.7097 3.8416 -0.0053 1.9673 1.6899 P8 LDGT1 66-89 4.48557 4.61047 2.0142 1.3439 1.7523 2.7365 0.09 0.1646 0.9526 2.9249 P9 LDGT1 86-89 4.44611 4.5639 2.1382 1.316 1.6099 2.7676 -0.2399 -0.7903 0.9655 2.4766 2.079 P10 LDGT1 86-69 4.41672 4.48084 2.0022 1.3046 1.5957 2.4537 -0.4021 .-0.7631 0.7899 2.5819 0.8104 2.0187 P11 LDGT1 86-89 3.05161 2.26609 2.8544 0.9359 0.8115 1.4536 0.41 0.3277 0.4047 2.1497 -0.2782 2.0562 0.9911 P12 LDGTl 86-89 2.51613 1.51088 2.8342 0.7464 0.9245 1.9441 0.5707 -0.2141 0.4114 2.0168 0.2073 1.9253 0.7436 1.2083 P13 LDGT1 86-89 1.68355 0.8064 2.1795 0.6233 0.6001 1.3103 0.4715 0.7177 0.6871 0.7545 0.7656 1.1501 0.7503 0.8943 2.2692 P14 LDGT1 66-89 1.31741 0.68081 2.1784 Ś 0.5849 0.6912 1.3598 0.4357 0.9373 0.6956 0.8278 0.6331 1.2087 0.7057 0.7686 1.824 1.1207 P15 LDGT1 86-89 0.93911 0.47 1.8779 0.6014 0.6301 1.3592 0.4395 0.6121 0.676 0.6638 0.7005 0.9787 0.7117 0.7437 1.0888 0.8731 1.3809 P16 LDGT1 86-89 0.88497 0.41787 1.6764 0.5971 0.6595 1.3638 0.5037 0.6323 0.8426 0.6198 0.6814 1.0802 0.7147 0.7671 1.022 0.8818 0.9654 1.3933 P17 LDGT1 86-89 0.76094 0.28167 1.1884 0.633 0.6369 1.0655 0.6729 0.7423 0.7527 0.7112 0.7965 0.8042 0.7055 0.7447 0.9245 0.8662 1.048 0.6815 1.7386 P18 LDGT1 86-89 0.51513 0.11517 1.107 0.6616 0.6251 1.1599 0.7126 0.7927 0.5228 0.8379 0.7639 0.7464 0.7046 0.7204 0.7302 0.7811 1.0809 0.4756 1.39 0.869 P19 LDGT1 86-89 0.22775 0.06824 0.9804 0.6792 0.6446 1.0212 0.6645 0.7468 0.6847 0.7006 0.7495 0.7734 0.7063 0.7111 0.7207 0.7067 0.7928 0.5475 0.9726 0.7062 0.7346 IM147 Phase 2 Regression Coefficients, CO, 1986 to 1989 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 86-89 5.73419 6.44585 1.4889 B12 . LDGT1 86-89 3.83261 3.46613 1.1555 1.6694 B13 LDGTl 66-89 2.36328 1.33398 1.0419 1.1807 3.2911 B14 LDGTl 86-89 1.90917 1.13186 0.9737 1.0107 2.6834 1.4946 B15 LDGT1 86-89 1.36531 0.75413 0.976 0.9839 1.5355 1.1464 1.973 B16 LDGT1 86-89 1.28845 0.69885 0.9841 1.0114 1.4453 1.1588 1.3867 1.9667 B17 LDGT1 86-89 1.04812 0.39116 0.9599 0.9873 1.256 1.1763 1.4736 0.6337 2.6609 816 LDGT1 86-89 0.7463 0.14799 0.9544 0.9576 1.0029 1.0365 1.5243 0.6229 2.1332 1.1447 B19 LDGT1 86-89 0.33544 0.07788 0.9581 0.9522 0.9677 0.9398 1.0857 0.7154 1.4816 0.9344 1.0357 ------- 0 1 )* ON I M147 Composite Regression Coefficients o o CD CO 0 to 1995 Model Year LDGTI's . Regress on Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 Cl1i C12 C13 C14 C15 C16 C17 C18 C19 Pi LDGT1 90-95 7.96708 6.47319 8.6737 P2 LDGT1 90-95 5.93561 3.60059 0.3029 2.9243 P3 LDGT1 90-95 5.49867 2.58569 0.1967 1.6676 5.73 P4 LDGT1 90-95 5.47555 2.41839 0.0472 1.683 4.5333 2.9942 i P5 LDGT1 90-95 5.34097 2.4681 0.3008 1.3842 2.8744 1.4949 3.7142 P6 LDGT1 90-95 5.23567 2.50459 -0.4457 1.3784 2.6883 0.8341 2.2673 4.9505 P7 LDGT1 90-95 5.12065 2.31146 -0.3312 1.2837 1.873 -0.1146 0.7803 4.7671 3.2802 P8 LDGT1 90-95 4.93002 2.15394 -0.1308 1.2614 1.6587 -0.7666 0.4112 1.9844 1.3783 5.3488 P9 LDGT1 90-95 4.92436 2.15209 -0.2521 1.2934 . 1.4663 -0.6767 0.1855 1.6357 1.4433 4.9503 1.3627 P10 LDGT1 90-95 4.83877 2.07448 0.0883 1.3418 1.2572 -0.6746 -0.0356 1.4432 1.0559 4.4422 -1.7507 4.57 . .. P11 LDGT1 90-95 3.05447 0.67888 0.7448 0.9163 0.7304 -1.1973 0.6625 1.9991 1.1363 2.2194 -0.3005 2.3958 1.1499 P12 LDGT1 90-95 2.59434 0.35922 0.8716 0.834 -0.2278 -0.1746 0.685 1.603 1.4913 1.7526 -0.5786 2.4993 0.9124 1.5835 P13 LDGT1 90-95 1.69051 0.3538 0.7846 0.7978 0.2453 0.7291 0.7109 0.6836 0.9317 0.9732 0.7569 0.3609 0.8116 0.8909 2.4633 P14 LDGT1 90-95 1.25118 0.29289 0.7084 0.7156 0.4878 0.5609 0.6826 0.8354 0.9735 1.1145 0.55 0.9761 0.7373 0.7493 1.6233 1.1532 P15 LDGT1 90-95 0.88666 0.17653 0.6832 0.7443 0.6247 0.7585 0.551 0.8411 0.8994 0.6594 0.516 0.6847 0.7688 0.6805 1.0955 0.8841 1.4675 P16 LDGT1 90-95 0.83688 0.14342 0.6546 0.7325 0.6626 0.7804 0.6 0.754 0.9659 0.5296 0.6743 0.5828 0.7741 0.6855 0.9886 0.964 1.0266 1.788 P17 LDGT1 90-95 0.75939 0.03084 0.6751 0.7626 0.4984 0.7727 0.6819 0.6512 0.8746 0.5588 0.7025 0.4405 0.774 0.6594 0.9682 0.9724 1.0041 1.3254 2.054 P18 LDGT1 90-95 0.46016 0.02847 0.6665 0.7378 0.5678 0.7356 0.81 0.7878 0.6105 0.7172 0.7969 0.5663 0.7101 0.721 0.8057 0.915 0.6348 1.2367 1.3997 1.0199 P19 LDGT1 90-95 0.16891 0.00569 0.7401 0.7243 0.6471 0.7729 0.6775 0.7704 0.7643 0.6818 0.7573 0.6839 0.7157 0.7187 0.7375 0.6868 0.717 0.6511 1.1384 0.7248 0.7498 IM147 Phase 2 Regression Coefficients, CO, 1990 to 1995 Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 90-95 5.48884 2.92303 1.7403 B12 LDGT1 90-95 3.74619 0.83254 1.3383 2.3792 B13 LDGT1 90-95 2.29939 0.47273 1.1019 1.1903 3.3532 B14 LDGT1 90-95 1.7181 0.48708 1.0162 1.02 2.3871 1.5186 B15 LDGT1 90-95 1.20254 0.23609 1.0388 0.9035 1.4703 1.2205 1.9615 B16 LDGT1 90-95 1.13572 0.20105 1.0447 0.9125 1.3168 1.3287 1.3674 2.3837 B17 LDGT1 90-95 1.04495 0.01286 1.0374 0.8678 1.2457 1.3709 1.2968 1.8468 2.5009 B18 L0GT1 90-95 0.65189 0.0106 0.9551 0.9717 1.0606 1.2615 0.8351 1.7471 1.7371 1.3845 819 L0GT1 90-95 0.219 0.00146 0.968 0.9695 0.9969 0.9364 0.9681 0.9047 1.5264 0.9898 1.012 ------- IM147 Composite Regression Coefficients, CO, 1996 and Newer Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 PI LDGT1 96+ 1.82409 1.39419 5.531 P2 LDGT1 96+ 1.63165 0.95561 -5.9413 3.0869 P3 LDGT1 96+ 1.57733 0.85422 -1.5783 1.5971 3.8856 P4 LDGT1 96+ 1.57696 0.84862 -1.5432 1.758 1.4232 4.1058 P5 LDGT1 96+ 1.55655 0.77387 -1.8497 1.5252 1.4375 1.5669 3.3072 P6 LDGT1 96+ 1.48358 0.83228 -5.5406 1.6396 -0.2011 -2.3153 2.1554 10.6111 P7 LDGT1 96+ 1.47504 0.7148 -5.2472 1.8872 0.9914 -0.8568 0.4965 7.7306 2.5842 . _ P8 LDGT1 96+ 1.48311 0.71859 -5.1507 1.8553 -0.9277 -0.9266 0.671 8.0227 2.621 -0.2793 P9 LDGT1 96+ 1.49155 0.72084 -5.0954 1.8331 -1.0008 0.6946 0.7565 7.9469 2.63 0.1935 -0.3333 P10 LDGT1 96+ 1.46528 0.70223 -3.9278 1.5364 -0.9574 -0.075 1.1515 8.3687 2.439 -0.1927 -2.1271 1.4542 P11 LDGT1 96+ 0.71796 0.22242 -1.2984 1.4739 0.5359 -1.1903 0.4023 2.2155 2.7505 0.2649 0.6008 1.0433 0.8928 P12 LDGT1 96+ 0.54429 0.13226 1.524 1.022 0.7868 -1.432 0.1372 2.482 1.6269 -0.3682 2.2202 0.4979 0.8781 1.6265 P13 LDGT1 96+ 0.43766 0.08336 1.4971 0.7999 1.0816 -1.4102 0.8701 2.4778 1.5114 -0.0111 0.8915 0.5019 0.8812 0.3603 3.2131 P14 LDGT1 96+ 0.27254 0.0707 1.3293 0.9437 0.7684 -0.6826 0.5074 1.6189 1.5858 0.2076 0.9221 0.5798 0.729 0.6284 2.4591 0.7018 P15 LDGT1 96+ 0.19651 0.03007 1.3733 0.8799 0.3876 0.2025 0.6389 2.4824 0.9733 0.4929 0.3408 0.8102 0.7261 0.7297 1.4475 0.6703 1.466 P16 IDGT1 96+ 0.17615 0.02979 1.0484 . 0.6101 0.7323 -0.205 0.5366 2.4872 0.7994 0.4735 0.5616 0.7865 0.7275 0.7712 1.4017 0.6818 0.9833 2.2195 P17 LDGT1 96+ 0.13011 0.05988 1.062 0.6811 0.3816 0.7616 0.4795 2.2042 Ś 0.9584 0.5832 0.1536 0.8447 0.7233 0.7198 1.1616 0.6987 1.0372 0.2683 1.8782 P18 LDGT1 96+ 0.08591 0.02837 1.3696 0.6559 0.7295 0.2618 0.8191 1.2599 0.5828 0.7744 0.6755 0.6858 0.712 0.7269 0.7236 0.7111 0.9386 0.4881 1.7121 1.0654 P19 LOGT1 96+ 0.03308 0.00956 0.6181 0.7187 0.7481 0.81 0.6377 0.8223 0.7051 0.6935 0.8481 0.6648 0.7101 0.7232 0.6944 0.7169 0.8125 0.226 1.1367 0.7471 0.7019 IM147 Phase 2 Regression Coefficients, CO, 1996 and Newer Model Year LDGT1 's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 96+ 1.23604 0.96823 0.9526 B12 LDGT1 964 0.74363 0.24386 1.1933 2.1843 B13 LDGT1 96+ 0.60473 0.1527 1.2056 0.5562 4.3384 B14 LDGT1 96+ 0.40405 0.15643 0.9921 0.9282 3.3758 0.9673 815 LOGT1 96+ 0.30462 0.06538 0.9969 1.0801 1.9431 0.914 2.0105 B16 LOGT1 96+ 0.26097 0.03614 0.9954 1.08 1.8842 0.9365 1.1866 3.7436 B17 LDGT1 96+ 0.2048 0.0653 0.9914 1.0331 1.5352 0.9586 1.2475 1.026 2.448 B18 LDGT1 96+ 0.12661 0.03391 0.9631 0.9575 0.9469 0.9698 1.2044 0.6083 2.4102 1.4239 B19 LDGT1 96+ 0.04352 0.01126 0.9604 0.9595 0.9395 0.9708 1.1345 0.2607 1.4873 1.0202 0.9453 ------- 0 1 >* OO I M147 Composite Repression Coefficients, CO, 1981 to 1985 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 010 r V C11- 012 C13 C14 C15 C16 017 018 019 P1 LDGT2 81-85 18.0705 25.7173 7.5612 P2 LDGT2 81-85 12.3795 12.9451 3.4376 3.6374 . P3 LDGT2 81-65 11.5598 10.6148 1.0212 2.3546 5.6508 * P4 LDGT2 81-65 11.1259 9.72659 -0.5375 2.5842 2.5322 5.5716 P5 LDGT2 81-65 11.1623 9.79162 -0.516 2.5693 2.415 5.44 0.2344 P6 LDGT2 81-85 10.5815 10.5149 -5.1729 2.5377 2.4642 3.5854 -2.439 9.5373 i! P7 LDGT2 81-85 10.5066 9.63447 5.1298 2.3487 2.1758 3.1663 -2.7919 8.8001 1.9586 . :i pa LDGT2 81-85 10.4986 9.46232 -4.5397 2.2916 2.2264 2.9027 2.935 8.1749 1.56 1.182 f P9 LDGT2 81-85 10.534 9.45853 -4.6546 2.2745 2.3176 2.9007 -2.96 8.3683 1.5537 1.2902 -0.3956 P10 LDGT2 81-85 10.5104 9.14764 -4.0569 2.3269 1.7791 3.0871 -3.3367 7.9403 1.0377 1.7694 -1.8655 2.3687 P11 LOGT2 81-85 6.70683 1.00692 -0.9625 0.8295 2.6212 0.599 -1.3472 6.2535 1.0658 1.6592 -0.2765 1.4949 1.0962 P12 LDGT2 81-85 5.84441 1.64077 0.9579 0.8376 2.0719 1.2901 -1.7228 6.1373 1.3423 1.3993 -0.7286 1.7599 0.6956 1.5499 P13 LDGT2 81-65 3.26892 0.61651 1.3225 0.6113 0.8575 1.1276 0.3105 2.4597 0.7224 0.0984 0.8037 1.0488 0.8185 0.7462 2.8417 P14 LDGT2 81-85 2.22826 0.53462 1.2795 0.6654 0.9477 1.2845 0.5586 1.7615 0.7222 0.4197 1.0753 0.8413 0.6372 0.838 1.8815 1.2908 P15 LDGT2 81-85 1.43068 0.16253 1.4763 0.6916 0.8805 1.0498 0.7614 1.593 0.6161 0.2931 0.9481 0.6408 0.759 0.6649 1.0119 0.9471 1.4604 P16 LDGT2 81-85 1.39761 0.09202 1.4913 0.6618 0.9629 0.9832 0.8115 1.4561 0.5963 0.4052 0.7374 0.7242 0.7696 0.6414 0.9815 0.9546 1.1597 1.2077 P17 LDGT2 81-85 1.30226 -0.04105 1.2639 0.683 0.8813 0.8942 0.7966 1.641 0.6092 0.4611 0.6586 0.3214 0.791 0.5936 0.982 0.9715 0.9473 1.1888 1.5709 P18 LOGT2 81-85 0.98169 -0.23459 1.3573 0.7081 0.8658 0.8401 0.6551 1.4524 0.5751 0.6074 0.7206 0.4241 0.7757 0.6063 0.7425 0.8257 0.8673 1.0036 1.3323 0.8645 P19 LDGT2 81-85 0.50955 -0.1442 1.2761 0.7138 0.6615 0.8723 0.6415 1.4442 0.6884 0.5997 0.8506 0.4704 0.7392 0.6704 0.7048 0.6711 0.7256 0.4192 1.1487 0.8223 0.7477 IM147 Phase 2 Regression Coefficients, CO, 1981 to 1985 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 08 09 C10 C11 C12 C13 C14 C15 016 C17 018 019 811 LDGT2 81-85 13.6662 10.3119 1.5327 B12 LOGT2 81-85 9.5454 6.3219 0.9685 2.6838 B13 L0GT2 81-85 4.74433 1.25606 1.0757 0.9968 4.0692 B14 LDGT2 81-85 3.50162 1.57213 0.8395 1.1681 2.9369 1.7248 B15 LDGT2 81-85 2.41832 0.59238 1.0109 0.918 1.4255 1.2686 2.1564 B16 LDGT2 81-65 2.31489 0.51543 1.0284 0.8761 1.3639 1.2777 1.5125 2.5385 B17 LOGT2 81-85 2.02742 0.13816 1.0611 0.7865 1.3133 1.3648 1.0331 2.4287 2.7411 B18 LDGT2 81-65 1.59498 -0.3698 1.0435 0.6038 0.976 1.1376 0.9439 2:1064 2.2992 1.2758 B19 LDGT2 81-65 1.13333 -0.2565 , 1 0.8755 0.9246 0.9223 0.7526 1.3895 2.0596 1.205 1.0168 ------- 0 1 VO IM147 Composite Regression Coefficients, CO, 1986 to 1987 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT2 86-87 10.9444 12.9835 9.9921 P2 LOGT2 86-87 7.55854 7.97406 5.9552 2.7507 P3 LDGT2 86-87 6.99501 6.13276 4.0152 1.8868 4.4945 P4 LDGT2 86-87 6.97238 5.95324 3.4929 1.8483 3.4182 2.3575 P5 LDGT2 86-87 6.96833 6.05169 3.0277 1.6681 2.9579 1.9105 1.5754 P6 LDGT2 86-87 6.92215 5.98054 2.189 1.8432 2.2804 1.5316 0.9365 3.0663 P7 LDGT2 86-87 6.94695 5.98146 2.2344 1.7997 2.1735 1.5149 0.7182 3.2075 0.349 P8 LDGT2 86-87 6.86254 5.92929 3.5275 1.4609 2.7997 0.5921 0.3135 i 2.0879 -0.2127 2.0942 P9 LDGT2 86-87 6.87158 5.95327 3.6626 1.5007 2.5734 0.597 0.1327 1.6367 0.0355 1.5977 1.3024 P10 LDGT2 86-87 6.8465 5.84494 3.819 1.5958 2.1327 0.6932 0.2752 1.0075 -0.5524 1.789 -0.4389 2.5615 P11 LDGT2 86-87 4,00509 1.79665 1.9731 0.449 2.1242 -1.7461 2.0304 1.3946 -0.7103 1.246 1.7513 2.4635 1.0775 P12 LDGT2 86-87 3.20981 1.45021 1.8521 0.2977 1.3666 1.06 2.0745 -0.8491 0.2216 1.3232 1.1353 1.6136 0.6802 1.8785 P13 LDGT2 86-87 2.18993 0.93828 1.6364 0.6089 1.2938 1.9515 0.3665 -0.5328 -0.0587 1.2006 0.4256 1.2367 0.6922 1.2207 2.0554 P14 LDGT2 86-87 1.37534 0.49488 1.2757 0.6254 1.0665 0.9774 0.576 0.938 0.198 1.1141 0.5629 1.3587 0.6742 0.7466 1.5811 1.3798 P15 LDGT2 86-87 0.82014 0.15633 1.0946 0.6111 0.9745 0.7593 0.7578 1.1728 0.4226 1.0261 0.3048 1.0888 0.7145 0.6425 0.9896 1.02 1.5309 P16 LDGT2 86-87 0.69514 -0.03757 1.0745 0.6025 1.0009 0.7342 0.6475 1.2043 0.645 0.9395 0.3972 0.8668 0.7284 0.6071 0.9474 1.1426 0.9553 2.2245 P17 LDGT2 86-87 0.60214 -0.0821 1.1799 0.6279 0.8864 0.6465 0.8573 0.9847 0.4956 1.0034 0.4603 0.6907 0.7232 0.6711 0.8613 1.1323 0.9966 1.4224 1.7537 P18 LDGT2 66-87 0.44333 0.03364 1.1598 0.6606 0.7253 0.5353 0.9864 1.1754 0.4065 0.9965 0.5489 0.8032 0.7002 0.7326 0.6569 1.0226 0.9838 0.6221 1.7755 0.6944 P19 LDGT2 86-87 0.18065 -0.04526 1.0637 0.7123 0.7409 0.8109 0.662 0.9665 0.652 0.7498 0.6309 0.6545 0.715 0.7245 0.7123 0.6921 0.7437 0.6817 1.153 0.6878 0.7348 IM147 Phase 2 Repression Coefficients, CO, 1986 to 1987 Model Year LDGT2's Repression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 86-87 7.78582 5.80047 1.4386 B12 LDGT2 86-87 4.68466 3.16394 0.9308 2.6872 B13 LDGT2 86-87 3.05474 1.67487 0.9606 1.5982 2.7862 B14 L0GT2 86-87 2.02059 1.1291 0.8908 1.1326 2.3405 1.8259 B15 LDGT2 66-87 1.24999 0.47151 0.9544 0.9442 1.4229 1.3063 2.1977 B16 LDGT2 86-87 1.06317 0.16747 0.9833 0.8605 1.3578 1.508 1.2803 3.3341 B17 LDGT2 86-87 0.90457 -0.06369 0.9757 0.9354 1.2019 1.5292 1.2917 2.1953 2.4754 B18 LDGT2 86-87 0.76129 -0.02736 0.9433 1.0434 0.9962 1.3721 1.2582 1.4319 2.4588 0.8181 B19 LDGT2 86-87 0.34719 -0.0285 0.9787 0.9885 0.9663 0.8916 0.9566 0.9606 1.5579 0.947 1.0918 ------- 0 1 K> O 1 M147 Composite Repression Coefficients, CO, 1988 to 1995 Model Year LDGT2's Reqression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT2 68-95 7.29843 8.32241 9.6375 P2 LDGT2 88-95 5.72678 5.74122 -0.9189 2.3772 P3 LD6T2 88-95 5.24439 4.67052 -1.3811 1.1541 5.5635 P4 LDGT2 88-95 5.14626 4.26813 -1.2272 1.0883 3.6162 4.9841 P5 LDGT2 88-95 5.09801 4.25023 -1.2746 0.9561 3.0791 3.914 1.7536 P6 LDGT2 88-95 5.03227 4.30487 -2.116 1.045 2.75 2.5651 1.1445 4.8242 P7 LDGT2 88-95 4.94785 3.9903 -1.6246 0.9734 2.2915 2.261 0.0539 4.0718 2.4087 P8 LDGT2 88-95 4.69866 3.89264 -1.7403 1.0233 1.931 1.7839 0.2408 2.7207 1.5017 2.1926 P9 LDGT2 88-95 4.85108 3.81187 -1.2151 1.0106 1.7096 1.9917 -0.0267 1.4417 1.4322 1.6155 3.0089 P10 L0GT2 88-95 4.82605 3.68452 -1.0168 1.0461 1.7253 1.7644 -0.2401 1.2295 1.1312 1.6015 1.909 2.3455 P11 LDGT2 88-95 3.00665 1.74476 1.0382 0.6435 2.2076 0.8361 0.297 0.8814 1.2011 -0.12 2.4158 1.0538 1.0849 P12 L0GT2 88-95 2.71202 1.40261 1.3589 0.5437 1.8138 1.0668 0.3264 0.7761 1.4578 0.0633 1.6352 0.6125 0.9029 1.3647 P13 L0GT2 88-95 1.9294 0.60523 1.2096 0.6003 1.3401 0.909 0.5644 1.2411 0.8508 0.3162 0.575 0.1662 0.8851 0.7097 2.7113 P14 L0GT2 88-95 1.26478 0.42753 1.2507 0.6101 1.151 0.906 0.608 1.707 0.8161 0.593 0.3194 0.7475 0.7497 0.8072 1.7136 1.2385 P15 L0GT2 88-95 0.8779 0.26841 0.9697 0.6415 1.0695 0.7999 0.5865 1.284 0.8121 0.5778 0.5467 0.957 0.7335 0.8178 1.057 0.6384 1.2288 P16 LDGT2 68-95 0.79251 0.18425 1.0108 0.6516 1.0689 0.8061 0.5356 1.3495 0.8301 0.6371 0.4534 0.9417 0.7116 0.8113 1.0192 0.8943 0.9962 1.2531 P17 LDGT2 88-95 0.69731 0.07985 0.6509 0.7135 0.6792 0.7703 0.5294 1.3068 0.6755 0.6246 0.4121 0.958 0.7118 0.8086 0.9187 0.9121 0.9794 0.751 2.056 P18 LDGT2 88-95 0.52944 0.00314 0.7855 0.7161 0.8485 0.6381 0.64 1.2114 0.5735 0.7301 0.4666 0.6616 0.7036 0.7599 0.6897 0.6686 0.9015 0.6581 1.2682 1.0027 P19 LDGT2 86-95 0.18228 0.02029 1.0068 0.7103 0.7155 0.8232 0.7148 0.8349 0.6247 0.8302 0.6614 0.7366 0.7094 0.724 0.6921 0.7083 0.7232 0.6986 1.0875 0.7309 0.7311 1M14 7 Phase 2 Regression Coefficients, o o 1988 to 1995 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 05 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 CIS B11 LDGT2 88-95 5.55431 4.86126 1.4596 B12 LDGT2 66-95 3.81315 2.54349 1.2451 2.0945 B13 LDGT2 88-95 2.63707 0.83295 1.1846 0.9524 3.6628 B14 LDGT2 88-95 1.77124 0.72633 1.0256 1.1004 2.4663 1.6377 B15 LDGT2 86-95 1.22283 0.50627 1.0014 1.1189 1.5439 1.0872 1.7128 816 LDGT2 88-95 1.11757 0.38075 0.9754 1.1094 1.4934 1.1553 1.4158 1.6429 B17 LOGT2 86-95 0.95396 0.13503 0.9673 1.1001 1.2482 1.2114 1.3643 0.9484 2.9337 B18 LDGT2 88-95 0.72512 -0.00576 0.9541 1.0331 0.9195 1.156 1.2511 1.1276 1.7748 1.3609 B19 LDGT2 88-95 0.28372 0.00704 0.9656 0.9834 0.9474 0.9412 1.0187 0.8827 1.65 0.9988 0.9646 ------- 0 1 to IM147 Composite Regression Coefficients, CO, 1996 and Newer Model Year LDGT2's Reqression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C16 C19 P1 LDGT2 96+ 1.3088 1.42622 17.1995 P2 LDGT2 96+ 1.1479 1.00342 -1.078 3.0953 P3 LDGT2 96+ 0.98949 0.98019 -2.6776 0.7119 4.3482 P4 LOGT2 96+ 0.82952 0.97186 -4.3164 0.4586 -0.0693 9.4052 P5 LDGT2 96+ 0.82272 0.92541 -6.1685 0.6234 0.0838 7.9052 1.1172 P6 LDGT2 96+ 0.60854 0.93687 -2.5537 0.4345 0.9927 5.0074 0.2744 5.1496 P7 L0GT2 96+ 0.78612 0.93219 0.8299 0.1902 0.867 2.2463 -0.6382 7.4747 1.8026 P8 L0GT2 96+ 0.7922 0.9145 -0.5421 0.3098 1.1613 -0.1172 -0.5476 8.6914 1.4725 1.1768 P9 LDGT2 96+ 0.79439 0.90764 0.4323 0.2804 1.57 1.9801 0.584 8.2734 1.281 1.4763 1.8463 P10 LDGT2 96+ 0.80176 0.89411 0.6291 0.4422 1.1925 -2.1872 -0.6255 7.6689 1.3637 1.6415 0.9728 1.2038 P11 LDGT2 96+ 0.45829 0.43587 2.7438 0.9615 1.1761 -4.6902 -0.051 7.5376 0.5427 2.6889 0.9224 1.6589 0.7053 P12 LOGT2 96+ 0.43361 0.27262 1.6129 0.8974 1.1808 -3.2179 0.1827 6.7307 0.6851 2.4296 1.0559 0.404 0.7027 1.0882 P13 LDGT2 96+ 0.35545 0.13528 0.1926 1.1224 -0.5669 2.979 0.8935 3.5682 0.4381 0.8157 0.0928 0.4589 0.7161 0.668 1.9197 P14 LDGT2 96+ 0.30207 0.07328 2.5856 1.3224 -0.4318 1.9157 1.1728 1.7862 0.501 0.7493 0.097 1.2673 0.6916 0.617 1.1886 1.8756 P15 LDGT2 96+ 0.11959 0.01946 0.6675 0.9676 -0.1805 1.423 0.6812 2.4687 0.9513 0.6769 0.0759 1.097 0.7126 0.6574 1.1486 0.3594 1.6893 P16 LDGT2 96+ 0.09136 0.01277 -0.0995 0.9293 -0.1463 2.0115 0.7771 2.1495 0.9056 0.4462 0.4595 0.6343 0.7138 0.6364 1.0287 0.625 1.3436 1.1992 P17 LDGT2 96+ 0.08571 0.00815 0.3122 0.914 0.0235 1.495 0.6462 2.28 0.6668 0.6935 0.6686 0.5081 0.7115 0.6609 0.9459 0.5995 1.3506 0.864 1.1952 P18 L0GT2 96+ 0.07115 -0.00155 0.1596 0.9576 0.1434 1.207 0.6318 1.7452 0.5467 0.8902 0.2782 0.9571 0.7097 0.6328 0.8136 0.7233 1.1938 0.7903 0.6252 0.8195 P19 LDGT2 96+ 0.02689 0.00249 -0.2617 0.9243 0.5171 0.6083 0.8825 1.183 0.5554 0.926 0.5734 0.7532 0.715 0.657 0.7458 0.7682 0.6645 0.7015 0.8289 0.6613 0.7521 IM147 Phase 2 Repression Coefficients, CO, 1996 and Newer Model Year LDGT2's Reqression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C16 C19 B11 LDGT2 96+ 1.20408 1.12226 0.7609 B12 LDGT2 96+ 0.62558 0.40305 0.9685 2.1567 B13 LDGT2 96+ 0.54186 0.23919 0.9716 1.5278 2.1489 B14 LDGT2 96+ 0.44605 0.15037 0.9303 1.3986 1.0695 2.924 B15 LDGT2 96+ 0.24307 0.03103 0.9646 1.2679 1.1889 0.9274 2.4187 B16- LDGT2 96+ 0.21899 0.02112 0.9694 1.1748 0.9741 1.2313 1.9455 1 1.756 B17 LDGT2 96+ 0.19501 0.00313 0.9536 1.2274 0.7628 1.1104 1.9911 0.8779 2.8529 B18 LDGT2 96+ 0.14188 -0.00797 0.9522 1.0935 0.6831 1.2308 1.578 1.0698 0.7732 1.802 B19 LDGT2 96+ 0.06636 0.01816 0.9663 0.9491 0.8971 1.0203 0.7701 0.8901 1.1191 1.0268 1.2815 ------- 0 1 to to IM147 Composite Repression Coefficients, CO, 1981 to 1982 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 V ii C11 C12 C13 C14 C15 C16 C17 C16 C19 P1 Lb6V 81-82 7.99839 11.8892 10.2939 P2 LDGV 81-82 6.58035 8.47137 3.9709 2.1482 . .I P3 LDGV 81-82 5.95523 7.35499 2.3706 1.1885 4.3273 ' . ;Ł P4 LDGV 81-82 5.84107 6.94384 0.9352 1.4 2.5147 4.4184 . 'ir P5 LDGV 81-82 5.67429 6.92733 -0.9051 1.3052 1.0082 3.8467 3.2208 P6 LDGV 81-82 5.60045 7.2234 -1.4742 1.4463 0.1894 3.6997 2.3343 2.4304 . :p P7 LDGV 81-82 5.42769 6.90375 -1.2377 1.0873 0.2076 2.5448 1.0667 3.0396 2.007 . .'I P8 LDGV 81-82 5.17703 6.82475 -1.395 0.841 -0.567 2.793 1.0784 2.1889 0.6894 3.583 'L P9 LDGV 81-82 5.11543 6.5602 -1.8042 0.9128 -0.049 1.6424 0.3327 1.1075 0.5581 3.277 3.6636 . 1! P10 LDGV 81-82 5.09166 6.36664 -1.8837 0.9302 -0.4412 1.7282 0.1522 1.3964 0.1539 3.943 1.4796 2.6588 . P11 LDGV 81-82 3.36176 2.42133 -1.7143 0.9431 0.3844 -1.067 0.3333 1.7489 0.1534 1.8436 3.6102 2.5886 0.995 P12 LDGV 81-82 2.45131 1.51191 -0.6053 0.8264 0.3739 0.1846 0.3965 1.281 0.4926 1.5431 2.6906 2.8681 0.6782 1.3748 P13 LDGV 81-82 1.64222 0.81857 0.547 0.7456 0.5366 1.486 -0.1283 1.4627 0.5971 1.0752 1.2221 2.1211 0.6873 0.8834 2.0965 P14 LDGV 81-82 1.30098 0.6002 0.761 0.7639 0.4632 0.8416 -0.3448 2.0537 0.7512 1.0435 0.8919 2.1412 0.6644 0.6957 1.6479 1.2133 P15 LOGV 81-82 0.99586 0.48274 1.4122 0.6691 0.5819 1.4446 0.1117 1.6713 0.9146 0.8677 0.8203 0.9725 0.6913 0.6885 1.0384 0.6235 1.3336 P16 LDGV 81-82 0.92334 0.37973 1.5086 0.6203 0.6059 1.5811 0.0371 1.4292 1.0143 0.8284 0.7229 0.8179 0.6963 0.6941 0.9588 0.9625 0.9091 1.8136 P17 LDGV 81-82 0.64728 0.19446 1.3677 0.6505 0.544 1.2293 0.3115 1.4563 0.9885 0.8554 0.4817 0.4999 0.7063 0.6726 0.9646 0.9625 0.9546 1.1516 2.0177 P18 LDGV 81-82 0.52289 0.04744 1.1672 0.7297 0.3484 1.2653 0.3087 1.1707 0.6643 0.8253 0.7848 Ś0.8903 0.6988 0.7311 0.7096 0.8592 0.8362 1.0156 1.7867 0.8424 P19 LDGV 81-82 0.31963 -0.00493 1.0652 0.7221 0.5675 1.1578 0.3797 0.9834 0.7238 0.6876 0.994 0.7578 0.7005 0.7001 0.7374 0.7792 0.7058 0.244 1.232 0.646 0.9011 IM14 7 Phase 2 Regression Coefficients, CO, 1981 to 1982 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C0 C7 C8 C9 C10 i; C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 81-82 7.87561 6.04896 1.2099 B12 LDGV 81-82 4.28681 4.12849 0.9421 1.067 B13 LDGV 81-82 2.91132 1.85708 0.9717 1.0199 3.5026 B14 LDGV 81-82 2.35783 1.43426 0.9278 0.7999 2.9182 1.5911 B15 LDGV 81-82 1.81322 0.97781 0.9942 0.8184 1.529 0.995 2.1475 B16 LDGV 81-82 1.64673 0.74808 1.0076 0.8215 1.3625 1.3316 1.1639 3.4285 B17 LDGV 81-82 1.35105 0.12268 1.0045 0.658 1.3822 1.29 1.2026 1.7464 4.1093 B18 LDGV 81-82 0.83894 0.09966 0.9679 0.9648 0.9202 1.1549 1.0845 1.5149 3.5169 1.2191 B19 LDGV 81-82 0.57427 -0.10409 0.9662 0.9148 0.9743 1.0447 0.9099 0.3112 2.66 0.9357 1.318 ------- 0 1 K) IM147 Composite Repression Coefficients, CO, 1983 to 1985 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 cie C19 Pi LDGV 83-85 5.65667 8.71977 10.0773 P2 LDGV 83-65 4.39498 6.05704 2.7834 2.4038 P3 LDGV 83-85 3.79704 5.05452 2.132 1.2291 4.6404 P4 LDGV 83-85 3.62371 4.67806 0.5952 1.3042 2.354 5.4862 P5 LDGV 83-85 3.59464 4.60345 0.0971 1.2134 2.0391 4.9125 1.269 P6 LDGV 83-85 3.59268 4.59476 0.3543 1.2106 2.0275 4.62 1.1146 1.0934 P7 LDGV 83-85 3.44031 4.32493 -0.5781 1.2446 0.8896 4.3893 0.6174 1.6598 1.5993 P6 LDGV 83-85 3.36343 4.2564 0.0459 1.093 0.9849 3.5699 0.5789 0.9952 1.2147 1.5455 P9 LDGV 83-85 3.33029 4.1852 -0.1659 1.1245 0.9751 3.3156 0.1104 0.6868 1.2398 1.3047 2.0264 P10 LDGV 83-85 3.29987 4.12599 0.0405 1.1476 0.8196 3.3153 -0.2824 0.8442 1.1246 1.22 1.1895 1.552 P11 LDGV 83-85 2.20291 2.25511 0.3918 0.6788 0.9189 1.6168 0.932 1.9937 0.5168 0.7853 1.9632 1.1107 0.7316 P12 LDGV 83-85 1.87972 1.58471 0.5439 0.7415 0.6536 1.6537 0.727 1.925 0.7033 0.7588 1.3654 1.0997 0.6415 1.1026 P13 LDGV 83-85 1.28242 0.91668 0.7111 0.7088 0.7627 1.8163 0.2956 1.6032 0.6711 0.6173 1.5002 0.6472 0.6692 0.8018 1.9269 P14 LDGV 83-65 1.03127 0.58797 0.6489 0.7423 0.3578 1.6308 0.4662 1.5147 0.8643 0.6344 1.5828 0.5435 0.6488 0.7632 1.5457 1.1794 P15 LDGV 63-85 0.69673 0.34058 0.7164 0.7324 0.354 1.9052 0.5776 1.1365 0.8592 0.5437 1.0426 0.8335 0.6765 0.708 0.9536 0.8489 1.4982 P16 LDGV 83-85 0.6836 0.3288 0.6397 0.7348 0.4045 1.8616 0.5991 1.1119 0.8675 0.5275 0.9881 0.8608 0.6754 0.7075 0.9317 0.628 1.269 0.931 P17 LDGV 83-65 0.61016 0.1403 0.7783 0.7181 0.5497 1.5209 0.4902 1.1225 0.8292 0.592 0.9313 0.8148 0.6905 0.7307 0.9393 0.8363 1.1503 0.6464 1.6992 P18 LDGV 83-85 0.35809 0.01542 1.0268 0.6988 0.5932 1.2395 0.5512 1.1677 0.6596 0.7122 0.6022 0.6293 0.6933 0.7668 0.6573 0.8294 1.0268 0.6176 1.4839 0.9529 P19 LDGV 83-85 0.25947 -0.01001 1.0127 0.7023 0.6463 1.1344 0.5259 1.0927 0.704 0.6265 0.7924 0.6521 0.7068 0.755 0.7028 0.7409 0.743 0.4739 1.3579 0.7294 0.7327 IM14 Y7 Phase 2 Regression Coefficients, CO, 1983 to 1985 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C18 C17 C18 C19 B11 LDGV 83-85 4.69212 5.7524 1.0243 B12 LDGV 83-85 2.83315 3.12116 0.9251 1.6564 013 LDGV 83-85 1.86354 1.56942 0.9295 1.1338 2.8182 B14 LDGV 83-85 1.55825 1.15742 0.9063 1.0736 2.325 1.5629 B15 LDGV 83-85 1.07233 0.69467 0.943 0.9688 1.4054 1.0868 2.1751 B16 LDGV 83-85 1.05389 0.68187 0.9437 0.9647 1.3864 1.0625 1.8313 '1.3471 B17 LDGV 83-85 0.89995 0.27297 0.9612 0.9964 1.3316 1.1059 1.5922 0.8992 2.8516 B18 LDGV 83-85 0.54856 0.0412 0.9415 1.0459 0.8624 1.115 1.3817 1.1042 2.1846 1.3008 B19 LDGV 83-85 0.4078 -0.005 0.9594 1.0223 0.9206 0.9952 0.9616 0.8242 1.9266 0.9809 1.0636 ------- o t Ni 4^ IM147 Composite Repression Coefficients, CO, 1986 to 1989 Model Year LDGV's Regression Coeffiq tients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 06 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGV 86-89 4.67926 5.71778 10.6405 r P2 LDGV 86-89 3.51709 3.81768 0.1364 2.695 P3 LDGV 86-89 3.15842 3.11069 1.0124 1.6189 3.9459 . f P4 LDGV 86-89 3.14894 3.04972 0.8613 1.649 3.2671 1.3569 . ' P5 LDGV 86-89 3.10204 3.03267 0.7486 1.4903 2.3894 0.6299 1.8322 S P6 LDGV 86-89 3.09685 3.06465 0.6252 1.4634 2.3373 0.5195 1.5731 0.9056 P7 LDGV 86-89 3.02573 2.66107 1.1098 1.2652 1.9531 0.4921 0.7359 0.8227 2.1087 P8 LDGV 86-89 2.94297 2.66329 0.7801 1.2178 1.8801 0.6138 0.7593 -0.8902 1.5027 2.0203 . < P9 LDGV 66-89 2.67652 2.54621 <0.0543 1.2924 1.523 1.0828 0.2369 -1.106 1.5137 1.3545 3.0417 P10 LDGV 86-89 2.74734 2.37808 -0.6515 1.4027 1.4393 1.1934 -0.1403 -0.5027 1.2073 0.9038 -0.333 4.4623 P11 LDGV 66-89 1.60566 1.04515 0.0874 0.8282 1.4645 1.1894 0.8248 -0.2838 0.6291 1.0042 1.4625 2.0759 0.8639 P12 LDGV 86-89 1.49356 0.71931 0.2804 0.7493 0.9793 1.513 0.8296 0.0469 0.689 1.0787 0.769 1.9598 0.7434 1.2423 P13 LDGV 66-89 1.06067 0.36461 0.8207 0.6529 0.9226 1.272 0.8321 0.1223 0.6425 0.6893 0.7679 1.6076 0.7379 0.7345 2.1093 P14 LDGV 86-89 0.82231 0.26307 0.8376 0.5953 0.933 1.0658 0.6038 0.4366 0.672 0.8766 0.8378 1.4217 0.7134 0.7172 1.4718 1.4518 P15 LDGV 86-89 0.57466 0.17901 0.6519 0.6644 0.7016 1.0893 0.6865 0.5537 0.8363 0.7644 0.7334 1.1584 0.7057 0.7004 0.9553 0.9039 1.5623 P16 LDGV 86-89 0.54701 0.18097 0.946 0.6657 0.6984 0.9951 0.7165 0.6267 0.8323 0.7529 0.6652 1.0979 0.7051 0.7127 0.9675 0.8974 1.1743 1.3482 P17 LDGV 86-89 0.49266 0.13124 0.9156 0.6933 0.6436 0.674 0.7016 0.686 0.7814 0.7572 0.5724 0.9245 0.7053 0.7153 0.9399 0.6866 1.0881 0.9489 1.6484 P18 LOGV 66-89 0.29132 0.05317 0.9002 0.685 0.7018 0.9872 0.6206 0.7756 0.6789 0.7453 0.7249 0.7506 0.7065 0.7322 0.7864 0.697 0.6566 1.1084 1.3762 0.8805 P19 LDGV 86-89 0.1491 0.02273 0.8668 0.6976 0.726 0.8665 0.6806 0.7684 0.7148 0.6669 0.6743 0.7484 0.7087 0.7177 0.7096 0.7249 0.7202 0.6852 0.9434 0.7022 0.8271 MU 17 Phase 2 Regression Coefficients, CO, 1986 to 1989 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 86-89 4.18172 3.9913 1.1684 B12 LDGV 86-89 2.21032 1.63478 1.0475 1.964 813 LDGV 86-89 1.5326 0.82575 1.0148 1.0859 3.0946 B14 LDGV 86-89 1.21271 0.6268 0.9705 1.0461 2.17 2.0218 B15 LDGV 86-89 0.82446 0.40601 0.9653 0.9766 1.3403 1.2022 2.2818 B16 LDGV 86-89 0.77323 0.3766 0.9636 0.9888 1.3495 1.1956 1.6328 2.1301 B17 LDGV 86-89 0.67252 0.23667 0.9624 0.9661 1.2888 1.1945 1.4722 1.3414 2.4799 B18 LDGV 86-89 0.40383 0.10653 0.9566 0.984 1.0723 0.9316 1.1674 1.5464 1.9865 1.1802 B19 LDGV 86-69 0.20724 0.05209 0.9611 0.965 0.9597 0.9776 0.9801 0.9182 1.3409 0.9424 1.1303 ------- o to IM147 Composite Regression Coefficients, CO, 1990 to 1995 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 ca C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGV 90-95 5.5829 4.52493 10.561 P2 LDGV 90-95 4.19016 2.55934 -2.4176 3.5176 P3 LDGV 90-95 3.59909 1.7089 -2.0167 1.8381 5.8658 P4 LDGV 90-95 3.52751 1.57279 -2.1878 1.7555 4.1493 4.0887 P5 LDGV 90-95 3.35967 1.52742 -1.8847 1.3658 3.0822 2.7595 3.0615 P6 LDGV 90-95 3.30911 1.53896 -2.1747 1.433 2.5763 2.1516 2.309 3.2307 P7 LDGV 90-95 3.20465 1.37753 -1.7351 1.2877 2.4869 1.7618 0.6153 2.8007 3.0019 P8 LDGV 90-95 3.08075 1.25584 -0.8703 1.3157 1.5808 1.0639 0.6544 1.6621 1.8801 3.4334 P9 LDGV 90-95 3.03308 1.28742 -0.0409 1.3472 1.2629 1.2814 0.1146 0.9574 2.0383 2.5679 2.6687 P10 LDGV 90-95 2.99284 1.29175 -0.6667 1.2906 1.1937 1.4946 -0.2007 0.8633 1.7797 2.4995 0.7394 2.5847 P11 LDGV 90-95 1.79334 0.39573 0.8907 0.875 0.9738 0.6525 0.6724 1.4441 0.9505 1.8704 1.0336 1.2774 0.946 P12 LDGV 90-95 1.48201 0.27031 0.8772 0.7905 0.9191 0.8699 0.6683 0.8286 0.9734 1.6847 0.9938 1.2059 0.7634 1.2265 P13 LDGV 90-95 1.22124 0.18956 0.9352 0.7049 1.1123 0.7943 0.6333 0.6174 0.819 1.3469 0.6633 0.9046 0.7518 0.8615 1.6924 P14 LDGV 90-95 0.96232 0.19909 1.2249 0.6231 0.6871 0.5243 0.6879 1.0339 0.9491 1.0285 1.2323 0.735 0.7153 0.7873 1.2566 1.1834 P15 LDGV 90-95 0.67059 0.09713 1.1865 0.6235 0.8229 0.6645 0.7023 0.9149 0.9983 0.9575 1.0487 0.5318 0.7254 0.7636 0.8921 0.7821 1.3034 P16 LDGV 90-95 0.61862 0.0845 1.1157 0.6363 0.8291 0.672 0.6393 0.9358 0.9492 0.8718 1.0229 0.6422 0.7279 0.7605 0.8704 0.8116 0.9084 1.6959 P17 LDGV 90-95 0.56789 0.04406 1.1034 0.6685 0.6934 0.7551 0.6966 0.8027 0.9189 0.7923 0.9078 0.6476 0.7299 0.7393 0.8716 0.8032 0.8822 1.1592 1.9586 P18 LDGV 90-95 0.38985 0.01218 1.0107 0.6685 0.7981 0.7109 0.6672 0.9253 0.6833 0.6064 0.9037 0.6447 0.7163 0.726 0.7487 0.7691 0.7468 1.0451 1.2236 0.9766 P19 LDGV 90-95 0.13949 0.01779 0.7784 0.7029 0.7108 0.7886 0.6798 0.8583 0.733 0.6745 0.7331 0.685 0.7139 0.7148 0.711 0.7002 0.7152 0.6362 0.9674 0.7322 0.7843 IM14 V7 Phase 2 Repression Coefficients, CO, 1990 to 1995 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 90-95 4.103 2.35754 1.4957 B12 LDGV 90-95 2.32229 0.91944 1.1523 2.0251 B13 LDGV 90-95 1.76794 0.5229 1.0723 . 1.2322 2.7642 B14 LDGV 90-95 1.40264 0.44364 0.9979 1.1201 2.0763 1.6438 B15 LDGV 90-95 0.97281 0.25022 1.0003 1.0639 1.4071 1.0512 1.8686 B16 LDGV 90-95 0.88873 0.20771 1.0023 1.0561 1.3449 1.0929 1.2656 2.5533 B17 LDGV 90-95 0.79003 0.10796 0.9981 1.009 1.2809 1.0795 1.1947 1.6393 3.0349 B18 LDGV 90-95 0.54771 0.04247 0.9734 0.9933 1.0674 1.0346 1.0249 1.4049 2.0212 1.3189 B19 LDGV 90-95 0.19036 0.03127 0.9671 0.9711 0.9708 0.9442 0.9644 0.8616 1.3588 0.9845 1.0769 ------- 0 1 ro ON IM147 Composite Regression Coefficients, CO, 1996 and Newer Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 Pi LDGV 96+ 4.2502 1.81884 53.0314 . u P2 LDGV 96+ 2.79375 0.66889 -6.9133 5.3642 . i P3 LDGV 96+ Ś2.31951 0.49886 -22.4698 2.9121 7.8787 P4 LDGV 96+ 2.30066 0.48852 -20.2357 2.6582 5.7046 6.7489 . i P5 LDGV 96+ 2.29856 0.4874 -19.0504 2.2906 4.9636 7.0228 1.4214 . ; P6 LDGV 96+ 2.17485 0.36586 -18.1249 2.1863 5.4292 5.3931 -0.6432 6.8037 P7 LDGV 96+ 2.17872 0.34724 -17.6869 2.1129 5.4618 5.4355 0.6956 6.7245 0.3983 P8 LDGV 96+ 2.17729 0.32839 -15.8387 1.9759 5.2235 5.2769 -1.1683 6.4131 0.2201 1.4309 P9 LDGV 96+ 2.16612 0.4024 -11.8473 1.4566 4.5337 6.4172 -1.2079 5.4459 0.3128 0.4287 3.2042 P10 LDGV 96+ 2.17081 0.40258 -11.4985 1.3861 4.5521 6.4182 -1.1694 5.5108 0.2871 0.3698 2.6985 0.5987 P11 LOGV 96+ 1.3457 0.04672 -1.5411 1.3212 1.6098 6.5926 -0.785 3.2891 0.7158 -0.2278 1.3044 1.2284 1.1171 P12 LDGV 96+ 0.94601 0.03677 0.1569 1.1325 0.7478 7.1762 -2.7331 2.8294 0.8581 0.5253 -0.3667 2.4476 0.8375 1.6915 P13 LDGV 96+ 0.83127 0.04161 6.5433 0.7725 0.7897 2.3429 -1.2999 2.5769 0.8026 1.0385 0.6814 0.7591 0.8719 1.3576 1.4743 P14 LDGV 96+ 0.36237 0.07318 -1.7846 0.7477 1.0289 0.323 0.8116 2.2639 0.6876 0.4145 1.1104 0.4144 0.6973 0.7747 0.877 1.8748 P15 LDGV 96+ 0.29315 0.05053 0.3484 0.5009 0.8684 2.0337 0.8189 2.1406 0.8321 0.2796 -0.0113 0.3684 0.7271 0.7433 1.1116 1.0359 0.8817 P16 LDGV 96+ 0.29205 0.04924 0.0274 0.4967 0.8209 2.0574 0.6019 2.0106 0.6654 0.1479 0.1851 0.3987 0.7276 0.7477 1.1209 1.0535 0.7882 0.5132 P17 LDGV 96+ 0.23667 0.02554 -0.2222 0.7177 0.3166 1.7918 0.9891 1.2672 0.7007 0.7511 0.1642 0.6983 0.7129 0.7751 0.8196 1.3698 0.5666 0.1481 2.2487 P18 LDGV 96+ 0.211 0.02076 1.6186 0.5321 0.4849 0.6707 1.2541 1.116 0.6247 0.9096 0.3981 0.9301 0.7173 0.7777 0.8774 0.7931 1.0214 0.0389 1.7607 0.6392 P19 LDGV 96+ 0.05419 0.00995 0.7814 0.7041 0.7989 0.4985 0.7204 0.9434 0.6972 0.728 0.6052 0.6829 0.7065 0.7486 0.7235 0.6863 0.7334 0.6843 0.7333 0.7391 0.7203 M147 Phase 2 Repression Coefficients, CO, 1996 and Newer Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 015 C16 C17 C18 C19 B11 LDGV 96+ 2.8132 0.6365 1.9389 B12 LDGV 96+ 1.55476 0.22139 1.2687 2.444 B13 LDGV 96+ 1.21697 0.12297 1.1955 1.6069 1.8026 B14 LDGV 96+ 0.54556 0.13004 0.9864 1.1698 1.3748 2.3969 B15 LDGV 96+ 0.4606 0.1147 0.9893 1 0617 1.3564 1.8394 0.7573 B16 LDGV 96+ 0.4633 0.09077 0.985 1.0604 1.4116 1.8485 0.4981 1.6802 B17 LDGV 96+ 0.33866 0.04766 0.968 1.0626 1.1629 1.9325 0.5734 0.6152 3.1735 B18 LDGV 96+ 0.30627 0.03898 0.9596 1.0743 1.0207 1.4501 0.9473 0.8378 2.7247 0.6646 B19 LDGV 96+ 0.07437 0.01481 0.9598 1.0098 0.9451 0.986 0.9371 1.0864 1.159 0.9807 0.9897 ------- 0 1 to IM147 Composite Regression Coefficients, NOx, 1981 to 1985 Model Year LDGTI's - Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT1 81-85 1.62529 2.96654 22.8751 P2 LDGT1 81-85 0.82038 0.84089 -0.5552 5.7627 P3 LDGT1 81-85 0.76521 0.73117 1.5629 4.5821 7.1986 P4 LDGT1 81-85 0.75286 0.70493 -1.1236 4.4969 4.9474 8.0174 P5 LDGT1 61-85 0.74049 0.73769 -3.3308 4.0049 3.8497 6.663 3.702 P6 LDGT1 81-85 0.73674 0.71486 -3.9127 4.0869 3.8046 3.6966 3.0011 4.8137 P7 LDGT1 81-85 0.71739 0.69385 -3.0212 3.5234 3.5081 3.9937 -0.1792 6.3496 3.9431 P8 LDGT1 61-85 0.70652 0.67353 2.6651 3.5435 2.1822 2.6706 0.0115 3.2904 2.6238 3.9683 P9 LDGTl 81-85 0.70505 0.67708 -1.8358 3.4819 2.4475 3.1809 0.6337 3.7784 2.952 4.1428 -2.2115 . : P10 LDGT1 81-85 0.70227 0.66689 -2.4177 3.494 2.3056 3.6942 0.2591 3.8357 2.4151 4.2563 -4.019 2.4819 P11 LDGT1 81-85 0.61346 0.34129 0.06 2.6963 1.5535 2.4383 1.5201 5.655 0.7401 2.8173 -2.4441 1.0401 1.3101 P12 LDGT1 81-85 0.59022 0.30932 -0.6809 2.658 1.3697 1.2656 1.4801 3.8951 1.356 2.1592 -2.7018 1.3911 0.8585 1.6633 P13 LDGT1 81-85 0.27799 0.12998 -0.1294 0.9449 1.4979 2.9854 0.7785 0.1823 -0.313 0.9976 0.5738 0.4624 0.694 1.076 2.7355 P14 L0GT1 81-85 0.19309 0.05416 0.6065 0.856 0.9251 1.6344 0.8251 -0.094 0.4201 1.4703 0.998 0.7742 0.5699 0.79 1.6886 1.7582 P15 LDGT1 81-85 0.10867 0.02502 0.5161 0.6742 1.1768 1.4558 0.2902 1.3526 0.7996 0.8754 0.732 0.6248 0.6617 0.8183 0.6141 1.2315 1.3505 P16 LDGT1 81-85 0.09725 0.01363 0.4935 0.6784 1.1208 1.1307 0.3923 0.9511 0.8969 0.5086 0.7311 0.6773 0.699 0.7813 0.8355 1.1375 1.1283 1.0972 P17 L0GT1 81-85 0.09344 0.01174 0.5252 0.7346 0.9639 1.1015 0.6838 0.7005 0.7356 0.6406 0.6359 0.4841 0.6953 0.7762 0.8511 1.1597 1.0383 1.0222 1.0523 P18 LDGT1 81-85 0.05721 0.00436 0.6442 0.6817 1.0499 1.0725 0.6056 1.15 0.5722 0.7474 0.8465 0.5131 0.6842 0.8341 0.6429 0.922 0.8319 1.1396 0.7657 0.8484 R19 LDGT1 81-85 0.0165 0.00076 0.7099 0.6986 0.7653 0.7747 0.7168 0.9007 0.7285 0.7007 0.7342 0.6359 0.7093 0.736 0.7337 0.6964 0.7033 0.7238 0.7471 0.7212 0.7201 IM147 Phase 2 Regression Coefficients, NOx, 1981 to 1985 Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 019 B11 LDGTl 81-85 1.076 0.85959 3.4954 B12 LDGT1 81-85 1.00048 0.67351 2.4254 3.4463 B13 LDGTl 81-85 0.38989 0.20314 0.8889 1.8075 3.7573 B14 LDGTl 61-65 0.27037 0.0891 0.8213 1.2937 2.454 2.3704 B15 LDGTl 81-85 0.15297 0.04337 0.8745 1.292 1.0958 1.7178 1.8359 B16 LDGT1 81-85 0.13123 0.01686 0.9513 1.0861 1.1153 1.5763 1.4991 1.5837 B17 LDGT1 81-85 0.12525 0.01412 0.9423 1.0698 1.128 1.6173 1.4024 1.4526 1.3599 B16 L0GT1 81-85 0.08127 0.0094 0.8966 1.2134 0.8469 1.3163 1.0875 1.7522 1.1639 1.0764 B19 LDGT1 81-85 0.01759 0.00132 0.9552 1.0089 0.0919 0.9568 0.9493 1.0003 1.0401 0.9479 1 ------- o K> oo IM147 Composite Regression Coefficients, NOx, 1986 to 1989 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 i C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT1 86-89 1.23043 2.06694 23.5015 P2 LDGT1 66-89 0.70187 0.74907 1.2554 4.6182 P3 LDGT1 86-89 0.65788 0.66459 1.9511 3.4741 6.6849 P4 LDGT1 86-89 0.65847 0.6622 2.0121 3.4718 6.5638 0.4801 P5 LDGT1 86-89 0.60304 0.65314 0.8979 2.5886 3.7778 0.693 7.82 P6 LDGT1 86-89 0.60221 0.64909 0.8986 2.5939 3.729 -0.3487 7.391 2.7912 P7 LDGT1 86-89 0.58554 0.63588 -0.2658 2-2631 2.7626 0.2519 4.6311 4.814 3.3278 P0 LDGT1 86-89 0.57068 0.60583 0.1372 2.1042 2.0549 -1.463 4.181 1.0084 2.9915 4.4107 P9 LDGT1 86-89 0.57122 0.60597 0.1342 2.1057 2.0299 -1.4503 4.1619 0.9698 2.989 4.3952 0.1053 P10 LDGT1 86-89 0.56454 0.59282 0.9316 2.1076 1.5254 -0.8557 3.4898 1.6834 2.3002 4.575 -1.5419 3.1754 P11 LDGT1 86-89 0.44918 0.31619 0.8774 1.547 0.3721 -1.2258 3.7047 1.8503 0.299 3.4571 -0.7772 1.6757 1.4183 P12 LDGT1 86-89 0.39015 0.23116 2.1899 1.3771 -0.3977 -1.386 2.7841 0.6668 0.7794 1.8436 -0.1039 1.861 0.955 2.7651 P13 LDGT1 66-89 0.21107 0.11066 -0.9618 0.6673 1.0909 0.613 1.2202 0.9175 0.3457 1.1499 1.3016 -0.1198 0.8075 1.3691 2.4303 P14 L0GT1 86-89 0.15824 0.06623 -0.4774 0.6974 0.7145 0.8449 1.214 0.9814 0.6256 0.7295 1.5692 0.6474 0.63S9 0.8685 1.8063 1.4835 P15 LDGT1 66-89 0.10582 0.03533 0.0217 0.6721 0.9237 0.9185 0.462 1.2051 0.9519 0.6397 1.0663 0.5564 0.7033 0.8047 0.8872 1.1532 1.2727 P16 LDGT1 86-89 0.09624 0.03427 0.4994 0.6974 0.6351 1.2145 0.5601 0.5254 1.017 0.3771 0.9662 0.7977 0.7267 0.8275 0.9401 1.0394 0.9564 1.2214 P17 LDGT1 86-89 0.09019 0.02973 0.3688 0.7491 0.5027 1.1374 0.7202 0.7519 0.9628 0.4449 0.9374 0.4114 0.7103 0.847 0.8964 1.087 0.896 1.1054 1.2517 P18 LDGT1 88-89 0.05194 0.01155 0.561 0.6874 0.7247 0.8944 0.4555 1.1666 0.7413 1.03 0.6717 0.4644 0.7014 0.8099 0.637 0.8736 0.8444 0.9947 0.7536 0.9298 P19 LDGT1 86-89 0.01448 0.0007 0.522 0.7051 0.7802 0.7612 0.7043 0.8674 0.7473 0.7107 0.6734 0.7097 0.7081 0.7257 0.7131 0.7176 0.7067 0.7115 0.6881 0.7158 0.7428 IM147 Phase 2 Regression Coefficients, NOx, 1986 to 1989 Model Year LDGTTs Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 ' C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT1 86-89 0.71447 0.56214 2.9488 B12 LDGT1 86-89 0.58562 0.33666 1.7796 4.5068 B13 LDGT1 86-89 0.29253 0.15935 1.0449 2.1599 3.1651 B14 LDGT1 66-89 0.22004 0.08924 0.9001 1.3657 2.486 1.9841 B15 LDGT1 86-89 0.14504 0.05 0.9776 1.164 1.1562 1.5604 1.7468 B16 LDGT1 86-89 0.13142 0.04594 1.0201 1.0708 1.2975 1.3785 1.3303 1.5302 B17 LDGT1 86-89 0.12255 0.04125 0.977 1.1131 1.218 1.4703 1.2356 1.4639 1.602 B18 LDGT1 86-89 0.0715 0.02293 0.937 1.1335 0.8403 1.2196 1.0929 1.5766 0.9306 1.2015 B19 LDGT1 86-89 0.01559 0.00168 0.9648 0.9858 0.9551 0,9715 0.9562 1.0147 0.9809 0.9807 0.9956 ------- 0 1 K> VO 1 V1147 Composite Regression Coefficients, NOx, 1990 to 1995 Model Year LDGH's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 Cl6 C17 C18 C19 P1 . LDGTl 90-95 1.15989 1.24128 32.9634 P2 LDGT1 90-95 0.72181 0.40146 0.679 4.5914 P3 LDGT1 90-95 0.69216 0.32475 2.1542 3.4912 6.7253 . i P.4. LDGTl 90-95 0.66749 0.3123 2.1371 3.4737 5.023 5.8563 P5 LDGTl 90-95 0.62455 0.33205 3.4043 2.3634 1.7239 5.0562 8.7446 . - P6 LDGTl 90-95 0.61978 0.33046 2.2722 2.3005 1.5599 2.9584 8.2913 6.1514 P7 LDGT1 90-95 0.58792 0.29749 2.0622 1.8473 0.0555 3.6238 4.2618 7.9809 5.3164 P8 LDGT1 90-95 0.55746 0.26129 5.9179 1.7445 -0.5662 1.235 4.3563 0.782 3.7446 7.1976 P9 LDGT1 90-95 0.54008 0.24653 4.1085 1.8313 -1.1591 2.0187 2.9278 -0.7348 3.4121 5.4986 7.8688 P.10 LDGT1 90-95 0.53603 0.2342 2.929 1.8441 -1.4993 2.6412 2.6552 0.0076 2.7177 5.458 6.0386 2.7176 P11 LDGT1 90-95 0.45579 0.05778 3.2806 1.051 0.2096 0.7143 2.8133 1.4291 0.9563 2.9071 6.0179 0.9428 1.4469 P12 LDGT1 90-95 0.36036 0.0273 4.3485 0.9563 -0.635 0.3902 2.43 0.4675 1.7444 0.5941 4.1616 1.3325 0.6814 3.9706 P.13 LDGT1 90-95 0.17758 0.04529 0.6553 0.7331 0.4798 1.1424 1.2661 0.9419 0.3362 0.8 1.2658 0.277 0.7426 1.2309 2.6602 P14 LDGTl 90-95 0.12741 0.02345 0.6061 0.7033 0.6424 1.2708 1.124 1.4638 0.5682 0.7251 0.9404 0.6888 0.6431 0.6253 1.6054 1.8253 P15 LDGT1 90-95 0.07822 0.01761 1.1782 0.6831 0.7996 1.3552 0.7806 0.9482 0.6863 0.8429 1.0081 0.3774 0.6949 0.7428 0.8665 1.0954 1.3651 P16 LDGT1 90-95 0.06969 0.01628 1.367 0.7128 0.824 1.0691 0.8406 0.6506 0.711 0.7051 0.914 0.6359 0.7085 0.724 0.8452 1.0797 1.0435 1.2099 PI 7 LDGT1 90-95 0.06569 0.00712 1.0171 0.713 0.7449 0.9912 0.8383 0.5662 0.7047 0.7856 0.8682 0.5184 0.7126 0.7606 0.8503 1.0453 1.0145 1.1224 1.218 P18 LDGT1 90-95 0.03909 0.0008 0.7599 0.7108 0.6532 1.2612 0.7148 1.122 0.5737 0.837 0.7408 0.6827 0.6897 0.8094 0.6626 0.8525 0.8302 1.1087 0.8134 0.8578 P19 LDGT1 90-95 0.01086 0.00079 0.757 0.7 0.7412 0.7991 0.7237 0.854 0.7206 0.6726 0.7413 0.7075 0.7135 0.7264 0.7129 0.729 0.7202 0.7122 0.7434 0.6941 0.73 IM147 Phase 2 Regression Coefficients, NOx, 1990 to 1995 Model Year LDGH's Regression Coefficients Segment dumber RMS Error Reg. . Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 cii C12 C13 C14 C15 C16 C17 C16 C19 B11 LDGT1 90-95 0.70725 0.11508 . 3.158 812 LDGT1 90-95 0.54083 0.02981 1.5474 5.8806 B13 LDGT1 90-95 0.24246 0.06072 0.9747 1.7935 3.6 B14 LDGT1 90-95 0.1753 0.03555 0.8927 1.2292 2.2351 2.4522 B15 LDGT1 90-95 0.10866 0.02903 0.9331 1.1278 1.1969 1.4839 1.8483 B16 LDGTl 90-95 0.09571 0.02747 0.9823 1.0076 1.1844 1.4428 1.4242 1.6458 B17 LDGTl 90-95 0.08918 0.01127 0.9705 1.0641 1.1573 1.415 1.3597 1.5734 1.7341 B18 LDGT1 90-95 0.05491 0.00454 0.9276 1.1839 0.9094 1.1571 1.0953 1.6597 1.2012 1.1201 B19 LDGT1 90-95 0.01259 0.0022 0.9623 0.9896 0.9651 0.9817 0.9732 0.9769 1.063 0.9406 0.9926 ------- n i U> O IM147 Composite Regression Coefficients, NOx, 1996 and Newer Model Year LDGTI's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT1 96+ 0.75596 0.48362 23.6975 P2 LDGT1 96+ 0.46673 0.18401 12.1233 3.0933 P3 LDGT1 96+ 0.46374 0.16288 12.749 2.6461 3.1462 P4 LDGT1 96+ 0.46514 0.16338 12.1028 2.534 2.3048 4.9422 P5 LDGT1 96+ 0.3791 0.14765 18.1804 1.5554 -1.8687 0.8812 12.6955 P6 L0GT1 96+ 0.38121 0.14804 18.193 1.5564 -1.9029 1.0072 12.8058 -0.5274 P7 LDGT1 96+ 0.3264 0.12947 19.6606 1.0854 -0.0898 -4.9504 5.5367 -2.3077 9.8773 P8 LDGT1 96+ 0.31962 0.14076 16.8826 0.9749 1.2061 -5.4908 2.7593 -4.5733 4.9464 8.5778 P9 LDGT1 96+ 0.32062 0.15138 18.5988 0.8971 1.8707 -4.2693 2.4773 -3.9369 4.9682 9.0399 -2.7492 P10 LDGT1 96+ 0.24391 0.13494 9.3292 0.7629 2.6156 -11.0979 1.3803 2.9882 1.7874 11.2329 -6.1993 6.6929 P11 LOGT1 96+ 0.18099 0.10785 0.7068 0.5528 3.4731 8.9995 -0.2732 4.7615 0.0782 7.2458 -2.3664 2.4504 1.28 P12 LDGT1 96+ 0.17159 0.09436 -0.6219 0.5365 2.5353 -3.0512 -0.8472 1.9087 -0.1596 6.9805 -2.7139 3.1761 0.9818 1.8863 P13. LDGT1 96+ 0.12899 0.07087 1.7268 0.3591 2.536 1.5628 0.5764 -2.7771 1.4626 2.9783 1.5186 1.4469 0.7662 1.0775 1.7081 PU LOGT1 96+ 0.10769 0.06233 -1.0403 0.4537 1.9505 0.0036 0.7496 1.6955 0.3857 3.6594 -0.8929 0.9085 0.7212 0.4779 1.0256 1.5265 P15 LDGT1 96+ 0.06499 0.02733 -0.039 0.6102 1.5778 0.9818 0.0528 0.1089 1.1166 2.0576 0.9366 1.0631 0.7396 0.86 0.6974 0.7671 1.3416 P16 LDGT1 96+ 0.05153 0.02163 0.4719 0.5741 2.1518 -0.4167 0.1408 -0.69 0.7472 2.1867 0.0185 0.9315 0.8149 0.7897 0.797 0.6681 0.6861 2.2482 P17 LOGT1 96+ 0.04794 0.01483 0.175 0.6024 1.5796 0.5006 0.714 -1.1578 0.8449 2.0981 0.6903 1.1098 0.7149 0.9709 0.8286 0.7074 0.7195 1.7387 1.6876 P18 LDGT1 96+ 0.02782 0.00211 0.6747 0.6416 1.5148 -0.1218 0.5752 0.4924 0.7534 1.8426 -0.2846 0.9695 0.7007 0.83 0.7552 0.7376 0.7664 1.1279 0.6032 0.758 P19 LDGT1 96+ 0.00403 0.00098 0.7432 0.7168 0.7163 0.6233 0.7292 1.0286 0.6325 0.9521 0.6402 0.7217 0.7125 0.7112 0.7066 0.7343 0.6734 0.7676 0.7274 0.7239 0.7481 IM147 Phase 2 Regression Coefficients, NOx, 1996 and Newer Model Year LDGTVs Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGTl 96+ 0.28558 0.12907 2.2171 B12 LDGT1 96+ 0.24779 0.09313 1.6406 3.4138 B13 LDGT1 96+ 0.18123 0.08816 1.252 1.6471 2.2975 B14 LDGTl 96+ 0.15739 0.06574 0.9727 1.3922 1.7725 1.7148 B15 LDGTl 96+ 0.09057 0.02458 1.0721 1.1839 0.9796 1.028 1.925 B16 LDGTl 96+ 0.07613 0.02379 1.1061 1.2115 1.0191 0.9869 1.1519 2.6649 B17 LDGTl 96+ 0.07139 0.00843 1.0352 1.3901 1.1173 0.956 1.1159 2.3796 2.0104 818 LDGT1 96+ 0.04333 0.00582 0.9841 1.2998 1.0418 0.9729 1.1585 1.4094 0.9684 1.1153 B19 LDGT1 96+ 0.00565 0.00178 0.9761 0.9749 0.9793 0.9737 0.9206 1.0687 0.9432 0.9717 1.0089 ------- IM147 Composite Regression Coefficients, NOx, 1981 to 1985 Model Year LDGT2's < Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 PI LDGT2 81-65 1.99804 3.69101 10.2569 P2 LDGT2 81-85 0.87409 0.91821 -7.9377 6.5305 P3 LDGT2 81-85 0.84122 0.83444 -8.613 5.4145 7.2247 P4 LDGT2 81-85 0.82892 0.89342 -10.2897 5.4633 2.5096 8.8657 P5 LDGT2 81-85 0.80333 0.89081 9.4068 4.7319 0.276 9.1558 5.3418 P6 LDGT2 81-85 0.7983 0.87735 -13.437 4.8675 0.2244 7.2171 4.8373 4.2644 P7 LDGT2 81-85 0.79972 0.89425 -13.7206 4.6923 -0.887 7.62 4.3577 4.6628 1.0456 P8 LDGT2 81-85 0.79473 0.84415 -11.4499 4.5789 -1.4261 5.7988 3.4916 2.7174 1.1653 3.2995 P9 LDGT2 81-85 0.79705 0.83069 -11.6615 4.6051 -1.4956 5.8933 3.2922 2.8161 1.1255 3.028 0.9261 i P10 LDGT2 81-85 0.79648 0.81111 -11.8131 4.561 -1.2986 5.9935 2.4713 3.4666 0.516 3.1827 -0.4726 2.7254 P11 LDGT2 81-85 0.6859 0.51142 -7.4827 3.0898 -1.5249 6.9352 3.5295 2.9385 1.4316 3.7346 1.3238 2.5498 1.2169 P12 LDGT2 81-85 0.54957 0.33422 -8.1307 2.4318 -0.7549 0.1755 1.141 5.7388 3.1487 2.2295 0.7669 0.0157 0.6425 3.1201 P13 LOGT2 81-85 0.28368 0.1926 -3.0729 0.495 1.2524 2.9861 -0.2928 2.9259 1.4715 0.5835 2.0376 -0.4876 0.712 1.7639 2.5466 P14 LDGT2 81-85 0.17079 0.09831 -1.9366 0.7267 0.5707 1.4814 -0.6512 2.8236 1.4133 1.5048 1.8051 0.7881 0.6265 1.0991 1.5732 1.6465 R15 LDGT2 81-85 0.13538 0.05816 -1.2593 0.6843 0.9631 1.699 -1.1762 2.6604 1.5116 1.2555 1.0619 1.1308 0.6644 0.9993 0.9615 1.3864 0.9139 P16 LDGT2 81-85 0.12927 0.04855 -0.6587 0.6411 0.9878 1.2005 -0.6445 2.468 1.4956 0.99 1.0154 1.2544 0.6922 0.8896 1.0162 1.349 0.717 0.6881 P17 LDGT2 81-85 0.12798 0.04485 -0.3609 0.6627 0.916 1.0272 -0.4442 2.2519 1.3213 1.1276 1.0486 1.1459 0.6823 0.8485 0.9897 1.3779 0.7091 0.682 0.6099 P18 LDGT2 81-85 0.08606 0.02155 0.3328 0.587 1.0724 1.0072 0.5742 1.9792 0.4674 1.4066 1.3569 -0.0211 0.6682 0.8649 0.7604 0.9339 0.6473 0.8806 1.1799 0.9731 P19 LDGT2 81-85 0.02488 0.0013 1.1429 0.6401 0.9555 0.8446 0.5989 0.8719 0.8311 0.6923 0.5888 0.7318 0.7118 0.6787 0.7213 0.7279 0.7365 0.6649 0.7072 0.679 0.7799 IM147 Phase 2 Regression Coefficients, NOx, 1981 to 1985 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 81-85 1.42447 1.55851 3.2137 B12 LDGT2 81-85 1.03629 0.77984 1.5673 6.4951 B13 LDGT2 81-85 0.38433 0.27378 0.8325 2.718 3.5162 814 LOGT2 81-85 0.26921 0.18815 0.8123 1.6716 2.4085 2.1998 B15 LDGT2 81-85 0.22469 0.11816 0.8718 1.4697 1.4685 1.8766 1.2618 B16 LDGT2 81-85 0.19939 0.09578 0.9679 1.128 1.6012 1.7461 0.8574 1.4532 017 LDGT2 81-85 0.19039 0.09229 0.9221 1.0819 1.4847 1.7994 0.8974 1.351 1.5644 B18 LDGT2 81-85 0.15301 0.07095 0.8347 1.2468 1.0994 1.3645 0.8676 1.6334 1.4952 1.0081 B19 LDGT2 81-85 0.03621 0.00936 . 0.943 0.9237 0.9993 0.9988 0.978 0.8541 1.0342 0.8947 1.1014 ------- 0 1 LO K) IM147 Composite Regression Coefficients, NOx, 1986 to 1987 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGT2 86-67 1.61179 3.13663 31.3453 P2 LDGT2 86-87 0.97795 1.10882 -3.0631 5.8632 P3 LDGT2 86-87 0.94502 0.9839 -5.0759 4.548 8.9192 P4 LOGT2 86-87 0.93501 0.93053 -3.3825 4.4591 5.2516 12.419 P5 LDGT2 86-87 0.85394 0.96285 -4.4203 3.6091 -1.7208 11.4179 9.1873 P6 LDGT2 86*87 0.84757 0.91259 -4.4095 3.6676 -1.5513 6.2687 8.3678 9.6507 P7 LDGT2 86-87 0.85029 0.91803 -4.0853 3.616 -2.1286 6.0176 7.6929 10.0876 0.7099 P8 LDGT2 86-87 0.85275 0.92194 -3.8044 3.5415 -2.4148 4.8872 7.7612 9.5053 0.6261 1.3726 P9 LDGT2 86-87 0.83178 0.86005 -3.778 4.3986 Ś6.3555 9.9437 4.6162 8.3808 -0.6741 -0.2851 9.9702 P10 LDGT2 86-87 0.80985 0.90764 -7.0196 4.1229 -5.9089 8.2497 2.6936 10.9511 -1.1815 0.3123 3.2549 6.8821 P11 LDGT2 86-87 0.76148 0.62189 -8.9566 3.094 -4.9268 6.2947 5.15 9.9243 -2.4765 -0.213 3.9969 6.4604 0.8832 P12 LDGT2 86-87 0.59003 0.3735 -2.4217 2.087 -2.6617 4.0616 2.224 1.9821 -1.3263 -2 4.7639 5.4191 0.2795 5.406 P13 LOGT2 86-87 0.33385 0.19007 0.2814 1.1386 0.5116 3.255 -0.1077 2.4058 0.5387 -1.7814 4.2622 -0.1798 0.5637 0.4606 3.2021 P14 LDGT2 86-87 0.20099 0.13811 1.8759 0.6935 0.5775 3.0569 0.9302 1.5194 0.7032 0.6974 2.2407 0.0121 0.4793 0.2025 1.9049 2.1646 P15 LDGT2 86-87 0.1036 0.0461 0.3765 0.6236 1.678 0.8805 1.0648 1.097 0.2424 -0.0264 1.7158 0.0601 0.6449 0.9194 1.0397 1.1878 1.3176 P16 LDGT2 86-87 0.09657 0.03819 0.6854 0.5604 1.7082 0.6146 0.6222 1.3221 0.5089 -0.0753 1.3097 0.8115 0.675 0.9002 0.9782 1.1823 1.0924 0.8604 P17 LDGT2 86-87 0.08202 0.02243 0.7341 0.6449 1.3141 -0.1109 0.6071 1.3461 0.4166 0.5635 1.6121 -0.0273 0.635 0.8997 0.9541 1.1481 1.0157 0.8336 1.7925 P18 LDGT2 86-87 0.05719 0.02206 0.9248 0.6705 1.2239 0.0494 0.6674 1.7312 0.3068 0.8156 1.3908 0.0199 0.6645 0.9323 0.8095 0.9223 0.7808 0.9324 1.241 . 0.7403 P19 LDGT2 86-87 0.01482 0.00786 0.6527 0.6745 0.6642 0.6799 0.7944 0.7706 0.6813 0.8969 0.6883 0.6733 on 0.7009 0.7385 0.7569 0.6679 0.7207 0.8026 0.7243 0.7208 IM147 Phase 2 Repression Coefficients, NOx, 1986 to 1987 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 68-87 1.44862 1.6677 2.507 B12 LDGT2 86-67 0.92821 0.61792 0.4543 9.6299 B13 LDGT2 66-87 0.45405 0.32299 0.78S2 0.5919 4.3588 Bl4 LDGT2 88-87 0.27745 0.22123 0.6178 0.2412 2.551 2.9603 B15 LDGT2 86-87 0.14699 0.0967 0.8032 1.2698 1.2876 1.6105 1.8179 016 LDGT2 86-87 0.1382 0.07565 0.8S9 1.1449 1.2732 1.5831 1.6009 0.9193 B17 LDGT2 86-87 0.1172 0.05283 0.8945 1.2281 1.135 1.5637 1.3545 1.3493 2.1364 B18 LDGT2 86-87 0.08824 0.05293 0.8887 1.3239 0.9624 1.2543 1.0381 1.5212 1.2998 0.9849 B19 LDGT2 86-87 0.0176 0.00876 0.958 0.9848 0.9632 0.9799 0.9647 0.9458 1.1217 0.9723 0.9846 ------- 0 1 U> U) IM147 Composite Repression Coefficients, NOx, 1988 to 1995 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 Pi LDGT2 88-95 1.24878 2.35108 23.5648 P2 L0GT2 88*95 0.82176 0.87411 0.1479 4.6355 P3 LDGT2 88-95 0.77845 0.71657 -1.7162 3.425 8.1364 P4 LDGT2 88-95 0.7742 0.70093 -1.8478 3.4367 6.2476 5.1308 . P5 LDGT2 88-95 0.73556 0.71294 -1.4757 2.6692 3.7613 3.2899 6.2663 P6 LDGT2 88-95 0.72738 0.72398 -2.5889 2.6139 3.8661 -0.5835 5.6306 6.8957 P7 LDGT2 88-95 0.70212 0.63963 -1.6939 2.0359 3.1437 0.5064 1.7183 8.4446 5.0667 P8 LDGT2 88-95 0.68022 0.60531 -0.4848 1.8702 2.347 -1.174 1.1716 2.9882 4.0341 6.3227 P9 LDGT2 88-95 0.67704 0.59719 -1.6152 1.9533 1.922 0.993 0.5073 2.3082 3.7294 5.9607 3.1202 . f P10 L0GT2 88-95 0.66525 0.59249 -1.2143 1.9942 1.1629 0.3777 0.1791 2.4339 2.4689 6.3272 -0.1222 4.5356 P11 LDGT2 88-95 0.49566 0.16456 -1.995 0.9652 1.6003 1.6143 0.2178 0.8508 -0.2399 3.1105 1.8601 2.7668 1.9044 P12 L0GT2 88-95 0.39011 0.13915 -0.5048 0.8942 0.6646 0.7011 0.5649 0.3036 0.9763 0.3149 0.4631 3.3575 0.9895 3.9381 P13 LDGT2 88-95 0.22071 0.13586 0.1295 0.5892 1.2326 1.6722 0.4435 0.7176 0.4615 0.651 1.5772 0.5428 0.7152 1.6198 2.5117 P14 L0GT2 88-95 0.16814 0.08085 0.4177 0.6242 1.2076 1.54 0.4306 1.0374 0.8469 0.7753 1.5915 0.5229 0.5567 1.0483 1.735 1.6545 P15 LDGT2 88-95 0.09792 0.04274 1.2914 0.6299 0.932 1.5301 0.7496 0.6925 0.7213 0.5876 .1.3142 0.4284 0.6686 0.9379 0.8221 1.0317 1.4681 P16 LDGT2 88-95 0.08515 0.04 1.3815 0.6274 0.8992 1.2443 0.777 0.5902 0.9112 0.3659 1.1635 0.6339 0.6996 0.6562 0.8431 0.9855 1.1161 1.2384 P17 LDGT2 88-95 0.0793 0.03 1.1604 0.6609 0.7332 1.166 0.8211 0.5607 0.8664 0.4764 1.1288 0.4527 0.7155 0.6591 0.8406 0.9915 1.0422 1.1105 1.2867 P18 L0GT2 88-95 0.04482 0.00528 1.0205 0.6891 0.8472 0.9762 0.6666 1.0589 0.5961 0.8411 0.7658 0.5158 0.703 0.6891 0.6603 0.6286 0.8216 1.0348 0.9346 0.8792 P19 LDGT2 88-95 0.0136 -0.00252 0.7213 0.711 0.747 0.8258 0.7089 0.7055 0.696 0.744 0.7329 0.7233 0.715 0.7382 0.7138 0.716 0.7117 0.7363 0.7333 0.6995 0.7243 IM147 Phase 2 Regression Coefficients, NOx, 1988 to 1995 Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 68-95 0.71642 0.34014 3.1658 B12 LDGT2 88-95 0.55127 0.22451 1.6945 5.1988 013 LDGT2 88-95 0.30245 0.20374 0.8854 2.4317 3.3283 B14 LDGT2 88-95 0.23377 0.14045 0.7529 1.6901 2.356 2.1936 B15 LDGT2 88-95 0.13663 0.07379 :. 0.881 1.4099 1.0647 1.3754 2.0165 B16 LDGT2 88-95 0.11912 0.06757 0.9498 1.1559 1.1515 1.31 1.5401 1.6409 B17 LDGT2 88-95 0.10918 0.04812 0.9645 1.1668 1.1296 1.3343 1.4075 1.494 1.8722 Q16 LDGT2 88-95 0.0635 0.01509 0.9291 1.3099 0.8516 1.1259 1.093 1.5612 1.2713 1.1432 019 LDGT2 88-95 0.01423 0.0016 0.9872 1.0052 0.9647 0.96 0.9599 1.0009 1.0227 0.9483 Ś1.0006 ------- n t U> 4^ IM147 Composite Regression Coefficients, NOx, 1996 and Newer Model Year LDGT2's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 013 014 C15 C16 C17 C18 C19 Pi LDGT2 96+ 0.578 0.71379 15.4654 P2 LDGT2 96+ 0.33821 0.37061 -21.218 3.256 P3 LDGT2 96+ 0.33307 0.33021 -17.4893 2.5344 3.6906 P4 LDGT2 96+ 0.30416 0.30038 -14.5682 2.2242 11.6872 -12.2279 i P5 LDGT2 96+ 0.29072 0.28842 -17.6088 1.8185 9.3908 -10.8536 4.734 P6 LDGT2 96+ 0.2883 0.30337 -11.6965 1.745 9.6903 -10.7171 5.8457 -7.1022 P7 LDGT2 96+ 0.24718 0.25136 -20.0849 1.7221 5.3396 -2.4287 -0.5675 -9.9875 8.4455 t P8 LDGT2 96+ 0.23492 0.2519 -22.4763 1.5623 5.2794 -3.3995 -2.4191 -8.9711 6.6514 7.9536 P9 LDGT2 96+ 0.23606 0.24316 -22.27 1.5055 5.3167 -3.4781 -3.0309 -9.6937 6.8514 8.1317 3.1882 < P10 LOGT2 96+ 0.2359 0.23077 . -20.9734 1.5584 5.6716 -3.4051 4.2894 -8.7832 6.8235 6.3126 1.8606 3.9221 N P11 LDGT2 96+ 0.18371 0.16444 -10.5542 0.5675 3.1636 -1.1457 -3.3023 0.5825 3.4287 3.1947 4.4137 2.9738 1.3085 P12 LDGT2 96+ 0.18578 0.16354 9.9999 0.6169 2.8046 -0.7267 3.2312 0.1935 3.5982 3.0439 4.2462 2.9303 1.2191 0.2662 P13 LDGT2 96+ 0.12127 0.05817 -4.2485 0.6116 4.4173 -3.0746 -1.9566 -5.1095 2.1593 3.236 5.8951 2.3221 1.0537 -0.1452 1.6328 P14 LOGT2 96+ 0.06953 0.0183 4.3668 0.7373 2.7372 0.575 -1.3744 -3.22 2.644 -0.5548 1.4212 3.1253 0.6404 1.0543 0.791 1.4102 P15 LDGT2 96+ 0.05854 0.01128 2.8175 0.7837 1.5604 1.7698 -0.6221 -2.7325 2.4484 0.2815 0.0066 2.3246 0.5592 1.0307 0.7388 1.1608 0.7986 P16 LDGT2 96+ 0.05584 0.00873 1.9819 0.7799 1.6432 1.5678 0.4135 -2.4395 2.0557 0.0994 0.2982 1.9349 0.6182 0.9336 0.7441 1.0662 0.7148 1.3475 P17 LDGT2 96+ 0.05251 0.00847 0.351 0.7816 1.4739 1.5467 -0.3431 -2.0093 2.0146 0.208 0.4812 1.4451 0.723 0.6611 0.7379 1.0331 0.6538 1.3684 2.1742 P18 LDGT2 96+ 0.01918 0.00084 1.8364 0.7759 1.0399 1.1145 0.2474 -0.0511 0.6696 0.6814 0.6723 1.4583 0.5956 0.9049 0.7647 0.6482 0.6884 1.3503 -0.5973 0.8893 P19 LDGT2 96+ 0.00325 0.00226 0.7097 0.7313 0.8044 0.6612 0.6632 0.833 0.6604 0.6611 0.5647 0.8016 0.72 0.6732 0.7118 0.7105 0.7272 0.7857 0.7169 0.7106 0.6983 IM147 Phase 2 Regression Coefficients, NOx, 1996 and Newer Model Year LDGT2's Regression Coefficients 1 Segment Number RMS Error Reg. Constant CI * C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 : C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGT2 96+ 0.25624 0.28161 2.2649 B12 LDGT2 96+ 0.25427 0.26879 1.8738 1.3389 B13 LDGT2 96+ 0.18429 0.10134 1.6678 0.8038 2.0982 B14 LDGT2 96+ 0.11213 0.05754 1.2535 1.3129 1,0561 1.9394 B15 LDGT2 96+ 0.08972 0.02934 1.1523 0.9849 1.0364 1.4467 1.3152 B16 LDGT2 96+ 0.08172 0.02161 1.1412 0.9687 1.0064 1.2969 1.0936 2.5063 B17 LDGT2 96+ 0.07624 0.01744 1.1996 0.6743 0.9948 1.295 1.0153 2.0425 3.1694 B18 LDGT2 96+ 0.03345 0.0095 0.929 1.0616 1.0355 1.1034 0.9077 1.821 -0.1023 1.2122 B19 LDGT2 96+ 0.00415 0.0038 0.9821 0.9175 0.9698 0.9467 0.9808 1.0103 1.1121 0.9437 0.9983 ------- 0 1 U> Lh IM147 Composite Regression Coefficients, NOx, 1981 to 1982 Model Year LDGV's Repression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LDGV B1-82 0.91799 1.65041 42.7692 P2. LDGV 81-82 0.63601 0.72708 17.5251 4.1787 P3. LDGV 81-82 0.52331 0.58513 -4.6789 2.4424 11.7817 P4 LDGV 81-82 0.52237 0.58386 -6.215 2.4752 10.4671 3.1134 P5 LDGV 81-82 0.50775 0.50405 -4.2723 2.2111 9.2772 2.5761 3.9991 P6 LDGV 81-82 0.50352 0.48347 -3.9075 2.1456 8.9952 1.1125 3.4211 5.3989 P7 LDGV 81-82 0.49839 0.48687 -4.3121 2.0136 6.6708 0.9449 1.0442 5.8614 2.4035 P8 LDGV 81-82 0.46669 0.46683 1.6928 1.6175 7.9174 -2.5315 1.6902 0.2702 0.6553 7.5772 P9 LDGV 81-82 0.46311 0.452 -0.6119 1.7211 7.5725 -2.1824 0.4138 -2.0068 0.7582 7.4243 3.9487 P10 LDGV 81-82 0.46376 0.44782 -1.0048 1.7433 7.4758 -2.0726 0.0943 -1.7963 0.4584 7.4711 3.3527 1.0886 P11 LDGV 81-82 0.35194 0.23815 -6.5826 0.7668 4.8431 -0.79 0.1979 2.4979 0.684 4.4696 3.7564 0.6973 1.4384 P12 LDGV 81-82 0.2927 0.17049 -9.8733 0.9681 2.776 2.0862 0.4024 1.4025 2.1207 3.7386 2.3379 1.5715 0.6911 2.5825 P13 LDGV 81-62 0.18109 0.07569 -4.5202 0.7632 1.5688 1.3953 0.0498 2.8038 -0.6444 2.0452 0.4039 1.7917 0.6671 1.3367 2.4626 Pt4 LDGV 81-82 0.14099 0.02431 -4.1674 0.7506 1.6956 1.9641 0.1583 1.4761 -0.6375 1.7248 -0.4851 2.7419 0.5236 0.9827 1.8418 1.5633 P15 LDGV 81-82 0.08737 0.01811 -1.382 0.8401 1.3947 1.6138 -0.2041 1.0019 0.7778 0.8836 1.2699 0.6037 0.6393 0.8763 0.6065 0.6832 1.7403 P16 LDGV 81-82 0.07312 0.0041 -1.1445 0.8151 1.058 1.5224 0.0582 0.8216 1.2094 0.5747 1.1231 0.8514 0.7173 0.8313 0.6082 0.7402 1.2902 1.5334 P17 LDGV 81-82 0.07264 0.00367 -0.9811 0.8032 1.0413 1.4118 0.1444 0.648 1.1681 0.6401 1.1443 0.7002 0.7202 0.816 0.6414 0.7463 1.2653 1.4284 0.5328 P18 LDGV 81-82 0.04243 0.00563 0.017 0.7339 0.8526 1.377 0.2716 1.0442 0.6919 0.8178 0.9258 0.5464 0.7223 0.7683 0.6426 0.8313 0.6107 1.4947 0.4033 0.8387 P19 LDGV 81-82 0.01342 0.00395 0.7869 0.7185 0.788 0.8447 0.6524 0.7441 0.6472 0.857 0.6045 0.7562 0.6997 0.7449 0.6929 0.7279 0.7185 0.7737 0.6197 0.7165 0.6907 IM147 Phase 2 Regression Coefficients, NOx, 1981 to 1982 Model Year LDGV's 1 Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 81-62 0.59924 0.60079 2.8438 B12 LDGV 81-62 0.50466 0.48014 1.6 3.6643 B13 LOGV 81-62 0.2646 0.11652 0.8595 2.316 3.4315 B14 LDGV 81-82 0.21268 0.04456 0.7104 1.676 2.6006 2.1072 B15 LDGV 81-82 0.13034 0.02244 0.8971 1.4034 0.8839 1.2572 2.3849 B16 LDGV 81-62 0.1083 0.01418 1.0311 1.1362 1.0575 0.9885 1.7362 2.0954 B17 LDGV 81-62 0.10478 0.0082 1.0248 1.1015 1.1113 1.0374 1.6264 1.8908 1.4012 B18 LDGV 81-82 0.06108 0.00004 0.9845 1.139 0.8671 1.145 0.9874 2.2913 0.654 1.1564 819 LDGV 81-82 0.01536 0.00328 0.9626 1.0202 0.9556 0.9706 0.9459 1.0635 0.9025 0.9816 0.9686 ------- 0 1 u> On IM147 Composite Repression Coefficients, NOx, 1983 to 1985 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 PI LDGV 83-85 0.96463 1.86521 18.7418 P2 LDGV 83-85 0.58177 0.74711 0.782 4.5182 P3 LDGV 83-85 0.57164 0.69924 1.7395 3.8215 3.9404 P4 LOGV 83-85 0.56866 0.69182 2.2485 3.7307 2.712 4.1135 Śr P5 LDGV 83-85 0.54912 0.68329 0.0309 3.6538 -0.5552 2.7207 4.4539 P6 LDGV 83-85 0.54056 0.65766 -0.2605 3.6825 0.6893 0.2755 3.3123 6.9261 P7 LDGV 8345 0.52478 0.64771 -1.2424 3.0961 -0.7131 -0.3096 1.2743 8.0479 3.3901 P8 LDGV 83-85 0.49039 0.64355 -0.2025 2.5593 -0.8873 -2.2555 0.9327 0.9917 2.648 6.9554 P9 LDGV 83-85 0.49081 0.6366 -0.2479 2.5692 -0.9586 -2.1078 0.7709 0.6195 2.6479 6.804 0.8258 P10 LDGV 83-65 0.48198 0.62502 -0.6852 2.5126 -1.7481 0.7564 0.3327 1.3097 1.8572 6.9911 -1.6919 3.7996 P11 LDGV 83-85 0.43238 0.38649 0.6471 1.837 -1.1074 0.237 1.3594 1.4974 0.1122 6.1798 0.33 1.9052 1.0346 P12 LDGV 83-85 0.35372 0.24907 1.0365 1.599 -1.5529 1.0464 1.5953 -0.2612 1.2403 3.835 0.566 1.7663 0.4528 3.3405 P13 LDGV 83-65 0.22239 0.11604 1.4435 0.8789 . 1.1138 3.1005 0.6475 1.7479 0.451 1.0221 0.5773 0.455 0.4514 1.7854 2.2469 P14 LDGV 83-65 0.16685 0.04792 1.1199 0.8083 0.993 2.3512 0.3301 2.1868 0.9638 0.6226 1.2885 1.0545 0.5328 0.7862 1.2774 2.0142 P15 LDGV 83-85 0.09344 0.00442 0.6972 0.724 1.0561 1.1547 0.647 1.6301 0.8766 0.4133 1.7045 0.185 0.7132 0.6701 0.6719 1.0248 1.5987 P16 LDGV 83-85 0.07618 0.01155 1.4799 0.6974 0.9307 0.8525 0.7603 0.9115 1.0175 0.3596 1.3353 0.4874 0.7501 0.6536 0.7419 0.6969 1.1713 1.4857 P17 LDGV 83-85 0.07136 0.00873 1.0722 0.7157 0.74 1.0551 0.6004 0.9688 0.8579 0.6022 0.9517 0.6109 0.7611 0.5992 0.7747 0.9329 1.1217 1.2072 1.6127 P18 LDGV 83-85 0.03952 0.0055 0.6259 0.6692 0.8058 1.0136 0.6208 1.1765 0.512 0.9231 0.7662 0.5965 0.7047 0.7879 0.6504 0.8038 0.8335 1.1663 1.2587 0.9322 P19 LDGV 83-85 0.01224 -0.00128 0.8422 0.7007 0.7698 0.724 0.7372 0.8013 0.7199 0.7142 0.6943 0.7149 0.7141 0.7285 0.7042 0.7381 0.7145 0.6689 0.7795 0.691 0.7508 IM147 Phase 2 Regression Coefficients, NOx, 1983 to 1985 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant 01 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C16 C19 B11 LDGV 83-85 0.73374 0.81776 .2.4382 B12 LDGV 83-85 0.54834 0.39309 1.1577 5.767 B13 LDGV 83-85 0.31629 0.20822 0.6375 2.8068 3.1803 814 LDGV 83-85 0.24412 0.1176 0.8023 1.3747 1.9218 2.6998 BIS LDGV 83-85 0.13375 0.03665 0.9663 1.0509 0.9033 1.3245 2.2612 B16 LDGV 83-85 0.10708 0.03554 1.0501 0.6984 1.0487 1.1265 1.6402 2.0329 B17 LDGV 83-85 0.09639 0.02013 1.0496 0.8104 1.0655 1.2345 1.5252 1.6719 2.3989 B18 LDGV 83-85 0.05614 0.0163 0.9253 1.1246 0.8814 1.096 1.1113 1.8248 1.4559 1.1912 B19 LDGV 83-85 0.01369 0.0022 0.9666 0.9844 0.9642 0.9894 0.982 0.9116 1.1138 0.9386 1.0241 ------- 0 1 U) -J IM147 Composite Repression Coefficients, NOx, 1986 to 1989 Model Year LDGV's Repression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 P1 LOGV 86-89 0.78565 1.40846 22.9983 P2 LDGV 86-89 0.57844 0.77237 9.21 3.0925 P3 LDGV 66-89 0.53576 0.63136 9.1703 2.0307 7.0173 P4 LDGV 86-89 0.52985 0.61276 9.5621 2.0154 5.1612 5.5354 P5 LDGV 86-89 0.48733 0.58411 7.6422 1.5437 3.1981 4.1123 5.8762 P6 LDGV 86-89 0.469 0.58 4.0676 1.525 2.9935 0.4973 4.8986 10.1283 P7 LDGV 86-89 0.44452 0.51948 2.5722 1.2222 2.2925 1.1125 2.3363 11.7286 3.4574 P8 LDGV 86-89 0.42664 0.46953 3.5526 1.2073 1.7483 0.3084 2.0313 6.7259 2.6637 5.3318 P9 LDGV 86-89 0.42322 0.45426 3.3018 1.301 1.139 0.1369 1.5033 5.9904 2.5273 4.6613 3.2834 P10 LDGV 86-69 0.42046 0.44434 2.7682 1.279 0.8669 0.8265 1.1624 5.9839 1.9554 4.9824 2.1016 2.3949 P.11 LDGV 86-89 0.3165 0.16273 1.6969 0.9234 0.2926 1.5843 1.5047 5.7857 0.6453 3.2517 2.4719 1.4771 1.3713 P.12 LDGV 86-89 0.28211 0.13132 0.4472 0.9805 -0.3268 1.5058 1.3641 3.9271 1.2167 1.5887 1.8636 1.9637 0.9702 2.2297 PI 3 LDGV 66-89 0.17181 0.05351 1.5839 0.7587 0.7909 1.2166 0.7145 1.5691 0.9527 1.0643 1.0876 0.6012 0.7526 1.15 2.3208 P14 LDGV 86-89 0.13065 0.04173 0.3316 0.7154 1.0709 0.6518 0.7486 1.2646 0.8718 0.8749 0.8522 0.9529 0.6065 0.9816 1.5163 1.6439 P15 LDGV 86-89 0.08069 0.01822 0.1209 0.6879 0.733 1.0408 0.7212 1.5437 0.8545 0.957 0.2801 0.7471 0.6975 0.8112 0.8905 1.0278 1.3107 P16 LDGV 86-89 0.06908 0.01317 0.5535 0.7013 0.7623 0.918 0.6392 1.068 0.9735 0.7864 0.2568 0.9626 0.7416 0.7472 0.8532 0.8939 1.013 1.467 P17 LDGV 86-89 0.06367 0.01125 0.3901 0.7232 0.6166 0.8153 0.6598 1.0147 0.8687 0.8487 0.2077 0.8136 0.7444 0.7667 0.847 0.9405 0.9462 1.3241 1.4006 P18 LDGV 66-89 0.03823 0.00394 0.7351 0.6866 0.7439 1.0847 0.6838 1.2044 0.6292 0.8002 0.5291 0.672 0.6852 0.8818 0.7191 0.82 0.7757 1.2183 0.9059 0.8386 P19 LDGV 66-89 0.01005 0.00003 0.7593 0.7055 0.7238 0.7901 0.6814 0.9355 0.7036 0.7211 0.6966 0.7005 0.7134 0.7342 0.7231 0.7021 0.7072 0.7254 0.7597 0.7156 0.7326 IM147 Phase 2 Regression Coefficients, NOx, 1986 to 1989 Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant CI C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 86-69 0.52925 0.42098 2.5542 B12 LDGV 86-69 0.43224 0.27635 1.6545 3.9022 B13 LDGV 86-89 0.23852 0.0957 1.0646 1.7542 3.3487 B14 LDGV 86-89 0.16118 0.07257 0.867 1.4108 2.1959 2.261 B15 LDGV 66-89 0.11203 0.02849 0.9505 1.1788 1.2539 1.4295 1.7812 B16 LDGV- 66-49 0.0963 0.02351 1.0428 0.9898 1.2047 1.2246 1.4141 1.9452 B17 LDGV 86-89 0.08734 0.01558 1.0275 1.0114 1.167 1.2958 1.2756 1.7751 1.9705 B18 LDGV 86-89 0.05354 0.0073 0.9133 1.2514 0.967 1.1377 1.0194 1.8001 1.2499 1.0893 B19 LDGV 86-89 0.01153 0.00104 0.964 0.9967 0.973 0.9592 0.9531 0.9973 1.0544 0.9572 1.0152 ------- 0 1 U> 00 IM147 Composite Repression Coefficients, NOx, 1990 to 1995 Model Year LDGV's Reqression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 ' C12 C13 C14 C15 016 C17 C18 C19 P1 LDGV 90-95 0.85265 0.98234 27.2602 P2 LDGV 90-95 0.55676 0.37911 3.4685 3.9198 . ' ii. P3 LDGV 90-95 0.5187 0.31019 2.4287 2.702 6.9157 P4 LDGV 90-95 0.51364 0.29256 2.566 2.669 5.4346 4.5269 P5 LDGV 90-95 0.47026 0.27475 0.7261 2.0325 3.4764 2.5823 7.1639 P6 LDGV 90-95 0.46835 0.27776 0.0795 2.0153 3.3673 1.5718 6.6904 3.6255 P7 LDGV 90-95 0.43598 0.24653 -0.2787 1.6476 1.7967 2.607 2.986 4.5264 4.7362 P8 LDGV 90-95 0.42397 0.22613 0.8668 1.576 1.4199 1.0315 2.7698 1.4176 3.8634 4.4221 P9 LDGV 90-95 0.41414 0.2177S 0.4251 1.6155 0.9945 0.7619 2.1999 0.108 3.4947 3.5743 5.7316 ť P10 LDGV 90-95 0.40712 0.2139 -0.4251 1.6685 0.371 1.1081 1.7587 0.4377 2.6567 3.7398 4.0116 3.1506 P11 LDGV 90-95 0.32541 0.07608 Q.4951 1.01 0.6419 0.3247 2.3515 0.3887 1.0667 2.0267 3.9523 1.1292 1.4038 P12 LDGV 90-95 0.27382 0.05866 1.1507 0.9457 0.7041 0.1603 1.9823 -0.7329 1.4374 0.757 2.5555 1.2857 0.8276 3.2666 P13 LDGV 90-95 0.1552 0.0465 1.245 0.7538 1.2793 0.8753 0.6342 0.2174 0.6015 0.6589 2.2733 -0.1902 0.7277 1.3434 2.4018 P14 LDGV 90-95 0.11701 0.02975 1.1371 0.7436 0.97 0.8665 0.7991 0.5198 0.7343 0.5517 1.7367 0,4654 0.6168 1.0241 1.4687 1.7969 P15 LDGV 90-95 0.07219 0.00927 0.8725 0.7153 0.8392 0.9635 0.7945 0.6765 0.7708 0.6798 1.0226 0.5631 0.7013 0.6864 0.8237 1.0439 1.3295 P16 LDGV 90-95 0.06398 0.00808 1.0798 0.7045 0.8792 0.8433 0.7229 0.5834 0.8931 0.5466 0.9541 0.7035 0.7254 0.6265 0.8134 0.9715 1.0589 1.191 P17 LDGV 90-95 0.05809 0.00266 0.7561 0.7292 0.6974 0.8682 0.8458 0.6179 0.6355 0.681 0.8312 0.5502 0.7239 0.822 0.6206 0.9771 0.9638 1.1024 1.5163 P18 LDGV 90-95 0.03251 0.00233 0.6937 0.7144 0.6856 0.9367 0.7529 0.9301 0.619 0.8318 0.7798 0.5858 0.6898 0.6309 0.7014 0.849 0.7786 1.1002 0.83 0.8865 P19 LDGV 90-95 0.00903 0.00112 0.8195 0.7065 0.733 0.7558 0.7072 0.8124 0.7212 0.7331 0.7165 0.7193 0.7121 0.7215 0.7187 0.7075 0.7149 0.7316 0.7374 0.7123 0.7172 IM147 Phase 2 Regression Coefficients, NOx, 1990 to 1995 Model Year LDGV's Reqression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 C8 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C16 C19 B11 LDGV 90-95 0.4955 0.16783 2.7648 B12 LDGV 90-95 0.39806 0.10265 1.5279 4.9874 B13 LDGV 90-95 0.21598 0.0764 0.9775 2.036 3.2274 B14 LDGV 90-95 0.16191 0.04667 0.8655 1.4947 2.0303 2.4727 B15 LDGV 90-95 0.09916 0.01709 0.9685 1.262 1.1221 1.4177 1.8321 816 LDGV 90-95 0.08828 0.01425 1.0137 1.1226 1.1413 1.3001 1.4762 1.565 B17 LDGV 90-95 0.07957 0.00619 0.9986 1.1256 1.132 1.322 1.3198 1.5162 2.0499 B18 LDGV 90-95 0.04511 0.00552 0.9273 1.1868 0.9219 1.1679 1.0538 1.5928 1.0608 1.1805 B19 LDGV 90-95 0.01076 0.00217 0.9663 0.9854 0.9735 0.9568 0.9719 1.0002 1.022 0.9682 0.9807 ------- 0 1 u> VO IM 147 Composite Regression Coefficients, NOx, 1996 and Newer Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 C5 . C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 PI LDGV 96+ 0.29865 0.37011 11.7738 P2 LDGV 96+ 0.22677 0.24297 -3.0327 2.4026 P3 LDGV 96+ 0.21591 0.222 0.4545 1.4421 5.448 P4 LDGV 96+ 0.21456 0.2173 -0.7302 1.4606 4.3233 3.1071 P5 LDGV 96+ 0.20702 0.21364 1.0698 1.2242 3.8494 -0.6109 4.3718 P6 LDGV 96+ 0.20746 0.2141 0.6213 1.2154 3.841 0.6842 4.2145 0.9383 P7 LDGV 96+ 0.20382 0.20285 2.5242 1.0207 3.6971 -1.4265 3.1047 0.3945 2.683 P8 LDGV 96+ 0.19807 0.19482 4.1563 0.9074 3.4911 -2.0343 2.3803 -0.0536 2.1013 3.3361 P9 LDGV 96+ 0.19619 0.19096 1.8739 1.0364 2.2721 -1.119 2.141 -0.5391 2.1297 2.5845 3.318 P10 LDGV 98+ 0.1935 0.18673 1.493 0.998 2.3525 -1.4944 2.0734 0.242 1.6938 2.8999 1.2192 1.9845 P11 LDGV 96+ 0.15206 0.11563 6.6686 0.3811 2.506 -1.1863 0.7057 -0.2193 2.4209 0.8267 1.5013 0.5498 1.1615 P12 LDGV 96+ 0.14242 0.09025 7.3264 0.4696 2.0937 -0.2105 0.3797 -0.6436 2.1032 0.8737 0.8521 0.6422 0.984 1.8962 P13 LDGV 96+ 0.1117 0.0621 4.8271 0.6696 2.0895 -2.2591 0.9164 Ś0.565 1.4492 -0.1321 0.1274 0.7773 0.8735 0.8257 2.0634 P14 LDGV 96+ 0.0781 0.04141 3.4134 0.5916 1.3165 -0.2995 1.3675 -0.2069 1.4165 0.0989 0.2565 0.3993 0.7549 0.6841 1.288 1.631 P15 LDGV 96+ 0.05148 0.02187 2.4596 0.6558 1.17 0.6931 0.7565 0.2435 0.7215 0.5341 0.5256 0.6843 0.7209 0.6775 1.1075 0.9069 1.122 P16 LDGV 96+ 0.04778 0.01987 3.1395 0.6091 1.203 0.5425 0.9077 0.4201 0.7848 0.5764 0.1403 0.7957 0.7431 0.6929 1.0516 0.9276 0.7894 1.4029 P17 LDGV 96+ 0.04525 0.01326 3.3943 0.6187 0.8612 0.7678 1.1673 0.4078 0.7682 0.5486 0.2441 0.6409 0.7404 0.726 0.935 1.0264 0.7278 1.3847 1.368 P18 LDGV 96+ 0.02353 0.00022 1.5178 0.6979 0.7651 0.9288 0.6991 0.3426 0.7996 0.6796 0.6094 0.5847 0.7147 0.8429 0.6623 0.895 0.6746 0.9795 0.7813 1.0666 P19 LDGV 96+ 0.004 0.00038 0.7646 0.7189 0.7176 0.7276 0.6851 0.7511 0.7279 0.6879 0.7681 0.6817 0.7174 0.7317 0.7291 0.7203 0.7125 0.7336 0.7584 0.7113 0.7181 I M147 Phase 2 Repression Coefficients, NOx, 1996 and Newer Model Year LDGV's Regression Coefficients Segment Number RMS Error Reg. Constant C1 C2 C3 C4 . C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 B11 LDGV 96+ 0.20953 0.18322 1.6724 B12 LDGV 96+ 0.19448 0.14434 1.3687 2.7382 B13 LDGV 96+ 0.155 0.09509 1.1262 1.2948 2.6037 B14 LDGV 96+ 0.10659 0.06065 0.9665 0.9898 1.6977 2.2378 B15 LDGV 98+ 0.06912 0.03298 0.9526 0.6932 1.4872 1.252 1.5305 816 LDGV 96+ 0.0649 0.03049 0.9772 0.9069 1.4267 1.2614 1.1325 1.7334 B17 LDGV 96+ 0.06203 0.02204 0.9694 0.9325 1.3118 1.3384 1.0793 1.6717 1.6273 B18 LDGV 96+ 0.03172 0.00214 0.9508 1.1071 0.9089 1.1952 0.9329 1.2914 1.0022 1.4675 B19 LDGV 96+ 0.00523 0.00111 0.9672 0.6874 0.9855 0.0701 0.8675 0.0838 1.0369 0.6689 0.9757 ------- Appendix D Excess Emissions Identification for Max CO Cutpoints ------- Excess Emissions with Fast-Pass, Retest, Fast-Fail Enabled Max CO Outpoints (Data Not Normalized for Model Year Distribution) Excess Excess HC Excess Excess CO Excess Excess NO Vehicle Model Year IM240 HC with fast- Excess HC IM240 CO with fast- Excess CO IM240 NOx with fast- Excess NOx Class Range (grams) pass (grams) Identified (grams) pass (grams) Identified (grams) pass (grams) Identified LDGV 81-82 0 - 0 N/A 56.74 49.73 87.6% 3.61 3.61 100.0% 83-85 6.13 6.13 100.0% 243.24 237.02 97.4% 3.49 3.49 100.0% 86-89 10.72 10.65 99.3% 301.8 294.27 97.5% 9.92 9.92 100.0% 90-95 8.96 8.68 96.9% 185.26 157.96 85.3% 6.59 6.02 91.4% 96+ 0 0 N/A 0 0 N/A 0 0 N/A ALL 25.81 25.46 98.6% 787.04 738.98 93.9% 23.61 23.04 97.6% LDGT1 81-85 10.99 8.32 75.7% 605.55 590.71 97.5% 19.02 18.72 98.4% 88-89 5.55 5.13 92.4% 533.37 485.61 91.0% 2.37 2.36 99.6% 90-95 0.42 0 0.0% 1.77 0 0.0% 1.65 1.65 100.0% 96+ 0 0 N/A 0 0 N/A 0.43 0.43 100.0% ALL 16.96 13.45 79.3% 1140.69 1076.32 94.4% 23.47 23.16 98.7% LDGT2 81-85 15.62 15.62 100.0% 441.44 441.44 100.0% 6.69 6.69 100.0% 86-87 10.39 10.39 100.0% 285.54 285.54 100.0% 4.11 2.91 70.8% 88-95 0.21 0.21 100.0% 26.15 15.16 58.0% 0.68 0.27 39.7% 96+ 0 0 N/A 0 0 N/A 0 0 N/A ALL 26.22 26.22 100.0% 753.13 742.14 98.5% 11.48 9.87 86.0% Total ALL 68.99 65.13 94.4% 2680.86 2557.44 95.4% 58.56 56.07 95.7% D-l ------- Excess Emissions with Fast-pass Enabled Max CO Cutpoints (Data Not Normalized for Model Year Distribution) Excess Excess HC Excess Excess CO Excess Excess NO Vehicle Model Year IM240 HC with fast- Excess HC IM240 CO with fast- Excess CO IM240 Nox with fast- Excess NOx Class Range (grams) pass (grams) Identified (grams) pass (grams) Identified (grams) pass (grams) Identified LDGV 81-82 0 0 N/A 56.74 49.73 87.6% 3.61 3.61 100.0% 83-85 6.13 6.13 100.0% 243.24 237.02 97.4% 3.49 3.49 100.0% 86-89 10.72 10.65 99.3% 301.8 294.27 97.5% 9.92 9.92 100.0% 90-95 8.96 8.68 96.9% 185.26 157.96 85.3% 6.59 6.02 91.4% 96+ 0 0 N/A 0 0 N/A 0 0 N/A ALL 25.81 25.46 98.6% 787.04 738.98 93.9% 23.61 23.04 97.6% LDGT1 81-85 10.99 8.32 75.7% 605.55 560.92 92.6% 19.02 18.72 98.4% 88-89 5.55 . 5.13 92.4% 533.37 485.61 91.0% 2.37 1.69 71.3% 90-95 0.42 0 0.0% 1.77 0 0.0% 1.65 1.65 100.0% 96+ 0 0 N/A 0 0 N/A 0.43 0 0.0% ALL 16.96 13.45 79.3% 1140.69 1046.53 91.7% 23.47 22.06 94.0% LDGT2 81-85 15.62 14.47 92.6% 441.44 441.44 100.0% 6.69 5.72 85.5% 86-87 10.39 10.39 100.0% 285.54 285.54 100.0% 4.11 2.31 56.2% 88-95 0.21 0.21 100.0% 26.15 15.16 58.0% 0.68 0.27 39.7% 96+ 0 0 N/A 0 0 N/A 0 0 N/A ALL 26.22 25.07 95.6% 753.13 742.14 98.5% 11.48 8.3 72.3% Total ALL 68.99 63.98 92.7% 2680.86 2527.65 94.3% 58.56 53.4 91.2% D-2 ------- Excess Emissions with Fast-pass Disabled Max CO Cutpoints (Data Not Normalized for Model Year Distribution) Excess Excess HC Excess Excess CO Excess Excess NO Vehicle Model Year IM240 HC w/o fast- Excess HC IM240 CO w/o fast- Excess CO IM240 Nox w/o fast- Excess NOx Class Range (grams) pass (grams) Identified (grams) pass (grams) Identified (grams) pass (grams) Identified LDGV 81-82 10.99 8.32 75.7% 605.55 603.26 99.6% 19.02 18.72 98.4% 83-85 5.55 5.55 100.0% 533.37 527.14 98.8% 2.37 1.69 71.3% 86-89 0.42 0.42 . 100.0% 1.77 0 0.0% 1.65 1.65 100.0% 90-95 0 0 N/A 0 0 N/A 0.43 0.43 100.0% 96+ 16.96 14.29 84.3% 1140.69 1130.4 99.1% 23.47 22.49 95.8% ALL 15.62 14.47 92.6% 441.44 441.44 100.0% 6.69 5.72 85.5% LDGT1 81-85 10.39 10.39 100.0% 285.54 285.54 100.0% 4.11 2.31 56.2% 88-89 0.21 0.21 100.0% 26.15 26.15 100.0% 0.68 0.38 55.9% 90-95 0 0 N/A 0 0 N/A 0 0 N/A 96+ 26.22 25.07 95.6% 753.13 753.13 100.0% 11.48 8.41 73.3% ALL 0 0 N/A 56.74 49.73 87.6% 3.61 3.61 100.0% LDGT2 81-85 6.13 6.13 100.0% 243.24 237.02 97.4% 3.49 3.49 100.0% 86-87 10.72 10.72 100.0% 301.8 301.22 99.8% 9.92 9.92 100.0% 88-95 8.96 8.87 99.0% 185.26 160 86.4% 6.59 6.02 91.4% 96+ 0 0 N/A 0 0 N/A 0 0 N/A ALL 25.81 25.72 99.7% Ś 787.04 747.97 95.0% 23.61 23.04 97.6% Total ALL 68.99 65.08 94.3% 2680.86 . 2631.5 98.2% 58.56 53.94 92.1% D-3 ------- Appendix E Second-by-Second CPP Variation Limits ------- IM147 CUMULATIVE POSITIVE POWER (CPP) VARIATION OUTPOINTS IM147 REFERENCE DATA TIME SPEED CPP (sec) (mph) fmph2/sec} 0 0.0 0.00 1 0.0 0.00 2 0.0 0.00 3 0.0 0.00 4 0.0 0.00 5 3.3 10.89 6 6.6 43.56 7 9.9 98.01 8 13.2 174.24 9 16.5 272.25 10 19.8 392.04 11 22.2 492.84 12 24.3 590.49 13 25.8 665.64 14 26.4 696.96 15 25.7 696.96 16 25.1 696.96 17 24.7 696.96 18 25.2 721.91 19 25.4 732.03 20 27.2 826.71 21 26.5 826.71 22 24.0 826.71 23 22.7 826.71 24 19.4 826.71 25 17.7 826.71 26 17.2 826.71 27 18.1 858.48 28 18.6 876.83 29 20.0 930.87 30 20.7 959.36 31 21.7 1001.76 32 22.4 1032.63 33 22.5 1037.12 34 22.1 1037.12 35 21.5 1037.12 36 20.9 1037.12 37 20.4 1037.12 38 19.8 1037.12 39 17.0 1037.12 40 17.1 1040.53 41 15.8 1040.53 42 15.8 1040.53 43 17.7 1104.18 44 19.8 1182.93 45 21.6 1257.45 POWER VARIATION OUTPOINTS (mph2/sec) "BASE" MULT. VARYING CPP LIMITS DELTA FACTOR DELTA LQM HIGH 69.7 3.500 244.0 715.31 1,203.41 72.8 3.424 249.3 752.44 1,251.08 75.1 3.386 254.2 778.47 1,286.79 75.4 3.348 252.4 784.71 1,289.53 75.4 3.348 252.4 784.71 1,289.53 75.4 3.348 252.4 784.71 1,289.53 75.4 3.348 252.4 784.71 1,289.53 75.4 3.348 252.4 784.71 1,289.53 75.4 3.348 252.4 784.71 1,289.53 75.4 3.348 252.4 784.71 1,289.53 75.6 3.311 250.4 790.16 1,290.90 75.6 3.311 250.4 790.16 1,290.90 75.6 3.311 250.4 790.16 1,290.90 80.3 3.273 262.7 841.53 1,366.83 86.0 3.235 278.1 904.80 1,461.06 91.4 3.197 292.2 965.27 1,549.63 E-1 ------- IM147 CUMULATIVE POSITIVE POWER (CPP) VARIATION OUTPOINTS IM147 REFERENCE DATA POWER VARIATION OUTPOINTS (mph2/sec) TIME SPEED CPP "BASE- MULT. VARYING CPP LIMITS (sec) (mph) (mph2/sec') DELTA FACTOR DELTA LOW HIGH 46 22.2 1283.73 93.3 3.159 294.8 988.97 1,578.49 47 24.5 1391.14 101.1 3.121 315.6 1,075.55 1,706.73 48 24.7 1400.98 101.8 3.083 314.0 1,087.02 1,714.94 49 24.8 1405.93 102.2 3.045 311.2 1,094.73 1,717.13 50 24.7 1405.93 102.2 3.045 311.2 1,094.73 1,717.13 51 24.6 1405.93 102.2 3.045 311.2 1,094.73 1,717.13 52 24.6 1405.93 102.2 3.045 311.2 1,094.73 1,717.13 53 25.1 1430.78 104.0 3.008 312.8 1,118.02 1,743.54 54 25.6 1456.13 105.8 2.970 314.3 1,141.83 1,770.43 55 25.7 1461.26 106.2 2.932 311.4 1,149.88 1,772.64 56 25.4 1461.26 106.2 2.932 311.4 1,149.88 1,772.64 57 24.9 1461.26 106.2 2.932 311.4 1,149.88 1,772.64 58 25.0 1466.25 106.6 2.894 308.4 1,157.84 1,774.66 59 25.4 1486.41 108.0 2.856 308.6 1,177.85 1,794.97 60 26.0 1517.25 110.3 2.818 310.8 1,206.47 1,828.03 61 26.0 1517.25 110.3 2.818 310.8 1,206.47 1,828.03 62 25.7 1517.25 110.3 2.818 310.8 1,206.47 1,828.03 63 26.1 1537.97 111.8 2.780 310.8 1,227.18 1,848.76 64 26.7 1569.65 114.1 2.742 312.9 1,256.78 1,882.52 65 27.3 1602.05 116.4 2.705 314.9 1,287.13 1,916.97 66 30.5 1787.01 129.9 2.667 346.4 1,440.65 2,133.37 67 33.5 1979.01 143.8 2.629 378.1 1,600.89 2,357.13 68 36.2 2167.20 157.5 2.591 408.1 1,759.09 2,575.31 69 37.3 2248.05 163.4 2.553 417.1 1,830.90 2,665.20 70 39.3 2401.25 174.5 2.515 439.0 1,962.29 2,840.21 71 40.5 2497.01 181.5 2.477 . 449.6 2,047.41 2,946.61 72 42.1 2629.17 191.1 2.439 466.2 2,163.02 3,095.32 73 43.5 2749.01 199.8 2.402 479.8 2,269.18 3,228.84 74 45.1 2890.77 210.1 2.364 496.6 2,394.15 3,387.39 75 46.0 2972.76 216.1 2.326 502.5 2,470.24 3,475.28 76 46.8 3047.00 221.5 2.288 506.7 2,540.32 3,553.68 77 47.5 3113.01 226.3 2.250 509.1 2,603.92 3,622.10 78 47.5 3113.01 226.3 2.250 509.1 2,603.92 3,622*10 79 47.3 3113.01 226.3 2.250 509.1 2,603.92 3,622.10 80 47.2 3113.01 226.3 2.250 509.1 2,603.92 3,622.10 81 47.2 3113.01 226.3 2.250 509.1 2,603.92 3,622.10 82 47.4 3131.93 227.6 2.212 503.6 2,628.37 3,635.49 83 47.9 3179.58 231.1 2.174 502.5 2,677.11 3,682.05 84 48.5 3237.42 235.3 2.136 502.7 2,734.73 3,740.11 85 49.1 3295.98 239.6 2.098 502.7 2,793.27 3,798.69 86 49.5 3335.42 242.4 2.061 499.5 2,835.88 3,834.96 87 50.0 3385.17 246.0 2.023 497.7 2,887.49 3,882.85 88 50.6 3445.53 250.4 1.985 497.1 2,948.47 3,942.59 89 51.0 3486.17 253.4 1.947 493.3 2,992.84 3,979.50 90 51.5 3537.42 257.1 1.909 490.8 3,046.58 4,028.26 91 52.2 3610.01 262.4 1.871 491.0 3,119.03 4,100.99 E-2 ------- IM147 CUMULATIVE POSITIVE POWER (CPP) VARIATION OUTPOINTS IM147 REFERENCE DATA POWER VARIATION OUTPOINTS (mph2/sec) TIME SPEED CPP "BASE" MULT. VARYING CPP LIMITS (sec) (mph) (mph2/sec) DELTA FACTOR DELTA LOW HIGH 92 53.2 3715.41 270.0 1.833 495.1 3,220.33 4,210.49 93 54.1 3811.98 277.1 1.795 497.5 3,314.53 4,309.43 94 54.6 3866.33 281.0 1.758 493.9 3,372.43 4,360.23 95 54.9 3899.18 283.4 1.720 487.4 3,411.82 4,386.54 96 55.0 3910.17 284.2 1.682 478.0 3,432.20 4,388.14 97 54.9 3910.17 284.2 1.682 478.0 3,432.20 4,388.14 98 54.6 3910.17 284.2 1.682 478.0 3,432.20 4,388.14 99 54.6 3910.17 284.2 1.682 478.0 3,432.20 4,388.14 100 54.8 3932.05 285.8 1.644 469.8 3,462.23 4,401.87 101 55.1 3965.02 288.2 1.606 462.8 3,502.17 4,427.87 102 55.5 4009.26 291.4 1.568 457.0 3,552.29 4,466.23 103 55.7 4031.50 293.0 1.530 448.4 3,583.09 4,479.91 104 56.1 4076.22 296.3 1.492 442.2 3,634.06 4,518.38 105 56.3 4098.70 297.9 1.455 433.3 3,665.39 4,532.01 106 56.6 4132.57 300.4 1.417 425.5 3,707.05 4,558.09 107 56.7 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 108 56.7 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 109 56.3 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 110 56.0 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 111 55.0 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 112 53.4 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 113 51.6 4143.90 301.2 1.379 415.3 3,728.63 4,559.17 114 51.8 4164.58 302.7 1.341 405.9 3,758.70 4,570.46 115 52.1 4195.75 305.0 1.303 397.4 3,798.38 4,593.12 116 52.5 4237.59 308.0 1.265 389.7 3,847.93 4,627.25 117 53.0 4290.34 311.8 1.227 382.7 3,907.64 4,673.04 118 53.5 4343.59 315.7 1.189 375.5 3,968.10 4,719.08 119 54.0 4397.34 319.6 1.152 368.0 4,029.31 4,765.37 120 54.9 4495.35 326.7 1.114 363.9 4,131.49 4,859.21 121 55.4 4550.50 330.7 1.076 355.8 4,194.70 4,906.30 122 55.6 4572.70 332.4 1.038 344.9 4,227.76 4,917.64 123 56.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 124 56.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 125 55.8 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 126 55.2 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 127 54.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 128 53.6 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 129 52.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 130 51.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 131 50.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 132 48.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 133 44.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 134 41.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 135 37.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 136 34.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 137 30.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 ------- IM147 CUMULATIVE POSITIVE POWER (CPP) VARIATION OUTPOINTS IM147 REFERENCE DATA POWER VARIATION OUTPOINTS (mph2/sec) TIME SPEED CPP "BASE- MULT. VARYING CPP LIMITS (sec) (mph) (mph2/sec) DELTA FACTOR DELTA LOW HIGH 138 27.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 139 23.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 140 20.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 141 16.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 142 13.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 143 9.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 144 6.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 145 2.5 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 146 0.0 4617.34 335.6 1.000 335.6 4,281.74 4,952.94 Cycle Sums 4617.34 1.000 4,281.74 4,952.94 E-4 ------- Appendix F Regression Summaries ------- Regression Summary - Composite HC, IM240 to IM147 Model Years 1981-1985 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.981278 R Square 0.962907 Adjusted R 0.962636 Square Standard 0.331332 Error Observations 139 ANOVA df SS MS F Significance F Regression 1 390.4283 390.4283 3556.441 6.9E-100 Residual 137 15.03995 0.109781 Total 138 405.4682 Coefficients Standard Error tStat P-vaiue Lower 95% Upper 95% Intercept 0.110694 0.036049 3.070666 0.002576 0.03941 0.181979 IM240 0.896629 0.015035 59.6359 6.9E-100 0.866899 0.92636 IM240 Line Fit Plot ~ IM147 Predicted IM147 F-l ------- Regression Summary - Composite CO, IM240 to IM147 Model Years 1981-1985 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.989621 R Square 0.979349 Adjusted R 0.979198 Square Standard 5.150236 Error Observations 139 ANOVA df SS MS F Significance F Regression 1 172333.7 172333.7 6497.045 2.6E-117 Residual 137 3633.916 26.52493 Total 138 175967.6 Coefficients Standard Error tStat P-value Lower 95% Upper 95% Intercept 0.858255 0.536573 1.599511 0.112011 -0.20278 1.919292 IM240 1.020463 0.01266 80.60425 2.6E-117 0.995428 1.045497 IM240 Line Fit Plot IM147 Predicted IM147 IM240 F-2 ------- Regression Summary - Composite NOx, IM240 to IM147 Model Years 1981-1985 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.988936 R Square 0.977995 Adjusted R 0.977835 Square Standard 0.322696 Error Observations 139 ANOVA . df_ SS MS F Significance F Regression 1 634.0613 634.0613 6088.969 2E-115 Residual 137 14.26619 0.104133 Total 138 648.3275 Coefficients Standard t Stat P-value Lower 95% Upper 95% Error Intercept 0.085613 0.045469 1.882899 0.061834 -0.0043 0.175525 IM240 1.065128 0.01365 78.03185 2E-115 1.038136 1.09212 F-3 ------- Regression Summary - Composite HC, IM240 to IM147 Model Years 1986-1989 All Vehicle Types SUMMARY OUTPUT Repression Statistics Multiple R 0.987927 R Square 0.975999 Adjusted R 0.975876 Square Standard 0.189037 Error Observations 198 ANOVA df SS MS F Significance F Regression 1 284.8199 284.8199 7970.309 1.1E-160 Residual 196 7.004082 0.035735 Total 197 291.824 Coefficients Standard Error tStat P-value Lower 95% Upper 95% Intercept 0.056509 0.015868 3.561206 0.000463 0.025215 0,087802 IM240 0.933646 0.010458 89.27659 1.1E-160 0.913021 0.95427 IM240 Line Fit Plot IM240 ~ IM147 Predicted IM147 F-4 ------- Regression Summary - Composite CO, IM240 to IM147 Model Years 1986-1989 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.984568 R Square 0.969374 Adjusted R 0.969218 Square Standard 5.123058 Error Observations 198 ANOVA df SS MS F Significance F Regression 1 162823.5 162823.5 6203.812 2.5E-150 Residual 196 5144.161 26.24572 Total 197 167967.7 Coefficients Standard Error tStat P-value Lower 95% Upper 95% Intercept 1.679632 0.403976 4.157756 4.8E-05 0.882936 2.476328 IM240 0.939067 0.011922 78.76428 2.5E-150 0.915554 0.96258 IM240 Line Fit Plot 350 300 250 Ł 200 T ^ 150 100 50 0 IM240 F-5 ------- Regression Summary - Composite NOx, IM240 to IM147 Model Years 1986-1989 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.977553 R Square 0.955611 Adjusted R 0.955384 Square Standard 0.298767 Error Observations 198 ANOVA df SS MS F Significance F Regression . 1 376.638 376.638 4219.491 1.6E-134 Residual 196 17.49525 0.089261 Total 197 394.1332 Coefficients Standard tStat P-value Lower 95% Upper 95% Error Intercept 0.058971 0.035732 1.650379 0.100467 -0.0115 0.129439 IM240 1.077932 0.016594 64.95761 1.6E-134 1.045206 1.110659 IM240 Line Fit Plot ~ IM147 Predicted IM147 F-6 ------- Regression Summary - Composite HC, IM240 to IM147 Model Years 1990-1995 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.974081 R Square 0.948834 Adjusted R 0.948723 Square Standard 0.100407 Error Observations 464 ANOVA df_ SS MS F Significance F Regression 1 86.37186 86.37186 8567.38 2.3E-300 Residual 462 4.657643 0.010081 Total 463 91.0295 Coefficients Standard t Stat P-value Lower 95% Upper 95% Error Intercept 0.026672 0.00526 5.071064 5.74E-07 0.016336 0.037007 IM240 0.963839 0.010413 92.56014 2.3E-300 0.943376 0.984302 IM240 Line Fit Plot IM240 ~ IM147 Predicted IM147 F-7 ------- Regression Summary - Composite CO, IM240 to IM147 Model Years 1990-1995 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.916596 R Square 0.840149 Adjusted R 0.839803 Square Standard 3.35827 Error Observations 464 ANOVA df SS MS F Significance F Regression 1 27384.98 27384.98 2428.183 4.6E-186 Residual 462 5210.424 11.27797 Total 463 32595.41 Coefficients Standard tStat P-vaiue Lower 95% Upper 95% Error Intercept 0.392486 0.180702 2.172 0.030364 0.037385 0.747586 IM240 1.037836 0.021061 49.2766 4.6E-186 0.996448 1.079224 IM240 Line Fit Plot 140 0 50 100 150 IM240 F-8 ------- Regression Summary - Composite NOx, IM240 to IM147 Model Years 1990-1995 All Vehicle Types SUMMARY OUTPUT Regression Statistics Multiple R 0.958073 R Square 0.917905 Adjusted R 0.917727 Square Standard 0.306825 Error Observations 464 ANOVA df SS MS F Significance F Regression 1 486.2964 486.2964 5165.597 6.2E-253 Residual 462 43.49332 0.094141 Total 463 529.7897 Coefficients Standard tStat P-value Lower 95% Upper 95% Error Intercept 0.048771 0.020544 2.374 0.018004 0.0084 0.089143 IM240 1.102698 0.015343 71.87209 6.2E-253 1.072548 1.132848 IM240 Line Fit Plot ~ IM147 Predicted IM147 0 2 4 6 8 IM240 F-9 ------- Appendix G Fleet Distribution Data ------- Fleet Distribution Data Vehicle Distribution Data - 9 months 2% random sample (initial test) Data Collected between 7/1/97 and 3/31/98 LDGV LDT1 LDT2 Model Year # Tests % of Fleet # Tests % of Fleet # Tests % of Fleet 1981 102 1.02% 31 0.31% 13 0.13% 1982 138 1.38% 33 0.33% 18 0.18% 1983 153 1.52% 52 0.52% 17 0.17% 1984 318 3.17% 80 0.80% 33 0.33% 1985 372 3.71% 129 1.29% 29 0.29% 1986 414 4.13% 172 1.71% 48 0.48% 1987 450 4.49% 167 1.66% 40 0.40% 1988 540 5.38% 205 2.04% 54 0.54% 1989 540 5.38% 235 2.34% 62 0.62% 1990 512 5.10% 183 1.82% 44 0.44% 1991 542 5.40% 229 2.28% 40 0.40% 1992 533 5.31% 201 2.00% 74 0.74% 1993 559 5.57% 270 2.69% 62 0.62% 1994 654 6.52% 306 3.05% 110 1.10% 1995 673 6.71% 310 3.09% 131 1.31% 1996 100 1.00% 45 0.45% 10 0.10% 1997 0 0.00% 0 0.00% 0 0.00% 1998 0 0.00% 0 0.00% 0 0.00% 1999 0 0.00% 0 0.00% 0 0.00% 2000 0 0.00% 0 0.00% 0 0.00% Total 6600 2648 785 Distribution Between Vehicle Types LDGV LDT1 LDT2 Total 10033 # Tests % of Fleet 6600 65.8% 2648 26.4% 785 7.8% G-l ------- |