For Presentation at the Air & Waste Management Association's 91st Annual Meeting & Exhibition, June 14-18, 1998, San Diego, California
Effects of Oxygenated Fuels Use on I/M240 Emissions Test Results in Colorado
98-WA61A.01
Larry G. Anderson
Department of Chemistry and Center for Environmental Sciences, University of Colorado at Denver, Denver, CO 80217-3364 and Department of Chemistry, U.S. Air Force Academy, USAF Academy, CO 80840
Edward B. Wilkes
Sentinel Environmental, 7818 S. Windermere Circle, Littleton, CO 80120
ABSTRACT
The effects of oxygenated fuels use on the emissions from motor vehicles is assessed using IM240 emissions data collected in the Denver metropolitan area during 1995 and 1996. In 1995, 30,842 vehicles passed their first emissions test of the year during the oxygenated fuels period, and 72,170 vehicles passed during the non-oxygenated fuels period. For this portion of the tested fleet CO emissions were reduced by 13%, HC emissions were reduced by 9% and NOx emissions were increased by 15% during the oxygenated fuels period. A similar analysis was conducted for 5,894 vehicles that failed their first emissions test of the year during the oxygenated fuels period, and 16,234 vehicles that failed during the non-oxygenated fuels period. CO emissions were reduced by 4%, HC emissions were increased by 6% and NOx emissions were increased by 18% during the oxygenated fuels period. In 1996, 25,662 vehicles passed their first test of the year during the oxygenated fuels period, and 65,491 vehicles passed during the non-oxygenated fuels period. The results are very different from the 1995 test year results, CO emissions were increased by 5%, HC emissions were increased by 15% and NOx emissions were increased by 37%. The benefits of oxygenated fuels on the reduction of CO emissions from the 1995 IM240 data are consistent with the results of ambient data analysis reported previously, but much lower than MOBILE 5A predictions. One of the very disturbing aspects of this work is the lack of consistency of the results between the first and second years of IM240 testing. Another very disturbing factor is the relatively large increase in NOx emissions that is seen in this study. This suggests that additional studies of the effects of reformulated gasoline on NOx emissions are necessary.
INTRODUCTION
In January 1988, the Colorado Front Range began the use of oxygenated fuels, in an attempt to reduce ambient concentrations of carbon monoxide (CO). In part based on the modeled success of oxygenated fuels in Colorado, the 1990 Clean Air Act Amendments mandated the use of oxygenated fuels in areas where ambient CO exceeds federal standards. Oxygenated fuels are used in 31 metropolitan areas of the U.S. to reduce winter CO problems. The Clean Air Act Amendments also required the use of reformulated gasoline (RFG) including at least 2.0% oxygen by weight in areas that have ozone problems. Currently, RFG is used in the most polluted urban areas across 17 states and the District of Columbia. These areas account for about 32% of the U.S. gasoline market (approximately 2.5 million barrels/day) (1). During November 1997, the total finished motor vehicle gasoline supply was 8.0 million barrels/day, while the production of fuel ethanol was 98 thousand barrels/day and that for methyl tertiary butyl ether (MTBE) was 203 thousand barrels/day. Fuel oxygenates account for over 3.7% by volume of the total finished motor vehicle gasoline available on a nationwide basis.
In Colorado, the oxygenated fuels program has remained a winter CO only program. All of the automotive fuel sold in the Colorado Front Range area was required to contain 1.5% oxygen by weight from January 1 through February 28, 1988. During this first year the oxygenate requirements were met with about 95% of the fuel sold being an 8% by volume mixture of MTBE and gasoline. The remainder of the fuel sold was a 10% by volume blend of ethanol and gasoline (2). In each subsequent year the program required motor vehicle fuels to be oxygenated during the winter months, November through February. The oxygen content requirement has been steadily increasing while the additive used has gradually shifted from largely MTBE to largely ethanol blended fuels (3). During the winter of 1997-1998, nearly all of the fuel being sold is blended with ethanol at 10% by volume. This program is modeled by MOBILE 5A to reduce the emissions of carbon monoxide from motor vehicles in the Denver area by about 30% (3). The average oxygen content of the fuel being sold is required to exceed 3.1% by weight. In recent years, the winter program has been shortened to end on February 21, 1997 and on February 7, 1998 (4).
For several years, our research group has been using ambient carbon monoxide data to assess the effectiveness of the oxygenated fuels program (5-12). Recently, a number of other research groups have begun to assess the effectiveness of oxygenated fuels for the reduction of ambient carbon monoxide concentrations. The status of this research has been reviewed recently (12). Most recently, we have extended our efforts to assessing the effectiveness of the oxygenated fuels program to an analysis of inspection and maintenance program data (11,13). Beginning in January 1995, a centralized IM240 dynamometer test program was begun in the Denver metropolitan area.
The basic inspection and maintenance (I/M) program in effect along the Colorado Front Range was changed, beginning in January 1995. An enhanced I/M area was established that included the Denver metropolitan area. In this area, all 1982 and newer light duty vehicles are required to undergo a centralized IM240 dynamometer emissions test. Older and larger vehicles are required to undergo an idle test. The IM240 test program measures emissions, not only for carbon monoxide, but also for hydrocarbons and nitrogen oxides. In this paper, we will discuss the results of the analyses of the 1995 and 1996 IM240 emissions data sets.
IM240 EMISSIONS DATA
In 1995, 817,846 IM240 emissions tests were conducted, excluding aborted tests. Of these, 575,003 of the tests were "fast pass tests". Vehicles are allowed to "fast pass" the emissions test if their emissions are sufficiently low after a minimum of 30 seconds on the dynamometer. These results have not been included in this analysis. The remaining 242,843 tests were full tests lasting the entire 4 minutes (240 seconds). We have divided this data set into those that passed the emissions test on the first attempt, those that failed on the first attempt, and multiple tests of the same vehicle. We have not analyzed repeated test results in this work. 167,307 vehicles passed the IM240 test on the first attempt during 1995 and 35,669 failed on the first attempt.
In 1996, 762,025 IM240 emissions tests were conducted. Of these, 545,178 of the tests were "fast pass tests". Again for the 1996 data set, we have not completed an analysis of these results. The remaining 216,847 tests were full tests lasting the entire 240 seconds. Of these, 134,251 vehicles passed the IM240 test on the first attempt during 1996 and 29,380 failed on the first attempt.
For the results presented in this paper, the data set was further filtered before conducting the analyses. We chose to eliminate specialty vehicles from these analyses, those vehicles for which less than 10 vehicles were tested during an oxygenated or non-oxygenated fuels period. Since, we were interested in determining an oxygenated fuel effect on emissions, the data set was broken into a winter oxygenated fuel period (January, February, November and December) and a non-oxygenated fuel period (April through September). Tests conducted during March and October were eliminated from these analyses, since these were viewed as transition months between the two fuel types. The results of other analyses will be discussed in the presentation.
Temperature Effects
Prior to the analysis of the data for an oxygenated fuel effect, we explored the data set for artifacts. This analysis concentrated on those vehicles that passed the full IM240 test on the first attempt. This subset of the data was relatively well behaved. It was not subject to wide variations in results that are more common for those vehicles failing the test.
One of the potential artifacts that was investigated was the effect of ambient temperature on the emissions results. Each test bay is equipped with an ambient temperature sensor, and this data is recorded in the data files generated with each emissions test. One expects to find a dependence of emissions on engine operating temperature, but the vehicles that are tested are expected to be warmed up. The vehicles that are tested are driven to the test station and probably idled while waiting to begin the test. Hence, we expected to be looking at the effect of ambient temperature on emissions from engines that are fully warmed up. This should not show much of an effect.
Data from the oxygenated fuel and the non-oxygenated fuel periods were analyzed separately. The emissions data from the two periods were segregated into 5oC wide bins. Figure 1 shows a plot of the percent difference in CO emissions from the mean of the data collected between 0 and 30oC for the entire period of both years as a function of ambient temperature. Each of the points plotted represent the mean for at least 100 emissions tests in the temperature range. For the emissions data collected during the non-oxygenated fuels period, for which the indicated ambient temperature was above 30oC, there are over 18,000 measurements each year that average more than 10% above the mean for the period. For the data collected at ambient temperatures less than 0oC, it is not nearly as convincing that there is a significant ambient temperature effect on CO emissions. We are convinced that ambient temperatures above 30oC have a real effect on the emissions results. But for the purpose of these analyses, we have restricted the data set to be analyzed to those ambient temperatures between 0oC and 30oC.
We believe that the high temperature effect on the emissions results may be related to inadequate cooling by the fans placed in front of the vehicle radiators during the dynamometer test.
Other Test Effects
The potential for other unexpected test effects were investigated. One artifact that was found in the 1996 data was unusually low emissions recorded for an eighteen day period during the late summer. It is not clear what the origin of this artifact was, but eliminating this data from the analyses had little effect on the results. The data sets that will be analyzed further include 30,842 full IM240 tests during the oxygenated and 72,170 full IM240 tests during the non-oxygenated fuels period for 1995, and 25,662 full IM240 tests during oxygenated and 65,491 full IM240 tests during non-oxygenated fuels periods for 1996. For each of these data sets the vehicles that passed the test on the first attempt and were tested at ambient temperatures between 0oC and 30oC were included. Similarly, we have analyzed 5,894 full IM240 tests during the oxygenated fuels period and 16,234 full IM240 tests during the non-oxygenated fuels period, for vehicles that failed the test on the first attempt in 1995 and were tested at ambient temperatures between 0oC and 30oC.
Representativeness of the Full Test Results
For the 1995 data set, about 80% of the IM240 tests resulted in a fast pass, about 15% were full test passes, and about 5% were failures. In this analysis, we are only looking at two parts of this data set, the full tests that resulted in either a pass or a failure on the first test of the vehicle during the year. Hence we are excluding 80% of the fleet that was tested from these analyses. It is necessary to see how well the vehicles used in our analyses represent the entire fleet. Our analyses tend to overrepresent the larger engine sizes, those with engine displacements of 5.0 liters or more. Similarly, our analyses include a larger percentage of older vehicles than the entire fleet tested. Vehicles newer than 1986 model year are slightly underrepresented in our full test analyses, while vehicles older than 1986 model year are overrepresented in our full test analyses.
ANALYSIS RESULTS
1995 Full Test Passes
Figure 2 shows the emissions of CO, HC and NOx for those vehicles that had the full IM240 test and passed the test on the first attempt during the year. The emissions data are shown for the oxygenated and non-oxygenated fuels seasons separately, as a function of model year. It is quite apparent from this figure that the emissions of the older vehicles are higher than the emissions of the newer vehicles. It is also quite apparent that the CO emissions during the oxygenated fuels period are lower than during the non-oxygenated fuels period. On the other hand, the NOx emissions are consistently higher during the oxygenated fuels period than during the non-oxygenated fuels period. The HC emissions are generally, but not always, lower during the oxygenated fuels period than during the non-oxygenated fuels period.
Figure 3 shows a plot of the percentage reduction of the emissions of CO, HC and NOx during the oxygenated fuels period compared to the non-oxygenated fuels period, as a function of model year. In Figure 2, we saw that the absolute change in emissions due to fuels differences got much smaller for the newer vehicles, but the percentage decreases in CO and HC emissions with oxygenated fuels use gets only slightly smaller as we move to the newer model year vehicles. The percentage increase in NOx emissions appears to get larger as we move to newer vehicles.
Overall, we have found that there is about a 13% reduction in CO emissions, a 9% reduction in HC emissions and almost a 15% increase in NOx emissions due to oxygenated fuels use, for the vehicles that passed the full IM240 test on the first attempt of the year.
1995 Full Test Failures
Figure 4 shows the emissions of CO, HC and NOx for those vehicles that had the full IM240 test and failed the test on the first attempt during the year. The emissions data are shown for the oxygenated and non-oxygenated fuels seasons separately, as a function of model year. It is quite apparent that the emissions shown in this figure for CO and HC are about triple those for the vehicles that passed the full test, shown in Figure 2. On the other hand, the NOx emissions are slightly lower for the vehicles that failed the emissions test. There is much greater variability in the emissions results for these vehicles that failed the test. This is partially due to it being a smaller data set. The CO emissions during the oxygenated fuels period are generally lower than during the non-oxygenated fuels period. But, the NOx emissions are consistently higher during the oxygenated fuels period than during the non-oxygenated fuels period. The HC emissions are often higher during the oxygenated fuels period than during the non-oxygenated fuels period.
Figure 5 shows a plot of the percentage reduction of the emissions of CO, HC and NOx during the oxygenated fuels period compared to the non-oxygenated fuels period for these vehicles that failed the test. The effect of oxygenated fuels is much more variable for these vehicles.
Overall, we have found that there is about a 4% reduction in CO emissions, a 6% increase in HC emissions and a 18% increase in NOx due to oxygenated fuels use for the vehicles that failed the full IM240 test.
1996 Full Test Passes
Figure 6 shows the emissions of CO, HC and NOx for those vehicles that had the full IM240 test and passed the test on the first attempt during the year. The emissions data are shown for the oxygenated and non-oxygenated fuels seasons separately, as a function of model year. As with the 1995 emissions data, it is apparent that the emissions of the older vehicles are higher than the emissions of the newer vehicles. It is also quite apparent that the CO emissions for the 1996 tests are consistently higher than those from 1995. This is particularly obvious for the 1982 and 1983 model year vehicles, where CO emissions were about 43 g/mi during both the oxygenated and non-oxygenated fuels period. The 1995 tests had CO emissions between 30 and 37 g/mi for the 1982 and 1983 model year vehicles. In the 1996 data, CO emissions during the oxygenated fuels period differed much less from those during the non-oxygenated fuels period. Both HC and NOx emissions were consistently higher during the oxygenated fuels period than during the non-oxygenated fuels period.
We do not currently understand the reason for the apparent differences between the emissions test results for 1995 and those for 1996. Overall, we have found that there is about a 5% increase in CO emissions, a 15% increase in HC emissions and a 37% increase in NOx emissions due to oxygenated fuels use for vehicles that passed the first full IM240 test of 1996.
CONCLUSIONS
This paper presents an assessment of the effectiveness of oxygenated fuels use for the largest fleet of real world vehicles to be analyzed. This analysis suggests that the oxygenated fuels program may reduce CO emissions by as much as 13%, although the current results of the analysis of the 1996 IM240 data suggest a small increase in CO emissions.
The CO emissions reductions observed in this study are, at best, about half those predicted by MOBILE 5A modeling for the Denver area (3). As has been suggested before, the results of MOBILE 5A modeling appear to overpredict the effectiveness of oxygenated fuels use (14). The MOBILE model is used for regulatory purposes, such as preparation of State Implementation Plans (SIPs). Another EPA model, COMPLEX, predicts effects on CO emissions that are much more consistent with those observed in this study and the results of assessments based on the analysis of ambient concentration data for CO (12, 14).
Probably the more disturbing result of this study, is the large increase in NOx emissions that was found. Most of the earlier emissions studies have found only a small increase in NOx emissions as a result of oxygenated fuels use. The COMPLEX model predicts a small decrease in NOx emissions. The only other large scale study of the effects of oxygenated fuels on NOx emissions for in-use vehicles found a 0 ± 10% effect in a tunnel study (15). These results are considerably different than those reported in this work.
Phase I RFG use which began January 1, 1995 is required to reduce air toxic and volatile organic compounds (VOCs) emissions from motor vehicle fuels by at least 15% with no increase in NOx emissions. Phase II RFG use which will begin January 1, 2000, is required to achieve at least a 25% reduction in VOCs, a 20% reduction in air toxics, and a 4 - 7 % reduction in NOx emissions. RFG must contain a minimum of 2% oxygen by weight (16). The data analyzed here is for an oxygenated fuel, not RFG, but does lead to a concern that NOx emissions may increase as a result of RFG use.
Another very disturbing aspect of the current work is the apparent lack of consistency in the results collected during these two years of the IM240 emissions testing program in Colorado. This leads one to question the relative value of emissions results from an expensive and relatively inconvenient centralized inspection and maintenance program.
This work suggests that similar assessments should be undertaken in areas that have a seasonal RFG program. It also suggests that the data from such enhanced inspection/maintenance programs should be looked at more closely.
REFERENCES
1. Clean Fuels Development Coalition, Reformulated Gasoline: History and Overview of the Reformulated Gasoline Program; 1997: http://www.cleanfuels.net/doc/TECHLIB/doc/techpapr/rfgbrf/RFGBRIEF.PDF.
2. Livo, K. B.; Miron, W.; Hollman, T. 1988-89 Oxygenated Fuels Program, Air Pollution Control Division Final Report to the Colorado Air Quality Control Commission, Colorado Department of Health, Denver, 1989.
3. Livo, K. B.; Miron, W. 1995 Oxygenated Gasoline Program: Air Pollution Control Division Report to the Colorado Air Quality Control Commission, Colorado Department of Public Health and Environment, Denver, 1995.
4. Colorado Air Quality Control Commission, Regulation 13, 1997.
5. Anderson, L. G.; Lanning, J. A.; Machovec, C.M.; Meglen, R. "The Effects of Colorado's Mandated High Oxygen Fuels Program on Atmospheric Pollution in Denver", in Proceedings of the 82nd Annual Meeting of the Air Pollution Control Association, 89-124.4, 1989.
6. Anderson, L. G.; Machovec, C.M.; Lanning, J.A.; Aloi, M. J.; Meglen, R. R. "The Effects of Using Oxygenated Fuels on the Concentration of Carbon Monoxide in Denver", in Proceedings of the 85th Annual Meeting of the Air & Waste Management Association, 92-91.07, 1992.
7. Wolfe, P.; Anderson, L.G.; Lanning, J. A. "A Structural Time Series Assessment of the Effectiveness of the Oxygenated Fuels Program in Reducing Carbon Monoxide in Downtown Denver", in Proceedings of the 86th Annual Meeting of the Air & Waste Management Association, 93-TP-41B.03, 1993.
8. Wolfe, P.; Anderson, L. G.; Lanning, J. A.; Jones, R.H. "A Structural Time Series Assessment of the Effectiveness of the Oxygenated Fuels Program in Reducing Carbon Monoxide in Five Western U S Cities", in Proceedings of the 87th Annual Meeting of the Air & Waste Management Association, 94-WP91.04, 1994.
9. Anderson, L. G.; Wolfe, P.; Barrell, R. A.; Lanning, J. A. "The Effects of Oxygenated Fuels on the Atmospheric Concentrations of Carbon Monoxide and Aldehydes in Colorado", in Alternative Fuels and the Environment, F. S. Sterret, ed., Lewis Publishers, Boca Raton, FL, Chapter 5, 75 - 103, 1994.
10. Wolfe, P.; Anderson, L. G.; Lanning, J. A.; Jones, R.H. "Techniques for Assessing the Effectiveness of Oxygenated Fuels", in Proceedings of the 89th Annual Meeting of the Air & Waste Management Association, 96-WP89.06, 1996.
11. Anderson, L. G.; Lanning, J. A.; Wilkes, E.; Wolfe, P.; Jones, R. H. Effects of Using Oxygenated Fuels on Carbon Monoxide, Formaldehyde and Acetaldehyde Concentrations in Denver", in Proceedings of the 90th Annual Meeting of the Air & Waste Management Association, 97-RP139.05, 1997.
12. Anderson, L. G.; Lanning, J. A.; Wolfe, P. "Oxygenated Fuels to Reduce Carbon Monoxide Concentrations", Designing Transportation Fuels for the Environment, M. R. Khan, D. C. Cronauer and J. G. Reynolds, eds., Taylor & Francis, submitted, 1998.
13. Anderson, L. G.; Wilkes, E. B. "Effects of Oxygenated Gasoline on Vehicle Emissions in the Metropolitan Denver Area", in Proceedings of the Seventh CRC On-Road Vehicle Emissions Workshop, pp. 3-83 to 3-95, 1997.
14. National Science and Technology Council. 1997. Interagency Assessment of Oxygenated Fuels. Washington, DC.
15. Kirchstetter, T. W.; Singer, B. C.; Harley, R. A.; Kendall, G. R.; Chan, W. Environ. Sci. Technol., 1996, 30, 661-670.
16. Clean Air Act, Section 211(k), (42 U.S.C. 7545).