The purpose of this study was to intercompare hydrocarbon (HC) measurements performed by a number of different instruments: a gas chromatograph (GC), a flame ionization detector (FID), a fourier transform infrared spectrometer (FTIR), a commercially produced non-dispersive infrared analyzer (NDIR), and two remote sensors. These instruments were used to measure total HC concentrations in a variety of samples, including (1) ten different individual HC species, (2) 12 different vehicle exhaust samples, and (3) three different volatilized fuel samples. The 12 exhaust samples were generated by operating two different vehicles on a dynamometer. Each vehicle was operated at different times with three different fuels. The vehicles were operated fuel rich, i.e., with low air/fuel ratios to encourage elevated exhaust HC levels. Some of the exhaust samples were obtained while operating each vehicle at a stoichiometric air/fuel ratio with one spark plug wire disconnected.To quantify the degree to which the various instruments agreed with the FID, a parameter called the response factor was used, where the response factor was defined as the HC/CO 2 ratio measured by each instrument divided by the HC/CO 2 ratio measured by the dynamometer bench. Of the various instruments, only the GC yielded response factors that were consistently at or close to one. The other instruments typically had values at or below one. For the ten individual HC species studied, the NDIR and remote sensors obtained response factors between 0.05 and 1.0, with the IMPLICATIONSThe data in this report suggest that remote sensors can accurately measure CO emissions but have large errors in measuring HC emissions. For this reason, remote sensors should not be relied upon to measure the inventory of inuse vehicle exhaust hydrocarbon emissions. Furthermore, when used to identify high emitting vehicles, remote sensors can be expected to have a higher error of omission for detection of high hydrocarbon emitters than that obtained for high CO emitters.highest response factors being obtained for the alkanes and the lowest response factors obtained for toluene and ethylene.For the exhaust samples, the NDIR and remote sensors obtained response factors between 0.23 and 0.68. For raw fuel samples, the response factors were between 0.44 and 0.68. NDIR and remote sensor measurements correlated very poorly with total HC in exhaust. INTRODUCTIONThe use of remote sensors to measure vehicle exhaust emissions is gaining widespread acceptance. These devices are being used to measure fleet average emissions and/or to identify high carbon monoxide (CO) and/or high hydrocarbon (HC) emissions from in-use vehicles.

No abstract

Remote sensing measurements of CO emissions from onroad vehicles were made in California in 1991 and in Michigan in 1992. It was determined that both fleets had a small linear increase in the high emitter frequency (vehicles emitting more than 4% CO) as a function of vehicle age for 1986 and newer model vehicles. Although high emitting vehicles were only a small minority of the fleet, they had a dominant impact on the mean CO and total CO emitted by the fleet. In Michigan, the highest emitting 5% of passenger cars generated 45% of the CO from cars. In California, the highest emitting 5% of passenger cars generated 38% of the CO from cars. There was a high correlation between the mean CO emitted by each model year of vehicle and the frequency of high emitting vehicles within the model year for both the Michigan and California fleets. The frequency of high emitters within any model year had no obvious relation to that model year's certification standards. The high emitter frequencies for vehicles less than nine years old were very similar for the California and Michigan fleets. An increase in the high emitter frequency in the ten-year-old and older Michigan passenger car fleet (relative to the California passenger car fleet), suggests, but does not conclusively demonstrate, that the rate of high emitters in Michigan and California is reduced by the inspection and maintenance (I/ M) programs. INTRODUCTIONRemote sensors have been in use to measure the CO and HC emissions from on-road vehicles for approximately five years. These devices have shown potential for measuring vehicle fleet average emissions 1,2 and for identifying high emitting vehicles.3,4 Studies using these devices have shown dramatic differences in average fleet emissions as a function of vehicle model year.Several large-scale remote sensing studies have been performed over the last few years, making it possible to begin to compare the emissions from model year classes of vehicles that have been measured at widely different times. Such comparisons might begin to show the effects of deterioration of emissions within specific model year classes of vehicles. These studies can also be used to look for measurable effects on tailpipe emissions in areas subject to inspection and maintenance (I/M) programs.We will present measurements of CO emissions from tens of thousands of on-road vehicles obtained during the Michigan Roadside Study (MRS) of 1992.5 Mean CO concentrations and high emitter frequencies measured during

No abstract