Real-time evaporative emissions test data were obtained for thirty-six 1977 to 1989 model year vehicles under conditions representative of a California mid-morning commute. The test procedure involved a cold soak followed by a driving event using the LA92 cycle, followed by a hot soak at 75°F and then a diurnal event where the ambient temperature was increased from 75 to 84°F over six hours. Ambient temperature then decreased from 84°F to 60°F over the next 13 hours and then was increased from 60°F to 75°F over the final five hours of the test.
Overall, the test results could generally be described by three types of responses:
High hot soak emissions that continued for much longer than one hour and only a minimal increase in the rate of evaporative emissions during the heat builds of the diurnal test;
High and extended hot soak emissions coupled with substantial emissions during one or both of the heat builds during the diurnal test; and
Low hot soak emissions coupled with only a minimal increase in the rate of evaporative emissions during the two heat builds of the diurnal test.
In general, vehicles that failed an inspection and maintenance (I/M) type canister purge test tended to exhibit the first response, while vehicles that failed an I/M-type evaporative system pressure test tended to exhibit the second response. As expected, vehicles passing both types of I/M tests and gas cap pressure tests tended to exhibit the third response, and had low hot soak and diurnal emissions.
The real-time evaporative emissions data obtained from the nineteen 1980-1985 model year vehicles in the 36-vehicle fleet were compared with emission predictions from draft versions of the evaporative emissions algorithms of EPA's MOBILE6 and CARB's EMFAC2000 models. This comparison indicates that, for the 19-vehicle fleet, both models underestimated the magnitude and duration of hot soak emissions. In contrast, both models overestimated the magnitude of partial-day diurnal emissions for the same fleet. Further, a comparison of model results with data representative of a mid-morning commute indicates that MOBILE6 underpredicted emissions from the 19 vehicles while EMFAC2000 overpredicted emissions.