The EPAct/V2/E-89 gasoline fuel effects program collected emissions data for 27 test fuels using a fleet of 15 high-sales cars and light trucks from the 2008 model year (all with port fuel injection). The test fuel matrix covered values of T50, T90, vapor pressure, ethanol content, and total aromatic content spanning ranges typical of market gasolines. Emission measurements were made over the LA92 cycle at a nominal temperature of 24°C (75°F). The resulting emissions database of 956 tests includes a particulate matter (PM) mass measurement for each. Emission models for PM fuel effects were fit based on terms for which the fuel matrix was originally optimized, with results published by EPA in a 2013 analysis report.
This paper presents results of a subsequent modeling analysis of this PM data using the PM Index fuel parameter, and compares these models to the original versions. Two related observations of interest are also discussed: the first is a significant reinforcing interaction between ethanol and PM Index; the second is the wide variation in sensitivity of PM emissions to fuel parameters across the 15 vehicles in the test fleet, indicating that fuel interacts in important ways with engine and vehicle design characteristics, calibrations, and control algorithms.