Effect of Fuel Parameters on Emissions from a Direct Injection Spark Ignition Engine During Constant Speed, Variable Load Tests

A 1998 Toyota Corona passenger car with a direct injection spark ignition (DISI) engine was tested at constant engine speed (2000 rpm) over a range of loads. Engine-out and tailpipe emissions of gas phase species were measured each second. This allowed examination of the engine-out emissions for late and early injection. Seven fuels were used for these tests: five blended fuels and two pure hydrocarbon fuels. These seven fuels can be divided into groups for examination of the effects of volatility, MTBE, and structure (an aromatic versus an i-alkane). Correlations between the fuel properties and their effects on emissions are presented. Use of steady state tests rather than driving cycles to examine fuel effects on emissions eliminates the complications resulting from accelerations, decelerations, and changes of injection timing but care had to be taken to account for the periodic regenerations of the lean NOx trap/catalyst. Statistically significant effects, at the 95% confidence level, of the fuel on the engine-out emissions for both early and late injection were found. For late injection, the engine-out Emissions Index for hydrocarbons was found to be a function of volatility and correlations were found with respect to a number of volatility parameters and other fuel properties. The strongest relationship to a volatility parameter was an inverse correlation with the fraction evaporated at 120°C (E120). Correlations between the engine-out EIHC during early injection and the properties of the blended fuels were also identified. In this case, the most important volatility parameter was E100. The results for engine-out NOx and CO during late injection are also discussed, as are fuel effects on the efficiency of the lean NOx trap/catalyst.