Explicit Equations to Estimate the Flammability of Blends of Diesel Fuel, Gasoline and Ethanol

Abstract

Blends of gasoline, diesel fuel and ethanol (�dieseline�) have shown promise in engine studies examining low temperature combustion using compression ignition. They offer the possibility of high efficiency combined with low emissions of oxides of nitrogen and soot. However, unlike gasoline or diesel fuel alone, such mixtures can be flammable in the headspace above the liquid in a vehicle fuel tank at common ambient temperatures. Quantifying their flammability characteristics is important if these fuels are to see commercial service. The parameter of most interest is the Upper Flammable Limit (UFL) temperature, below which the headspace vapour is flammable. In earlier work a mathematical model to predict the flammability of dieseline blends, including those containing ethanol, was developed and validated experimentally. It was then used to study the flammability of a wide variety of dieseline blends parametrically. Gasolines used in the simulations had DVPEs (Dry Vapor Pressure Equivalent) varying from 45 to 110 kPa. The parametric study revealed that the UFL temperatures of all alcohol-free dieseline blends were well correlated by blend DVPE, and the same correlation was found regardless of which gasoline or diesel fuels were used in the blends. For dieselines containing ethanol, the UFL temperatures could also be correlated using blend DVPE, but the UFL temperatures with ethanol present were different than for the alcohol-free blends at any given DVPE and varied with ethanol content. The results were presented graphically, from which UFL temperature could be estimated for any specific dieseline of interest. However, using the graphical data for broader analyses would be extremely time consuming. In the work reported here, explicit correlation equations have been derived that allow UFL temperature to be determined directly for dieseline blends with or without ethanol. The results have been shown to agree with the predictions of the full mathematical model on which they are based with an RMS error of less than 1�C. The equations can be used to quickly compare the UFL temperatures of a wide variety of dieseline formulations and to evaluate the impact of practical issues arising in-service, such as blending errors or regional variations.