Development of emissions predictor equations for a light-duty diesel engine using biodiesel fuel properties

In this experimental study, the effects of pertinent biodiesel fuel properties on exhaust emissions from a light-duty diesel engine were investigated. To achieve this, 31 different fuel blends were tested under a steady-state emissions test cycle. They consisted of coconut methyl ester (CME), palm methyl ester (PME), soybean methyl ester (SME) and their blends with fossil diesel at 10 vol% intervals. The combinations covered all biodiesel-diesel blend levels and also the full range of the saturation-unsaturation proportions typical of biodiesel fuels. From this study, the measurements of exhaust NO, CO, UHC and smoke opacity were correlated against that of key fuel properties such as cetane number, kinematic viscosity, density, carbon residue and fuel-bound oxygen content. It was found that the cetane number of a fuel strongly influences the engine-out NO and CO levels. The carbon residue and fuel-bound oxygen contents have the greatest effects on smoke opacity levels. Meanwhile, the kinematic viscosity values are construed to affect CO and UHC concentrations. Fuel density plays a minor but significant enough role in influencing most emissions, especially CO. Based on experimental data points and fitted linear trendlines, a set of four predictor equations for emissions were proposed using a non-statistical graphical method. The empirical equations are designed to be applicable for all present and immediate future fuels. Validation against the experimental points showed that the average discrepancies between experimental and predicted values are 1.98%, 2.47%, 10.34%, and 13.85% for NO, CO, smoke opacity and UHC, respectively. The predictor equations are invaluable for rapid analysis of engine-out emissions as they utilise only common fuel properties easily attainable from fuel specification sheets, without the need for expensive and time-consuming experiments. The method used to develop the predictor equations proposed in this study is also transferable to diesel engines of any class. (C) 2011 Elsevier Ltd. All rights reserved.