Artificial neural network prediction of performance and emissions of a diesel engine fueled with palm biodiesel

Abstract

Increasing of energy consumption, depletion of petroleum fuels and harmful emissions have triggered the interest to find substitute fuels for diesel engines. Palm ethyl ester was synthesized from palm oil through transesterification process. The physicochemical properties of palm biodiesel have been measured and confirmed in accordance with ASTM standards. The aim of the paper is to show the effect of different diesel-palm biodiesel blends on performance, combustion and emissions in diesel engine at engine load variation. Artificial Neural Network was used for the prediction of engine performance, exhaust emission and combustion characteristics parameters. Palm ethyl ester and diesel oil were blended in 5, 10, 15 and 20 by volume percentage. The maximum decreases in thermal efficiency, fuel–air equivalence ratio for B20 were 1.5, 3.5, 6 and 8% but the maximum increases in BSFC, exhaust gas temperature and NOx emission for B20 at full load about diesel fuel were 9, 8 and 10%, respectively. The highest decreases in CO, HC and smoke emissions of B20 about diesel oil at full load were 2, 35 and 18.5% at full load, respectively. Biodiesel blend B20 achieved the maximum declines in peak HRR, cylinder temperature and combustion duration about diesel fuel. The results of ANN were compared with experimental results and showed that ANN is effective modeling method with high accuracy. Palm biodiesel blends up to 20% showed the highest enhancements in engine performance, combustion and emission reductions compared to diesel fuel.