An experimental based artificial neural network modeling in prediction of optimum combustion, performance, and emission from diesel engine operated with three biodiesels

Abstract

Purpose This study aims to improve the performance and to regulate the harmful emission from the diesel engine. For this purpose, palm oil biodiesel (POBD), waste cooking biodiesel (WCBD) and animal fat biodiesel (AFBD) are used for examination. Design/methodology/approach The transesterification process was followed to convert the three raw oils into biodiesels and the experiments are conducted at various loads with fixed 25 rps. Diesel as a reference fuel and three neat biodiesels are tested for emissions and performance. By training the experimental results in an artificial neural network (ANN), the best biodiesel was predicted. Findings The biodiesels are tested for significant fuel properties with the American Society for testing and materials standards and observed that kinematic viscosity, density and cetane number are recorded higher than diesel fuel. The fatty acid composition (FAC) from chromatography reveals the presence of unsaturated FAC is more in POBD (70.89%) followed by WCBD (57.67%) and AFBD (43.13%). The combustion pressures measured at every degree of crank angle reveal that WCBD and AFBD exhibited on far with diesel fuel. Compared to diesel fuel WCBD and AFBD achieved maximum brake thermal efficiency of 31.99% and 30.93% at 75% load. However, there is a penalty in fuel consumption and NOx emissions from biodiesels. On the other hand, low carbon monoxide, unburnt hydrocarbon emissions and exhaust smoke are reported for biodiesels. Finally, WCBD was chosen as the best choice based on ANN modeling prediction results. Originality/value There is no evident literature on these three neat biodiesel applications with the mapping of ANN modeling.