Artificial neural network approach for parametric investigation of biodiesel synthesis using biocatalyst and engine characteristics of diesel engine fuelled with Aegle Marmelos Correa biodiesel

In this present study, biodiesel was produced from an eco-friendly and non-edible AMC seed oil using a biocatalyst. The optimum biodiesel yield was obtained as 92% by undergoing microwave transesterification with 20 min time, 4.5 wt% catalyst amount and 1:12 oil-methanol ratio at 55 °C. The activation energy needed for the reaction was 51.9 kJ/mol. The thermodynamic parameters for the transesterification process, such as enthalpy and entropy were 56.4 kJ/mol and −0.091 kJ/mol. Further, the engine studies were carried out for different fuel injection pressures and injection timing. Performance results reveal that BSEC and BTE of biodiesel are lower and higher than that of diesel fuel for 400 bar FIP and 27° CA bTDC FIT at full load respectively. However, except NO, composite emissions of CO, UBHC and dry soot are comparatively lesser than that of standard emission norms. It is thereby inferred from the experimental results that the optimum fuel injection pressure and timing are 400 bar and 27° CA bTDC. The developed ANN model precisely predicted the out data with a higher R2 value for biodiesel synthesis and engine characteristics. Hence, it can be concluded that ANN is the best tool for predicting output data. [Display omitted] •Transesterfication reaction found pseudo 1st order with Ea of 51.9 kJ/mol.•Ca(CH3O)2 catalyst undergone 5 cycles of operation with only 7.6% loss of activity.•Increasing FIP and FIT minimizes BSEC, CO, UBHC and Smoke except NO.•FIP of 400 bar and FIT of 27° bTDC meets Indian genset emissions except NO.