Prediction of single cylinder direct injection diesel engine performance fuelled with lemon peel oil biodiesel using artificial neural network

Abstract

The aim of the present work is to predict the performance of single cylinder direct injection diesel engine fuelled with lemon peel oil biodiesel (B25) loaded by ceria nanoparticle additives and coated with yttria-ceria stabilized zirconia using Levenberg-Marquardt algorithm of single-input multiple-output artificial neural network (ANN). The dependent variables are brake power, indicated power, frictional power, brake thermal efficiency, indicated thermal efficiency, volumetric efficiency, fuel flowrate, air flowrate and specific fuel consumption at varying loads from 0 to 100%. The prediction ability of ANN was evaluated by determination coefficient (R 2 ) and root mean square error (RMSE). The average R 2 and RMSE of 0.93 and 0.82 show that the ANN fitted well to the experimental values of response variables at varying loads. The results also validate through correlation coefficient (R) of 0.99254.