PERFORMANCE PREDICTION OF A CI ENGINE USING ARTIFICIAL NEURAL NETWORK FOR VARIOUS SME AND DIESEL FUEL BLENDS

Abstract

This study deals with predicting various performance parameters and exhaust emissions of a four-stroke, four-cylinder, direct injection diesel engine fuelled with soybean oil methyl ester (SME) and its 5%, 20% and 50% blends with jet fuel, marine fuel and No.2 diesel fuel using artificial neural networks (ANNs). In order to acquire data for training and testing the proposed ANN, the test engine was operated at steady-state conditions while varying the engine speed and torque for each fuel case. Using some of the experimental data for training, an ANN model based on standard back propagation algorithm for the engine was developed. This model was used to predict various performance parameters and exhaust emissions of the engine, namely the brake specific fuel consumption, break thermal efficiency, mechanical efficiency, exhaust gas temperature, and emissions of CO, NOx, and CO2. Then, the performance of the ANN predictions were measured by comparing the predictions with the experimental results. It was observed that the ANN model can predict the engine performance and exhaust emissions quite well with correlation coefficients in the range of 0.937–0.989%, mean relative errors in the range of 0.36–16.72% and very low root mean square errors. The results reveal that the ANN approach can accurately predict the performance and emissions of diesel engines using various diesel and biodiesel fuels