Prediction of emissions of a diesel engine fueled with soybean biodiesel using artificial neural networks

Recently, the usage of biodiesel as an alternative energy source instead of fossil-based fuels becomes very popular because biodiesel is totally renewable and has more favorable combustion emission profile, however; to determine exhaust emission values at different loads and engine speeds is an important challenge and requires both time consuming and expensive experiments. Instead of conducting experiments, artificial neural network (ANN) models which are computing systems composed of neurons are used to solve complex functions can be used. Therefore, in this study an ANN model was prepared in order to predict the exhaust emissions values of 100% soybean biodiesel using diesel engine for different engine speeds at varying load conditions. Engine speed, torque and exhaust temperature values were used as input in order to predict CO, CO2, NOx and NO2 emissions and coefficient of correlation (R), mean absolute percentage error (MAPE) values were calculated in order to define correlation between the target value and output value and identify the convergence between the target and the output values. Calculated R values are in the range of 0,9979-0, 9999 and MAPE values are in the range of 0,69-2,55%. According to results, the usage of ANNs is highly recommended to predict the engine's emissions of a diesel engine fueled with pure soybean biodiesel.