Soot emission prediction of a waste-gated turbo-charged DI diesel engine using artificial neural network

摘要

This study is about soot emission prediction of a waste-gated turbo-charged DI diesel engine using artificial neural network (ANN). For training the ANN model, six ranges of experimental data in previous study were used, and one range of data was kept for testing the accuracy of ANN predictions. The input parameters for the ANN are inlet manifold pressure, inlet manifold temperature, inlet air mass flow rate, fuel consumption, engine torque, and speed. Output parameter is the density of soot in the exhaust. The results show the ANN approach can be used to accurately predict soot emission of a turbo-charged diesel engine in different opening ranges of waste-gate (ORWG). Root mean-squared error (RMSE), fraction of variance (R2), and mean absolute percentage error (MAPE) for predictions were found to be 1.19 (mg/m~3), 0.9998, and 6.4%, respectively.