Artificial neural networks to predict the performance and emission parameters of a compression ignition engine fuelled with diesel and preheated biogas-air mixture

In recent days, artificial neural network (ANN) is seen as a potential tool to perform mathematical modeling and prediction. In this analysis, the Levenberg-Marquardt backpropagation training algorithm is used to map the actual and the predicted value with the tansig activation function. The ANN model is developed to predict the engine performance and emission parameters for varying intake biogas-air mixture temperatures ranging from 55 +/- 5 to 85 +/- 5 degrees C under various load conditions. Around 70% of the total experimental data has been used for training, 15% for validation, and 15% for testing. The findings show higher coefficient of correlation (R, R-2 and adjusted R-2 values are in the range of 0.96-0.99) and lower mean square error (MSE = 0.0003515-0.00544501) for the developed ANN models. This study proves that ANN is an excellent tool for determining and optimizing the performance and emission parameters of compression ignition engines fuelled with alternative gaseous fuels. [GRAPHICS] .

Automated summary

Artificial neural network (ANN) is widely used for mathematical modeling and prediction, with Levenberg-Marquardt backpropagation training algorithm effectively mapping actual and predicted values to determine engine performance and emission parameters. The study shows that developed ANN models have higher correlation coefficients and lower mean square errors, providing an efficient way to optimize performance and emission parameters.