Biogas engine performance estimation using ANN

Abstract

Artificial neural network (ANN) method was used to estimate the thermal efficiency (TE), brake specific fuel consumption (BSFC) and volumetric efficiency (VE) values of a biogas engine with spark ignition at different methane (CH4) ratios and engine load values. For this purpose, the biogas used in the biogas engine was produced by the anaerobic fermentation method from bovine manure and different CH4 contents (51%, 57%, 87%) were obtained by purification of CO2 and H2S. The data used in the ANN models were obtained experimentally from a 4-stroke four-cylinder, spark ignition engine, at constant speed for different load and CH4 ratios. Using some of the obtained experimental data, ANN models were developed, and the rest was used to test the developed models. In the ANN models, the CH4 ratio of the fuel, engine load, inlet air temperature (Tin), air fuel ratio and the maximum cylinder pressure are chosen as the input parameters. TE, BSFC and VE are used as the output parameters. Root mean square error (RMSE), mean absolute percentage error (MAPE) and correlation coefficient (R) performance indicators are used to compare measured and predicted values. It has been shown that ANN models give good results in spark ignition biogas engines with high correlation and low error rates for TE, BSFC and VE values.