Artificial Neural Network Based Prediction of Biogas Generation through Wastewater Treatment Using Anaerobic Baffled Reactor

Abstract

A novel prediction method based on Levenberg–Marquardt (LM) algorithm and back propagation neural network (BPNN) is proposed to predict biogas generation through textile wastewater using anaerobic baffled reactor (ABR). Then, the proposed BPNN (eight input nodes consisting of average influent COD, flow rate, influent COD, pH, HRT, OLR, VFA and VSS; one output node consisting of biogas generation through textile wastewater using ABR) is used. Results indicate that BPNN is an effective prediction method for biogas generation through textile wastewater using ABR. The values of squared correlation coefficient (R²) and mean absolute percentage error (MAPE) are 0.9846, and 2.3117. Statistical information exhibit that BPNN has excellent forecast ability and high precision, and the correlation between predicted and experimental data is acceptable.