Neural Network for Performance Prediction of Biogas Production from Municipal Solid Wastes

Abstract

The study includes identifying different parameters affecting biogas production from municipal solid wastes (MSW). Their performance is measured for optimizing the productivity of methane content in biogas. The experiments have been carried out by collecting different samples from municipality garbage and tested under standard measuring techniques. The exploration of parametric optimization is conducted by the use of artificial neural networks models and helps to predict the enrichment of biogas generation from MSW using different input parameters. The results analysis demonstrates the feed-forward backpropagation model in the neural network with input neurons, and its hidden unit produced an efficient predictive output. In order to change the simulation, it designs the neural system by following various performance functions such as MSE, MSEREG, and SSE associated with different transfer functions as defined Tansig, Logsig, and Purelin to valid the outcomes. An ANN model was developed using 80% of the experimental results for training and used ANFIS model to simulate the given data and validate the result.