Prediction of Direct Methanol Fuel Cell Using Artificial Neural Network

Abstract

An artificial neural network has been applied for performance prediction of direct methanol fuel cell. In order to optimization of neuron number in hidden layer, average relative deviation of neural network versus number of neurons has been plotted. Results showed 5-6-2 is the best topology of the neural network. Present study showed the proposed model has an average relative deviation about 2.95 % and 3.4 % respectively for training and testing subset.