A PROPOSED ARTIFICIAL NEURAL NETWORK MODEL for PEM FUEL CELLS

Abstract

Fuel cells convert the chemical energy directly to the electrical energy and hence they are a very favorable alternative energy source. In the literature, there are many studies related to the modeling of fuel cells. Artificial neural networks (ANNs) is one of the promising techniques for modelling nonlinear systems such as fuel cells. The proposed model in this study doesn't require many parameters like other studies. Firstly, training and testing data was obtained the dynamic model of a PEM fuel-cell. Then, proposed ANN model outputs are compared with dynamic model ouputs Simulation results shows that the proposed ANN model can be used very efficiently for PEM fuel-cells without using many parameters.