Neural network controller for PEM fuel cells

This paper considers that any system of production is subjected permanently to load steps change variations. In our case, we consider a static production system including a PEMFC is subjected to variations of active and reactive power. The goal is then to make so that the system follows these imposed variations. We were thus interested in control of the powers by using the neural networks controllers. Simulation requires the modelling of the principal element (the Proton Exchange Membrane Fuel Cell) in dynamic mode. The model used is that described by J. Padulles with a modification concerning the addition of losses of activation and concentration. For the neural network, various network design parameters such as the network size, Levenberg-Marquardt training algorithm, activation functions and their causes on the effectiveness of the performance modeling are discussed, the Quasi-Newton neural networks was described. Results from the analysis as well as the limitations of the approach are presented and discussed.