Black-box modeling of proton exchange membrane fuel cell generators

Proton exchange membranes are one of the most promising fuel cell technologies for transportation applications. Considering this aim, a simulation model of the whole fuel cell system is a binding milestone. This would lead in the optimization possibility of the complete vehicle (including auxiliaries, output electrical converter and their control laws). So a "minimal behavioral model" of a fuel cell system, able to evaluate the output variables and their variations, is presented in this paper. In a fuel cell system, electrical quantities are dependent both on the fuel cell power level and also on the auxiliaries and gases used. Artificial neural networks are here used to model the whole system. To carry out an efficient model, it is necessary to identify the number and the nature of input and output parameters. The proposed model is implemented on Matlab/Simulink® software and will be integrated in a complete vehicle powertrain. Thus, it is possible to carry out the development and the simulation of the command laws intended to manage the energy transfers aboard fuel cell vehicle.