Hybrid model of fuel cell system using wavelet network and PSO algorithm

Fuel cell power system has attracted significant attentions from many researchers. A fuel cell system model, which is accurate and applicable in engineering practice, plays a key role in controller design, fault diagnosis and power management for fuel cell power systems. In this paper, a hybrid model of a Polymer Electrolyte Membrane Fuel Cell (PEMFC) power system is proposed. The hybrid model contains a mechanism submodel and a black-box submodel. The adjustable parameters in the mechanism submodel are optimized using PSO algorithm. The black-box submodel is expressed in NARX form, which is approximated by a wavelet network using the real test data. From practical experiments, it is shown that, the hybrid model acceptably approximates the practical system. This hybrid PEMFC model can be used for mass flow controller design, fault diagnosis and power management.