Multiple model predictive control for oxygen starvation prevention of fuel cell system

The purpose of this work is to prevent oxygen starvation of fuel cell system with the compressor surge and choke constraints considered. It has been already known that the compressor inertia causes the oxygen excess ratio to drop rapidly when the current drawn out of the fuel cell steps up dramatically. Thus we introduce a load actuator model to extract the load current out of the fuel cell as a performance variable. Then we are able to regulate the input of load actuator not only to smooth the transitions of load current to prevent the oxygen starvation, but also to track the desired current of fuel cell. Moreover, air supply system of fuel cell has strong nonlinear behaviors while working in a wide range. Thus a multiple multivariable model predictive controller has been introduced based on linearized models obtained from the fuel cell´s nonlinear model linearized at operation points. The simulation results have demonstrated that the proposed control strategy is effective when applied to the fuel cell´s nonlinear system.