Real-time optimization of net power in a fuel cell system

This paper proposes the use of extremum seeking algorithms to control the air flow into the cathode of a fuel cell system. Through experimental testing of an eight-cell, hydrogen-fueled polymer electrolyte stack, it shows that extremum seeking is a viable method to control the air flow rate. The algorithms do not rely on knowledge of system modeling parameters, and adapt to changes in those parameters that occur due to disturbances and degradation. An additional penalty on the ohmic resistance is proposed for the objective function to increase the life of the cell.