Data based modeling of a PEM fuel cell for control design purposes

In this paper we consider the problem of data based modeling of Proton Exchange Membrane (PEM) fuel cells for optimal control purposes. Similar to a combustion engine the fuel cell needs some control algorithm for the air path and cooling unit. For an optimal control design a mathematical model of the system is essential. In contrary to a common solution we present a data based method for modeling the PEM fuel cell. Furthermore we discuss the steps in order to achieve a well tuned model. The mathematical model describes the dynamical behavior from the control and the disturbance input to the performance and measurement outputs. All these quantities will be discussed throughout the paper. The model quality is validated using validation data and control strategies are presented.