Dynamic modeling of SOFC stack using ARX identification algorithm

This paper reports a dynamic modeling study of solid oxide fuel cell (SOFC) stack using autoregressive with exogenous input (ARX) identification algorithm. It is well known that SOFC is a nonlinear, multi-input and multi-output system that is hard to model by traditional methodologies. By far, the most existing dynamic models are based on conservation laws which are too complicated to be used for control system design. For the purpose of dynamic simulation and control, this paper presents two black-box ARX identification models relating output voltage to stack current and to hydrogen flow rate respectively. The validities of the established ARX models are verified when they are used to predict the output voltage responses to: 1) step change in the stack current, and 2) step change in the hydrogen flow rate. The simulation results reveal that it is feasible to establish dynamic models of the SOFC stack using ARX identification algorithm, and the ARX models presented are accurate and valid. Furthermore, based on these ARX models, valid control strategy studies such as predictive control, robust control can be developed.