Health-conscious MPC for PEM fuel cells considering main degradation mechanisms of cathode catalyst

Abstract

The integration of fuel cell systems into hybrid electric vehicles presents unique challenges due to the highly dynamic power trajectories typical of automotive applications. The dynamic operation of fuel cells leads to performance degradation limiting the lifetime of fuel cells. This paper proposes a model predictive control strategy that optimizes the balance between fuel cell longevity, efficiency, and performance on the system level by adapting current, hydrogen flow, airflow, and the dew points at the inlet of the electrodes. Using a non-linear, physically motivated model that incorporates the key degradation mechanisms of the cathode catalyst layer, this approach provides a framework to optimize the operating strategy of the fuel cell system. By leveraging the information from the physical model, the controller can make health-conscious control decisions outperforming traditional and rule-based strategies.