Predictive hydrogen pressure control for an automotive proton exchange membrane fuel cell system

Maintaining a consistent hydrogen supply and pressure to the fuel cell system with recirculation is crucial. The small volume in an automotive hydrogen supply system, combined with comparatively high gas flows, creates high dynamics in the hydrogen supply pressure. Even a slight dead-time of the electric actuation, combined with the hysteresis of the actuator, can result in strong nonlinearities in the plant, a factor underestimated in simulation-based literature. As main control objectives, a stable pressure must be assured, and the gaseous purge must be assessed. We propose a Smith Predictor by modeling the injector valve, capturing gas dynamics, and hysteresis. Compared to a full-scale test bench, the PI control leads to high oscillating pressure. In contrast, our proposed approach demonstrates smoother stationary and dynamic pressure profiles with a clear purge detection through effective dead-time compensation. • Real-time capable prediction of the anode pressure in a fuel cell system with recirculation. • Enhanced pressure control with a Smith Predictor, incorporating hysteresis and gas dynamics. • Reliable purge evaluation through the hydrogen supply pressure response.