Experimental dataset: Two-Stage Model Predictive Stack Control for PEM Fuel Cell Systems
Description
Proton-exchange membrane fuel cells (PEMFCs) play a pivotal role in decarbonizing the transport sector due to their high efficiency and zero local emissions. This work proposes an embedded, nonlinear, two-stage model predictive control (MPC) that optimally allocates the stack supply from the air, hydrogen, and coolant subsystems. A physics-based process model of the stack captures the dominant dynamics of membrane hydration, temperature, and reactant pressures, representing key internal states that influence performance and durability. The two-stage structure separates the optimal economic planning from the real-time tracking MPC, (1) ensuring constraint-compliant regulation of the considered states under highly dynamic operation, (2) maintaining efficient stack operation, and (3) resolving the over-actuated nature of the multi-input system. Experimental validation on a full-scale PEMFC test bench demonstrates efficiency gains of up to 8% and reduced reactant consumption compared to a nominal stack control. At the same time, membrane temperature and water content remain within prescribed limits.
Files
Institutions
- RWTH Aachen UniversityNorth Rhine-Westphalia, Aachen