Predictive coordination control algorithm for cooling system of a 140-kW proton exchange membrane fuel cell stack based on decoupling of component function

Operating temperature dominates performance and durability of proton exchange membrane fuel cell, and optimal control of operating temperature is essential for high performance and longer lifespan. Hence, this paper proposes predictive coordination control algorithm for cooling system of fuel cell stack. Lumped-parameter dynamic model of cooling system is established to quantify time constant of stack temperature and influence of time latency on temperature at stack inlet and outlet. Isolation-based analysis of heat exchange between heat generation and dissipation decouples component functions by considering coolant recirculation transport, and relationship between actuators and control targets is determined. Time latency and radiator heat dissipation are compensated predictively, and predictive coordination control algorithm is developed for three temperature objectives and three actuators. Simulation shows that proposed algorithm reduces temperature overshoot by more than 50% and settling time by 30% compared with PI controller, with relative errors within ±1.23% and ±2.05% for coolant temperature at stack inlet and outlet respectively. Experiment demonstrates that maximal overshoots are within 0.6% and 0.8% respectively, while maximal steady-state relative errors are within 0.4% and 0.5% respectively, with settling times below 28 s and 24 s respectively. Proposed algorithm effectively improves steady-state accuracy and transient response of cooling system.