A survey of PEM fuel cell system control models and control developments

Proton Exchange Membrane (PEM) fuel cell system performance can be significantly improved with suitable control strategies. Control appropriate models of the fuel cell stack and balance of plant are presented along with current control research. Fuel cell stack models are zero dimensional and range from simple empirical stack polarization curves to complex dynamic models of mass flow rates, pressures, temperatures, and voltages. Balance of plant models are also zero dimensional and can be used individually to build a complete system around a stack. Models of this type are presented for the air compressor air blower manifolds, reactant humidification, fuel recirculation, air cooling, and stack cooling. Current control work is surveyed with regard to feedforward, feedback, observers, optimization, model prediction, rule based, neural networks, and fuzzy methods. The most promising fuel cell stack model is evaluated. Additionally, improvements to the balance of plant models are recommended. Finally, future control work is explored with a desire for system control that leads to greater output power.