Thermodynamic modelling and intelligent control of fuel cell anode purge

Dead-ended anode Polymer Electrolyte Membrane Fuel Cell (PEMFC) is promising in small-scale power generation applications due to its simple structure, high reliability, and low price. However, it necessitates the anode purge to remove the impurities accumulated in the anode channel, making the system operation suffer from the sudden anode pressure drop during the purge. In most cases, a widely used solution of many fuel cell manufacturers is to open up the anode inlet valve regularly to recover the fuel cell terminal voltage. Although simple, the performance would degrade in case of the operation condition change. To this end, this paper proposes a modified Iterative Learning Control (ILC) strategy, which can repeatedly improve the control performance by learning from the previous actions. A mathematical model is developed to analyse the variations of the voltage and the anode pressure, by taking into account the membrane water transfer and the nitrogen permeation. The purge action is triggered when the single-cell terminal voltage drops by 0.1 V. The dynamics of the anode pressure is investigated under different openings of the purge valve. The simulation results have demonstrated the efficacy of the modified ILC strategy for anode pressure regulation in mitigating the periodic purge influence and the model uncertainty.