Modeling of Proton Exchange Membrane Fuel Cell Stack and Start Strategy Optimization with a Multi-Objective Genetic Algorithm

Abstract

This paper proposes a non-dominated sorting genetic algorithm II (NSGA-II) for optimizing the startup strategy of a proton exchange membrane fuel cell (PEMFC) stack to improve the dynamic response capability, output voltage, and net power. First, a Simulink model of the PEMFC stack including the anode module, cathode module, water transfer module, output voltage module, and output net power module is established, and the accuracy of the stack model is verified through experiments. The three performances are then optimized simultaneously based on NSGA-II. The results show that the optimized start-up loading strategy results in a PEMFC stack that outperforms the base model in steady-state voltage, undershoot percent, and net power these three indicators with the same response time, demonstrating the success of the method in solving multiple optimization problems. This study presents an effective approach for the multi-objective optimization of the PEMFC stack, which is of guidance for engineering practice.