Design and optimization of periodic waveform flow channels for HT-PEMFC using a surrogate-based optimization framework

High-temperature proton exchange membrane fuel cells (HT-PEMFCs) are constrained by limited mass transfer and non-uniform reactant distribution in straight flow channels. This study proposes an optimization framework for waveform flow channels by integrating a 3D multiphysics model with simulated annealing-based stratified maximin design (SA-SMD), Gaussian process regression (GPR), and genetic algorithm (GA). Four periodic waveform channels were compared, and the rectangular waveform channel exhibited the highest performance among the investigated configurations by enhancing oxygen transport and distribution. The GPR surrogate model accurately predicted the net power density of rectangular waveform channels, and the optimized configuration with a waveform center amplitude of 0.6307 mm and a waveform number of 16 increased the net power density by 4.16% compared with the straight flow channel. The proposed framework provides an efficient strategy for designing high-performance HT-PEMFC flow channels.