Lung-inspired hybrid flow field to enhance PEMFC performance: A case of dual optimization by response surface and artificial intelligence

The design and optimization of flow field play a crucial role in the development of proton exchange membrane fuel cells (PEMFC). This study presents a modular tri-layer lung-inspired hybrid flow field (LHFF) design that incorporates 2D and 3D flow field advantages. The key structural parameters of LHFF mainly encompass G, D, and S of reactant distribution layer and A of directional transport layer. The LHFFs with different G have been investigated, and the G = 2 LHFF exhibits a 16.55% enhancement in maximum net power density compared to conventional parallel flow field. Then the response surface methodology (RSM) and artificial intelligence methodology (AIM) have been employed to optimize the D, S, and A structure parameters of LHFF to determine the optimal inlet position of water removal layer. The LHFFs optimized by RSM and AIM show a further increase in maximum net power density by 3.58% and 4.10%, respectively. The optimized LHFFs achieve a trade-off among species distribution, water management, and pressure drop, with high consistency between numerical and experimental results. It demonstrates the reliability of artificial intelligence in optimizing PEMFC flow field. Therefore, the optimization strategies presented here hold a promising solution to improve the flow fields in other electrochemical systems.