Multi-objective optimization of proton exchange membrane fuel cell flow channel baffle based on artificial neural network and genetic algorithm

This research aims to improve the performance of proton exchange membrane fuel cells (PEMFCs) by multi-objective optimization of the flow channel structure. The accomplishment involves employing a three-dimensional multi-phase PEMFC model for performance estimation, optimizing it using artificial neural networks (ANN), and validating the water removal capacity through the volume of fluid (VOF) method. By investigating the simulation of different necking acceleration structures with their arrangements of PEMFC, it is found that U-shaped array arrangement performed better compared to other structures. The optimization of key structural features employs ANN to attain improved performance, specifically targeting enhancements in net power density and the uniformity index of oxygen. The optimized case predicts a net power of 0.8659 and simulates a net power of 0.8675 with an error of 0.184 %, an increase of 3.84 % compared to the conventional (Conv) straight channel simulation result of 0.8354. The predicted oxygen uniformity index is 0.6932, while the simulated oxygen uniformity coefficient is 0.6919, resulting in a slight error of 0.188 %. It represents a 1.91 % improvement compared to Conv’s simulation result of 0.6789. Furthermore, according to the VOF model, the presence of turbulence caused by the structure significantly decreased the saturation of liquid water, and the drainage period of the optimized case was nearly half as short as that of the straight channel configuration.