Structural gradient optimization of diffusion layer based on finite data mapping method for PEMFC performance improvement

Abstract

Optimized design of the full structure is a key solution to improve the overall performance of proton exchange membrane fuel cells and advance their commercialization. This paper established a three-dimensional mathematical model of PEMFC, and obtained a series of simulation data collection under different structural parameters. On this basis, we derived the mapping relationship between the thickness of the diffusion layer and the output power using response surface analysis, and then obtained the optimal structural parameters by the BP neural network. The optimization provided a new type of gradient diffusion layer used in the PEMFC, i.e., a tapered diffusion layer. Compared with the fuel cell using a conventional diffusion layer under the same flow channel structure and operating conditions, the results show that the average power of the fuel cell using a gradient diffusion layer has been increased by 3.5 %. The overall heat transfer capacity increased by 10.3 %, and the oxygen utilization increased by 8.7 %. Finally, this paper analyzed the multi-physics field synergistic performance of PEMFCs with different diffusion layer structures, and found that the fuel cell with tapered diffusion layer has better synergistic performance in terms of temperature and concentration fields, and temperature and velocity fields.