Multi-objective optimization of proton exchange membrane fuel cell geometry and operating parameters based on three new performance evaluation indexes

Abstract

Key issues such as performance, durability, and cost limit the large-scale commercial application of the proton exchange membrane fuel cell (PEMFC). The distribution of reactants in the PEMFC is not uniform, which will make the distribution of current density uneven, and then affect its durability. Different from the output power of the PEMFC as the performance evaluation index, the flow uniformity of the reactants, diffusion flux, and ohmic resistance are proposed as the comprehensive performance evaluation index. Firstly, the physical model of three-dimensional two-phase non– isothermal PEMFC was established. Secondly, six parameters were selected and the first-order finite-difference sensitivity of each parameter of the flow uniformity of the reactants and the diffusion flux was calculated. Five parameters were selected as optimization objects through sensitivity analysis. Then, the five parameters were scanned and parameterized, and the obtained data were used to train the neural network as a surrogate model. Finally, based on the fast non-dominated sorting genetic algorithm with elite strategy (NSGA-II), three PEMFC performance indexes including the flow uniformity of the reactants, the diffusion flux, and the ohmic resistance were optimized simultaneously. In this study, from Pareto solutions obtained by multi-objective optimization, the point with the largest diffusion flux is selected as the optimal solution, that is, the flow uniformity of reactants is 0.458411026, the diffusion flux is 0.049903274, and the ohmic resistance is 0.047476574. The optimized model has a higher polarization curve, which means better fuel cell performance. The flow uniformity of the reactants determines that the current density of the optimized model is more uniform than that of the basic model, which makes the fuel cell have better durability. Therefore, the fuel cell performance optimization method proposed in this paper is feasible. The three new performance evaluation indexes can provide a reference for the study of fuel cell performance and durability in the future.