Surrogate-Assisted Multi-objective Optimization of a Serpentine-Flow-Field Proton-Exchange-Membrane Fuel Cell for Power and Oxygen Uniformity

Serpentine flow fields enhance reactant transport in proton-exchange-membrane fuel cells (PEMFCs), but pressure loss and nonuniform oxygen supply can reduce useful power. This study develops a surrogate-assisted framework for the coupled optimization of flow-field geometry and operating conditions, with net power and cathode oxygen-distribution uniformity as competing objectives. A three-dimensional CFD model was assessed against published polarization data, and 125 cases were generated over five design variables. Four surrogate models were evaluated using repeated five-fold cross-validation, and response-specific models were selected for the two objectives. The selected surrogates were coupled with NSGA-II to obtain the Pareto front, and entropy-weighted TOPSIS identified a compromise design. CFD re-evaluation yielded a net power of 1.0412 W and an oxygen-concentration standard deviation of 0.4154 mol/m3. Compared with the maximum-net-power design, the compromise solution reduced oxygen variability by 42.19% with a 9.71% power penalty. Field analysis further related this trade-off to reactant transport, porous-layer water saturation, pressure loss, and current-density distribution. The framework provides an efficient and physically interpretable approach for balancing PEMFC power output and oxygen-distribution uniformity.