Multi-objective optimization of helical baffle flow field structure for fuel cell under multiple performance indicator constraints

To enhance the output performance and durability of fuel cells (FC), a novel helical baffle flow field (FF) is proposed, which can enhance mass transfer in both in-plane and through-plane directions. To improve the efficiency of FF structure design and optimization, and to further optimize FF structure to obtain the best mass transfer capability, an efficient and high-precision optimization analysis process is constructed. The sample data set of the baffle is obtained using optimal Latin hypercube sampling. The established numerical model was used to calculate the FC performance under different baffle structures, and artificial neural network surrogate models were trained to accurately predict two performance indicators under two relative humidity (RH) conditions, namely area and mass power density. The baffle structure parameters and output values that maximize the two performance indicators are obtained for different RH cases. Relative to the original FF, the new FF can make a 5.12 %∼12.68 % improvement in the FC performance. Since the two performance indicators are contradictory and constrained, to obtain the baffle structure parameters with better performance under different conditions, the FC's comprehensive performance under multi-objective performance constraints is optimized using a genetic algorithm to obtain the optimal baffle structure parameters under the specified conditions.