Multi-objective optimization of orifice-shaped cathode flow field design in polymer electrolyte membrane fuel cells

It is well known that the orifice-shaped flow field design of the Polymer Electrolyte Membrane (PEM) fuel cell has advantages in oxygen supply and water removal. This study introduces research that optimizes the key design variables of the orifice-shaped flow field using the previously developed multi-scale multi-phase PEM fuel cell model. Key design variables include reduced channel depth, width, length in the orifice channel region, and the orifice-to-total flow field ratio, optimized to maximize cell voltage (Vcell), and minimize pressure drop (ΔP) and the standard deviation of the current density distribution (ISD). Data for these variables were generated through three-dimensional PEM fuel cell simulations and used to train an Artificial Intelligence-based Multi-layer Perceptron (MLP) model. The trained MLP model, in conjunction with the Non-dominated Sorting Genetic Algorithm-II (NSGA-II), analyzed the impact of design variables to determine the optimal points. Through multi-objective optimization study, it has been successfully demonstrated that compared to the baseline orifice-shaped flow field design, at a current density of 2.5 A/cm2, Vcell can be increased by 13 mV, ΔP can be reduced by 47.85 Pa/cm, and ISD can be decreased by approximately 0.119 A/cm2, indicating that further performance improvement is possible.