Uniform reactant distribution within the bipolar plate of polymer electrolyte membrane fuel cells is critical to reach targeted lifetimes, requiring well designed gas distribution zones. While various distribution zone topologies are in use, the fundamental design mechanisms and their interdependencies with manifold and flow field design are not fully characterized. In this study, we utilize a dataset of 2048 computational fluid dynamics (CFD) simulations of circular dot matrix gas distribution zones to identify the correlations governing flow distribution. We introduce space velocity v s and pressure drop ratio Π between flow field and distribution zone pressure drop as primary design metrics. Our results show that Π > 3 is required to achieve a distribution error of less than ten percent, ensuring the flow field pressure drop remains the dominant resistance. Using Symbolic Regression, we derive a four-step framework of human-readable empirical formulas to estimate distribution zone pressure drop and relative flow distribution deviations. Validation against an independent dataset confirms the model's stability. Finally, we demonstrate the proposed approach through a case study for an active area of 250 cm 2 , sequentially investigating the external geometry, manifold dimensions and internal geometry to meet the Π > 3 criterion. The approach provides a computationally efficient addition to iterative CFD simulations for optimizing reactant distribution and ultimately reach targeted lifetimes.
Schuckert et al. (Thu,) studied this question.