Fuel cell model reduction through the spatial smoothing of flow channels

A commonly invoked postulate in fuel cell modeling involves solving for a two-dimensional (2D) instead of a three-dimensional geometry (3D). Often, however, this postulate affects the fidelity of model predictions, since not all geometrical features are captured. To achieve such a reduction in dimensionality, we introduce a methodology based on spatial smoothing over the flow channels in the flow field, coupled with correlations that account for variations in pathways due to ribs. The derived mathematical framework is demonstrated on a flow field comprising parallel flow channels, and verified for a detailed, mechanistic fuel cell model: overall, good agreement is achieved. Finally, we highlight how one can account for other types of flow channels and how a spatially smoothed 2D model that captures the main geometrical design parameters of a 3D counterpart can be solved in seconds. The latter opens up avenues for mechanistic modeling of large fuel cell stacks.