Abstract
Suitable flow field designs significantly affect the performance of high temperature proton exchange membrane (HT-PEM) fuel cells. Previous designs both in terms of shape and size optimisation are still inadequate in the category of experiential design method, depend on the experience of designers and lack of explainable theories, leading to uncertainties in the design process. To address this issue, a novel flow field design method for innovative flow fields using combined topology and surrogate models is proposed, in which a topology model is built to optimise to obtain the basic shape of flow field, then a numerical model is established to train a surrogate model for the optimum design. Finally, a radial basis function (RBF) model is built to determine the optimum channel parameters. With this method, a 25 cm2 leaf-like flow field with a diagonal inlet outlet is designed and investigated which outperforms the initial serpentine design with a 7.95% increase in current density, a 8.47% increase in actual power, and a 97.8% reduction in pressure drop at 0.50 V. This novel flow field design method combines the advantages of the topology and surrogate models in shape and size optimisation, respectively. It significantly reduces the design threshold and can be widely used in innovative flow field design in energy and other engineering fields.
•A nature-inspired leaf-like flow field is designed for HT-PEM fuel cells.•Topology and surrogate models are combined for shape and geometry optimisation.•Novel leaf-like flow field achieves improved performance over traditional serpentine design.•Current density and real power are improved by 7.95% and 8.47%, respectively, at 0.5 V.•Pressure drop is reduced by 97.8% in comparison with serpentine flow field.