3D+1D modeling approach toward large-scale PEM fuel cell simulation and partitioned optimization study on flow field

A 3D+1D PEM fuel cell model is developed in order to implement large-scale simulation with enhanced calculation efficiency. The model consists of the three-dimensional (3D) part and the one-dimensional (1D) part, which are related by adding two extra layers in the middle of the 3D computational domain. Bipolar plate (BP), gas channel (GCH) and gas diffusion layer (GDL) along with the extra layer (EL) form the 3D computational domain. Other components, micro-porous layer (MPL), catalyst layer (CL) and membrane (MEM) are treated as 1D computational domain and integrated into the grids of extra layer. The 3D sub-model solves conservation equations and provides scalar data for the 1D submodel to obtain solutions of flux equations, in turn updating parameters of the 3D domain to proceed iterations. The 3D+1D model considers the strongly-coupled physicochemical phenomena comprehensively inside PEM fuel cell including mass transfer (reactant gas and liquid water), electrochemical reaction, membrane water balance and heat transfer. The trade-off between model accuracy and calculation efficiency is evaluated with detail by comparing the simulation results and time cost of the 3D+1D model with those of the whole 3D model. The calculation speed is found to be greatly boosted via the 3D+1D approach and acceptable accuracy is obtained at the same time. Specifically, the simulation time can be shortened by 20 folds for the large-scale case in this study. Then, three flow field designs are compared on a 345 cm2 PEM fuel cell domain using the proposed 3D+1D model, namely the parallel-serpentine design (PS design), the parallel design with dots in the distribution zone (PD design) and the parallel design with dots in the distribution zone and waves in straight-flow zone (PDW design). Owing to the addition of wavy structure, the PDW design gives excellent performance under high current density with low external humidification and stoichiometric ratio due to enhanced gas convection and self-humidification effect. This proves the feasibility and potential of partitioned optimization design on PEM fuel cell flow field, meanwhile emphasizing the suitability of the 3D+1D modeling approach for occasions where the full morphology of flow field layout should be considered. (C) 2020 Elsevier B.V. All rights reserved.