Towards the 3D modeling of the effective conductivity of solid oxide fuel cell electrodes - II. Computational parameters

摘要

The effective conductivity of a thick-film solid oxide fuel cell (SOFC) electrode is an important characteristic used to link the microstructure of the electrode to its performance. A 3D resistor network model, the ResNet model, developed to determine the effective conductivity of a given SOFC electrode microstructure was introduced in earlier work (Rhazaoui et al., 2013. Chem. Eng. Sci. 99, 161-170). The approach is based on the discretization of each structure into voxels (small cubic elements discretizing the microstructure). In this paper, synthetic structures of increasing complexity are analyzed before an optimum discretization resolution per particle diameter is determined. The notion of Volume Elements (VEs), based on the Volume-Of-Fluid method, is then introduced in the model to allow larger structures to be modeled and is used to analyze synthetic structures as well as an experimental Ni/10ScSZ electrode. The behavior of the model output is examined with respect to increasing aggregation resolutions for several synthetic microstructures of varying compositions, with the aid of extracted skeletonized paths of charge-conducting pathways. A ratio of VE size to voxel size of 5 is shown to be appropriate. The first comparison of calculated and measured effective conductivities is presented for the Ni/10ScSZ electrode considered. The computed effective conductivities are found to be consistent with observations made on the microstructure itself and skeletonized network paths, and support the findings of earlier work with respect to the minimum sample size required to characterize the entire anode from which it is extracted.