Abstract
This paper introduces the stacking scheme for the flat-chip solid oxide cells (FCSOCs). The FCSOCs are easy for manufacturing and stacking, combining the advantages of planar SOCs and tubular SOCs. The FCSOC stacks are sealed at the cold-ends (around 150℃) of single cells, enabling an easy, reliable, and flexible sealing with silicone sealants.
R&Ds reveal that, however, single cells in an FCSOC stack exhibit non-uniformity caused by the distribution of operating conditions. Fully-coupled 3D stack models validated with experiments are powerful tools for investigating the non-uniformity, which, however, require high computation costs, especially for large stacks with up to about 100 cells. Inspired by that the stack model’s focus is the stack-level non-uniformity but not the cell-level non-uniformity, it is possible to simplify the fully-coupled model by ignoring cell-level non-uniformity. Therefore, a data-driven method, called the adaptive polynomial approximation (APA), is used to build surrogate models of single cells. The cell-level surrogate models are then integrated into the stack-level model to form a hybrid model that avoids computing the current distribution and fuel-side mass distribution inside the single cells. The computational cost is reduced, making it possible to simulate the models of large stacks. The simplification error is analyzed. Based on the simplified stack models, the design and operation factors inducing the non-uniformity among single cells are investigated and optimized.