An Accurate and Reliable Computational Approach for the Mechanical Stress Analysis of Planar SOFC Stack

Mechanical damage induced by operation under high stress conditions is one of the major factors limiting the long-term stability of SOFCs. Stress analysis is therefore important for the development the SOFC technology. To date, it is a common practice to focus the stress analysis at some interested component of an SOFC stack, e.g., PEN, glass-ceramic sealant, interconnector plate, etc. This approach is questionable as it is very difficult or even impossible to properly define the boundary conditions for the isolated component. To conduct a reliable stress analysis, it is desirable to carry out the stress computation on the entire stack. For that purpose, it is important to devise algorithm that confines the computational resource requirement at an affordable level. Here we report a grid matching method that enables the stress evaluation of the whole multi-cell SOFC stack. The basic idea of the grid matching scheme is to assign adequately fine grids for the interested stack components and their neighboring materials, while distant material components are assigned with coarse grids to reduce the computational burden. With appropriate free stress temperatures of materials, initial state of stress, material properties, boundary conditions and temperature profile of the SOFC stack, stress distributions inside the interested stack components can be calculated by solving the elastic matrix.

The method is tested with a 5-cell planar stack. The 5-cell stack is chosen so that the stress analysis with fine grids for all stack components is also affordable. The stack is of practical dimensions and material components such as the cells, frame, interconnects and glass-ceramic sealants, etc. For typical material combinations of SOFC stack and operation conditions, the stack temperature profiles were generated by a high resolution multi-physics numerical model. The test shows that the error for the maximal stress using coarse grids is limited to the order of 10% in comparison with that with fine grids for all stack components. In other words, the grid matching scheme is confirmed to be a computationally efficient and satisfactorily accurate method for the stress analysis. However, the maximal stress of a stack component evaluated with the commonly adopted method using a free boundary condition for the interested component is one order of magnitude smaller than the correct result. On the other hand, the stress of a stack component evaluated with a stiff boundary condition can be a few orders of magnitude higher than the true result. Clearly, the conventional stress analyzing approach with isolated stack components is prone to produce misleading results.

The proposed stress analysis method is valuable for reliably identifying stack components susceptible to mechanical damage. It can be used to improve the design of SOFC stack and to make informed selections of operating conditions.