A Thermal-Electrochemical Model of a Fuel Cell Stack Block

RRFCS is developing SOFC fuel cell stack and system technology for a hybrid power generation system. RRFCS is also developing state of the art numerical prediction tools that will be used to support fuel cell product design. One such tool, the Block Thermal Model, is based in the FEMAP-TMG thermal model generator package. This paper describes the constituents of the block and the physics modelled, presents the initial results and states the computer resources used. Full validation could be a subject for later publication. In the RRFCS stack block, aerothermal physics in the anode and cathode gas flows is coupled with electrochemistry, distributed throughout a highly repeated three-dimensional solid geometry. The requirement was to create numerical modelling capability to accurately solve the main coupled behaviours throughout the entire domain of the block, as a function of externally applied conditions. The key distributions are thermal, electrical, chemical and hydraulic. The strength of the couplings determines the robustness of the design. The geometry, mesh density and fidelity required create considerable scope for model size to be enormous, with all the attendant issues around model setup, solution time and stability and post-processing. A finite element (FE) pre- and post-processor with a dedicated thermal solver extended by user-defined subroutines provides detailed thermal mapping and will be able to capture the coupled physics interactions. The FE mesh includes every component in the block including the end fittings and fuel pipes, as the sensible energy in the circulating fuel has significant thermal effects particularly at inlet. These components are less exposed to cooling airflow and will have significant effects in dynamic solutions. In order to capture the non-uniform distribution of fuel cell heat release, anode and cathode mass balances and electrochemical calculations are required at adequate resolution. Distributions of electrical current and fuel flow are currently user-input. Considerable attention was given to careful structuring of the mesh in order to minimise the element count. Full advantage was taken of the reasonable simplifications that can be made of the solid structure, fluid flows and fluid-solid heat exchanges. It was decided at the outset to make these simplifications as without them the model would have required orders of magnitude greater computational capability, probably without appreciably greater accuracy of the thermal distribution. Control of the radiation model setup was also both reasonable and necessary. Some validation of the tube mesh and initial steady-state results will be presented.