Lifetime prediction of SteelCell® stacks using advanced multiphysics modelling

A major challenge in the development of SOFC stacks designed for long operating lives of the order of 50-100kh is predicting through-life performance without running durability tests under a particular set of operating conditions for 5-10 years. Whilst this is the final aim it is not practical for system design purposes or in the situation where product lifecycles are much longer than product development cycles. A particular challenge is that Ceres Power provides stack technology to multiple customers, most of whom have their own requirements for stack operating conditions, which means that degradation rates measured under one set of operating conditions are not necessarily directly applicable to a different set, and undertaking long term verification testing under every set of stack boundary conditions is not practical.

To address this issue, Ceres Power has invested significant effort into understanding mechanisms of performance degradation at the fundamental cell level, generating models of degradation as a function of time, temperature and current density. Most degradation mechanisms observed are common to all SOFC cells, such as poisoning of both electrodes and increases in ohmic resistance due to oxide scale growth.

These degradation models are not in themselves sufficient to enable through life performance simulation, as in a full stack there are significant variations in temperature and local current density, meaning different areas of the cell will be degrading at different rates. In addition, because of performance degradation through life, significant changes in the local temperature and current density distribution may occur, and the stack will often generate more heat meaning the air flow and/ or inlet temperature will need to change to maintain the nominal stack temperature.

To allow simulation of these effects, a full multiphysics stack model has been developed using COMSOL Multiphysics, incorporating fluid dynamics, heat transfer, electrochemical and heterogeneous chemical reactions. The electrochemical model at start of life is calibrated from extensive experimental performance mapping using coupon cells with high resolution AC impedance spectroscopy, and the resulting full multiphysics stack model validated against test data from real stacks.

A dynamic solver has then been applied to the multiphysics model applying the models of cell degradation to be applied to a full stack, including changing the air flow through life to maintain a constant nominal stack temperature. The model has been validated against through-life performance tests. This is a powerful tool which allows rapid evaluation of the likely impact of operating a stack in a condition other than ones in which extensive test data already exist.