Multi-scale model to describe the local degradation and mechanical failures in an SOC stack

A multi-physical 3D model of the stack is a useful tool to increase the lifetime of an SOC stack, as this can be used to study the impact of modifying operating parameters, design and/or predict the requirements to the stack component materials. The multi-physics model can estimate the variation of physical quantities (e.g. temperature, overpotentials, thermal stresses), which could be used to assess local degradation or mechanical failures. Multi-physical 3D models describing all relevant physical phenomena (current, gas flows, heat transport, mass balances and mechanical stresses) with the geometric details are computationally expensive to run, i.e. days on a cluster is required for a single steadystate simulation. However, by using multi-scale models or so-called homogenized models, the computational resources are tremendously decreased, such that a steady-state model can be solved within minutes on a workstation.

A wide range of models have been developed to describe the degradation and failure of cells and other components in the stack for both electrolysis and fuel cell modes of operation. These include Ni migration and agglomeration, Chromium poisoning, corrosion and failure in the oxygen electrode / interconnect interface. Most of these relate simple physical models or empirical models to the observed degradation in actual cells and stacks by fitting a few unknown parameters. These models are rather useful to describe the local degradation in SOC stack, but the local evolution will influence the entire stack, and a full 3D stack model is needed to predict the spatial variation of the degradation.

In this paper, we present a 3D model of an SOC stack, where the local degradation and mechanical failures are described everywhere in the stack over the stack lifetime. The model is thus able to describe all relevant physics (currents, gas flows, heat transport, mass balances and mechanical stresses) coupling in space and time. Simulations of tens thousands of hours of operation can be computed within an hour or two on a high-end workstation, a task that would have taken months of simulation time with conventional models. This makes it possible to assess the lifetime, effect of design modifications and material requirements for the SOC stack technology.