High-Temperature
Solid Oxide Electrolysers (SOECs) are a promising technology with the potential to reduce the electrical energy consumption by 30% compared to conventional low temperature electrolysers. The elevated operating temperatures, typically above 600°C, also allow for synergies with industrial production processes (e.g. steel, ammonia, etc.) where waste heat or steam is available. However, key challenges remain for the successful deployment of the technology associated to costs, durability and scale up.
A potential option to address some of these challenges is enlarging the active cell area from the state-of-the-art of 100 cm², up to 1000 cm², as fewer cells are then necessary to produce the same amount of hydrogen, thereby reducing costs. As part of the ~50 Mio€ ‘HyPRO’ project,
the largest ever R&D project on green hydrogen in the Netherlands bringing together 58 partners from research and industry,
we are looking for a PhD candidate to develop an electrochemical + thermomechanical cell and stack model of the large active area SOEC technology developed by project partner, TNO. The model will need to incorporate gas manifolds and hot-box design into a
hierarchical modelling methodology developed previously for the integrated multiscale modelling of SOEC cells and stacks The selected PhD candidate will then use the model to study the effect of various design factors such as electrode microstructure, layer thickness, and active area at the cell level, and rib and channel width for interconnect design, flow distribution and pressure drop through the gas manifolds, and sufficient mechanical stress on the interconnect and cells for effective sealing, at the stack level. The multiscale cell and stack model will then need to be coupled to process models developed by project partner University of Groningen, to perform a techno-economic analysis of the large active area SOEC technology for use cases selected by industrial project partners such as Bosal, Shell, and Lyondell Basell. The coupling of the multiscale multiphysics cell and stack model to the process model may require a model reduction step, e.g. via the development of a neural network surrogate of the multiphysics model.