Grant: Alliance Grants

The main objective of this project is to address the challenges in commercializing solid oxide fuel cell (SOFC) technology by focusing on minimizing long-term degradation and maximizing performance. This will be achieved by the development of innovative machine-learning techniques for SOFC control and diagnostics along the design of novel power electronics to increase the observability and controllability of SOFC cells in a stack. This project is well aligned with the Hydrogen Strategy of Canada for the need for sustained support in the research and development of fuel cells to reduce costs and discover breakthrough technologies. Issues such as microstructural failure, thermal stress, temperature gradients, leakage, corrosion, and fuel maldistribution among cells lead to SOFC stack degradation. The main challenge in SOFC technology lies in achieving long-term stability and a low power degradation rate, which should be below 0.2% per 1,000 hours. This challenge is amplified by the limited observability and controllability of SOFC cells in a stack that often necessitates the replacement of the entire stack due to limited faulty cells which is very costly. To address these challenges, this project aims to create artificial intelligence driven diagnostics and control methods to monitor SOFC cells performance in real-time and control the cells using powerful and cost-effective power electronic units. The proposed approach is unique and builds upon experimental setup design, testing of SOFC single cells and stacks under various conditions, model-based diagnostics and controller design, development of time series machine-learning methods, and building cell-level power electronics. Cummins, a global leader in power generation with products in fuel cells and engines, is the industrial partner in this project and is involved in various stages of the project. The collaboration between the University of Alberta and Cummins strengthens Canada's position in developing low-GHG solutions.