A validated multiscale modelling approach to solid oxide fuel cell system design

Fuel cells play a crucial role in the energy sector's efforts to mitigate greenhouse gas emissions by efficiently generating electricity from H₂ and O₂ through electrochemical reactions for a diverse range of applications. Among these, solid oxide fuel cells (SOFCs) stand out as electrochemical technologies capable of converting fuels such as hydrogen into electricity through oxidation reactions.

Despite successful small-scale demonstrations of fuel cell technology, significant challenges remain in reducing costs, extending lifetimes, and scaling power for industrial applications. Over the last decade there has been a significant maturation in fuel cell technology to meet commercial demands. However, there remains a lack of long-term field data from representative system testing.

Multi-domain 1D models are developed at system, sub-system, and component levels to simplify the complexity of the problem, identify key interfaces, and define requirements. These models are supported by existing literature, computational fluid dynamics (CFD), and experimental methods. Validation of the models is performed at representative operating conditions using data from highly instrumented test rigs.

Experimental methods validate flow and pressure distributions within the system, while ejectors developed to recirculate anode and cathode off-gas meet required pressure and recycle ratios demonstrated by the experimental methods. These validations confirm the modelling approach developed from published literature.

A health monitoring system, developed from validated models, is implemented to run real-time on the fuel cell system demonstration test. The modelling methods optimise fuel distribution within the system and stacks to achieve the overall fuel utilisation and system efficiency target for the 250kW system demonstration test.

The validated models successfully diagnose faults on test rigs in real-time and are implemented for 1800 hours of operation on the system demonstrator. Field testing of the system achieves reliability and degradation targets, validating the overall modelling methodology developed within this thesis.