Solid oxide fuel cells represent a promising technology for distributed power generation due to their high electrical efficiency and fuel flexibility. However, thermal management remains a significant operational challenge. Temperature variations within the SOFC stack can cause mechanical stress, material degradation, and performance losses. Direct measurement of internal temperatures is difficult because of the harsh operating environment, typically 800–1000°C, and the complexity of sensor integration within the stack architecture.
This research addresses the temperature estimation problem through a dual-observer framework combining the Luenberger observer with sliding mode control and the Kalman filter algorithm. The approach begins by developing a discrete-time state-space mathematical model of the SOFC stack based on one-dimensional thermodynamic principles, incorporating system noise characteristics. To optimize observer performance, the research team filters input variables and analyzes the measurement matrix condition number to identify the best sensor input combinations.
The Luenberger-sliding mode observer handles model uncertainties and system disturbances through robust control theory, while the Kalman filter observer processes measurement data to estimate states and suppress noise. Both designs incorporate error system decoupling and vibration suppression techniques to enhance stability and reduce estimation lag—the time delay between observed and actual system states.
Simulation experiments validate the method against multiple performance criteria, demonstrating improved estimation accuracy compared to traditional approaches. The technique bridges the gap between observable measurements and internal stack conditions, enabling better thermal control strategies and predictive maintenance. For fuel cell system operators, accurate temperature estimation translates to safer, more efficient operation and extended component life. This work is particularly relevant for industrial SOFC installations and emerging fuel cell-based grid support applications where thermal management directly impacts system availability and cost-effectiveness.
