Adaptive and Multiobjective Fault-Tolerant Control of a Solid Oxide Fuel Cell System Under Combinational Faults

The Solid Oxide Fuel Cell (SOFC) system offers high energy efficiency and low-carbon emissions, making them promising for transportation electrification. However, they are prone to various faults, especially combinational failures. To address this, a new electrochemical model validated is developed and experimentally validated. It simulates the effects of reformer carbon deposition, heat exchanger rupture, and their combinations on system performance. An adaptive and multi-objective fault-tolerant control scheme is then proposed. A grid search-optimized ensemble learning algorithm diagnoses 5 fault types. Based on diagnosis, 5 control strategies, including fuzzy control and neural network control, are adaptively applied to ensure multiple performance indicators remain within safe limits. Test results show fault diagnosis accuracy exceeds 97%. Under fault conditions, temperature and efficiency deviations are kept within 5%, and voltage deviation within ±3. The method ensures stable SOFC operation under combinational fault scenarios.