Fault prognosis control of solid oxide fuel cell system based on health evaluation

Abstract High cost and short life-span hinder the large-scale commercial application of solid oxide fuel cell technology. System performance inevitably degrades and gradually evolves into faults in the long run, leading to the change of system input-output characteristics. The control system designed based on initial conditions faces with problems such as parameter mismatch, inaccurate control and even wrong control laws. This paper proposes an adaptive control strategy to enable the system controller to cope with degradation and fault phenomena in real systems. The system characteristics under different degradation and fault conditions are investigated. A novel health evaluation system is developed, where an early fault identification and pretreatment approach based on optimal operation point optimization model and fuzzy neural network are adopted for control optimization. Simulation results show that the enhanced system controller can improve system efficiency by about 2% and prolong life-span significantly with the reality of degradation and fault mechanisms.