Hybrid Fault Isolation and Power Control of Wind Turbine Generators Based on Stacked UIO

The stable operation of megawatt-class wind turbines directly impacts grid connection performance. As complex electromechanical systems operating in variable environments, time-varying hybrid faults frequently occur, posing challenges for fault-tolerant control (FTC). This paper proposes novel approaches for hybrid fault diagnosis and fault-tolerant controller design. First, a collaborative architecture for a stacked unknown input observer (UIO) is constructed for integrated fault diagnosis of hybrid faults in wind turbine generators (WTGs). The designed stacked UIO strategy can decouple and separately estimate the dynamic of speed and torque, achieving effective time-varying hybrid fault diagnosis. Second, given the disturbances, uncertainties, and discrete-time nature of WTGs, the asymptotic stability of the proposed UIO is established using Lyapunov's theorem. Finally, combining fault information, a two-degree-of-freedom model predictive control strategy is adopted to achieve FTC while also considering power stability tracking. Simulation results show that the proposed method can estimate time-varying faults in WTG actuators and sensors, and maintain reliable control performance under hybrid fault conditions, thereby verifying its effectiveness as a comprehensive FTC solution.