Fault Detection of DFIG Using an Improved Fractional-Order PID Sliding Mode Observer Based on the Adaptive Golden Eagle Optimizer

This paper presents a fault detection approach for doubly-fed induction generators (DFIG) utilizing an improved fractional-order PID sliding mode observer (IFOPID-SMO). The IFOPID-SMO eliminates chattering and demonstrates superior stability effectively and has a faster convergence speed compared to other observers. Firstly, the stator voltage-oriented vector control method is employed to establish the nonlinear space state vector equation of the DFIG. Then, the IFOPID-SMO is constructed based on the DFIG’s mathematical model, and the d-q axis currents of the rotor are estimated. Secondly, the observer parameters are optimized using the adaptive golden eagle optimization algorithm, which enhances the convergence speed and steady-state accurateness of the observer significantly. Furthermore, the asymptotic stability of the IFOPID-SMO is analyzed by Lyapunov’s second law, which guarantees that the system state converges to a steady value rapidly and stably. Finally, three faults of DFIG including stator inter-turn short-circuit fault, rotor inter-turn short-circuit fault, and rotor current sensor faults are detected respectively. The effectiveness of the IFOPID-SMO is demonstrated through the data recorded of a doubly-fed wind power test bench under unfaulty and three faulty operation conditions.