<p>This paper proposes a fault detection method for doubly fed induction generators (DFIG) based on the combination of a novel sliding mode observer (SMO) and the bilateral cumulative sum (CUSUM) algorithm. The newly proposed sliding mode observer significantly improves the motor state tracking accuracy and convergence ability compared to the traditional power reaching law observer. For fault detection, the core of the method lies in using the residual between the estimated current value of the SMO and the actual output current as the fault characteristic signal. The residuals generated by the observer under normal and various fault conditions are uniformly used as the input to the detection algorithm, achieving an organic integration of the sliding mode observer and the cumulative sum algorithm. To distinguish the types of faults more clearly, while retaining the real-time advantage of the traditional CUSUM algorithm, a bilateral cumulative sum algorithm is innovatively proposed. Simulation analysis verifies that this method can accurately reflect the fault status of the motor, effectively convert weak fault characteristic signals into digital output signals, and significantly enhance the fault detection capability for DFIG.</p>