Multiple fault detection in induction generator for grid-connected wind turbine application

While numerous studies have investigated the diagnosis of squirrel-cage induction machines, most have focused on the motor mode. This work extends the diagnostic scope to the generator operation. Two fault scenarios are addressed: a mechanical unbalance and a combined condition where this defect coexists with incipient Rotor Electrical Asymmetry (REA). A theoretical framework is developed to analyze the spectral behavior of stator currents under these fault conditions across both operating modes. To enhance detection sensitivity, we introduce an innovative frequency offset spectrum combination technique, which combines the stator current spectrum with its frequency-shifted versions to amplify fault-related components. Experimental validation on a laboratory test rig confirms that the proposed approach significantly improves sensitivity to incipient REA faults in generator mode operation, highlighting its potential for early fault detection in real-world applications.