Application of MCSA and SVM to induction machine rotor fault diagnosis

A fault diagnosis system using support vector machine (SVM) based classification techniques is developed for cage induction machines rotor fault diagnosis. The proposed algorithm uses the motor current signal analysis (MCSA) as input. By using FFT method, the frequency spectrums of the stator current signal are derived, and several features are extracted. A support vector machine based multi-class classifier is then developed and applied to distinguish health condition from different rotor fault conditions. A series of experiments using a three phase cage induction machine performed in different fault conditions, such as broken bars, broken end-rings, eccentricity etc., are used to provide data for training and then testing the classifier. Experimental results confirm the efficiency of the proposed algorithm for diagnosing different rotor faults