Diagnosis of turn-to-turn insulation failure of induction motor winding with aid of Support Vector Machine

Short-circuit fault of stator winding due to aging of electrical insulation is one of the most probable faults in motor drive systems. The authors have proposed a simple method for diagnosing the fault probabilistically by considering distributions of features, i.e. magnitude and phase of load current. Laboratory experiments proved that the method was promising. In order to enhance accuracy of diagnosis, especially in the case of a slight insulation failure such as one turn-to-turn short circuit, this paper proposes a method with the aid of Support Vector Machine using the features. Effectiveness of the new method is evaluated compared with the probabilistic one.