Grey Clustering Based Diagnosis of Induction Motor Faults

In this paper, a fault classification method based on grey clustering is proposed for fault detection of induction motors. The amplitudes of rotor frequency related sideband components obtained through fourier transform of one phase stator current are used for broken rotor bar faults. Park's vector components are extracted from three phase motor currents and then new feature is obtained using principal component analysis on park vector components. Obtained features constitute the inputs of grey clustering algorithm. One broken rotor bar, stator faults and stator and multiple faults are diagnosed.