Identification of broken rotor bar fault and degree of loading in induction motor using neuro-wavelets

This paper presents a methodology for the detection of broken rotor bar fault in induction motor at different load conditions. Wavelet transform is applied to the stator current, for the extraction of the signature of the fault. These wavelet coefficients are fed as input to a feedforward neural network. The output of the neural network classifies the health of the rotor of the induction motor (healthy/ faulty), and also the load at which the machine is operating. The entire simulation is carried out using MATLAB. The proposed network has performance efficiency of 93.75%.