Detection methods of broken rotor bar fault in squirrel cage induction motor with light-load

The broken rotor bar is one of the frequent faults in squirrel cage induction motors. Fault detection based on a common steady-state analysis technique, such as FFT, is significantly dependent on the loading conditions of induction motors. At light load, it is difficult to distinguish between healthy and faulty rotors. This paper provides two detection methods based on digital signal processing system for the steady and the starting states of induction motors. One is based on the self-adaptive filter and frequency-spectra analysis of the stator current suitable for steady operation state. The other is the short time Fourier transforms (STFT) analysis of the starting stator current suitable for starting operation state. The simulation and experimental results show that the methods are practicable.