A novel method for the early detection of broken rotor bars in squirrel cage induction motors

This paper emphasizes to develop a hybrid detection scheme for broken rotor bar fault in induction motors. Since the power spectrum density estimation of classical multiple signal classification (MUSIC) possesses higher resolution with short-time samples compared with FFT, this paper applies MUSIC to detect broken rotor bar fault. And thus, the impacts of the fluctuation of stator current can be decreased to a certain extent because only short-time samples are necessary. The hybrid detection scheme can be realized by using continuous subdivision Fourier transform, self-adaptive filter, rotor slot harmonics based slip estimation and MUSIC techniques. Fault detection instances in laboratory demonstrate that the presented scheme can assure the detection sensitivity and reliability of broken rotor bar fault in induction motors.