A complex spectral analysis of the stator current dedicated to the defects monitoring of squirrel cage induction motors

Nowadays, the induction motor is an important element in industrial process in terms of safety and efficiency. In order to limit the cost induced by a defect, we have to make an early detection. Consequently, the earlier incipient faults will be detected the cheaper remediable faults will be. So, one of the widespread method to monitor the process is the spectral analysis of the stator current. Usually, the real fast Fourier transform (FFT) is used. Despite its waning achievement in regards of on-line methods, the complex FFT gives more information about the healthy state of the motor. So, a comparison of these two approaches in this paper will allow the users to better conclude or not to failure of the induction motor. Moreover, a sliding version of the discrete Fourier transform is conceivable even if the signal to analyse is real or complex. This technique is highly recommendable when updated spectrum is necessary at each sampling time. Experimental results show the efficiency of the complex FFT in comparison to the classical one.