Induction Motor Stator Fault Diagnosis by Rotor Slots Harmonics Tracking Using Prony Improved Approach

Abstract

The short-circuit fault diagnosis is essential in order to avoid irreversible damages that could affect the stator circuit of the induction motor. For this purpose, this paper addresses a new approach based on Prony's high resolution spectral analysis method to avoid the drawbacks of the well-known method based on the power spectral density estimation by Periodogram technique.  This new approach, called Root-Prony, allows a better frequency resolution even on a very short acquisition time as well as a better detection even on very small magnitudes. Moreover, and in order to reduce the computation time of this approach, this paper proposes to analyze only the band around the first rotor slot frequencies. The choice of this band allows avoiding any confusion with the frequency signatures of external phenomena which appear most often in very low frequencies. The experimental results obtained show the efficiency and the reliability of this new approach in tracking the induction motor stator winding degradation.