Comparison of techniques based on current signature analysis to fault detection and diagnosis in induction electrical motors

Abstract

The wide use of electrical induction motors in industries throughout the world requires, increasingly, more precision in fault diagnosis. Techniques of predictive maintenance such as Motor Current Signature Analysis (MCSA) and Motor Current Square Signature Analysis (MSCSA) are used to detect and diagnose faults patterns, characterized by the stator current spectrum, in induction motors. In this article, these techniques are applied and compared for different faults in real motors, such as inter-turn short circuit in the stator winding and eccentricity in the air gap. To assist in the comparison of these patterns of the stator current spectrum with and without faults, a theoretical model of a healthy electrical induction motor was used, with the same values of the real supply voltages, which generated the frequency spectrum patterns. The results presented in this article, which should be emphasized, demonstrated that the techniques mentioned above were suitable for the cited faults, whose comparison between the techniques showed the suitability of each one.