Fault Diagnosis of Induction Machines With Rotor Asymmetries Using the Backward-Rotating Transient Current Component

Abstract

Induction machine fault diagnosis using current spectral analysis is a well-established diagnostic technique based on the identification of the characteristic harmonic components generated in the machine current by each type of fault. However, one of the main problems with the application of this technique to the diagnosis of rotor asymmetry faults in induction machines is that the fault components have much lower amplitudes than the fundamental component and can be very close to it, making their detection difficult, especially in transient regimes. To improve the detection of fault harmonics, this work proposes a new diagnostic current signal, the backward-rotating transient current signal, which is generated in the time domain using the Hilbert transform of the stator currents and is free of the strong influence of the fundamental component. The key novelty of this proposal is the combination of the analytical current signals and the symmetrical components method, which produces a purely backward-rotating transient current signal that cannot be obtained using the raw phase current signals. This proposal is presented theoretically and validated in transient regime using a commercial induction motor with rotor asymmetries.