Monitoring of Wound Rotor Induction Machines by Means of Discrete Wavelet Transform

Abstract

This article studies the efficacy of a time-scale technique based on DWT for both stator and rotor asymmetries faults detection in WRIMs. The preliminary effort on the RAF detection in IMs relied mainly on the energy evaluation of a known low frequency bandwidth through DWT analysis. It was illustrated that the SCSVM and the IMSC are both good indicators for the RAF monitoring in steady state working conditions of WRIMs. In this work, it will be shown that in the case of any load variation, there is a time-varying DC component that may affect the energy evaluation criterion, resulting in a false alarm. Thus, an efficient method is proposed to estimate and then, to remove this DC component from both SCSVM and IMSC, before the energy evaluation through DWT analysis. Furthermore, by introducing a new frequency bandwidth, associated with SAF in the rotor current signature, this method can be extended to a universal slip independent stator and rotor asymmetries faults detection technique in WRIMs. The proposed approach is validated by experimental tests using a 7.5kW WRIM test rig.