Wavelet analysis for stator fault detection in induction machines

Abstract

The main objective of the proposed analysis is the detection of inter-turn short circuits in the stator windings of an induction machine. The analysis of the space vector current modulus of an induction motor is presented in this paper. This analysis is based on Daubechies 8 wavelet with seven decomposition levels. The 5th decomposition-level detail signal for a 4 kHz sampling frequency is chosen as a fault indicator, based on simulation results that show different behaviors of the energy contained in the detail signals independent of the percentage of load and fault levels. Experimental results that validate the proposed strategy are also presented. These results also show that the strategy is in addition immune to load variations as well as to feeding voltage unbalances. © 2011 World Scientific Publishing Company.