Abstract
This paper attempts to present a concise review of the research done around fault detection methods for IMs. Various methods to diagnose faults researched over past few years are reviewed and presented with the focus on state-of-the-art fault diagnosis that use on-line (while the machine is running) and non-intrusive (without requiring sensors) techniques. The speed and accuracy of the fault diagnosis requires signal analysis using DSPs. The signal analysis algorithms such as Fourier Transform and Wavelet Transform are studied and compared. Also for the sake of completeness the other traditional fault diagnosis techniques have been discussed. Current research is exploring Partial discharge, Winding Function Theory, and Expert Systems for fault diagnosis and these techniques are also mentioned. Reviewing of the main options for induction machines fault detections and comparing has been the main consideration of this paper. Mainly, in case of traction systems the new techniques are essential to avoid inadvertent shutdown.

This publication has 17 references indexed in Scilit: