Diagnosis of stator shorted-turn faults in induction machines using discrete wavelet transform

Abstract

Incipient detection and diagnosis of stator winding shorted-turn faults in induction machines is essential for reliable and economical operations in industries. The problem of detecting shorted-turn faults in stator windings has been difficult. Although, from the supply currents, major winding faults can easily be detected, minor faults with less than five per cent of winding turns are not easily detectable. If not detected early, such faults can lead to major winding faults which may further lead to the disruption of the machine and causes production shutdowns. This paper describes the discrete wavelet transform (DWT) based diagnosis technique that analyses stator currents under faulty (stator shorted-turn) and healthy conditions. The developed technique depends on the band pass filtering carried out by the DWT, facilitating the extraction of the harmonic components produced by the shorted-turn faults. The discrepancies energy spectrum for three frequencies from the residual of de-noising analysis enables us to classify the fault index. From 38 seconds, the shorted turn begins to show significant difference in the energy spectrum, and if this is not detected in time, it may lead to a more serious fault.