Pattern recognition-based fault detection of a PM synchronous motor under stator winding short circuit faults

Abstract

Stator winding short-circuit faults arising from winding insulation faults are among the most frequent faults in permanent magnet synchronous motors (PMSMs). If left undetected, such fault may be rapidly propagated, resulting in phase-to-phase or phase-to-ground faults, even the breakdown of the whole motor. A powerful fault diagnosis method requires the computation of a fault sensitive quantity and an appropriate method to get a diagnostic index and a threshold which present the edge between faulty and healthy conditions. This is particularly critical for stator short-circuit faults, especially in PMSMs which can cause catastrophic damage to the machine in a very short time. This paper proposes a new detection fault approach based on pattern recognition analysis for detecting the stator inter-turn fault in two phases. Firstly, an image of αβ stator currents in healthy and faulty conditions is composed in the 2D plane. Then, the extracted parameters according to the obtained image as areas and angles of rectangle shapes are used to detect the stator winding faults. Finally, a fault severity index (FSI) gives a slight or a serious degree of faults in PMSM. The experimental results are presented in this paper to show the usefulness of the proposed approach.