Application of Support Vector Machine to stator winding fault detection and classification of permanent magnet synchronous motor

Abstract

This paper deals with the topic of detecting and classifying one of the most common faults of permanent magnet synchronous motor – interturn short circuits. The idea of using spectral analysis of the stator phase current signal and its envelope to symptom extraction and the support vector machine to classify the fault is proposed. To assess the effectiveness of the proposed diagnostic method, experimental tests were conducted. The object of the experimental verification was a 2.5kW permanent magnet synchronous motor operating in a closed-loop control structure. The impact of changes in motor operating conditions on the effectiveness of failure classification has also been tested.