Fault detection and classification in permanent magnet synchronous machines using Fast Fourier Transform and Linear Discriminant Analysis

The main objective of this paper is to propose a method to detect the presence of a fault in Permanent Magnet Synchronous Machines (PMSMs), determine the type of that fault and estimate the severity in the case of eccentricity fault. In this paper, three types of faults are discussed: static eccentricity, inter-turn short circuit, and demagnetization faults. The machine is controlled using a three phase current source and the harmonics of the stator voltage are used as detailed features for the classifier for fault detection. Two dimensional (2-D) Finite Element Analysis (FEA) is used to model and simulate the machine under healthy and faulty conditions. Fast Fourier Transform (FFT) analysis is performed to the phase voltage signal to detect the frequency spectrum and Linear Discriminant Analysis (LDA) is chosen as a classification method. To validate the method, two different types of PMSMs are tested: the first with a concentrated winding and the second with a series distributed winding. The tests are applied for different operation loads.