Predictive maintenance in the permanent magnet synchronous motor (PMSM) is of paramount importance due to its usage in electric vehicles and other applications. Recently various deep learning techniques are applied for fault detection and identification (FDI). However, it is very hard to optimally train the deeper networks like convolutional neural network (CNN) on a relatively fewer and non-uniform experimental data of electric machines. This paper presents a deep learning-based FDI for the irreversible-demagnetization fault (IDF) and bearing fault (BF) using a new transfer learning-based pre-trained visual geometry group (VGG). A variant of ImageNet pre-trained VGG network with 16 layers is used for the classification. The vibration and the stator current signals are selected for the feature extraction using the VGG-16 network for reliable detection of faults. A confluence of vibration and current signals-based signal-to-image conversion approach is also introduced for exploiting the benefits of transfer learning. We evaluate the proposed approach on ImageNet pre-trained VGG-16 parameters and training from scratch to show that transfer learning improves the model accuracy. Our proposed method achieves a state-of-the-art accuracy of 96.65% for the classification of faults. Furthermore, we also observed that the combination of vibration and current signals significantly improves the efficiency of FDI techniques.

Contemporary research has shown impetus in the diagnostics of permanent magnet (PM) type machines. The manufacturers are now more interested in building diagnostics features in the control algorithms of machines to make them more salable and reliable. A compact structure, exclusive high-power density, high torque density, and efficiency make the PM machine an attractive option to use in industrial applications. The impact of a harsh operational environment most often leads to faults in PM machines. The diagnosis and nipping of such faults at an early stage have appeared as the prime concern of manufacturers and end users. This paper reviews the recent advances in fault diagnosis techniques of the two most frequently occurring faults, namely inter-turn short fault (ITSF) and irreversible demagnetization fault (IDF). ITSF is associated with a short circuit in stator winding turns in the same phase of the machine, while IDF is associated with the weakening strength of the PM in the rotor. A detailed literature review of different categories of fault indexes and their strengths and weaknesses is presented. The research trends in the fault diagnosis and the shortcomings of available literature are discussed. Moreover, potential research directions and techniques applicable for possible solutions are also extensively suggested.

Fast switching and common-mode voltage of the space vector pulse-width modulation inverter output create several parasitic capacitances according to the geometry of the motor. The windings emit the electric and magnetic flux inside of the motor. They create capacitance links and electromotive force (EMF) in the shaft, respectively. Those capacitance links and EMF generate a current through two bearings of the shaft and directly reduces the bearing lifetime. In this study, we propose a new design of equivalent circuit, taking into account all parasitic capacitor components.

A comprehensive analysis of inter-turn-short fault (ITSF) in a fractional slot concentrated winding (FSCW) interior permanent magnet synchronous motor (IPMSM) is presented in this paper. During online fault detection, the fault signatures can be influenced by the controller action in a closed-loop control system. Therefore, this study focused on the comparison of the ITSF characteristics in an IPMSM using open-loop and closed-loop control drives. The variation rate in various parameters, such as flux density, input current, circulating current, torque ripple, and vibration under healthy and ITSF conditions are compared. Conventional six-step square-wave control (SSC) and field-oriented control (FOC) drives are respectively used for open-loop and closed-loop analysis of ITSF. Furthermore, this study also focuses on mitigating capability of the control drives for achieving extended post-fault life of the motor. Based on our analysis a fault mitigating strategy for the benchmark IPMSM is also suggested. A multi-domain co-simulation technique is used to integrate the finite element method based model of IPMSM with the control drives under different severities of ITSF for the simulation. Finally, experiments on a three-phase 400-watt IPMSM are performed for verification. The result shows that in the FSCW machine the ITSF can be mitigated and FOC has a better capability of ITSF mitigation than SSC.

Condition monitoring of the stator and rotor problems such as inter-turn-short fault (ITSF) and partial irreversible demagnetization fault (PIDF) in an interior permanent magnet synchronous motor (IPMSM) is essential for achieving high efficiency and reliability. This paper presents a torque angle based inverter-embedded technique for the online detection and identification of ITSF and the PIDF. This technique monitors the deviation in the torque angle caused due to ITSF and PIDF at a certain operating point. In steady-state operation, the torque angle decreases from its original value (healthy state) due to ITSF while increases in the case of PIDF. This particular trend of torque angle is used for the detection and the identification of ITSF and PIDF. In this paper, the torque angle for healthy IPMSM is first obtained using multivariate regression analysis over the entire operating region and then compared with the torque angle during operation for any potential variation. The proposed method is verified using finite element method based simulation and experiment on a 400-watt IPMSM. The results suggest that the proposed method can be a viable candidate for efficient ITSF and PIDF detection and identification in real-time.

This paper presents the online detection, separation and severity estimation of the uniform and partial irreversible demagnetization fault (IDF) in interior permanent magnet synchronous motor (IPMSM) using the frequency spectrum of the input-voltage. Due to IDF the fundamental component of the voltage reduce and the partial IDF cause additional even order harmonics in the input-voltage. In this study, the fundamental component and the even order harmonics are used as the fault indicators. Furthermore, the increase in the ratio of fundamental to fifth order harmonic due to IDF is used to train an artificial neural network based algorithm and estimate the severity of a relatively small degree of IDF in the IPMSM. The proposed method has the ability to cover most of the operating points of the benchmark IPMSM and also able to distinguish between IDF and eccentricity fault. Finite element method based simulation and experiment on a 400-watt IPMSM is performed to verify the proposed method.