This paper deals with the finite element analysis of the demagnetization process of the line start permanent magnet synchronous motor. Special attention has been paid to demagnetization risk assessment after resynchronization during a short-term supply power outage. The current and torque waveforms have been determined assuming the difference depending initial rotor position angle. It has been demonstrated that the highest demagnetization risk occurs when resynchronization (motor reclosing) is performed whe induced electromotive forces are in anti-phase to the supply voltage waveforms. The effect of cage winding resistance on the risk of demagnetization is examined and discussed.

The paper focusses on the analysis of the demagnetisation process of permanent magnets in line-start synchronous motors in dynamic states related to start-up and resynchronisation. A field-circuit model of electromagnetic phenomena was used to analyse the demagnetisation process, taking into account the influence of temperature on the properties of permanent magnets and their resistance to demagnetisation. The results of the conducted research have shown, among other things, that the process of resynchronisation of the motor is much more dangerous from the standpoint of the risk of demagnetisation than the start-up itself.

Modern drives with Permanent Magnet Synchronous Motors (PMSMs) require both efficient control structure to ensure excellent dynamics and effective diagnostic algorithms to detect the motor faults that can occur. This paper shows the combination of both mentioned aspects – the direct-axis based signals of the Field Oriented Control (FOC) structure are proposed as diagnostic signals to allow diagnosing the interturn short-circuit failure that can appear inside stator windings. The amplitudes of second order harmonics are selected as the fault indicators. Different modelling methods are analysed and compared in detail in this paper: an analytical mathematical model, a Finite Element Method (FEM)- based model and next verified using a laboratory setup. The results obtained using all the mentioned models proved that the proposed fault indices are increasing significantly with the number of shorted turns and are independent on the load torque level.