Irreversible Demagnetization Fault Prognosis in a Permanent Magnet type Machines

Irreversible demagnetization fault (IDF) is considered a constraint in permanent magnet (PM) types machines used in traction applications. Variation in temperature, load, field weakening, and mechanical stresses are the main reasons that can permanently demagnetize the PMs during steady-state operation. In this paper, a prognostic approach for the prediction of the IDF is presented. A machine learning method is used to monitor the variation in the operating point of the machine and predict the remaining margin of temperature that a machine can endure before the occurrence of IDF. The operating point of the machine is achieved by monitoring the average magnetic flux density at the edges (weak points) of PMs for the instantaneous operating point and estimate the remaining temperature margin that can initiate the IDF in real-time. The temperature margin information obtained using the proposed method helps the user to adjust the operation of the machine before the IDF occurs. The proposed method is verified using the finite element-based simulation and experimental data.