FEM-Based Demagnetization Fault Analysis in Interior Permanent Magnet Synchronous Motor Using Machine Learning Approach

Interior Permanent Magnet Synchronous Motors (IPMSM) are increasingly used in the automobile industry owing to their high-speed operation and high efficiency. Demagnetization is one of the crucial problems in IPMSM, which requires detection and diagnosis at earlier stages, otherwise leading to permanent rotor damage. In this work, three different severities of demagnetization fault, such as 5%, 10%, and 25%, are incorporated on a 0.55 kW, 220 V IPMSM and analyzed using ANSYS Maxwell finite element tool. Various machine parameters, such as stator current, torque, flux density, and magnetic field distribution, are compared with healthy IPMSM. Stator current motor data under healthy and demagnetized conditions are acquired from finite element tool and examined by applying machine learning algorithms in MATLAB. The algorithms were trained and could predict the demagnetization at an earlier stage of 5%. The number of faulty samples increases based on the severity of the fault.