Online Fault Detection of Irreversible Demagnetization Using Current Injection and dual Extended Kalman Filter In Decoupled Vector Space Decomposition control

This paper presents an Irreversible demagnetization fault (IDF) detection method using parameter estimations for dual three-phase permanent magnet synchronous motors in a decoupled vector space decomposition control. The conventional diagnosis methods of IDF are noise sensitive and computationally burden. Current injection is performed to estimate temperature and winding resistance without the interference of magnetic saturation and output torque. Since estimating the temperature and the winding resistance, dual extended Kalman filter is used to reduce the burden of computation and to be robust to noise. The suggested IDF detection was carried out through flux linkage estimation considering the temperature.