Uniform and Localized Magnet Demagnetization Detection of Permanent Magnet Motor based on On-line Flux Estimation

The focus of this paper is on permanent magnet reliability issues in permanent magnet synchronous machines (PMSM). Considering that the motor operates in various working environments, the motor may face different demagnetization faults, and there are two common ones: uniform demagnetization (UD) and local demagnetization (LD). Based on the mathematical model of standard PMSM, this paper analyzes the change of flux linkage under UD&LD fault to find out the fault signatures for diagnosis. The analysis results will show that the fault characteristics can be predicted by prior knowledge of the target PMSM specification. And to reduce the computational burden of the microcontroller, the frequency frame method is used to extract the fault features. Furthermore, this study combines PWM measurement techniques with flux observers that enable accurate flux estimation and true motor isolation. In other words, a non-invasive and easy-to-implement magnetic fault diagnosis method is established.