Online Diagnosis of Partial Demagnetization Fault in PMSMs Based on Kurtosis of Air-Gap Flux Density Signal Difference

Permanent magnet synchronous machines (PMSMs) working in hostile environments face a potential risk of partial demagnetization fault (PDF). Once the fault occurs, it reduces the machine performance and accelerates component aging. Therefore, timely and accurate diagnosis of PDF is crucial for ensuring reliable operation of machine systems. For this purpose, this article proposes an online PDF diagnosis method for PMSM based on the kurtosis of the air-gap flux density signal difference. Initially, the frozen permeability technique (FPT) is adopted to linearly decompose the air-gap magnetic field. Subsequently, the established equivalent magnetic circuit (EMC) model is simplified, and the variation of the air-gap flux density under each pole of main magnetic path after PDF is analyzed. Furthermore, the fault characteristic and reference are presented. Through the equiangular interval resampling (EIR) method, the time-domain air-gap flux density signal excited by permanent magnets (PMs) is converted into angular domain signal. The PDF can be diagnosed by comparing the fault characteristic value with the reference value, and the fault degree will be calculated through the trend of flux density variation. Finally, simulated and experimental results validate effectiveness and superiority of the proposed method for diagnosing PDFs.