Online monitoring of internal ground-wall insulation in permanent magnet synchronous machines

For inverter-driven permanent magnet synchronous machines (PMSMs), the transient overvoltage phenomenon at the machine line-end has attracted widespread attention because of its potential to degrade the winding insulation, thereby causing premature failure. For this issue, various line-end ground-wall (GW) insulation monitoring methods have been proposed. However, recent research indicates that, in addition to the line-end, significant overvoltage also occurs inside the windings. In some cases, the internal overvoltage even exceeds the line-end overvoltage. Therefore, it is necessary to develop an online internal GW insulation monitoring method. This paper first establishes a high-frequency winding model, which can be used to study the effects of GW insulation aging at different positions. This study demonstrates that when the internal GW insulation degradation occurs, the common-mode (CM) impedance antiresonance frequency changes. It also shows that the natural frequency of the CM circuit is equal to the CM impedance antiresonance frequency. Therefore, this paper proposes an insulation monitoring method based on the system’s natural frequency. This method can quantitatively detect the internal GW insulation aging. To accurately extract the natural frequency of the system, a transfer function-based frequency extraction method is proposed. The performance of the proposed method is not affected by switching frequency and modulation index, demonstrating its strong robustness. Both distributed and concentrated winding PMSMs are employed for experimental validation.