A Detection Method for Insulator Leakage Current in Transmission Lines Based on Time-Varying Coupling of Pollution Layer Equivalent Impedance

To address the issues that existing methods neglect the physical mechanism of leakage current and struggle to decouple transient precursors under strong noise, a novel insulator state detection method based on a time-varying coupling model of pollution layer equivalent impedance is proposed. An impedance model considering surface wetting and ion polarization relaxation is established to reveal its amplitude-modulation and frequency-modulation coupling mechanism. Accordingly, a dynamic frequency band correction factor and convergence criterion are constructed to suppress energy leakage. The joint kurtosis entropy and singular value difference spectrum is interpreted as a separation criterion for random pre-discharge and periodic micro-arcing, and a robust criterion is constructed by incorporating fuzzy entropy. Experimental results demonstrate that the proposed method improves the feature signal-to-noise ratio to 14.2 dB and reduces the misjudgment rate to 3.11%, providing a new physics?driven approach for insulator state awareness.