Multi-physical field simulation of GIS partial discharge based on UHF signal propagation characteristics

Abstract Gas-insulated switchgear (GIS) is critical in power systems, yet partial discharge (PD) caused by insulation defects threatens its safety. This study investigates ultra-high frequency (UHF) signal propagation in GIS and analyzes PD behavior under insulation defects via a multi-physics COMSOL model. The coaxial wave guide structure of GIS reveals that TEM waves (non-dispersive, no cutoff frequency) dominate with radial components exceeding axial/normal peaks by over 2, validating radial UHF sensor deployment for enhanced sensitivity. TE/TM waves exhibit cutoff frequencies, propagating only at high frequencies. Simulations of three defects — metal protrusions, free particles, and surface contaminants—show distinct time-domain signatures: free particles exhibit 15% faster Z-component decay than protrusions, while contaminants display static-field characteristics at pouring holes, aiding defect identification. Optimized scattering boundaries and mesh settings (0.02 m unit size) improve computational efficiency by 30%, demonstrating the model’s adaptability to complex GIS structures. These findings provide theoretical insights for UHF-based PD detection, sensor optimization, and defect diagnosis, supporting enhanced GIS reliability in power grids.