GIS composite apparatus insulation fault diagnosis method based on ultra-wideband magnetic signal

The invention belongs to the field of high-voltage equipment state accurate detection, and particularly discloses a GIS combined electric appliance insulation fault diagnosis method based on ultra-wideband magnetic signals, which comprises the following steps: installing a plurality of MC magnetic coupling sensors on a GIS combined electric appliance shell, and acquiring discharge signals in a GIS combined electric appliance acquired by the MC magnetic coupling sensors; carrying out dimension reduction processing on the characteristic space of the discharge signal in the GIS combined electric appliance; and utilizing feature clustering to obtain GIS combined electric appliance insulation fault classification: four fault models of internal suspension discharge, point discharge, creeping discharge and air gap discharge. Compared with the prior art, through high-voltage equipment accurate detection technology research based on magnetic sensor detection and KPCA kernel principal component analysis technology, int