A Study on the Extraction Method of Partial Discharge Features in Gas Insulated Switchgear Based on Ultra-High Frequency Signal Envelope

With the help of the Hilbert transform, low-frequency signal that is modulated by high-frequency signal can be demodulated. A method to extract time-domain features of ultra-high frequency (UHF) signals emitted by gas insulated switchgears (GIS) partial discharges (PD’s) is presented. Utilizing the Hilbert transform, the envelope of a PD signal, and further the time-domain key points relevant with the features, can be acquired. A platform for artificial insulation fault experiments is developed. Experiment data of 3 typical types of PD signals are gained, with which the presented algorithm was analyzed and examined. The result shows that different types of PD are well recognized. It can be concluded that the method presented is practical and effective