Online source recognition of partial discharge for gas insulated substations using independent component analysis

To perform reliable insulation diagnosis for gas-insulated substation (GIS), detectable partial discharge (PD) should be identified quickly and effectively. With the increasing application of high voltage DC transmission, PD identification in such systems becomes more and more important. Therefore, a novel technique based on the analysis of ultra high frequency (UHF) resonance waveforms is proposed in this paper to meet the requirement. With the help of independent component analysis, the most dominating features are identified directly from UHF resonance signals without phase information. Using the identified features as the input, a neural network is implemented for recognizing sources of PD in SF
₆
and separating from the corona in air within a very short time