Study on the fractal characteristics of ultra-wideband partial discharge in gas-insulated system (GIS) with typical defects

In order to improve the correctness of the on-line monitoring and diagnosis of partial discharge (PD) for gas-insulated system (GIS), the characteristic of the typical PD defects in GIS is studied. Since PD in the insulation of GIS has a high frequency distribution and a wide frequency band, it is hard to analyze the characteristic of the PD by using conventional PD statistic analysis method based on IEC60270. So a new signal processing method called fractal analysis of ultra-wideband PD signals based on wavelet is introduced for analyzing the characteristic of ultra-wideband discharge in this paper. The fractal dimensions of PD signals of typical defects in each frequency-band can be calculated. So characteristics of PD signals can be expressed in numerical quantities with fractal dimensions. And we obtain some discrepancy of characteristic of different PD defects on surface of insulator. Moreover, it is the base for the further study to apply ultra-wideband PD measuring to the on-line monitoring PD for GIS.