GIS PD Type Recognition Algorithm Based On Wavelet And Its Implementation

The most common feature of Gas Insulated Switchgear (GIS) electrical fault isbefore the insulation was completely broken, a partial discharge (PD) occurs.Through the online monitor of GIS PD, we can find the interior defect of GIS in timeand avoid serious accidents. It is significant to research on the GIS PD monitor andthe method of its pattern recognition. It is important not only for the safe and reliablerunning of GIS, for the diagnosis of GIS insulation and the character of its defecttype, and also for the repair of GIS.Based on recent research results of domestic and international GIS PD detectionand recognition, this paper proposes a method of GIS PD type recognition algorithmbased on wavelet. transform Paper introduces Mallat algorithm of wavelet and liftingwavelet algorithm briefly, discusses the hardware of the online monitor of GIS PDand the problem of GIS PD signal de-noising; gives the method of GIS PD grayimage construction; decomposes four kinds of typical defect of GIS PD gray imageusing db3wavelet and haar wavelet, and uses the method of coarse grid for GISPD gray image feature extraction, establishes the feature library; uses minimumdistance classifier to identify defect type of GIS PD. This paper tests the recognitionalgorithm and the test shows that the algorithm is effective for four kinds of defect ofGIS PD.