Partial discharge detection and type identification of typical defects of XLPE cables under AC voltage

Abstract

According to the discharge characteristics of the typical defects of power cables, four common physical models of cable simulation defects are designed and produced under laboratory conditions. The high-frequency current method is used to measure the discharge signals of the four common cable simulation defect models, the data is collected and stored, a two-dimensional discharge map is drawn, and the support vector machine method (SVM) is used for classification.The results show that: the signal distribution range of levitation discharge is the widest, and the signal distribution is long; the tip discharge only generates the discharge signal in the negative half of the cycle; the negative half of the creeping discharge is more dense than the positive half of the cycle; the number of discharges in the positive and negative half of the internal air gap discharge is roughly the same ; discharge of different types of defects is obviously different. Extracting statistical features from its two-dimensional spectrogram for identification, the identification accuracy rate is above 90%, and different types of defects can be effectively identified.