Feature Extraction of Partial Discharge Signals of Cable Terminal in Distribution Network Based on Discrete Level Acquisition Technology

Partial discharge on-line monitoring device can detect and identify partial discharge signals, which can be used as an important basis for evaluating the insulation condition of cable terminals. However, the accuracy of the signal feature spectrum is mainly limited by the sampling bandwidth of the high-speed data acquisition board, which is expensive and difficult to be widely used. To solve the above problems, an online partial discharge monitoring device based on digital acquisition technology of discrete level is designed in this paper. The continuous partial discharge signal with frequency up to tens of megahertz can be converted into discrete level signals through the signal processing module, which is transmitted to FPGA for reconstruction analysis. In order to realize the effective and accurate identification of partial discharge types of cable terminals in the distribution network, four typical partial discharge models of cable terminals are designed. The PRPD and $\Delta t$ characteristic spectra under different discharge models are constructed through a large number of discharge tests. Feature parameters are extracted from the fractal feature, moment feature, and texture feature respectively, which can be used to identify PD types effectively.