Partial Discharge Pattern Recognition Method of Distribution Cabinet Equipment Based on Higher Moment Feature

Switchgear, ring cabinet, cable branch box and other cabinet equipment are the important components for high voltage distribution network. Its stable operation can stabilize the distribution network system. Reliable power supply is of great significance. Partial discharge (PD) is the main cause of switchgear failure. At present, transient earth voltages (TEV) detection method is one of the more effective partial discharge detection methods for switchgear. But the accurate pattern recognition of partial discharge signal is still one of the more difficult research tasks. Based on the traditional waveform features, moment features, typing features, statistical features and wavelet methods, this paper proposes to add higher-order moments as feature parameters, using support vector machine and probabilistic neural network as classifiers for comparative study. The typical discharges such as tip discharge, surface discharge, internal air gap discharge and suspension discharge were simulated on the switchgear in the laboratory, and 50 groups of data were collected for analysis. The experimental results show that the recognition accuracy of typical discharge types in switchgear is 86.25% and 88.75% respectively when classifier is probabilistic neural network and support vector machine. They are 15% and 27.5% higher than that of traditional five identification parameters. It can be seen that the identification accuracy can be effectively improved by incorporating high-order moments into the characteristic parameters. Therefore, this method has guiding significance for the identification of typical partial discharge signals.