Dynamic Frequency-selection Clustering of Automatic Multiple Source Separation based on UHF PD Detection

Abstract

Partial discharge (PD) ultra-high frequency (UHF) on-line monitoring technique is an important resort to evaluate the insulation condition of the high-voltage power equipment. The presence of a large number of on-site interference affects the detection sensitivity and reliability, and the interference generated by discharges are the crucial bottleneck for effective PD detection under complicated electromagnetic environment, because they have similar time-frequency characteristics as real PD. Therefore, in order to solve the problem of mutual existence of multiple discharge and their interference to each other, a method of auto separation of multiple PD is studied in this paper, and the rules and a combined strategy is presented to make accurate multi-PD separation. The technique of dynamic automated separation of multi-PD is developed based on digital RF chip by using these rules, then theoretical and experimental verification is carried out. The results indicate that the clustering technology presented in this paper could realize automated separation of multiple PD and its accuracy can up to 90 %.