Automatic Recognition of Periodic Pulse Shaped Interferences in Partial Discharge Online Monitoring

摘要

The partial discharge (PD) online monitoring is an effective way to timely detect the insulation faults in the high voltage equipments and it is significantly important to the safety of the power system. It is critical for PD monitoring to' detect the weak PD pulses from strong interferences. The periodic pulse interferences are produced by the periodic switches in the power system, and always occur at specific phase angles. The recognition of the periodic shaped interferences plays an important role in partial discharge online monitoring. In the paper based on the cluster analysis, the residual power and the instantaneous curvature are used for identifying the end of the pulses. The pulses can be clustered on the basis of the similarity of the pulse waveform by using the longest-distance algorithm. Thereafter, the periodic pulse interferences can be distinguished from the PD pulses by the pulse-phase histograms.