Cluster pulse criteria for recognition of partial discharge

AbstractStatistical calculating programs simulating partial discharge appearance are developed and applied to evaluations of criteria for discrimination of the partial discharges from noises. According to the simulation, clusters of sequential pulses appear at the partial discharge inception voltage, and have never been utilized at the usual pattern recognition of the partial discharges. Two original indexes that are a pair pulse rate and a maximum sequential pulse rate are defined to describe the time sequential features of the partial discharge and their signal distinguishable capabilities from noises are evaluated. A new logical index describing the signal distinguishable capability of the usual ø‐q pattern of the partial discharges from the noises also is proposed and compared with those of cluster pulses. The maximum sequential pulse rate gives most sensitive criteria in the bounds of calculating conditions.A 66‐kV cross‐linked polyethylene cable with an artificial defect is used to detect cluster pulses. By adding twice as many noise pulses as measured partial discharge pulses, the usual ø‐q pattern criteria becomes powerless, but cluster pulse criteria are still informative