Different types of Partial Discharge (PD) patterns have been proposed in the past to synthesize the information provided by thousand PD-pulse signals recorded during a measurement session. The efficiency in defect identification of each pattern type has been singularly verified but less work has been done to compare their performances. For this purpose, four types of PD patters have been evaluated in the identification of PD occurring in micro-voids distributed within the ground-wall insulation of new bars and coils for rotating machines. This choice is motivated by the rapid changes in PD activity during the transition from initial to steady-state conditions of the discharges within the voids. It was found that, despite some shape modifications, the Phase Resolved PD pattern provides the most effective identification among the four patterns considered in this investigation.
