Cluster Classification of Partial Discharges in Oil-impregnated Paper Insulation

Abstract

Recognition of multiple partial discharge (PD) sources in high voltage equipment has been a challenging task until now. The work reported here, aims to recognize multiple PD sources in oil-impregnated paper using Cluster Analysis (CA) and Fuzzy Logic (FL). The typical sources of PD in transformer are identified and the corresponding single source PD defect laboratory models are fabricated. From the measured PD signals, the necessary statistical parameters are extracted by applying CA for classification. A Fuzzy based algorithm has been developed to recognize single source PDs. The developed algorithm has also been applied to recognize multiple PD sources.

Keywords

• PDs
• multiple PD sources
• oil-impregnated paper insulation
• PD classification
• cluster analysis