Classification of partial discharge using PCA and SOM

Partial discharge (PD) is harmful to the insulation of electrical power equipment. Classification of PD plays an important role in determining the level of harmfulness for the PD. In this paper, the PD patterns in the forms of univariate phase-resolved distributions are analysed. An alternative method to that using statistical moments for characterizing the PD patterns is proposed. Principal component analysis (PCA) is used for the purpose of feature extraction and dimensionality reduction. Self-organizing map (SOM) is used as the tool for better visualisation of classification for different types of PD.