Application of ACF-wavelet feature extraction for classification of some artificial PD models of power transformer

Abstract

In this paper, 7 different artificial partial discharge (PD) models of power transformer defect types arebuilt in a high voltage laboratory, and then PD signals are recorded in the whole power frequency cycle to develop apreliminary PD knowledge base. A specific technique is used to extract single PD events from the recorded knowledgebase. By application of wavelet transform, different decomposition levels of extracted PD signals are investigated. Featureextraction is performed by application of statistic moments on the autocorrelation function of PD signal decompositions.“Best” features are selected based on the F-test. To evaluate the performance of features, a naive Bayes classifier is appliedto selected features. Results show that nearly 100 percent accuracy in PD discrimination is achieved using multivariatemultinomial distribution estimation of the feature space. A case study is carried out to show the contribution of theproposed approach. The combined PD detection-classification system proposed in this paper can serve as a complementto conventional PD monitoring systems.