Classification of defects and evaluation of electrical tree degradation in cable insulation using pattern recognition method and weibull process of partial discharge

The purpose of this paper is to recognize partial discharge (PD) sources and evaluate of electrical treeing degradation for cable insulation. To acquire PD data, three defective tree models were made. And the data are shown by the phase-resolved partial discharge method (PRPD). As a result of PRPD, tree discharge sources have their own characteristics. If other defects (void, metal particle) exist at internal power cable, their characteristics are shown very differently. This result is related to the time of breakdown and this is important point of cable diagnosis. To apply PD data classification, methods of different type are selected. Those are, multi layer perceptions (MLP-BP), adaptive neuro-fuzzy inference system (ANFIS) and principle component analysis-linear inference system(PCA-LDA). As a result, ANFIS shows the highest rate which value is 100 %. Finally, we performed classification of tree progress using ANFIS and that result is 99 %. To evaluate degradation of electrical tree, weibull distribution was used. The time of each degradation stage (initiation, middle, breakdown) was measured to classify electrical tree degradation with each model by. Using the result, parameters are presumed by the each model and stage and it is possible to calculate time difference of each degradation stage and estimate the lifetime.