Partial discharge detection and localization on the medium voltage XLPE cables with multiclass support vector machines

Abstract

In medium voltage cables, partial discharges (PD's) are the major problems that trigger electrical insulation failures. Therefore, classification of PD source type and failure localization in medium voltage cables are significant issues of medium voltage engineering. Therefore, in this study, both detection and localization of PD are studied. As a first step, 4 different kind of defects are artificially generated at the same length of the same kind of medium voltage cross-linked polyethylene (XLPE) cables. Consequently, an experimental setup is built. During the experiments, different medium voltage levels are applied to the cables, then the PD signals are measured and recorded. To classify the signals of different defects, different statistics of frequency spectrum of the signals are considered as features. As a final task of this step, multiclass support vector machine is employed and the PD signals are classified. In the second step, one kind of defect is generated at different locations of same kind of longer XLPE cable. Consequently, the cable exposed to different medium voltage levels and PD signals are measured and recorded. The statistics of the data are employed as features. Finally, PD signals measured from different lengths are classified by the help of multiclass support vector machine.

Partial Discharges Classification Methods in XLPE Cable: A Review

“…The category sample of PD data was classified during multi-level SVM training as one class and the other samples as another class [68]. Serttaş [77] proposed multi-class SVM to classify PD defects with additional statistical features as a parameter of the SVM. The best part of the research is that no noise filtering was applied, and all the data used were based on the measuring signal desired in the signal processing technique.…”