Partial Discharges Identification and Localisation within Transformer Windings

• Published in: IEEE transactions on dielectrics and electrical insulation 2020-12, Vol.27 (6), p.2095-2103
• Main Authors: Ali, N. H. Nik, Ariffin, A. Mohd, Rapisarda, P., Lewin, P. L.
• Format: Article
• Language: English
• Subjects: Clustering; Clustering algorithms; Coils (windings); Condition monitoring; Cross correlation; Current measurement; Insulators; Localization; Mathematical morphology; partial discharge (PD); Partial discharges; signal processing; transformer windings; Transformers; Waveforms; Wideband; Winding; Windings
• ISSN: 1070-9878; 1558-4135
• DOI: 10.1109/TDEI.2020.008706

This study proposes a technique that is based on the hypothesis that distinct PD sources have unique PD waveform characteristics at the point of measurement. Two different forms of PD pulses are injected into a model transformer winding at different points and by using wideband radio frequency current transformers (RFCTs), measurement data are obtained by positioning the RFCTs at the neutral-to-earth point of the winding and the bushing tap-point-to-earth. Mathematical morphology energy analysis using ordering points to identify clustering structure and signal cross correlation methods are utilized to extract information from the raw measurement data. Assessment of the relative performance of the analytical methods are examined and characterized in this study. The proposed approach allows automatic separation of PD data by source and localization of source site within the winding.