PARTIAL DISCHARGE RECOGNITION IN XPE CABLES USING ARTIFICIAL NEURAL NETWORKS

Iraida Kolcunová - Radoslav Haluška

   The submitted paper deals with partial discharge (PD) recognition in high voltage (hv) insulating systems using artificial neural networks (ANN). Electrical treeing and discharges in cavities were measured in the specimens of XPE cables. Results of partial discharge phase resolved analysis (PRPDA) were statistically processed and evaluated using discriminating analysis. To classified PD defects, artificial back-propagation networks (BPN) and restricted Coulomb energy system (RCE) were used.

Keywords: artificial neural networks, cavity discharges, diagnostics of XPE cables, electrical treeing, partial discharge classification, statistical operators

[full-paper]