Abstract
Partial discharges (PD) within the dielectric of High-Voltage solid dielectric power cable can lead to serious insulation damage and reduce the cable lifetime. Therefore, it is important to detect and locate the PD in the cable before breakdown occurs. In this paper, we propose a study on the detection and analysis of internal PDs within a polyethylene-insulated cable used under DC voltage. This work aims to build a database from PD measurements and to build classifiers using machinelearning techniques to recognize the degradation state of cables under High-Voltage Direct Current (HVDC) conditions. For this purpose, an original experimental setup is implemented. The experiments are performed on a long length 100 m coaxial cable subjected to high electric fields. PD events are detected by direct coupling and collected with a digitizer. Multiple insulation failures can be located, both at the cable boundaries and in the cable that reveal intrinsic defects in cable insulation. The PD signals associated with these two types of failures can be distinguished from each other. From experimental PD measurements, variables are extracted to establish a database. The implementation of a supervised classification method permits a good recognition of two different ageing states, or degradation levels of the cable insulator.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call