Neural‐network‐based approach for secondary arc recognition on HV overhead lines

Abstract

AbstractAn alternative approach to the present situation for automatic reclosure of high‐voltage (HV) lines, in particular problems of repetitive reclosure onto permanent faults and lack of wider system co‐ordination, is possible thanks to modern technology. In order to determine the appropriate response to single‐phase‐to‐ground fault, it is necessary to measure conditions on the HV line during the fault and adapt the response to correspond to the most favoured outcome. This paper presents a numerical algorithm for blocking autoreclosure during permanent faults on overhead lines. It is based on one‐terminal input data processing. The fault nature (permanent or transient) is determined from the voltage and current signals of faulted and tripped line phase using neural network approach. The algorithm presented was successfully tested using the signals recorded on the real power system and signals obtained through laboratory tests. The tests demonstrate the ability of the presented algorithm to determine the secondary arc duration. The algorithm can be successfully used to define the recloser reclaim time and blocking automatic reclosing of power lines with a permanent fault.