Recognition of Surface Discharges in High-voltage Lines Insulators using Artificial Intelligence

Abstract

Environmental contamination and pollution under specific meteorological conditions can cause surface discharges on the insulators in high-voltage lines and stations, with consequent possible service interruptions. In this paper, it is presented the first attempt of a novel approach in the detection of surface discharges, based on AI and computer vision state of the art techniques. The research is carried out in the frame of an active collaboration with TERNA (the Italian TSO). The videos collected by the camera of the LANPRIS testing station for the purpose of ageing monitoring, have been used to explore the potentials of an object detection approach using an artificial neural network to recognize insulators in a video and to distinguish between normal behaviour insulators from those subject to surface discharge phenomena. The input data are described in the context of the LANPRIS experimentation, the model, chosen for the object detection, and the pipeline, built to process the video files, are introduced; the tools, used for this work, are widely illustrated and the results discussed. This study is the initial part of a larger work aimed at experimenting artificial intelligence and computer vision techniques in systems that monitor very important electrical system components, such as high-voltage line insulators.