Měřicí zařízení pro detekci a klasifikaci částečných výbojů

Abstract

Partial discharges serve as an indicator of insulation degradation. They have a cumulative nature, and prolonged exposure can lead to failures, financial damages, and in extreme cases, even injuries. This thesis focuses on the development and implementation of a system for detecting and classifying partial discharges in the power grid. The proposed solution utilizes edge computing, which enables us to assess the presence of partial discharges using a neural network directly on-site. In case of partial discharge detection, an automatic alert is sent to the control center. Unlike current solutions, our approach is wireless, contactless, and battery-powered.