Efficient Partial Discharge Detection in Online Gas Insulated Switchgear Monitoring: Characterization Insights
Loading...
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This paper focuses on investigating the detection of partial discharge (PD) in gas insulated switchgears (GIS) through the use of advanced deep learning models and signal processing methods. The aim is to effectively differentiate PD-induced noise from other noises in high-voltage environments using these sophisticated techniques. PD, often a precursor to serious faults, emits distinctive acoustic signals during its occurrence. The collaboration with Qatar General Water and Electricity Corporation (KAHRAMAA) enriches our research by providing valuable insights into real-world applications and specific considerations in GIS systems. The paper also includes a comprehensive overview of GIS systems, highlighting the complexities of their operations and the vital role of PD detection in ensuring their reliability. Furthermore, the study extensively explores various machine learning techniques, examining their effectiveness in identifying unique patterns in PD noise, thereby enabling quick and accurate detection of potential faults in the system. The paper also investigates several designs in the field of deep learning, including K-Nearest Neighbors (KNN). This model is taught to identify the distinctive features of noise caused by PD, setting it apart from other kinds of noise that are prevalent in high-voltage situations. The study also explores the useful uses of PD detection in preserving the integrity of GIS systems. It emphasizes how early detection of PD can save system failures, lower maintenance expenses, and increase equipment longevity. KAHRAMAA's real-world case studies demonstrate how successful these deep learning models are in practical environments. The paper concludes by discussing the future directions of this field of study. It raises the possibility of making model training better by utilizing bigger and more varied datasets and investigating cutting-edge deep learning methods.
Description
Keywords
Partial Discharge, Gas Insulated Switchgear, Ultra High Frequency