AI-Driven Electrostatic Modeling for Improved Electronic Reliability: Case of Electrical Substations of Kyrgyzstan

This article presents the use of a neural network to model the distribution of the electrostatic field in existing electrical substations and stations in the Kyrgyz Republic. Data were collected using direct measurements (March 2021 – May 2023) at the 220/110/10 kV Davydov substation and the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">500kV</tex> open switchgear Toktogul hydroelectric power station; Today, the widespread use of electronic components increases the need to study reliability problems of electronic components, as well as control systems (CS) in general, and to find solutions to improve their noise immunity. During operation, control systems of various power equipment are subject to various electromagnetic interference. The paper describes an artificial intelligence system for detecting, classifying, and localizing electrostatic discharge (ESD) in electronic components of electrical equipment of an outdoor switchgear with a voltage of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$220$</tex> and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">500kV</tex> using deep neural networks based on CNN and LSTM models, using specialized software to obtain a large set data for training deep neural networks, and outlines the procedure for preparing the data necessary for training.