Year: 2025
Type: conference-paper
Abstract: Electrical equipment plays a critical role in industrial electrical systems, where unexpected mechanical failures during operation can lead to severe consequences, such as disruptions in technological processes, reduced quality and quantity of produced goods, and the occurrence of emergency situations. To ensure the reliable operation of these systems, it is essential to conduct regular inspections and maintenance of electrical equipment using modern computer technologies for the timely detection of faults. To address this issue, a comprehensive approach is proposed, based on the processing of color images obtained from thermal imaging devices, black-and-white graphs generated from sensor data stored in Excel format, as well as simple color photographs. The core of the image, graph, and photograph processing (recognition) lies in artificial intelligence technologies, specifically convolutional neural networks (CNNs). The developed CNN outputs a class number corresponding to the current state of the equipment. Additionally, an algorithm for a comprehensive approach to fault detection in electrical equipment using CNNs is proposed. The results obtained allow for a reduction in the number of unplanned failures by providing early warnings about the development of faults and potential equipment failures..
Authors: A.E. Kolodenkova, S.S. Vereshchagina
Institution: Samara State Technical University
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Funder: Russian Science Foundation
OpenAlex ID: W4410317145