The Methodology for Overhead Power Transmission Lines Technical State Assessment Based on Machine Learning

• Published in: 2023 Belarusian-Ural-Siberian Smart Energy Conference (BUSSEC) pp. 94 - 99
• Main Author: Stepanova, Alina I.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 25.09.2023
• Subjects: Boosting; decomposition model; Documentation; High-voltage techniques; Insulation; machine learning application; overhead power transmission lines; Poles and towers; Power transmission; Power transmission lines; technical state assessment
• DOI: 10.1109/BUSSEC59406.2023.10296227

The paper presents the methodology for assessing the technical state of high-voltage power transmission lines, based on the decomposition method with the application of gradient boosting on decision tress. The decomposition method presupposes separate assessment of the technical state of the different structural parts of the high-voltage power transmission line: insulation, tower, span, tower footing, high-frequency filter. The total score of the structural parts' technical state is calculated by aggregation of the sub-elements' values of each structural part. The article demonstrates the application of the decomposition method, based on data obtained as a result of technical diagnostics of the high-voltage power transmission lines of 110 kV. As a result of the assessment of the technical state of the high-voltage transmission line, editorial changes were noted for the formation of the database. The paper highlights the existing barriers in regulatory documentation in the field of determining the boundaries between different technical states. The authors provide suggestions and recommendations on regulatory documentation improvement to foster the implementation of intelligent monitoring systems in power grids.