A detection method for transmission line icing via improved K-means clustering
WENG Bingjun1, YANG Gengjie1, GAO Wei1, ZHENG Weicou2
1. College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108; 2. Ningde Power Supply Company of State Grid Fujian Electric Power Co., Ltd, Ningde, Fujian 352100
Abstract:Icing on transmission lines is a hidden danger that affects the normal operation of the power grid in winter. Therefore, it is essential to detect the icing status of the transmission line to find the icing phenomenon in time and then take countermeasures. This research proposes a method for detecting the icing status of transmission lines via improved K-means clustering. Firstly, images of transmission line are obtained and preprocessed. After that, line segment detector algorithm, improved K-means clustering, and least square fitting are used in turn to determine the position of the conductor. Finally, the width of the conductor is calculated to evaluate the icing status of transmission line based on the change in width before and after icing. Comprehensive analysis of the experimental results of the field and experimental environment, the proposed method can quickly and reliably identify the icing status of the conductor with high accuracy.