Research on Feature Extraction and Monitoringearly Warning Technologies for Overhead Line Galloping

Galloping of overhead lines is a low-frequency, large-amplitude vibration phenomenon that occurs under external excitation, with a long duration. It is especially prone to occur under conditions of ice accumulation and strong winds. To effectively monitor the operational status of overhead lines, this study proposes a feature extraction method for galloping and a monitoring and early warning system. Firstly, a weatherbased line galloping grading early warning method is proposed, utilizing generalized linear regression networks and a particle swarm optimization-based support vector machine (PSO-SVM) algorithm. Secondly, the DeepLabv3+ network is employed for the segmentation of overhead conductors, and by locating the conductor's center of mass, the galloping trajectory is extracted. The displacement variation curve and amplitude-frequency characteristics of the galloping are then obtained, enabling realtime monitoring of the overhead lines and verifying the reliability of the monitoring system. Finally, based on these methods, an online monitoring system for overhead line galloping is designed, integrating 5 G communication technology and multi-source data fusion. The results show that the system exhibits high accuracy and timeliness, making it suitable for large-scale application in engineering. This study provides a theoretical basis and technical support for the development and implementation of galloping prevention and control technologies for overhead lines.