Ultra-Low-Power Multi-Sensor Fusion-Based Online Monitoring System for Fault Detection in Power Transmission Line Strain Clamps

Abstract

Failures of strain clamps on power transmission lines can undermine power system stability and disrupt electricity supply. To improve fault diagnosis accuracy, this study proposes an online monitoring approach based on a micro-power sensing device, employing a collaborative multi-sensor fusion framework for operating condition identification. Thermosensitive resonators, torque sensors, and strain gauges are used to collect temperature, resistance, torque, and deformation data, which are then fused using Support Vector Machines (SVM), Decision Trees, and Dempster–Shafer (D–S) evidence theory for fault classification. Laboratory tests involving 25,200 cases across seven measurement points showed that the proposed fusion model achieved an accuracy of 94.8% and an F1-score of 94.4%, representing improvements of 8.84% and 10.9%, respectively, over single-sensor models and outperforming baseline approaches. Field validation on 66 kV transmission lines operated by the State Grid Siping Power Supply Company confirmed its effectiveness in detecting major fault types such as clamp cracking, deformation, and bolt loosening. Furthermore, the predictive-assisted wake-up strategy effectively balanced energy consumption and response time, extending system endurance. Overall, the proposed method enhances diagnostic accuracy and energy efficiency, providing a practical solution for long-term online monitoring and smart grid applications.