Conductor Aging Monitoring and Intelligent Analysis Strategies for Power Grid Status Visualization
Abstract
INTRODUCTION: Continuous sensing, unified analysis, and spatial representation of conductor aging remain challenging in power-line monitoring involving computer vision, edge computing, and digital twins.
Objectives
This study develops a cloud-edge collaborative framework for power-grid status visualization to improve conductor defect detection and aging-risk assessment under complex operating conditions.
Methods
The sensing layer collects visible-light, infrared, electrical, vibration, and micrometeorological data; edge nodes perform time alignment, quality assessment, lightweight defect inference, and event filtering; and the cloud establishes conductor digital-twin instances through object-level coding.
Results
The proposed lightweight network achieves an mAP@0.5 of 96.05% with a single-frame inference time of 16.3 ms. The aging-state recognition accuracy reaches 95.16%, while the health-index mean absolute error is 0.041, the 7-period forecast RMSE is 0.047, and the risk-classification F1 score is 94.12%.
Conclusion
Dynamic overlay visualization reduces the anomaly localization time to 34.7 s and increases the task completion rate to 96.5%. Quality-aware fusion and digital-twin object mapping establish a closed-loop connection among multi-source monitoring, state prediction, and 3D interactive decision support.
