Visual Inspection System for High-Voltage Switchgear based on Digital Twin

As a core component of substations, the safe and stable operation of high-voltage switchgear relies on efficient inspection methods. Addressing the inefficiencies of traditional manual inspections, this paper constructs a remote visual inspection system based on digital twin technology, enabling virtual-physical mapping and three-dimensional monitoring of switchgear status. Firstly, a multilevel digital twin system framework is established, and a hierarchical coding rule is designed to assign unique identification codes to switchgear components, ensuring a precise correlation between physical equipment and digital models. Secondly, a switchgear component state recognition detector called GB-FFCA-YOLO for is proposed, enhancing the mAP for small target component detection to 97.55%. Finally, a visual inspection system for high-voltage switchgear is developed, which integrates intelligent data acquisition terminals, three-dimensional models, and data analysis modules, supporting remote status verification and anomaly alerts. The system has been installed and validated on a 110 kV substation, demonstrating its practical engineering value.