ABSTRACT
Traditional IoT-based health monitoring systems for cable-stayed bridges can integrate multisource sensor data but often lack direct interaction with finite element (FE) mechanical models, thus relying on indirect model updating or manual model adjustments. To address this limitation, this study presents a streaming-based digital-mechanical twin framework that enables near-real-time coupling between sensor data and FE analysis with minimal manual intervention. An intelligent inspection vehicle equipped with RTK positioning provides dynamic load and locational data to the cloud, which relays essential parameters to a fog layer running ANSYS APDL. Updated FE results are then fed to a Unity 3D–based visualization platform, where C# scripting facilitates secondary representation (e.g., dynamic mesh construction, deformation logic, and color-mapped rendering) for comprehensive, interactive monitoring. Despite low sampling rates (<200 Hz) and moderate measurement errors (~13.64%) relative to wired sensors, the low-cost, IoT-based edge devices effectively capture mid- to low-frequency vibrations. Furthermore, 5G networks and MQTT ensure high-speed, bidirectional communication, supporting near-real-time system responsiveness.
