Edge-fog-cloud-based digital twin network for autonomous and distributed structural health monitoring of a mega dam cluster.

Structural health monitoring (SHM) of mega engineering is huge, complex, and time-consuming. To address these challenges, this paper proposes an edge-fog-cloud-based digital twin network and provides its application on a mega dam cluster consisting of three dams along a river. Primary features of the network include an intelligent seismograph signal identification algorithm with Convolutional Neural Network (CNN) in the edge computing layer, a streaming finite element analysis (FEA) method for cumulatively simulating effects of water pressure and continuous seismic ground motion in the fog computing layer, and a real-time 3D virtual model visualization approach on Web driven by FEA response in the cloud computing layer. All processes are automated. Performance analysis indicates that the seismograph signal identification algorithm achieves an impressive accuracy of 95 %, virtual model spatial mapping deviation is only 5 %, and SHM processing speed is 9 times faster than the previous manual work. This digital twin network provides high-efficiency, autonomous and distributed SHM for the mega dam cluster, effectively minimizing labor costs, economic expenses and energy consumption. • Edge-fog-cloud-based distributed network. • Intelligent seismic signal identification algorithm using CNN for system triggering. • Streaming FEA of static water pressure and multiple dynamic seismic ground motion for high-fidelity simulation. • Real-time 3D virtual model visualization algorithm on the Web driven by FEA response. [ABSTRACT FROM AUTHOR]