A smart monitoring approach based on decentralized digital twins

Citation
International Conference on Computing in Civil and Building Engineering, ICCCBE 2024
Contribution to Conference
International Conference on Computing in Civil and Building Engineering, ICCCBE 2024
Publisher Link
https://www.xcdsystem.com/proceedings/isccbe/JRvh2J7/presentation/89630.cfm?uuid=E58ED4C0-E623-A7C8-8B823D2663B5ACDD
Scopus ID
2-s2.0-105001239756
The abundance of data, collected by sensors installed in civil engineering structures, has been fostering smart monitoring approaches that leverage digital twins, i.e. virtual representations that “live” and “evolve” in tandem with the physical structures. Conventional digital twins usually reside in centralized servers, which also serve as data collection hubs of sensor networks, installed in the structures. Therefore, the centralized servers facilitate the data exchange between the structures and the respective digital twins, essentially constituting the “single source of truth” regarding the structural condition. However, centralized digital twin (DT) approaches are inherently vulnerable to malfunctions of the server, which may end up being a “single point of failure”. This paper aims to circumvent this problem by decentralizing digital twins into the microcontrollers of smart wireless sensor nodes. In particular, building on previous studies that utilize the embedded processing capabilities of smart wireless sensor nodes, a decentralized DT approach is proposed, in which the DT is segmented into partial digital twins that are embedded in the microcontrollers of the sensor nodes. By defining interfaces for data exchange, the wireless sensor nodes use the partial digital twins to collaboratively analyze the structural condition. Furthermore, since each wireless sensor operates as an independent data acquisition unit, the decentralized DT approach foresees alternative collaborative analysis routes in the event of failure of a wireless sensor node, thus enhancing the robustness of the DT. The proposed approach is validated through experiments using a laboratory test structure, in which the DT is represented by a finite element model. The results of the tests demonstrate the capability of the decentralized DT approach to evolve with the physical structure, thus serving as an attractive solution for modern smart monitoring systems.
DDC Class
600: Technology