PSI - Issue 84

Vincenzo Gattulli et al. / Procedia Structural Integrity 84 (2026) 41–48

44

(Brownjohn, 2007). This architecture provides a scalable and robust framework for managing heterogeneous data streams and supporting near real-time operation of the Digital Twin. From a conceptual standpoint, the proposed Digital Twin can be interpreted as a multi-layer system composed of: (i) a physical layer, including the monitored structure and sensing devices; (ii) a data layer, responsible for acquisition, transmission and storage; and (iii) a model layer, where data are interpreted through structural analysis and system identification techniques. The interaction between these layers establishes a continuous feedback loop between measurements, data interpretation, and model-based assessment. Within this architecture, predictive analytics should be understood as an evolving capability of the DT framework, enabled by the progressive integration of historical data, system identification, and future model updating procedures. 3.1. Monitoring and Data Acquisition The monitoring system, which relies on distributed sensor nodes installed within the tunnel, performs local processing at the sensor or SHM card level to reduce data volume and optimize transmission efficiency, as appropriate, while simultaneously obtaining measurements of structural response and environmental conditions. Sensor nodes are connected via Ethernet local area networks with Power over Ethernet (PoE) to a local aggregation unit, referred to as the SHM Gateway. This configuration enables continuous data acquisition while ensuring a reliable power supply and stable communication in underground environments. 3.2. Data Aggregation and Transmission The SHM Gateway acts as an intermediate node that collects data streams from multiple sensor nodes and manages their transmission to the centralized platform. Preliminary data handling is performed at the gateway level to ensure stable communication between the underground monitoring network and external systems. Data collected at the gateway are transmitted through the network infrastructure to a remote server for storage and analysis. This workflow follows the reference architecture illustrated in Fig. 2, where sensing nodes send data to a local gateway, which forwards them to a centralized server. The architecture supports both continuous data streaming for near-real-time monitoring and reliable data transfer for subsequent offline analysis. This data flow forms the operational backbone of the Digital Twin, enabling continuous synchronization between the physical system and its digital representation.

Fig. 2. Reference architecture for sensor-based monitoring and data management adopted for the DT system at LNGS.

3.3. Centralized Storage and Analysis The remote server is responsible for storing, visualizing and analyzing the acquired monitoring data. Real-time data streams are used for event detection and reporting, while historical datasets are retained to enable offline analysis of structural behavior over time. The availability of historical data supports the identification of gradual changes in structural response that may lead to the development of anomalies. By combining distributed sensing, local data aggregation and centralized analysis within a single system, the DT provides a coherent digital representation of Hall B, directly linked to on-site physical measurements. Within this framework, the centralized platform serves as the analytical core of the Digital Twin, transforming raw sensor data into meaningful indicators of structural performance, enabling continuous assessment and supporting data-driven decision-making.