PSI - Issue 84

Vanni Nicoletti et al. / Procedia Structural Integrity 84 (2026) 638–644

643

be achieved even when directly using modal force profiles instead of displacement modes. In this second case, data processing times for monitoring are drastically reduced, since reconstructing displacement modal shapes is no longer required, and monitoring costs are also lowered, as load cells alone can achieve the dual purpose of monitoring both

the stays and the overall bridge structure. 5. Results of stay-cable force monitoring

The bridge is equipped with a permanent SHM system, installed in 2022, capable of static, dynamic, and seismic monitoring, with more than three years of continuous data now available. A comprehensive description of the monitoring system and procedures may be found in Nicoeltti et al. (2023). For this study, the focus is on measurements from the four deck accelerometers (with a total of seven measurement directions) and the four load cells installed on the stay cables connected in proximity to the deck connection. Continuous 30-minute AVT recordings are stored on the server and processed using standard pre-processing procedures, followed by AutoOMA, which automatically identifies the modal parameters of the structure with the support of clustering analysis. Load-cell data undergo an additional post-processing step to compute the average axial force of each stay cable over the recording period. Analysis of the monitoring data shows clear seasonal and daily fluctuations in stay cable forces around their average values, which correspond to the design forces ( Fig. 5a ). Importantly, no sudden or significant changes are observed, indicating that no damage occurred during the monitoring period. Global bridge frequencies extracted from both load cell and accelerometer data are compared ( Fig. 5b and c ): modes identified by both sensor typology are matched, while modes detected by only one sensor are also noted (grey lines). Although frequency estimates from load cells show slightly higher variability, they closely align with the results from high-sensitivity, low-noise accelerometers. This validates the effectiveness of load cells for dynamic monitoring of cable-stayed bridges. Observed frequency trends exhibit both seasonal and daily variations across different modes, with no abrupt drops, further confirming the absence of structural damage during the monitoring period.

Fig. 5. Results of the first three years of monitoring: (a) load on stays obtained through load cells; (b) frequencies identified starting from accelerometer measurements; (c) frequencies identified starting from load cell measurements.

6. Conclusions This study presents an innovative SHM approach for cable-stayed bridges that uses load-cell data for both static and dynamic monitoring. High-frequency force measurements enable modal identification, and displacement-based mode shapes can be reconstructed from these force profiles using a compliance matrix derived from a calibrated finite element model. An optimal sensor placement strategy was also developed, combining static and dynamic criteria to determine the best number and layout of monitored stays. This approach minimizes an objective function based on stay damage sensitivity and modal force information, using evolutionary algorithms for optimization. The methodology was validated through extensive experimental measurements and numerical analyses on a real cable-stayed bridge in Central Italy, demonstrating its feasibility and efficiency. Overall, results show that load-cell-based SHM provides accurate static and dynamic monitoring while reducing costs and resource requirements, confirming its effectiveness as a practical, sustainable solution for long-term bridge management.