PSI - Issue 84
Ana Avramova et al. / Procedia Structural Integrity 84 (2026) 560–568
567
Table 3 Comparison between the identified frequencies of VI03 and VI04 viaducts. Mode f VI03 (Hz) f VI04 (Hz) Δ f (%) f 2 −f B 1 1.541 1.410 − 8.50% 0.781 T 1 1.825 1.678 − 8.05% 0.968 B 2 1.945 1.860 − 4.37% 0.903 T 2 2.193 2.112 − 3.69% 0.970 B 3 2.481 2.405 − 3.06% 0.957 T 3 2.684 2.636 − 1.79% 0.965 B 4 2.869 2.734 − 4.71% 0.873 T 4 3.029 2.946 − 2.74% 0.965 B 5 5.299 4.820 − 9.04% 0.921 T 5 5.578 5.159 − 7.51% 0.972 B 6 5.963 5.520 − 7.43% 0.856 T 6 6.185 5.805 − 6.14% 0.951
Fig. 9 MAC matrix between the reference mode shapes of VI03 and VI04 viaducts.
In addition to the frequency comparison, the similarity between the mode shapes of bridges VI03 and VI04 was evaluated through the MAC parameter. The MAC matrix shown in Errore. L'origine riferimento non è stata trovata. compares the mode shapes identified in the reference datasets for bridges VI03 (05/06/2024, 16:00-17:00) and VI04 (08/06/2024, 16:00-17:00). Figure 9 confirms the expected correspondence of the dynamic characteristics of the two similar structures, with most of MAC values being above 0.91. 7. Conclusions The paper illustrates the OMA-based strategy adopted for the SHM of two similar steel-concrete composite viaducts. Both bridges—referred to as Olona East (VI03) and Olona West (VI04) in the paper—includes 4 spans and belong to Autostrada Pedemontana Lombarda in Northern Italy. The results collected in about 16 months of continuous dynamic monitoring allowed the following main conclusions: 1. Within the frequency range of 0-7 Hz, a large number of normal modes (12 for VI03 and 14 for VI04) were automatically identified and tracked during the addressed monitoring period. 2. For both bridges, the air temperature turned out to be a dominant driver of daily and seasonal fluctuations in the natural frequencies of all modes, with torsion modes generally exhibiting a larger dependence on temperature variations than bending modes. 3. As the mode shapes of the two viaducts were not affected by EOVs, a mode shape-based metrics was used to check the invariance of the structural condition since the beginning of the monitoring activity. 4. A PCA-based cleansing strategy effectively minimised the impact of EOVs, and the application of Hotelling’s T² control charts on the residuals demonstrated that no significant structural anomalies occurred during the monitoring period. 5. The comparative analysis between the dynamic characteristics of the two bridges indicates an almost linear relationship as well as the possibility of accurately predicting the frequency variations in one bridge from the knowledge acquired on the other bridge. Acknowledgements The support of Autostrada Pedemontana Lombarda (APL) − funding the PhD Scholar ship of the first author − is gratefully acknowledged.
Made with FlippingBook flipbook maker