PSI - Issue 84

Michele Morici et al. / Procedia Structural Integrity 84 (2026) 89–96

93

Fig. 2. Time series of rotations measured by inclinometers: a) sensor T1-b5y; b) sensor T2-b5y; and vertical displacement measured by LVDT d) sensor LV-b5.

a)

b)

Fig. 3. Bridge pictures portrayed in early January ‘25 in correspondence of the abutment B: a) debris in the proximity of supports and b) closure of the expansion joint.

In contrast, Figs. 4b) –e)–h)–k)–n) (T3- b5y) exhibit an increase up to approximately 10 °C, followed by a slight decrease. This behavior supports the interpretation that, at higher temperatures, rotations near abutment B are restrained. This suggest that the expected thermal expansion behavior of the bridge during the summer season may be affected by debris accumulation at abutment B, as well as and by boundary conditions associated with the slab–deck connection. Consequently, a purely linear regression model may not be sufficient to represent the observed temperature– response relationship, and the inclusion of nonlinear terms may be required to adequately capture the trends. This nonlinearity could be associated with the altered boundary conditions from those of a simply supported deck towards a more continuous structural behavior. The transition between these response regimes seems to occur around a temperature of 20 °C for T1-b5y and of 10° for T3-b5y. It is worth noting that the knots, i.e., the temperature values at which the slope changes, differ significantly among the sensors. This highlights the importance of accurately modelling and removing environmental effects and relating them to the specific local conditions identified on site. Figure 5 presents Hotelling’s multivariate control chart for beam #5. The chart was generated by evaluating the residuals of the rotations measured by tiltmeters T1-b5y and T3-b4y, together with a corresponding reconstructed signal obtained from the regression model.