PSI - Issue 78

Lorenzo Ciccarelli et al. / Procedia Structural Integrity 78 (2026) 1428–1435

1433

4. Seismic vulnerability assessment and influence of corrosion After modelling the primary elements and assigning their mechanical material parameters, the subsequent step involved determining and applying the loads. Within this process, the self-weights of the structural elements and the variable traffic load were automatically calculated by the software in accordance with NTC2018, while the non structural permanent loads were manually computed and then introduced as nodal and surface forces. These actions were considered for the safety assessment under traffic loads through a linear static analysis. Additionally, a preliminary seismic vulnerability assessment was carried out by performing a modal analysis and applying the response spectrum method, using a behaviour factor, q, equal to 1.5. The linear static analysis was not satisfied for the "adequate bridge" conditions, i.e., those related to the indications of NTC2018, with the loads and partial factors foreseen therein. Following the guidelines, the analysis was repeated under "operational bridge" conditions, where the partial safety factors are reduced, considering a reference period (time for which the verifications are valid) of 30 years. This latter verification was found to be satisfied. For a detailed description of the aforementioned conditions, the interested readers is referred to (Gara et al., 2025). For the dynamic analysis, the verification was also not satisfied, with a safety factor, ζE , of 0.33. This outcome was expected, considering the bridge’s design period and the high seismicity of the area. The elements identified as critical are the piers, particularly in the biaxial bending check. Following the initial linear static and dynamic analyses under operational and seismic loadings, a nonlinear dynamic analysis was conducted to better assess the seismic performance of the bridge. To achieve this, the target spectrum corresponding to the br idge’s coordinates was used to generate three groups of artificial spectrum compatible accelerograms. Each group consists of one accelerogram in the bridge’s longitudinal (x) direction, one in the transverse (y) direction, and one in the vertical (z) direction. The same approach was used to assess the influence of corrosion, by repeating the analysis for the three scenarios, i.e., NC, C1 and C2. In all three cases, the section did not meet the verification requirements, as the ultimate strain of the concrete was reached. As shown in Figure 5, failure initiated at a corner and propagated in different directions depending on the model. For models NC and C2, the failure predominantly developed along the Y direction, while for model C1 it occurred along the Z direction.

(a)

(b)

(c)

Fig. 5. Failure evolution: (a) NC section at 6.58, 7.00 and 7.81 seconds; (b) C1 at 5.12, 5.62 and 6.21 seconds; (c) C2 at time 4.29, 4.40 and 5.50 seconds.

However, the most significant finding is that the introduction of degradation led to an earlier onset of failure. Figure 6a shows the failure instant mapped onto the accelerogram, further highlighting the increased vulnerability of the degraded structure. Specifically, in the model NC, failure occurred at load step 658, corresponding to 6.58 seconds. In C1, failure occurred at load step 512 (5.12 seconds), while in C2, more severe degradation scenario, failure was reached at load step 429, corresponding to 4.29 seconds.