PSI - Issue 78

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

1430

Regarding the support devices, each beam features a reinforced elastomeric bearing on the right and a metal pendulum bearing on the left. The piers are made of ordinary reinforced concrete with a hollow cross-section, having external dimensions of 2.5 m × 7.7 m and a variable height ranging from 13.48 m to 47.24 m (Figure 1c). A defect analysis was performed on the structure, which revealed severe corrosion on the beams and piers (Figure 1b). An inspection plan has been designed to define the key tests and investigations required to accurately determine the geometry of the structural elements, identify the technical details, and characterise the mechanical parameters of the materials. Pending the execution of these tests, a simulated project has been conducted. This simulation, consistent with design specifications from the 1960s, serves as the basis for the Level 4 verification presented in this study. 3. Modelling structural behaviour and corrosion 3.1. Finite Element Model The FE model of the bridge was developed using Midas Civil software (Midas, 2024), Figure 2. Here, girders, crossbeams, and piers were modelled through beam elements, whereas plate elements have been used for the slab. For connections (i.e., bearings and beam-slab interactions), elastic links were calibrated adopting an equivalent stiffness to ensure an optimal model response. Additionally, the bearing bases were connected to the piers using rigid links. Abutments and foundations were modelled with fixed supports.

(a)

(b)

Fig. 2. (a) Axonometric view of the bridge model; (b) detail of the bearings.

As the piers were identified as the critical elements in the linear dynamic analysis, nonlinearlies were modelled only for them, so to avoid excessively high computational costs. To this end, a fibre element approach with distributed plasticity was employed (Figure 3). The element’s cross -section was discretised into a set of longitudinal fibres and each one was assigned a uniaxial nonlinear constitutive law representative of the material at that point (i.e., confined concrete, unconfined concrete and reinforcing steel). During the analysis, the response of the entire element was derived from the moment-curvature response of the sections which was obtained by integrating the contribution of the fibres (Taucer et al., 1991). For the nonlinear behaviour of concrete, the Kent and Park (1973) constitutive model, subsequently extended by Scott et al. (1982), was employed. This model is suitable for cyclic loading and allows for the quantification of section confinement. For steel, the Menegotto and Pinto constitutive model was utilised. This model incorporates isotropic hardening and is recommended for applications involving cyclic loading (Menegotto & Pinto, 1973). The parameters entered into the software, reported in Table 1 and 2, were calculated using the formulas reported in (Taucer et al., 1991). The fibre section, based on distributed plasticity, was incorporated at the Gauss-Lobatto integration points, where the element is subdivided, such that the first segment corresponds to the plastic hinge length.