PSI - Issue 62

Francesco Mariani et al. / Procedia Structural Integrity 62 (2024) 955–962 Mariani et al / Structural Integrity Procedia 00 (2019) 000 – 000

956

2

1. Introduction The structural performance of prestressed segmentally constructed concrete bridges is heavily influenced by time-dependent phenomena such as concrete shrinkage, creep, and steel relaxation. Prediction of vertical deflection in these structures becomes a complex task due to the combination of loads, time-varying material properties, and accumulated deformations, that can impact serviceability, durability, and reliability (Di Luzio et al 2015, Madsen et al 1983). The rheological properties of concrete, particularly shrinkage and creep, typically contribute to the long term total deformation (Danon et al 1977). The uncertainties associated with material properties and FE models make quantifying all these effects challenging. Numerous studies address uncertainty handling in this context, emphasizing the importance of incorporating experimental data within models to improve the accuracy and reliability of predictions ( Bažant et al 2018, Galassi Sconocchia et al 2023). In cast-in-place segmental bridges, all these issues can significantly impact the normal stress distribution, particularly when the girder continuity is achieved by post-imposed constraints, leading to incremental positive bending moments (Cho et al 2008). Additionally, the beams undergo prestress losses caused by the reduction of the con crete elements’ volume combined with the steel relaxation phenomenon (Barthélémy et al 2015). Finite Element (FE) models are commonly employed to perform long-term analysis aiming to predict the response of such structures. Literature studies suggest that reliable results require analyses that consider construction stages, static and loading configuration variations, rheological effects, and variable concrete properties depending on aging (Cruz et al 1998, Joyklad et al 2022). This paper focuses on understanding the influence of different types of in situ tests on the reduction of the uncertainty related to the complex time-varying response of a bridge. A 40-year-old multi-span post-tensioned box girder viaduct, constructed using a segmental cantilever approach, with four vertically prestressed internal joints in the middle span, is adopted as the case study. Visual inspections and digital laser scanning surveys reveal significant vertical deflection, attributed to the underestimation of delayed phenomena, potentially impacting serviceability. To comprehend the viaduct’s actual conditions, a FE model is constructed based on design documentation and numerical analyses are carried out incorporating prestress steel relaxation, concrete creep, and shrinkage. To preliminary simulate the uncertainty of materials mechanical properties, a sensitivity analysis on concrete elastic modulus variation is performed to investigate its influence on deformation and modal properties. The results are compared to those obtained with elastic modulus resulting from core compression tests. Additionally, Operational Modal Analysis, considering results from Ambient Vibration Tests, is performed to compare experimental modal properties with those obtained from the FE model in correspondence of different values of elastic modulus. Results allow to understand the pivotal role of field tests in enhancing the reliability of predictions by mitigating uncertainties related to material properties, especially the elastic Young's modulus of concrete that significantly influences both long-term deflection and modal features. 2. The case study viaduct The chosen case study, displayed in Figure 1(a), is a post-tensioned concrete box girder viaduct with ten spans. Situated in central Italy, construction started in 1984 using the balanced cantilever method. The spans measure 70 m in length, except for the first and last spans, which are 35 m long, resulting in a total length of approximately 630 m as represented in Figure 2. The substructure comprises two abutments and eight piers, varying in height from 7,5 m to 18,5 m to follow the ground slope. The superstructure is made of a continuous box girder connected to the substructure through fixed rolling and sliding supports, alternately. Continuity between the cantilevers is ensured by four vertical prestressed internal joints located at spans n.3-5-7-9. These elements, shown in Figure 1(b), connect the upper and the lower segments of the box girder, accommodating thermal dilatations by two systems of sliding supports and ensuring the transfer of shear stresses. Bending moments are transmitted through 30 prestressed vertical Dywidag bars situated near the lower supports system. The prestressing system comprises three sets of tendons positioned at different heights within the box girder, each consisting of 24 wires with a diameter of 8 mm. The first group, consisting of over 600 tendons, is located between the webs and the upper slab. These tendons are tensioned after two symmetrical casts to minimize deformation under dead load during construction stages when the independent girders act as cantilevers. A second group of 160 tendons are positioned in the upper slab, while the last 180 tendons, which constitute the continuity tendons system, are situated inside the lower slab and counteract