PSI - Issue 84

Diego Esposito et al. / Procedia Structural Integrity 84 (2026) 1198–1205

1201

Compliance with the original design documentation was verified through direct measurements of the girder and typical deck geometries, GPR surveys, Mazzatura et al. (2024), and exploratory inspections of the anchorage zones Bernardi et al. (2022) to confirm the layout of the prestressing tendons. This activity revealed substantial correspondence with the original design, confirming the presence of six prestressing tendons for each girder, composed of 12 strands of 0.6 in. A total of 18 endoscopic inspections, Rapaport (2024) were performed to tendon state assess. These revealed highly heterogeneous conditions: in some sections, grout injection was absent or incomplete, while the prestressing tendons, although showing oxidation phenomena and, in two cases, initial signs of pitting, did not exhibit significant cross-section reductions. The evaluation of the residual stress of prestressing tendons was carried out through detensioning tests. Six strands from different tendons were checked. These tests allowed estimation of the residual stresses of prestressing tendons. Recorded values ranged between 672 MPa and 923 MPa, except for an anomaly detected in tendon 2 of girder 8, where a stress of 177 MPa was measured. The results are reported in Table 1, indicating the percentage loss of prestress compared to the design value of 996 MPa. The diagnostic investigations confirmed that, despite the presence of visual anomalies, no immediate safety-critical conditions were detected in the prestressing system at a local level: no broken wire or significant reduction in cross sectional areas were detected in the prestressing reinforcement.

Table 1. Summary of the outcomes of detensioning tests and stress comparison with theoretical value (from design documentation).

Deck

Beam

Sect.

Tendon

Stress σ (MPa)

Stress losses (%)

1 4 4 9 9

1 8 8

¼L ¾L ½L ¼L ½L ¾L

3 6 2 3 2 1

923 672 177 872 909 888

7%

33% 82% 12%

17 17 50

9%

25

11%

4. FEM and static load test definition The FEM of a single span was carried out to design the load test. The deck was schematized using a grillage composed of frame-type elements operating in the linear elastic range. The model consists of longitudinal and transverse elements: the former represents the main girders, modelled using a section designer to define a box-girder section that includes the collaborating portion of the slab and lateral edge beam (Figure 4). The end diaphragms were modeled with real characteristic section, as well as elements representing the transverse slab. The slab elements were modeled by a strip 1.0 m wide, but assigned zero weight and mass to avoid duplication of loads and masses already considered in the box girder elements. The function of the slab elements is to distribute the truck axle loads of the truck and make the connection between the two box girders Given the extensive flexural cracking observed in the girders, it was deemed appropriate to adopt a modelling approach that accounted for the detected defects and to evaluate the actual structural behavior of the bridge. The initial mechanical properties were defined using the concrete characteristics estimated from the diagnostic tests. These values form the basis of the theoretical model referred to as E100%. The effect of degradation was considered through a progressive reduction of the material’s elastic modules; two additional degradation levels were introduced by reducing the stiffness of the model, identified as E70% and E50%. The load test was designed to achieve a maximum midspan bending moment equal to approximately 85% of the theoretical value from the original design, as inferred from the applicable standard at the time, CSLLPP (1962). All structural checks were performed to ensure test safety, Italian Ministry of Infrastructure and Transport (2018, 2019). Element strengths were calibrated based on the extensive prior investigations, which confirmed the satisfactory condition of the prestressing tendons. The test was conducted using four 4-axle trucks, each weighing approximately 40 t, arranged in eight configurations shown in Figure 5.