PSI - Issue 78

Valentina Picciano et al. / Procedia Structural Integrity 78 (2026) 1167–1174

1171

heterogeneous dataset, some predictions still exhibited significant errors, including an underestimation as high as - 49.6%, though overestimations were limited to about +16.1% (Fig. 3b). Overall, dividing the database based on concrete strength led to more homogeneous groups and substantially reduced prediction errors. Discrepancies exceeding 100–200%, related to considering the entire dataset, were lowered to within ±20% for overestimations and about -50% for underestimations, often resulting in conservative capacity predictions that support safety in structural assessment. These models demonstrated good predictive performance, offering a promising tool for preliminary assessments and prioritisation in large-scale bridge management systems.

(a)

(b)

Fig. 3. Regression results for fc > 37.15 MPa: a) experimental vs. predicted load capacity; b) residuals on test set.

4. Intervention techniques Due to ageing infrastructure, increasing traffic loads, and material degradation, many existing bridges with Gerber saddles now exhibit signs of structural distress and reduced load-bearing capacity. Strengthening interventions are therefore essential to ensure the continued safety, durability, and functionality of these bridge elements. In the literature, two main categories of strengthening techniques for Gerber saddles can be identified: those developed specifically for bridge structures and others more suitable for dapped-end beams commonly used in industrial buildings or precast structures. To facilitate the analysis, the techniques can be broadly classified into two groups: load redistribution methods and proper strengthening interventions. The first group includes methods aimed at transferring the load from the saddle region—typically where the simply supported span rests—to less stressed areas. This can be achieved by installing additional steel columns beneath the suspended beams to redirect the load to the foundation; by using steel brackets that transfer loads across the saddle, or through post-tensioned external steel cables anchored at specific points to achieve the desired load distribution. The second group includes interventions designed to restore or enhance the structural performance of the saddle by compensating for degradation over time. A direct approach involves the demolition and reconstruction of damaged areas, including the replacement or addition of reinforcement bars. However, due to its invasive nature and high cost— especially when full bridge closure is required—this solution is not always feasible. Less intrusive alternatives include the use of external post-tensioned high-strength steel bars anchored to steel plates placed above and below the undapped zone of the saddle. These bars enhance both ultimate strength and crack control, and their angle and anchor placement can be adjusted to mitigate specific failure modes (Fig. 4a). Other techniques include jacketing the saddle’s lateral surfaces with externally anchored steel plates using transverse steel bars (Fig. 4b), or bonding fibre-reinforced polymers (FRP or CFRP) in the form of plates or sheets to the external concrete surfaces (Fig. 4c). Each technique presents advantages and limitations, which must be evaluated case by case. Lateral surface interventions are often impractical due to the presence of transverse beams and curbs connecting longitudinal girders near the saddles. Additionally, they often require lifting the suspended span, interrupting traffic and increasing overall costs. In contrast, the use of post-tensioned steel bars is more versatile and quicker to apply—even in the presence of