PSI - Issue 84

Alessandro Scala et al. / Procedia Structural Integrity 84 (2026) 489–496

495

lost its horizontality, displaying the features of the so-called “nutcracker mechanism” (Simeone et al. 2024). Additionally, the pier foundations were exposed, likely due to riverbed erosion, creating a potentially dangerous combination of slope instability and undermining. This critical state of structural distress is reflected in the coefficient of 0.516, indicating the need for urgent investigation and interventions. The third structure (Fig. 4c), representative of simply supported bridges interacting with a rapid landslide, is a girder bridge located in Northern Italy. During the census phase, no information was available indicating a potential interaction between the bridge and a landslide, resulting in a very low Level 0 coefficient (0.080). During the inspection, no particularly evident signs were observed on the structure itself, while the surrounding context displayed numerous indicators of precursor or currently active events. Notably, debris accumulations were observed near the foundations, a visibly unstable rock wall close to one of the abutments, and, most importantly, a superficial landslide that, if advancing, could interact with two piers. The detachment niche of the movement is clearly visible (Fig. 5c), with areas devoid of vegetation, allowing estimates of both the volume and direction of the movement. The only mitigation measures identified during the inspection, which do not directly interact with the other phenomena but highlight the predisposition of the site to instability, were bentonite coverings and a rockfall gallery at the opposite abutment. For these reasons, the PI pos t increased to 0.448. The fourth structure (Fig. 4d), representative of isostatic bridges interacting with a slow-moving landslide, is a large highway bridge located in central Italy. Maps indicate a large rotational/translational landslide, which, combined with the high traffic volume, resulted in a high pre-inspection index value. The inspection confirmed the information reported in the maps, with the observation of a clear detachment niche (Fig. 5d), although its extent suggests a landslide smaller than initially estimated. In addition, a slight rotation of the piers and a significant cracking pattern on the abutment affected by the phenomenon were observed. These findings, combined with the complete absence of mitigation measures or slope monitoring systems, resulted in the PI post of 0.425, still high.

Fig. 5. The most influential signs observed during the inspection of the cases studies: (a) the mitigation structures protecting Viaduct A; (b) the nutcracker mechanism observable on Viaduct B; (c, d) detachment niches under viaduct C and over Viaduct D. 5. Conclusions This paper addressed a key limitation of current landslide risk assessment frameworks for existing bridges, namely the difficulty of translating inventory-based classifications and qualitative inspection outcomes into effective prioritization strategies at the network scale. While the Italian Guidelines provide a robust multilevel structure for risk classification, the exclusive use of discrete Classes of Attention may lead to a flattening of priorities when large numbers of structures fall within the same class despite markedly different interaction scenarios. To overcome this limitation, a two-stage prioritization framework was proposed, introducing continuous priority indices defined before and after inspection activities. The pre-inspection index supports initial screening of bridges affected by mapped