PSI - Issue 78

Andrea Nettis et al. / Procedia Structural Integrity 78 (2026) 1404–1411

1410

By transitioning from fragility to vulnerability curves, loss curves can be constructed. As shown in Fig. 4, the loss curve corresponding to the corroded condition shifts markedly upward. The computed EAL reveals that, while a pristine pier would require an annual reserve of about 0.04% of the reconstruction cost, the corroded condition demands nearly 0.09% - indicating that the asset manager must allocate twice the annual reserve required under pristine conditions. This finding underscores the critical importance of incorporating actual structural degradation into risk assessments of bridge portfolios, as such defects can substantially alter expected losses.

Fig. 4. Loss Curves for Pristine and Corroded conditions.

5. Conclusions This study presents an integrated framework that combines AI-based computer vision (CV) with probabilistic seismic assessment for RC bridge piers affected by corrosion. By automating the classification of corrosion severity from images and linking each class to corresponding mass loss range, the proposed framework enables the incorporation of observable damage into streamlined structural analyses. The method quantifies the impact of localized corrosion on fragility and expected losses, with case study results showing that corrosion significantly increases seismic fragility — by up to 70% for severe damage states — and doubles the expected annual losses (EAL) compared to the pristine condition. These findings highlight the need to account for in-situ deterioration in bridge maintenance planning. The use of CV enables fast quantification of corrosion severity, enhancing rapid seismic risk assessment of bridge portfolios to support maintenance prioritization and optimize resource allocation within infrastructure management systems. Acknowledgements This study was supported by FABRE – “Research consortium for the evaluation and monitoring of bridges, viaducts and other structures”. Any opinion expressed in the paper does not necessarily reflect the view of the funder. Sergio Ruggieri thanks funding by FABRE, within the research grant ”Sviluppo e implementazione di strategie basate sulla Intelligenza Artificiale per l’analisi e il monitoraggio del rischio strutturale di ponti e viadotti esistenti”. Andrea Nettis and Giuseppina Uva thank MOST – Sustainable Mobility National Research Center and received funding from the European Union Next-GenerationEU (PIANO NAZIONALE DI RIPRESA E RESILIENZA (PNRR) –