A new open-access article published by our research team in the journal "AI in Civil Engineering"
10/11/2026
As part of STRENGTH project, a new open-access article titled "Machine learning surrogates for predicting and interpreting flood fragility parameters of masonry walls" has been published in AI in Civil Engineering by Stefanos Georgiadis, Marco Nale, Marco Faggioli, Željana Nikolić, and Elena Benvenuti.
Key highlights from the publication include:
- The study introduces a mechanics-informed surrogate modelling framework to rapidly and accurately assess the vulnerability of unreinforced masonry walls to flood-induced out-of-plane collapse.
- Researchers trained machine-learning surrogates, such as kernel-based methods and ensemble models, on a synthetic dataset generated from mechanics-based limit-analysis simulations.
- The methodology integrates explainable artificial intelligence (XAI) to quantify the influence of geometry, openings, vertical connectivity, and boundary restraints on wall stability.
- The research bridges the gap between computationally expensive mechanics-based simulations and opaque empirical models, offering a scalable tool for regional flood-risk assessments.
By bridging the gap between structural physics and artificial intelligence, this breakthrough from the STRENGTH project unlocks scalable, lightning-fast, and highly transparent flood risk assessments to better protect our built environment.
The full article is available at the link provided in the right column.
