Using Greece’s Polyfytos Bridge as a case study, the research demonstrates how iterative ensemble learning can rapidly identify hidden prestress losses in concrete bridges, achieving high predictive accuracy across three bridge zones.
The findings highlight the potential of artificial intelligence to support infrastructure assessment while reaffirming the essential role of engineering judgement in safety-critical decisions.
The research was developed with Prof Stergios-Aristoteles Mitulis, Prof Ivan Izonin, Prof Athanasia Kazantzi, Francesco Pentassuglia, Illia Nesterenko and Roman Tkachenko.
