A model-based damage detection approach using genetic algorithm optimization

Resumo

The identification of structural damages based on numerical model updating has been of increasing
interest in several areas of engineering, and basically consists on an optimization process relying on the
minimization of the residuals between the measured and estimated numerical responses. Most of the existing
studies using this methodology are applied to simple structures such as beams, frames and trusses, existing a scarce
of studies applied to large and complex structures. In this context, this article aims to evaluate the performance of
this methodology using genetic algorithms with application to a case study of a real railway bridge, to assess its
performance when applied to complex and large-scale structures. The results obtained proved the reliability of the
proposed methodology in detecting, locating and quantify multiple damages, with a considerable number of
updating parameters and scarce number of target responses.