Combining operational modal analysis and genetic algorithms to understand the actual structural behavior of historical constructions

Evolution of technologies both in the fields of in situ investigations and Finite Element (FE) modeling strongly enhanced the possibility to understand the structural dynamic behavior of masonry historical constructions, allowing a periodic or continuous analysis of their response to environmental, anthropic, and exceptional actions and the formulation of accurate hypotheses about their future behavior, which is fundamental for Cultural Heritage preservation. In order to improve the assessment of the health status of historical buildings, inverse methods re-sorting to dynamic identification techniques are often used to provide experimentally verified data for the accurate calibration of FE models representative of the investigated structures. In this paper, vibration-based identification methods are coupled with an automatic FE model updating procedure to study the dynamic behavior of the Civic Tower of Ostra, Italy, and obtain baseline information for future comparative analyses. The experimental data obtained from two different campaigns of ambient vibration tests are used to update the mechanical characteristics of the detailed FE model of the tower. The updating process, unsolvable via common calibration procedures, is automatically managed through a powerful bio-inspired tool, i.e. the Genetic Algorithm (GA), allowing to closely reproduce the actual behavior of the tower.