Fragility curves of masonry buildings in aggregate accounting for local mechanisms and site effects

Abstract Masonry buildings in aggregate are largely diffused within the existing Italian and European building stock, especially in historical centres. Their seismic evaluation represents a difficult and open task that has not been exhaustively investigated so far. The study proposes a procedure aimed at evaluating the potential impact of the combination of local mechanisms and site-amplification in terms of fragility curves on an existing masonry aggregate. The latter consists of five adjacent structural units mutually interacting with each other during seismic sequences. The aggregate is inspired by the unreinforced masonry (URM) built heritage of the historic centre of Visso struck by the Central Italy 2016/2017 earthquakes. It was proved to be negatively affected especially by the site-amplification phenomena. The in-plane (IP) response of URM buildings was simulated through nonlinear dynamic analyses performed on a 3D equivalent frame model of the structure, whereas out-of-plane (OOP) mechanisms were analysed by adopting the rigid-block assumption but assuming as seismic input the floor accelerations derived from the post-processing of data derived from the global 3D model. Also an innovative procedure considering the pounding effect to the global response of the building was presented. Two soil conditions were assumed with (freefield) and without (bedrock case) site amplification. In the first case, a site response analysis based on a 1D soil model aimed to reproduce the actual soil profile under the examined aggregate has been carried out. The results showed that site effects strongly affected the seismic vulnerability of the aggregate, also possibly altering the combination between IP and OOP mechanisms with respect to what expected on basis of evidence from other historical centres. More specifically, for bedrock condition, the overall seismic response of the buildings in aggregate was strongly affected by such a combination, especially for medium-high damage levels where local mechanisms were prevailing with respect to the IP response. Conversely, for freefield condition, IP mainly governed the overall behaviour for all the damage levels, consistently with the field evidence. The result appears also affected by the architectural configuration of such buildings, characterized by quite large openings at ground floor.