Equivalent frame modelling of an unreinforced masonry building in finite element environment

Equivalent frame method (EFM) is a viable modelling option for global seismic analysis of masonry buildings in comparison to more refined techniques, such as finite elements (FE), especially in professional practice [1–3]. EFM takes advantage of a building’s geometric regularity, both in plan and elevation, as well as of the good quality of masonry and floors stiffness, as required for the activation of box-like behaviour under seismic loads. However, typical vulnerabilities in existing unreinforced masonry (URM) buildings, e.g. highly flexible floors, openings too close one another, poor quality masonry, isolated pillars or non-vertically aligned walls, limit the effectiveness of EFM application. Recently, many studies have been devoted to expanding the possibilities of applying EFM to buildings which do not meet box-behaviour hypotheses [4–6]. The paper describes the procedure for implementing an EF model of an existing URM building in Midas GEN, a FE software commonly used for design of steel and r.c. structures. The equivalent frame (piers and spandrels) consists of a system of mono-dimensional, lumped plasticity beam elements [7]. In MIDAS, the frame is defined by the user, who also have to control the modelling process, by using the theoretical criteria available [8–10] and adapting their results to a building’s characteristics. Therefore, some peculiar vulnerabilities of the original building may be specifically implemented, thus obtaining a more refined model. The case study is Palazzo Carraro, a cultural heritage building, located close to the old town of Noale (Venezia). The palace complies with the main requirements of EFM except for floors stiffness, although horizontal connections may be considered sufficient at this level of analysis. Pro and cons of the specific procedure are here discussed, also referring to other state-of-the-art techniques, such as continuum models [11] implemented through the DIANA FEA code. Finally, the work explores the response of the EF model to different modelling choices, but also its reliability in overall analyses.