PSI - Issue 44

Marco Gaetani d’Aragona et al. / Procedia Structural Integrity 44 (2023) 1052–1059 M. Gaetani d’Aragona, M. Polese, A. Prota / Structural Integrity Procedia 00 (2022) 000–000

1053

2

instruments to predict the potential impact of earthquakes, and to plan effective mitigation strategies at the large scale need to be provided. While at the building level, the seismic performances of buildings can be reliably predicted by adopting Nonlinear Response History Analyses (NRHAs) performed on refined finite element models (FEMs), the expertise for structural modeling and the computational effort required makes these strategies not suitable to be extended for large scale assessment purposes. To reduce the computational effort, different strategies were proposed either consisting of the adoption of less refined approaches to estimate the seismic response (MAEviz, HAZUS-MH) or the adoption of simplified models of buildings (Gaetani d’Aragona et al., 2018a, 2019a; Polese et al., 2018; Borzi et al., 2008). Recent studies proposed the adoption of condensed Multi-Degree-Of-Freedom (MDOF) fish-bone (Luco et al., 2003; Khaloo et al., 2013; Soleimani et al., 2019; Jamsek and Dolsek, 2020) or stick models (Lu et al., 2014, 2018; Xiong et al., 2016, Gaetani d’Aragona et al. 2020, 2021a, 2022a, b) . The latter consists of a system of masses lumped at the story level connected in series by nonlinear shear springs simulating the interstory nonlinear shear-interstorey drift response for each main building direction, which accuracy in prediction depends on the proper calibration of the nonlinear behavior of interstory springs (Lu et al., 2014; Gaetani d’Aragona et al. 2019b). Gaetani d’Aragona et al. (2020) proposed a calibrated Stick model to represent the behavior of RC Infilled frames Typologies (Stick-IT). In such paper, simplified formulations allow determining model parameters based on few geometrical and structural data that can be easily retrieved for large scale studies. The reliability of the calibrated model in predicting damage and repair costs was validated both at the building level Gaetani d’Aragona et al. (2021a) and at the large-scale Gaetani d’Aragona et al. (2022a) using real damage and repair cost data collected in the aftermath of the 2009 L’Aquila earthquake. However, the Stick-IT model (Gaetani d’Aragona et al., 2020) was specifically calibrated to represent the behavior of gravity load designed buildings, while other frequent buildings typologies were neglected. To account for a wider range of masonry-infilled RC building typologies, Gaetani d’Aragona (2022b) extended the procedure also accounting for the presence of possible retrofit structural intervention designed to increase the building structural safety also reducing future damage and losses Gaetani d’Aragona (2017a, 2018b). To reduce the computational effort required to characterize the proposed curves, an alternative to the more complex procedure based on the cyclic pushover analysis of refined FEMs was also proposed in Gaetani d’Aragona (2022b). The procedure relies on the basic hypothesis of shear-type behavior to generate Stick interstory backbones curves for buildings designed and constructed in different periods, regions and according to different design requirements (gravity load only, design according to obsolete seismic code provisions) and accounting for possible retrofit interventions both at the global (RC jacketing) and the local level (FRP wrapping). The procedure allows to explicitly simulate typical failure modes of non-conforming RC elements, brittle failures induced by local interactions, and typical existing RC infilled building global collapse modes. This paper synthetizes the main results obtained in Gaetani d’Aragona et al. (2022b) and highlights the usefulness of the model to evaluate potential beneficial effects of retrofit strategies for building typologies. 2. Generation of interstory backbone curves Fig. 1 illustrates the flowchart of a simplified procedure to generate interstory backbone curves for Stick models implemented in a built-in Matlab ® (2021) routine. At the large scale, no information regarding geometric/structural parameters are available at the building level. To obtain these information, a simulated design routine is adopted to characterize the geometry and structural details of RC members (Verderame et al., 2010, Gaetani d’Aragona et al., 2019a, Polese et al., 2018). Based on building information data such as the number of stories (n s ) and the building surface area (SA), a number of compatible geometrical configurations are obtained (1.1). Then, based on the simulated building geometry, the structural configuration is obtained (1.2) and the structural members of the frames are dimensioned and designed (1.3) complying with the building codes and the design rules in force at the time considered for the construction. If seismic design is considered, the building location also needs to be included as a building information data. Depending on the age of construction and the design code, the framing system can consist of planar moment-resisting frames (connected by flat transverse beams) or spatial frames. During the simulated design process, forces acting in columns and beams are calculated according to approximate design schemes common at the time of the construction. For instance, beam