PSI - Issue 44

Michele Angiolilli et al. / Procedia Structural Integrity 44 (2023) 2074–2081 M. Angiolilli et al./ Structural Integrity Procedia 00 (2022) 000 – 000

2076

3

analytical approaches, the aggregate-effect is often meant as the effect that boundary conditions, provided by adjacent SUs, may have on the seismic response of an individual building belonging to an aggregate. Despite the modeling of adjacent SUs is often wrongly neglected, most of the analytical study reported above concluded that aggregate-effect was beneficial for the individual units, although this outcome can be also associated to the type of the analysis. Indeed, in Angiolilli et al. (2021) it was shown that nonlinear static analyses (NSA) are not appropriate, as compared to the nonlinear dynamic analyses (NDA), to gather accurately the seismic response of mutual interacting SUs. On the other hand, from most of the empirical observations, one can conclude that isolated buildings are less vulnerable than aggregated buildings. This is because isolated buildings are usually characterized by a structural typology and an architectural configuration different from buildings belonging to aggregate. The challenges related with the aggregate-effect, as well as the combination of local and IP mechanisms, will be highlighted in this work. In the current literature, there are various valid computational models and methodologies for assessing URM structures, each with its own set of benefits and drawbacks in terms of representativeness, computer efficiency, and application (Roca et al. (2010), Asteris (2015)). Modelling historic URM structures, on the other hand, necessitates special considerations and a design philosophy distinct from that used for modern URM and other materials. The ability to appropriately represent the interaction between IP and OOP responses, as well as the influence of diaphragm failures on them, is a must. Although sophisticated detailed models capable of directly representing local mechanisms within the analysis of the global response can be used (e.g. Lourenço et al. (2011), Castellazzi et al. (2017), Malomo et al. (2020), Grant et al. (2021)), their high computational burden prevents their use for large numbers of nonlinear analyses on complete building models. Equivalent Frame (EF) models requires a limited number of degrees of freedom, with a reasonable reduction in computational effort, allowing the analysis of complex models of URM structures. In the proposed procedure, the IP response of the buildings is based on the use of EF models whereas the local mechanisms are evaluated in a separate way. The combination of responses can be then performed according to the procedure described in Angiolilli et al (2021) and Lagomarsino et al. (2022). Within this general context, this work presents the results obtained for two numerical models representative of URM aggregate belonging to small to large Italian historical centres. The models were developed in previous studies (Angiolilli et al. (2021), Brunelli et al. (2022)) and results are herein post-processed to highlight the effect mutual interaction between adjacent SUs during seismic events in terms of fragility curves, explicitly accounting also for local mechanisms. 2. Main outcomes obtained for the investigated case studies 2.1. Features and modelling criteria of the case studies The two studied case-studies consists of URM-SUs belonging to aggregate located in Catania and Visso (Italy). They are representative of very complex building conglomerates in which the interaction between SUs is fundamental to be evaluated to accurately estimate their seismic response. In particular, regarding the Catania’s aggregate, a corner structural unit of a L-shaped aggregate formed with two adjacent SUs was taken as reference building. It is assumed that the two SUs were built after the construction of the reference building, simply built against it without doubling the boundary walls. The 3-storey URM building is characterized by a thickness of the external lavic stone walls ranging from 0.55 m to 0.7 m, as well as an internal wall thickness ranging from 0.25 m to 0.4 m. The same geometrical features were assumed also for the two adjacent buildings. The case study was recently studied in Angiolilli et al. (2021) and Lagomarsino et al. (2022). Regarding the Visso’s aggregate, four of five adjacent SUs belonging to the so-called "row housing" (consisting of a series of buildings aggregated in lines) were studied. The geometric and structural details of the units were assumed on the basis of field survey and, in some cases, were also based on the characteristics of neighboring damaged buildings that showed almost clearly the type of the masonry, the diaphragm system, and the distribution of the internal space. In particular, the load-bearing walls were characterized by two-leaf stone masonry, with rough stones sizing about 90-120 mm in height and 360-400 mm in length. Floor diaphragms were assumed to be composed of concrete slab (not reinforced) system 150 mm tick. Only sporadic tie-rods were present, making some of structural units possibly susceptible to the activation of OOP mechanisms, especially at the upper building floor. The Visso ’s aggregate was struck by several mainshocks in 2016 causing widespread damage to the SUs (Brunelli et al. 2022). In