PSI - Issue 44

Christian Salvatori et al. / Procedia Structural Integrity 44 (2023) 520–527 Christian Salvatori et al./ Structural Integrity Procedia 00 (2022) 000–000

522

3

Construction details of the historical center of Basel were reproduced. Walls consisted of double-leaf undressed stone masonry and river pebbles, with a thickness starting from 35 cm at the first story and decreasing to 25 cm at the third one. Through stones provided a better connection between the masonry leaves only at opening edges and corners. Flexible timber diaphragms were adopted for floors and roofs. Three-dimensional timber frames, consisting of trusses and diagonal braces, constituted the roof structure, whereas 10x16 cm timber joists with a single 2-cm-thick layer of planks formed the flooring systems. The joists of the first two floors were arranged in the shaking direction (North-South), while the roof trusses were oriented orthogonally (East-West). The truss tie beams acted as joists for the third floor, spanning perpendicular to the shaking direction. Most building aggregates originated from a linear growth along the street. Consequently, the connection between adjacent units might be weak and vulnerable to local damage due to their mutual interaction. To reproduce a similar situation, the North (shortest) unit of the specimen was built first, whereas the South (tallest) unit was realized with some delay, providing a through-stone every third masonry course. 2.2. Material properties and masses Complementary characterization tests on materials and components constituting the prototype were conducted at the University of Pavia (Guerrini et al., 2017). Vertical and diagonal compression tests allowed evaluating the mean Young’s modulus (3460 MPa), shear modulus (1520 MPa from vertical and 1900 MPa from diagonal compression), Poisson’s ratio (0.14), compressive strength (1.30 MPa) and tensile strength (0.17 MPa). Moreover, cyclic shear compression tests were performed on masonry piers (Senaldi et al, 2018). The double-leaf stone masonry had an average density of 1980 kg/m 3 . In addition to the self-weight of walls and diaphragms, mortar bags were evenly distributed over each floor level to emulate superimposed dead and live loads without influencing their stiffness. 2.3. Testing protocol The shake-table input consisted of a series of ground motion records to simulate increasing levels of intensity, up to ultimate conditions of the specimen. Three different natural accelerograms were chosen. The first two were low intensity records from recent seismic events in Basel and Linthal, CH, respectively, and were applied with their actual acceleration amplitude. The third signal came from the Ulcinj station during the 1979, M w 6.9 Montenegro earthquake, with PGA of 0.224 g, and was selected due to its spectral displacement compatibility with the 475-years-return-period design spectrum for Basel; it was applied scaling its acceleration amplitude from 25% to 175% in increments of 25%. The time step of all records was compressed by a factor λ 1/2 = 0.707 to satisfy the chosen similitude relationship. 3. Numerical simulations The response of the specimen was simulated through nonlinear static (pushover) analyses with the equivalent frame modeling approach implemented in the software TREMURI (Lagomarsino et al., 2013). Advanced and conventional 3D models, as well as single-wall 2D models, were analyzed to assess the influence of explicitly modeling the walls excited out-of-plane and the diaphragm stiffness. 3.1. In-plane masonry macroelements The overall behavior of a building can be obtained by assembling vertical walls and horizontal diaphragms, considering only their in-plane strength and stiffness contributions (Fig. 2a). Within the equivalent frame modeling framework, each wall is discretized into macroelements capable of simulating the response of piers and spandrels, and rigid nodes, which define portions of masonry less sensitive to deformations and damage (Fig. 2b and c). In this work, the two-dimensional macroelement proposed by Penna et al. (2014) and the improved version described by Bracchi et al. (2021a,b) were adopted to model spandrel and pier elements, respectively. In fact, the capability of these macroelements to capture the main failure mechanisms of a masonry panel, together with their analytically integrated formulation, make them particularly suitable for performing static and dynamic analyses.