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

526

7

400

200

200

300

150

150

200

100

100

100

50

50

0

0

0

Test Modal 3D (OOP) Uniform 3D (OOP) Modal 3D Uniform 3D Modal 2D Uniform 2D

Test Modal 3D (OOP) Uniform 3D (OOP) Modal 3D Uniform 3D Modal 2D Uniform 2D

-100

-50

-50

Test Modal 3D (OOP) Uniform 3D (OOP) Modal 3D Uniform 3D

Base Shear [kN]

Base Shear [kN]

Base Shear [kN]

-200

-100

-100

-300

-150

-150

-400

-200

-200

-60 -40 -20

0

20

40

60

-60 -40 -20

0

20

40

60

-60 -40 -20

0

20

40

60

Third Floor Displacement [mm]

Third Floor Displacement [mm]

Third Floor Displacement [mm]

(a)

(b)

(c)

Fig. 4 - Comparison between numerical and experimental results: (a) entire building; (b) East façade; (c) West façade.

Numerical and experimental results were compared in terms of capacity and backbone curves, respectively. The latter were derived as the envelope of the hysteretic responses of each dynamic test, taking the point of maximum base shear with the associated average third-floor displacement. The overall experimental base shear was determined by summing the inertia contributions at each accelerometer location, as the product between the recorded acceleration and its tributary mass. For 2D analysis discussion the average displacement was calculated for each individual façade; the base shear was evaluated from the masses of that façade, plus half the out-of-plane contributions of the transverse walls. Fig. 4a compares the global response from both 3D models to the experimental one. For this building, the modal force distribution underestimated the initial stiffness and, in the negative verse, also the lateral strength. Instead, the uniform one provided better results and successfully caught the degrading behavior due to the failure mechanisms. With this force distribution, explicitly modeling the out-of-plane response led to a particularly good prediction in the positive direction, with a little stiffness underestimation in the negative one. On the other hand, the conventional 3D model resulted in a minor strength overestimation in the positive verse, while yielding more accurate results in the negative one. Fig. 4b and c show the comparison between numerical and experimental responses for the individual East and West walls. For each façade, the total base shear of its piers was plotted against its third-floor mean displacement; in the unconventional 3D model, the out-of-plane force contributions were distributed evenly among the two longitudinal walls. Again, the uniform force distribution resulted in better accuracy than the modal one. For both walls, the capacity curves from 2D models were very close to the corresponding ones extracted from the conventional 3D model. Focusing on the uniform distribution results, both 3D models provided good agreement with the experimental backbone curve for the East wall. Moreover, despite the two large openings in the first story introduced a source of irregularity, the numerical models proved reasonably accurate also in predicting the West façade response; in this case, the 3D model with out-of-plane features underestimated the stiffness slightly more than the conventional 3D or single-wall 2D models. 5. Conclusions This paper discussed the numerical simulation of the experimental response of a masonry building aggregate with timber floors and roofs subjected to a unidirectional incremental dynamic shake-table test. Two structural units constituted the specimen, which incorporated structural and architectural features of the historical center of Basel, CH. The test was simulated by nonlinear static (pushover) analyses, adopting the equivalent-frame approach implemented in the software TREMURI, with nonlinear macroelements for masonry members and linear orthotropic membranes or truss elements for floor or roof diaphragms, respectively.