PSI - Issue 44

Gabriele Guerrini et al. / Procedia Structural Integrity 44 (2023) 1877–1884 Gabriele Guerrini et al. / Structural Integrity Procedia 00 (2022) 000 – 000

1883

7

wall. Floor and roof diaphragms were modelled through 4-node orthotropic membrane elements characterized by a linear-elastic behavior. The Rayleigh damping formulation was implemented in the software, and a selected damping ratio of 5% was assigned to the first and last natural frequencies. The input ground-motion signal consisted of a single component accelerogram labelled EQ-NPR, with a peak ground acceleration of 0.31 g, applied in the building longitudinal direction. The acceleration amplitude of the original signal was progressively scaled from 20% to 266%, accordingly to the experimental test. Additional considerations were also necessary to properly simulate the experimental response of the retrofitted building. With the increasing of the external excitation, the prototype exhibited such a significant lateral stiffness degradation that could not be captured by TREMURI. Accordingly, a new numerical model with the same geometry but with a reduced masonry shear modulus G red = 0.2 G m was generated to simulate the last test runs (where G m = 0.35 E m ). The coefficient of 0.2 was found as the minimum that allowed numerical convergence. Fig 5a and 5b show the overlap of numerical and experimental specimen hysteretic response, in terms of base shear coefficient ( BSC ) and global drift ratio ( θ ), corresponding to half of the maximum scaling applied to the specimen (i.e., 133%) and at high level of damage, with the input motion scaled at 200%. The BSC was obtained normalizing the base shear by the weight of the contributing mass (35.4 t), while the global drift ratio was defined as the ratio of the average second-floor displacement to the corresponding floor height (5.40 m). It can be noted that the response of the building with a limited accumulation of damage was well simulated by the software, while in an advanced state of damage the lateral stiffness could not be fully captured. The comparison of backbone curves reported in Fig 5c shows that the evolution of global dynamic properties throughout the experimental tests was well matched by the numerical model, except for the last testing run (i.e., 266%) where the actual masonry stiffness degradation could not be simulated. This implied an underestimation of the specimen ultimate global drift ratios. 6. Conclusions The experimental responses of a retrofitted masonry pier quasi-statically tested in plane and a retrofitted full-scale building prototype incrementally tested on the shake-table were numerically simulated. The timber system installed on the first specimen consisted of timber frames mechanically connected to the masonry and to floor and foundation systems, on which OSB are nailed. The flexural resistance of masonry elements is improved through the mechanical anchoring of vertical timber posts to the floors through tie-down anchors, while the shear resistance was improved by the nailed OSB layer. The application of the retrofit system to the entire building was designed to also improve the overall building connections. The numerical models were generated using the software TREMURI, which relies on macroelements based on the equivalent frame method. A modelling approach to explicitly model the timber-based retrofit system is proposed. Specifically, the masonry elements were modelled through discretized macroelements, while flexural and shear contributions offered by the timber retrofit system were modelled using equivalent non-linear beam elements. The simulation of component and full-scale building tests resulted in a good match between experimental and numerical results. Besides the good simulation of displacement and strength capacities, the numerical models captured the evolution of the dynamic properties of the specimens up to a significant level of damage. The presented study discussed the numerical simulation of the contribution of a newly proposed retrofit system and proposed a numerical strategy based on the simulation of a quasi-static and an incremental dynamic test on retrofitted specimens. The proposed procedure appears to be promising to assist in the assessment of the effect of the intervention at large scale. Different geometrical and material properties of masonry and timber elements could be also investigated. In any case, for a complete assessment on the effectiveness of the retrofit system also further studies on the out-of plane behavior should be carried out, from the experimental and numerical points of view. Acknowledgements The work presented in this paper is part of the EUCENTRE project “Study of the vulnerability of m asonry buildings in Groningen” within the research framework on hazard and risk of induced seismicity in the Groningen province, sponsored by the Nederlandse Aardolie Maatschappij BV (NAM). The modelling activity was funded by DPC ReLUIS Work Package 5 (2022 – 2024 ) “Interventi di rapida esecuzione a basso impatto ed integrati”. The authors