Abstract

Modern code-oriented elastic floor response spectra formulations for RC framed structures do not take into account effects of non-negligible nonstructural components in terms of mass and stiffness, such as masonry infills (MIs). MIs nonlinear behaviour can be represented through the combination and mutual interaction between the in-plane (IP) and out-of-plane (OOP) responses. The present work is aimed at identifying the effect of IP and OOP nonlinear modelling assumptions on floor acceleration response spectra, consistently with the required seismic intensity level for simplified verification of life-threatening nonstructural elements. To this end, a spatial one-bay multi-storey shear-type model is considered as equivalent to infilled RC framed buildings with common double-leaf MIs. Additional variability of the following design parameters is considered: number of storeys (three, five and seven); behaviour factor (low, 1.5, medium, 3, and high, 4.5); OOP strength of MIs, with lower and upper bound values corresponding to oneand two-way arching mechanisms, respectively. A recently proposed computer code, that includes a five-element nonlinear infill macro-model comprising four diagonal OOP beams and one (horizontal) central IP truss, is considered for the numerical investigation. The proposed algorithm modifies stiffness and strength values of MIs in the OOP direction on the basis of simultaneous or prior IP damage and vice versa. Moreover, a lumped plasticity model describes the inelastic behaviour of RC frame members. Biaxial spectrum-compatible accelerograms are considered at life-safety limit state provided by the Italian seismic code. A simplified code-oriented formulation for the evaluation of floor response spectra of infilled RC framed structures is proposed. Nonstructural maximum acceleration is firstly evaluated by means of vertical and nonstructural amplification factors. Continuous wavelet transforms are used to calibrate parameters that define the resonance region width, accounting for moving resonance due to nonlinearity and higher modes effects. Parabolic and Gaussian curves are considered in order to reproduce preand post-resonance regions, respectively. Finally, a code-oriented proposal is compared to exact elastic and inelastic floor spectra of MIs evaluated over their common range of OOP vibration periods.