SEISMIC ASSESSMENT OF MASONRY BUILDINGS FROM SHAKING TABLE TESTS AND NONLINEAR DYNAMIC SIMULATIONS BY THE PROPER ORTHOGONAL DECOMPOSITION (POD)

The paper focuses on the interpretation and numerical simulation of the results of shaking table tests on two masonry buildings mock-ups, performed at Laboratório Nacional de Engenharia Civil (LNEC) in Lisbon, within the purposes of the SERIES project Transnational Access activities (Mendes et al. 2014, Lu et al. 2014). Shaking table tests have produced a huge amount of remarkable data that need to be properly interpreted, in order to get general and synthetic information on the seismic behaviour. To this aim, different techniques of data interpretation are applied, consisting in some traditional approaches (e.g. the frequencies analysis or the identification of an equivalent linear system) and in the innovative use of the Proper Orthogonal Decomposition (POD) technique. Together with the interpretation of experimental results, the paper also illustrates the nonlinear dynamic simulation of these two models by Tremuri program (Lagomarsino et al. 2013), which is based on the equivalent frame approach. The results achieved showed a good agreement between experimental and numerical results, highlighting that the tools adopted are quite effective in interpreting the nonlinear response of URM masonry structures. INTRODUCTION Current trends in the field of the performance-based seismic assessment of new and, in particular, existing buildings are oriented to the use of nonlinear analyses, both static, as proposed in international codes (ASCE SEI 41-06 2007, EN 1998-3 2005), and dynamic, by the execution of Incremental Dynamic Analyses – IDA (Vamvatsikos and Cornell 2002). In the case of masonry buildings, the use of nonlinear approaches is even more necessary due to their seismic behavior characterized by the occurrence of cracking even for low levels of seismic demand. It is evident that reliable numerical models are needed for performing such complex analyses: to this aim, the support of experimental results constitutes an essential tool to validate and corroborate their use. In particular, information on the dynamic response are valuable, which can be achieved by shaking table tests, but the difficulty in interpreting the results in an effective and synthetic way arises. 1 Assistant Professor, DICCA University of Genoa, Genoa (Italy), serena.cattari@unige.it 2 Graduate student, University of Genoa, Genoa (Italy), ambra.chioccariello@yahoo.it 3 Full Professor, Hasselt University (Belgium), herve.degee@uhasselt.be 4 Research engineer, Catholic University of Louvain, (Belgium), catherine.doneux@uclouvain.be 5 Full Professor, DICCA University of Genoa, Genoa (Italy), sergio.lagomarsino@unige.it 6 PhD Student, University of Liege (Belgium), cmordant@ulg.ac.be