GNGTS 2018 - 37° Convegno Nazionale

398 GNGTS 2018 S essione 2.2 Aviable alternative is to use synthetic accelerograms generated from a simulation of the source rupture and wave propagation. In this work, a direct link between hazard and response-history analysis is established. Synthetic seismograms are used to define the hazard as described by the Neo Deterministic Seismic HazardAssessment (NDSHA) (Panza et. al. , 2001, 2012; Fasan et al. , 2016) and, as a logical consequence, to perform NLTHA on a selected building. A comparison of the results of NLTHAs obtained with natural and synthetic records confirms that physics-based simulations are a valuable tool in structural analysis. Moreover, the NDSHAmethod is applied to the site of Norcia and predicted spectral acceleration are compared with the recorded one during the event of the 30 th of October 2016. Using NLTHAs, structural demands predicted using the real records and the synthetic ones used in the NDSHA are compared, showing that simulated accelerograms can be used to predict real non-linear demands of future earthquakes. Method. Using the procedure proposed by Fasan et al. (2016), the so-called Maximum Credible Seismic Input (MCSI) has been defined at bedrock (MCSI BD ) for the site of Trieste (Fig. 1a). MCSI is a multi-scenario assessment and its computation briefly consists of (Fasan, 2017): identifying all the (known) sources that may affect the site of interest; assigning each source the maximum credible magnitude; performing physics-based computations considering variability by simulating different directivity, rupture velocity, distribution of slip on the fault plane and soil layers; identifying, at each structural vibrational period, the most hazardous source in terms of median spectral acceleration and developing the statistics of all simulated response spectra for the corresponding source (see Fig. 1a). Therefore, the controlling source (and the simulated accelerograms) can be different at each vibrational period, and a selection of accelerograms conducted by imposing the spectrum compatibility within a range of periods would involve different scenarios and sources. For each vibrational period, MCSI is set equal to the 95th percentile of the spectral accelerations and represents a sort of Uniform Hazard Spectrum (UHS). To each point of the MCSI “cloud” shown in Fig. 1a corresponds one accelerogram, therefore it is natural to use these accelerograms if it is needed to perform NLTHAs. Using the MCSI (95 th percentile) at bedrock (MCSI BD ) acceleration response spectrum, the 4-storeys Steel Moment Resisting Frame (S-MRF) shown in Fig. 1b has been designed according to EC8 (CEN, 2004). Only the 2D MRF along the x direction is analysed using time-history analysis. Interior columns have HE300B cross-section whereas that of the external ones is HE280B. The floor beams are IPE300, on the upper floor an IPE270 cross section is used instead. The length of the spans is of 6 m. The ground storey height is 4 m and 3.5 m in the others. This planar four storey steel MRF has a first vibrational period of 1.5 s with 85% of mass participation. Non-linear dynamic analyses are performed using the software ADAPTIC (Izzuddin, 1991) adopting a non-linear fibres model for the cross-sections and including large displacements effects. The steel material is class S235 as per EC8 and is modelled as bilinear with kinematic hardening. A direct-integration numerical analysis with the Newmark-β method is used to resolve the equation of motion adopting a Rayleigh proportional damping matrix. The constants α and β necessary to define the damping matrix are chosen to have a critical damping ratio of 1% at target periods of two times the first translational vibrational periods (3 s) and the fourth translation periods (0.17 s). This choice avoids a possible overdamping of short periods due to the matrix damping definition. Five spectrum compatible sets of 11 recorded accelerograms are selected from the ESM database according to the following criteria dependent on the selected site: a magnitude range from 6 to 7; an epicentral distance range from 10 km to 30 km; EC8 soil class A or B; a period range for compatibility from 2 times the fundamental vibrational period T 1 to 0.2T 1 ; a maximum deviation of spectral accelerations from the target spectrum ranging from 90% to 130% of the target value. No scaling is applied. The same criteria are adopted to select 5 sets of synthetic accelerograms from a database created for this purpose. As an example, a set of natural (NAT) and a set of synthetic (SIM) accelerograms are shown in Fig. 1d and Fig. 1e respectively; the other sets can be found in Fasan (2017).