GNGTS 2019 - Atti del 38° Convegno Nazionale

386 GNGTS 2019 S essione 2.2 tuff curves have identified the variation zone of G/G 0 -γ and D- γ for cliffs and lava rocks; finally, for travertine plateau on alluvional plains, the decay curves from Rieti SM study have been adopted. As example, Fig. 1 shows one hundred randomized Vs inversion profile for each geometric configuration and the corresponding m and m ± σ nonlinear soil properties predicted for the fluvial and marine terraces. Once defined the set of 1D columns for each geological domain, three groups of 20 non-scaled real accelerograms, recorded at outcropping stiff soil were selected for three different PGA levels: Class 1 (PGA<0.1g), Class 2 (0.1g≤ PGA <0.2g), Class 3 (0.2g≤ PGA <0.4g), respectively. In this way, for each geological setting, 100 Vs profiles have been generated for 4 different geometrical configurations, for a total of 400 1D profiles → 400profiles x 60signals = 24000 analyses for each geological setting for a total of 72000 performed analyses. The results have been processed in terms of shear deformation profile, γ max -z, and respose spectra, T-Sa. Fig. 2 for instance shows, as example, g max -z profiles and T-Sa relation for Class 3 of PGA, associated to the fluvial and marine terraces, as the envelope of the minimum and the maximum reached value. It can be appreciated an attenuation effect of the seismic motion (Fig. 2b) at low periods and an amplification mode at higher periods, that tends to be more emphasized when H soft >H stiff and with the increasing of H soft . This trend highlights the influence of the stiffness contrast due to Vs profile inversion on the 1D seismic behavior of the column. Note that the stiff layer exerts a confining action on the soft layer that exhibits higher level of deformation (see Fig. 2a) with respect to the case of increasing value of V S with depth; the soft layer thus represents the controlling element of the 1D column in these complex geological conditions through the variation of V s,soft and H soft. Finally, a comparison between the above defined geological settings and the three different PGA classes is proposed in Fig. 3 in terms of spectral acceleration and spectral ratio (i.e. Sa surface /Sa input ). As expected, in the case of travertine plateau in alluvional plains Fig. 1 - (a) Randomized Vs profiles and (b) variability of G/G 0 -γ and D-γ curves for fluvial and marine terraced deposits.