GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 2.2 385 available for consultation on www.webms.it and created by CNR IGAG for the Italian National Civil Protection Department (DPC 2018), has been conducted in order to identify, for both stiff and soft layer: 1) the variation, in terms of minimum and maximum values, of the shear wave velocity, Vs, and layer thickness, H stiff and H soft ; 2) the soil unit weight (g soil ). In particular, SMs of the municipalities of the centre of Italy, SMs of Sicilia and Calabria regions were used for the case of fluvial and marine terraces; Orvieto and Catania cliffs were considered for the case of cliffs and lava rock layer; finally, the case of Rieti plain was considered studying the case of travertine plateau in alluvional plains. Tab. 1 synthesizes the identified ranges of values for the geometrical and geotechnical properties. Table 1 - Geometrical and mechanical characterization of the stiff and soft layers, in the case of Vs inversion, for 1) fluvial and marine terraces, 2) cliffs and lava rocks, 3) travertine plateau on alluvional plains in Italian territory. GEOLOGICAL SETTINGS FLUVIAL AND MARINE TERRACES CLIFFS AND LAVA ROCKS TRAVERTINE PLATEAU ON ALLUVIAL PLAINS LAYER PROPERTIES H min -H max (m) V s,min -V s,max (m/s) γ soil (kN/m 3 ) G/G 0- γ and D - γ curves H min -H max (m) V s,min -V s,max (m/s) γ soil (kN/m 3 ) G/G 0- γ and D - γ curves H min -H max (m) V s,min -V s,max (m/s) γ soil (kN/m 3 ) G/G 0- γ and D - γ curves STIFF LAYER 5-20 320-530 21 Rollins et al. , 1998 upper-medium-lower bound 15-50 600-1000 15 Pagliaroli et al. , 2014 tuff curves 20-50 550-1000 20 Curves from Rieti SM study SOFT LAYER 10-30 200-460 19 Vucetic and Dobry, 1991 PI 0-15-30 10-30 200-460 19 Vucetic and Dobry, 1991 PI 0-15-30 50-400 200-350 18 Vucetic and Dobry, 1991 PI 50-100-200 One dimensional site response analysis: Monte Carlo simulations. The characterization proposed in Tab. 1 has been adopted to carry out a set of stochastic equivalent-linear site response analyses, performed with STRATA code (Kottke et al. , 2013), for one-dimensional columns set up by combining the minimum and the maximum value of H stiff and H soft for a total of four different geometrical configurations. To take into account the natural variability of Vs profile in the range identified in Tab. 1 and the uncertainties related to the nonlinearity of the stiff and soft soil layers, Monte Carlo simulations were performed. V s profiles were generated using Toro (1995) random field model described by a log-normal distribution, while the nonlinarity uncertainties were considered adopting Darendeli (2001) model, where the variability around the mean value is assumed to be normally distributed through the mean ( m ) and standard deviation ( σ ) values. For the scope, literature G/G 0 -g and D-g curves, listed in Tab. 1, have been selected. For the soft layer in particular, Vucetic and Dobry (1991) curves were used for a variation of the plasticity index, PI, ranging between 0 and 30 (PI=15 represent the target curve) in the case of fluvial and marine terraces and the cliffs and lava rocks domains, while the curves for PI ranging between 50 and 200 (PI=100 represent the target curve) were used in the case of travertine plateau on alluvional plains. In this case, the use of very high PI values allows to reproduce a pronounced linear and low dissipative behaviuour that is typical of very deep deposits and organic clays frequently encountered in alluvial plains (Pagliaroli et al. , 2014). For the stiff layer, Rollins et al. (1998) upper-medium-lower curves were adopted for fluvial and marine terraces (the medium curve is the target curve); Pagliaroli et al. (2014)