Finite element model for the probabilistic seismic response of heterogeneous soil profile

Site- and Motion-Dependent Parametric Uncertainty of Site-Response Analyses in Earthquake Simulations

“…Additional work includes the published results by Rahman and Yeh (1999) for one base profile with ground motion being simulated as a stationary random process with fixed frequency content; Wang and Hao (2002), who included the effects of groundwater level on the ground surface response variability; Nour et al (2003), who investigated the effects of correlation distance of the soil V S , ξ, and Poisson ratio (V) for a 2D configuration; Assimaki et al (2003) and Bazzurro and Cornell (2004), who investigated the effective stress transient nonlinear response of cohesive and cohesionless sites subjected to multiple recorded ground motions; Andrade and Borja (2006), who compared the ground response variability due to soil parameter uncertainty predicted at two sites by means of the equivalent linear (Idriss and Sun, 1992) and time-domain nonlinear (Borja et al, 2000) models; and Stewart and and Kwok et al (2008), who evaluated the effects of soil parameter uncertainty on the response of La Cienega and Turkey Flat vertical array profiles to strong ground motion events using the nonlinear site-response computer code DEEPSOIL as part of a study that looked at several site-response codes and their prediction variability.…”

“…More specifically, Hwang and Lee (1991), Tian and Jie (1992), Assimaki et al (2003), Andrade and Borja (2006), and Stewart and used a very limited number of ground motions; Wu and Han (1992), Rahman and Yeh (1999), and Nour et al (2003) studied simplified pulses instead of true seismic excitations; Suzuki and Asano (1992) and Field and Jacob (1993) limited their study to weak ground-motion recordings. In the majority of these studies, results illustrated the effects of uncertainty in the low-strain (visco-elastic) soil properties (Suzuki and Asano, 1992;Tian and Jie, 1992;Wu and Han, 1992;Filed and Jacob, 1993;Rahman and Yeh, 1999;Assimaki, 2003;Nour et al, 2003). Also, the variability statistics of visco-elastic and nonlinear soil parameters in these studies are by and large described by simplified probability distribution functions and correlation structures, while typical near-surface geologic formations tend to exert more complex spatial variability characteristics.…”

Attenuation and Velocity Structure for Site Response Analyses via Downhole Seismogram Inversion

“…Since scattering attenuation is the mathematical expression of energy redistribution, geoprocesses shown to be significantly affected by micro-scale heterogeneities, cannot be accurately simulated based on the assumption of scattering energy loss. Examples include the soil response during liquefaction (POPESCU and PREVOST, 1996;POPESCU et al, 1997;KOKUSHO, 1999); the spatial variability of surface ground motion (ASSIMAKI et al, 2003;NOUR et al, 2003); slope instability (YONG et al, 1977;TONON et al, 2000), settlement (PAICE et al, 1996) and seepage (GRIFFITHS and FENTON, 1993;FENTON and GRIFFITHS, 1996) in porous media; and soilstructure interaction (ZERVA, 1991). In particular for high-frequency components of seismic waves propagating through the near-surficial, highly heterogeneous formations, deviation of our predictions from the physical process might be quite Figure 21 Comparison of: (a) the optimum synthetic and observation ground surface motion for event 1 at station iwth04, and (b) the empirical transfer functions obtained from the recorded and synthetic surface and borehole seismograms for the optimization time window, and (c) the theoretical transfer function obtained using the average site response and travel time, and the one obtained using 2 s window wavelet-domain inversion, to the frequency response computed using 10 s-windows of 28 events recorded at the site.…”

Estimation of stochastic nonlinear dynamic response of rock-fill dams with uncertain material parameters for non-stationary random seismic excitation

“…In the current research, model for the shear modulus consists of spatially random fields and uses a lognormal distribution to represent the heterogeneous characteristics of dam soil. The motivation for this choice is the fact that these parameters are positive, and lognormal distribution enables analysis of its wide range of variability [12]. The average shear modulus at low strain only correlates in the horizontal direction for each layer.…”

“…In this study, the assumption is that the low-strain shear modulus is a random variable within a specified mean and standard deviation. Significant prior studies [4][5][6][7][8][9][10][11][12][13], related to stochastic finite-element analysis of the geotechnical area, included parametric uncertainties as well as random loads.…”