Evaluation of the Effects of Spatial Variability of Shear Wave Velocity on Seismic Ground Response

Abstract

This thesis describes a Monte Carlo simulation of one-dimensional nonlinear site response analysis to evaluate the effect of spatial variability of site velocity profiles on propagated ground motion. The spatially variable site velocity profiles are generated using two randomizing models and new developed model. One model assumes that the shear wave velocities are log-normally distributed and generates randomized profiles based on the probabilistic model proposed by Toro, 1995, which uses the concept of correlation coefficient of interlayer. Second method generates randomized profiles using a correlation structure with variance reduction effect. The other method is newly developed by combining these two models into a model. The spatial variability of natural soils is represented by the three model and the effects of depth to bedrock and distribution of shear wave velocity are investigated. The results are presented by comparing probabilistic analyses and deterministic analyses using baseline shear wave velocity profiles. The results show that the average of peak ground acceleration is more decreased as depth to bedrock increases. This is because at layer boundaries for which the impedance ratio is smaller than unity, pronounced shear strain induces low shear stiffness and high damping, thus lowering the peak ground acceleration. The baseline shear wave velocity profiles with Vs inversion causes more variability in surface responses due to higher contrast in shear wave velocities.