A Simplified Soil Model for Seismic Site Response Analysis of Liquefaction

Session: Near-Surface Effects: Advances in Site Response Estimation and Its Applications [Poster]
Type: Poster
Date: 4/28/2020
Time: 08:00 AM
Room: Ballroom
Description: 

Numerical modeling of soil liquefaction remains quite challenging. Several constitutive models were developed and implemented in numerical analysis platforms to evaluate liquefaction triggering and its effects. However, available programs have several limitations including: (a) determination and calibration of constitutive model parameters are usually complex and expensive, (b) computational efficiencies of programs with advanced soil models are insufficient to perform large scale analysis and (c) many programs are limited to one-dimensional (1-D) or two-dimensional (2-D) analysis. This study introduces a unified, simple and fully-coupled analysis platform to perform seismic site response analysis with high pore-water pressure generation based on a nonlinear three-dimensional (3-D) soil constitutive model, I-soil, implemented in the general-purpose analysis software LS-DYNA. The I-soil model was initially developed to evaluate the volumetric strain of dense to very dense sands under unidirectional and multi-directional seismic loadings. This model was shown to provide good estimations for seismic settlements and excess pore-water pressure generation of saturated dense sand under free-field and soil-structure interaction conditions. This study will extend and validate the capabilities of I-soil model for soil liquefaction analysis. The calibration of the I-soil model is relatively straightforward whereby only two parameters are needed to describe the volumetric response of the soil. In this study, the performance of the soil constitutive model with excess pore-water pressure generation is assessed with single element tests, downhole array recordings from Port Island and selected Liquefaction Experiments and Analysis Projects (LEAP) experiments. The performance of the model relative to measured behavior will be presented. Computational considerations and efficiency of the numerical analyses will also be described.

Presenting Author: Guangchao Xing

Authors