A Critical State Two-Surface Plasticity Model for Simulation of Flow Liquefaction and Cyclic Mobility Open Access
Downloadable ContentDownload PDF
AbstractA Critical State Two-Surface Plasticity ModelforSimulation of Flow Liquefaction and Cyclic MobilityA major part of damage to buildings and port infrastructure during seismic events is due to soil-structure interaction during ground shaking. When saturated granular soils are subjected to severe shaking, due to potentially large increase in pore water pressure, the soil may lose its shear strength and behave like a fluid. This phenomenon is called liquefaction which may lead to relatively large displacements and/or tilting in the structures supported by the liquefied soil. Since granular materials such as sands, gravels and silts are commonly found in the site of many buildings, bridges, retaining structures, and port facilities, probability of liquefaction in these soils is an important design issue that needs to be carefully considered. The main objective of this research is to develop a validated methodology for simulation of the response of granular soils under cyclic and seismic loads. To this end, the evolving nature of soil fabric and its impact on volume change response of granular soils are investigated and a constitutive model is proposed based on modification of an existing constitutive model of granular soils that was introduced by Dafalias and Manzari in 2004. The original Dafalias-Manzari model has been modified to add new features such as hysteretic behavior in cyclic loading at small strain levels, the effects of volumetric plastic strain on fabric evolution of the soil, and the effect of accumulated deviatoric plastic strain on plastic hardening modulus. Another aspect of this research is to investigate the effect of the change of permeability during the three phases of earthquake simulations, i.e. pore water pressure build up, liquefaction, and excess pore water pressure dissipation. It is unrealistic to expect that a constitutive model can capture all features of soil stress-strain behavior. Hence, this research attempts to provide a model that can be used for simulation of the essential aspects of the stress-strain-strength of granular soils under cyclic loading conditions with reasonable accuracy. In this dissertation, the original and modified constitutive models are first introduced and then the integration schemes and implementation of material model in finite element codes are discussed. After calibrating the constitutive model parameters for three different types of sandy soils, the performance of the modified constitutive model is evaluated by simulating a number of triaxial and direct simple shear tests. The constitutive model is also used to simulate a few boundary value problems including a level ground deposit of Nevada sand (VELACS#1 test), centrifuge model of a waterfront structure (VELACS#11 test), a multilayer system of soil (CUB test), and a sloping ground of Ottawa sand (LEAP test). The simulation results in terms of accelerations, displacements, and excess pore water pressures at different locations of the system are compared to the reported results from the laboratory tests. While the numerical simulations over-predict the acceleration time histories in the liquefied soil, the simulated excess pore water pressures and settlements show reasonably good correlations with the experimental observations in majority of the analyzed cases.