The 2011 off the Pacific coast of Tohoku Earthquake wreaked enormous damage to many earth structures. In particular, the damage to river dykes has attracted much attention. Conventionally, seismic damage to river dykes had been thought to be caused mainly by the liquefaction of sandy foundation grounds. During the above earthquake, however, collapse of several river dykes built on cohesive foundation grounds was observed. It is believed that the observed collapse occurred as a result of liquefaction of “enclosed saturation area” within the dykes formed by consolidation settlement of the cohesive foundation ground during construction. For interpretation/realization of the collapse mechanism through numerical analysis, it is essential not only to handle both saturated and unsaturated conditions seamlessly but also to carry out static/dynamic analyses and finite deformation analyses for simulating the processes of enclosed saturation area formation due to consolidation settlement and the subsequent dyke collapse. In addition, the constitutive equation that allows a wide spectrum of soils ranging from sandy soils up to clayey soils is also be required. Using a static/dynamic soil-water-air coupled finite deformation analysis code incorporating the SYS Cam clay model as the constitutive equation, which satisfies the above requirement, the behavior of an unsaturated embankment on cohesive soil ground during its construction and during/after an earthquake was examined. The results indicated that it is possible to simulate the formation of an enclosed saturation area within the embankment during its construction and the liquefaction of that enclosed saturation area during an earthquake. Furthermore, new knowledge that the water level within the embankment rise after the earthquake and then return back to the original ground water level, was gained.

View full abstract

Understanding the mechanical behavior of hydrate-bearing sediments during earthquakes is important for the safe and long term gas production from methane hydrates. Additionally, long term and widespread gas production may have effects on the mechanical properties of the seabed ground. In the present study we have analyzed the dynamic behavior of hydrate-bearing sediments considering gas production by the depressurization method. Firstly, we simulated gas production process using a chemo-thermo-mechanically coupled method for different levels of depressurizing pressure. Then, we conducted dynamic analyses by using the results of the production analyses as the initial conditions in order to consider the effects of gas production on dynamic behavior of the sediments. We have modeled the hydrate bearing-sediments as a multiphase mixture composed of gas, water, soil, and hydrates. For the constitutive model for soils, we used a cyclic elasto-viscoplastic model for clayey soils with nonlinear kinematic hardening. A predicted earthquake motion at Kumanonada along the Nankai trough was input. From the numerical results, we have found that the production of natural gas may affect the ground motion during strong earthquakes.

View full abstract

The slope failure so far has been occurred due to the heavy rainfall, and induced losses of a human life and properties. In order to prevent these damages, the measurement and the evaluation for the strength characteristics of the ground such as the shear strength and the angle of internal friction should be required. The direct shear test (DST) which is one of testing methods for this has been generally used, and has several experimental limitations. In particular, it should be recognized that there is no the standard of the opening size between the shear boxes in the DST, and the strength characteristics can vary according to the opening size. Thus, Kim et al. (2012) suggested that the opening size is determined based on the Threshold-Line. In this study, the Threshold-Line (TL) proposed by Kim et al. (2012) was examined through the distinct element method (DEM) simulation. As a result, the shear behaviors in the DEM simulation are similar to the experimental results by Kim et al. (2012). Through the results of the DEM simulation, it can be confirmed that the TL can be a standard for the opening size in the DST.

View full abstract

A series of solute transport simulations in physically heterogeneous aquifers were conducted to assess the impacts on macrodispersion under saturated flow conditions. The aquifer system of concern relied on the hydrogeological data in a site in the Netherlands and was generated geostatistically as randomly correlated hydraulic conductivity. Ensemble computations in 100 realizations were based upon random walk particle tracking linked with temporal moments to estimate three macrodispersivities. The results showed the scale dependence of macrodispersivities and the asymptotic nature of longitudinal macrodispersivity. From a practical standpoint, the simulations demonstrated that control plane and its middle element provide almost identical macrodispersivity estimates in terms of transverse mixing, but differ markedly from estimates of longitudinal mixing.

View full abstract

The breaking of dam can cause sudden debris flow which leads to hazardous consequences. This paper presents a coupled Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) approach to study the dam break problem. The CFD is employed to investigate the fluid flow by solving the locally averaged Navier-Stokes equation, while the DEM is used to simulate the granular particle system based on the Newton’s equation of motion. The fluid-particle interaction is accounted for in the modeling by exchanging interaction forces such as the drag force and the buoyancy force between the CFD and the DEM. In simulating the dam break problem, a mixture of viscous fluid and uniform particles is initially confined within a cubic container with a removable side gate which is subsequently initiate the dam breaking. Four comparison cases are investigated, including a Bingham fluid-particle sample, a water-particle sample, a pure Bingham fluid sample and a dry particle sample. The Bingham model is employed to simulate a viscous fluid consisting of water and fine particles, which allows the DEM simulate big gravels and boulders only and lends a great computational efficiency. The simulations enable us to examine the flow patterns of different samples and interactions between particles to understand the mechanisms of dam break better.

View full abstract

Numerical simulations of triaxial compression tests are performed in order to investigate the applicability of the material point method (MPM) to large-deformation geotechnical problems. A Mohr-Coulomb constitutive model is adopted as a geotechnical nonlinearity, for which parameters are obtained by experimental results. The stress-strain relationships obtained in the simulations show good agreement with the experiments at high levels of strain, thus demonstrating the effectiveness of MPM. Parametric studies are also performed in order to investigate factors that influence the relationship between the initial and final shear bands observed during simulations of triaxial compression tests. The main findings of the study are followings: (1) initial shear bands are generated near the constrained edges of a specimen; (2) when both top and bottom edges of a specimen are constrained, initial shear bands near the edges at which the external forces are applied develop finally into clear shear bands; (3) specimen geometry has a major influence on the formation of shear bands in the case of rectangular specimens.

View full abstract

The finite element analysis, in which an elastoplastic Cosserat continuum model for the soil is incorporated, is implemented to simulate the strain localization phenomena due to strain softening or non-associated plasticity of the material. Based on the finite element procedure developed for proposed pressure-dependent elastoplastic Cosserat continuum model, progressive failure phenomena which occurring in the shear layer and the full scale test of filling embankment on soft foundation characterized by strain localization due to material softening and the material dilatancy, i.e. non-associated plasticity, are numerically simulated respectively. Numerical results indicate that the classical continuum finite element may suffer from pathological mesh dependence and be incapable of completing the analysis of the whole failure process, while Cosserat continuum finite elements possess better performance in preserving the well-posedness of the localization problems and in completing the simulation of the entire progressive failure process occurring in geotechnical engineering structures.

View full abstract