International Journal of Geomechanics

The differential settlement between the newly built embankment and the old one is becoming a serious problem in the extension construction of roadways in some areas. To solve this problem, cement fly-ash gravel (CFG) piles are used to control settlements because of their simple construction. Sheet piles, which are mainly used in railway subgrade construction, can provide greater load performance and reduce settlement. The main objectives of this study were (1) to compare the settlement-controlling performance of sheet piles and CFG piles and (2) to evaluate the application of sheet piles to highway extension engineering. To compare the treatment effect, a case study was conducted. The settlements occurring at the instrumented locations were periodically monitored through the embedded settlement plates and horizontal inclinometers. The settlements in the construction processes were continuously observed. It was found that the settlement rate in the sheet-pile area lessened after surface construction, and the settlement was much smaller than that in the CFG-pile area, even though the initial settlement in the sheet-pile area was larger than that in the CFG-pile area before surface construction. The results suggest that the sheet pile has a significant effect on total and differential settlement control in extension construction. The insights in differential settlement indicate that the sheet pile could improve the subgrade bearing capacity and strengthen the stability of the embankment. Sheet piles could be one of the favorable ground-improvement methods to treat the settlement in highway extension constructions when the construction time and deformation are strictly restricted.

Piles subject to soil lateral movement are often referred to as passive piles. The response of passive piles in clayey soils adjacent to surcharge is not only dependent on the soil intrinsic behavior but also on the loading schedule and consolidation situation. In this study, the effect of consolidation on the behavior of a 2 × 2 pile foundation involved in a warehouse collapse in 1998 in Shanghai, China, was investigated. An experimental study on passive piles in Shanghai soft clay subject to incremental surcharge loading was reviewed. Then, numerical modeling was conducted using two-dimensional (2D) finite-element (FE) models with two different pile–soil interaction methods: embedded and link-element methods. The FE models were validated against the in situ passive-pile test, and then a parametric study on the rate of loading was conducted. The responses of passive piles to fast loading (e.g., linearly applied 120 kPa in 30 days) and staged loading were compared. The former corresponds to the working condition of the collapsed industrial warehouse, and the latter corresponds to the loading schedule of the field experiment. The results show that the displacement of the pile top under the condition of 1-month fast loading was approximately 1.4 times that of the staged loading. The proportion of immediate displacement was also studied with FE modeling. The load transfer between soil and pile and the consolidation process were also investigated.

In contrast to the conventional discrete element method (DEM), no dynamic or damping variables are needed in the static DEM. This study aims at investigating the potential of the static DEM in modeling the quasi-static deformation of stiff granular materials, in which the conventional DEM performs poorly. To represent flexible boundary conditions in the static DEM, the variational formulation for the bonded particle approach is developed, and these formulations are implemented with two-dimensional numerical examples. The boundary conditions are of great importance in numerical simulations. Their effects on the mechanical behavior of granular media are studied at the particle scale. The results show that the force transmission from the boundary to the specimen is more efficient under the rigid boundary condition than under the flexible boundary condition. Finally, the quasi-static behavior of perfectly rigid granular media is investigated with biaxial numerical tests. The peak and residual strengths and strain localization are analyzed by varying the confining stresses and the rolling resistance coefficient. The numerical results are consistent with the common experimental observations in terms of macroscopic stress-strain response, volume-change behavior, strain localization, and force transmission. The macroscopic mechanical response of perfectly rigid particles is a hardening-softening plastic behavior governed by a Mohr-Coulomb yield function. It is also shown that the width of the shear band decreases significantly when the rolling resistance coefficient increases.

Compared with the three-dimensional (3D) simple slope, actual slopes often have irregular slope topography. The 3D convex and concave slopes are two important types of actual slopes. To analyze the stability of 3D convex and concave slopes, it is very important to construct reasonable 3D slip surfaces. Usually, spheres and ellipsoids are used; however, engineering examples show that these specific types of surfaces may be quite different from an actual slip surface. Therefore, based on the study of 3D convex and concave slope models, this work first establishes a new method to generate an arbitrary 3D curved slip surface. According to the generation characteristics of an arbitrary 3D curved slip surface, the 3D sliding body can be divided into several columns during the forming process of the slip surface. Therefore, the generated 3D slip surface is easily combined with the limit equilibrium (LE) column method for calculating the slope factor of safety (FOS) to achieve the purpose of analyzing slope stability. By simulating approximately a number of regular surfaces, such as the sphere, ellipsoid, and so on, the proposed method’s rationality is verified. Additionally, the results of some examples of 3D convex and concave slopes show that compared with a 3D simple slope, the 3D convex slope has poorer stability, whereas the 3D concave slope has better stability.

This paper numerically examines the bearing capacity, settlement, and failure kinematics of two closely spaced circular footings on reinforced soil. A number of large-scale tests are performed to identify the influence of the tilt of interfering footings on their ultimate bearing capacity and settlement, and these experimental results are used to verify the numerical model. Because the mobilization of shear strength in soil and tensile resistance in the geogrid depend to a large extent on the strain level (especially in small strains), a nonlinear elastic-plastic constitutive model in conjunction with a nonassociated flow rule is proposed. In addition, the model accounts for the dependency of the friction and dilation angles on the strain level in the plastic domain. The constitutive parameters are calibrated for the triaxial loading test, whereas the numerical model for the closely spaced footings is verified with reference to the large-scale test results. Thereafter, the ultimate bearing capacity and settlement of interfering circular footings on reinforced soil are studied for different configurations, and the critical size and position of reinforcements that maximize the bearing capacity are characterized. Results show that the ultimate bearing capacity increases up to a maximum of 40 and 90% by the use of one and two layers of geogrid, respectively. Beyond the bearing capacity, the settlement of adjacent circular footings increases up to 45% (compared with a single footing with the same safety factor). Finally, the influence of reinforcing soil on failure kinematics and soil deformation pattern is investigated.

Piping poses a great threat to dam and levee systems across the globe. Although various straightforward approaches have been used to assess piping resistance, a fundamental understanding of the complex soil–flow interactions in this process is still lacking. Practitioners have traditionally relied on Terzaghi’s theory to evaluate piping potential. In this study, the factors affecting piping resistance are revisited using a coupled computational fluid dynamics and discrete element method (CFD-DEM) approach. A series of simulations were conducted to investigate the effects of specific gravity, initial void ratio, particle size distribution, sample aspect ratio, and frictional coefficients. Analyses of simulation results show good agreement with experimental results in the existing literature providing demonstrative explanations for deviations of the experimental results from Terzaghi’s theory. A key finding is that for laboratory experiments with soil samples in a container, the wall friction and the aspect ratio significantly affect the critical hydraulic gradients. The effects of the various soil properties are first analyzed on an individual basis; based on the numerical simulation results, a theoretical model for piping is developed that considers the previously mentioned factors.

Spaces for backfills are often constrained and limited to a narrow width when retaining walls must be built close to existing stable walls in urban areas or near rock faces in mountainous areas. Estimation of lateral earth pressure acting on the back of the retaining wall has been typically based on classical earth-pressure theories such as Coulomb’s theory or Rankine’s theory, which assumes that the backfill is sufficiently wide to allow the full development of failure planes in the backfill. It has been generally recognized that the classical earth-pressure theories tend to significantly overestimate the lateral pressures for retaining walls with narrow backfill width. However, data from experimental studies are scarce, preventing a close examination of how narrow backfill spaces and various wall movement modes impact the forming of the failure surfaces and the distribution of lateral pressure acting on the retaining wall. In this study, a rigid retaining wall model was custom designed and built to investigate the effects of narrow backfill spaces on lateral pressures and backfill failure under active earth-pressure conditions. A series of tests was conducted on combinations of different backfill widths and wall movement modes [translation (T) mode, rotation about bottom (RB) mode, and rotation about top (RT) mode]. The failure surfaces in the backfill and distribution of lateral pressures acting on the back of the wall were recorded for each of the tests. Results of the tests demonstrated that the failure surfaces were continuous and nonlinear and limited by the width of narrow backfills. It was also found that the mode of retaining wall movement plays a role in forming the failure surfaces and mobilizing the lateral pressures. The coefficient of active pressure K_a based on measured lateral pressure decreases along with the depth. Whereas the RT mode showed much higher K_a values at the upper part of the wall, K_a values for the T mode and RB mode were smaller than those found by Coulomb’s theory. The resultant of lateral pressure for all three wall movement modes was significantly smaller than that of Coulomb’s solution, with its point of application generally higher than one-third from the wall base. A simple and practically useful analytical solution of the lateral pressure based on arching theory is presented in this article and shows a reasonably good agreement with the measured lateral pressure for the translation and rotation about the bottom modes of retaining wall movements.

A finite-layer method (FLM) based on consolidation theory is presented for land subsidence due to groundwater withdrawal. The groundwater flow in aquifer systems and the displacement within soil skeletons are approximated by an integration of the standard finite-element method (FEM) in the vertical direction and analytical techniques in the other two spatial directions. By virtue of the analytical eigenfunctions introduced in the formulation, the FEM-weighted residual equations can be decoupled into small-scale linear subsystems, and the formulas for coefficient matrices and flow vectors can be obtained explicitly. Numerical examples are presented to verify the validity of the presented solution through comparisons with available analytical and experimental results. The study also shows that the compressibility of pore water has a significant influence on the pumping-induced subsidence process. The numerical applications to multilayered aquifer systems and horizontal wells are presented to demonstrate the applicability and efficiency of the present method.

Presented in this paper is a modified strain-potential method for the prediction of earthquake-induced permanent deformation of concrete face rockfill dams (CFRDs). The concept of decomposing the total strain of rockfill materials into a viscoelastic part and a plastic part during dynamic loading was explained based on experimental observations. Use of the strain-decomposition concept in the finite-element method resulted in two residual strain–related terms on the right-hand side of the dynamic equilibrium equation. An approach to predict the equivalent number of loading cycles, which is an independent variable in a residual strain model, was proposed. It was assumed that the energy dissipated in an element during an irregular strain history equals that dissipated in an equivalent regular one. The energy-based approach was proved effective in numerically evaluating the equivalent number of loading cycles by simple column tests considering both compression and shearing. The modified strain-potential method can be easily implemented in an equivalent linear analysis program, and the history of permanent deformation accumulation can be directly obtained without recourse to sophisticated constitutive models. Effectiveness of the method was demonstrated by a successful application in a high CFRD.

Based on an extensive list of research work, expansive soil is renowned for its swelling-shrinkage property. Despite the wealth of research, the strength characteristics of expansive soil, which are more complicated than common clayey soil and characterized by a considerable loss of postpeak strength, have not been fully investigated. Within this area of study, soil reinforcement with discrete, short natural fibers is attracting increased attention in geotechnical applications, in part thanks to the unique advantages that this material possesses. Because of these possible advantages, the objectives of the study presented in this paper are (1) to investigate strategies for improving the strength characteristics of expansive soil through the introduction of jute fibers into soil mass and (2) to ascertain jute fibers’ reinforcing effects in relation to several important parameters known to control the effectiveness of fiber-reinforced expansive soil. The study presents discussions related to fiber content, fiber length, fiber orientation, water content, and dry density. More specifically, an experimental program of direct shear and triaxial compression tests was performed on unreinforced and reinforced specimens. The results demonstrate the beneficial role of jute fibers in enhancing strength parameters, enhancing deviator stress-strain behavior, and, if the considered parameters are properly selected, in reducing the postpeak strength loss of expansive soil. The study also presents a detailed discussion regarding the variation of the reinforcing effects and regarding the mechanisms accounting for how fiber-reinforcing effects are affected at the fiber and grain scale. Finally, limitations of this study and potential work for the future are contemplated.

Whether a granular assembly of frictionless particles has shear strength is a very interesting but not well-understood question. This study addressed this question using discrete element method (DEM) simulations along with an energy-based analysis. It is shown that the use of artificial damping in DEM simulations leads to a frictionless assembly exhibiting normal quasi-static shear behavior, with the overall angle of friction at the critical state being nonzero. However, when this artificial damping is absent, the frictionless assembly cannot achieve a quasi-static state but rather exhibits a stress oscillating state, with all particles in vibration, and the shear strength is expected to be zero. From an energy perspective, it is shown that the artificial damping used in DEM simulations plays the sole role in energy dissipation for the frictionless assembly and that it facilitates the establishment of a quasi-static state from which shear strength is mobilized. Therefore, the nonzero angle of shear resistance reported in the literature for frictionless granular assemblies under quasi-static shear should be regarded as a false rather than a true strength parameter.

The development and implementation of a constitutive model used to predict the behavior of unsaturated sands are described. The model is based on critical state and bounding surface concepts and incorporates parameters to account for sand fabric and suction. The model was implemented into finite-element software using user-defined mechanical material behavior (UMAT) subroutine. The model performance was verified with a wide range of triaxial test data that included different drainage conditions, sample preparation methods, densities, confining stresses, and modes of shear. Simulation results show that saturation level significantly affects sand performance and a modified suction-based state parameter is proposed to describe better such sand behavior. This parameter was found to correlate well with an increase in peak stress caused by a decrease in saturation level.

This paper reviews the energy-based solutions that have been proposed for boundary value problems involving pile foundations subjected to lateral loads and presents semianalytical solutions that are applicable to both elastic and elastoplastic soils. Generally, the displacements in the soil are directly linked to the pile deflections through formulations in either Cartesian or cylindrical coordinates. A system of differential equations for the pile deflection and soil displacements can be derived based on the principle of minimum total potential energy or the principle of virtual work in a total or incremental form. The system of equations can be solved analytically or numerically using an iterative algorithm, producing profiles of pile deflections, shear forces, and bending moments for the pile(s) and displacements within the soil domain. Comparisons between results from energy-based analyses using various combinations of the energy principles and displacement formulations show that a stiffer pile response is observed when degrees of freedom in certain directions are constrained in the soil. By adopting suitable displacement formulations and appropriate constitutive models for the soil, the method produces realistic predictions of laterally loaded pile response at an affordable computational cost.

One of the main disadvantages of the advanced tunneling technology of the double-O-tube (DOT) shield is unavoidable rolling of the shield machine during construction. In this study, the influence of the DOT shield rolling on soil deformation during tunnel construction was analytically investigated. Based on the generalized Sagaseta’s solution, the equations for calculating vertical and horizontal soil deformations due to DOT shield tunneling–induced ground loss caused by shield rolling were derived. The theoretical results show that the deformation field of the surrounding soil is obviously changed from symmetry to asymmetry, and the maximum deformation increases observably with rolling. In the case of different rolling angles of the DOT shield, on any horizontal plane in the range of the burial depth of the shield top, an interesting rule was found where the vertical deformation curves have a focal point, whereas the horizontal deformation curves hold two focal points at which the corresponding deformation is constant with rolling. The balanced point of horizontal deformation, at which the horizontal deformation is zero and the maximum settlement occurs, moves gradually toward the rolling side with rolling. These findings from the analytical investigation provide a reliable background in predicting the soil deformation induced by DOT shield tunneling.

Pore-water pressure buildup strongly affects seismic permanent displacement of the slopes constituted from saturated soils. Previous attempts simulating the generation of excess pore pressures (or decrease in soil strength) have been applied in the sliding block models, but the models assumed a rigid block. In the present study, a simplified procedure is presented to account for this effect in sliding block analysis. A correlation between excess pore pressure ratio and cumulative absolute velocity is obtained through the fully coupled solid-fluid effective stress analysis of level ground. The effect of pore-water pressure variation on critical acceleration is investigated in this paper, and the available Newmarkian sliding block methods, such as the rigid block, decoupled, and coupled approaches, are modified to estimate seismic permanent displacement in the presence of excess pore pressure buildup. The present study is original because the method not only simulates the generation of excess pore pressures but also considers the coupled procedure. In addition, centrifuge test results of lateral spreading in the infinite sloping ground are used to demonstrate that the proposed modification considerably improves prediction of seismic sliding displacement. The results clearly demonstrate that ignored pore-water pressure buildup in sliding block procedures can lead to unconservative estimates of seismic permanent displacements during seismic loading.

Deformation localization of granular geomaterials is critical in analyzing the stability of earth structures. This paper presents a new and simple method for quantitatively identifying deformation localization in the granular geomaterials simulated by the discrete element method (DEM) from a new perspective of the microscale. The method involves a contact-deformation-based indicator, ξ, to describe the deformation field and a corresponding threshold, ξ_thr, to highlight the localized deformation from the surrounding zone. The indicator ξ is determined by dividing the relative translation of the two contact points in a contact by the sum of the radii of the two contacted particles. On the basis of this determination, the ξ can always be computed whether the two particles stay in contact or subsequently separate (it can be understood as the disturbance degree of an initial contact). The threshold ξ_thr is quantitatively determined according to the inflection point on the N_ξ'/N_tot−ξ curve (where N_ξ' is the number of contacts with ξ larger than a specified value ξ', and N_tot is the total initial contact number). The zone of ξ > ξ_thr was proved as the explicit shear band via a series of biaxial tests on sands. The modeling results indicate that a smaller ξ_thr corresponds to a more concentrated or narrower localized deformation zone. The proposed method was also extended to search the failure surface of a granular slope and to recognize crack growth and coalescence in a sandstone. The good agreement of the simulated and experimental results suggests the capability of the method in further research of granular system deformation and failure problems.

The Latrobe Valley in Victoria, Australia has the largest brown coal deposit in the world, where large-scale brown coal mines were established in the early 1970s. The geotechnical stability of mining batters requires a comprehensive understanding of the mechanical behavior of the brown coal. Therefore, an experimental study was undertaken to evaluate mechanical properties of brown coal in different loading conditions. In doing so, effects of intermediate loading histories on the mechanical properties of brown coal in a multistep loading (ML), consolidated undrained (CU), triaxial compression (TC) test using a single specimen were evaluated. A set of single-step loading (SL) tests at different confining pressures, with postpeak ML with unload/reload cycles, were also performed. It was found that peak and residual strength and stiffness values and stress–strain relations of brown coal can be estimated accurately by a relevant ML test. Therefore, cyclic loading and accumulated irreversible axial strain do not have any significant influence on the mechanical properties of brown coal. By observing pore-pressure behavior of samples during SL loading, it was found that, with an increase in confining pressure, the amount of damage to the specimen as a result of axial loading decreases, and therefore, the sample experiences contraction even after the peak stress is achieved. The ML increasing-confining-pressure test was found to generate more accurate peak and residual strength values as compared with the ML decreasing-confining-pressure test. In addition, it was concluded that, during the ML decreasing-confining-pressure test, maintaining undrained conditions through the entire test will produce more accurate results.