Search

This study combined a neural network and Latin hypercube sampling (LHS) to calibrate soil parameters. The Monte Carlo parameters were calibrated by generating different numbers of training samples for pressuremeter tests and excavations. The results showed that when the number of samples was 25 or 50, the parameter calibration accuracy was very high. However, the improvement in accuracy did not increase significantly with a further increase in the number of samples, but tended to be stable. The number of training samples was set at 50 to strike a balance between the calibration accuracy and efficiency for four parameters. For 25 groups of samples, the calibration results using LHS were better than those using orthogonal sampling. Compared to stochastic optimization algorithms, a neural network combined with LHS could significantly reduce the calibration time. This method was applied to actual foundation pit engineering in China. The results showed that using the proposed calibration method clearly improved the accuracy when predicting the deformation induced by the excavation.

A uniform arrangement of individual piles is commonly adopted in the conventional pile group foundation, and basin-shaped settlement is often observed in practice. Large differential settlement of pile groups will decrease the use-safety requirements of building, even cause the whole-building tilt or collapse. To reduce differential settlement among individual piles, non-uniformly arranged pile groups can be adopted. This paper presents a finite element analysis on the response of pile groups with different layouts of individual piles in pile groups. Using the user-defined subroutine FRIC as the secondary development platform, a softening model of skin friction and a hyperbolic model of end resistance are introduced into the contact pair calculation of ABAQUS software. As to the response analysis of a single pile, the reliability of the proposed secondary development method of ABAQUS software is verified using an iterative computer program. The reinforcing effects of individual piles is then analyzed using the present finite element analysis. Furthermore, the response of non-uniformly arranged pile groups, e.g., individual piles with variable length and individual piles with variable diameter, is analyzed using the proposed numerical analysis method. Some suggestions on the layout of individual piles are proposed to reduce differential settlement and make full use of the bearing capacity of individual piles in pile groups for practical purposes.

Based on the ‘t-z method’, this paper presents a simplified approach to analyzing the nonlinear response of a single bored pile embedded in sand. The lateral pile-sand interaction of the pile shaft is modeled using the cylindrical cavity expansion theory, and the axial load-settlement behavior of the pile base is analyzed using the spherical cavity expansion theory. The softening model and hyperbolic model are employed to describe the strain-softening and strain-hardening behavior, respectively, of the axial pile-sand interaction observed in elemental tests. The constrained shear-dilatancy response near the pile shaft and the particle-crushing behavior beneath the pile base are subsequently incorporated into the parameters of the present load-transfer models. Using the proposed load-transfer functions, an efficient iterative program is then developed to predict the nonlinear response of a single bored pile embedded in sand. Comparisons of the responses of the bored pile embedded in sand demonstrate that the present calculated results are generally in good agreement with the well-documented observations and the computed results derived from other approaches. A parametric study is also carried out to assess the influence of some key parameters related to the proposed model on the load-displacement response of a single bored pile embedded in sand.

How to accurately estimate the permanent deformation of soils due to large number cycles of cyclic loads is essential in many engineering projects, such as offshore wind turbine foundation and high-speed railway. It is still a challenging issue, although many works have been carried out. In this study, a series of cyclic triaxial tests with different levels of initial mean effective stress, relative density, initial static stress ratio, and cyclic stress ratio were performed on saturated Toyoura sand. Test results clearly showed that the vertical permanent strain increased with increasing cyclic stress ratio and initial mean effective stress. Moreover, the shear strain amplitude during the initial loading of the first load cycle was used to describe the intensity of soil particles rearrangement which is the mechanism of permanent deformation. A unique relationship between the shear strain amplitude and the vertical permanent strain was established. A four-parameter model was subsequently proposed for estimating the permanent strain of sand under cyclic loading, which inherently considered all the main influencing factors. Finally, the comparison between the estimation and the measurement in laboratory tests convincingly illustrates the adequacy of the proposed model in estimating the permanent deformation of sand under cyclic loading.

Recently traffic induced geotechnical problems have been receiving increasing attention. Yet the silty clay from the third layer of Shanghai, which is within the main influence depth of traffic loading and overconsolidated, had not been sufficiently studied. A series of monotonic and cyclic undrained triaxial tests were thus carried out on reconstituted Shanghai silty clay with various overconsolidation ratios (OCRs). The silty clay specimens showed phase transformation behavior, which was characterized by stages of initial contraction, temporary phase transformation and later dilation. Such behavior could be termed as a transitional mode between the behavior seen for typical clays and that for sands. The dilation behavior in the later stage was shown more significant for overconsolidated specimens. Six different failure criteria were employed to study the friction angle. There was no observable effect of OCR on the friction angle for this silty clay as if it was a granular material. The undrained shear strength could be normalized with only a small deviation. The normalized shear strength increased with OCR and it behaved more as typical silts than clays at high OCRs. The development of permanent strain under cyclic loading was shown to depend significantly on the cyclic stress ratio and OCR. A higher OCR resulted in smaller permanent and resilient strains, and the decrease of the permanent strain with OCR was remarkable for lightly overconsolidated specimens. Such effect of OCR would be more significant in real soil condition where the soil is at least partially drained.

A series of cyclic torsional shear tests using hollow cylinder apparatus (HCA) were performed to investigate the effect of principal stress rotation (PSR) on the stress–strain behaviors of saturated soft clay. The traffic–load–induced shear stress path was used in the cyclic test and the investigation mainly concerned the influence of PSR on the shear stiffness and non-coaxiality. It indicated that the effects of PSR substantially depends on the magnitude of deviatoric stress (q = {[(σ1 − σ2)2 + (σ2 − σ3)2 + (σ1 − σ3)2]/2}1/2) as well as the intermediate principal stress ratio (b = (σ2 − σ3)/(σ1 − σ3)). At low deviatoric stress, the trajectory envelope of deviatoric strain path translates with a nearly constant size, showing constant shear stiffness and strong non-coaxiality. However, at high deviatoric stress, the trajectory envelope of deviatoric strain rapidly expands towards instability, showing degenerating shear stiffness and weak non-coaxiality. Moreover, the excess pore water pressure increases and the shear stiffness decreases more rapidly as b value increases. The results can provide an experimental basis for constitutive modelling of clays under traffic–induced loadings.

With wind farm development, a new type of foundation, the piled beam–slab foundation, was invented to support wind turbines on land. Although this foundation enables a lower-cost investmen...

The evolution of toppling deformation of anti-dip slope is essentially a process of energy dissipation and transformation. Aiming to study the characteristics of energy evolution in different stages, the DEM (discrete element method) software PFC (Particle Flow Code) was utilized to establish a two-dimensional numerical model for a bank slope in Chongqing based on geological background data and field investigation. The DEM model was proven to be reliable not only because the deformation discrepancy between the numerical model and actual bank slope was not large but also because some obvious fractures in the actual bank slope can readily be found in the numerical model as well. In this article, content about displacement in the shallow layer was analyzed briefly. Special effort was made to analyze the energy field and divide the toppling deformation process into three stages. (1) Shear deformation stage: this is an energy accumulating stage in which the strain energy, friction energy, and kinetic energy are all small and the deformation is mainly shear deformation in the slope toe. (2) Stage of main toppling fracture surface hole-through: all three kinds of energy present the increasing trend. The shear deformation in the slope toe expands further, and the toppling deformation also appears in the middle and rear parts of the bank slope. (3) Stage of secondary toppling and fracture surface development: strain energy and friction energy increase steadily but kinetic energy remains constant. Deformation consists mainly of secondary shearing and a fracture surface in the shallow layer. Secondary toppling and fracture surface develop densely.

This paper gives a set of time-dependent displacement solutions for the case of a vertical concentrated force acting in the interior of fractional derivative viscoelastic soils. First, a t...

Anchor cables are now widely used to stabilize many kinds of geostructures in practice. However, the prediction of the anti-pullout performance of anchor cables still involves uncertainty for designer. This paper presents an analysis procedure for the evaluation the ultimate pullout capacity of anchor cables embedded in rock. A unified rupture shape model is proposed to describe the failure surface shape of anchor in rock mass by incorporating typical failure mode and failure surface shape. Using the limit equilibrium method, a formula of ultimate pullout capacity of a single anchor (cable) considering layout pattern, grouting pressures, rock masse quality and construction disturbance is then established with the assumption that surrounding rock follows the generalized Hoek–Brown criterion. Performance of the present method is assessed using comparison of the theoretical and well-documented filed test results. Parametric study is also carried out to clarify the influence of some key indices for rock mass on the ultimate pullout capacity of anchor cables embedded in rock. The present analyses shows that the geological strength index has great influence on ultimate pullout capacity of a single cable and ultimate pullout capacity of a single cable could be improved to a certain extent by reducing construction disturbance.

Shakedown solution is widely used to analyse the elastic-plastic behaviours of structures in pavement design. Previous studies only concerned the shakedown theorem under traffic loads for an isotro...

The deformation of the gravel soil is an important component of the total settlement of the pavement under traffic load. The traffic load induces a heart-shaped stress path in the deviatoric stress space characterised by the rotation of principal stress. However, cyclic triaxial tests are usually carried out to investigate the deformation of gravel soil, which cannot account for the effect of principal stress rotation. In this study, discrete element method (DEM) is used to investigate the response of granular soil subjected to a heart-shaped stress path, and the results are compared with cyclic simple shear tests and biaxial tests. The results indicate that the accumulated volumetric strain of the soil increases with increasing cyclic stress ratio and initial confining pressure. Meanwhile, the accumulated volumetric strain changes from contractive to dilative when the cyclic stress ratio increases from 0.3 to 0.7. At the particulate level, the degree of fabric anisotropy increases and the princip...

With the expansion of existing networks of underground system, many metro tunnels are excavated in a pair undercrossing existing tunnels. The new excavations will affect the stress distribution and settlement behavior of the existing tunnels, particularly when the clearance is small. Generally, the construction of the second undercrossing tunnel induces more settlement in the upper tunnels compared to that caused by excavating the first one, due to the interaction between the new excavations. The degree of interaction makes it difficult to predict the settlement of existing tunnels. In this paper, a simplified two-stage analysis method considering the interaction between the undercrossing twin tunnels is proposed to estimate the induced response of existing upper tunnel. In the first stage, the greenfield settlement caused by the first excavation is estimated using Peck’s empirical function, and that caused by the second excavation is modified with a modification factor to consider the influence of the first tunnel. In the second stage, the additional stress calculated using the greenfield displacement is imposed on the existing tunnel resting on Pasternak foundation model. The calculated settlement profile and bending moment of the existing tunnel using the two-stage analysis method are compared with those obtained from finite element analysis and historical field monitored data. The effects of the characteristic parameter of Pasternak foundation model, the spacing between the new twin tunnels, and the modification factor of greenfield settlement, on the settlement profile and bending moment of the existing tunnel are also investigated.

For tunnels built in the saturated silty sand ground, fine particles may be migrated into tunnels through seams of tunnel segmental joints and then seepage erosion is triggered, which may induce ground settlement. However, the process from fine particles erosion to the stress redistribution and soil properties’ change surrounding the tunnel and ground settlement has not been clarified up to now. For this purpose, five numerical tests of seepage erosion in granular soils around the tunnel are conducted using the Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) coupling method. The influences of buried depth and grain size distribution (GSD) of gap graded soils (mainly controlled by the fines content and mean particle size ratio from coarse to fine) on the seepage erosion around the tunnel are investigated. Eroded mass, fines loss mode, surface vertical displacement, stress redistribution, fabric anisotropy, soil behavior and water pressure around the tunnel during the seepage erosion process for five tests are presented and compared. The following results can be upscaled to the practical tunnel engineering, such as: (1) the number of fines loss, the eroded zone and the ground settlement increase with buried depth and mean particle size ratio; (2) the earth pressure near the crack significantly increases due to the stress redistribution induced by fines loss, and the stress redistributed area expands with buried depth; (3) the strength and stiffness of granular soils around the crack are significantly reduced by the seepage erosion. All results revealed that the CFD-DEM simulations provide a new sight on understanding the mechanics of tunnel seepage erosion from a microscopic perspective.

Since cross-anisotropic sand behaves differently when the loading direction or the stress state changes, the influences of the loading direction and the intermediate principal stress ratio (b = (σ 2 − σ 3)/(σ 1 − σ 3)) on the initiation of strain localization need study. According to the loading angle (angle between the major principal stress direction and the normal of bedding plane), a 3D non-coaxial non-associated elasto-plasticity hardening model was proposed by modifying Lode angle formulation of the Mohr–Coulomb yield function and the stress–dilatancy function. By using bifurcation analysis, the model was used to predict the initiation of strain localization under plane strain and true triaxial conditions. The predictions of the plane strain tests show that the major principal strain at the bifurcation points increases with the loading angle, while the stress ratio decreases with the loading angle. According to the loading angle and the intermediate principal stress ratio, the true triaxial tests were analyzed in three sectors. The stress–strain behavior and the volumetric strain in each sector can be well captured by the proposed model. Strain localization occurs in most b value conditions in all three sectors except for those which are close to triaxial compression condition (b = 0). The difference between the peak shear strength corresponding to the strain localization and the ultimate shear strength corresponding to plastic limit becomes obvious when the b value is near 0.4. The influence of bifurcation on the shear strength becomes weak when the loading direction changes from perpendicular to the bedding plane to parallel. The bifurcation analysis based on the proposed model gives out major principal strain and peak shear strength at the initiation of strain localization; the given results are consistent with experiments.

The laboratory tests had revealed that the fabric anisotropy plays an important role in the mechanical behaviour of the granular materials. However, the fabric anisotropy includes two aspects, the degree and the principle direction, and their contributions to the mechanical behaviour cannot be separated in the laboratory tests. In this study, granular soil specimens with different degrees and principle directions of inherent anisotropy were successfully produced. The mechanical behaviours of these specimens were numerically investigated by drained biaxial tests using discrete element method (DEM). Meanwhile, the evolutions of the microstructure, including coordination number, anisotropy of the contact normal and the clump orientation, were monitored during the shearing. It was found that not only the principal direction of inherent anisotropy can significantly influence the strength and the deformation characteristics, but also the degree of inherent anisotropy. The peak shear strength first decre...

This paper compares the pre-failure instability of sands under axisymmetric proportional strain and axisymmetric proportional stress conditions. Based on a deviatoric hardening plasticity model, th...

Foundation soils are often under non-proportional cyclic loadings. The deformation behaviour and the mechanism of non-coaxiality under continuous pure principal stress rotation for clays are not clearly investigated up to now. In order to study the effect of pure principal stress rotation, a series of cyclic undrained tests on Shanghai soft clay subjected to cyclic rotation of principal stress directions keeping the deviatoric stress constant under the pure rotation condition were conducted using hollow cylinder apparatus. Based on this, the evolutions of excess pore pressure and strains during cyclic loading were investigated, together with the effects of the intermediate principal stress parameter and the deviatoric stress level on stress–strain stiffness and non-coaxiality. The result can provide an experimental basis for constitutive modelling of clays describing the behaviour under non-proportional loadings.

Discrete element modeling was used to investigate the effect of particle size distribution on the small strain shear stiffness of granular soils and explore the fundamental mechanism controlling this small strain shear stiffness at the particle level. The results indicate that the mean particle size has a negligible effect on the small strain shear modulus. The observed increase of the shear modulus with increasing particle size is caused by a scale effect. It is suggested that the ratio of sample size to the mean particle size should be larger than 11.5 to avoid this possible scale effect. At the same confining pressure and void ratio, the small strain shear modulus decreases as the coefficient of uniformity of the soil increases. The Poisson’s ratio decreases with decreasing void ratio and increasing confining pressure instead of being constant as is commonly assumed. Microscopic analyses indicate that the small strain shear stiffness and Poisson’s ratio depend uniquely on the soil’s coordination number.

The aim of this study is to numerically investigate the influence of particle breakage on the mechanical behavior of granular materials using a discrete element method (DEM). To enable particle crushing, non-crushable elementary particles are boned together to represents the granular aggregates which can be crushed when the external force exceeds its strength. The flaw of the aggregate was also modeled by randomly distributed void. Single particle crushing tests were carried out to determine the distribution of particle strength. The results of single particle crushing tests illustrate that the simulated single particle fracture strength and pattern agree well with the Weibull’s distribution equation. Conventional oedometer tests, drained monotonic and cyclic triaxial tests were also carried out to investigate the crushing of the aggregates and the associated mechanical behaviors. The effect of confining pressure on the crushing of aggregates and the mechanical behavior was also analyzed. It was found that the peak stress and dilation decrease significantly when particle crushing was considered. The deformation behavior of the specimen is essentially controlled by two factors: particle rearrangement - induced dilation and particle crushing - induced contraction. The increase of permanent strain and the reduction of dilation were observed during cyclic loading and they tend to reach a stable state after a certain number of cycles. The crushing of aggregate is most significant in the first two cycles. The results also indicated that for the same axial strain the volumetric strain and the bound breakage in the cyclic loading tests are significantly larger than that in the monotonic loading tests, especially at high cyclic stress ratio.

Cast in situ and grouted concrete helical piles with 150-200 mm diameter half cylindrical ribs have become an economical and effective choice in Shanghai, China for uplift piles in deep soft soils. Though this type of pile has been successful used in practice, the reinforcing mechanism and the contribution of the ribs to the total resistance is not clear, and there is no clear guideline for the design of such piles. To study the inclusion of ribs to the contribution of shear resistance, the shear behaviour between silty sand and concrete slabs with parallel ribs at different spacing and angles were tested in a large direct shear box (600 mm × 400 mm × 200 mm). The front panels of the shear box are detachable to observe the soil deformation after the test. The tests were modelled with threedimensional finite element method in ABAQUS. It was found that, passive zones can be developed ahead of the ribs to form undulated failure surfaces. The shear resistance and failure mode are affected by the ratio of rib spacing to rib diameter. Based on the shape and continuity of the failure zones at the interface, the failure modes at the interface can be classified as "punching", "local" or "general" shear failure respectively. With the inclusion of the ribs, the pull out resistance can increase up to 17%. The optimum rib spacing to rib diameter ratio was found to be around 7 based on the observed experimental results and the numerical modelling.

In situ cast concrete piles with helical ribs have been used in Shanghai, China for uplift piles. The piles are about 600 mm in diameter with half-cylindrical ribs of 75 to 100 mm in radius. Pull-out tests on model helical ribbed nylon piles of different rib spacing have been carried out in a large pull-out box under different confining pressures. Pull-out resistance and displacement of the piles revealed that the resistance varies with rib spacing and is at a maximum when the ratio of rib spacing to pile diameter is 1, and the resistance is twice that of a smooth pile with the same outer diameter. Three-dimensional finite-element analysis of the model piles revealed that the inclusion of ribs changed the soil and pile interaction mechanism. Shear wedges may form ahead of the ribs. The size of the wedges varies with rib spacing, and is the largest when the ratio of rib spacing to pile diameter is about 1. The wedges may connect to each other to form larger shear zones around the piles to increase the pile resistance. It is suggested that rib spacing is one of the most important factors that affect the pull-out resistance of the piles with the designed geometry.

This paper presents a microscopic investigation on the effects of initial anisotropy, drainage condition, and consolidation state on the mechanical behavior and strain localization of dense sands. Discrete element simulations with non-spherical clumps were carried out to simulate drained and undrained biaxial tests of isotropically and K0 consolidated anisotropic sands. In drained tests, the stress–strain relationship shows an initial hardening and subsequent softening behavior, and the peak shear stress decreases as the bedding plane angle increases. K0 consolidation has slight influence on the peak friction angle and does not affect the friction angle at zero dilatancy. In undrained tests, softening behavior occurs when the bedding plane angle is small, while for higher bedding plane angle, the strengthening response takes over. As the bedding plane angle increases, the peak friction angle decreases initially but increases afterwards. The relative displacement and rotation angle of clumps as well as the void ratio distribution within the specimen indicate the appearance of shear band. Shear bands leads to the inhomogeneous deformation field within specimens. Excessive dilation inside of shear band is produced, and it may induce re-contraction behavior under drained condition and may re-increase the pore water pressure under undrained condition. The appearance of shear band reduces the peak shear strength, and the specimen with a low bedding angle results in a larger reduction of shear strength than that from the specimen with a high bedding angle. Particle rotation mode and force chain network change along with the formation of a shear band. As the longest axis of clumped particle varies from vertical to parallel with respect to the loading direction, the majority of particle contacts inside of shear band changes from multi-point mode to single contact mode. The obtained shear band width and inclination angle were computed, and their variations with bedding angle were obtained.

Traffic loading causes the rotation of principal stress axis and generates a heart-shaped stress path in the deviatoric stress space. Using dynamic hollow cylinder apparatus (DHCA), a series of cyclic heart-shaped and cyclic triaxial undrained tests were performed on Shanghai clay through simultaneously varying the torsional shear stress and the normal stresses. Experimental results show that the accumulated undrained responses of clay at different stress levels can be described by the shakedown approach. The permanent strain and energy dissipation were used to classify the accumulated deformation patterns, including plastic shakedown, cyclic plastic creep and ratcheting (incremental instability). For the plastic shakedown, the IPC (increment per cycle) of permanent axial strain becomes negligible and the IPC (area of hysteretic loop per cycle) of energy dissipation tends to reach a low constant level after cycling. For the cyclic plastic creep, the IPC of permanent axial strain (plastic strain increment per cycle) and the IPC of energy dissipation become nearly constant after a large number of cycles. For the ratcheting, the IPCs of strain and energy dissipation increase after a limited number of cycles. Based on these results, a new method is proposed to define the boundaries of shakedown behaviors using the effective cyclic stress ratio.

Shakedown theory is often used to analyze the plastic behaviors of structures subjected to variable complex loads. It is an effective way to predict the maximum load under which the failure due to plastic collapse or excessive permanent deformation of pavement will not occur. Based on Melan's lower-bound shakedown theory, this work has proposed a numerical method for estimating the shakedown limit involving the effect of traffic moving speed. The dynamic response of elastic stress to traffic moving speed is computed using combined Finite Element-Infinite Element (FE-IE) method. The shakedown limits for a two-layered pavement system have been investigated at various traffic moving speeds. It shows that the shakedown limit early reduces and subsequently turns to grow as the moving speed increases. The shakedown limit decreases to the minimum when the traffic speed reaches the Rayleigh wave speed of subsoil. In order to provide an optimized design of the pavement system, the dependence of shakedown limits on the material properties of pavement system has also been estimated. Eventually, the characteristic distribution of critical residual stress is discussed.

The deformation and the degradation of ballast significantly affect the performance of the railway track. In this study, the discrete element method (DEM) was used to investigate the deformation behavior and degradation characteristics of ballasts under monotonic and cyclic loading. The ballasts were simulated as crushable aggregate which is modeled by bounded discrete elementary particles together and it can be crushed under external forces. The flaw of the aggregate was also modeled by randomly distributed void. Conventional monotonic and cyclic triaxial tests were carried out to investigate the crushing of the aggregates and the associated mechanical behaviors. The effects of confining on the crushing of aggregates and the mechanical behavior were also analyzed. It was found that the permanent deformation of the aggregates significantly increases when particle crushing is considered. The crushing of aggregate is most significant in the first two or three cycles.

This study aims to investigate the dynamic responses of layered poroelastic ground underlying rough pavement subjected to traffic loads. Particular attention has been given to the roles of the pavement roughness idealized as sinusoidal waves. The vehicle is simplified as a multi-rigid-body vibration system and the pavement roughness results in dynamic traffic loads in addition to the static component due to the vehicle weight. The pavement layer is modelled as a thin Kirchhoff plate, while the underlying base and the subgrade are treated as saturated two-layered poroelastic medium obeying Biot's dynamics theory. Frequency-wavenumber domain responses of the vehicle-pavement-underlying ground are solved by Fourier integral transform, transfer matrix method and considering stress and displacement compatibility conditions at the pavement-ground interface. The corresponding time-domain solution is numerically obtained from the inverse Fourier transform. Based on this semi-analytical solution, the influences of surface roughness, vehicle velocities and mechanical properties of pavement and base materials are discussed to provide practical guideline for the design of rough pavement system.

This paper presents a micromechanics-based approach to investigate the effects of fabric anisotropy on the behavior of localized failure in granular materials. Based on a micromechanical analysis, the origin of deviatoric stress is decomposed into two components: contact force anisotropy and fabric anisotropy. Using a micro–macro approach, the back stress is interpreted as an contribution to the change of the fabric’s principal direction. The evolution of the back stress is deduced from the stress–fabric relationship and determined with reference to the deviation of the principal directions between the rate of the reduced stress tensor and the actual reduced stress tensor. With this micro–macro framework, a mixed (isotropic–kinematic) hardening model is developed based on the classical isotropic hardening theory. A laboratory simple shear test is first analyzed to validate the proposed model and illustrate the kinematic-hardening mechanism which is usually displayed under non-proportional loading. The analysis further focuses on the anisotropic aspect of localized failure. It has been discovered that the fabric anisotropy can play an important role in the occurrence of shear banding. An increasing degree of fabric anisotropy tends to delay the initiation of the strain localization and result in higher failure strength. The effects of fabric anisotropy have also been illustrated by comparing the theoretical predictions and measured results on the shear band inclination angle, shear strain level and dilatancy at bifurcation.

The discrete element method was adopted in this paper to numerically investigate the macroscopic behaviors of granular material with initial anisotropic fabric in simple shear tests. The evolutions of the stress-strain relationship, the non-coaxiality referring to the difference of principle directions between the strain rate and the stress increment, and the fabric anisotropy were analyzed during the tests. The results indicated that inherent anisotropy affects the shear strength. The directions of principal strain rate and principal stress are non-coaxial at the start of the shear loading and they become coaxial gradually as the shear strain increases. Both the principle directions of particle orientation and contact normal change towards 45°, but the later is much faster than the other.