International Journal of Geomechanics

The geosynthetic-reinforced pile-supported embankment is one of the favorable ground improvement techniques used in the construction of earth structures over a compressible soil when limited construction time is available and limited deformation is permissible. Various methods are available for the design of the geosynthetic-reinforced platform based on various load transfer mechanisms from the embankment to the piles and the compressible soil. The existence of the geosynthetic layer makes the mechanisms more complex. This study focuses on the behavior of geogrid-reinforced embankments over piles compared with the behavior of unreinforced embankments. The numerical simulations of the unreinforced and reinforced pile-supported embankments were conducted using the discrete element method (DEM). The embankment fill was simulated using unbonded graded aggregates of diameters ranging from 9.2 to 20.8 mm and the geogrid was simulated using bonded particles. This study investigated the changes of vertical and horizontal stresses and porosities, the vertical displacements within the embankment fill, and the deflection and tension in the geogrid. The simulation results showed that the coefficient of lateral earth pressure in the embankment fill changed from an initial at rest condition to a passive condition at certain locations after the compression of the compressible soil. The embankment fill dilated during the development of soil arching. The embankment load was transferred to the piles owing to the reorientation of the principal stresses. The results also showed that the geogrid reinforcement significantly reduced the total and differential settlements at the top of the embankment.

Considering that most geosynthetic-reinforced soil structures (GRS) are backfilled with well-compacted granular soils exhibiting evident strain-softening, it is expected that the mobilized soil strength beyond the peak would affect the response of GRS structures if the earthquake-induced deformation is large. This issue is particularly relevant if GRS structures are to be designed against seismic loading based on permanent displacement. In this study, a calibrated finite-element procedure was used to investigate the influences of strain softening of backfill soils on the deformation and reinforcement load of wrapped-face GRS walls. Dense Toyoura sand, a medium dense Japanese silty sand, and loose Fuji River sand were used as backfills of model GRS walls, which were subjected to sinusoidal excitation ranging from 0.1 to $0.6g$. From the study, it is found that the permanent deformation of GRS walls was attributed to the compaction of backfill, smeared shear deformation of soil, shear deformation along slip surfaces, and free-field displacement in the retained earth if it was adequately deep. The slip surfaces were related only to the peak soil strength. The maximum reinforcement load was directly related to the strain softening of backfill soil. Soils with larger peak strength but smaller residual strength could result in larger reinforcement load in strong earthquakes. The results imply that using residual strength along the slip surface determined from peak strength may be more appropriate if GRS structures are to be designed against rare-event strong earthquake loading.

Using a high-strength geosynthetic for encasement of granular columns increases the strength of a given column and improves its stress-displacement response. This paper describes the results from a series of three-dimensional finite-element analyses that were performed to simulate the behavior of a single geosynthetic-encased column in soft clay. To examine the sensitivity of the results to the constitutive model that was used to simulate the behavior of the encased granular soil, analyses were performed using models for the encased soil possessing various levels of sophistication. For each model that was studied, comparative analyses were performed to simulate the behavior of a dense and a loose granular soil within the encasement. The results demonstrate the importance of selecting a constitutive model that accurately captures the shear-induced volume-change behavior of the encased granular soil. Additional findings provide guidance for other researchers seeking to model the behavior of geosynthetic-encased columns and are useful for practitioners looking to enhance their understanding of the performance of different constitutive models for simulating the behavior of similar foundation elements.

Mechanistic-empirical modeling and design solutions have been routinely used for conventional flexible pavements and have recently entered mainstream design practice through the issue of AASHTO design software. Recent work has been performed to incorporate mechanisms of base reinforcement with geosynthetics into mechanistic-empirical modeling and design. The purpose of this paper is to examine the use of this work for mechanistic-empirical modeling of reinforced unpaved roads by using the same model components used for paved roads and incorporating new components that account for the significant influence of pore-water pressure generation in the subgrade. Results from instrumented unpaved road test sections are used to calibrate and compare with the model predictions. The model provides reasonable predictions of rutting behavior for test sections where rut development is stable, where stable is defined as a rate of rutting that decreases with increasing traffic. The importance of excess pore-water pressure development in the subgrade is illustrated by the test sections and accounted for by the inclusion of steps in the model.

The long-term performance of embankments with creep susceptible geosynthetics as a basal reinforcement layer over rate-sensitive soils is examined. Various factors, including reinforcement type and stiffness, soil viscosity, construction rate, and allowable long-term reinforcement strain, affecting the time-dependent behavior of reinforced embankments are investigated. Increasing the reinforcement stiffness (for a given allowable strain) allowed construction of higher embankments from a stability perspective; however, it also resulted in larger deformations and potential serviceability issues. Lateral toe spreading, maximum settlement, and differential settlement increased with the increasing rate of construction. Reducing the allowable long-term reinforcement strain proved to be an effective means of limiting embankment deformation. The analyses also showed that if the long-term reinforcement strain can be limited to $<5\%$ and combined with the practical rate of construction, the long-term deformations of reinforced embankments can be controlled.

Geogrids are commonly used in road construction for stabilization and reinforcement purposes. Factors affecting the interaction or interlock mechanisms between geogrids and aggregates may include, but are not limited to, aggregate size and shape and geogrid types and properties. To better quantify these effects, an aggregate image aided discrete element method (DEM) modeling approach is introduced in this paper. DEM simulations of laboratory direct shear tests carried out on both unreinforced and geogrid-reinforced aggregate shear box samples indicate that the aggregate imaging aided DEM can accurately predict both unreinforced and geogrid-reinforced aggregate strength properties. The use of geogrids increased the shear strength of the aggregate assembly by constraining the movement of the aggregates in the shear zone, which is often referred to as the geogrid’s stiffening effect in this aggregate-geogrid composite system. Preliminary findings on the effects of geogrids with various opening shapes and geometries on the mechanical interlock are also presented to demonstrate the effectiveness of the aggregate image aided DEM model and its potential for quantifying the individual effects of geogrid aperture size and shape relative to aggregate size and shape, gradation, and density, as well as the shape and stiffness of the ribs and the stiffness of the junction between the ribs of various geogrid products.

The results of load and resistance factor design (LRFD) calibration are reported for the pullout limit state in geogrid reinforced soil walls under self-weight loading and permanent uniform surcharge. Bias statistics are used to account for the prediction accuracy of the underlying deterministic models for load and pullout capacity and the random variability in the input parameters. The paper shows that the current AASHTO simplified method to calculate reinforcement loads under operational conditions is overly conservative leading to poor prediction accuracy of the underlying deterministic model used in LRFD calibration. Refinements to the load and default pullout capacity models in the AASHTO and Federal Highway Administration guidance documents are proposed. These models generate reasonable resistance factors using a load factor of 1.35 and give a consistent probability of pullout failure of 1%. A comparison with the allowable stress design (ASD) past practice shows that the operational factors of safety using a reliability-based LRFD approach give factors of safety greater than 1.5. Regardless of the design approach (ASD or LRFD), the analysis results demonstrate that the current empirical minimum reinforcement length criteria will likely control the design for pullout.

The coupled hydromechanical state in soils coming from consolidation/subsidence processes and undergoing plasticity phenomena is evaluated by means of the subloading surface model. The most important feature of this theory is the abolition of the distinction between the elastic and plastic domains, as in conventional elastoplastic models. This means that plastic deformations are generated whenever there is a change in stress and a smoother elastoplastic transition is produced. The plasticity algorithm has been implemented in the PLASCON3D FE code, coupling hydromechanical fields within a saturated porous medium (locally partially saturated at the reservoir level owing to the possible presence of a gas phase) subjected to external loads and water/gas withdrawals from deep layers (aquifers/reservoirs). The three-dimensional (3D) model is first calibrated and validated against examples taken from the literature and then subsidence analyses at regional scales owing to gas extractions are developed to predict the evolution of settlements and the pore pressure in soils for long-term scenarios.

The freezing process in soils is generally recognized as a coupled problem of heat and moisture transfer. Especially in the process of transfer, it is essential to understand the status of unfrozen water. In this paper, the freezing process in unsaturated soils has been studied with the hypothesis that initially unsaturated soil maintains the apparent unsaturated condition during freezing. Accordingly, the concept of the soil-water characteristic curve (SWCC) of unfrozen soil could be also applied to soil in the freezing stage. Therefore, a methodology for estimating a correlation between the suction and the amount of unfrozen water content, named cryo-SWCC, which considers the volume change of soil pores owing to ice formation, is proposed with the assumption that any deformation in unsaturated frozen soils owing to freezing can be negligible. On the basis of a relationship between a statistical pore-size distribution and the SWCC represented by the Mualem model, the variation of characteristic parameters $a$ and $n$ in the cryo-SWCC with respect to ice content is evaluated based on an incremental formulation using optimized average unsaturated soil data values. According to the analysis results, the predicted parameters $a$ and $n$ tend to decrease as ice content increases, and the status of unfrozen water during freezing moves across each cryo-SWCC during freezing, which indicates the proposed method is able to describe variations of the pore system, matric suction, and freezing temperature depression induced by ice formation during freezing.

This paper describes the development of a numerical model for simulating the shaking table tests on wrap-faced reinforced soil retaining walls. Some of the physical model tests carried out on reinforced soil retaining walls subjected to dynamic excitation through uniaxial shaking tests are briefly discussed. Models of retaining walls are constructed in a perspex box with geotextile reinforcement using the wraparound technique with dry sand backfill and instrumented with displacement sensors, accelerometers, and soil pressure sensors. Results showed that the displacements decrease with the increase in number of reinforcement layers, whereas acceleration amplifications were not affected significantly. Numerical modeling of these shaking table tests is carried out using the Fast Lagrangian Analysis of Continua program. The numerical model is validated by comparing the results with experiments on physical models. Responses of wrap-faced walls with varying numbers of reinforcement layers are compared. Sensitivity analysis performed on the numerical models showed that the friction and dilation angle of backfill material and stiffness properties of the geotextile-soil interface are the most affecting parameters for the model response.

This paper presents several analytical solutions for the consolidation analysis of a soil layer with fairly general laws of variation of permeability and compressibility for both a single-drained condition and a double-drained condition. The solution is described in detail. A normalized time, $T$, is suggested. The average degree of consolidation exhibiting very good normalized behavior is shown. It is also revealed that the excess pore-pressure distribution of a nonhomogeneous soil layer may be quite different from the conventional Terzaghi’s solution. For the double-drained condition, some symmetrical characteristics of the average degree of consolidation and excess pore pressure are proved. Special cases of constant permeability, constant compressibility, or constant coefficient of consolidation, in which the other parameter is depth dependent, are analyzed on the basis of general analytical solutions. Results are presented in figures and tables for practical applications.

The settlements of structures resting on saturated sand deposits are examined in shaking table tests. Two series of tests are conducted to examine the effects of the duration of seismic shaking and the group effects of closely spaced foundations. The models of rigid circular foundations are founded on the surface of saturated clean fine sand, and seismically excited. The settlements of model foundations and the distributions of excess pore-water pressures induced around the model foundations are observed. In one of the test series, the settlements occurring during shaking are found to increase as the durations of shaking increase; however, the settlements occurring after the end of the shaking remain almost the same, regardless of the durations of shaking. The total settlements are, therefore, found to increase as the durations of shaking increase. In the other test series, the settlements occurring during shaking remain almost the same, regardless of the spacing. However, the settlements occurring after the end of shaking are different, where the more closely spaced foundations tend to be subjected to smaller settlements.

Zonal disintegration around a deep excavated tunnel is a characteristic of the rock mass intense fracturing phenomenon. In this paper, three-dimensional numerical tests on the failure processes of rock samples containing vertical wall semiarched tunnels are performed with the three-dimensional Realistic Failure Process Analysis (RFPA3D)–Parallel system model running on a Lenovo 1800 cluster, in order to study the configuration of zonal disintegration within the rock mass around the deep excavated tunnels. This study focuses on the zonal disintegration phenomenon in the tunnel longitudinal direction and the time-lag effect. Numerical results show that the microcracking in the disintegration zone within the model decreases gradually from the tunnel surface to the interior of the model along the tunnel longitudinal direction; the disintegration zone within the model also reduces, but the diameter of the disintegration zone becomes larger. Numerical results also indicate that an apparent time lag exists in the forming of the zonal disintegration. In other words, the zonal disintegration does not occur instantly after the excavation. Therefore, this time lag has to be considered carefully in the support design of the deep excavated tunnels.

Rolling friction plays an important role in the heap formation of granular materials. In this paper, three rolling friction models, which are incorporated into the discrete element method, are compared with simulation of sandpile formation in detail. The experiments were first implemented for cement particles to verify the reliability of simulated results. An approximate approach, which can be used to determine the friction coefficient between particles and between particles and a boundary, is also proposed. Then, numerical simulations parallel to the experiments were carried out by using the three rolling friction models. The simulated results are detailed and discussed compared with the experimental results. In addition, the rolling friction between the particles and boundaries was also studied through an example of a sphere moving on a horizontal plate with an initial translational velocity. The numerical results suggest that the rolling friction model described by the relative angular velocity can well reflect the rolling friction behavior and that the rolling friction tangential force has little effect on the simulated result.

This paper proposes an analytical approach predicting the development and progress of the plastic zone around a wellbore drilled in linear hardening or softening Drucker-Prager rocks in a theoretically consistent way. With some reasonable assumptions and simplifications on the definitions of stress invariants as well as the equivalent plastic strain, the wellbore boundary value problem is fully linearized, which leads to a simple ordinary differential equation with the radial displacement being the basic unknown. Explicit solutions for the stress and displacement fields due to wellbore drilling can thus be easily obtained. Comparisons between the analytical solution and the ABAQUS numerical simulation verify the reliability and applicability of the present closed-form solution. Extensive numerical results for the tangential and radial stress distributions are then provided to explore the influences of the friction angle and the plastic tangential modulus, for both strain hardening and softening cases. The important issue of critical mud density equivalent to wellbore radial controlled stresses that is required to maintain wellbore stability is also addressed.

Two types of models, coupled and uncoupled, are currently used to determine the extent to which it is necessary to bury subsea pipelines deeper than the maximum expected depth of ice gouges. In the uncoupled model, the soil is modeled by nonlinear Winkler springs attached to the pipe at one end, with the subgouge displacement imposed at the other end of the springs. In coupled models, the soil is modeled as a three-dimensional (3D) continuum, simultaneously capturing the processes of gouging (with associated very large deformations) and the pipeline resisting the soil displacements. This paper pinpoints the main reason for differences in predictions between the coupled and uncoupled model. It is not the coupling errors (attributable to directional coupling between Winkler springs in the axial, lateral, and vertical directions, and slice-to-slice coupling), but, rather, the superposition error, which arises in the uncoupled model by adding the soil displacements attributable to the load the pipe exerts on the soil to the subgouge deformations, despite the strongly nonlinear behavior of the soil.