Journal of Geotechnical and Geoenvironmental Engineering

The paper presents a description of the methods used to model rock as discontinuous media. The objective of the work is to bring to the geomechanics community recent advances in numerical modeling in the field of rock mechanics. The following methods are included: (1) the distinct element method; (2) the discontinuous deformation analysis method; and (3) the bonded particle method. A brief description of the fundamental algorithms that apply to each method is included, as well as a simple case to illustrate their use.

The results of 167 full-scale field load tests were used to examine several issues related to the load-displacement behavior of footings in cohesionless soils under axial compression loading, including (1) method to interpret the “failure load” from the load-settlement curves; (2) correlations among interpreted loads and settlements; and (3) generalized load-settlement behavior. The $L_1\text{-}{L}_2$ method was found to be more appropriate than the “tangent intersection” and “10% of the footing width” methods for interpreting the failure load. The interpreted loads and displacements indicate that footing load-settlement behavior is less elastic and more nonlinear than that of drilled foundations. The results show that the footing behavior will be beyond the elastic limit for designs where a traditional factor of safety between 2 and 3 is used. A normalized curve was developed by approximating the load-settlement curve for each load test in the database by hyperbolic fitting, and the uncertainty in this curve was quantified. This normalized curve can be used in footing design that considers capacity and settlement together. Where possible or warranted, the normalized curve can be subdivided as a function of initial soil modulus.

An extensive database of full-scale field load tests was used to examine the bearing capacity for footings in cohesionless soils. Each load test curve was evaluated consistently to determine the interpreted failure load (i.e., bearing capacity) using the $L_1\text{-}{L}_2$ method. This test value then was compared with the theoretical bearing capacity, computed primarily using the basic Vesić model. The comparisons show that, for footing widths $B>1\text{m}$ , the field results agree very well with the Vesić predictions. However, for $B<1\text{m}$ , the results indicated a relationship between $B$ and the predicted-to-measured bearing capacity ratio. Accordingly, a simple modification was made to the bearing capacity equation, and the resulting predictions are very good.

The development of strain in two 31 and 56 m long instrumented postdriving grouted cylinder piles at a site west of Busan, South Korea, were monitored during 200 days after construction, whereupon a static loading test was performed. Initial strain measurements showed unexpected elongation of the pile, probably due to swelling from absorption of water but as the soil reconsolidated, the elongation changed into shortening, probably due to imposed residual load in the pile. The resulting compression of the pile eventually offset the swelling of the pile. To investigate the cause of the strain changes more closely and enhance the evaluation of the field data, two short pile pieces were prepared and placed free-standing above ground in an outside laboratory. One piece was from a cylinder pile of which central void was grouted and one was made up by grouting inside a temporary casing. The monitoring showed that the short pieces appeared first to shorten and then to elongate due to the heating and cooling from the hydration process. When strains and temperature had stabilized 150 days after start of the study, both pieces were submerged to introduce swelling of the concrete. For the first 100 days after submerging, the swelling strains in both short pieces amounted to $100\mu \epsilon$ . Seven hundred days after submersion, the total swelling strains were 150 and $250\mu \epsilon$ .

This study discusses the new earth retention system with prestressed wales. The prestressed wale provides a high flexural stiffness that resists bending caused by lateral earth pressures. The prestressed wale system consists of wales, steel wires, H-beam support legs, and hydraulic jacks. The system allows for a wider spacing between supports and is more economically efficient than conventional systems. In addition, the construction process is relatively simple and the system demonstrates excellent performance. This paper explains the basic principles, mechanisms, and design procedure, as well as the advantages and limitations of this new earth retention system. It also presents the results for the performance of the field application and the measured performance is investigated and discussed. These results demonstrated that this new earth retention system with prestressed wales performed successfully in an urban excavation.

In many tunneling and excavation projects, free-field vertical ground movements have been used to predict subsidence, and empirical limits have been employed to evaluate risk. Validity of such approaches is largely unknown given that ground movements are in fact not one-dimensional and that adjacent applied loads are known to have an impact. This paper employed analytical and large-scale experimental efforts to quantify these issues, in the case of excavation adjacent to a reinforced concrete frame with tieback anchors and a sheetpile wall in dry sand. With this flexible system, a disproportionate amount of the soil and building movements occurred prior to installation of the first tieback, even when conservative construction practices were applied. Furthermore, free-field data generated a trough as little as one-half the size of that recorded near the building frames. Empirically based relative gradient limits generally matched the extent and distribution of the damage, while the application of various structural limits did not fully identify local damage distribution but did generally reflect global response. The use of fully free-field data or a failure to include lateral soil displacements both underpredicted building displacements by as much as 50% for low-rise concrete frames without grade beams on sand.

Volume changes due to wetting may occur in naturally deposited soils as well as earthen construction (e.g., compacted fills or embankments). Depending on the stress level, some soils exhibit increase in volume upon wetting (swell) while others may exhibit decrease in volume upon wetting (collapse). The work described in this paper focused on wetting-induced volume changes in compacted soils. Motivation for this work stemmed from observations of earthen structures that exhibit problematic behavior under wetting conditions, even though soils were compacted to engineering specifications (i.e., at or above minimum density and within moisture content ranges). Not only is this problematic behavior a concern but also the laboratory tests used to predict settlement of constructed facilities may not properly model the actual behavior of soil compacted under field conditions. For example, settlements experienced by compacted fills may be different from settlement predictions based on one-dimensional oedometer tests. These differences are partly related to the variations in the soil structure in tested specimens that arise because soil clods compacted in the laboratory are smaller than soil clods compacted in the field. The term “soil structure” includes the combined effects of soil fabric and interparticle forces. “Fabric” generally refers to the geometric arrangement of particles, whereas interparticle forces include physical and physicochemical interactions between particles. The soil structure in this case is associated with specimen preparation methods and is influenced by several factors including soil composition (including pore water chemistry), compaction method, clod sizes, initial moisture condition of clods, dry density or void ratio, and compaction moisture content. A laboratory research study was conducted to investigate the influence of variations in clod-size and structure on one-dimensional volume change, with emphasis on wetting-induced volume change, for nine different fine-grained soils. The results of the study suggest that the influence of structure in one-dimensional oedometer tests depends on soil type and nature of the clods in the compacted soil. Clayey soils appear to be influenced more by differences in structure, whereas silts or clayey sands of low plasticity $\left(PI<10\right)$ do not appear to suffer as much from structure effects in one-dimensional oedometer tests. This is attributed to more extensive clod development in clayey soils. Furthermore, the moisture condition of clods appears to have an important influence on volume change behavior.

The paper discusses finite element models for predicting the elastic settlement of a strip footing on a variable soil. The paper then goes on to compare results obtained in a probabilistic settlement analysis using a stochastic finite element method based on first order second moment approximations, with the random finite element method based on generation of random fields combined with Monte Carlo simulations. The paper highlights the deficiencies of probabilistic methods that are unable to properly account for spatial correlation.

This paper provides an assessment of tactile pressure sensors for geotechnical applications. A tactile pressure sensor is an array of small sensing units, called sensels, embedded in a polymeric sheet or pad that measures the magnitude and distribution of stresses normal to the sheet surface. Methods for minimizing the effects of shear on sensor measurements are discussed and the efficacy of these methods are demonstrated by laboratory experiments. The time-dependent characteristics of the sensors are evaluated and recommendations are provided for measurements that account for time-dependent effects. Tactile pressure sensor measurements in response to vertical loading and unloading and to lateral loads on full-scale pipelines affected by large horizontal ground movements are compared with independent measurements of the loads. Sensor measurements are used to show the distribution of normal stress on pipelines subject to large lateral soil movement.

Difficulty in predicting the transfer of load from a structural element to the surrounding soil has limited the reliability of geotechnical design and performance. The remaining uncertainty in load transfer mechanics is primarily due to the localized nature of the mechanism. This study examines localized soil-structure interaction through a series of monotonic direct interface shear tests. Parameters investigated include relative density, particle angularity, particle hardness, surface roughness, normal stress, and normal stiffness. The soil-structure interface behavior is quantified in terms of the local two-dimensional displacement and strain distributions within the test specimens using particle image velocimetry. In addition, the localized zone of soil adjacent to the structural surface within which shear and volumetric strains occur is quantified. The relative density of the soil, and the relationship between particle characteristics (angularity and hardness) and surface roughness are shown to have the greatest effect on local interface behavior, followed by confining stress and stiffness conditions.

Numerical experiments were conducted to understand the effect of geometric and transport characteristics of a geomembrane-geosynthetic clay liner (GM/GCL) composite liner on gas leakage rate through a circular defect in the geomembrane (GM). The originality of the approach proposed in this paper rests on the use of a new conceptual two-layered system for modeling of GM/GCL composite liners where the interface zone between the GM and geosynthetic clay liner (GCL) has been merged with the GCL cover geotextile and handled as one layer; the GCL bentonite layer was considered the second layer. The role of the carrier geotextile layer was ignored since it can be considered as a no pressure loss layer. Analysis of numerical simulation results shows the existence of a constitutive leakage flow surface which enables evaluation of the leakage flow state for different geometric and transport properties of GM/GCL composite liners. Furthermore, the determined surface was also exploited to evaluate gas leakage rates under the framework of the Forchheimer’s analytical solution. The gas leakage rate predictions were found to be in good agreement with experimental results obtained at different GCL moisture content.

Finite-element (FE) simulations are increasingly providing a versatile environment for conducting lateral ground deformation studies. In this environment, mitigation strategies may be assessed in order to achieve economical and effective solutions. On the basis of a systematic parametric study, three-dimensional FE simulations are conducted to evaluate mitigation by the stone column (SC) and the pile-pinning approaches. Mildly sloping saturated cohesionless strata are investigated under the action of an applied earthquake excitation. For that purpose, the open-source computational platform OpenSees is employed, through a robust user interface that simplifies the effort-intensive pre- and postprocessing phases. The extent of deployed remediation and effect of the installed SC permeability are investigated. The influence of mesh resolution is also addressed. Generally, SC remediation was found to be effective in reducing the sand stratum lateral deformation. For a similar stratum with permeability in the silt range, SC remediation was highly ineffective. In contrast, pile pinning appeared to be equally effective for the sand and silt strata permeability scenarios. Overall, the conducted study highlights the potential of computations for providing insights toward the process of defining a reliable remediation solution.

Data from over 30 sites in 5 countries are analyzed to develop updated factors for correcting liquefaction resistance for aged sand deposits. Results of cyclic laboratory tests on relatively undisturbed and reconstituted specimens suggest an increase in the correction factors of 0.12 per log cycle of time and an average reference age of 2 days for the reconstitute specimens. Laboratory and field test results combined with cyclic resistance ratio (CRR) charts suggest an increase in the correction factors of 0.13 per log cycle of time and an average reference age of 23 years. A reference age of 23 years seems appropriate for the commonly used CRR charts derived from field liquefaction and no liquefaction case history data. Because age of natural deposits is often difficult to accurately determine, a relationship between measured to estimated shear-wave velocity ratio (MEVR) and liquefaction resistance correction factor is also derived directly from the compiled data. This new MEVR-liquefaction resistance correction factor relationship is not as sensitive to MEVR as in the relationship derived indirectly in a previous paper.

Recent criteria have been developed to describe the onset of static liquefaction in constitutive models. This paper expands the theory to a finite-element framework in order to predict potentially unstable regions in granular soils at the engineering scale. Example simulations are presented for two plane strain tests and a submarine slope to demonstrate the applicability of a proposed liquefaction criterion to boundary value problems. In addition, loading rate and mesh size effects on the liquefaction prediction are examined. The methodology presented herein shows promise as a means of predicting soil liquefaction based on solid mechanical theory rather than empiricism.

Mitigation of liquefaction potential in loose granular soil can theoretically be achieved through permeation and subsequent gelation of dilute colloidal silica stabilizer. However, practical application of this technique requires efficient and adequate delivery of the stabilizer to the liquefiable soil prior to gelation. The purpose of this research was to evaluate colloidal silica transport mechanisms and to determine if an adequate concentration can be delivered to a soil column prior to gelation. The laboratory work consisted of grouting 15 short (0.9 m) columns tests packed with Nevada No. 120, Ottawa 20/30, or graded silty sand to identify the variables that influence stabilizer transport through porous media. It was found that colloidal silica can be successfully delivered through 0.9-m columns packed with loose sand efficiently and in an adequate concentration to mitigate the liquefaction potential. Transport occurs primarily by advection with limited hydrodynamic dispersion, so Darcy’s law can be used to predict flow. The Kozeny-Carmen equation can be used to include the effect of increasing viscosity on transport by incorporating the power law mixing rule of Todd. The primary variables influencing stabilizer transport were found to be the viscosity of the colloidal silica stabilizer, the hydraulic gradient, and the hydraulic conductivity of the liquefiable soil.

Pipe bursting is a construction technique that involves the replacement of an existing buried pipe with potentially much less surface disturbance than traditional cut and cover construction. However, excessive ground movements associated with pipe-bursting operations may lead to damage to surrounding infrastructure. A static pipe-bursting experiment was performed in sand and gravel within an 8-m-long, 8-m-wide, and 3-m-deep test pit to quantify the ground displacements from pipe bursting. An existing unreinforced concrete pipe buried 1.385 m below the ground surface was replaced with a high-density polyethylene pipe. Pulling force and the three-dimensional nature of surface displacements associated with pipe bursting are examined. The 4-m wide surface response had a peak vertical displacement of 6 mm. In addition, transverse displacements of 1.2 mm resulted in the formation of a tension crack in the ground above the concrete pipe. This experiment offers data that improves the understanding of the mechanisms of ground disturbance, and provides unique experimental data for calibration of numerical models.

Two 381 mm (15 in. nominal) diameter fiber reinforced cement pipes have been tested under embankment loading conditions to study pipe response in both low stiffness, fine grained backfill, and a high stiffness graded granular backfill. Pipe deformations and strains were measured and interpreted to provide insight into the effect of soil backfill on the deformations and moments that develop. Not surprisingly, the use of silty clay backfill resulted in greater pipe deflections while the stiffer granular backfill lead to greater load transfer to the surrounding ground. Calculations using elastic soil-pipe interaction theory were effective in estimating the observed changes in pipe diameter at typical service loads (overburden pressures of 100 kPa, i.e., 14.4 psi in the lower stiffness backfill and 200 kPa, i.e., 28.8 psi in the high stiffness backfill). Measured strain distributions show that the fiber reinforced pipe exhibited ovaling response similar to that seen for flexible and semiflexible pipes. As expected, tensile strains were observed on the outer surface at the springlines and the inner surface at the crown. Strains observed at the haunch were negligible, indicating that the bending moments within the pipe have conventional “hourglass” distribution, with negligible moments at shoulders and haunches. Differences in strain measured at the inner and outer surfaces were used with the elastic pipe modulus to calculate the experimental bending moments. Comparisons of those experimental bending moments with the bending moment calculated for a rigid pipe indicate that these FRC pipe structures are semirigid so that moments are reduced as a result of support provided by the surrounding soil. A design expression for moment arching factor (MAF or moment divided by the rigid pipe moments) developed in an earlier paper was found to provide reasonable estimates for the experimental moment values. Moment estimated using the design soil moduli of McGrath and MAF provide moment values that are reasonable and conservative relative to those that were observed.

The measurement of track stiffness, or track modulus, is an important parameter for assessing the condition of a railway track. This paper describes a method by which the dynamic track modulus can be determined from the dynamic displacements of the track during normal train service, measured using geophones. Two techniques are described for calculating the track modulus—the inferred displacement basin test (DBT) method and a modified beam on an elastic foundation (BOEF) method. Results indicate that the viscoelastic response of the soil will influence the value of track modulus determined using the DBT method. The BOEF method was therefore used to calculate the apparent increase in axle load due to train speed. Hanging or partly supported sleepers were associated with a relatively small increase in dynamic axle loads with train speed.

An experimental testing program was undertaken to investigate failure mechanisms induced by the active movement of a deep rectangular trapdoor underlying a granular soil. Reduced-scale models were tested under normal gravity as well as under an increased gravitational field using a centrifuge facility. Some models were used to evaluate the performance of both flexible and rigid pipes undergoing a localized loss of support. Failure mechanisms in the longitudinal direction of the models were characterized by a single, well-defined failure surface that developed within the limits of the trapdoor. However, failure mechanisms in the transverse direction of the models were characterized by multiple failure surfaces extending outside the limits of the trapdoor. Significant dilation of the soil located immediately above the trapdoor was identified in the failure of the models. The pattern of the failure mechanisms was found to be affected by the stress level and backfill density. Higher stress levels were found to lead to well-developed failure zones. The influence of backfill density was found to be more relevant in models involving flexible pipes. Pipes embedded within loose backfill were severely damaged after loss of support, while pipes embedded in dense backfill experienced negligible deformations. These results indicate that damage to pipelines caused by ground loss of support can be significantly minimized by controlling the compaction of the fill.

Underground mine caving may induce collapse of the ground during ore caving, particularly when there is a loose material on top of the mine. In the present paper, a laboratory test is performed to model this mining problem which is followed by a particle flow simulation to simulate the interaction and movement between the two materials. For this nonhomogeneous problem, an interesting columniform channel for the loose material is formed during the drawing in both the experiment tests and the numerical modeling. The formation of this channel greatly affects the efficiency of ore drawing and the stability of the mine caving. The numerical modeling can be used for the optimization of ore drawing, ore drawing management and the prediction of ore loss and dilution indexes. The importance of the moisture content on ore drawing process and the efficiency of ore drawing are also investigated by both experimental and numerical tests, and useful results are found from the present study.

Although the cone penetration test (CPT) and flat-plat dilatometer test (DMT) have been used for over 30 years, relatively little has been published regarding comprehensive correlations between the two in situ tests. This paper presents preliminary correlations between the main parameters of the CPT and DMT. The key to the proposed correlations is the recognition that the main DMT parameters are normalized and hence, should be correlated with normalized CPT parameters. The suggested correlations are developed and evaluated using published records and existing links to various other parameters as well as comparison profiles. The suggested correlations may guide future more detailed correlations between these two in situ tests.

The technique presented deals with the assessment, based on drained test behavior and formulation, of the undrained postcyclic stress-strain behavior of sands under limited or complete (full) liquefaction and its associated strength. At present, there is no particular procedure that allows assessment of such undrained postcyclic behavior that could develop full (pore-water pressure ratio, $r_u=1$ ) or limited $\left(r_u<1\right)$ liquefaction. The prediction of the undrained postliquefaction (full or limited liquefaction) response presented here is based on basic properties of sand such as its relative density $\left(D{r}_c\right)$ [or ${\left(N_1\right)}_{60}$ blowcount], the effective angle of internal friction $\left(\phi \right)$ , the roundness of the sand grains $\left(\rho \right)$ , and the drained axial strain at 50% stress level $\left({\epsilon }_{50}\right)$ . The technique presented accounts for the excess pore-water pressure induced by cyclic loading $\left(\Delta u_c\right)$ and the postcyclic excess pore-water pressure generated under undrained monotonic loading $\left(\Delta u_d\right)$ .