Journal of Geotechnical and Geoenvironmental Engineering

The migration of contaminants through a 10-mm-diameter hole (0.785 cm²) in a geomembrane in direct contact with a geosynthetic clay liner (GCL) and adjacent silty sand is examined. Experiments were conducted in four 0.6-m-diameter cells at a vertical stress of 100 kPa and hydraulic head differences of 0.3 and 1 m. The system was first permeated with distilled water until steady-state flow was attained (the reference case). After 280 days the permeant was changed to a NaCL solution. After 800 days of permeation with 0.14 M NaCl solution there was only a 3% increase in the flow (leakage) compared with the reference case despite up to almost an order of magnitude increase in GCL permeability near the hole. The wetted radius at the end of the experiments was inferred by injection of dye and was found to be about 0.1–0.15 m. This provides the first experimental evidence in support of theoretical predictions that, when the geomembrane is in direct contact with a GCL, leakage through a hole is primarily controlled by the interface transmissivity rather than the GCL hydraulic conductivity when there is interaction between the permeant and the GCL. The observed chloride distribution in the silty sand at the end of the experiments is reported.

Settlement of soil layers during and after earthquake shaking is a major cause of damage to buildings and geotechnical structures. The available empirical design methods to consider seismically induced settlement focus on sands in dry or water-saturated conditions, and there is currently a gap in the basic understanding of the mechanisms of seismically induced settlements of partially saturated sands. An effective stress-based empirical methodology is proposed to estimate the seismically induced settlement of a free-field layer of sand in partially saturated conditions. This approach estimates the settlement by separately considering the volumetric strains caused by compression of void space during strong shaking (seismic compression) and dissipation of excess pore water pressures generated during earthquake shaking (postcyclic reconsolidation). A parametric evaluation of the methodology indicates that the small strain shear modulus, the parameters of the modulus reduction curve, the approach to estimate the upper bound on volumetric strain during liquefaction, and the pore water pressure generation parameter can have significant impacts on the predicted settlement. The model predictions were validated using results from a newly developed centrifuge physical modeling system that involved the use of steady-state infiltration to maintain a uniform degree of saturation with depth in the sand layer. Both the model and experimental results show a nonlinear trend in surface settlement with degree of saturation, with a minimum value obtained for sand at a degree of saturation between 0.3 and 0.6.

Improvement of soft clay deposits by the installation of stone columns is one of the most popular techniques followed worldwide. The stone columns not only act as reinforcing material increasing the overall strength and stiffness of the compressible soft soil, but they also promote consolidation through effective drainage. The analytical and numerical solutions available for ascertaining the response of column-reinforced soil have been developed on the basis of the equal strain hypothesis. For typical surcharge (embankment) loading, the free strain analysis appears to give more realistic results comparable to field data. The paper presents a novel numerical model (finite-difference method) to analyze the response of stone column–reinforced soft soil under embankment loading, adopting the free strain approach and considering both arching and clogging effects. Apart from predicting the dissipation of excess pore water pressure and the resulting consolidation settlement with time, the load transfer mechanism and the extent of ground improvement are some of the salient features captured by the proposed model. The proposed model is validated by comparing with existing models and field data, which indicate the suitability and accuracy of the solutions. The proposed model is also applied successfully to selected case studies.

A Bayesian framework using field observations for back-analysis and updating of soil parameters in a multistage braced excavation is presented. Because of the uncertainties originating from the poorly known soil parameters, the imperfect analysis model, and other factors such as construction variability, the soil parameters can only be inferred as probability distributions. In this paper, these posterior distributions are derived using the Markov chain Monte Carlo (MCMC) sampling method implemented with the Metropolis-Hastings algorithm. In the proposed framework, Bayesian updating is first realized with one type of response observation (maximum wall deflection or maximum settlement), and then this Bayesian framework is extended to allow for simultaneous use of two types of response observations in the updating. The proposed framework is illustrated with a quality excavation case and shown effective regardless of the prior knowledge of soil parameters and type of response observations adopted.

Shear-wave velocity ($V_s$) offers a means to determine the seismic resistance of soil to liquefaction by a fundamental soil property. This paper presents the results of an 11-year international project to gather new $V_s$ site data and develop probabilistic correlations for seismic soil liquefaction occurrence. Toward that objective, shear-wave velocity test sites were identified, and measurements made for 301 new liquefaction field case histories in China, Japan, Taiwan, Greece, and the United States over a decade. The majority of these new case histories reoccupy those previously investigated by penetration testing. These new data are combined with previously published case histories to build a global catalog of 422 case histories of $V_s$ liquefaction performance. Bayesian regression and structural reliability methods facilitate a probabilistic treatment of the $V_s$ catalog for performance-based engineering applications. Where possible, uncertainties of the variables comprising both the seismic demand and the soil capacity were estimated and included in the analysis, resulting in greatly reduced overall model uncertainty relative to previous studies. The presented data set and probabilistic analysis also help resolve the ancillary issues of adjustment for soil fines content and magnitude scaling factors.

This paper examines the influence of the initial state of sands on the potential for undrained instability. The main goal is to illustrate how advanced constitutive modeling of sand behavior can be used to evaluate the susceptibility for static liquefaction. The methodology is based on the concept of latent instability, in which the potential for collapse is contingent on particular boundary conditions. A generalized effective stress soil model, MIT-S1, is used to support the analysis and is combined with a theoretical approach for identifying loss of control owing to undrained shear perturbations. The theory is evaluated using experimental evidence available for Toyoura sand to point out the key role of void ratio and consolidation history and to provide experimental validation for the theory. Model predictions are then used to disclose the subtle role of drained preloading paths in promoting liquefaction instabilities. The ability of the constitutive model to reproduce the interplay between undrained kinematic constraints and material failure is fundamental in predicting potential instabilities arising from changes in drainage conditions. The examples shed light on the mechanics of static liquefaction and set a framework for applying the principles of material stability to the triggering analysis of flow slides induced by undrained shear perturbations.

Earthquake-induced deformation of piled bridge abutments in approach embankments underlain by liquefied soils may be reduced by the restraining forces provided by the piles and bridge superstructure. Design methods accounting for these pinning effects have been applied in practice but lack significant comparison with actual physical data. Three dynamic centrifuge model tests were performed to investigate pile-pinning effects, and the results are presented in this paper. Each centrifuge model was composed of two identical embankments underlain by liquefiable soil: one with a pile group and the other without. The tests demonstrated that the embankment with a pile group experienced less lateral movement and settlement than the embankment without a pile group. To better understand and extend the results of these centrifuge tests, nonlinear deformation analyses of the three centrifuge tests were made using a fully coupled effective stress model. The ability of the analyses to model the centrifuge tests is assessed primarily through comparison of dynamic histories of acceleration, pore water pressure, displacement, and pile bending moments. The usefulness of the modeling and dynamic analysis procedures is evaluated, and key lessons are described.

Based on an advanced micromechanics-based nonlinear creep model for shotcrete and thin shell kinematics, measured displacement data are first converted into in-plane stress fields throughout a shotcrete tunnel shell driven according to the New Austrian Tunneling Method (NATM). Subsequently, the partial differential equations for the local force equilibrium in the cylindrical coordinate directions are solved for out-of-plane shell stresses, inclusive of the tractions at the ground-shell interface. Results obtained for an NATM-tunnel driven through clayey ground at moderate depth show that the maximum shear tractions at the ground-shell interface may even exceed the maximum normal tractions (ground pressure). At the same time, even a young top heading tunnel shell may act as an arching thrust. The authors regard this method as a further step in the continously refined of interpretation of displacement measurements stemming from tunnel monitoring systems for the NATM, on a well-defined mechanical basis. This results in two structural models for top heading excavation stages, which might have the potential to support NATM predesign activities in the future.

Because low-compacted trench backfill around a manhole is normally liquefiable, the manhole could suffer uplift damage associated with soil liquefaction during a strong earthquake. In this study, 22 dynamic centrifuge models were tested to investigate the response of a buried manhole subjected to a dynamic load. The models were tested under 20g, and a substitute pore fluid was used to avoid the scaling law conflict between the dynamic and diffusion processes. It was found that excess pore water pressure is one of the contributing factors to the magnitude of the manhole uplift. Using this result, new mitigation methods against the uplift in liquefied ground were developed. Their effectiveness was also examined through the tests. A model manhole mitigated with the proposed methods was tested alongside regular model manhole. From the test results, the magnitude of manhole uplifts with the mitigation methods decreased as buildup of the excess pore water pressure was restrained in high-compacted backfill or excess pore water was dissipated into the manhole during strong shaking.

In this paper, experimental studies were conducted using a true triaxial apparatus with a bender element system to examine the mechanisms of aging-induced, small-strain shear modulus changes in sand samples under isotropic and anisotropic loading. Numerical simulations based on the discrete element method (DEM) were also carried out in parallel. In the isotropic loading cases, the three measured shear moduli, $G_{xy}$, $G_{yz}$, and $G_{zx}$, and associated aging rates, in terms of the modulus changes, are similar in every loading stage. DEM simulations reproduced the experimental findings and suggested a general trend. A sample with a lower shear modulus before aging, because of a greater percentage of weak forces, allows more forces to be redistributed from the strong force network to the weak force network through the process of contact force homogenization during aging and therefore can have a higher aging rate. In the anisotropic loading cases where ${\sigma }_z>{\sigma }_x={\sigma }_y$, the measured modulus increase (i.e., the aging rate) is greater in $G_{yz}$ (or $G_{zx}$) than in $G_{xy}$. This behavior can be attributed to the increase in both the strong and weak forces in the z-direction during aging, because of arching breakdowns.

This paper presents results of a comprehensive investigation on the utilization of carpet waste fibers in reinforcement of clay soils. Effects of adding proportionate quantities of two different types of shredded carpet waste fibers to clay soils (i.e., 1, 3, and 5% by dry weight of the soil) were investigated and evaluated. The investigation was conducted on specimens prepared at their maximum dry unit weight and optimum moisture content, as well on specimens prepared at variable conditions of dry unit weight and moisture content. A comparison was also made on specimens prepared at the same fiber content by changing dry unit weight while moisture content was kept unchanged or by changing both dry unit weight and moisture content. The investigation revealed that inclusion of carpet waste fibers into clay soils prepared at the same dry unit weight can significantly enhance the unconfined compression strength (UCS), reduce postpeak strength loss, and change the failure behavior from brittle to ductile. The results also showed that the relative benefit of fibers to increase the UCS of the clay soils is highly dependent on initial dry unit weight and moisture content of the soil. Failure patterns were gradually transformed from the apparent classical failure for unreinforced soil specimens to barrel-shaped failures for reinforced specimens at 5% fiber content.

The cone tip resistance (${\mathit{q}}_{\mathit{c}}$) from cone penetration tests (CPTs) is widely used to determine the bearing capacity of piles. Although it is widely known that soil properties are spatially correlated, the spatial correlation of ${\mathit{q}}_{\mathit{c}}$ is not considered in the current methods for predicting the bearing capacity of piles. In this paper, a probabilistic approach is presented for predicting the bearing capacity of driven piles in clay by considering the spatial correlation between ${\mathit{q}}_{\mathit{csVn}}$ (the spatial average of ${\mathit{q}}_{\mathit{c}}$ of the bottom soil layer within the pile length) and ${\mathit{q}}_{\mathit{cbV}}$ (the spatial average of ${\mathit{q}}_{\mathit{c}}$ over an interval near the pile base). Parametric studies are conducted to evaluate the effect of the spatial correlation between $q_{csVn}$ and $q_{cbV}$ on the bearing capacity of piles. The results indicate that it is important to consider the spatial correlation between ${\mathit{q}}_{\mathit{csVn}}$ and ${\mathit{q}}_{\mathit{cbV}}$ in the probabilistic prediction of the bearing capacities of piles. Ignoring the spatial correlation between ${\mathit{q}}_{\mathit{csVn}}$ and ${\mathit{q}}_{\mathit{cbV}}$ will underestimate the probability of failure and lead to unsafe design. Finally, 14 field test piles are analyzed with the presented approach to demonstrate the probabilistic prediction of the bearing capacity of piles by considering the spatial correlation of the CPT data.

This paper deals with the design of rigid piles under lateral load. In many practical situations pile displacements may be the limiting factor in the design; thus, an allowable displacement analysis instead of an allowable stress analysis may be more appropriate in the foundation design. A nonlinear elastoplastic closed-form solution for the evaluation of the response of rigid piles under lateral load is presented. The derived equations allow determining the lateral displacement and load distribution for a given combination of the applied forces, $H_0$ and $M_0$, at the pile head. The approach can be applied either for cohesive or cohesionless soils. For greater generality, the solution has been given in nondimensional form. A comparison with the Broms solution for a free rigid pile in cohesionless soil is also presented. Comparisons with experimental results show that the proposed approach may fit in a reliable way the nonlinear load-displacement response of a free rigid pile.

Laboratory tests to examine hysteresis in the hydrologic and mechanical properties of partially saturated soils were conducted on six intact specimens collected from a landslide-prone area of Alameda County, California. The results reveal that the pore-size distribution parameter remains statistically unchanged between the wetting and drying paths; however, the wetting or drying state has a pronounced influence on the water-entry pressure, the water-filled porosity at zero suction, and the saturated hydraulic conductivity. The suction stress values obtained from the shear-strength tests under both natural moisture and resaturated conditions were mostly bounded by the suction stress characteristic curves (SSCCs) obtained from the hydrologic tests. This finding experimentally confirms that the soil-water retention curve, hydraulic conductivity function, and SSCC are intrinsically related.

A series of pullout tests is presented in this paper and is used to identify the kinematics of failure and the uplift response of circular anchor plates embedded in sand-cement stabilized layers at distinct normalized embedment depths ($\mathit{H}/\mathit{D}$), where $\mathit{H}$ is the thickness of the treated layer and $\mathit{D}$ is the diameter of the anchor plates. Experimental results show that the uplift capacity of anchor plates embedded in sand backfill layers increases considerably after mixing 3% cement with the backfill material. Distinct failure mechanisms observed for anchor plates embedded in both sand and cement-stabilized backfills are shown to be a function of $\mathit{H}/\mathit{D}$. The addition of cement to the sand backfill leads to an increase in uplift capacity of 9 times for an $\mathit{H}/\mathit{D}$ ratio of 1.0 and of 13 times for an $\mathit{H}/\mathit{D}$ ratio of 2.0. For sand backfill with $\mathit{H}/\mathit{D}=1.0$, the failure surface had a truncated cone shape with a vertical inclination of 22°, whereas for $\mathit{H}/\mathit{D}$ of 1.5 and 2.0, radial cracking was observed, and final failure surfaces had inclinations of 26 and 30°, respectively. Pullout of anchor plates in cement-stabilized backfills at $\mathit{H}/\mathit{D}$ ratios ranging from 1.0 to 2.0 exhibit two distinct characteristics: (a) a linear elastic deformation response at small pullout displacements and (b) a later stage where radial fracturing of the stabilized backfill leads to hardening just prior to failure. Radial cracks starting at the top of the layer near the center of the anchor plates start to propagate only at about 90% of the final uplift failure load, irrespective of $\mathit{H}/\mathit{D}$.

Although many studies have been done to investigate the axial behaviors of open-ended piles in sands, few studies have been reported for weak clayey silts. To develop reliable models for the design of open-ended steel-pipe piles driven into 29-m-thick varved clayey silt deposits, a series of full-scale field load tests including large-strain dynamic tests and static cyclic axial-compression-load tests was conducted on two groups of instrumented piles. Through analysis of the test data, soil parameters were back-calculated for estimation of pile capacities using the static-bearing-capacity formulas and cone-resistance-based methods. The comparisons between the calculated results and the field load test data demonstrated that the following considerations can be adopted in the design of static compression capacities of an open-ended pipe pile penetrating through thick varved clayey silts to end-bearing in dense cohesionless soils: (1) a fully plugged condition can be assumed, (2) cone resistance with an upper limit of 4,788 kPa (100 ksf) can be used for unit base resistance on the soil plug, and (3) exterior unit shaft resistance can be estimated using two-thirds of the total unit shaft resistance.