Journal of Geotechnical and Geoenvironmental Engineering

In recent years, several empirical and semiempirical relationships have been proposed to predict the small-strain shear modulus of unsaturated fine-grained soils along different hydraulic and mechanical loadings paths. However, a major deficiency of these relationships is the absence of a coupled linkage between hydraulic and mechanical processes that occur in unsaturated conditions. Specifically, the void ratio and effective stress are considered uncoupled, and changes in soil volume are rarely considered when implementing soil water retention curves in these equations. This study aims to address these deficiencies by discussing the coupled effect of hydraulic and mechanical processes in unsaturated soils and presenting a semiempirical model to predict the small-strain shear modulus, $G_{\max }$, of unsaturated low plasticity soils subjected to volume and effective stress changes along different mechanical and hydraulic stress paths. Predictions from this model and three other recently proposed models in the literature are compared with experimental results obtained from a series of suction-controlled bender element tests on silty soil specimens to validate the proposed model. The comparison reveals that the model proposed in this study provides more consistent predictions of the small-strain shear modulus during hydraulic hysteresis, as well as different paths of loading and unloading.

Multiple barrier system has become a widely adopted method to mitigate debris flows, which consists of smaller barriers that progressively retain the debris to reduce the acceleration and carbon footprint. Different from single barriers, the impact force on multiple barriers depends on the overflow from the upstream barrier. Overflow may substantially accelerate the flow by converting potential energy into kinetic energy with limited energy dissipation. While installing a deflector at the crest of a single barrier suppresses overflow, the efficacy of a deflector in reducing impact force in a multiple barrier system is not well understood yet. In this study, physical tests were conducted using a 28-m-long flume to investigate the impact dynamics of debris flow against dual rigid barriers with a deflector installed at the first barrier. A smoothed particle hydrodynamic (SPH) model was calibrated by back analyzing the flume tests. A numerical parametric study was then conducted to investigate the efficiency of a deflector in reducing impact force from different volumes of debris flow against dual barriers. Newly modified equations are proposed that can reasonably predict the overflow velocity, the velocity after landing, and landing distance considering a deflector. Physical and numerical results reveal that a deflector at the first barrier can efficiently reduce overflow by redirecting the flow to roll back and reduce the landing distance by up to 75%. A deflector at the first barrier can reduce the impact force at the second barrier by up to 90%.

In a conventional Newmark slope model, a soil block is assumed to slide along a rigid plane in a perfectly plastic mode when an input motion exceeds yield acceleration. In reality, a small deformation may well be expected to occur immediately before sliding due to shear deformation in a thin layer underneath the sliding block. In this research, a spring-supported Newmark model (SSNM) has been developed by adding a linear spring for the presliding yield displacement $u_0$ to the slider of the conventional Newmark model (CNM), demonstrating the reproducibility of key behavior of slope sliding in model tests. Dynamic responses of slopes to harmonic and earthquake motions have been calculated to find that in the SSNM the acceleration for slide initiation tends to overshoot the yield acceleration of the CNM by a larger margin due to a tiny yield displacement $u_0$ of millimeters. The overshoot is also found to become larger with increasing input frequency, which has also been observed in model shaking table tests. According to example calculations on a typical slope with realistic design parameters shaken by recorded strong earthquake motions, the number of slide repetitions and the associated cumulative slope displacements are calculated much smaller than in the CNM due to the overshooting of yield acceleration. Thus, the newly developed SSNM can be more realistic than the conventional method in evaluating seismic slope behavior by providing a small yield deformation $u_0$ as a specific slope-dependent parameter.

The prediction of time-dependent deformations of embankments constructed on soft soils is essential for preloading or surcharge design. The predictions can be obtained by Bayesian back analysis methods progressively based on measurements so that practical decisions can be made after each monitoring round. However, the effect of creep is typically ignored in previous settlement predictions based on Bayesian back analysis to avoid the heavy computational costs. This study aims to fill this gap by combining the Bayesian back analysis with a decoupled consolidation constitutive model, which accounts for creep to perform long-term settlement predictions of the trial embankment with prefabricated vertical drains (PVDs) constructed in Ballina, Australia. The effect of creep on settlement predictions is illustrated by the comparisons of the cases with and without considering creep. The results show that good settlement predictions could be obtained if creep is ignored and could be further improved if creep is incorporated when the monitoring settlement data is applied in the Bayesian back analysis. Ignoring creep could lead to an underestimation of the ultimate consolidation settlement. The swelling index $\kappa$ and the compression index $\lambda$ need to be adjusted to larger values to match the measurements if creep is ignored. Four updating schemes (using surface settlement data only, using settlement data at all monitoring depths, using pore water pressure data only, and using both settlement and pore water pressure data) are applied to study the effects of monitoring data on the accuracy of settlement prediction. The results show that the variability introduced by the noisy pore water pressure data result in fluctuating settlement predictions. Incorporating both settlement and pore water pressure observations into the Bayesian updating process reduces the variability in the updated soil parameters.

The effect of spatial variability on the liquefaction prediction and liquefaction-induced ground failure of a soil profile with thinly interbedded layer of sand and clay was examined using simplified procedures and fully coupled two-dimensional (2D) nonlinear dynamic analysis (NDA). The Palinurus Road site in Christchurch, New Zealand, subjected to the 2010 Darfield and 2011 Christchurch events, was selected for the analyses. The liquefaction vulnerability of the site was estimated using one-dimensional (1D) liquefaction vulnerability indices (LVIs), and the results were compared with the results of NDAs and field observations. Spatially correlated random fields conditioned on corrected cone penetration tests (CPTs) were generated based on sand-like and claylike portions of the interbedded soil profile. CPT data were corrected for the thin-layer and transition zone effects. The generated random fields were assigned to the interbedded layer of the stochastic model to examine the effect of spatial variability on the system dynamic response, cross-interaction between layers, and liquefaction-induced diffusion during the evolution of liquefaction and postliquefaction. The potential of sediment ejecta manifestation using the artesian flow potential (AFP) and ejecta potential index (EPI) also was investigated for the stochastic models and compared with those of deterministic models. This study demonstrates the value of advanced 2D NDA modeling with realistic soil spatial variability for understanding inconsistent ejecta predictions from simplified tools. The approach can guide future refinements to assessment procedures and provide insights into key factors controlling manifestation in complex stratified sites.

Given their high cost, servo steel struts typically are used only in specific sections of foundation pits in which nearby buildings are sensitive to excavation-induced deformation. Servo steel strut adjustments affect the mechanical behaviors of adjacent ordinary-steel-strut-supported areas. However, the plane strain assumption is used in current designs, ignoring this influence. To propose a more reasonable design, the effects of servo steel strut adjustments should be examined. In this study, a model support system with adjustable strut axial forces was constructed (scale: $1:20$, under $1g$), and several groups of tests with different adjustment schemes were conducted. The lateral earth pressure, lateral wall deflection, and strut axial forces were monitored and analyzed. The results indicated the following: (1) the first level of struts is not suitable to be actively adjusted because this can increase the maximum lateral wall deflection; (2) increasing the strut axial forces can lead to redistribution of lateral earth pressure along the horizontal direction, which can be divided into two areas with different characteristics, and whereas the lateral wall deflection decreases in both areas, the lateral earth pressure increases in the area adjacent to the adjusted strut and decreases in the other area; and (3) a transition section can be designed between the servo-steel-strut-supported area and the ordinary-steel-strut-supported area, within which strut adjustment diminishes as one approaches the ordinary-steel-strut-supported area. However, the horizontal length of the transition section should exceed a threshold value, which is approximately 3 times the horizontal strut distance, to effectively mitigate the axial force reduction in ordinary steel struts.

The influence of anisotropic consolidation on the cyclic liquefaction resistance of granular materials is explored using 3D discrete element method simulations. In this study, the term anisotropic consolidation was defined as the ratio of initial horizontal and vertical normal stresses, and the hypothesis was that the conflicting results from previous laboratory experiments could be attributed to differences in inherent fabric. To test this hypothesis, three unique sample preparation protocols were employed to construct polydisperse spherical particle samples with varying inherent fabrics, as quantified by coordination number and contact-normal fabric anisotropy, under consistent initial mean stress and density conditions. The results were intriguing, as they revealed that anisotropic consolidation had a consistent impact on the cyclic liquefaction resistance of loose and medium-dense samples, regardless of preparation protocol. However, this relationship was not as straightforward in dense samples. In addition, the study assessed the correlations between various parameters, including initial shear wave velocity, state parameters associated with both void ratio and coordination number, fabric anisotropy, and their impact on the cyclic liquefaction resistance of the samples. The findings enhance the understanding of the intricate interplay between anisotropic consolidation and the resistance of granular materials to cyclic liquefaction, providing valuable insights that can inform the development of accurate models for predicting and mitigating cyclic liquefaction in various applications.

A pore network model (PNM) idealizes the pore space in a soil as voids (network nodes) connected by pore throats (network edges). Along each edge the fluid flow rate is linearly related to the pressure drop by a hydraulic conductance. This study demonstrates the benefit of using a pore network model (PNM) with an appropriate conductance model in coupled particle–fluid simulations that use the discrete-element method (DEM) to simulate the particle phase. PNM simulations and fully resolved finite-volume method computational fluid dynamics (CFD) simulations are used to obtain the fluid–particle interaction force vectors on particles in virtual samples of sand created using DEM simulations. Linearly graded and bidisperse samples are considered. The study assesses the predictive capabilities of existing conductance models considering local flow rates, global permeability, and particle–fluid interaction force magnitude for the packings. A new refined conductance model that is developed upon models available in the literature is also proposed. Taking the fully resolved CFD data as a benchmark, the PNM approach is shown to better capture the heterogeneity in the magnitude of the particle–fluid interaction force acting on particles with a similar size than the coarse-grid CFD-DEM approach for all the samples considered. The orientation of particle–fluid interaction force vectors obtained from the fully resolved CFD is compared with the direction of the force vector predicted by a coarse-grid CFD-DEM and a PNM with the novel conductance model, where the PNM demonstrates a better accuracy. This work enables more realistic and more accurate coupled simulations of phenomena including liquefaction and internal erosion than has hitherto been possible.

In recent decades, research on renewable energy has been boosted by the emerging awareness of energy security and climate change and their consequences, such as the global cost of adapting to the climate impacts. Both onshore and offshore wind turbine farms have been considered as one of the main alternatives to fossil fuels. Their development currently involves seismic-prone areas, such as the Californian coastline and East Asia, where the risk of soil liquefaction is significant. Onshore wind turbines (OWTs) typically are founded on shallow rafts. Their operation can be affected strongly by the simultaneous presence of intense earthquakes and wind thrust, which may cause remarkable permanent tilting and loss of serviceability. In these conditions, accurate evaluation of the seismic performance of these structures requires the development of well-validated numerical tools capable of capturing the cyclic soil behavior and the build-up and contextual dissipation of seismic-induced pore-water pressures. In this paper, a numerical model developed in OpenSees, calibrated against the results of dynamic centrifuge tests, was used to evaluate the influence of some ground motion intensity Measures of the seismic behavior of OWTs included the amplitude, frequency content, strong-motion duration, and Arias intensity (energy content) of the earthquake, together with the effect of a coseismal wind thrust, which is not well explored in the literature. The seismic performance of an OWT was assessed in terms of peak and permanent settlement and tilting, the latter of which was compared with the threshold of 0.5° typically adopted in practice.

This paper reports numerical investigations on the bearing performance of a novel finned suction caisson foundation under combined vertical, horizontal, moment, and torsional loading in clay. From comprehensive comparison analyses between finned and nonfinned suction foundations, improvements in uniaxial and combined loading capacities due to the assembled fins are examined and interpreted by insights into the failure mechanisms. Effects of non-coplanar torsional loading on the combined loading capacity surfaces are investigated and evaluated with the aid of the failure envelope method. A set of closed-form expressions accommodated to three different loading scenarios (zero-, low- and high-torsion conditions) are proposed based on a detailed design chart to estimate the bearing performance under combined vertical ($V$), horizontal ($H$), moment ($M$), and torsional ($T$) loading (denoted as VHMT loading) in clay. The critical considerations in design including effects of loading direction and foundation–soil interface are also discussed.

This study investigated the effect of gradation, especially of mean particle size and uniformity of gradation, on the shear modulus degradation of natural marine sands compared to regraded clean sands. A series of fixed-free resonant column tests were performed on graded clean sands and marine sands having similar uniformity coefficient values, $C_u$. It is observed that the modulus degradation of both clean sands and marine sands increased with the increase in $C_u$. However, the marine sands have degraded less than clean sands for the same $C_u$ and effective confining stress. The reasons can be attributed to the nonlinearity in the grain size distribution of marine sands above 60% finer and mean particle size, $D_{50}$, compared to graded clean sands. The overestimation of the modulus degradation of marine sands by the existing empirical equations is demonstrated. Considering the effects of uniformity coefficient, $C_u$, and mean particle diameter, $D_{50}$, a generalized calibrated modulus degradation model for a wide range of clean sands and marine sands is proposed with a high coefficient of determination. Good agreement is demonstrated between the estimated and experimental data from the literature.

The empirical expressions for predicting the small-strain shear modulus ($G_0$) of granular soils in current engineering practice are established mainly on experimental data under isotropic stress conditions. In most geotechnical applications, however, soils are subjected to anisotropic stress conditions. The impact of stress anisotropy on $G_0$ is a critical concern but is not yet fully understood. In this paper, we present a specifically designed experimental study to address the question. Various principal stress ratios were applied to isotropically consolidated sand specimens in a triaxial apparatus, and the elastic shear waves were generated by the bender elements installed in the apparatus such that the variations of $G_0$ from isotropic stress states to anisotropic stress states were determined. Three quartz sands with different particle shapes were tested under a range of states in terms of void ratio, axial stress, and radial stress. The study shows that the impact of stress anisotropy is much more complicated than commonly thought. It depends on the magnitude of the stress ratio and the loading mode. A simple model that accounts for two primary mechanisms associated with the impact of stress anisotropy is proposed, and its performance is evaluated using various sources of data in the literature.

This paper re-evaluates the Edenville earth dam that failed in 2020 after a rapid rise in reservoir level caused by prolonged significant rainfall. The dam instability and possibility of static liquefaction–induced flow failure are assessed in the framework of critical state soil mechanics (CSSM). Both undrained loading and drained unloading stress paths are considered for flow liquefaction assessment. Using documented triaxial compression tests, the Norsand constitutive model is calibrated for finite-element numerical models. This study supplements the analyses conducted by an independent forensic team (IFT) in charge of failure investigation. The advances achieved by tailings dam practitioners are leveraged here to study flow liquefaction as recommended by the IFT. Multiple ways are employed to generate reasonable initial field stresses. Simulation of dam construction coupled with steady-state seepage analysis creates an appropriate starting condition for further flow liquefaction assessment. The numerical outputs confirm the result of the IFT’s limit equilibrium analyses for an undrained boundary condition. Stress analyses with a drained boundary condition show that a major stress change with a constant-shear-drained (CSD) stress path develops below the dam crest due to a rising reservoir level. The results of our models with a CSD stress path and drained-to-undrained transition better explain the evolution of the Edenville dam failure. The current study provides a practical approach to earth dam design and evaluation.

Energy piles are often partially or fully embedded in unsaturated soils, particularly in arid and semiarid regions; however, suction effects on their thermomechanical behavior have not been fully understood. In this study, a series of physical modeling tests were carried out to investigate the behavior of energy piles in unsaturated soils. The soil used for testing is completely decomposed granite, which is a type of clayey sand sampled from Hong Kong. Five constant-temperature pile load tests and three constant-load heating and cooling tests were conducted under different conditions of pile surface roughness and initial soil suction. It is observed that the ultimate capacity of the energy pile ($Q_{\mathrm{ult}}$) increases as the initial suction increases from 0 to 90 kPa, the pile roughness increases from 0.05 to 1, and the temperature increases from 10°C to 21°C. Moreover, the effects of roughness are more significant in unsaturated conditions than in saturated states. A suction increment can increase both shaft and toe resistance, while temperature and roughness mainly affect the shaft resistance. During cyclic heating and cooling, suction and roughness increment reduces the irreversible pile head settlement due to the increment of shaft resistance. Furthermore, the irreversible pile head settlement is greatly affected by vertical load. When the working load is relatively small (smaller than $0.5Q_{\mathrm{ult}}$ in this study), the settlement increases with the number of thermal cycles but at a decreasing rate for both saturated and unsaturated conditions, and finally reaches a stable state. When the working load is relatively high ($0.7Q_{\mathrm{ult}}$ in this study), the response of the pile does not reach a stable state in saturated conditions even after five thermal cycles. These observations should be treated with caution because they may be specific and only applicable to the case of no overburden pressure. Exercising caution is crucial when interpreting these observations, particularly at a quantitative level, because they may be limited to the test soil and the low stress level utilized in the physical model tests.

Abstract

Air-booster vacuum preloading (AVP) without sand blanket and surcharge vacuum preloading (SVP) methods were used to treat two adjacent areas in a foundation reinforcement project on Donghai Island, Zhanjiang City, Guangdong Province, China. In this study, engineering characteristics of the AVP foundation were obtained through analyzing monitoring data. Results showed that the AVP method obtained superior foundation consolidation results compared to SVP method. This study proposed a new simplified mechanical model for AVP, with a single PVD (prefabricated vertical drain) centered and the air-booster pipes symmetrically distributed around the PVD, applied equivalent air-booster pressure to the drainage boundary. Furthermore, this study established a mathematical analytical model for predicting consolidation settlement based on the proposed simplified model. By comparing the theoretical settlement value with the measured value, this paper verifies the rationality of the analytical solution. Finally, this study analyzed the effects of well resistance, smearing effect, compression index, and permeability index ratio on consolidation characteristics.

Practical Applications

Land reclamation projects will generate a large amount of dredged ultrasoft soil foundation. The AVP method can achieve superior foundation treatment effects by injecting gas into soil mass through air-booster pipes. The results of this case study demonstrate that the AVP method has substantial advantages compared to the conventional vacuum preloading method. The AVP method can increase foundation settlement, shorten the time for foundation treatment construction, and improve the foundation reinforcement effect. In addition, the AVP method can alleviate PVD clogging and maintain soil permeability, while simultaneously avoiding the excessive use of dredging sand. Therefore, the AVP method can lead to potential economic benefits and protect the natural environment, and it is a method worth considering. Moreover, the mathematical model for predicting consolidation settlement of the AVP method was proposed based on new mechanical assumptions, which can be applied as a reference for settlement prediction in relevant engineering, and the rationality has been verified through comparison with measured settlement results.

A geosynthetic clay liner (GCL) was exhumed from a composite liner (geomembrane over GCL containing granular sodium bentonite) in the base of a landfill cell after 12 years of atmospheric exposure. The GCL was altered appreciably during exposure, despite being overlain by a geomembrane. Water content, hydration, and bentonite erosion varied considerably along the slope. GCL samples from the top of the slope were dry, with visible bentonite granules comparable to those in a virgin GCL as if the bentonite had not hydrated. No substantial change in exchange complex or swell index (SI) occurred in samples from the top of the slope, and hydraulic conductivity of these samples to water was low ($\approx {10}^{-11}\mathrm{m}/\mathrm{s}$). GCL samples from the toe of the slope varied considerably, from moist and soft to dry and cracked, but all had undergone hydration, and had low SI, a small mole fraction of monovalent cations in the exchange complex, and high hydraulic conductivity (${10}^{-8}\mathrm{m}/\mathrm{s}$ or higher). Bentonite in GCL samples from midslope had a cracked structure commonly observed in GCLs that have undergone wet–dry cycling. Midslope GCLs typically had hydraulic conductivity greater than ${10}^{-7}\mathrm{m}/\mathrm{s}$. A sample from the top of the slope from a location near a GCL panel separation was comparable to other samples from the top of the slope. GCL samples from the anchor trench differed considerably, being very permeable or having low hydraulic conductivity. Hydraulic conductivity was strongly correlated with SI; GCL samples with $\mathrm{SI}<15\mathrm{mL}/2\mathrm{g}$ were highly permeable, and those with $\mathrm{SI}>15\mathrm{mL}/2\mathrm{g}$ were comparable to a virgin GCL. The variation in SI was related directly to replacement of sodium by calcium in the exchange complex of the bentonite. Hydraulic conductivity also was affected by the thinning of the GCL.

Abstract

Strut-free retaining walls are an efficient and cost-effective technology for large-scale excavations, but their application at excavation depths of more than 10 m in soft soils has rarely been reported. An innovative multibench retaining system composed of an inclined–vertical framed retaining wall (IVFRW) and a cantilever wall with anchors was used in a 14.7-m-deep excavation in soft clay in Tianjin, China. To analyze the pile–soil interaction considering the influence of the second stepped excavation, indicators were monitored to provide comprehensive measured data. Field measurements included the lateral deflection and vertical displacement of the retaining wall, the ground water level variation, the ground settlement, and the axial force of the piles within the IVFRW. The measured results indicated that the wall displacements and the ground movements caused by the excavations were small in comparison to those of previous projects and empirical formulas. The influence of the second stepped excavation on the soil|structure interaction and its influence on the axial force of the IVFRW were analyzed. The optimal values of the first stepped excavation depth ($H_1$) for different site conditions were further investigated.

Practical Applications

This case study reports an application of multibench excavation with IVFRW in a 14.7-m excavation in soft soil area. Compared with propped excavations, the multibench excavation with IVFRW has comparable wall deflection while greatly reducing the material consumption and the construction duration. In addition, the amount of pile wall and pile wall length is greatly reduced in comparison to conventional multibench excavations in similar soil condition and excavation depths. This case history provides an economic and efficient strategy for professional engineers to integrate a similar solution in design. During the excavation process, the compressive stress accumulated on the inclined piles within IVFRW restricting the displacement of retaining structure. Due to the second stepped excavation affecting the development of the axial force of the inclined piles, the optimal excavation depth $H_1/H_{\mathrm{e}}$ varies. In engineering design, a width of earth berm should be first determined according to the available workspace. An optimal stepped excavation depth can then be chosen based on this study to achieve the best deformation performance or a minimum material consumption.

This technical note presents a relationship between soil behavior type index, $I_c$, obtained from cone penetration testing (CPT), and soil fines content, FC, obtained from laboratory testing of samples from colocated borings. Liquefaction evaluation based on CPT testing is best informed by independent evaluation of FC and plasticity characteristics by retrieving samples from the site. However, sampling is not feasible for many projects, in which case correlations between CPT data and FC are used despite the added uncertainty. The data set used to develop the relationship is from the Next Generation Liquefaction (NGL) project and consists of 227 colocated CPT soundings and boring logs from 111 sites, and a total of 2,714 $I_c\text{-}\mathrm{FC}$ pairs. This global data set is larger and spans a broader range of locations and geological settings than has been considered in previous models. A model was fit to the data using linear least-squares regression following logistic transformation of the FC data, and standardization of both $I_c$ and the transformed FC data. The model is presented in a probabilistic framework based on the computed standard deviation of the residuals. An influence of soil plasticity on model residuals was found, but PI is not included as an independent variable in the model, mainly because plasticity measurements require sampling, in which case FC should be measured directly rather than being computed from $I_c$.