Journal of Geotechnical and Geoenvironmental Engineering

Soil suction–based analyses are paramount to proper design of foundations in expansive soils, yet availability of field soil suction profiles from direct measurement is limited. A method was developed, using directly measured suction profiles obtained from sites across the US, by which routine geotechnical tests can be used to estimate soil suction to expand the database for estimation of field suction profiles. Herein, the expanded database is used to develop best-estimate clay soil suction envelopes for boundary conditions associated with covered or uncovered ground surface subjected to natural climatic surface flux. Where appropriate for site conditions, the proposed soil suction profile envelopes can be used with any suction-based heave/shrinkage analysis. Key features of the field suction envelopes, consistent with current popular approaches, include climate-dependent estimates for equilibrium suction at depth, depth to equilibrium (stable) suction, and variation in soil suction at the ground surface. Comparisons to existing soil suction envelope recommendations are made. Finally, a limited number of cases of developed/irrigated sites are presented for comparison to soil suction profiles obtained for natural climatic surface flux conditions.

Particle granulometry plays an important role in the engineering behavior of many sands. However, the evaluation of particle shape and size has historically been a tedious and labor-intensive process. The recent availability of dynamic image analysis (DIA) makes it possible to evaluate many particle shape and size parameters, quickly and conveniently. These shape parameters include sphericity, roundness, aspect ratio, circularity, and convexity; while size descriptors include the diameter of a circle of equal projection area (EQPC), a variety of Feret diameters, as well as inscribed and circumscribed circle diameters. The terms roundness and sphericity are commonly used to describe how close a particle resembles a sphere, with many definitions in common use. However, it is not immediately evident how these roundness descriptors correlate. The correlation of nine shape and six size descriptors was investigated for six sands that reflect the breadth of particle shapes and sizes that may be encountered. The analysis was based on 1,000 images of each sand obtained using two-dimensional DIA apparatus. The study demonstrates that there is no correlation between size and shape parameters, and that shape descriptors can be reduced to four independent shape parameters representing the granulometry of sand at different scales. The use of size and shape descriptors for classification of sand was explored using six machine learning algorithms including support vector machines (SVMs), random forest, decision tree, bagging tree, $k$-nearest neighbors (KNN), and bagging KNN. Classification accuracies of 77% and 66% were achieved using size and shape features, respectively. The mean accuracy improved to 87% when combining both size and shape descriptors using bagging KNN and random forest classifiers. The analysis also revealed an important hierarchy of size and shape features employed, with EQPC and Wadell’s roundness alone classifying sands with 70% accuracy.

The delineation of vertical geological cross-sections is an essential task in geotechnical site characterization and has a profound impact on subsequent geotechnical designs and analyses. It is a long-lasting challenge, particularly for complex geological settings, to properly produce a subsurface geological cross-section from limited boreholes that are usually encountered in engineering practice. Emerging machine learning methods, such as the convolutional neural network (CNN), provide a fresh perspective of this challenge and effective alternatives for exploiting the complex stratigraphic relationships between different soil deposits. In this study, a novel iterative convolution eXtreme Gradient Boosting model (IC-XGBoost) is proposed. This model interpolates a subsurface geological cross-section from limited site-specific boreholes and a training geological cross-section obtained from previous projects with similar geological settings. This direct application of previous geological cross-sections for training is based on the assumption of similar local spatial connectivity or stratigraphic relationships between soils in areas with similar geological settings. The proposed method can learn stratigraphic patterns from a training image in an automatic manner. In addition, the proposed method is purely data-driven and does not require the specification of any parametric function form. The model performance is illustrated using both a simulated example and real data from a tunnel project in Australia. The proposed method not only infers the most probable geological cross-section but also quantifies the associated interpolation uncertainty from multiple realizations. The effect of the borehole number on the interpolation performance is also explicitly investigated.

A recently developed damping calculation method for undrained load-controlled cyclic tests was implemented on some 70 cyclic triaxial and direct simple shear tests conducted by various researchers on a range of different sands for a correct understanding of the high-strain (0.1%–4%) damping behavior of sands. The effect of various controlling parameters was investigated, including the shear strain, confining stress, number of cycles, void ratio, gradation, nonplastic fines content, and overconsolidation ratio. The data analysis shows that general trends at high strains are in good agreement with the damping behavior of sands at small strains. The results also indicate a direct correlation between the damping behavior of sands and cyclic strength. The recently proposed guideline on load-controlled damping ratio bounds for clean sands at high strains was also in good agreement with the damping behavior of sands with nonplastic fines content up to 25%.

A half-scale field test was conducted on a pile group installed at a stiff clay site. The pile group consisted of eight steel H-piles, with a battered pile at each corner, and connected to a precast pile cap using preformed socket connections. The pile foundation system was tested under different combinations of vertical and lateral loads. The magnitude of the loads, as well as the height of the lateral loads, were varied to produce different overturning moment-to-lateral load ratios. These load combinations enabled the researchers to investigate the foundation system for both nonseismic and seismic conditions as well as its behavior under extreme loading. The pile group produced excellent resistance during different phases of testing although there were differences in the behavior of battered and vertical piles. Under simulated seismic loading, the battered piles sustained 17% more axial compression forces, 100% more axial tension forces, and 33% lower bending moments than the vertical piles. A numerical model developed to investigate the pile behavior produced a good agreement between the measured and computed pile responses.

To investigate the liquefaction mitigation capability of rammed aggregate piers (RAP) in silty sand, blast liquefaction testing was performed at a soil profile treated with a full-scale RAP group relative to an untreated soil profile. The RAP group consisted of 16 piers in a $4\times 4$ arrangement at $2\mathrm{m}$ center-to-center spacing extending to a depth of $9.5\mathrm{m}$. Blasting around the untreated area induced liquefaction ($r_u\approx 1.0$) from a 3 to $11\mathrm{m}$ depth, producing several large sand boils and causing a settlement of $10\mathrm{cm}$. In contrast, the installation of the RAP group reduced excess pore water pressure ($r_u\approx 0.75$), eliminated sand ejecta, and reduced the average settlement to between 2 and $5\mathrm{cm}$ when subjected to the same blast charges. Although the liquefaction-induced settlement in the untreated area could be accurately estimated using an integrated cone penetration test (CPT)-based settlement approach, settlement in the RAP treated area was significantly overestimated with the same approach, even after considering RAP treatment-induced densification. Analyses indicate that settlement after RAP treatment could be successfully estimated from liquefaction-induced compression of the sand and RAP acting as a composite material. This test program identifies a mechanism that explains how the settlement was reduced for the RAP group despite the elevated $r_u$ values in the silty sands that are often difficult to improve with vibratory methods.

The soil freezing curve (SFC) is a fundamental constitutive relationship between liquid water content and temperature under subzero (0°C or 273.15 K) conditions. SFC governs mechanical and hydrologic behavior of soil in freezing and thawing environments. The state-of-the-art SFC paradigms have been established empirically based on the capillary pressure-based Clapeyron equation. Two practical challenges prevent the rigorous use of the capillary pressure-based Clapeyron equation for realistic prediction of the SFC: (1) unable to use the governing pressure (intermolecular water pressure) for defining water phase change; and (2) unable to account for variations in latent heat of fusion and water density. A new paradigm based on soil sorptive potential (SSP) to predict the SFC from the soil water retention curve is developed, directly using the intermolecular water pressure distribution and pure water phase diagram in lieu of capillary pressure and the Clapeyron equation. The latest theory of SSP is used to quantify intermolecular water pressure distribution. Experimental validation demonstrates that the proposed paradigm yields excellent matches to the experimental SFC data for different soil types, and is a significant improvement over the predictions by the capillary pressure-based Clapeyron equation paradigm. The proposed paradigm reveals that the SFCs for various soils below 273.15 K (0°C) are mostly dominated by adsorptive water. Furthermore, the proposed paradigm can fully explain the practically encountered phenomenon that the SFC for soils with high clay content depends on the initial water content, whereas it does not for sandy soils. Practical significance of the new paradigm in geotechnical engineering problems is demonstrated through predicting soil moisture profiles under freezing and thawing, and permafrost environments.

Pile-supported (PS) embankments have been used increasingly to support highways and railways on soft subsoils. In addition to the self-weight of the embankment, this embankment system is often subjected to surface localized loading, such as traffic loading. In this embankment system, soil arching is a key load transfer mechanism. Stability of soil arching under localized surface loading is important because traffic loading applied on the embankment surface can transfer onto and between pile heads and affect the degree of mobilization and degradation of soil arching. Conventional trapdoor systems have a rigid trapdoor, control its displacement manually/automatically, and cannot represent load-induced subsoil settlement below the embankment. This study utilized a trapdoor supported on low-stiffness or high-stiffness compression springs that moved under the load above the trapdoor (called a spring-based trapdoor) to evaluate the effects of continuous trapdoor displacement on the soil arching stability under static footing loading. To investigate the trapdoor rigidity effect, a trapdoor consisting of three segments (called a flexible trapdoor) was utilized in this study as well. The trapdoor test results showed that soil arching was mobilized during fill placement as the fill height and the trapdoor displacement increased. Subsequently, under static footing loading, the degree of soil arching increased at a low applied pressure; however, it degraded under higher footing loading that caused a larger trapdoor displacement. The high-stiffness trapdoor increased the degradation pressure required to eliminate soil arching even though it reduced the degree of soil arching under a low applied pressure compared with the low-stiffness trapdoor. The flexible trapdoor resulted in a uniform stress distribution on the trapdoor but reduced the total load transferred to the supports. The conventional trapdoor resulted in a lower soil arching ratio compared with the spring-based trapdoor during the soil arching mobilization, but had a higher soil arching degradation rate than the spring-based trapdoor due to continuous soil movement.

This paper studies the effect of a high effective overburden pressure [${\sigma }_{v0}^{\prime }=\sim 600$ kPa (6 atm)] under two drainage conditions on the field liquefaction behavior of saturated Ottawa sand. A series of eight centrifuge experiments with relative densities $D_r=45\%$ and 80% and base shaking are considered that include a 5-m saturated sand layer under a pressure of either ${\sigma }_{v0}^{\prime }=\sim 100$ kPa (1 atm) or $\sim 600$ kPa (6 atm). Four of the tests had single drainage at the top of the layer (SD), whereas the other four tests had double drainage (DD) at top and bottom. The four SD test results had been reported before, whereas the four DD tests are new. A novel centrifuge technique was developed to achieve the double-drainage boundary condition of two pervious boundaries at the top and bottom of the sand layer, using a geocomposite at the bottom. Measured responses are compared at the same ${\sigma }_{v0}^{\prime }$ between SD and DD tests having the same input acceleration, as well as between SD and DD tests where the shaking induced a similar maximum excess pore pressure ratio $(r_u{)}_{\max }\approx 0.8$. These comparisons include acceleration time histories, excess pore pressure time histories and profiles during and after shaking, and stress ratio and shear strain time histories. Comparisons between corresponding tests at $\sim 100$ and $\sim 600$ kPa (1 and 6 atm) revealed significantly more partial drainage at $\sim 600$ kPa (6 atm) than at $\sim 100$ kPa (1 atm), with even more significant variation in excess pore pressures in the DD than in the SD tests. Best estimates of field overburden pressure correction factors at $\sim 600$ kPa (6 atm), $K_{\sigma }$ were obtained, were obtained from the centrifuge results with two independent methods for a failure criterion of $(r_u{)}_{\max }=0.8$. Those $K_{\sigma }=1.2–1.3>1.0$ for both SD and DD drainage conditions due to the significantly lower compressibility of the sand at $\sim 600$ kPa (6 atm). The results further emphasize the important role partial drainage may play in the field during shaking at high ${\sigma }_{v0}^{\prime }$ on the excess pore pressures and values of $K_{\sigma }$.

Common practice for evaluating the seismic performance of earth dams uses design ground motions selected to be consistent with a target design response spectrum, which are subsequently used in dynamic analyses that estimate seismically induced displacements as an index of performance. This approach involves selecting the probability of exceedance of a ground-motion parameter rather than the probability of exceeding the seismically induced displacements. This implicitly assumes that the different response spectra levels (e.g., different $S_A$ hazard levels or $50\mathrm{t}\mathrm{h}$ percentile versus $84\mathrm{t}\mathrm{h}$ percentile ground motions) is correlated to the different levels of seismically induced displacement. This may not always be the case and can be particularly problematic when evaluating earth structures located in a tectonic setting with multiple source types (e.g., subduction and shallow crustal earthquakes). This can lead to high variability in the estimated displacements, making the selection of a representative overall displacement computed from the different source types not immediately clear. In this study, we propose simplified approaches to select representative displacements for dams affected by earthquakes from multiple source types and we evaluate their performance by constructing displacement hazard curves that rely on the conditional scenario spectra (CSS) framework. We illustrate the application of the proposed procedures for a fictitious dam located in Vancouver, British Columbia, and offer recommendations for using the proposed procedures in practice. Finally, we propose a new procedure for selecting a subset of ground motions for use in complex dynamic analyses [e.g., FEM or finite-difference methods (FDM)] based on the deaggregation from displacement hazard curves rather than the deaggregation from elastic response spectra hazard curves.

Current simplified analysis methods for rectangular underground structures subjected to earthquake action present in US design codes or analogous documents are based on the structure shear distortion assessment by means of closed-form solutions of the so-called racking coefficient, and then evaluation of the structure internal forces through the analysis of a simple frame subjected to imposed displacements. However, current formulations for these methods provide inaccurate results in terms of internal forces when compared with numerical analyses and experimental measures, especially for the more flexible structures compared with surrounding ground corresponding to large underground spaces like metro stations. This paper addresses this issue and proposes an improved yet simple method to compute internal forces for deeply and shallowly buried rectangular underground structures subjected to earthquake action. The proposed method is proven to be more accurate predicting internal forces than current simplified methods when compared with finite-element analyses for structure flexibility ratios $0<F_R\le 120$.

Pullout of plate anchors from granular sands is investigated using a novel computational multiscale approach. We employ the material point method (MPM) to solve a large deformation boundary value problem and adopt the discrete element method (DEM) to derive the history-dependent material responses required for each material point of the MPM domain. The continuum-discrete hierarchical coupling between MPM and DEM not only helps to bypass the assumption of complicated phenomenological constitutive models for sand, but also facilitates the handling of large displacement movement of the anchor and its ensuing complicated interactions with surrounding soil. This multiscale method is used to simulate the pullout of both horizontally and vertically placed plate anchors in sand by a large displacement, and to examine the roles of key factors, including the relative density of sand and the embedment depth, on the bearing capacity and pullout behavior. For a horizontally placed plate anchor, a truncated cone shape of soil body is mobilized upward at shallow embedment depth, whereas at greater depths, the surrounding soil may flow from the top to the bottom of the anchor, forming an interesting circulating circular shape. For a vertically placed plate anchor, the failure pattern of soil evolves gradually from a general shear failure mode to a local rotational failure mode when the embedment depth is increased. The study also provides cross-scale insight for the macroscopic observation on anchor pullout and comparisons with past studies.

Verification core holes are important tools in ascertaining the properties of the bed rock at the bottom of drilled shafts. However, it is unknown whether verification core holes fill during concrete placement; if they fill, the mechanical properties of the filling are unknown. The diameters of the core holes are in the $15–30\mathrm{cm}$ range, and core holes are at least $1.5\mathrm{m}$ long; empty core holes or core holes filled with weak and deformable materials may significantly decrease the capacity of drill shafts. The paper describes the experiments performed to answer these two questions by using 15- and $30\text{-}\mathrm{cm}$ verification core hole models, and drilled shaft depths between 9- and 30-m depth. The smallest slump allowed by the Texas Department of Transportation (TxDOT) was used: 15 and $18\mathrm{cm}$ in dry and wet conditions, respectively. In dry conditions, verification core holes completely filled with concrete, whose compressive strength exceeded the TxDOT specifications. The compressive strength of the concrete obtained in the $30\text{-}\mathrm{cm}$ core holes was higher than the concrete strength in the $15\text{-}\mathrm{cm}$ core holes by 60% to 150%. In wet conditions, unless the tremie or pump pipe was inserted into the bottom of the core hole, the bottom half of the verification core filled with uncemented gravel–sand mixture (angle of internal friction, $\phi =52°$), while the upper half of the verification core filled with weakly cemented material whose P-wave velocity was found to be $610\mathrm{m}/\mathrm{s}$.

The influences of consolidation conditions and rotation patterns of principal stress on the liquefaction susceptibility of saturated coral sand have emerged as an interesting problem in recent years. This paper presents results from a comprehensive experimental study comprising undrained monotonic shear tests and undrained cyclic shear tests subjected to various patterns of principal stress rotation. A remarkable finding is that a virtually unique correlation exists between a generalized shear strain amplitude ${\gamma }_{ga}$ and the excess pore water pressure ratio $r_u$ irrespective of consolidation conditions and cyclic loading patterns. A simple formulation is then proposed to relate ${\gamma }_{ga}$ and $r_u$. Another significant finding is that the liquefaction susceptibility of coral sand and the correlation between the conventional cyclic stress ratio (CSR) and the number of cycles to failure $N_f$ (corresponding to ${\gamma }_{ga}=2.5\%$) are strongly affected by the couplings of consolidation conditions and cyclic loading patterns. By introducing a generalized unit cyclic stress ratio ($US{R}_g$) as a new proxy for liquefaction resistance, a strong correlation is found between $US{R}_g$ and $N_f$ for all data sets from the experiments. An explicit relationship is then proposed for practical application. The wide applicability of this relationship is well demonstrated using the literature data for various undrained cyclic laboratory tests and different sands.

Recent earthquake events have indicated that liquefaction histories can greatly affect the reliquefaction resistance of saturated sandy ground. This paper reports a fundamental experimental study of medium-dense/dense Toyoura sand with a relative density of 70% using a hollow cylinder torsional shear apparatus (HCTSA) to provide a better understanding of the integrated effects of strain history and soil fabric on reliquefaction resistance of Toyoura sand. A series of postliquefaction monotonic tests was conducted to investigate the degrees of induced fabric anisotropy at different states of liquefaction and subsequent reconsolidation. A series of reliquefaction tests was performed on sandy specimens that had experienced medium to large residual shear strains (${\gamma }_{\mathrm{res}}$) and that had reconsolidated in different states, under various cyclic stress ratios (CSRs). It was shown that reliquefaction resistance of Toyoura sand is influenced highly by the residual shear strain during liquefaction and fabric anisotropy after reconsolidation. The experimental results were used to predict excess pore-water pressure (EPWP) buildup during liquefaction and reliquefaction as a function of shear strain or dissipated energy. The liquefaction/reliquefaction resistance of Toyoura sand was quantified using an energy-based criterion considering the integrated effects of strain history and soil fabric.

Assessing the stability of rock slopes is essential to ensuring the typical performance of adjacent transportation infrastructures but simulating the existing field topography and boundary conditions for stability analysis is challenging. This paper describes a case study in which highly weathered rock slopes adjacent to a railroad track in Texas were analyzed using close-range photogrammetry data obtained from an unmanned aerial vehicle with close-range photogrammetry (UAV-CRP). A highly weathered rock mass with low intact strength was considered as a single continuum that undergoes circular failure and was analyzed using Spencer’s two-dimensional (2D) and three-dimensional (3D) limit equilibrium method (LEM) analyses. A 3D critical slip surface was obtained to identify the critical sections on which to conduct the individual 2D stability analysis and obtain a safe resloping angle. These angles were further used to reslope the whole rock slopes and conduct a 3D analysis to arrive at a safe and economic slope angle for each side of the rock-cut. The realistic geometrical slopes developed from the aerial imagery were used to develop a framework for comprehensively analyzing the stability of a rock slope undergoing circular failure.

The need for a reliable estimate of the actual failure probability of existing river embankments under changing boundary conditions represents an ever-demanding task for researchers and designers, and those involved in their maintenance and management. Uncertainty and variability of soil suction and water content spatial and temporal distributions, together with the definition of a suitable soil model that accounts for the partially saturated state of embankment materials, are among the most critical aspects to be possibly included in the advanced analysis and design of such linear earthen infrastructures. The use of specialist integrated monitoring can be here functional to enable calibration and to enhance reliability and consistency of predictive analyses. Site measurements of the main variables at relevant depths are typically rarely available, whereas an accurate soil characterization under partially saturated conditions is only performed in research applications, thus producing limited confidence on stability conditions. To provide a useful and innovative tool to evaluate realistic stability conditions of levees under transient flow conditions, a full-scale monitoring system has been implemented on an existing section of the River Secchia (northern Italy) 11-m-high flood defenses. The complementary use of specific laboratory tests, innovative field measurements and numerical analyses presented in the paper aims at providing a suitable methodological approach to the performance assessment of these vital geotechnical systems throughout their entire lifetime and highlighting possible limitations of typically used methods of analysis.

The upstream or waterside slopes of water-retaining structures, such as earth dams and levees, must be designed for rapid drawdown (RDD). While multiple instances of RDD failure have occurred, only two case histories, Pilarcitos Dam and Walter Bouldin Dam, have sufficient shear strength data to be useful for validation of analysis procedures. This paper documents a RDD failure at Sparmos Dam in Northern Greece in 2016, including the results of index, compaction, and triaxial compression tests on soil from the dam. Based on the collected data, the factor of safety following RDD is calculated using two effective stress and two undrained methods. The $\overline{B}$ effective stress method and both undrained methods are able to predict the Sparmos Dam failure, if a conservative interpretation of shear strength is employed. In contrast, the effective stress evaluation of RDD using uncoupled transient seepage analysis produces invalid results when appropriate values of the coefficient of consolidation are used. The Sparmos Dam case history highlights the need for ongoing study of RDD analysis procedures and shows that pore pressures from uncoupled transient seepage are unreliable for RDD.

This paper proposes analytical solutions and a normalized interaction diagram for the lateral responses of a rigid pile subjected to a combined horizontal force and moment. The analysis model considers a rigid pile embedded in homogeneous cohesive soil (assuming that the horizontal subgrade stiffness and yield strength are constant with depth). Different states of soil yielding, as well as the ultimate state, are defined according to the development of the yielding range of soil around the ground surface and pile tip. Through the equilibria of moment and horizontal force at the pile head, the analytical solutions of the failure envelope and the displacement and rotation responses for different soil states are derived. These solutions can be used to build the force-displacement and moment-rotation curves at the pile head. The results of the laboratory and field tests reported in the literature are simulated for demonstration. Furthermore, the normalized interaction diagram between the pile-head moment and horizontal force is built, using which the state of soil yielding and corresponding displacement mode of the pile under a pile-head loading condition can be traced.

This technical note investigates the effect of footing embedment depth on tunnel–structure interaction using geotechnical centrifuge testing. A 2-story framed building on separate footings, either resting directly on the surface or embedded in the soil, and subjected to tunneling-induced displacements is modeled. Measurements of the displacements of the footings and underlying soil, ground deformations, and structural distortions are presented. Results show that footing embedment increases foundation differential settlements and horizontal displacements, thereby causing a greater level of distortion within the frame. Furthermore, the embedded footings result in a larger magnitude of ground displacements and shear strains of the soil. Finally, modification factors and relative stiffness parameters are presented, indicating a greater effect of the embedment on horizontal deformations than the angular distortion of the bays.

This note reports the physical and mechanical properties of two postwildfire sands collected in Southern California and chemically induced hydrophobic sand. Hydrophobicity decreases the attraction between water molecules and solid surfaces and manifests a modification of hydraulic soil properties that lead to postwildfire rain soil erosion and mudflows. Wildfires induce different hydrophobicity levels in shallow soil layers based on fire severity, vegetation, and soil chemistry. This note assesses the applicability of laboratory-made hydrophobic soils for studying mudflows by comparing them to natural hydrophobic soils collected from two burned sites that will help understand and predict postwildfire soil erosion and mudflow mechanisms. Results show that soil water retention curve (SWRC) can be unreliable for assessing the hydraulic properties of burned soils on slopes impacted by rain because suction develops with significant time delay or is absent. Flash floods on inclined hydrophobic soil surfaces occur rapidly during rainfall, and the overflow water does not pond on slopes. The water entry value is linearly related to the water drop contact angle and characterizes the immediate hydrophobic soil response, which is relevant to mudflow onset.