Journal of Geotechnical and Geoenvironmental Engineering

The paper begins with a historical review of reinforced soil technology, beginning with nature and the ancients, the development of Terre Armée or Reinforced Earth by Vidal and Lee, and ending with the early uses of geotextiles for soil reinforcement in France, Sweden (Wager and Broms), and the United States (U.S. Forest Service, Federal Highway Administration, J. R. Bell, T. A. Haliburton, B. R. Christopher, and others). Then the advantages and basic behavior of geosynthetic reinforced soil (GRS) are described. An overview of current design procedures, modified with the results of analytical research conducted at the University of Washington, leads to practical suggestions for dealing with creep, pullout, and backfill drainage. Next the properties of reinforcement geosynthetics are discussed, and important conclusions of research on soil–geosynthetic interaction at the University of Washington are presented. Although GRS is quite a mature development, a few technical and several professional issues remain. The technical issues are relatively straightforward but the professional issues are not: they are costly, potentially tragic, and threaten a wonderful technology and our profession. Some suggestions as to what the profession can do about these issues are given.

Although semi-open H piles have been widely used, geometric effects (GEs) on H-shaped piles in consolidating ground subjected to negative skin friction (NSF) and downdrag are still not well investigated or understood. This paper reports three centrifuge model tests to investigate the responses of one semi-open pile (i.e., H-shaped cross-section) and two solid piles (i.e., circular and rectangular cross-sections) in a consolidating clay stratum. In addition, corresponding three-dimensional (3D) numerical back-analyses considering the elastoplastic slip at the pile–soil interface were carried out and discussed. It is found that the normalized downdrag of the H pile is the largest because the dragload distributed on the end bearing area is the largest among the three piles. The neutral planes of the circular, rectangular, and H piles are located at the normalized depths of $z/H=0.86$, 0.82, and 0.75, respectively. This is because the H pile is subjected to “hang-up” effects due to vertical shearing inside the flanges and web, resulting in a shallower depth of the neutral plane and hence the smallest dragload. Even the average NSF of the H pile is smaller than those of the solid circular and rectangular piles; the largest downdrag induced by consolidating soil should be addressed carefully in the design.

Hydraulic hysteresis, including capillary soil water retention (SWR), air entrapment SWR, and hydraulic conductivity, is a common phenomenon in unsaturated soils. However, the influence of hydraulic hysteresis on suction stress, and subsequently slope stability, is generally ignored. This paper examines the influence of each of these three types of hysteresis on slope stability using an infinite slope stability analysis under steady infiltration conditions. First, hypothetical slopes for representative silty and sandy soils are examined. Then a monitored hillslope in the San Francisco Bay Area, California is assessed, using observed rainfall conditions and measured hydraulic and geotechnical properties of the colluvial soil. Results show that profiles of suction stress and the corresponding factor of safety are generally strongly affected by hydraulic hysteresis. Results suggest that each of the three types of hydraulic hysteresis may play a major role in the occurrence of slope failure, indicating that ignoring hydraulic hysteresis will likely lead to underestimates of failure potential and hence to inaccurate slope stability analysis.

Sampling affects all soils, including frozen soils and hydrate-bearing sediments. The authors monitor the stiffness evolution of frozen sands subjected to various temperature and stress conditions using an oedometer cell instrumented with P-wave transducers. Experimental results show the stress-dependent stiffness of freshly remolded sands, the dominant stiffening effect of ice, creep after unloading, and the associated exponential decrease in stiffness with time. The characteristic time for stiffness loss during creep is of the order of tens of minutes; therefore it is inevitable that frozen soils experience sampling disturbances attributable to unloading. Slow unloading minimizes stiffness loss; conversely, fast unloading causes a pronounced reduction in stiffness probably attributable to the brittle failure of ice or ice-mineral bonding.

This paper presents the results of large-size cyclic simple shear testing on three uniform gravels (Pea Gravel, 8-mm Crushed Limestone, and 5-mm Crushed Limestone), which varied in particle size and angularity. Monotonic, cyclic, and postcyclic shear response was assessed for these materials at two relative densities and initial vertical effective stresses up to 400 kPa. Shear wave velocity was measured in each specimen so that laboratory and field conditions could be compared. Particle angularity was shown to be an important parameter that affects peak, phase transformation (PT), and ultimate state (US) response of uniform gravels. As particle angularity increased peak, PT, and US friction angles increased. Particle size was shown to have a lesser impact on these friction angles. Peak and US friction angles were dependent on relative density, while the PT friction angle was unique for each material. Results of cyclic tests on uniform gravels in this study showed that gravels will liquefy at normalized shear wave velocities ($V_{S1}$) of up to approximately 230 meters per second ($\mathrm{m}/\mathrm{s}$). Increasing particle size, angularity, and relative density led to an increase in postcyclic shear strength.

This paper presents probabilistic analyses of slopes and strip footings, with spatially variable soil modeled by the random field theory. Random fields are simulated using Latin hypercube sampling with dependence (LHSD), which is a stratified sampling technique that preserves the spatial autocorrelation characteristics. Latin hypercube sampling with dependence is coupled with polynomial chaos expansion (PCE) to approximate the probability density function of model response. The LHSD-PCE approach is applied to probabilistic slope analyses for soils with cross-correlated shear strength parameters, and is shown to be more robust than raw Monte Carlo simulations, even with much smaller numbers of model simulations. The approach is then applied to strip footing analyses with conditioned random fields of Young’s modulus and shear strength parameters, to quantify the reductions in settlement uncertainty when soil samples are available at different depths underneath the footing. The most influential sampling depth is found to vary between 0.25 and 1 times the footing width, depending on the strength mobilization and spatial correlation features. Design charts are established with practical guidelines for quick estimations of uncertainty in footing settlements.

This study examines the feasibility of using data fusion techniques to predict the overconsolidation ratio (OCR) and undrained shear strength ($s_u$) from piezocone penetration test (PCPT) measurements, in situ stresses, and additional features created from available data. Several data fusion models were developed using two feature-level fusion techniques: regression trees and model trees. Hybrid models were developed using these feature-level fusion techniques and two decision-level fusion techniques, namely, bootstrap aggregation and stacked generalization. After training and testing the data fusion models, the predicted values of OCR and $s_u$ were compared to the reference values, obtained from laboratory or in situ tests, and to the values estimated using several existing interpretation methods. The model trees, which predict continuous values, performed better than the regression trees, which predict discrete values. The decision-level fusion algorithms tended to improve the results of the regression trees, but had little effect on the results of the model trees. Overall, the data fusion models were found to perform well and tended to perform better than the corresponding interpretation methods in estimating values of OCR and $s_u$.

The present paper proposes a limit equilibrium method for the stability analysis of reinforced tunnel faces in purely cohesive soils exhibiting constant or depth-dependent undrained shear strength. The method can be applied to arbitrary reinforcement layouts in relation to bolt spacing and length. The computational predictions are compared to the results of other models and numerical stress analyses in terms of the so-called load factor. Finally, design nomograms are presented that provide a rapid assessment of the stability of uniformly reinforced faces in soils with constant or depth-dependent undrained strength.

Full-scale, controlled blasting field tests on driven displacement pile–improved ground were conducted to study the response of densified and reinforced ground to blast-induced excess pore pressures. In order to make appropriate comparisons to the baseline response of the native, unimproved ground, explosive charges sufficient to induce liquefaction were detonated in a control zone and the resulting postliquefaction settlements were measured. Excess pore pressures generated in the improved ground were observed to be significantly smaller than that in the unimproved ground, and resulted in settlements that were generally one-sixth to one-third of that measured in the unimproved ground. Piles tipped into a dense bearing layer settled significantly less than the surrounding soil and piles that were floated above the bearing layer. Importantly, measured excess pore pressures pointed to a change in soil response from contractive to dilative during blasting, indicating that the improved ground mobilized significant strength during blasting, similar to the response expected from cyclic mobility of dense soils. The energy of scaled ground motions developed from velocity measurements are used to relate the observed soil response to blasting to that expected from earthquake-induced ground motions. The paper concludes with a comparison of shear strains expected from shear strain compatibility (SSC) between the improved ground and the displacement piles to those implied by the measured pore pressures. The comparison indicated that some portions of the improved ground responded in an incompatible manner during the blast-induced ground motions and that the assumption of SSC may not be appropriate for design of some reinforcement-type ground improvements.

Three-dimensional synchrotron X-ray diffraction (3DXRD) and synchrotron microcomputed tomography (SMT) techniques were used to measure and monitor the lattice strain evolution and fracture behavior of natural Ottawa sand particles subjected to one-dimensional (1D) compression loading. The average particle-averaged lattice strain within each sand particle was measured using 3DXRD and then was used to calculate the corresponding lattice stress tensor. In addition, the evolution and mode of fracture of sand particles was investigated using high-resolution three-dimensional (3D) SMT images. The results of diffraction data analyses revealed that the major principal component of the lattice strain or stress tensor increased in most of the particles as the global applied compressive load increased until the onset of fracture. Particle fracture and subsequent rearrangements caused significant variation and fluctuations in measured lattice stress-strain values from one particle to another and from one load stage to the next load stage. SMT image analysis at the particle scale showed that cracks in fractured sand particles generally initiated and propagated along the plane that connects the two contact points. Fractured particles initially split into two or three major fragments, which in some cases was followed by disintegration into multiple smaller fragments. Microscale analysis of fractured particles showed that particle position, morphology, and the number and location of contact points played a major role in the occurrence of particle fracture in confined comminution of the sand assembly.

Prefabricated vertical (PV) drain arrays have been proposed as a minimally intrusive technique for mitigating seismically induced ground deformations in sandy liquefiable slopes. This paper describes the representation of individual PV drains as line elements within a finite-element program. The elements can represent laminar or fully turbulent discharge regimes, based on classic Darcy-Weisbach pipe flow, as well as fluid storage above the water table. Two-dimensional, plane strain simulations of coupled flow and deformation are performed using the proposed PV drain elements, with equivalent permeability properties for the surrounding soil mass, and a constitutive soil model describing the nonlinear effective stress-strain behavior of the sand. The numerical predictions are evaluated through comparisons with pore pressures and deformations measured in a centrifuge model test. The results highlight the role of the PV drains and their discharge characteristics in controlling deformation mechanisms.

Eight column tests were performed with specimens comprising unamended and zeolite-amended sand-bentonite backfills for vertical cutoff walls permeated with solutions of either 35 mM KCl (five tests) or 20 mM ${\mathrm{ZnCl}}_2$ (three tests) to evaluate the potential use of zeolite amendments for enhancing the sorption capacity of ${\mathrm{K}}^{+}$ and ${\mathrm{Zn}}^{2+}$. The column tests used flow-pump systems to maintain constant flow rates and flexible-wall cells to minimize the potential for side-wall leakage and short circuiting of solutes. However, the use of low flow rates ($2.06–8.05\mathrm{mL}/\mathrm{d}$) to prevent membrane blow-out and initial permeation with deionized water to control soluble salts contributed to extensive test durations ranging from 385 days (1.05 year) to 1,370 days (3.75 year). The column specimens were characterized by soil:solution ratios ranging from $3.3:1$ to $3.7:1$ and hydraulic conductivity, $k$, based on permeation with a salt solution that were $<1.0\times {10}^{-9}\mathrm{m}/\mathrm{s}$. Thus, the $k$ of the backfill specimens was consistent with that typically required of backfills for vertical cutoff walls used for containment of contaminated groundwater. Also, effluent electrical conductivity (EC) and pH attained the values for the influent in the case of the 35 mM KCl solution, whereas for the low pH (1.83), 20 mM ${\mathrm{ZnCl}}_2$ solution, the buffering capacity of the backfill specimens reduced the contribution of ${\mathrm{H}}^{+}$ to the EC of the effluent, resulting in a slight increase in the effluent pH ($2.80\le \mathrm{pH}\le 3.74$). Finally, delayed breakthrough of ${\mathrm{K}}^{+}$ and ${\mathrm{Zn}}^{2+}$ was consistent with preferential cation exchange of ${\mathrm{K}}^{+}$ and ${\mathrm{Zn}}^{2+}$ relative to exchangeable ${\mathrm{Na}}^{+}$ and/or ${\mathrm{Ca}}^{2+}$ initially bound to the exchange sites of the bentonite and zeolite components of the backfills.

Eight, long-term column tests (1.05–3.75 year) involving soil-bentonite (SB) backfills with 5 or 10% (by dry weight) of one of three types of zeolite amendment (clinoptilolite, chabazite-UB, chabazite-LB) and permeated with either 35 mM KCl or 20 mM ${\mathrm{ZnCl}}_2$ solutions were conducted to determine the effects of type and amount of zeolite amendment on the hydrodynamic dispersion coefficient, $D$, and the retardation factor, $R_d$, of ${\mathrm{Cl}}^{-}$, ${\mathrm{K}}^{+}$, and ${\mathrm{Zn}}^{2+}$. The superimposed effluent breakthrough curves for ${\mathrm{Cl}}^{-}$ were virtually coincidental indicating excellent reproducibility of all eight column specimens, and solute transport generally was dominated by diffusion relative to mechanical dispersion owing to low imposed seepage velocities ($1.60\times {10}^{-8}\mathrm{m}/\mathrm{s}\le v_s\le 6.88\times {10}^{-8}\mathrm{m}/\mathrm{s}$). Also, $D$ of ${\mathrm{Cl}}^{-}$ tended to be greater for the zeolite-amended backfills relative to the unamended backfill, which was attributed to the replacement of the coarse-grained sand with fine-grained zeolite amendments, leading to more tortuous pathways. The $R_d$ of ${\mathrm{Cl}}^{-}$ was essentially unity ($1.00\le R_d\le 1.03$) for seven of the eight column tests, suggesting the absence of any effective porosity. Retardation of ${\mathrm{K}}^{+}$ and ${\mathrm{Zn}}^{2+}$ correlated reasonably well with the measured cation exchange capacity of the backfills, as expected on the basis that cation exchange was the dominant attenuation (sorption) mechanism. Overall, as little as 5% zeolite amendment was effective in enhancing the sorption capacity ($R_d$) of the SB backfill by a factor ranging from 2.4 to 3.2 for ${\mathrm{K}}^{+}$ and 1.4 to 2.2 for ${\mathrm{Zn}}^{2+}$, whereas the backfill with 10% chabazite-UB amendment was 4.7 times more effective than the unamended backfill in sorbing ${\mathrm{K}}^{+}$. Thus, the column test results support the hypothesis that SB backfills for vertical cutoff walls that are amended with zeolite can significantly increase the sorption capacity for metals, thereby enhancing the containment function of the cutoff walls.

Dynamically installed anchors are torpedo-shaped anchors that are installed by dropping them through the ocean such that they self-bury in the soft seabeds typically encountered in deep water. This paper presents and considers field data from reduced scale anchor tests at two sites to validate a new release-to-rest model for dynamically installed anchors. This model extends existing studies by considering the motion of the anchor from the point of release in the water column, modeling the drag resistance that acts on the anchor and its mooring line. Simulations from the model, together with data from the field tests, demonstrate the importance of considering the free-fall in water phase of installation, because the drag developed on the mooring line can be significant for larger release heights, reducing the anchor velocity as it arrives at the seabed by up to 44%. The soil embedment phase of the model also considers drag resistance on the anchor—and optionally on the mooring line—and accounts for the effect of the high anchor velocities on soil strength using a simple power law. Model simulations are shown in the paper to be in very good agreement with the anchor motion data measured in the field tests, resulting in predicted anchor embedments that are in agreement with the field database of over 100 anchor installations to an accuracy of $\pm 10\%$.

A general closed-form solution for the reliability index (or probability of failure) of a simple linear limit-state design function with one load term and one resistance term is used to compute the resistance factor expressed in a load and resistance factor design (LRFD) format. The solution considers method bias, bias dependencies, and uncertainties in choice of nominal values of load and resistance determined as part of the project-specific design process. Uncertainty in the choice of nominal values for design is linked quantitatively to the concept of project level of understanding that has been recently adopted in Canadian design practice. All random variables are assumed to be lognormally distributed. Parametric analyses are carried out to show that ignoring possible correlations between random variables can lead to conservative (safe) values of resistance factor and in other cases to nonconservative (unsafe) values. Example LRFD calibrations are carried out using different load and resistance models for the pullout internal stability limit state of steel-reinforced soil walls together with matching bias data reported in the literature. The results demonstrate the practical influence of model type, method bias statistics including dependencies, and operational factor of safety on computed resistance factors.

A soil-shrinkage curve (SSC) can be defined as the characteristic volume–water content relation under free external stress conditions. A SSC is traditionally divided into three characteristic states: normal shrinkage when soil is saturated, residual shrinkage when soil is unsaturated and subjected to volume reduction, and zero shrinkage when soil approaches the dry state. Using the measured SSC, soil water-retention (SWR) curve, and suction stress characteristic curve of various silty and clayey soils, the traditional states are found to be not well defined and incorrect, particularly when soil approaches the dry state. A conceptual model for SSC is proposed in which SSC can be divided into four states: capillary, pendular, adsorbed, and tightly adsorbed; each is governed by one or/and two SWR mechanisms: capillary and adsorption. The shrinkage rate, defined here as the change in void ratio due to the change in moisture ratio, is found to be not universally zero in adsorptive states. It is shown that the shrinkage rate in the adsorption SWR regime is highly correlated to SWR characteristics—namely, the specific surface area—or to cation-exchange capacity, or that total amount of adsorption water follows the same exponential form with the shrinkage rate.

A case study is presented describing the effects of age (14 years) and differential settlement (≈0.3 m vertical over ≈0.4 m horizontal along a horizontal distance of ≈10 m) on the engineering properties of a soil barrier layer, a geosynthetic clay liner (GCL), and a geomembrane within a composite cover. Samples of the soil barrier layer had hydraulic conductivity below the design requirement of $5.0\times {10}^{-7}\mathrm{m}/\mathrm{s}$, except in areas that were cracked because of differential settlement. Tests showed that the geomembrane exceeded design specifications for tensile yield strength ($\ge 22.9\mathrm{kN}/\mathrm{m}$) and elongation at tensile yield ($\ge 13.0\%$), and current standard specifications for oxidative induction time ($\ge 100\min$) and stress crack resistance ($\ge 500\mathrm{h}$). Geomembrane seams also exceeded design specifications for peel strength ($\ge 15.9\mathrm{kN}/\mathrm{m}$) and shear strength ($\ge 22.9\mathrm{kN}/\mathrm{m}$). Geosynthetic clay liner samples showed a reduction in swell index relative to the as-built condition (from 27.9 to $21.0–24.5\mathrm{mL}/2\mathrm{g}$) because of cation exchange. However, all GCL samples had hydraulic conductivity below the design requirement of $4\times {10}^{-11}\mathrm{m}/\mathrm{s}$.

Climate change is expected to alter statistics of extreme events in the future. Adapting geotechnical infrastructure to a changing climate necessitates quantitative assessment of the potential climate change impacts on the performance of infrastructure. This study numerically investigates the hydromechanical response of a mechanically stabilized earth (MSE) wall constructed with marginal backfill to extreme rainfall events under a changing climate. The need for investigating the effects of extreme precipitation on marginal backfill is more pronounced because larger matric suction can be developed in such backfills. To address this need, this paper compares the performance of an MSE wall using two sets of rainfall intensity-duration-frequency (IDF) curves, denoted as baseline and projected, for the Seattle area. The baseline IDF curves are provided by the National Oceanic and Atmospheric Administration (NOAA) and currently used for design purposes, and the projected IDF curves are obtained using 20 climate model simulations of the future. The results show that use of the baseline IDFs can lead to underestimation of the wall deformation and loads carried by reinforcements. The results highlight the importance of site-specific assessments to quantify the potential impacts of climate change on the performance of current and future MSE walls. Such consideration gains even more importance considering the increasing interest in using marginal backfills in earth retaining structures due to economic and environmental considerations.

The amount of exchangeable cation (EC) and cation types play important roles in fundamental soil behavior such as mechanical, hydraulic, chemical transport, and soil water retention (SWR). A methodology is proposed for the EC quantification of individual exchangeable cations using the SWR curve (SWRC) in $\mathrm{matric}\mathrm{potential}<-100\mathrm{MPa}$. The methodology is based on a SWRC model for homoionic soils that links soil water content to the cation exchange capacity (CEC) and the type of the exchangeable cations by considering the explicit effect of cation hydration on soil water content. Using the principle of local thermodynamic equilibrium, the authors extend the SWRC model to include soils with multiple cation species, leading to the quantification of exchangeable cations in soils. A suite of natural and artificially mixed homoionic soils are used to examine the validity of the methodology. The results reveal that the predicted exchangeable cations agree with those obtained by an independent experimental method, all within $\pm 15\%$ difference in CEC, indicating the potential applicability of the methodology in quantifying exchangeable cations in soils.

Variation of the resilient modulus ($M_R$), elastic modulus ($E$), and unconfined compression strength ($q_u$) with the gravimetric water content ($w$) and soil suction ($s$) for four compacted fine-grained subgrade soils from Canada was determined from experimental studies and investigated in this paper. $M_R$ was determined from cyclic triaxial tests and $E$ and $q_u$ were derived from modified unconfined compression tests. These tests were conducted on identical specimens that were wetted or dried to achieve different values of $w$ and $s$. Soil suction was either imposed using the axis-translation technique prior to performing the tests or measured using the filter paper method after conducting the tests. The measured $M_R–w$, $E–w$, and $q_u–w$ relationships were found to exhibit similar characteristics when normalized using an approach proposed in this paper. The similarities in the normalized stiffness/shear strength–moisture content relationships found for the four compacted fine-grained soils were also corroborated by the experimental data for several other soils published in the literature. In this paper, the $M_R–w$ relationships, which are cumbersome to determine, have been successfully predicted from the easy-to-obtain $E–w$ or $q_u–w$ relationships based on such similarity.

Triaxial and oedometer tests were used to demonstrate that a critical state framework can be applied to sand–rubber mixtures of similar soil grain and rubber sizes. It described well the behavior of a crushable sand and a quartz sand with either rubber fibers or granules of a variety of quantities, from small to large strains. Together with additional oedometer tests on soils of a wider variety of gradings, the work enabled the influences of sand particle type, grading, and rubber shape to be established. The sand particle type, specifically whether the grains were weak or strong, was found to be a key factor. It affected the yield in compression, even when large quantities of rubber were added. It controlled the critical state stress ratio, except for those mixtures with the highest content of rubber fibers, as well as the stress strain behavior. Sand particle type also determined the critical state line (CSL) location in the volumetric plane for lower rubber contents, but at higher rubber contents the behavior tended to converge for the two sand types. The grading and rubber type were not found to affect the compression or swelling indices significantly, which were mainly controlled by rubber content. Gradings that had nonconvergent compression paths without added rubber tended to retain this feature with rubber. The addition of both types of rubber led to higher volumetric compression in isotropic or one-dimensional compression but reduced volumetric strain during shear, altering the shapes of the state boundary surfaces.

The data format developed by Robertson and his coworkers to assess soil behavior type with cone penetration test data using a series of concentric circles can provide an unreliable assessment in overconsolidated soils. With increasing overconsolidation, the normalized net-tip stress-normalized friction ratio, $Q_t–F_r$, data for uniform-texture, but variably overconsolidated soils follow the shape of a hyperbola as shown by the generalized soil behavior type chart developed by Schneider and his coworkers. To improve soil classification from cone penetration test/piezocone penetration test (CPT/CPTu) data, a method is proposed that plots $Q_t$ against $f_s/{\sigma }_{vo}^{\prime }$, where $f_s/{\sigma }_{vo}^{\prime }$ is sleeve friction normalized by effective vertical stress. In $Q_t–f_s/{\sigma }_{vo}^{\prime }$ space, CPT/CPTu data for a given soil yields a linear relationship with a slope designated ${\mathrm{\Delta }}_Q$. The linear $Q_t–f_s/{\sigma }_{vo}^{\prime }$ relationships from many variably overconsolidated soils converge to a common origin that is offset from zero. When plotted in $Q_t–F_r$ space, the linear ${\mathrm{\Delta }}_Q$ relationships become hyperbolas, similar to the Schneider hyperbolas, but the position of the individual hyperbolas in $Q_t–F_r$ space are defined by the offset origin location. The ${\mathrm{\Delta }}_Q$ index provides a numerical value that can be linked to soil index and engineering properties. Using ${\mathrm{\Delta }}_Q$ values computed for uniform-texture, variably overconsolidated soils, and empirical correlations between ${\mathrm{\Delta }}_Q$ and soil index properties, a CPT/CPTu assessment chart is presented to improve soil classification and the generalized $Q_t–F_r$ framework.

This paper presents results of testing good-quality piston tube samples and parallel block samples from three silt sites in Ireland and Norway. Theoretical work and laboratory strain damage tests suggest that good-quality tube samples can be obtained provided the sampling process remains undrained and the tubes have a sharp cutting edge. Laboratory tests confirm that at two of the sites the piston tube samples are effectively equivalent to the block samples. However, at the third site densification of the piston tube samples occurred and the laboratory tests produced nonconservative design parameters. Given the uncertainties of the effects of piston tube sampling of the silts, it seems prudent to assess likely disturbance effects using parallel in situ tests, for example piezocone penetration testing (CPTU) or shear wave velocity testing. The CPTU pore pressure data can be used to assess the degree of drainage which is likely to have occurred during sampling, and shear wave velocity can be used to assess possible densification.

The Becker penetration test (BPT) is a widely used tool for characterizing gravelly soils, especially for liquefaction assessment. Interpretation of BPT data is complicated by the variable energy transferred from the hammer to the drill and by the shaft resistance that develops along the drill string which generally increases with penetration depth. Existing BPT interpretation methods that use above-ground measurements have had limited success in accurately separating the shaft and tip resistance. Therefore, penetration resistance with depth cannot be reliably predicted. An instrumented Becker penetration test (iBPT) that measures acceleration and force directly behind the driving shoe in order to compute the energy delivered to the driving tip was developed and integrated into the standard closed-ended Becker drill system. The equipment and data acquisition system are described in this paper. The analysis procedure used to compute energy-normalized blow count values and produce continuous penetration resistance profiles is outlined. The energy-normalized blow count profiles generated are shown to be independent of penetration depth, predrilling depth, shaft resistance magnitude, and hammer operating conditions. In particular, the efficacy of a residual energy-based normalization scheme is demonstrated. Finally, a discussion regarding energy measurements, locked-in stress at the drill string tip, and the use of the pull-back and redrive procedure is presented.

The Becker penetration test (BPT) is a widely used tool for the characterization of gravelly soils, especially for liquefaction assessment. An instrumented Becker penetration test (iBPT) was developed and integrated into the standard, closed-ended Becker drill string. The iBPT produces a continuous profile of energy-normalized blow count values $N_{B30}$, which are computed using the acceleration and force measurements recorded directly behind the driving shoe. The $N_{B30}$ profile is repeatable and unaffected by hammer driving energy or accumulated shaft resistance. This paper presents the correlation between iBPT $N_{B30}$ and standard penetration test (SPT) $N_{60}$ values, which is necessary for performing liquefaction assessment in gravelly soils. In developing this correlation, field variability was addressed by comparing median iBPT $N_{B30}$ and SPT $N_{60}$ values from adjacent soundings over geologically consistent depth intervals. A framework was also developed to assess and, when appropriate, correct for gravel influence on measured SPT blow count values. This framework uses SPT blows per inch as well as physical evidence from SPT and adjacent Sonic samples. The correlation factor between iBPT $N_{B30}$ and SPT $N_{60}$ is shown to be a constant value of 1.8 and independent of soil type or penetration resistance magnitude.

Determination of the base contribution to total pile capacity is an important aspect in the design of end-bearing piles. For driven piles, dynamic load testing and signal matching are the predominant approach for estimating total pile capacity. This includes separating the base and shaft resistance, and differentiating between static and dynamic contributions. Despite its widespread usage, the approach is susceptible to uncertainty and nonuniqueness of solutions. The new instrumented Becker Penetration Test (iBPT) configured as a reusable test pile (RTP) is capable of directly measuring the dynamic pile response at the base and along the shaft during driving. In this paper, RTP measurements at the base are presented and used to guide the selection of models and parameters available in signal-matching methods. The direct measurements at the base depict pile driving as a steady penetration process with unload–reload cycles, and consistency of locked-in residual force between subsequent blows. The results show that when fundamental mechanical constraints are satisfied, simple existing models are adequate for capturing the measured response, and uncertainty and nonuniqueness at the base are curbed.

The state-of-the-practice uses the “simplified procedure” for evaluating liquefaction susceptibility of soils. Based on this procedure, liquefaction charts have been developed that correlate soil resistance to earthquake-induced stresses. These charts are based on case histories of past earthquakes and have proven to be useful while evaluating liquefaction susceptibility at a new site. However, these charts are inherently empirical, which makes extrapolation into regimes with insufficient data difficult. In addition, they do not inform an engineer about the effects of liquefaction. This work hypothesizes that the lower end of liquefaction charts corresponds to soils that are susceptible to unstable flow liquefaction. A numerical investigation is undertaken, the results of which support this hypothesis. This implies that if test data at a new site correspond to the lower end of liquefaction charts, then the site may be susceptible to flow liquefaction. This in turn could provide an engineer with some predictive power regarding the effects of liquefaction.

The 2010–2011 Canterbury earthquake sequence included a number of events that triggered recurrent soil liquefaction at many locations in Christchurch, New Zealand. However, the most severe liquefaction was induced by the $M_{w}7.1$ September 4, 2010, Darfield and $M_{w}6.2$ February 22, 2011, Christchurch earthquakes. The combination of well-documented liquefaction surface manifestations during multiple events, densely recorded ground motions during these events, and detailed subsurface characterization information at the selected sites provides an unprecedented opportunity to add quality case histories to the empirical soil liquefaction database. The authors have already documented and published 50 high-quality liquefaction case histories from these earthquakes using cone penetration test (CPT) data. This paper examines 46 of these case histories using shear-wave velocity (Vs) profiles derived from surface wave (SW) methods and a Christchurch-specific Vs correlation based on CPT tip resistance. The Vs profiles have been used to evaluate the two most commonly used Vs-based simplified liquefaction evaluation procedures (i.e., Andrus and Stokoe and Kayen et al.). An error index ($E_I$) has been used to quantify the overall performance of these two procedures in relation to liquefaction observations. Although the two procedures are essentially equivalent for sites with normalized Vs (i.e., Vs1) $<180\mathrm{m}/\mathrm{s}$, the Kayen et al. procedure, with 15% probability of liquefaction, provides better predictions of liquefaction triggering for sites with Vs1 greater than $180\mathrm{m}/\mathrm{s}$. Additionally, total $E_I$ values obtained using Vs profiles from surface wave testing in conjunction with the Kayen et al. procedure are lower than two other CPT-based triggering procedures but higher than the total $E_I$ value obtained using the Idriss and Boulanger CPT-based procedure.

A series of major earthquakes near Christchurch, New Zealand, produced different levels of liquefaction-induced ground failure and damage to multistory buildings in the city’s central business district. Observations of seismic performance during the 2010–2011 Canterbury earthquake sequence provide a unique data set for evaluating design procedures. Nonlinear dynamic soil-structure interaction analyses of two shallow-founded multistory buildings that were damaged by moderate liquefaction-induced settlements provide useful insights. Nonlinear effective-stress fully coupled soil-structure interaction analyses were performed using the finite-difference method with a widely used constitutive model for soil liquefaction calibrated with field and laboratory test data. The results show good agreement between observed and calculated responses of the ground and the structure during the earthquake events. Shear-induced ground deformation was a key mechanism of liquefaction-induced settlement of the structures. This mechanism was captured well in the two-dimensional analyses. Liquefaction-induced volumetric-induced ground deformation was also important and captured in the analyses. However, the influence of ejecta-induced building settlement was not captured.

The hydraulic performance of geosynthetic clay liner (GCL) overlapped seams relying on hydration of panel bentonite through a melted groove in one of its geotextiles to seal the overlap is reported for the situation where the seam is below a wrinkle in an overlying geomembrane. The amount of hydrated bentonite extruded through the groove at the seam depended on whether or not the groove melt was continuous. Flow along a seam having complete groove melt with well-hydrated bentonite was 2–3 times higher than that through a single panel of GCL below a wrinkle, but two orders of magnitude lower than that through a seam with incomplete groove melt. The effect of heat tacking on the hydraulic performance of an overlapped seam is examined. The method of heat tacking, the width of the heat-tacked zone, and its positioning at the seam were found to impact seam flow.

The seismic response of a stiff levee structure placed on soft peat foundation soil is studied using centrifuge testing. The model levee was 5.1 m tall in prototype dimensions and rested atop 6.1 m of peat. The model was shaken with a suite of earthquake ground motions scaled to various intensities. Vertical and horizontal displacement records obtained from accelerometer arrays embedded in the peat were interpreted using bilinear quadrilateral interpolation to obtain in-plane components of the Cauchy strain tensor at each time increment. A direct-simple-shear-equivalent shear strain invariant, ${\gamma }_{\mathrm{DSS},eq}$, was computed from the tensors. Values of ${\gamma }_{\mathrm{DSS},eq}$ as high as 7% were observed for input accelerations of $0.52g$. Residual excess pore pressures were mobilized at shear strain amplitudes higher than 1%, reaching maximum residual pore pressure ratios near 0.2. The observed relationship between the residual excess pore pressure ratio and mobilized shear strain agrees reasonably well with results from a laboratory simple shear testing program. Vertical strains in the levee toe region were 2.5 times larger than beneath the crest attributable to levee rocking under seismic loading. These vertical strains contribute to ${\gamma }_{\mathrm{DSS},eq}$ and therefore constitute an important demand on the underlying peat.

Results from full-scale physical experiments are reported to identify and quantify short-term lateral response around the perimeter of a modular polymer stormwater collection structure during backfilling and under applied vertical pressures equivalent to the maximum and maximum-factored burial depths. The physical simulation was such that lateral stresses developed in the side fill for stiff native trench conditions and the particular soil, compaction, and resulting lateral displacements of the side panels (i.e., soil-structure-compaction interaction). Two types of coarse-grained soil, two types of compaction, and two structure heights were examined. Tensile strains in edge PVC columns were induced by lateral bending from soil placement and lateral earth pressures. Of the conditions examined, rammer-compacted 19-mm poorly graded gravel resulted in the largest short-term edge column tensions (0.4%); however, even these values were 4.5 times smaller than the measured short-term PVC rupture strains.

This paper describes and explains the spectacular mobility of the 2014 Oso landslide, which was the cause of its fatal consequences. A geomorphic interpretation of the site conditions is used to reconstruct the landslide failure mechanism. Two numerical models are used to conduct an inverse runout analysis. The models implement a newly defined rheology appropriate for liquefied soils. It is shown that this landslide occurred in two phases, characterized by different material strengths. Although the temporal sequencing of the two phases remains somewhat ambiguous, it is clear that the distal phase underwent significant undrained strength loss (liquefaction) and travelled more than 1.4 km over a nearly horizontal surface. The proximal phase underwent brittle failure, with much less strength loss than the first phase. The parent material forming the slide mass was composed of insensitive, overconsolidated glaciolacustrine silt and clay, material not traditionally recognized as liquefiable. It is hypothesized that a substantial volume of liquefaction-prone soil was formed by colluvial softening of the parent material during the process of slope development prior to 2014.

Resonant column tests were conducted to examine the variation in shear modulus ($G$) and damping ratio ($D$) of dry sand subjected to 10 batches of vibration cycles. Each batch increment comprised 1,000 cycles. Tests were performed by varying void ratio, confining pressures, and input driving voltage. The vibration cycles cause a continuous increment in shear strain, which eventually leads to a decrease in shear modulus and an increase in damping ratio without affecting the basic variation of shear modulus and damping ratio with shear strain. The effect of each subsequent batch of 1,000 cycles reduces continuously as compared to the previous batch increment. The effect becomes more appreciable for (1) lower confining pressure and (2) greater void ratio and high input driving voltage.

In this paper, the nonlinearity of the gravelly and clayey soils in Japan is investigated based on the Kiban Kyoshin network (KiK-net) downhole array seismic observations. Shear modulus reduction ratios at 18 sites under 165 strong motions are calculated by comparing the response amplification spectra (surface response normalized by bedrock response) subjected to strong motions with those subjected to weak motions and are plotted versus the corresponding seismically induced shear strains. The back-calculated shear modulus reduction ratios are categorized into several groups under different effective confining pressures, and a best-fit line is proposed for each group as the corresponding empirical shear modulus degradation curve. The obtained results are compared with the existing studies based on both in situ and laboratory tests, highlighting the influence of different testing techniques on nonlinear soil behavior. Furthermore, a modified cyclic nonlinear model is used to formulate the proposed empirical curves to account for the shear modulus degradation under a wide range of effective confining pressure. This modified model can then be used to specify material constitutive relationships for the future numerical analysis of geotechnical structures under seismic loads.

This paper investigates the effect of uncertainty in the hydraulic conductivity and van Genuchten fitting parameters used for unsaturated flow modeling of transient seepage through embankment dams. Random variables of the parameters are generated using a number generation algorithm and are then coupled with the finite element software SEEP/W. Stochastic analyses are conducted via Monte Carlo simulation. The statistical and probabilistic properties of the seepage rate passing through different embankment dams and transient boundary conditions are assessed and the degree of uncertainty associated with the seepage is evaluated. It is found that the probabilistic properties of the seepage strongly depend upon the boundary conditions of the dam. The seepage through embankment dams is found to follow the generalized extreme value or three-parameter lognormal distributions.

A methodology for virtualizing irregularly shaped grains is described here. The principle, largely inspired by computed tomography, is simple and accessible because only the three-dimensional (3D) outline of the grain is required. The volumetric object is obtained by reconstructing the planar projections of the grain acquired at different angles of rotation using a standard camera. Depending on the lens system, the resolution of the images can be as good as a few microns. A numerical representation of the real grain can be obtained by meshing the 3D image. The influence of grain morphology on the contact behavior of quartz sand is investigated here as an application of this novel technique. Numerical simulations using a finite-element model were carried out to reproduce the experimental data from normal compression single-grain tests. The results show the contribution of the initial grain rearrangement on the normal force-displacement response and its strong dependency on the shape of the grain. This study demonstrates that particle shape is a critical parameter for calibration of the contact behavior of sand.

The widely acknowledged Mononobe-Okabe (MO) model is not recommended for seismic passive earth pressures analysis because it yields nonconservative results when the ratio of the wall’s to the soil’s frictional angle is equal to or greater than 0.5. A more advanced analytical approach, termed the log-spiral-Rankine (LSR) method, was recently introduced and is capable of producing more conservative solutions for the aforementioned scenario by considering soil cohesion, a curvilinear failure surface, and so on. Implementation of the LSR method, however, is quite complicated, as it requires several iterations to obtain the final converged solution. This study proposes an alternate way to approximate the more accurate LSR solutions from the simple MO equation for evaluation of seismic passive earth pressures in cohesive backfills. The approach uses two modification factors to account for the effects of soil cohesion and curvilinear failure surface, respectively. The overall error of the proposed equation is approximately 2%. More importantly, the new equation is easy to implement and thus suitable for routine design practice.

Geogrids have been increasingly used for stabilization of base courses and subgrade. In the design of the geogrid-stabilized roads, the benefit of geogrids is usually quantified by a modulus improvement factor (MIF) so that the geogrid-stabilized base course can be simplified as a nonstabilized base course with an equivalent modulus. In previous studies, the equivalent moduli of base courses were usually back-calculated by using the measured vertical stresses and/or resilient deformations of the road surface. However, these responses (i.e., stress or resilient deformation) fail to or cannot fully capture the benefits of the geogrid in reducing the permanent deformation of roads. In this study, the equivalent moduli of the geogrid-stabilized base courses were back-calculated by using the measured permanent deformations. Burmister’s layered elastic solution and a modified mechanistic empirical pavement design guide (MEPDG) soil damage model were used for this back-calculation. The results show that the MIFs of the geogrid-stabilized bases determined by using the measured permanent deformations were mostly higher than those determined by using the measured vertical stresses at the interface. The equivalent modulus of the geogrid-stabilized base back-calculated by using the permanent deformation captured the benefits of the geogrid at both the loading and unloading stages.

There are some attempts to evaluate Hardin’s equation for transition soils, i.e., sand mixed with a systematic increase of fines ($\text{particle size}\le 0.075\mathrm{mm}$) content, $f_c$. The most common finding is that maxmium shear modulus, $G_{\max }$, decreases with increasing $f_c$, and there are attempts to capture the effect of $f_c$ on $G_{\max }$ by considering $G_{\max }$ of clean sand as a reference. These are done by (1) developing empirical relations of $A$, $n$, $c$, or $d$ with $f_c$; (2) using equivalent granular void ratio, $e^{*}$, instead of void ratio, $e$ in Hardin’s equation; and (3) using a critical state (CS) approach. This paper presents 288 resonant column $G_{\max }$ data for clean Hostun sand and Hostun sand mixed with 5, 10, 20, and 30% nonplastic quartz powder. It is found that when $e$ in a pre-established Hardin equation for clean Hostun sand is replaced by $e^{*}$, Hardin’s equation can predict $G_{\max }$ for Hostun sand with $f_c$ with good accuracy. Only soil grading properties, e.g., $D_{10}$ and $d_{50}$, are required as inputs to convert $e$ to $e^{*}$. This is a significant advantage over a recently proposed CS approach which requires critical state lines (CSLs) data for each $f_c$.