Journal of Geotechnical and Geoenvironmental Engineering

This paper presents a new geotechnical design concept called robust geotechnical design (RGD). This new design methodology, seeking to achieve a certain level of design robustness in addition to meeting safety and cost requirements, is complementary to traditional design methods. Here a design is considered robust if the variation in the system response is insensitive to the variation in noise factors (mainly uncertain soil parameters). To aid in selection of the best design, a Pareto front that describes a tradeoff relationship between cost and robustness at a given safety level can be established using the RGD methodology. The new design methodology is illustrated with an example of drilled-shaft design for axial compression. The significance of the RGD methodology is demonstrated.

Analysis of laterally loaded piles or drilled shafts is a complex soil–structure interaction problem for which the $p–y$ method has been widely accepted as a computational tool for predicting the load-displacement behavior. This paper presents a sampling-based method that can be applied to reliability analysis of the load-deflection behavior for laterally loaded piles. Various sources of uncertainties arising from soil properties, analysis model, and $p–y$ criteria can be taken into account in the proposed method. In particular, soil properties needed as input for determining site-specific $p–y$ curves were modeled as random fields. To statistically characterize the random field of a soil parameter, the field mean, the field variance, and the correlation structure are required. The performance of a pile under lateral loading is evaluated using a different set of random samples in each realization. The probability of failure, in which failure is defined as the event of not meeting the prescribed performance criteria, is calculated as the number of failures divided by the total number of samples. A numerical example is presented at the end of paper to demonstrate practical applications of the developed sampling-based method.

This paper describes finite-element (FE) modeling, validated by large-scale tests, to simulate the lateral force versus displacement relationship of pipelines under plane-strain conditions in both dry and partially saturated sand. The FE model is based on an elastoplastic characterization of the soil, with Mohr-Coulomb strength parameters to determine the soil yield surface. Direct shear test data and strain softening models are used to represent peak and postpeak strength behavior. A methodology for defining a strain-compatible secant modulus is also presented. The analytical results are compared with numerous large-scale experimental test results, showing excellent agreement in terms of prepeak, peak, and postpeak performance. The modeling process is used to show the relationship between the maximum lateral force and pipe depth and to explain decreased dimensionless maximum lateral forces mobilized by large diameter pipes at low depth to diameter ratios in dense sand.

Rock-socketed piles are often used to transfer heavy loads from a superstructure to competent underlying rock layers. The loads are transferred by the pile to the surrounding rock mass through shaft and base resistance. Several researchers have investigated the behavior of rock-socketed drilled shafts and related the uniaxial compressive strength of intact rock to pile-shaft resistance. However, the load-transfer behavior and load-settlement response of micropiles are different from those of drilled shafts because of the large slenderness ratio ($\text{pile}\hspace{0.17em}\text{length}/\text{pile}\hspace{0.17em}\text{diameter}$) of micropiles. This study presents results from a fully instrumented field-scale load test on a 0.2-m-diameter micropile socketed 4.2 m into limestone layers (2.7 m into weathered limestone and 1.5 m into hard limestone). The results show that practically no base resistance is mobilized until the pile-head settlement reaches approximately 7% of the diameter of the test micropile. The measured limit shaft resistance values are compared with values predicted using methods available in the literature. The comparison indicates that use of the intact strength of the rock for estimation of limit shaft resistance may lead to an unconservative design when the rock mass quality is very poor, as indicated by, for example, low rock-quality designation values.

This paper addresses a three-dimensional consolidation problem through experimentation in a geotechnical centrifuge. During a discontinuous footing installation, which is typical for mobile offshore drilling rigs, periods of consolidation are encountered as part of the installation process. While this is acknowledged to expose the platforms to potentially severe consequences due to the risk of rapid uncontrolled leg penetration, to date no method has been established to predict this risk and thus enable its mitigation. The findings of this experimental study show that while the peak penetration resistance postconsolidation is mainly influenced by the duration of the consolidation period, as is expected, the load magnitude held during consolidation also significantly influences the severity of capacity reduction postpeak. The consolidation and postconsolidation behavior is discussed for various penetration depths relevant to typical jack-up installations in the field. The key findings of this study are summarized in nondimensional graphs to facilitate application in practice.

This paper presents a generalized approach for predicting the diameter of jet grout columns based on the theoretical framework of turbulent kinematic flow and soil erosion. The proposed calculation method is applicable to all conventional jet-grouting systems and takes into account the full range of operational parameters, fluid properties, soil strength, and particle size distribution, including the effect of the injection time on erosion distance. It was demonstrated that the increase in the jet grout column diameter arising from the use of a compressed air shroud in the double and triple fluid systems is approximately 27–81% for the typical range of air pressure of 0.5–1.5 MPa. The proposed method was applied to four case histories involving four variants of jet-grouting systems, i.e., single fluid, double fluid, triple fluid, and an enhanced triple fluid system. Comparison between the calculated and the measured jet grout column diameters indicated that the proposed method can produce reasonable predictions for a variety of soil conditions. It was shown that jet grout columns formed by the enhanced triple fluid system are larger than those formed by the conventional triple fluid system by approximately 36% on average. The proposed generalized approach allows all the key variables to be considered and is a useful means for the design of ground improvement by jet grouting.

The microstructure, compressibility, and shear strength of an unsaturated widely graded coarse soil were investigated through a series of laboratory tests. The results show that coarse soil has a dual-porosity structure that has both intra- and interaggregate pores. The soil-water characteristic curve (SWCC) shows bimodal features. Accordingly, the shear strength and compressibility over the entire suction range also show two distinct modes. The pore water drains from the interaggregate pores in the first mode while the intraaggregate pores remain saturated. The shear-strength behavior during this drainage period resembles that of uniform coarse sand. The apparent cohesion reaches a peak and then decreases with increasing suction, whereas the compressibility decreases first and then increases. The clay aggregates become unsaturated in the second mode as the pore water drains from the intraaggregate pores. The compressibility decreases and the apparent cohesion increases with suction as in a typical fine-grained soil. A shear-strength model is proposed for granular soils with a bimodal SWCC. The proposed equation can also be degenerated to predict the shear strength of a wide range of soils.

The merits of incorporating the spatial variability of soil properties into slope reliability analyses have been demonstrated by many researchers. However, the suitability of the use of limit equilibrium methods (LEMs) for calculating the minimum factor of safety (FS) of a spatially variable slope modeled by a random field is not well understood. The present paper investigates this topic by performing a series of analyses using circular and noncircular LEMs and a strength reduction method based on finite-element analysis (FE SRM). The LEMs are found to encounter three possible issues in random fields, including (1) noncircular slip surfaces, (2) a significant increase in the complexity of an optimization problem, and (3) failure to converge. The LEM analysis results in random fields can only be reliable when these three issues are adequately addressed. It is found that a noncircular rigorous LEM with a carefully selected interslice force function may work adequately in random fields. Moreover, the FE SRM is suitable for random field cases because of its robustness, accuracy, and versatility.

This paper presents the investigation of consolidation for radial drainage under linear time-dependent loading with varying loading-dependent coefficients of radial consolidation by using a viscoelastic approach. By extending Barron’s solution for radial consolidation of small strain sustained constant load, the convolution integral with time as the variable was used to analyze the consolidation under time-dependent loading. Four different loading rates were applied in the consolidation tests on three types of remolded clay with various plasticity indices to study the behavior of radial consolidation. The variation of the coefficient of radial consolidation versus effective stress was determined by explicitly performing falling head radial permeability tests at various loading stages in a modified consolidometer through a series of radial consolidation tests. The aforementioned relationship was converted to the coefficient of radial consolidation with the loading time of the corresponding pressure applied to the specimen for each specific loading rate test. The convolution integral of the viscoelastic theory was then applied for the prediction of consolidation settlement versus time for each loading rate test. The findings indicated that the viscoelastic settlement prediction, in terms of the average degree of radial consolidation using the loading-dependent coefficients of radial consolidation, is more consistent with the experimental results compared with those when assuming a constant coefficient of radial consolidation.

Slurry trenches are often used to impede groundwater flow or to contain soil and groundwater contamination. The stability of the slurry trench during excavation and prior to backfilling is a major concern in design. A horizontal slice method for stability analysis of slurry trenches, which allows consideration of soil stratification, is presented. Vertical force equilibrium for each slice of the sliding mass and horizontal force equilibrium for the whole sliding mass are considered on the basis of limit equilibrium. The factor of safety is obtained via the Newton–Raphson method and the critical slip surface corresponding to the minimum factor of safety is located by the pattern search method. The proposed method is verified via consideration of a previously reported problem. It is then applied to a full-scale field experiment in three-layer soils. The obtained minimum factor of safety and location of the slip surface of the sliding mass are in excellent agreement with the experimental measurements. Analyses of trenches having a weak layer at various depths demonstrate potential variations in slip surface inclination between soil layers.

A series of numerical simulations, including transient seepage analyses and stress deformation analyses, was performed to predict the behavior of a mechanically stabilized earth (MSE) wall subjected to surface-water infiltration. In this research, two mechanisms to cause the deformation because of wetting were considered: (1) the deformation induced by shear strength decreases and (2) the volumetric deformation (swell or collapse) because of wetting. The effects of low as-compacted water content and a low-quality compaction zone behind the wall face on the wall behavior were investigated. As result of the simulations, the wall deformations (face deflections and reinforced-soil settlements) and reinforcement tensions (maximum tensions) are presented at the end of construction and after periods of surface-water infiltration.

This study investigates basal heave stability of deep excavations in soft clay using FEM with reduced shear strength. Three failure and one excessive deformation case histories are utilized. Three methods of estimating factors of safety against basal heave are employed, including the convergence criterion method involving numerical convergence, the angle method of quantifying the kick-out phenomenon of the wall, and the intersection method of examining the behavior of nodal displacements. Results show that the convergence criterion method and the angle method overestimated the factors of safety (FS) against basal heave. The intersection method, on the other hand, resulted in good agreement with the field observations because the calculated factors of safety were close to unity. The intersection method was also verified by a parametric study wherein its factors of safety increased linearly with the undrained shear strength of soil.

In this study, the behavior of gold tailings is evaluated by a series of drained and undrained triaxial tests carried out in compression and extension and sheared under monotonic loading, allowing the critical state line (CSL) to be established. The observed highly nonlinear shape of the CSL under undrained loading was used to investigate the mechanics of tailings that were dictated by state transitions captured by both the state parameter and the measured values of shear-wave velocities. The definition of a zone is where state conditions may develop strain hardening (stable zone) or strain softening with high compressibility while maintaining the effective stress condition or, at the limit, a strong compressibility condition that may lead to collapse with loss of effective confinement and liquefaction. A set of original test data provides new evidence on how the small-strain shear modulus and shear strength can be coupled in the critical state framework to estimate the state parameter. This coupling is regarded as a key ingredient in characterizing the state of tailings because both the state parameter and the shear stiffness are proxies for flow liquefaction.

Resonant column procedures may be used to quantify a soil’s shear modulus degradation and damping ratio curves along the small strain ($\text{∼}{10}^{-4}\hspace{0.17em}\text{to}\hspace{0.17em}\text{∼}{10}^{-1}\%$) range. However, resonant column procedures often cannot provide measurements of very small strain ($\text{<}\text{∼}{10}^{-4}\%$) dynamic mechanical properties. The objective of this study is to determine if free-free resonance (FFR) procedures may be used to provide complementary very small strain, maximum dynamic moduli and minimum damping ratio to small strain shear moduli degradation and damping ratio curves from resonant column testing procedures. A plastic control specimen in six different free-free configurations was used to determine appropriate free-free boundary conditions for axial FFR procedures. Five cohesive soil specimens were then subjected to axial FFR and torsional fixed-free resonant column procedures to evaluate the consistency of the resulting strain-dependent modulus and damping measurements. It can be concluded that: (1) the specific free-free boundary conditions used in axial FFR testing do not significantly influence measured elastic or dissipative properties, and (2) axial FFR and torsional fixed-free resonant column procedures can be used to generate consistent and complementary shear modulus data but cannot be used to generate consistent and complementary damping ratios, based on the inherent issues involved in approximating nonlinear, microstructural sources of energy losses with macroscopic, equivalent linear models of dissipation.

In this paper, a global inversion algorithm for surface wave dispersion curve is proposed. An improved weighted average velocity (Wave) method is used as the forward model and generalized reduced gradient (GRG) method is used for minimization of misfit function. The development of the proposed algorithm is detailed. The proposed inversion algorithm is then used to invert synthetic Rayleigh wave dispersion curves of known one-dimensional (1D) soil profiles and results show that the initial soil profile population size of 10,000 is sufficient to obtain the global minimum solution. Finally, field Rayleigh wave dispersion curves are inverted to show that the 1D shear wave velocity ($V_{\mathrm{s}}$) profiles from the proposed inversion algorithm agree well with independently measured $V_{\mathrm{s}}$ data.

An ideal solution for disposal of large volumes of unwanted dredged clays is to stabilize and use them as fill materials for land reclamations. This kind of stabilized dredged fill (SDF) requires a lower cement amount ($C_m$) compared with traditional cement-treated soils. Strength behavior might be different for mixes within the inactive zone (lower $C_m$) and those within the active zone (higher $C_m$). The SDF is fully/partially within the inactive zone. A new set of unconfined compressive strength ($q_u$) data as well as $q_u$ data compiled from literature covering both lower and higher $C_m$ are analyzed. First, main parameters governing $q_u$ of cement-stabilized clays are identified: (1) cement amount ($C_m$), and (2) water content ($W$). Then, $q_u$ behaviors in the inactive zone and active zone are compared. Results indicate that a nonlinear normalized $q_u\text{-}{C}_m$ trend is commonly observed in the inactive zone whereas a linear one is observed in the active zone. In both zones, the normalized $q_u\text{-}W$ curve (or normalized $q_m{\text{-}\mathrm{C}}_m$ curve) is unique for a given soil type, regardless of $C_m$ (or $W$). The key conclusion is that $C_m$ and $W$ should be considered as separate variables in the estimation of $q_u$, rather than in a combined form such as water:cement ratio ($W/A_w$). The practical outcome is the development of an empirical formula as a function of $C_m$ and $W$ for estimation of $q_u$ of cement-stabilized clays. Validation results show the proposed formula is applicable to both inactive zone and active zone.

A hybrid subsea foundation (HSF) is defined here as the combination of shallow and deep foundations, designed such that both shallow and deep foundation elements contribute to the total foundation capacity. In this project the type of HSF investigated is a shallow mat foundation connected to short piles, with pinned connections made close to the mat corners. The motivation for this research is the potential economic benefit to using HSFs in deepwater subsea developments, where mat sizes to resist typical subsea foundation design loads (e.g., from pipeline end terminations, manifolds, and riser bases) require ever larger installation vessels with higher associated cost. This paper assesses the design feasibility of a HSF for subsea facilities in soft clay and provides recommendations for design. The focus here is on analytical and numerical modeling, although corroborating results from physical modeling are provided. Numerical and physical modeling has shown that the proposed analytically based design approach is suitably conservative, with deformations well within typical serviceability limits at the calculated capacities. Most significantly, the paper demonstrates that significant reduction in the foundation footprint can be realized by supplementing a mat with short piles at each corner.

Mineral dissolution is a common chemomechanical digenetic process in geological systems. The penetration resistance in sediments that have experienced dissolution is studied using a laboratory-scale cone penetration test device and a calibration chamber. Variables include the initial sediment density and mass fraction of soluble grains. Results show that the void ratio increases with the extent of mineral dissolution; the magnitude of the void ratio change is higher in initially dense sediments. A terminal void ratio is found for dissolution; the void ratio after dissolution will not exceed this terminal void ratio regardless of the extent of dissolution. For boundary conditions applied in this study, the terminal void ratio for dissolution corresponds to a relative density of $Dr\approx 15\%$ , which is attained when dissolution exceeds a mass fraction loss of 10%. While the tip resistance decreases after dissolution, the drop in tip resistance is most pronounced in initially dense sands. A single penetration resistance versus density trend is observed for all tests, regardless of the changes in lateral stress and fabric that soils may have experienced as a result of mineral dissolution.

A grouting technique for enzymatic calcite precipitation is evaluated. Urea and calcium salt, at various concentrations, are mixed with a concentration-fixed enzyme to obtain the optimal precipitation of ${\text{CaCO}}_3$. The optimally combined solution is injected into sand samples in small PVC cylinders. Then, the improvement in small-scale samples is observed. The combination, approved for small-scale tests, is further used for larger-scale tests. The porosity distribution within the soil is evaluated by sampling the treated sand at different locations. A precipitation ratio up to 80% can be obtained using a small amount of the enzyme. The results show that the in situ enzymatic ${\text{CaCO}}_3$ precipitation technique may be feasible for use in larger-scale applications. A multiphysics simulator that considers the calcite precipitation reaction during the transport of the solution is adopted to predict the evolution of the porosity. The predicted porosities are compared with the measured porosities. The results show that the numerical predictions can replicate the actual changes in porosity relatively well and that the numerical model should be helpful in assuming these changes caused by the precipitated ${\text{CaCO}}_3$ induced by the grouting technique examined in this work.

Stability charts for homogenous soil slopes are still routinely used in practice to determine the safety factor. A new graphical approach is employed to produce a set of stability charts that eliminates any iterative process when calculating the safety factor and can provide the type of failure mode. Stability charts are presented based on the limit equilibrium method, and effects of pore water pressure and a slow-drawdown reservoir process are considered approximately. These charts are quite simple and straightforward to use in engineering evaluation of homogeneous slopes. Two examples are presented to illustrate the use of the slope stability charts and the reliability of the limit equilibrium method, respectively. Finally, the work is extended to establish the relationships between probability of failure and the development of reservoir drawdown.

This paper presents a field and laboratory investigation of pullout resistance of steel anchors in rock. The field testing involved pullout tests of six steel anchors installed in rock by gravity grouting and pressure grouting. The laboratory testing involved unconfined axial compression tests of six grout specimens prepared in custom-made molds using the two grouting methods to investigate the effect of grouting method on the mechanical properties of grout. The pressure-grouted specimens had a higher density and compressive strength, which was a result of their denser microstructure with significantly fewer and smaller voids from air bubble inclusions than the gravity-grouted specimens. The field pullout tests suggest a progressive failure mechanism, which was manifested through the progressive elongation of the bond zone. The pressure-grouted anchors yielded higher pullout resistance than did the gravity-grouted anchors. The increase in rock-grout bond strength from pressure grouting is likely due to a combination of several factors, including the resultant higher grout strength, higher rock-grout contact stresses, and grout injected under pressure into the fractures/fissures that existed in the bedrock.

Dynamic response of a railway embankment to a high-speed train is simulated for two cases: the soft ground is improved by cast–in situ concrete pipe (PCC) piles, and the soft ground is not improved. The obtained results are compared to evaluate the effectiveness of ground improvement in mitigating embankment vibration from a high-speed train. The study shows that ground improvement significantly reduces embankment vibration at all considered train speeds ($36–360\hspace{0.17em}\text{km/h}$). The possibility of vibrational resonance when the train speed approaches the critical speed governed by the soft soil is completely excluded. However, vibrational resonance still happens when the train speed approaches the critical speed governed by the embankment material, and this suggests the following implication. Even when the soft ground of a railway embankment system has already been improved, vibrational resonance can still happen at high train speeds. Furthermore, for a given site, each ground improvement scheme results in a different change in the natural vibration properties of the system and hence a different behavior of vibrational resonance. Therefore, besides design issues concerning stability and settlement of the embankment system under static loads, its vibrational resonance behavior is another issue of concern that should be carefully evaluated because it is associated with the operational safety of high-speed trains.