Journal of Geotechnical and Geoenvironmental Engineering

Current reinforced earth structure designs arbitrarily distinguish between reinforced walls and slopes, that is, the batter of walls is 20° or less while in slopes it is larger than 20°. This has led to disjointed design methodologies where walls employ a lateral earth pressure approach and slopes utilize limit equilibrium analyses. The earth pressure approach used is either simplified (e.g., ignoring facing effects), approximated (e.g., considering facing effects only partially), or purely empirical. It results in selection of a geosynthetic with a long-term strength that is potentially overly conservative or, by virtue of ignoring statics, potentially unconservative. The limit equilibrium approach used in slopes deals explicitly with global equilibrium only; it is ambiguous about the load in individual layers. Presented is a simple limit equilibrium methodology to determine the unfactored global geosynthetic strength required to ensure sufficient internal stability in reinforced earth structures. This approach allows for seamless integration of the design methodologies for reinforced earth walls and slopes. The methodology that is developed accounts for the sliding resistance of the facing. The results are displayed in the form of dimensionless stability charts. Given the slope angle, the design frictional strength of the soil, and the toe resistance, the required global unfactored strength of the reinforcement can be determined using these charts. The global strength is then distributed among individual layers using three different assumed distribution functions. It is observed that, generally, the assumed distribution functions have secondary effects on the trace of the critical slip surface. The impact of the distribution function on the required global strength of reinforcement is minor and exists only when there is no toe resistance, when the slope tends to be vertical, or when the soil has low strength. Conversely, the impact of the distribution function on the maximum unfactored load in individual layers, a value which is typically used to select the geosynthetics, can result in doubling its required long-term strength.

In 2001, the Utah Department of Transportation completed a 4-year $1.4 billion I-15 reconstruction project in Salt Lake City, Utah. This project included widespread use of expanded polystyrene geofoam as lightweight embankment at important utility crossings and where close proximity to existing buildings necessitated minimizing consolidation settlement. This paper presents construction and long-term monitoring results for some of these embankments with numerical modeling of the field measurements. Fast Lagrangian Analysis of Continua, a finite-difference program, was used to estimate the complex stress distribution and the displacements (i.e., strain) that developed in select geofoam embankments. The writers used a bilinear elastic model to produce reasonable estimates of gap closure, block seating, and the subsequent elastic compression of the geofoam embankment at higher stress levels. Such estimations are important for modeling and designing geofoam embankments and potential connections with other systems. The calculation of the complex stress distribution and displacements that develops in a geofoam embankment has application to settlement, lateral earth pressure against retaining and buried walls, slope stability, and seismic design of geofoam embankments.

The results of an accelerated aging test program to evaluate the effect of thickness on the depletion of antioxidants from high-density polyethylene (HDPE) geomembranes and subsequent degradation of the physical properties are reported. Three commercially available HDPE geomembranes having nominal thicknesses of 1.5, 2.0, and 2.5 mm were examined. The geomembranes were immersed in a synthetic leachate at 85, 70, 55, and $22°\text{C}$ and tested for oxidative induction time, crystallinity, melt index (MI), tensile properties, and stress-crack resistance. The antioxidant depletion rate for the 1.5 mm geomembrane was faster than for the 2.0 and 2.5 mm geomembranes. Antioxidant depletion time was predicted at representative landfill temperatures of $20–60°\text{C}$ using Arrhenius modeling and was found to increase with geomembrane thickness for the three geomembranes examined. Based on the results of crystallinity, MI, and stress-crack resistance, the degradation of the geomembrane was slowest for the thickest geomembrane. These results suggest that a thicker geomembrane may have a longer service life (other things being equal).

Many poorly graded granular materials of engineering importance can be characterized as gap-graded binary mixtures. Such mixtures display a volume-change response at a threshold value of the coarse fraction that is reminiscent of systems described by percolation theory. An experimental investigation on a sand-clay mixture is presented that clearly displays threshold behavior and sheds light on the role that each soil fraction plays in transferring loads through the medium. There are two key effects. First, an analysis of void ratio of the interpore clay fraction for varying compaction energies reveals an abrupt reduction in clay density at the threshold fraction of sand, whereby it is virtually impossible to impart compaction on the clay fraction at sand contents exceeding this threshold. Second, although force chains cannot be observed directly, analysis of the sand in terms of its component void ratio, computed based on treating the clay as part of the void space, shows that the sand carries a majority of the load at component void ratios that are too high to form stable force chains. The traditional interrelationship between mean stress and void ratio based on critical state theory breaks down when the sand content nears its threshold fraction. When the sand content is near the threshold limit, increasing mean stress results in a greater dilative tendency. Results are compared with findings on consolidation of sand-bentonite mixtures, and so-called reverse behavior of sand-silt mixtures.

Pile groups are frequently designed with equal or similar pile lengths. However, the significant interaction effects among equal-length piles imply that this may not be the optimized configuration. This paper presents the optimization analyses of piled rafts and freestanding pile groups, where pile lengths are varied across the group to optimize the overall foundation performance. The results of the analyses are applicable in cases where the piles derive a majority of the capacity from the frictional resistance. It is demonstrated that, with the same amount of total pile material, an optimized pile length configuration can both increase the overall stiffness of the foundation and reduce the differential settlements that may cause distortion and cracking of the superstructure. The benefits of the optimization can be translated to economic and environmental savings as less material is required to attain the required level of foundation performances. The reliability of the optimization benefits in relation to construction-induced variability is also discussed.

Soil-pile interaction remains to be the most ambiguous yet one of the most crucial aspects in the design of laterally loaded soil-pile systems subjected to embankment-induced movements. This paper proposes a new method that is capable of producing soil stiffness degradation curves, which are the outcome of real field behavior through free-field measurements. Soil-pile interaction mechanism can be solved with the proposed method for any possible case either the piles are constructed before the embankment construction or during and after. For any time considered, the method enables the computation of resultant stress effects on the pile cross section and the accompanying deflections. To provide a basis of comparison, an example problem has been solved with the proposed method and with two well-known commercial finite-element softwares. Obtained results indicated the capability of the proposed method to disclose real field behavior, which can be attributed to its inherent property of being also an observational method.

In this paper, the seismic behavior of embedded cantilevered retaining walls in a coarse-grained soil is studied with a number of numerical analyses, using a nonlinear hysteretic model coupled with a Mohr-Coulomb failure criterion. Two different seismic inputs are used, consisting of acceleration time histories recorded at rock outcrops in Italy. The numerical analyses are aimed to investigate the dynamic behavior of this class of retaining walls, and to interpret this behavior with a pseudostatic approach, in order to provide guidance for design. The role of the wall stiffness on the dynamic response of the system is investigated first. Then, the seismic performance of the retaining walls under severe seismic loading is investigated, exploring the possibility of designing the system in such a way that during the earthquake the strengths of both the soil and the retaining walls are mobilized. In this way, an economic design criterion may be developed, that relies on the ductility of the system, as it is customary in the seismic design of structures.

Three-dimensional large deformation finite-element (FE) analyses were performed to investigate plate anchor capacity during vertical pullout. The remeshing and interpolation technique with small strain approach was expanded from two-dimensional to three-dimensional conditions and coupled with the FE software, ABAQUS. A modified recovery of equilibrium in patches technique was developed to map stresses after each remeshing. Continuous pullout of rectangular plate anchors was simulated and the large deformation results for strip, circular, and rectangular anchors were compared with model test data, small strain FE results, and plastic limit solutions. Interface conditions of no breakaway (bonded) and immediate breakaway (no tension) were considered at the anchor base. The effects of anchor roughness, aspect ratio, soil properties, and soil overburden pressure were investigated. It was found that the anchor roughness had minimal effect on anchor performance. For square and circular deep anchors under immediate breakaway conditions, the maximum uplift capacity increased with soil elastic modulus, which suggests that lower bound limit analysis and small strain FE analysis may overestimate the capacity. The soil beneath the anchor base separates from the anchor at a certain embedment depth near the mudline, once tensile stresses were generated. The ratio of separation depth to anchor width was found to increase linearly with the ratio of soil undrained shear strength to the product of soil effective unit weight and anchor width and was independent of the initial anchor embedment depth.

Shear band formation is an important factor in understanding failures in soil. In this paper, shear localization and shear band formation and evolution are examined using ring shear tests performed on three sands prepared by air pluviation. A transparent outer confining ring was used to visualize formation and evolution of the entire shear band. By comparing the ring shear stress paths with visual observations made during shearing, the writers show that the specimen shears uniformly over its entire height prior to shear localization. Bifurcation under constant volume and drained conditions occurs as the soil fully mobilizes its effective friction angle, and subsequent shear displacements occur only within the shear band. Consistent with previous studies, the final thickness of the observed shear band ranged from 10 to 14 times the median particle diameter. Substantial particle damage occurred within the shear band after large displacements, particularly for dilative specimens, causing additional strain-softening in contractive specimens and a second phase transformation and considerable strain-softening in dilative specimens.

A numerical model set up to simulate rapid flowlike landslide motion across three-dimensional terrain has been used to investigate the capability of various constitutive relationships to model the dynamics of complex events characterized by a changing type of substrate and morphology (e.g., glacier, bends). The numerical procedure is based on a continuum mechanics approach and on depth-averaged St. Venant equations for shallow flows. The developed RASH3D code includes the possibility of using several rheological laws, whose parameter values can vary along the runout path. Two rock avalanche cases, with some morphological peculiarities along the propagation path, have been numerically back-analyzed with both a frictional and a Voellmy rheology. Of the two considered rheologies, the Voellmy model produces the most consistent results in terms of runout area as well as velocity values. The main drawbacks of the frictional model are the tendency to predict excessive spreading of the mass and to overestimate the velocities. The results show that, when a complex problem of runout of rapid flowlike landslides has to be analyzed, it is necessary to have detailed knowledge of the geological and morphological features and to resort to increasingly complex rheologies.

During the 2002 eruption, the NW flank of Stromboli volcano (Southern Italy) experienced large landslides that produced tsunami waves with a maximum runup of 10 m in height. This paper focuses on the geomechanical behavior of the loose volcaniclastic layers, the weakest component of the deposit forming the volcano flank. Tests on the coarse-grained volcaniclastic soil and on the rock material forming the grains were conducted in dry conditions, a feature of the subaerial slope. The rock material, in spite of its high porosity, exhibits relatively high stiffness and strength due to the remarkable continuity of the solid skeleton. Accordingly, the volcaniclastic soil is characterized by high shear strength. During shear tests, however, significant grain crushing occurs, which partly explains the nonlinear shear strength envelope and the strain softening associated with contractive behavior. Grain crushing was also found to affect compressibility and its time dependency, investigated under oedometric conditions.

Sand infiltration into gravel frameworks poses significant engineering and ecological difficulties. Ten flume experiments were conducted to quantify a sand bridging threshold in a static gravel bed. The $D_{15\text{substrate}}/d_{85\text{sand}}$ ratio was computed for each of 37 unique sand-substrate pairs and the data were plotted, with previously published flume data, to determine the threshold between bridging and unimpeded static percolation. The process threshold boundary between bridging and unimpeded static percolation fell in the range of $12<D_{15\text{substrate}}/d_{85\text{sand}}<14$ , substantially higher than predicted by standard geotechnical filter criteria. Plotting the results against the critical constriction size of the substrates provided a more definitive process separation along the line $D_c^{\ast }/d_{85}=3$ .