Journal of Geotechnical and Geoenvironmental Engineering

Compacted and undisturbed expansive soils have widely been modeled as dual-porosity materials with two levels of pore systems: micropores and macropores. The micropores accommodate the soil activity of swelling and shrinkage behavior. However, the macropores experience a volume change mainly resulting from the structural rearrangement. Due to the deformation of the macropores, a plastic volumetric strain takes place. However, after several wetting–drying cycles, an equilibrium state is reached. Beyond the equilibrium, the volume change becomes elastic and mainly is led by the micropores. This study investigated the volume change characteristics of expansive soils from the viewpoint of the soil shrinkage curve. The slope of the normal shrinkage curve at equilibrium was found to be parallel to the saturation line. Based on these observations and laboratory test results, a conceptual framework for the interrelationship between the virgin line of the soil-water characteristic curve and the reversible volume change of the micropores is presented. Experimental data demonstrated the applicability of the presented framework to slurry and compacted soils.

Despite the widely accepted notion that water potential is the fundamental variable for describing soil–water interactions in soil under unsaturated conditions, it is unable to describe several basic soil properties and behaviors such as pore-water pressure, soil water density, and phase changes of soil water. A variable with greater explanatory power, sorptive potential, is conceptualized as the origin of matric potential and pore-water pressure. The sorptive potential is the sum of the locally varying electromagnetic potentials comprising van der Waals attraction, electrical double-layer repulsion, and surface and cation hydration. Local thermodynamic energy equilibrium dictates that the sorptive potential is always transformed or equal to matric potential minus the pressure potential within a representative elementary volume of matric potential. Limited verification was demonstrated by reducing the sorptive potential to the two well-established concepts of disjoining pressure and osmotic swelling pressure. A parametric study was conducted to illustrate how soil and pore-fluid properties affect sorptive and pressure potentials, indicating that the pore-water pressure under unsaturated conditions can be as high as 0.6 GPa. Such locally high pore-water pressure is induced by the sorptive potential and provides an explanation for phenomena such as abnormally high soil water density, supercooling, and decreased cavitation observed in fine-grained soils.

In this paper, an elastoplastic two-stage analysis method is proposed to model tunneling-induced soil-structure interaction and incorporated into a computer program called Analysis of Structural Response to Excavation (ASRE). This solution allows considering both vertical and horizontal greenfield ground movements, gap formation and slippage, continuous or isolated foundations, and a variety of structural configurations and loading conditions. After introducing the proposed formulation, the model predictions are first compared with previously published data for validation. Then, to isolate the effects of various structural characteristics (relative beam-column stiffness, presence of a ground-level slab, column height, and number of stories) and foundation types (continuous versus isolated), several example structures are analyzed. Results demonstrate the value of the proposed analysis method to study a broad range of building characteristics very quickly, and show how the soil-structure interaction occurring due to underground excavations is altered by both foundation and superstructure configurations. In particular, the difference in behavior between equivalent simple beams and framed structures on separated footings is clarified.

A direct axisymmetric cone-penetration model developed for use with a user-written implementation of the MIT-S1 constitutive model is presented. The penetration model uses a finite-difference program with an Arbitrary Lagrangian Eulerian algorithm that couples the program’s large-deformation Lagrangian formulation with user-written algorithms for rezoning and second-order Eulerian advection remapping. Numerical examples illustrate the performance of the remapping and advection algorithms and cone-penetration simulations. Cone penetration at a Boston blue clay site is simulated with the Mohr-Coulomb, modified Cam clay, and MIT-S1 constitutive models and compared with measured cone-penetration test profiles. Single-element simulations illustrate that the MIT-S1 constitutive model captures the significant undrained shear-strength anisotropy exhibited by Boston blue clay, whereas the modified Cam clay and Mohr-Coulomb models do not. Penetration simulations demonstrate the important effect of undrained shear-strength anisotropy on the cone tip resistance, as well as on stress and pore pressure fields around the cone tip and rod.

The sleeper/ballast interface is an essential stage in the transfer of train loads from a railway track structure into the ground. Generally, only a small number of ballast grains support the sleeper base. The resulting localized contact stresses can be very high, especially for modern concrete sleepers on hard igneous ballast. This may result in damage to both sleepers and ballast and reduce the stability of the interface. This paper presents results from cyclic loading tests carried out to explore the potential for performance improvement through the adoption of different sleeper types and modifications to the sleeper/ballast interface. Measurements of resilient performance, plastic settlement, sleeper/ballast contact number and area, shoulder movement, ballast breakage and attrition, and the development of ballast longitudinal pressure are used to explore and explain the effect of each intervention. It is shown that twin-block sleeper types and under-sleeper pads (USP) have the potential to reduce maintenance requirements and whole-life costs.

Effects of hydration fluid and long-term field exposure on moisture-suction relationships for three types of needlepunched geosynthetic clay liners (GCLs) containing granular sodium bentonite were investigated using the pressure plate, filter paper, and relative humidity methods over a wide range of suction (0–400,000 kPa). Tests were conducted on virgin GCLs hydrated with deionized (DI) water, tap water, and a 0.07 M ${\mathrm{CaCl}}_2$ solution as well as a GCL exhumed from a landfill cover test plot 9 years after installation. Hysteresis in the moisture-suction relationship decreased progressively for GCLs hydrated with DI water (maximum hysteresis) to tap water to the ${\mathrm{CaCl}}_2$ solution to the field exhumed conditions (minimum hysteresis), with the drying and wetting paths of the moisture-suction relationship converging for the tests with ${\mathrm{CaCl}}_2$ solution and the exhumed GCL. The liquid limit, plasticity index, swell index, bulk void ratio, and water content at zero suction decreased as the hydration condition varied from DI and tap water to the ${\mathrm{CaCl}}_2$ solution and the field exhumed specimen. Both crystalline and osmotic swelling occurred in the GCLs hydrated with DI and tap water, whereas only crystalline swelling occurred in the GCLs hydrated with ${\mathrm{CaCl}}_2$ and in the field-exposed GCL. The field-exhumed GCL exhibited nearly complete cation exchange. Mole fractions of ${\mathrm{Na}}^{+}$ bound to the bentonite decreased from 61% in DI water to less than 1% for the exhumed GCL.

Laboratory and field-scale tests were conducted to determine the saturated hydraulic conductivity ($K_s$) of fresh and degraded municipal solid waste (MSW) at multiple dry unit weights and void ratios ($e$). Field-scale $K_s$ was obtained from numerical inversion of data from the Deer Track Bioreactor Experiment (DTBE). Reasonable prediction of field-$K_s$ can be achieved from large-scale laboratory tests if laboratory specimens are prepared from MSW with composition and dry unit weight similar to field conditions. In both the laboratory and field, $K_s$ of MSW decreased with increasing dry unit weight as the voids become smaller. At constant dry unit weight, more decomposed MSW had higher void ratio and $K_s$. The relationship between $K_s$ and void ratio of MSW was consistent for a range of decomposition states, suggesting that void ratio is a better indicator of $K_s$ than dry unit weight. An empirical relationship between $K_s$ and void ratio was developed that predicted the measured $K_s$ of MSW in this study within one order of magnitude. At field scale, however, the effects of decomposition on $K_s$ of MSW are more complex because decomposition and settlement occur simultaneously and have opposite effects on void ratio, complicating prediction of $K_s$.

A 27 m deep excavation in a pyroclastic sand has been constructed in the urban area of Napoli (Italy). The main excavation is supported by T-section reinforced concrete (RC) diaphragm walls. The presence of valuable historical edifices on one side of the excavations suggested the selection of a top-down construction procedure. Maximum settlement recorded at the building façade was about 25 mm [i.e., 0.09% of the free excavation depth (He)]. Full settlement troughs were observed along three different alignments. As a matter of fact none of the monitored alignments could be interpreted exactly as a normal or a parallel settlement trough. A shape function based on the complementary error function was thus proposed to fit the observed settlement data in the plane considering both the distances x and y respectively parallel and normal to the excavation. This approach is actually a combination of existing independent methods.

Deep foundations may be subjected to combined torsional and lateral loads. However, there are few full-scale observations to inform a complete understanding of the possible foundation response. This paper describes the generation and transfer of lateral loads that developed within drilled shaft foundations subjected to quasi-static and cyclic torsional loads. Due to differences in stratigraphy, one drilled shaft experienced geotechnical torsional failure, whereas the other did not. Torsional loading to failure produced shear cracks that softened the soil–structure interface and produced a softened response to the lateral load that developed. Cyclic loading resulted in increased lateral displacement at near-constant lateral load amplitude. The maximum bending moment observed in the shaft that experienced torsional failure was significantly greater than that in the shaft in which geotechnical failure did not occur, and greater than that anticipated from lateral loading alone.