The foundation soil of an offshore Arctic gravity structure is subjected to complex loading conditions. In addition to radial shear stresses induced by the weight of a structure, large horizontal shear forces from ice loads can produce large rotations and even reversal in the direction of horizontal shear stresses acting on soil elements within the foundation soil. The Multidirectional Direct Simple Shear (MDSS) apparatus was developed to simulate, at the element level, these states of stress. The MDSS first consolidates a circular specimen under both a vertical stress (σ′_vc) and a horizontal shear stress (τ_hc=τ₁; setdown conditions). The specimen is then sheared undrained by applying a second independent horizontal shear stress (τ₂; ice loading) at an angle θ relative to the horizontal shear stress τ₁. The experimental program of MDSS and Geonor Direct Simple Shear tests on normally consolidated Boston Blue Clay used τ_hc=τ₁= 0.2σ′_vc and varied θ in 30° increments from zero (shear in the same direction) to 180° (shear in the opposite direction). The peak undrained resistance decreased almost twofold from θ= 0° to θ= 120°, and the deformation behavior varied from very brittle at low θ angles to becoming ductile at higher angles. These dramatic changes in foundation response are an important design issue.

The paper describes results from 25 K _o consolidated-undrained triaxial compression tests on resedimented Boston blue clay using a computer-automated triaxial apparatus with lubricated end platens and a midheight pore-pressure measurement device. Specimens were consolidated to four overconsolidation ratios (OCR = 1, 2, 4, or 8), and for each OCR, undrained shear was performed using four axial strain rates (0.05%, 0.5%, 5%, and 50%/h). The results show that the undrained strength ( s _u ) rate sensitivity (percent increase in s _u per log cycle strain rate) across the two fastest strain rates does not vary with OCR and equals about 9%. However, across the three slower rates, increases in OCR cause a consistent decrease in the rate sensitivity that reaches zero at OCR = 8. For high OCR clay, increases in s _u (if they occur) are caused by lower shear-induced pore pressures since the effective stress envelope at the peak strength does not vary with strain rate. For low OCR clay, increases in s _u are caused by both lower shear-induced pore pressures and increases in the mobilized friction angle. A simple technique is proposed for modeling the rate dependent stress-strain curves of overconsolidated clay.

The results of an experimental study of granular materials at high pressures are presented. Axisymmetric specimens of dense Cambria sand were tested in drained triaxial compression tests between confining pressures of 0.05 and 52 MPa. Drained triaxial extension tests between confining pressures of 0.25 and 52 MPa were also performed. As the confining pressure is increased, it was found that the stress-strain curve, and volumetric and axial strains to failure rapidly increased at a certain stress magnitude. This was shown to be directly related to a rapid increase in particle crushing. Beyond a certain higher value of stress magnitude, the stress-strain curves steepen, and the volumetric and axial strains to failure decrease. This was also shown to be directly related to the cessation of particle crushing and is typified by a zero rate of volume change at failure. The Mohr-Coulomb secant friction angle was found to be related to the rate of volume change at failure, whether the soil was dilatant or highly contractive and subject to large amounts of particle crushing.

The results from an experimental study of undrained granular material behavior at high pressures are presented. Axisymmetric specimens of dense Cambria sand were tested in undrained triaxial compression tests between initial effective confining pressures of 6.4 and 68.9 MPa. Undrained triaxial extension tests between initial effective confining pressures of 12.0 and 52.0 MPa were also performed. Uniform strains in extension specimens were enforced by creating a rigid jacket around the specimen, which prevented necking, but allowed unimpeded deformations in the specimen. It consisted of two layers of spaced steel plates, which were bent to fit the contours of the specimen. The layers of plates are separated by greased rubber membranes. The high confining pressures cause large amounts of particle crushing to occur, which in turn cause the development of large positive pore pressures. This caused the effective stress path to move rapidly toward the effective stress failure envelope. The Mohr-Coulomb, effective stress, secant friction angles in both compression and extension appear to be approximately the same. The applicability of critical state soil mechanics at high pressures is explored.

A method has been developed to predict the load-displacement relationship for single piles subjected to lateral load, embedded in sands, considering soil nonlinearity using subgrade reaction. This is the first systematic study to develop lateral load-deflection curves using modulus degradation with strain. Based on the analysis of 14 full-scale lateral pile-load tests, an empirical equation of modulus degradation with strain and a range of k _hmax values, i.e., k _h at 0.002 strain at 1 m (3.28 ft) depth, has also been proposed. An example problem is included to demonstrate its use. Based on this method, the computed load deflections have been compared with the observed ones and with those based on p - y analysis. This method shows promise over the p - y solutions and predicts upper- and lower-bound load-deflection curves, which are valuable guides to making informed engineering decisions.

In many existing elastoplastic plane-strain solutions for a circular tunnel excavated in a homogeneous, isotropic rock mass under a hydrostatic stress field, the longitudinal axial stress (or the out-of-plane stress) σ_z is not correctly evaluated. This paper describes the concepts and a numerical method to calculate this stress component and proposes analytical solutions for practical applications. The computed numerical solutions are compared with theoretical solutions where possible. The relations between rock-mass dilatancy, tunnel displacements, and the in-situ stress field are investigated. Mohr-Coulomb and Hoek-Brown yield criteria with appropriate plastic potential functions and nonassociated flow rules are used in the brittle plastic calculations. The effects of various boundary conditions on the numerical analyses are discussed and several important practical conclusions are drawn from the investigations.

A one-dimensional testing apparatus was developed to test soils to axial stresses up to 850 MPa. The apparatus was instrumented with strain gauges such that lateral soil stresses, and therefore K₀ , could be inferred from measured circumferential strains. Three different initial densities of quartz, Cambria, and gypsum sands were tested and it was found that the effect of initial density was eliminated at high stress magnitudes. This stress magnitude was higher for mineralogically harder grains than for softer grains. The inferred values of K₀ for the mineralogically harder Cambria sand was found to be constant at high pressures, but slightly below that indicated by Jaky's equation. However, the softer gypsum sand indicated increasing values of K₀ as the stress magnitude increased. This apparently was caused by inelastic, viscous flow during shearing.

A simplified block analysis to evaluate yield accelerations of a refuse mass founded on a weak interface is presented. Comparison with results obtained from conventional slope stability evaluations indicate excellent agreement. The simplified solution reduces independent variables to interface friction angle, backfill slope, face slope, and height and length of the refuse mass. Design charts are presented for two typical landfill configurations; backslopes of 2 and 4 horizontal (h) to 1 vertical (v) and face slopes of 2h:1v during staged construction.