The results of two‐dimensional (axisymmetric) finite element analyses of a clay sea bed located in the Canadian Beaufort Sea are described herein. The sea bed is subjected to loads associated with excavation, construction of a sand island, and placement and infilling of a caisson. This construction process is numerically simulated using an”incremental construction approach. The sea bed is characterized using the bounding surface model for isotropic cohesive soils. Values for the parameters associated with the model are determined by matching laboratory data for similar soils. The initial state of the sea‐bed soil and the construction/loading time history employed are, however, obtained directly from field data. The numerical predictions generated agreed quite favorably with in situ measurements. In addition to providing information concerning the response of the sea bed to the construction of a sand island, the present study provides additional evidence that validates the bounding surface model for cohesive soils.

A data base of 70 loading tests on expanded‐base piles with compacted concrete shafts has been collected to develop empirical design correlations based on standard penetration N‐values. Both shaft and point resistances, $f_s$ and $q_p\text{,}$ were found to increase with depth and with increasing N‐value. It appears that limiting values of $f_s$ and $q_p$ might be approached for initially loose sands at D/B ($D_b/B_b$) ratios greater than 12; however, at greater relative densities, both $f_s$ and $q_p$ continue to increase with depth, but at a decreasing rate, indicating an important scale effect. A modification to traditional bearing capacity theory is used to show that the so‐called scale effect arises because the bearing capacity factor $N_q\text{,}$ which is a function of the friction angle of the soil, decreases with depth or increasing effective overburden pressure; consequently, point resistance increases with depth at a decreasing rate. About 40% of the data base was used as an independent means of evaluating the accuracy of the empirical correlations developed.

Results of a laboratory experimental program conducted to evaluate three different criteria for the selection of the required screen slot width for gravel pack relief wells are presented. A widely graded filter is used in conjunction with continuous slot screen sections based on the $D_{20}$ size and the $D_{50}$ size of the filter. A third test involves a uniform filter with a continuous slot screen section based on the $D_{10}$ size of the filter. The two filters are compared in terms of (1) Entrance head losses; (2) amount of material entering the screen; and (3) effective radius of the well. The investigation indicates that all three systems result in minimal entrance head losses at the screen. The largest loss of filter material occurs in the well designed with the $D_{50}$ size of the filter. A larger effective well radius is obtained using the widely graded filter. The conclusion applies to conditions shortly after the well is installed and does not consider changes in entrance losses after clogging of well screens and filters that may take place with time.

Stress waves generated by local excitation from machines travel away from a source and eventually are absorbed in the earth. Centrifuge models are plagued by the fact that such waves reflect off the sides of the sample container and distort the desired experimental conditions. The use of Duxseal in treating the boundaries of the container is tested and evaluated as to its efficacy in alleviating the reflection problem for a class of foundation vibration problems. The results are positive for small‐size sand centrifuge models subject to free surface vertical excitation, especially if the wall coverings are grooved. The prototype size corresponds to a depth and radius of 25–50 ft. A formula for the influence of a rigid substratum on natural frequency of vertical oscillation of a circular foundation is found.

A theoretical solution for the subcritical growth of a crack in a quasi‐brittle solid is applied to predict the time for progressive failure to develop in a heavily overconsolidated clay slope. The solution utilizes the principles of delayed fracture in a linear viscoelastic plastic quasi‐brittle material to determine the initiation time and rate of propagation of the failure surface in terms of the loading, material properties, and boundary conditions. A hypothetical example of a slope 17 m high in an overconsolidated soil, using the actual properties of an overconsolidated soil, is given to illustrate the method of solution. An excavation at the toe of the slope causes an overstressed zone from which a crack is assumed to be initiated along the failure surface. The initiation time and the rate of crack propagation are in accordance with the theory of fracture mechanics, in which creep plays a dominant role. The residual shear strength is used along the crack surface and the peak shear strength is used along the remainder of the failure surface to determine the factor of safety against slope‐stability failure.

A constitutive model for the stress‐strain‐time behavior of cohesive soils is developed using Cam‐clay plasticity theory extended to include time‐dependent effects. The model adopts the concept of separating the total deformation into immediate and delayed components. The immediate plastic deformation is evaluated by employing the associative flow rule on each of two distinct yield surfaces defined by the ellipsoid of the modified Cam‐clay theory and the Von Mises cylinder inscribed in the Cam‐clay ellipsoid. The delayed component of deformation is evaluated by employing the normality rule on equivalent ellipsoidal and cylindrical yield surfaces associated with the current state of stress of the soil and forcing the resulting creep strain rate tensor to satisfy phenomenological creep laws. In a companion paper by the same writers, the resulting constitutive equation is shown to predict the stress‐strain‐time behavior of wet clays more accurately than an earlier version based on a single‐yield‐surface formulation.

A double‐yield‐surface constitutive model for the stress‐strain‐time behavior of cohesive soils is implemented into a nonlinear finite element program based on Biot's three‐dimensional consolidation theory. The coupled soil deformation‐fluid flow model allows creep effects to be modeled concurrently. The soil's hydraulic conductivity is considered a state variable, which varies with the void ratio so that it decreases as the soil compacts. Parametric, laboratory, and field‐case studies are performed on various cohesive soils such as Weald clay, undisturbed bay mud, and Boston blue clay to validate the model. Numerical simulations include drained, undrained, consolidation, creep, stress‐relaxation, and combined stress‐relaxation and creep tests under triaxial and plane‐strain stress conditions. The constitutive model is shown to predict the stress‐strain‐time behavior of “wet” clays more accurately than did an earlier version based on a single‐yield‐surface criterion.

Fill in slope areas is generally placed and compacted near optimum moisture, which is often well below saturation. At optimum moisture content, compacted clayey fill has high shear-strength characteristics because of negative pore pressures. After construction of the slope, additional moisture is introduced into the fill by irrigation, rainfall, or other sources. As water infiltrates the soil, the slope softens as the pore spaces fill with water and the pore pressures tend toward zero. The results of this study indicate that compacted clays are most susceptible to the slope-softening process while granular soils are relatively unaffected. If the factor of safety of a compacted clay slope is overestimated, as is commonly the case with surficial instability, then significant deformation of the slope will occur as moisture infiltrates the soil. The requirements for building setbacks should take soil types into account, with building setbacks being greater for compacted clays than compacted granular soils.