The Cam‐clay models, or any other plasticity‐based models, do not make distinction between the mode of stress transfer in coarse‐ and fine‐grained soils. An examination of behavior at micro level in fine‐grained soils, from the consideration of load transfer through physico‐chemical interactions, suggests that the plastic compressions result from the grouping of particles into larger clusters and that elastic compressions result from the decrease in the spacing between particles. During shearing, these clusters gradually get dismembered, releasing the locked‐in energy. The effect of such dismembering of clusters can be easily incorporated into the original Cam‐clay model, and better predictions can be obtained with the associated flow rule, itself, for both normally and overconsolidated states. The method essentially defines the hardening of yield surfaces with internal changes in the spacing between particles, instead of changes in externally observed plastic strains. The approach describes the behavior of overconsolidated soils as yielding along successively hardening Roscoe surfaces with gradually varying plastic properties.

Slope failures on hillside slopes are frequently caused by increases in pore‐water pressures following periods of heavy rainfall. This paper presents a simplified lumped‐parameter model for estimating piezometric levels due to infiltration of rainfall. The model requires a minimal number of parameters and is applicable to hillside slopes with a layer of soil over an impermeable base. Individual components are built in the model for evapotranspiration, infiltration, and drainage due to gravity flow. The model is used to estimate piezometric levels of small basins in the watersheds of Kennel Creek, Alaska, and Perkins Creek, Oregon. The estimated piezometric levels are compared with observed values, and the agreement is satisfactory. An uncertainty model based on the first‐order, second‐moment method is used to estimate the uncertainty about the moisture contents of the unsaturated zone and piezometric levels in terms of uncertainties about the input parameters. Bayesian updating is formulated and the observed piezometric levels are used to update the parameters.

A microplane model, in which the constitutive properties are characterized independently on planes of various orientations (microplanes), is presented. It is found that the basic scheme previously developed for concrete is also valid for clays, but with certain modifications. The micro‐macro constraint is kinematic, and the stresses on each microplane are defined as explicit functions of the volumetric and deviatoric normal and shear components of the macroscopic strain tensor on the microplane. To account for undrained behavior, a pore water pressure term that affects only the volumetric equations is introduced. This makes possible the uncoupling of the stress‐strain and the pore water pressure formulations. The model is calibrated and verified by comparisons with numerous data for both drained and undrained tests, and good agreement is attained, including volume changes, pore water pressure evolution, and various stress‐strain diagrams. Although the model involves nine material parameters, four (or five in some cases) can be fixed constant for all soils, and only five (or four) need to be determined by data fitting. The fact that the stress is given as an explicit function of strain makes the model suitable for finite element codes.

An anisotropic hardening plasticity model based on the concept of bounding‐surface plasticity is formulated for prediction of the three‐dimensional stress‐strain‐volume‐change behavior of sand. The model employs the anisotropic invariants of the second‐order stress and rotational tensors as a formalism to account for the material anisotropy. The rotational tensor is linked to the fabric ellipsoid of particulate sand. The model is shown to satisfy the objectivity requirement from the form‐invariance principle. Twelve material constants required in the model can be determined from triaxial experiments. A study of the three‐dimensional behavior of Hostun sand serves as a basis for evaluating the capability of the model. Overall, the model's predictions are acceptable and can capture the variations of soil behavior with the rotation of principal stress direction and the relative magnitude of intermediate principal stress.

Analyses and performance of groups of step‐tapered piles located adjacent to a 50‐ft‐deep tieback excavation are described. The excavation is made through primarily granular soils within an existing framed structure. The main columns are supported by groups of 70‐ft‐long, unreinforced or lightly reinforced concrete piles. The temporary tieback sheet‐pile wall is located as close as 2 ft to the pile caps. Field observations, including lateral deformations of the sheet pile and lateral and vertical deformations of the main columns, reveal movements of several of the pile caps as large as 3 in. toward the excavation. This movement is twice as large as normally expected for excavations made under similar conditions. An analysis of the existing foundation conditions and a finite element computation of the bending moments induced in the piles during construction of the cut indicate that the resulting moments are not large enough to cause cracking in the piles. Sheet‐pile extraction as close as 2 ft to the main‐column pile caps cause the pile caps to translate toward the excavation by an amount no larger than one‐half the displaced width of the sheet pile; it does not significantly alter the magnitude or distribution of computed, excavation‐induced bending moments in the piles.

Realizing the inadequacy of the classical Winkler model to represent an elastic continuum, Vlasov and Leont'ev developed a two‐parameter model for a beam on an elastic foundation. The Vlasov model required an estimation of a parameter, 7, that controls the decay of stresses in the continuum. Vallabhan and Das developed a unique iterative technique to determine a consistent value of the y parameter. Their model is called a “modified” or “refined Vlasov” model. Here, the finite different technique is employed to solve the fourth‐order differential equations. The results of sample problems obtained by using the modified Vlasov model are compared with more sophisticated finite element solutions.

The primary factors governing the expansion of a compacted clayey soil are the amount and activity of the clay particles. In many cases, the compacted soil may contain gravel-sized particles that, by themselves, will not expand in the presence of water. For swell tests on gravelly clay at a constant matrix dry density, normalized swell data are plotted near the volume relationship. This indicated that the swell behavior of the compacted gravelly clay was based solely on a reduction in volume caused by the presence of gravel. For swell tests on gravelly clay at a constant compactive effort, the normalized swell data approached the weight relationship. The gravel particles interfered with compaction, which caused a lower dry density of the soil matrix as the gravel content increased. When predicting the percentage of swell of a gravelly clay, the results of this study indicate that swell depends on whether the soil is compacted to a constant matrix dry density, or whether it is compacted by a constant compactive effort.

Reinforced concrete cantilever retaining walls bearing on soil usually have a shear key incorporated into their footing base. The intended function of the key is to increase the sliding resistance of the wall. A study of the efficiency of shear keys and their location on the footing (toe, stem, or heel) was performed. A finite element model of a 20-ft-high (6-m) wall bearing on sand was used, and the results indicated that the presence of a shear key or sloped-footing bottom did not significantly improve performance over a simple flat-bottom footing. The reason is believed to be that actually measured soil-concrete shear parameters were used in the analysis, while more conservative shear parameters are used in routine design.