A study is carried out of the anticipated effects on the rock surrounding a storage cavern due to the storage of water at elevated temperatures. Properties of granite and andesite at temperatures between 15 ° C and 100 ° C determined in laboratory experiments are then measured. The rock properties are compressive strength, tensile strength, tangential Young's modulus, Poisson's ratio, coefficient of thermal expansion, and thermal diffusivity. Results are determined for rock specimens prepared in both a dry and a wet condition. In general, the rock properties are adversely affected by moisture and by an increase in temperature. The effect of an increase in temperature is more significant in the wet‐rock specimens. The experimental rock properties determined were used in a numerical analysis to determine the effect of the storage of heated water at a temperature of 100 ° C in a 10‐m‐diameter circular underground opening at a depth of 100 m below the ground surface. For the conditions analyzed, the results indicated that rock surrounding the opening would not be overstressed.

Piles supporting compliant offshore structures subject the foundation soils to large cyclic shear stresses superimposed on significant static (sustained) stresses. In order to investigate the behavior of foundation clays during critical undrained storm‐loading conditions, an extensive experimental program involving $K_o$‐consolidated undrained cyclic direct simple shear tests was performed on normally consolidated samples of resedimented Boston blue clay. Results are presented in a normalized format to allow generalizations and comparisons with other clays obeying normalized behavior. A simplified method is then proposed for predicting important aspects of shaft performance for a rigid rough cylindrical pile subjected to the combined static and undrained cyclic loading conditions imposed by tension leg platforms.

Undrained cyclic shearing of normally consolidated (NC) clays develops positive excess pore pressures and causes a reduction in effective stresses. As a result, initially cycled NC clay subjected to subsequent undrained monotonic shearing behaves as an overconsolidated clay. This observation is extended herein to form the basis of the apparent overconsolidation (AOCR) hypothesis and to develop a framework for predicting the undrained cyclic behavior of slightly overconsolidated clays from the NC behavior. In addition, the AOCR framework allows predictions of clay response due to cyclic shear with variable intensities. Comparisons of predictions with results of cyclic direct simple shear tests on Boston blue clay show generally favorable agreement.

An initial design methodology is presented to estimate post‐construction reliability of soil liners, where reliability is defined as the probability that the post‐construction effective hydraulic conductivity K of a liner will be less than $1\times {10}^7\mathrm{cm}/s\text{.}$ In the initial design phase, liner hydraulic conductivity is estimated from: (1) A regression relationship between soil properties and K; (2) laboratory K‐measurements; and (3) compaction tests. A Bayesian methodology is used to combine the information from these three sources. The initial design reliability is shown to be influenced by the variability of the liner soil, laboratory measurement error, the error in the prediction relationship used, the number and nature of samples, and the dimensions of the liner. A numerical example is presented to illustrate the methodology.

A previously described finite element model for analyzing the effects of pumping and recovery in deep aquifers has been implemented on a micro‐computer. The system includes the effects of leaky aquifers, wellbore storage, skin effects at the well, and single or multiple impervious or recharge boundaries. The recovery of pressure following the cessation of flow into the well is plotted against an inverse function of time (θ). The shape of the curve can be related to values of the parameters describing the hydrogeologic properties. By iteratively modifying the parameters of the model, the engineer can develop a representation of a porous soil or rock layer that is consistent with the results of the pumping tests. Examples show how each parameter affects the calculated results. Applying the model to three sets of tests in deep rock formations shows that consistent results are obtained but that the match between the field data and the computed results is sometimes quite approximate. Nevertheless, the model gives rapid and useful insight into the hydrogeological properties of geotechnical media.

A stress‐strain model for one‐dimensional strain in normally consolidated cohesive soils is compared to test results covering the stress range 0–950 atm. The model is called the “l/e vs. ${\sigma }_{\nu }^{\prime p}$ model” because the implied relationship between void ratio e and vertical effective stress ${\sigma }_{\nu }^{\prime }$ is linear in the normally consolidated range. The slope of the linear relationship is nearly independent of soil disturbance and its variation is relatively small for a wide variety of soils with different compressibilities. Because of this the effects of aging and soil disturbance are isolated to a single model parameter, providing a quantitative measure of these effects. The liquid limit is easily related to the model because the ${\sigma }_{\nu }^{\prime }=0$ state is included. These features provide for better interpretation of test results. An undisturbed sample can be tested as usual and then remolded to a water content slightly above the liquid limit. The liquid limit should be measured as water is added and the soil remolded. The final remolded sample should then be subjected to a 1‐D strain test. The results for both undisturbed and remolded samples define the model slope and the liquid limit and remolded test results are beneficial in choosing the best value of p. By following this procedure, all three tests are conducted on identical material.

When a bored pile is formed in a soft or weak rock, the contact between the concrete of the pile and the wall of the rock socket is equivalent to a rough rock joint. When the pile is loaded, the side shear resistance develops as a function of the shear characteristics of the rough interface between the concrete and the rock. The work reported herein and in a companion paper represents one of the steps of an investigation, into the behavior of these concrete/rock joints, wherein the shear behavior of regular triangular concrete/rock joints is considered. This work presents the development of the various analytical concepts necessary to predict the shear response.

A series of analytical expressions was developed in the companion paper to describe in detail the shear behavior of regular triangular concrete/rock joints. However, in order to evaluate the relevancy and accuracy of these expressions, a detailed laboratory test program was undertaken with the conditions of each test maintained as closely as possible to the conditions of the analysis. The test series made use of a synthetic rock material, of well‐defined properties, which was formed into a joint with concrete and tested in a specially developed large direct shear machine. The results demonstrate that the analytical expressions are capable of describing the various stresses and displacements experienced by the joint for conditions of both constant normal stiffness and constant normal load.