The causes of geotechnical engineering failures are listed and discussed. Many of these could have been prevented or damages mitigated if redundancy had been used in design and construction. Redundancy is defined in special terms applicable to geotechnical engineering. How redundancy can and should be used in preliminary exploration, soil borings, laboratory testing, analysis, design, and construction is discussed. Eight cases of failure are discussed and analyzed in which lack of redundancy led to the failures. In each case it is shown how redundancy in at least one of the forms discussed could have prevented failure.

Five model tests were performed to qualify the concept of a monobase gravity platform for the Troll Field in the North Sea. These tests were designed to simulate the conditions for a platform situated in deep water $(\simeq 330m)$ and on soft clay. The program included one static, one static after cyclic, and three cyclic model tests. A description of the procedures and selected test results are presented. The results provide a basis for a critical evaluation of the foundation design procedures presently used for offshore gravity structures in the North Sea. The results indicate that cyclic loading results in a reduced static bearing capacity, cyclic bearing capacity is lower than static bearing capacity, the cyclic stiffness, decreases with number of cycles, and for the relatively large moment arms employed, cyclic rotation is the dominant displacement mode. The results also show that for the conditions of these tests, design storm compositions with the largest cyclic load first, followed by decreasing cyclic loads, may be more critical than the traditional design storm composition with increasing cyclic loads (largest load last).

The paper presents the interpretation of five model tests on soft clay performed to qualify the concept of a monobase gravity platform for the Troll Field in the North Sea. The test program contained one static, one static after cyclic, and three cyclic model tests. The cyclic tests were run with different moment arms and with widely different cyclic load history compositions. In general, the interpretation supported the calculation procedures that are presently used in foundation design of offshore deep‐water gravity platforms on soft clay. The calculated bearing capacities, failure surface locations, and failure modes (large cyclic rotations) agreed well with the measurements for all the model tests. The calculated static and cyclic displacements also agreed reasonably well with the measurements. The deviations that did occur indicated smaller calculated displacements than measured. This should be kept in mind in design since it implies that the calculations may underestimate the displacements. However, the deviations primarily occurred at small and moderate loads and may be explained by small imperfections in the model tests (false deformations). Both the calculated and measured cyclically induced settlements were small for the vertical static load applied in these model tests.

The paper provides a methodology to estimate the spatial variation of potentiometric surfaces, allowing for possible uncertainties such as instrument errors, human errors, and statistical errors. The spatial variation is expressed in terms of the mean estimation, the coefficient of variation of the estimation, and the variance of the data scatter. The methodology is illustrated by a case study. The paper provides the criteria to decide the best estimation model when a number of theoretically consistent models are available to estimate spatial variation. A procedure to detect anomalous data is discussed in the paper through an example. The paper also provides a procedure to estimate the confidence of the mean spatial variation estimation, and describes steps one should take to improve this confidence, if it is needed. A technique is also given to estimate the random‐error component of the measurements.

The ultimate oblique pull of bored piles in sand was examined by a field testing program. Tests included axial uplift tests, lateral load tests, and oblique tests at an angle of 30° with the pile axis. The piles used were 101 mm in diameter by 1.5 m long and were installed in predominantly fine, medium‐dense sand. Lateral failure occurred in the oblique tests of the present program; however, both the axial and lateral failure must be considered in practical design. The maximum oblique pull was 74% of the axial pull since the smaller lateral capacity was reached first. Both the lateral and axial capacities were predicted with reasonable accuracy. The critical angle was defined as the angle of inclination of the load from the pile axis at which the failure mode changes from axial to lateral failure. It can be determined if the soil strength parameters and the pile type and geometry are known.

This paper presents the results of a study to improve the prediction of the in‐service behavior of several field‐compacted Indiana soils. The data base was created from testing of field‐compacted samples taken from test embankments and ongoing construction. The most important characteristic of the achievement with assurance of the best possible in‐service behavior parameters is the range of water content allowed in the lift at time of compaction. Selected charts and diagrams prepared from the data base are presented, as are the details of the created procedures. Examples are shown for the use of the procedures in: (1) Design engineering, where one creates the specification for compaction to ensure the presence of specific property parameters magnitudes; and (2) quality assurance, where one does not know the borrow area in advance of construction, but where one can establish the magnitudes of parameters of the compacted product from inspection testing results. It is believed that if implemented with a sufficiently large data base, these procedures would markedly increase the reliability of, and confidence in, the use of compacted soils for engineering purposes.

A new approach for the interpretation of pressuremeter tests in sands is proposed that permits the determination of shear strength characteristics, dilatancy properties, and shear modulus from the anlaysis of pressuremeter test results. A relatively simple soil model has been developed that takes into account the contracting or dilating behavior of the soil, or both, during the pressuremeter cavity expansion. The soil is assumed to be an elasto‐plastic material with a nonassociated flow rule. To evaluate the proposed interpretation procedure, laboratory cavity expansion tests on Fontainbleau sand specimens have been conducted in a specially designed hollow‐cylinder cell. The comparison between the sand properties (shear modulus, peak friction angle, dilation angle) derived from the cavity expansion tests and those obtained from conventional tests illustrates the applicability of the proposed interpretation procedure.

This paper discusses the development of a mapped‐type compositeinfinite element for modeling the response of unbounded saturated‐soil media. The coordinate ascent mapping technique that uses conventional shape functions and the Gauss‐Legendre integration scheme is used for the formulation. The constructed element preserves the compatibility between the variations in effective stress and pore pressure. This compatibility is a basic requirement for the accurate numerical modeling of saturated media. The accuracy of the element has been evaluated by comparing the numerical solutions of problems involving unbounded media with the corresponding analytical solutions. The effects of both the remoteness of the truncated boundary and the location of the infinite element coupling to the finite elements on the solution accuracy has been investigated. It is observed that accurate numerical responses of unbounded media can be obtained by using infinite elements at a much lesser computational effort than the conventional finite‐element methods in which unbounded media are represented by large mesh configurations.

During the phosphate beneficiation process, substantial quantities of waste clays are produced that are typically disposed of in large impoundment areas. Due to the low permeability and self‐weight of these clay slurries, years are required to achieve a consistency conducive to reclamation. Accordingly, the objective of this research is to compare reclamation schemes of sand/clay (s/c) mix versus capping for enhancing consolidation of these waste clays. These comparisons are accomplished analytically using a closed‐form solution to estimate magnitudes and a finite strain consolidation computer program to determine rates, while centrifugal model tests provide experimental data. From these comparisons, it is found that the descending order of reclamation scheme effectiveness in achieving the highest final solids content is: (1) Sand capping (staged); (2) s/c‐mix capping; and (3) using s/c mix, provided surface drainage is maintained and caps do not become buoyant. The time required, however, is proportional to the solids content; e.g., higher solids contents require longer times.

A fracture-grouting test section is a 7.1-m-high homogeneous dam of compacted silty clay on a foundation of the same material. Four test holes with different depths and lengths of casing are drilled at the dam crest along the dam axis. The grouting pressure variation curves are measured while the holes were grouted, and the shape and orientation of grout vein are recorded while the test section is excavated. The fracturing pressure is found to be directly related to the stress of the dam where fracturing started. The fracturing in an uncased hole starts near the middle height of the dam and expanded laterally and vertically. If long casing restricts fracturing within the lower part of the dam, the grout vein will have uniform thickness along the depth and be able to extend into foundation soil. All the grout veins are parallel to the dam axis and vertical.

In this paper, a novel approach is proposed to predict the pressure-settlement and pressure-tilt characteristics of eccentrically-obliquely loaded footings using the hyperbolic stress-strain curves of soils as nonlinear constitutive laws. The analysis is developed for strip footing. Stresses in the soil mass were obtained using the theory of elasticity. It is capable of incorporating the effect of shape, base roughness, and flexibility of footings. The results are obtained for the rigid strip footing resting on clay for load inclinations of 0°, 10°, and 20°, and eccentricity width ratios of 0, 0.1, and 0.2. The values of ultimate pressure obtained using the double-tangent method were found to decrease with the increase in eccentricity and inclination of load, and its comparison with the work of previous investigators showed good agreement.