No abstract

The accurate predictions of deformations around geotechnical structures depend on the correct definition of the stiffness–strain curve of the soil, and in particular on the elastic stiffness at very small strains, which, being the starting point of the curve, may be regarded as a benchmark parameter for a given soil. Whereas for clays the factors controlling stiffness are well understood, for sands the current analyses tend to be empirical. The paper presents the results of bender element tests examining the elastic stiffness of sands, which is then related to the stiffness at larger strains determined in triaxial tests using a new system of local axial strain measurement. Three sands with very different geological origins were tested over a wide range of stresses, allowing a new general framework for stiffness to be established. By means of correct normalization of the data, unique relationships are derived for each soil which demonstrate that the confining stress and volumetric state relative to the normal compression line are principal controlling factors as they would be for clays. The paper shows, however, that for sands the means of arriving at its initial volume–stress state is also important, in particular whether this is by geological overconsolidation or compaction during the depositional process. La précision des prévisions des déformations autour de structures géotechniques repose sur une définition correcte de la courbe rigidité/ déformation du sol, et en particulier sur la rigidité élastique à très faibles contraintes, qui, vu qu'elle constitue le point de départ de la courbe, pent être considérée le paramètre de référence d'un sol donné. On comprend bien les facteurs qui régissent la rigidité des argiles, mais, dans le cas des sables, les analyses actuelles tendent à être empiriques. L'article présente les résultats d'essais de résistance à la flexion réalisés pour examiner la rigidité élastique des sables et la mettre ensuite en corrélation avec la rigidité à des contraintes plus fortes, établie dans des essais triaxiaux à I'aide d'un nouveau système de mesure locale des contraintes axiales. Les essais ont porté sur trois sables d'origines géologiques très différentes, dans un large éventail de contraintes, ce qui a permis de formuler un nouveau schéma général pour la rigidité. Grâce à une normalisation correcte des données, les auteurs ont pu établir pour chaque sol des rapports individuels qui montrent que la contrainte de confinement et la contrainte hydrostatique par rapport à la ligne de compression normale sont les principaux facteurs déterminants, tout comme dans le cas des argiles. L'article montre toutefois que, dans le cas des sables, le processus qui aboutit à I'état initial de contrainte hydrostatique est aussi un facteur important, qu'il s'agisse de préconsolidation géologique ou de tassement pendant la sédimentation.

A detailed investigation was carried out on a silty sand from Boˇstanj, Slovenia, in order to identify the role that particle breakage plays in test interpretations and mechanics. The soil was tested up to a high pressure in the triaxial and oedometer apparatus. Unexpected for a silty sand, the basic patterns of behaviour, in terms of strength and stiffness, were found to be similar to those of many previously investigated clean sands, with unique and parallel Normal Compression and Critical State Lines at higher stress levels and a horizontal asymptote to the Critical State Line at lower stress levels. The stiffness was controlled primarily by the state of the soil relative to the location of these lines. Despite the better grading, there was still very significant particle breakage. However, the breakage was largely confined to the coarse fraction. The grading curves after various tests all showed convergence with the initial grading towards 1 mm, which may therefore represent the comminution limit. It is believed that it is the particle breakage that caused the behaviour to follow a simple Critical State type of framework and why transitional or non-convergent compression behaviour was not found.

This study was conducted to characterize a coarse alluvial vadose zone. The characterized sediment was used to derive data across a wide moisture range on different size fractions combining the simplified evaporation experiment and the dew point method. The obtained data were compared with the mass-based gravel correction to test its accuracy.

An investigation into the behaviour of North Sea glaciomarine clays was carried out in which triaxial tests were conducted on both natural and reconstituted samples to assess the effects of structure. Although the tests were conventional CID tests, high quality instrumentation was used. The tests were also technically difficult both because of the very long test durations and because some of the samples were swelled back to very low effective stresses so that the effect of swelling on the influence of structure on the soil behaviour could be assessed. A "Class A" prediction of the behaviour in these tests was carried out using the BRICK model. Although the model is not designed to account for the influence of structure, it was found that its effects could be simulated by allowing the soil to have artificially high overconsolidation ratios (OCRs) so that the high undrained shear strengths resulting from structure could be modelled. Making the simple assumption that the decay of stiffness could be scaled from that of London Clay, reasonable predictions of the behaviour were made. The discrepancies between the predictions and the measured behaviour became significant only at the lowest stresses, where structure dominates the behaviour of the natural soil.Key words: laboratory testing, small strain stiffness, stiff clays, structure of soil, numerical predictions, shear deformation properties.