Michael J. Gunn

Water Retention and Compressibility of a Lime-Treated, High Plasticity Clay

The paper presents a laboratory investigation, which quantified the beneficial effect of lime on the water retention, compressibility and overall volume changes of a high plasticity clay (London Clay). The study comprised oedometer testing (including suction controlled testing), shrinkage tests, and filter paper testing to determine the soil water retention curves of the drying soils. The latter tests showed that the shallow-mixed lime-treated soils are likely to be partially saturated in situ depending on the environmental conditions; thus, unsaturated soil mechanics concepts are applicable for the description of their behaviour. The oedometer results quantified the marked decrease in the swelling and compressibility of the lime-treated soil compared to the untreated clay and the increase in the vertical effective yield stress (due to suction and cementation bonding). The lime enabled the soil to maintain a more open structure with respect to the untreated soil throughout the compression process, and appears to be the main factor influencing the compressibility of the unsaturated lime-treated soil (rather than suction) for the ranges of suction tested.

A New Laser Sensor Volume Measurement System for the Triaxial Testing of Unsaturated Soils

The paper demonstrates a novel non-contacting volume change measurement system for triaxial soil testing. The system uses an ultra high accuracy laser displacement sensor that measures water volume changes resulting from the volume change of the specimen during testing. The paper presents the design, working principles, system calibrations, and the validation of the system through comparative measurements of volume changes of soil specimens during triaxial shearing, against results of conventional volume measurement techniques.

Efficient Laboratory Measurements of the Soil Water Retention Curve

The knowledge of the soil water retention curve (SWRC) is of particular importance for unsaturated soils. However common techniques to determine SWRC may be much more time-consuming than other soil property testing techniques and may lack flexibility in controlling the stress state of the soil in the field, which is of most relevance for geotechnical engineering applications. The paper presents a transient testing technique for the continuous measurement of SWRC in a triaxial cell during which testing rates are adjusted based on numerical modeling results simulating the testing. This enables the optimization of the experimental process and hence a more efficient determination of the SWRC. The technique is then successfully applied to determine the SWRC of statically compacted silica flour specimens, tested under different confining stresses.

Mechanical Behaviour of Unsaturated Lime-Treated Clays: Experiments and Modelling

This paper gives a brief account of ongoing research on the mechanical behaviour of lime-treated clays. First indicative results of a series of suction-controlled triaxial tests using the axis translation technique are presented based on which the effect of cementation is assessed. The results showed the overall beneficial effect of the lime on the compressibility behaviour of the soil as well as the dramatic increase in strength gained for ranges of strain relevant to engineering applications. A constitutive model for the behaviour of lime-treated soils is then introduced to account for the combined effects of suction and cementation.

Numerical Modelling of Flow in Variably Saturated Porous Media with Fluctuating Shallow Water Tables

The paper discusses various models which have been proposed to simulate the hysteresis of the water retention curve. Three of these models were implemented into a finite element computer program for flow in variably saturated porous media and were applied to the solution of transient flow through a soil column subject to drying-wetting cycles. The performance of these models was assessed through comparisons with previously published experimental results. A linear hysteresis model was shown to reproduce satisfactorily the experimental results, although it was simple to implement and required minimal experimental data to be formulated. Overall, the results indicate that the fact of including hysteresis in the numerical model improved the predictions with respect to the case when single-valued functions were used.