Soil behaviour in suction controlled cyclic and dynamic torsional shear tests

Studies carried out on unsaturated soils have focused on identification of general rules governing their behaviour. Great care was devoted as well to constructing a general framework capable of fitting experimental observations in the large strain range (Alonso et al., 1990; Fredlund & Rahardjo, 1993; Wheeler e Sivakumar, 1995). Despite the need of analysing small strain range behaviour of unsaturated soils was felt from several years ago (Wu et al., 1984; Quin et al. 1991), thus far very little efforts have been devoted to this subject. This paper contributes to reduce this lack of evidence, exposing the results of suction controlled RCTS tests performed at the University of Naples Federico II. The resonant column-torsional shear device (RCTS) developed at the Dipartimento di Ingegneria Geotecnica (DIG) uses the axis translation technique in order to control suction (Mancuso et al. 1999, Vinale et al. 1999). Tests have been performed on a Proctor compacted silty sand at two different moulding water contents: optimum value and 2.5% wet of optimum.The experimental program was conceived in order to analyse the effects on small strain behaviour of both suction and compaction induced fabric. Several compression and torsional shear stages were carried out under constant suction conditions, varying (ua–uw) in the range 0400 kPa for both the optimum and the wet compacted soil. Each test included three isotropic compression stages, reaching a mean net stress (p–ua) equal to 100, 200 and 400 kPa. At the end of all the compression stages, several series of RC and TS tests were performed. In order to analyse strain rate effects, constant torque TS tests were carried out at loading frequencies of 0.06, 0.1, 0.5 and 2 Hz. In each series of RC and TS tests, the torque was progressively raised. For the intermediate applied net stresses, the elastic threshold strain of the soil was not crossed, while at the highest (p–ua) value torque was increased up to the maximum allowable (0.43 Nm) in order to obtain the G: and D: curves. Experimental results show a significant increase of shear stiffness with increasing suction. Both for the optimum and for the wet compacted material, most of the suction effects are observed for (ua-uw) from 0 to about 200 kPa. The stiffness values are well grouped for constant stress levels and show an S-shaped increase in stiffness with suction. This shape is explained in the paper by considering the progressive change of soil behaviour from bulk to menisci-water regulated mechanical response. Increasing the moulding water content causes a significant reduction in shear stiffness, affects the ratio between the saturated and unsaturated threshold values of Go, and influences the suction values characterising transition between bulk- to menisci-water regulated behaviour. Data referring to initial damping ratio Do show similar trends for the optimum and the wet soil and seem to indicate that this quantity is not affected by specimen preparation procedure. Increasing suction causes a significant reduction in damping values. The curves G: and D: highlight that the effects of suction are non-negligible in all the investigated strain range ( = 10-410-1 %). The shape of these functions, however, seems to be not significantly affected by suction level, while a higher moulding water content produces a slight increase of the linear threshold. Overall, the results as far collected indicate a significant effect of matric suction and specimen preparation procedure on shear stiffness and damping ratio. Furthermore, they confirm the functionality and the great potential of the new apparatus projected at the University of Naples.