Effects of molding water content on the behaviour of an unsaturated silty sand

Unsaturated soils can be often encountered in practical applications. In particular, even though index properties of a soil as construction material are left unchanged, the compaction procedure (moulding water content, compaction energy, etc.) influences the mechanical response of the resulting material, its after compaction degree of saturation (i.e. suction) and, more generally, the earth constructions behaviour (Vinale et al., 1999). This paper tries to contribute to the discussion of the above mentioned aspects of soil behaviour presenting the results of a series of suction controlled triaxial tests performed on an optimum water content and wet of optimum compacted soil. Details on the used apparatus are reported in Rampino et al. (1999). The experimental program was conceived in order to analyse the effects of both suction and fabric on soil behaviour. Overall, it consisted of 23 tests, i.e. isotropic compression, constant mean net stress (p-ua) shear and “standard” shear tests, all carried out under drained conditions and constant suction. The investigated suction levels were 0, 100, 200, 300 kPa for the optimum compacted soil and 0, 100, 200 kPa for the wet compacted material. The optimum and the wet compacted specimens exhibited very different behaviour in all the stages of the tests. For example, during the equalisation stages, the optimum material swelled and absorbed water, while the wet soil shrunk and expelled water. This behaviour is explained by the measurements of the after compaction suction (ua-uw)o. Using the Imperial College tensiometer (ua-uw)o was measured as high as 800 kPa for the optimum material (then causing wetting during equalisation) and as low as 60 kPa for the wet specimens (then causing drying during equalisation) – a relevant effect of moulding procedure. The results of the isotropic compression tests highlight a strong effect of suction on soil compressibility. In particular, the compressibility index decreases about 25% from 0 to 200 kPa of suction. The greatest amount of this effects occurs in a limited suction range (0-100 kPa) while tends to a threshold value for (ua-uw) greater than 200 kPa. Comparing experimental findings on the optimum and the wet specimens, it is observed that an increase in moulding water content induces a severe increase in compressibility. In addiction, data indicate that both the shape and the position of the after compaction loading-collapse loci (LC) are largely affected by moulding water content. Thus, soil susceptibility to collapse strongly depends on preparation water content. The data of shear tests show that suction significantly affects large strain shear stiffness and resistance. For instance, in the range 0-200 kPa of suction, the strengthening with (ua-uw) is expressed by an apparent cohesion varying from 0 to 200 kPa for the optimum and from 0 to 150 kPa for the wet compacted soils. The results seems to be well grouped for suction levels and can be interpreted by straight fitting lines having the same M coefficient in the q:(p-ua) plane. The effect of moving from the optimum to the wet compaction water content is then a decrease in stiffness and in apparent cohesion, as well as a reduction of the size of the after compaction yield surface. The large series of experimental results obtained on Metramo silty sand can be satisfactorily modelled by the modern theories for unsaturated soils (Alonso 1990, Wheeler & Sivakumar 1995), if it is assumed that the different preparation procedures yield to different soils. The influence of suction and moulding water content on the observed behaviour clearly emerged in all the tests phases and highlights the need of carefully considering these two variables in handling engineering problems involving compacted soils as construction materials.