Evolution of the properties of lime-treated silty soil in a small experimental embankment

International audienceIn order to study the evolution of a lime-treated silty soil under in-situ conditions of mixing, laying out, compaction and environment, an experimental dike was built near Rouen, in northwest France; this dike is 25 m long, 4 m wide at the crest and 1.8 m high. While special care was taken to mix the soil and lime very thoroughly, to precisely measure the quantity of water, to compact the soil using a vibrating sheepsfoot roller, the building conditions were those usually used to construct embankments in all other respects. Samples were taken from the dike approximately 1 month, 6 months and 12 months after the completion of the work and underwent triaxial and oedometric testing. To compare the properties of treated soil with those of untreated soil, another smaller "reference" embankment was built nearby under similar conditions with silty soil.Small-strain triaxial tests were performed on unsaturated specimens of treated and untreated soil using an original strain measurement device. The change in axial secant moduli can be described by a power law E = beta sigma(m)(1), where beta is a parameter that depends on curing time and m is nearly constant (m approximate to 0.45).CU and CD large-strain triaxial tests carried out on saturated treated and untreated specimens, consolidated under confining stresses ranging from 10 to 300 kPa, showed a definite increase in cohesion as a function of Curing time, with nearly unchanged friction angles. The possibility to use a non-linear failure criterion was examined. The compressibility coefficient C-iso,C- slope of the critical state line (CSL) in the [log(p'), e] coordinate system, is significantly lower in treated than untreated specimens, but its value remains constant whatever the curing time. At the same time, the CSL of the treated specimens is above that of the untreated silt. The brittleness of the material, derived from the peak to residual deviator stress ratio q(peak)/q(res) is the same for treated and untreated specimens and is independent of curing time and confining stress. However, the axial strain corresponding to the peak of CD curves is much higher in untreated than in treated specimens. The permeability of the saturated lime-treated soil (10(-9) m/s) is the same as that of the untreated soil. The value measured one month after the embankment was constructed, does not change with time or with the confining stress. The results of the present study confirm those of previous studies on laboratory-prepared unsaturated specimens: they have demonstrated the effectiveness of lime-treatment in the much more difficult environment of a real working site and shown that lime-treated samples present a significant gain in their engineering properties, even when saturated, provided that the dosage of lime and water, mixing and compaction conditions are carefully controlled. The new material thus obtained presents lower plasticity, improved workability, increased stiffness and cohesion without any loss of performance in terms of brittleness or permeability. (C) 2015 Elsevier B.V. All rights reserved