Effects of in situ wetting–drying cycles on the mechanical behaviour of an intact loess

Abstract

In situ wetting–drying cycles significantly and repeatedly influence the soil water content in the active zone but not in the steady zone. To investigate effects of in situ wetting–drying cycles on the mechanical behaviour of intact loess, a series of pressure plate tests was carried out with an intact loess sampled from the active and steady zones. Results show that the specimens from the steady zone have higher yield stresses at given suctions and a larger inclination of normalized loading collapse (LC) curve than the specimens from the active zone. As evidenced by the scanning electron microscopy results, the clay particles accumulating around the contacts between silt particles are widely observed in the former specimens but not in the latter specimens. The distribution of clay particles in the former specimens results in a stabilized skeleton with mechanical properties sensitive to suction variations. On the other hand, the maximum collapse potential of the specimen from the steady zone is 37% larger than that of the specimen from the active zone, mainly because of the larger inclination of normalized LC curve owned by the former specimen.

A constitutive model for structured soils under saturated and unsaturated conditions

“…In contrast, the unsaturation effects on preconsolidation pressure are pronounced for the soils with clay particles accumulated around the contacts between silt particles, which enlarges the contact area and enhances the capillary effects. 19,24,22 Therefore, the difference between Δ𝑒 3 and Δ𝑒 1 is larger than Δ𝑒 2 and Case-II is observed. Additionally, Case-III indicates that the structure effects on the normalised LC curve are negligible (e.g., 25 ), suggesting that Δ𝑒 3 − Δ𝑒 1 = Δ𝑒 2 .…”

“…Additionally, experimental data in the literature (e.g., 21,24,27,36 ) suggest that the normalised LC curve of structured soils changes towards that of reconstituted soils as structure degrades. This behaviour can be modelled by considering varied 𝑐 value with structure degradation, as shown in the following equation:…”

“…Intact loess tested by Mu et al 24 To investigate the mechanical behaviour of intact loess, Mu et al 24 conducted suction-controlled consolidation tests on the intact loess sampled from two depths (1 and 5 m). Table 6 illustrates the calibrated model parameters.…”

“…One example is intact loess, in which most clay particles accumulate around the contacts between silt particles (e.g., 22,23 ). The stiffening effects of clay particles on the soil skeleton are sensitive to moisture conditions (e.g., 24 ). Additionally, the normalised LC curve of some soils can be insensitive to the structure (i.e., Case-III) (e.g., 25 ).…”

“…Therefore, a unified approach that can model these three cases (Case-I/II/III) of the LC curve is necessary to capture the behaviour of various structured soils well. Furthermore, experimental results show that the normalised LC curve of structured soils evolves towards that of reconstituted soils as the structure degrades (e.g., 27,24 ). Additionally, Liu et al 7 suggested that structured soils generally have a higher value for the plastic strain increment ratio 𝑑𝜀 𝑝 𝑣 ∕𝑑𝜀 𝑝 𝑞 than reconstituted soils.…”