Effects of Cyclic Freezing and Thawing on the Shear Behaviors of an Expansive Soil under a Wide Range of Stress Levels


 The focus of this paper is directed towards investigating the influence of multiple freeze-thaw (FT) cycles on the stress-strain relationships during undrained shearing for an expansive soil under a wide range of confining stresses (σc) from 0 to 300 kPa. Different numbers of FT cycles were applied to compacted specimens. The influence of FT cycles on the soil’s structure was investigated using mercury intrusion porosimetry (MIP) and scanning electron microscope (SEM) tests. FT impacted specimens were subjected to consolidated undrained (CU) shear tests with pore pressure measurement (σc = 10 to 300 kPa) and unconfined compression (UC) tests (σc = 0 kPa) to derive the shearing stress-strain relationships and the associated mechanical properties including (i) failure strength (qu), elastic modulus (Eu), effective and apparent cohesion (c’ and c), and effective and apparent friction angle (ϕ’ and ϕ) obtained from CU tests and (ii) qu and reloading modulus (E1%) and stress (Su1%) at 1% strain obtained from UC tests. Testing results show that FT cycles mainly influence the soil’s macropores with diameters between 5 and 250 microns. Cracks develop during FT cycles and result in slight swelling which contributes to an increase in the global volume of the soil specimens. There is a significant reduction in the investigated mechanical properties after FT cycles. They typically achieve equilibrium after about 6 cycles. The shearing stress-strain curves transits from strain-softening to strain-hardening as the confining stress increases. An empirical model is developed to describe the strain-softening behavior of the specimens under low confining stresses. The model is simple to use and well describes all stress-strain curves obtained in this study that show strain-softening characteristics.

In this study, the volumetric characteristics, the stress versus strain relationships, and shear  An expansive soil collected from Qiqihar, Heilongjiang, China is used for this research study.

112
The sampling site is a typical seasonally frozen region characterized by hot summer and cold 113 winter. The average temperature ranges from 23.1°C in summer to −18.6°C in winter. The  Table 1.  127 The collected expansive soil samples were air-dried, pulverized with a rubber mallet, and 128 then passed through a 2mm sieve to remove gravels and large size particles. To simulate the 129 field condition, the air-dried soil was wetted with water such that the natural water content of 130 26.3% is achieved. The moist soil was statically compacted at the natural dry density of 131 1540kg/m 3 into cylindrical specimens (38mm in diameter and 76mm in height) for 132 microstructure investigations and triaxial tests. Compacted specimens were sealed in layers of 133 Saran wrap and stored in plastic containers for at least 72h at room temperature of 25±1°C to 134 ensure that water was evenly distributed within the specimens.  The volumetric strain (ɛv) of specimens during FT cycles can be defined using Eq. 1.

Specimen preparation and application of FT cycles
where V0 is the initial volume of the untreated specimen, VN is the volume of the specimen  Table 2.

286
With an increase in the FT cycles, axial strain at the peak strength (denoted as ɛp) increases 287 from approximately 2% of the untreated specimen to 3% after 10 FT cycles, for both 288 saturated and unsaturated specimens, and the pre-peak stress-strain curves tend to become flat.

289
The stress-strain curves of specimens with different NFT seem to have an identical residual   Table 3. The variation in the qu, E1%, and Su1% with NFT is well described by 303 the proposed power function with R 2 > 0.9. The following information can be derived from 304 Figure 7 and   Table 4.  Fitting parameters and R 2 are summarised in Table 5.   The model presented in this paper could be a useful tool in the numerical analysis of the 430 strain-softening behaviors that can be derived from the stress-strain curves. The schematic diagram of the stress-strain relationship and the determination of the E1%, Su1%, and qu during the revised UC tests Figure 3 The volumetric variations of specimens during FT cyclic treatments Figure 4 The (a) CI and (b) PSD curves of specimens after 0, 1, 4, 6 and 10 FT cycles Figure 5 SEM images of (a) untreated specimens and specimens after 1, 4, 6, 10 FT cycles (b-e) Figure 6 The UC stress-strain curves of specimens after 0, 1, 4, 6 and 10 FT cycles under (a) unsaturated and (b) saturated conditions Figure 7 Variation of qu, E1%, and Su1% of specimens after 0, 1, 4, 6 and 10 FT cyclic treatments Figure 8 The CU stress-strain relationships of specimens after 0, 1, 4, 6 and 10 FT cycles under various con ning pressures (i.e. 10, 20, 30, 50, 100, 200 and 300kPa) Figure 9 Evolution of Failure strengths obtained from CU tests with the increase of FT cycle numbers Figure 10 Evolution of Elastic modulus of specimens obtained from CU tests with the increase of FT cycle numbers Figure 11 Evolution of effective cohesions and effective friction angles obtained from CU tests with the increase of FT cycle numbers Evolution of apparent cohesions and apparent friction angles obtained from CU tests with the increase of FT cycle numbers