To understand the structural damage evolution process of loess under the action of dry-wet cycles, a triaxial test of a dry-wet cycle was performed by considering three influencing factors: initial moisture content, amplitude of the dry-wet cycle, and number of dry-wet cycles. The stress-strain curves of undisturbed loess samples at different cycling times vary under different compacted loess cycles. Under the same axial strain, the stress value of the undisturbed loess is higher than that of the loess sample after a dry-wet cycle, indicating that such cycle can reduce the strength of loess. As the number and amplitude of dry-wet cycles increase, the shear strength of the loess sample and the value of cohesion (
In Northwest China, loess is distributed continuously and widely, with thick strata. Loess in this region is typical Quaternary loose sediment with the best development and the most complete strata [
Recent studies on dry-wet cycles have focused on expansive soil and mud-gravel roadbeds. Zhao et al. [
In the current study, a large number of dry-wet cycle triaxial compression tests were conducted in Luochuan loess to comprehensively analyze the effects of initial moisture content and the amplitude and number of dry-wet cycles on the strength degradation of loess. From the micro and meso perspectives, CT images and numerical analyses were utilized to perform the meso-structure inspection of crack development and structural degradation inside a loess structure under the action of a dry-wet cycle. The damage evolution law of loess under the action of a dry-wet cycle from the macro and meso perspectives was comprehensively analyzed. The findings can provide a reference for analyzing the damage and degradation of loess during dry-wet cycles and predicting a loess slope disaster.
Luochuan loess was used as soil sample for laboratory tests in this study. The collected soil samples were subjected to laboratory geotechnical experiments to analyze the physical properties of the loess samples. The physical properties of soil were obtained as listed in Table
Experimental statistics of the physical and mechanical properties of the soil samples.
Statistics | Moisture content | Natural density | Void ratio | Granule proportion | Dry density | Liquid limit | Plastic limit | Plasticity index |
---|---|---|---|---|---|---|---|---|
Maximum value | 21.55 | 1.8 | 1.18 | 2.73 | 1.54 | 32.5 | 20.7 | 11.8 |
Minimum value | 13.11 | 1.45 | 0.76 | 2.71 | 1.25 | 27.5 | 18.4 | 9.1 |
Average value | 16.5 | 1.6 | 0.98 | 2.72 | 1.38 | 30.16 | 20.81 | 9.35 |
The primary test equipment is composed of the following: (1) an electric heating constant temperature drying oven, (2) a standard unsaturated soil triaxial test system manufactured by the British Government Digital Service (GDS) Company, and (3) a SIEMENS SOMATOM Plus X-ray spiral CT machine. The test equipment is shown in Figure
Test equipment. (a) Electric heating constant temperature drying box. (b) GDS three-axis instrument. (c) SIEMENS SOMATOM Plus X-ray spiral CT frame. (d) Computer supporting the CT machine.
The test scheme used Luochuan loess as a background. The crack evolution and strength degradation of loess under repeated dry-wet cycles were investigated.
The in situ soil sample, indoor measurement of moisture content, and data collection indicated that moisture content exhibits a large variation (
Schematic of the dry-wet cycle process.
Four dry-wet cycle amplitude tests were designed on the basis of the initial moisture content of the soil samples to compare and analyze the tests with different cycle times and amplitudes. Magnitude was determined through the meteorological and hydrological data of Luochuan County and the depth of the steep layer of atmospheric influence. The four dry-wet cycle amplitudes were as follows: 20% (5%–25%), 15% (5%–20%), 10% (5%–15%), and 5% (5%–10%) for five dry-wet cycles. To reflect the influence of dry-wet cycle amplitude on the shear strength parameters and reduce test strength, this study considered initial moisture content as the moisture content for the test.
The majority of the triaxial undisturbed soil samples were prepared and subjected to water distribution methods to prepare multiple groups of triaxial samples with different initial moisture contents. Using conventional triaxial tests, dry-wet cycle tests were performed for each group of soil samples under different confining pressures (
The GDS triaxial apparatus was used to perform a triaxial test on the undisturbed loess in Luochuan. The diameter and height of the conventional triaxial sample were 3.91 cm and 8.0 cm, respectively. To study the strength degradation law of the loess under the action of a dry-wet cycle, the stress-strain curve of unsaturated loess under such cycle was analyzed under the control of confining pressure.
Two samples (LC-1 and LC-2) were selected for this test, which was performed to observe the changes in the internal microstructure of the samples after the dry-wet cycles and analyze the effects of different numbers and amplitudes of dry-wet cycles on the structure of loess. The soil samples were Luochuan silty clay. For LC-1, density
In this section, dry-wet cycle tests with different times and amplitudes were performed on triaxial samples. As shown in Figures
Surface characteristics of Luochuan loess macrostructure during a dry-wet cycle when cycle amplitude (
Surface characteristics of Luochuan loess macrostructure during a dry-wet cycle when
The box-counting dimension method was used to analyze variation in the surface cracks of soil samples under the action of a dry-wet cycle from a macro perspective. This method covers the measured area with square grids (
Yin et al.’s [
The logarithm of equations (
In the double logarithmic coordinate system, the
The photos were processed using Adobe Photoshop, and the average crack rate of the sample was calculated using MATLAB to obtain the average value and the fractal dimension of the surface cracks of the sample. The correlation curve between the fractal dimension of the surface cracks of the loess and cycle times under different cycle amplitudes was drawn using cycle times as the abscissa and fractal dimension as the ordinate as shown in Figure
Curves of the fractal dimension of the fissures on the surface of the loess samples, number of cycles, and amplitude.
When the volume of the combined water and gravity water in loess clay particles increases, the distance between clay particles wedged by water increases, reducing the magnitude of the force between particles. When the softening degree of the cement is large, the decrease in the friction coefficient and the friction between particles is considerable, increasing the shear strength of the loess sample. Cracks will occur when water pressure between the aggregates of the loess body is greater than the shear strength. When water dries up, the moisture content of the soil sample decreases, the suction of the matrix in the soil sample increases, and the original cracks in the soil sample are closed. Notably, cracks are easily generated in the soil and only temporarily close but will not disappear due to the unrecoverable tensile strength of the soil. New cracks will occur when the tensile force caused by the hydraulic gradient of the dehumidification process is greater than the tensile strength of the soil. Therefore, the total number of cracks increases during a dry-wet cycle. The original cracks will occur again during rehumidification. The hydraulic gradient required for these cracks to occur is smaller than that in the previous cycle due to the increase in the number of cracks, and the tensile stress generated during the dry-wet cycle is gradually reduced. Consequently, the amplitude of cracks generated during the dry-wet cycle is reduced. Cracks stop developing and expanding when the tensile stress generated by the hydraulic gradient due to the dry-wet cycle is less than the tensile strength of the soil.
Figures
Stress-strain curves of the samples with an initial water content of
Stress-strain curves of the samples with an initial water content of
Stress-strain curves of the samples with an initial water content of
Stress-strain curves of the samples with an initial water content of
The stress-strain curves of the undisturbed loess samples under different cycle times considerably vary. Under the same axial strain, the stress value of undisturbed loess is higher than that of the loess sample after a dry-wet cycle, implying that dry-wet cycles exert a degrading effect on the strength of loess. When confining pressure and water content are low, the difference in the principal stress of undisturbed loess gradually decreases under the same axis, and the change amplitude decreases as the number of dry-wet cycles increases. Typical structural soils exhibit a certain degree of shear shrinkage. When the confining pressure is less than the structural strength, the stress-strain curve demonstrates a softened peak point; the corresponding structural strength plays a major role. The first stage of deformation is approximated. During the linear elastic stage, the loess body exhibits a certain amount of shear shrinkage. During the second stage of deformation, the structural strength of the loess body begins to deteriorate. Slips occur between loess particles, the structure is damaged, and dilatancy occurs. The stress-strain curve displays a nonlinear increasing trend. The third stage is called the stage of stable equilibrium. During this stage, shrinkage and dilatancy complement each other to attain balance, volume change stabilizes, and the stress-strain curve tends to decline gently and slowly. As the number of dry-wet cycle increases, the stress-strain curve exhibits no peak point, the shape of the curve changes from weakly softened to hardened, and the amplitude of the hardened curve becomes increasingly evident. The position of the strain curve decreases, the change in the principal stress difference follows a decreasing trend, indicating that the strength of loess is gradually deteriorating. The primary reason for this phenomenon is that loess has vertical joints, its structure is relatively strong, and its permeability is good. Therefore, as the number of dry-wet cycle increases, soil skeleton changes irrecoverably, seriously reducing the strength of the original soil structure, and eventually the structural strength of loess wanes. In summary, the effect of dry-wet cycles is the major factor that decreases the strength of undisturbed loess.
Mohr circles and strength envelopes of the samples with 25% initial moisture content at different numbers of cycles. Mohr circle and strength envelope (a) during the initial state, (b) in a dry-wet cycle, (c) in three dry-wet cycles, and (d) in five dry-wet cycles.
The total shear strength of undisturbed loess with different water contents is obtained from this. The curves of the shear strength of the loess sample under different dry-wet cycle conditions are shown in Figure
Relationship between shear strength
Figure
Relationship between the strength parameters of the soil samples with different moisture contents and numbers of cycles in Luochuan. (a) Relationship between cohesion (
Figure
A direct shear test exhibits certain comparability with the triaxial shear sample; thus, this section of the soil sample undergoes consolidated direct shear test. For the direct shear test, a consolidated fast shear test was performed using a ZJ strain-controlled direct shear instrument produced by Nanjing Soil Instrument Factory, with a shear rate of 0.15 mm/min. For soil samples with varying moisture contents, different dry-wet cycle amplitude tests were performed according to Section
Relationship between the shear strength indexes of different dry-wet cycle amplitudes and
As shown in Figure
The effect of a dry-wet cycle is attributed to the repeated loading and unloading of matrix suction in soil, resulting in an irreversible damage process to the internal structure of soil. The test results indicated that the effect of the dry-wet cycle of loess is closely related to the control parameters of the cycle. Therefore, under the action of a dry-wet cycle, the control of a series of related parameters, such as soil body and cycle order, should be strengthened to explore and discuss the attenuation law of the effect of dry-wet cycles on loess, particularly the change law of intensity.
This section presents the LC-1 and LC-2 samples as examples. The samples were subjected to six dry-wet cycles with amplitudes of
Figures
CT images of the
CT images of the
CT images of the
CT images of the
The results in the figures show that the initial soil sample before drying exhibits nonuniform micropores and microcracks and other initial damage media. Under the action of a dry-wet cycle, the microporous cracks in the dried soil samples gradually develop and open. With an increase in the number of dry-wet cycles, the microcracks gradually elongate and widen, exhibiting irregular bifurcation and, finally, penetration and soil failure. This process reflects the entire process of the damage evolution of the soil’s internal structure.
To accurately explain the development and evolution of the damage fractures in soil during a dry-wet cycle, the quantitative analysis of the CT numbers and the obtained changes in such numbers are used to clearly illustrate the full range of joint damage and fracture expansion evolution of loess. The change law of the mean values of CT number ME and SD is illustrated in Figure
Relationship between the number of dry-wet cycles and CT number. (a) Change curve of CT number ME. (b) CT number SD curve.
After the first dry-wet cycle, the ME and SD of the soil sample exhibited the highest decrement and increment, respectively, indicating that this cycle exerts the most evident effect on porosity, cracking, and degree of development inside loess. After the soil sample is humidified, the size of the micropores inside the soil sample gradually increases under the action of water. During the drying process, the water evaporates from the inside of the soil sample, and the microcracks expand (i.e., elongate and broaden). This phenomenon explains why the ME of the soil sample decreases and the SD increases. After two dry-wet cycles, the ME value increases, suggesting that the soil sample exhibits a dense realization during the dry-wet cycle. However, the increase in the SD value implies that the damage expansion characteristics continue and, thus, demonstrates the soil damage evolution. What mentioned above demonstrates the irregular characteristics of the soil damage evolution. After five dry-wet cycles, ME reaches the maximum value, but the SD value still increases linearly. This result indicates that before this, the soil sample was compact (i.e., the number of closed micropores in the soil sample was decreasing). Moreover, ME tends to decrease, whereas SD linearly increases, indicating that the meso damage of the soil sample is gradually evolving; that is, a new damage begins to develop based on the initial damage. The evolution reflects that cracks continue to open, and ME is decreasing. With an increase in the number of dry-wet cycles, when all the cracks are penetrated, the damage evolution of the internal cracks and fissures in the soil intensifies and eventually leads to the failure of the soil samples. The larger the amplitude, the larger the change in the ME and SD values, and the fewer the cycle times, leading to soil sample damage. In addition, the larger the amplitude, the faster the development of the fissures and cracks in the soil sample, exacerbating the damage evolution of the soil sample.
The causes of fissures in loess under the action of a dry-wet cycle were analyzed by combining the collapsibility of loess and the mechanical mechanism of fissure generation. Crack generation involves a change in the stress state of soil during the dewetting process. Negative stress appears on the surface of soil, inducing stress on soil. In accordance with theory, cracks are generated during the conversion of saturated soil to unsaturated soil. Therefore, three sets of stress tensor can be used for the analysis:
The isotropic and linear elasticity assumptions can be expressed as
Before cracking, the following condition must hold:
Substituting this expression to derive the net horizontal stress yields
Surface soil
For any unsaturated soil, the strength
In unsaturated soils, matrix suction exerts a certain effect on soil strength. The tensile strength of soil is calculated as
The condition wherein the surface layer is under critical cracking is mathematically defined as
From the preceding derivation, the stress state equation of soil cracking can be obtained as
The loess itself has collapsibility, disintegration, and permeability. These characteristics cause the soil to deform and destroy, and uneven deformation and fracture will produce cracks under certain conditions. According to the previous section and the theoretical analysis of the above-mentioned fissures, it is believed that the decrease of the water content in the soil increases the matrix suction
In this study, the structural degradation of loess is analyzed through the physical action of water, i.e., dry-wet cycles. The effects of this cycle on the structural and strength degradation of loess are investigated. The following conclusions are obtained from the micro and meso tests using high-technology techniques. Using fractal theory and MATLAB to calculate fractal dimension, the evolution of the surface cracks of the soil sample is observed from a macro perspective. Studies have shown that the maximum fractal dimension The stress-strain curves of the undisturbed loess samples under different cycles are obtained by varying the stress-strain curves of the static triaxial tests on undisturbed loess under different numbers of dry-wet cycles and confining pressures. Evident differences are observed. Under the same axial strain, the stress value of undisturbed loess is higher than that of the loess sample after a dry-wet cycle, implying that the effect of the dry-wet cycle on loess strength is degraded. Using the obtained stress-strain curve, the relationship between the shear strength and index of loess and the number and amplitude of dry-wet cycles is determined. As the number and amplitude of dry-wet cycles increase, the shear strength of the loess sample, the cohesion value of the strength index, and the amplitude gradually decrease. With an increase in the number and amplitude of dry-wet cycles, the change in the internal friction angle of the strength index is inevident, indicating that the effect of dry-wet cycles on the internal friction angle of loess is insignificant. Through the CT scan test, the effect of dry-wet cycles on the development and evolution of existing cracks in the sample was analyzed from a meso-scale perspective. Studies have shown that, with an increase in the number and amplitude of dry-wet cycles, the ME value of CT decreases, whereas its SD value increases. The highest decrement of ME and the highest increment of SD are observed during the initial stage of a dry-wet cycle. In conclusion, dry-wet cycles promote the development of cracks. According to experiments and theoretical analysis, it can be known that the generation of cracks in soil is closely related to the suction of the matrix inside the soil, and the change in water content (saturation) in the soil is the main factor leading to the generation of cracks. Of course, the generation of cracks in the loess body is also related to the deterioration of the strength of the loess. The soil structure has various degrees of damage, and these cracks provide good channels for water infiltration or evaporation.
The data used to support the findings of this study are available from the corresponding author upon request.
The authors declare no conflicts of interest.
The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (Grant nos. 41672305 and 41902299) and the Key Science and Technology Program of Shaanxi Province (Grant no. 2017ZDXM-SF-082).