Abstract
Exploring the development of multiscale behaviors of red-bed mudstone is beneficial for in-depth evaluation of swelling capacity and its progressive failure mechanism. In this paper, a macro-micro study was performed on rock samples in central Sichuan, China. The influence of environmental factors such as water content, degree of soaking, and gaseous moisture absorption was discussed. Spatial characterizations were finally described from pore structure and distribution. Experimental results suggested that deformation curve under water immersion had three stages of rapid growth, deceleration growth, and gradual stabilization. It can enter stable deformation within 60~80h and has short-term swelling effects, while gaseous moisture absorption presented long-term performance during 0~700 h. The difference in deformation between partial and complete immersion may be the release of swelling potential caused by humidity gradient and capillary suction. The strain-time curves during water immersion all tended to short-term swelling. Moreover, repeated drying and wetting promoted the time of shrinkage adjustment in the second cycle to be significantly shortened, resulting in a large amount of irreversible plastic deformation. Although initial water content influenced the final stable strain, it only changed the degree of deformation at initial stage. For the microscale structures, the closer to the end, the greater the porosity and the smaller the middle. The average porosity after 0, 1, and 2 dry and wet cycles reached 0.227%, 4.027%, and 6.121%, respectively. Geometric parameters such as volume and surface area of the pores enlarged with the increase of the number of cycles. Small pores gradually evolved into microcracks and large fissures. Besides, the pores developed from “tree-like” to “net-like” structures with continuous cycles.
1. Introduction
As a special type of rock mass in the construction of high-speed railways, red beds are mostly deposited in an environment prone to weathering and oxidation. From the perspective of geographical distribution, there are a large number of typical Jurassic and Cretaceous red beds in southwestern China, accounting for about one-third of its total area [1]. In the environment of fault zone in the distribution area, the structures of red-bed mudstone are broken, and their engineering properties are unstable. Affected by the climate of basin and seasonal rainfall, it often results in roadbed deformation, slope instability, and landslides [2, 3], severely restricting local economic development and road network construction. Therefore, studying the main causes and objective laws of its deformation has important practical significance for guiding engineering design and construction and improving transportation conditions.
Red-bed mudstone usually contains certain clay minerals, which will swell and even disintegrate when it encounters water [4–6]. In recent years, some scholars have carried out a series of field and laboratory experimental studies on the deformation of high-speed railway subgrade. Aiming at the upward arching of the red-bed foundation, Wang et al. [7] took the mudstone at the second line of Lanzhou-Xinjiang Railway as the research object, discussing the effects of the top and side flooding methods on the swelling of mudstone on-site. Wang et al. [8] believed that its swelling characteristics are the main cause of the arching of roadbed. Besides, the rainfall infiltration on the ground has increased water absorption and swelling. Based on the test data of hydrological survey and ground stress, Zhong et al. [9] attributed the long-term deformation of roadbed to the stress concentration caused by deep excavation. So far, most previous studies have focused on swelling soft rocks in the traditional sense, but they have not paid enough attention to the time-dependent swelling and degradation mechanisms of typical red-bed mudstones in central Sichuan.
Regarding the main factors affecting the deformation of red-bed mudstone, some scholars considered the influence of temperature, humidity, saturation, water pressure, and confinement conditions on the swelling law of rock samples [10–15]. Other researchers have also explored in terms of applied load, anisotropy, softening, and creep behaviors [16–18], but most of them are limited to short-term deformation performance. The above research mainly started from the macrolevel, so that there are few systematic and comprehensive explanations of the continuous swelling and shrinking effects of red-bed mudstone in the cyclic dry and wet environment from the microperspective. Considering that the actual engineering environment is a combination of complete water immersion, capillary moisture absorption, and air moisture absorption [19–22], it is urgent to implement corresponding investigations on the time-dependent deformation regulation according to natural conditions.
To this end, the typical red-bed mudstone in central Sichuan, China, was selected as the research object. Deformation behaviors of rock samples under different initial water content, partial water immersion, and gaseous moisture absorption were specifically conducted by designing a set of test equipment to control environmental humidity. Subsequently, combined with X-ray CT scanning digital core technology, the asymptotic deformation laws and degradation mechanisms of mudstone in cyclic wet and dry environments were, respectively, explained qualitatively and quantitatively.
2. Methodology
The purpose of this study is to clarify the time-dependent swelling and shrinking characteristics of red-bed mudstone, starting from two aspects and two levels in turn. The overall research framework of the experiment is shown in Figure 1.
2.1. Materials and Sampling
In the current engineering background, the deformation section of ballastless track subgrade at Neijiangbei Station in Sichuan, the excavation slope area near the West Station of Yibin high-speed railway, and the excavated mountains at Qianjiang Station in Chongqing were determined as sampling sites (Figure 2), followed by on-site geological survey and collection of rock mass (Figure 3). The three selected sampling areas are located in the middle, north-southwest, and east edges of the basin, showing regional characteristics. The geological distribution ages are mainly from the Middle and Lower Jurassic periods (Figure 2), with differences in the formation ages.
(a)
(b)
(c)
(d)
The red beds in central Sichuan belong to soft rocks, which have poor cementing ability and are easy to weather and disintegrate. Although the weathered layer on the surface was cleaned up when the samples were collected, initial microcracks and fissures of the original rock developed under the influence of existing environment and geological structures, which were inevitably disturbed during the process of collection and transportation. Considering the difficulty of sample preparation, the undistributed rocks which meant not yet disturbed were processed into several standard cylindrical samples (diameter and height: 50 mm) by using a wire cutting process with an error of ±1 mm (Figure 3).
To analyze the natural properties and swelling abilities of red-bed mudstone, a series of parallel physical and mechanical tests have been carried out successively (Figure 4). Specifically, the tests of water content and dry saturated absorbed water ratio followed the standards of “Specification of soil test” SL237-1999 and “Code for Rock Test of Railway Engineering” TB10115-2014, and the density test referred to “Specification for rock tests in water conservancy and hydroelectric engineering” SL264-2016. The confined swelling ratio and free swelling ratio of samples can be obtained by unidirectional dilatometer and three-dimensional dilatometer, respectively. The swelling force was measured by a digital explicit swelling force tester. All the experiments were conducted in strict accordance with the “Code for Rock Test of Railway Engineering” TB 10115-2014. The statistics of physical parameters are shown in Table 1.
(a)
(b)
(c)
(d)
(e)
(f)
2.2. XRD and FT-IR Pattern
The swelling property of rock mass depends largely on the content of clay minerals and the degree of cementation. In general, the higher the content of minerals such as montmorillonite, illite, and kaolinite, the greater the swelling potential of mudstone [23, 24]. Mineral composition and quantitative proportion of samples were characterized by XRD diffraction pattern and infrared spectrum, respectively (Figure 5).
(a)
(b)
Undistributed samples collected on the site were ground and passed through 0.05-mm sieve, dried into 10~20 g powder. During the experiment, XRD test (SmartLab9, Japan) was used to determine the composition and content of rock powder (Table 2), while FT-IR analysis (Vertex 80 V, Germany) could infer the chemical composition and structural differences of rock mass by detecting the changes of spectral absorption and absorption form (Figure 5).
Typical red-bed mudstone in central Sichuan is mainly composed of clay minerals such as montmorillonite, illite and kaolinite, and nonclay minerals such as quartz, albite, and calcite (Table 2). The diffraction peaks of some clay and nonclay minerals are marked in Figure 5(a). Observing Figure 5(b), the IR spectra of three samples are very similar. Taking Neijiang sample as an example, there are two obvious absorption bands in the high-frequency region of the infrared spectrum. One band near 3623.08 cm-1 is caused by the stretching vibration of Al-O-H, and the other band near 3423.73 cm-1 can be attributed to the stretching vibration of H-O-H in interlayer water molecules, corresponding to the bending vibration of H-O-H near 1639.63 cm-1 in the middle band region. In the mid- and low-frequency range, the strong absorption bands around 1030.19 cm-1, 779.06 cm-1, and 778.09 cm-1 are generated by the antisymmetric stretching vibration of Si-O-Si in montmorillonite. The absorption rate of the Neijiang sample to the infrared spectrum at this location is stronger, indicating that its content of montmorillonite is relatively high. In addition, the two strong absorption bands in the low-frequency region are at 520.42 cm-1 and 470.49 cm-1, respectively, which can be considered as the coupling vibration of Si-O-M (metal cation) and M-O of montmorillonite. The above results are consistent with the quantitative data, and the content of montmorillonite in Neijiang samples is the highest, up to 19.67%-20.7%, followed by Qianjiang and Yibin. On the whole, the mudstone contains a large proportion of clay minerals, about 60%~75%, so it has the strong swelling potential.
2.3. Scheme of Swelling Test
Indoor swelling tests are conducted to study the time-dependent deformation of red-bed mudstone under natural conditions. It not only monitors the macroscopic swelling of undisturbed rock mass under static water immersion from different initial water content, partial water immersion, and gaseous moisture absorption, but also uses controlled humidity technology to simulate the dynamic process of samples changing with dryness and wetness.
2.3.1. Initial Hydrated State
The test first processed the selected undisturbed rock mass into a cylindrical rock sample with a height-to-diameter ratio of 1 : 1 and discussed the relationship between water content and swelling capacity. Among them, the control group was the natural undisturbed sample, and the water content corresponded to the natural water content. In order to reduce disturbance, 4 sets of parallel swelling tests were quickly carried out after sealing and sampling. Initial water content was all 4.5%. The test group used natural air-dried samples and placed the samples in a ventilated manner at room temperature. The experiment was started when the water content stabilized to 2%. In view of the fast rate of softening and disintegration of mudstone after being immersed in water, the sample was confined with iron foil and hoop. Besides, permeable stones were placed at both ends. The vertical deformation was monitored by a dial indicator with an accuracy of 0.01 mm. When the change in the readings of the sample for two consecutive times is less than 0.01 mm, it can be considered that the mudstone deformation has been basically stabilized, but the length of time immersed in water should not less than 100 h.
2.3.2. Way of Soaking
Two groups of undistributed Neijiang samples were chosen for deformation tests of different way of soaking. Each group contained 4 parallel samples. Specifically, the first group was partially immersed, and the height of the immersion reached half of the total. The second group was completely immersed in water, and the sample was in complete contact with water. The rest of the conditions were consistent with the above test and will not be repeated.
2.3.3. Gaseous Moisture Absorption under Controlled Environmental Humidity
To eliminate the influence of air humidity on the swelling performance of mudstone, it was proposed to adopt controlled humidity technology to keep the environmental humidity constant during the test. Since the electrolyte is dissolved in water to form a solution, the water, air, and electrolyte in the closed container space will be in a three-phase equilibrium state [25]. When the saturated solution and the air above coexist in a closed environment, the humidity in the container tends to a stable value. At this time, the suction can be controlled by adjusting the environmental humidity.
The mathematical expression of air humidity can be described with reference to the common suction of unsaturated soils, conforming to the principles of thermodynamics [26]. The chemical potential of liquid water in an unsaturated medium can be expressed as
Similarly, the chemical potential of steam-water in an unsaturated medium is where and represent chemical potentials of liquid water and steam-water in the current state, respectively, while and , respectively, are chemical potential of liquid water and steam-water in the reference state. Accordingly, and refer to the pressures of liquid water and steam-water, respectively, in the current state, while and , respectively, indicate the standard atmospheric pressure and saturated vapor pressure in the reference state. Furthermore, , , and are the ideal gas constant, environmental temperature, and relative mass of water molecules, respectively.
When environmental temperature is the same, liquid water will be in full contact with steam-water. Thereby, solid, liquid, and gaseous media have thermodynamic dynamic equilibrium. The chemical potentials of them are equal:
According to the description of standard relative humidity of the saturated solution by the International Organization [27], the K2SO4 saturated salt solution was determined as the constant humidity solution by considering the chemical properties of electrolyte, stability of temperature, economy, and other factors. When the temperature ranges from 5°C to 50°C [28], the relative humidity changes by ±1%. To avoid interference from other ions in the water, distilled water was used for the saturated salt solution. Since the closed container will be indirectly or directly in contact with electrolyte during the test, the material of constant humidity box must have the characteristics of high-temperature resistance, corrosion resistance, nonhydrophilicity, and easy cleaning. Figure 6 is a schematic diagram of the constant humidity box. The box was made of transparent polypropylene, and the size was . In this experiment, cylindrical samples with a high content of swelling clay minerals (50 mm in diameter and height) from Neijiang were selected for the study. Confined treatment was adopted to limit lateral deformation. Before the test, the undistributed samples and the mudstones dried at 60°C to constant weight were regarded as the research objects so as to examine the gaseous moisture absorption law. Because mudstone usually deforms very slowly in the air by moisture absorption, a higher-precision dial gauge was operated, and the needle was placed in the middle of the top surface of the sample to monitor its vertical deformation.
Existing research suggested [29] that the humidity field is not uniformly distributed inside the container. The closer to the liquid level, the more the environmental humidity meets the international recommended standard. To this end, saturated salt solutions were installed symmetrically on both sides along the length of box. It is not only beneficial to ensure the contact between the upper air and the solution, but also facilitates timely replacement without disturbing the sample. The box adopted a double-layer sealing method, and the lower sealing layer and the upper sealing layer were, respectively, a large-size transparent cling film and a matching cover, thereby preventing the exchange of humidity with the air. By monitoring the dial indicator readings from the outside in real time, data disturbance can be reduced.
2.3.4. Dry-Wet Cycle
The above-mentioned experiments analyzed the factors that caused the deformation of red beds in central Sichuan and the resulting swelling and shrinking characteristics. In fact, mudstone in its natural state has been in a variety of external environments, such as temperature alternation, surface rainfall, and evaporation. When the external environment changes dynamically, the changing water content will induce the structure of mudstone to continue to deteriorate, which will produce swelling that evolves over time.
Deformation development law of mudstone was restored by dry-wet deformation test. The test also adopted a standard cylindrical sample cut from a typical Neijiang sample (Figure 3), and the process of moisture absorption was confronted with the above-mentioned swelling test. In the dehumidification stage, the sample was separated from the water, and the ambient humidity was controlled by a saturated salt solution to lose water. When the shrinkage deformation becomes stable, record the amount of dry deformation and again immerse in water for next test. In principle, the sample can be continuously tested for multiple dry and wet cycles without disintegration and peeling. However, after the second dry-wet cycle, it was found that there were crumbs and peeling on its surface. Therefore, it is no longer suitable to repeat the cycle, but the follow-up microscopic inspection can be carried out.
2.4. Microscopic Observation
Mineral size, arrangement, and morphology inside the sample before and after the dry-wet cycle can be qualitatively described by the Zeiss Xradia 410 Versa high-resolution micron CT scanner produced by Carl Zeiss, Germany. Three-dimensional pore, pore size, and connectivity information of rock structures under dry-wet cycle was quantitatively obtained by Dragonfly software. The working parameters of this CT scan are working voltage 120 kV, current 150 μA, exposure time 0.8 s, scanning accuracy 27.8 μm, and scanning time 40 min. Before further analysis of the pore structure in the images acquired by computed tomography, binary segmentation threshold processing should be required [30–32].
The macroscopic behavior of geotechnical materials strongly depends on the microstructure [33]. Through image processing technology, qualitative and quantitative postprocessing can be performed on the CT samples. On the one hand, the distribution of porosity along the height of the samples and the evolution of porosity with drying-wetting cycles can be analyzed based on the 3D pore-scale model. On the other hand, the geometric information of pores can also be quantitatively characterized from the volume, surface area, aspect ratio and orientation. Accordingly, the changes of the above parameters under the dry-wetting cycle will be explained, which is of great significance for further clarifying the microscopic features of macroscopic deformation.
3. Results and Discussion
3.1. Swelling Test
3.1.1. Impact of Initial Hydrated State
Figure 7(a) depicts the swelling results of first set of water immersion tests. On the whole, the deformation of three types of specimens gradually increased with time and eventually stabilized. At the initial stage of deformation, the vertical strain rose sharply and linearly, suggesting that clay minerals and water in the mudstone may have a strong physical and chemical reaction. Within 9~12 h, the sample entered the stage of slowing down the swelling rate, corresponding to the inflection point of linear deformation and before the stable strain. In this stage, it deformed slowly, and the overall rate began to decrease. With the increase of time, the vertical deformation of mudstone immersion in water had been basically completed within 60~80h. Observing the time-dependent swelling curves of undisturbed mudstone in the three regions in Figure 7(a), it was found that the Neijiang sample had the largest stable deformation, reaching 1.78%. The other two groups were relatively close (Table 3), which may be related to the content of clay minerals such as montmorillonite.
(a)
(b)
Figure 7(b) characterizes the swelling behaviors of natural air-dried sample over time. The curve as a whole can also be divided into three stages: rapid growth, decelerating growth, and gradual stabilization. Comparing the deformation curve in Figure 7 and the summary data in Table 4, it can be concluded that natural air-dried sample has a shorter duration in the rapid growth phase. From 5.5 to 7.3h, it entered the stage of slow growth, and the swelling strain had stabilized within 14–48h, reaching 2.06%–2.25%. The reason for the above phenomena is that natural air-dried sample has a lower water content and a greater swelling potential.
3.1.2. Different Way of Water Immersion
Figure 8 records the infiltration of the mudstone surface under partial immersed in water through snapshots. At the beginning of the test, there was no trace of wetting on the surface of the sample after being immersed in water for 1 min. After 15 minutes, the rock mass showed only a small amount of infiltration at the side edge. 3 hours later, it was found that the upper surface of the sample was completely wet, implying that as the immersion continues, even if the mudstone is not in direct contact with water, it will gradually be infiltrated under the action of the humidity gradient and capillary suction.
(a)
(b)
(c)
Figure 9 presents the comparison of the swelling curves of Neijiang samples in the complete and partial immersion test. Obviously, the deformation of partial immersed in water was less than that of completely immersed in water. Moreover, the development time required for the swelling of partially immersed samples was slightly shorter than that of completely immersed samples, and its stable swelling strain was about one-half of the latter. The above phenomena have proved that although the upper part of samples is not in direct contact with water, it still swells under the condition of capillary moisture absorption, which promotes the release of swelling potential to a certain extent. From the perspective of stable swelling time, rapid growth and decelerating growth of the partially immersed sample had a shorter duration and entered the stage of gradual stabilization relatively earlier.
Observing Figure 9, partially immersed samples still had a slow and slight growth within 25-100 h after the stage of decelerating growth, but the completely immersed sample had remained unchanged within 40-60 h, suggesting that slow moisture absorption may be the main reason for the difference in deformation between the two. In order to verify this conjecture, the following will discuss the law of moisture absorption and deformation of mudstone in a controlled environmental humidity.
3.1.3. Controlled Environmental Humidity
Figure 10(a) plots the measured values of relative humidity RH and temperature inside the device in the range of 0~225 h. RH in the container reached more than 90% at 1 h and was always close to 99% after 3 h, which was in line with the range of environmental humidity in central Sichuan. By monitoring the temperature in the container, the distribution interval was 18.4~21.1°C, which had a limited influence on the RH. Judging from RH during the test, the expected controlled effect has been achieved. Therefore, the designed controlled humidity box is reasonable and can strictly adjust the relative humidity in the closed container.
(a)
(b)
Gaseous moisture absorption curve under controlled humidity is shown in Figure 10(b). The test lasted more than 1 month. During 0-700 h, the two samples continuously absorbed moisture from the environment and produced aging deformation. Throughout the test period, the sample always kept a slow swelling. There was no obvious trend of convergence before the end of the experiment. In detail, the swelling strain of natural air-dried sample was 0.471% after 1 month, while the moisture-absorbing swelling rate of undistributed sample was 0.382%, again indicating that swelling potential of the dried mudstone was fully released. As for the changes of the curve, it is found that the swelling trends of two samples were similar. Swelling strain continued to rise slowly with time, and swelling rate after 1 month was about the same. The difference is that since natural air-dried sample has a stronger swelling potential, its initial swelling rate is relatively large. In general, gaseous moisture absorption of Neijiang mudstone experiences an extremely slow process with long-term time-dependent deformation characteristics.
Based on the above results, initial water content and way of soaking greatly determined its amount and rate of swelling, displaying a short-term effect. The curve roughly included three stages: rapid growth, decelerating growth, and gradual stabilization. Partial water immersion test demonstrated that nondegraded gradual swelling created by humidity had a certain time effect, while the moisture-absorbing deformation process of mudstone under controlled humidity conditions was longer. If there is a certain humidity gradient between mudstone and surrounding environment, red-bed roadbeds, slopes, and tunnels are likely to bring about long-term deformation and damage.
3.2. Dry-Wet Deformation Test
3.2.1. Cyclic Deformation Curve
Figure 11 presents the dry-wet deformation curve under different initial water content. Next, take the Neijiang samples with an initial water content of 5% as an example to discuss segmentation characteristics of the curve in the first and second cycles. For the first cycle, the swelling strain increased rapidly at the initial stage of water immersion. The strain was in a stable state between 10 and 100 h. The first dehumidification started at 91.83 h and ended at 311.83 h. When the sample was just out of contact with water, a small increase in strain was observed. On the one hand, the sample may generate uneven shrinkage stress due to the anisotropy of spatial distribution of dehumidification rate and clay minerals in the initial stage of dehumidification, resulting in rock dilatancy. On the other hand, moisture was continuing to be distributed throughout the sample, finding unswollen pockets of clay minerals elsewhere. However, with further dehumidification, shrinkage strain accelerated and declined in the interval of 114-192 h. After a small adjustment of the deformation at 192 h, decreasing trend of the curve slowed down.
According to the overall characteristics of curve in the first cycle, it can be approximately divided into three stages. The first stage was the adjustment stage of shrinkage, corresponding to section A in Figure 11. At this stage, the sample has just undergone dehumidification, in which the outside was the first to dehumidify so that its speed was faster than the inside. The resulting uneven tensile stress delayed the shrinkage strain at a large part, manifested as a small range of deformation growth. The second stage was the accelerated shrinkage stage, as shown in section B of Figure 11. This stage was characterized by the gradual acceleration of shrinkage. At this time, the rock formation that inhibited the external shrinking in the early stage gradually began to dehumidify, prompting the rapid release of accumulated shrinking potential. The third stage was the stable stage of shrinkage, namely, section C in Figure 11, which lasted about 120 hours. On the basis of the first two stages, dehumidification rate will be further reduced because of the decrease of humidity gradient, leading to the shrinkage tending to be stable over time. This may be due to the first loss of water in the fissures, but it has little effect on the bound water film between mineral layers. Only further water loss can induce the shrinking of layer spacing.
In view of the fact that the shrinkage deformation at room temperature was steady after 220 h, the start time of the second water immersion was set at 312 h. Comparing the curves of the two dry and wet cycles, it is found that the stable swelling strain of the second water immersion was 2.62%, which increased by 36% in contrast with the first time. It is confirmed that repeated drying and wetting has a strong promoting effect on the structural deterioration and the release of swelling potential. Similarly, deformation of the second water immersion stabilized in the range of 312-452 h. The curves of two dehumidification stages were basically consistent, and they have also gone through three stages. The difference lied in that shrinkage adjustment time of the latter was significantly shorter than that of the former. This may be because the wet-dry process causes further damage to the structure, so the moisture on the surface spreads faster. The stable shrinking strain after the second dehumidification was 1.85%, which increased by 127% compared to the first time. Thereby, under the effect of continuous drying and wetting, a large amount of irreversible plastic strain will be produced.
As shown in Figure 11, the curves of different initial water content were similar. Swelling strain of the sample with the initial water content of 3% after the first water immersion was steady was 2.46%, significantly greater than that of the rock sample with the initial water content of 5%. Moreover, the deformation after the first dehumidification was also higher than the latter, suggesting the lower the initial water content, the greater the swelling potential of mudstone and the stronger its swelling capacity. In contrast, the dehumidification curve of the sample with an initial moisture content of 3% also contained three sections, in which section A had a longer time. The sample did not undergo deformation growth at the initial stage of dehumidification. It may be that its surface has a higher degree of peeling. The shrinkage generated by dehumidification was greater than the deformation growth aroused by its own crack propagation. Therefore, the curve fluctuated up and down and shrunk slowly. The durations of sections B and C were close, and the boundary between the two stages was not obvious, meaning that the deformation of the inner and outer layers was coordinated during the dehumidification. At the same time, the stable swelling strain in the second cycle was not much larger than the first one. It could be due to the limited swelling capacity of mudstone under the current confined conditions, and its swelling potential has been fully released before the second dry-wet cycle. Based on this, it can be concluded that initial water content can make the mudstone deform more thoroughly in the early dry and wet, but it has limited influence on the final stable swelling and shrinking strain.
3.2.2. Pore-Scale Model
CT scan pictures were arranged according to a certain algorithm. Then, a binary three-dimensional digital model was generated after three-dimensional reconstruction (Figure 12). Accordingly, average porosity and porosity of the sample at different depths can be calculated (Table 4) to analyze the changes with depth and the number of cycles.
(a)
(b)
(c)
Figure 12 displays the visualized pore space of mudstone after several wet and dry cycles. From 500 continuous two-dimensional slices of statistics, 100 CT images were extracted sequentially from top to bottom; thus, the sample was classified into 5 layers, and the average value was taken to calculate the porosity. The change of porosity along the axial direction suggested that porosity of the slices showed significant anisotropy in the vertical direction. The closer to the end, the greater the porosity and the smaller the middle. The average porosity of the end under 0, 1, and 2 dry and wet cycles were as high as 0.227%, 4.027%, and 6.121%, respectively (Table 4). In longitudinal comparison, with continuous development of dry-wet cycle, porosity of the sample at each position gradually enlarged, and average porosity of the rock sample after 0, 1, and 2 cycles were 0.169%, 3.495%, and 5.054%, respectively (Table 4), proving repeated dry-wet cycles will promote uneven shrinkage stress inside the sample. The final result was an increment in the number of pores and pore volume (Figure 12).
3.2.3. Spatial Distribution and Statistical Characteristics
By statistically extracting three-dimensional visualized pore network model, the following structural information can be further acquired: the number, size, volume, orientation, and shape of pores (Figure 13). Due to the constraints of test conditions, parallel samples were used at different cycles. Controlled by original depositional environment and unloading, although undistributed sample contained a small amount of horizontal lamellar cracks, it did not affect the subsequent discussion of the evolution of the pores with several drying and wet cycles.
(a)
(b)
(c)
(d)
(e)
For Figures 13(a) and 13(b) and 14(a)–14(c), the number and density of pores obviously climbed after wet and dry cycles. The distribution range of its volume has changed from 0.02~2.51 mm3 to 0.02~2.58 mm3. The distribution range of corresponding surface area has also expanded from 0.41~37.70 mm2 to 0.41~47.18 mm2. Hereby, under the action of wet and dry cycles, pore size has been upgrading. Pores, microcracks, and cracks will enlarge and continue to generate more small pores. As shown in Figure 14, after one cycle, small pores ( and ), microcracks ( and ), and large cracks ( and ) accounted for approximately 87.42%, 10.64%, and 1.94%, respectively. Similarly, when the second cycle ended, the proportions of the three were as high as 90.70%, 8.22%, and 1.08%, respectively, reflecting that the pores inside the rock sample were mainly in the form of small pores, followed by microcracks and finally large cracks. By observing spatial geometric distribution of the pores along the axial direction in Figures 13(a) and 13(b), it could be further found that the number and density of pores gradually decreased from the ends to the middle, which is consistent with the quantitative results in Table 4. Simultaneously, volume and surface area of the pores were the largest in the outer range, while they were relatively smaller closer to the center of the circle, which may be related to the pore connectivity caused by external water seepage.
(a)
(b)
(c)
(d)
(e)
(f)
Geometric indicators such as Phi and Theta jointly determine the specific orientation of the pores in the spatial network, referring to the angle between the projection of vector on the XOY or XOZ plane and x- or z-axis, respectively. Combining Figures 13(c) and 13(d), original cracks in the sample before the wet-dry cycle were parallel to the XOY plane. When the first cycle was finished, the Theta of pores was mainly concentrated near the polar angle. Distribution intervals of Phi were approximately 0°~10°, 35°~50°, and 78°~90°. This stage is roughly close to the result of a tree-like fissure. After the second cycle, Phi and Theta were spatially isotropic. The range was similar to that after one cycle, and at this time, it can be approximated to a relatively uniform net-like pore structure.
The above analysis conveyed that with the participation of wet and dry cycles, the pores of mudstone transition and develop from “tree-like” to “net- like”. In terms of the changes in the diameter-to-length ratio (the ratio of diameter to length) (Figure 13(e)), the intervals corresponding to the first and second cycles were 0.02~0.76 and 0.01~0.87, respectively, with 0.22~0.44 dominating. In addition, the diameter-to-length ratio of pores was inversely proportional to the number and density, implying that the number of pores of small size was the largest, followed by the larger size of microcracks, and the largest size of fissures were the least. This finding was confronted with the results of Figures 13(a) and 13(b) and 14, fully verifying the reliability of the foregoing conclusions.
4. Conclusions
Based on the real environment that mudstone may encounter, effects of swelling of the red-bed mudstone have been discussed. Through CT scanning, macroscopic deformation and deterioration mechanism were explained from a microscopic view. The key findings are summarized as follows: (1)Time-dependent deformation curve of the red-bed mudstone during water immersion generally included three stages: rapid growth, deceleration growth, and gradual stabilization. Different initial water-bearing conditions will promote the difference in the swelling characteristics. For the undistributed and natural air-dried sample, the latter had a lower water content, greater swelling potential, and stronger short-term swelling ability(2)Comparative tests of partial and complete water immersion suggested that humidity gradient was a key factor affecting swelling performance. Moisture-absorption deformation curve in liquid water also displayed short-term behaviors. Swelling strain of undisturbed sample and natural air-dried sample at 700 h under controlled humidity were 0.382% and 0.471%, respectively. The entire process of gaseous moisture absorption exhibited long-term characteristics(3)Dehumidification curve can be roughly divided into three stages of adjusted shrinkage, accelerated shrinkage, and stable shrinkage. The difference in spatial dehumidification rate and anisotropy distribution of clay minerals will produce uneven shrinkage stress, leading to a slight rise in the initial stage. The number of cycles and initial water content jointly influenced the deformation of mudstone. Initial water content determined initial deformation by controlling the release degree of swelling potential, but it had a limited effect on the final stable strain(4)Affected by external water seepage, porosity of the slices had significant anisotropy along the axial direction. With the participation of dry and wet effects, average porosity after 0, 1, and 2 cycles were 0.169%, 3.495%, and 5.054%, respectively. The specific manifestations were as follows: the number and density of pores increased, the volume and surface area enlarged, and the pore size was upgraded. Orientation of the pores before and after the cycles were all near the polar angle, following the development law from “tree-like” to “net-like”
For the swelling test of gaseous moisture absorption of mudstone in a controlled humidity environment, it was found that the strain has not converged during a month. In the future, we will investigate its long-term deformation under different humidity. Accordingly, a constitutive model of the red-bed mudstone with humidity will be established, which can provide a necessary theoretical basis for the inversion and prediction of long-term uplift deformation of high-speed railway subgrade in central Sichuan red-bed.
Data Availability
The data used to support the findings of this study are included with the article.
Conflicts of Interest
The authors declare that they have no known competing financial Interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This research was financially supported by the National Natural Science Foundation of China (Grant Nos. 42077270 and 41702337).