Experimental Study on the Mining-Induced Water-Resistance Properties of Clay Aquicludes and Water Conservation Mining Practices

The Yushen mining area contains thin bedrock and a shallow buried coal seam, where JingLe group Hipparion clay and Lishi loess serve as a high-quality cement insulation cover. This study investigates the properties of the clay layer to determine the eﬀect of the clay aquiclude on the mining water variation and fracture characteristics. Unloading hydraulic jack experiments were performed to test the physical and mechanical properties of the clay layer and the structure was analyzed in detail. The experimental results show that mining aﬀects the soil cracks, leading to crack opening and subsequent bridging. The permeability coeﬃcient of the soil layer initially increases with increased unloading and then decreases. A theoretical model is developed to determine the recovery mechanism of the clay layer water insulation based on the spatial movement of the clay. The results indicate the formation of a waterproof cover type of coal mud protection. Design methods are proposed to optimize the coal pillar size. Mining damage leads to the formation of a mud-covered bridge belt, which can be designed to appropriately reduce the protective layer thickness. The model is applied to the Hao Jialiang 2301 working face. The results provide important insight on the variation characteristics of the mining water insulation in clay layers and an important reference for accurately calculating the size parameters of waterproof protective coal pillars under mud-cap conditions to increase the upper mining limit of the working face.


Introduction
Coal resources and mines are abundant in Yulin-Shenmu (Shanxi Province, China). e mining area contains a large amount of shallow buried coal seams (<200 m depth) and thin bedrock layers (30-50 m or less). e regional geological conditions are generally simple and the coal quality offers special advantages. However, high intensity mining under these conditions is not conducive for controlling water. Mining under water bodies poses unique challenges, in addition to maintaining aquifer water stability and protecting the surface ecological environment [1][2][3][4][5]. Cohesive soil layers above a coal seam offer good water resistance and can form a high-quality cement barrier cap that plays an important role in protecting the surface aquifer, which differs from common loose sand layers in terms of water conservation and coal mining practices.
Mining research has mostly focused on ecological water levels [6][7][8][9], rock overburden height, crack development [10][11][12][13][14][15][16], geological conditions, water partitioning [17,18], and the application of water-coal technology [9,[19][20][21] and groundwater level monitoring networks [22]. Clay aquicludes exposed to mining deformation have received less research attention and only in terms of the engineering geological characteristics of the water-resistant layer and its physical mechanical properties. Yu tested the physical and mechanical properties of the clay layer in the Yushuwan coal mine and pointed out that the red and yellow soil layer has good water resistance and swells easily upon exposure to water [23]. Li et al. described the mining-induced variation characteristics of loess from the Lishi Formation and laterite of the Baode Formation after large-scale coal mining operations in northern Shaanxi [24], and Xu proposed the concept of an effective water barrier thickness [25]. Zhang et al. tested the hydrological properties of the overlying rocks in the Nalinhe area and showed that the argillite has good water stability, nonswelling, and nondisintegration properties and was not easily recovered after cracking and compression [26]. Miao et al. proposed the concept of a water barrier key layer and classified the hydrogeological structure types of the Shendong mining area [27].
e Yulin-Shenmu coal mine area is known for the occurrence JingLe group Hipparion red clay and Lishi loess that form a clay aquiclude owing to their high clay mineral contents. is causes the soil to crack to a lesser degree, which suppresses the lead water cracking [28][29][30][31][32] and resistance of the clay layer bridge. e waterproof effect after mining recovery remains poorly understood. A calculation method is required to optimize the waterproof and protective coal pillar size for thin bedrock and thick mud-cap drilling conditions. is paper takes the Hao Jialiang coal mine of the Yushen mining area as the engineering background and studies the isolation of the mining water clay layer based on fluid-structure coupling experiments. e results are applied to practical water conservation mining operations.

Physical and Mechanical Property Tests
Drill core samples of the Lishi loess and JingLe group Hipparion red clay were collected from the Hao Jialiang coal mine 2301 face in their original state. An MIP scanning electron microscope and X-ray powder crystal diffractometer were used to determine the composition of the clay minerals.
e test results are listed in Table 1. Red loess samples with different dry densities (1.5-1.9 g/m 3 ) were then selected to test the physical and mechanical parameters including the soil shear strength, water dilatancy, and saturated permeability coefficient using a GDS direct shear tester, an expansion-measuring instrument, and a GDS hydraulic controller, respectively. e test results are shown in Figure 1.
e red clay of the JingLe Formation is deep red in color and the Lishi Formation loess is earthy yellow. e soil particles are relatively loose with a small particle size and good homogeneity, and both contain abundant quartz and clay minerals. e clay mineral composition is dominated by Yimeng mixed beds of mostly illite. Figure 1 shows that the maximum shear stress of the red loess sample is 450-850 kPa under a normal stress of 100 kPa and gradually increases with increasing dry density. e red loess sample contains more hydrophilic minerals and its swelling ability is strong when exposed to water. e maximum swelling rate varies from 4.82% to 13.73% and also gradually increases with increasing dry density. e saturated permeability coefficient of the red loess samples is relatively small (<1 d/m), which indicates weak permeability that gradually decreases with increasing dry density. Compared with the sandy soil layer, the JingLe laterite has generally high shear strength, strong expansion, extremely weak water permeability, and good water insulation properties.

Mining-Induced Fluid-Solid Coupling Experiments
Fluid-solid coupling experiments were performed to analyze the variation characteristics of the water insulation performance, recovery of the water insulation ability, and variation of the soil permeability coefficient of the clay water insulation layer after mining damage.

Experimental Design.
e experimental design is shown in Figure 2. To simplify the experimental system, a clay layer was used in the lower coal seam, and bedrock and aeolian sandstone were used in the upper layer. e coal bed mining system is modified using a hydraulic jack with only the soil viscosity identified as the research object during the test. e experimental system mainly contains a model box structure, hydraulic jack, and test bench. e hydraulic jack is placed in the experimental stage through the grooves with the casing supporting plate. e external water sources are supplied by a water pipe to directly fill to the top to fill the model casing.
e model box is composed of toughened glass with dimensions of 90 × 45 × 45 cm (length × width × height). e box wall thickness is 1 cm and is well sealed to maintain the strong bearing capacity. e lower part has six supporting plates arranged in parallel with dimensions of 45 × 14.7 × 1.7 cm. ere is a 3 mm gap to allow a water channel between the two supporting plates. e experimental system contains six hydraulic jacks as support (labeled 1-6 from left to right in Figure 2). e hydraulic jack fuselage height is 25 cm and has a lift range of 0-13 cm. e top disc has a 5.5 cm diameter, 11 cm diameter base, and maximum lifting capacity of 3.2 t, which meets the experimental requirements.

Experimental Procedure.
Mining water isolation experiments were carried out on the clay samples according to the following procedure.
(1) Design of the experimental scheme. Hipparion red clay of the JingLe Formation was collected from the unmodified soil of the 2301 working face of the Hao Jialiang coal mine. Two experimental schemes were designed to compare and analyze the effect of the clay layer thickness on the mining-induced water insulation. e model dimensions for experimental    (2) Adjustment of the experimental system. Prior to carrying out the experiment, the system was adjusted according to the designed scheme. e hydraulic jack was first adjusted in a reasonable position with appropriate horizontal spacing to maintain the designed conditions. e height was then adjusted to allow 4 cm of space between the supporting plate and bottom of the box. e horizontal spacing of the supporting plate was adjusted to maintain a 3 mm gap between the two plates as a water channel.
(3) Building the experimental model. e materials used in the models were collected from the undisturbed red soil on-site. e raw materials were placed in the model box, compacted, paved to the designed design height, and left stationary for 24 h. A layer of waterproof material was applied around the inner wall of the box to prevent downward water flow. Water was slowly injected into the box through the external water injection pipeline. e designed water height of both experiment groups was 4.5 cm, and the water was kept at rest for 16 h. e height of the upper water body was held fixed to measure the initial permeability coefficient prior to mining. e experimental model is shown in Figure 3. (4) Unload the hydraulic jack. e hydraulic jack was unloaded to simulate the coal seam mining process. Starting from jack 1 (Figure 2), the valve was opened to allow slow unloading until reaching 4 cm of the supporting plate to make complete contact with the bottom of the box. e conditions were then maintained for 8 h in a static state. e water penetration test was performed and the permeability was recorded. Jacks 2-6 were unloaded in turn and held for 8 h in a static state, and tests were performed to measure the k 2 -k 6 permeability coefficients.

Test Analysis.
After the water had fully permeated into the soil layer, the permeability coefficients of the experimental schemes 1 and 2 were determined to be k 10 � 1.45 × 10 −6 cm/s and k 20 � 1.414 × 10 −6 cm/s, respectively. e results show that the water permeated very slowly into the cohesive red soil layer and that the soil permeability coefficient was small for both schemes under static conditions. is indicates the good water insulation performance of the cohesive red soil layer. Experimental scheme 1 was taken as an example to further explain the variation characteristics of the cohesive soil layer and upper water during different unloading stages, as shown in Figure 4. Figure 4(a) shows that when jacks 1 and 2 were unloaded, the supporting plate at the top of the soil showed notable sinking and the water level significantly increased.
Supporting plates 2 and 3 above the vertical soil displacement responded poorly and the two plate transition positions above the soil show notably opened cracks, which can be observed in the experimental flow along the open fracture that infiltrated the soil and outflow through the gap in the plate.
When four jacks were unloaded, the soil above supporting plates 1-4 subsided significantly and the extent of vertical subsidence of the top interface was essentially equal. e opened cracks generated in Figure 4(a) had healed and decreased in size. However, notable vertical displacement differences occurred in the soil above the fourth and fifth supporting plates along with the formation of vertical cracks. e developed cracks were large in both width and depth. Water rapidly flowed into the soil body along the cracks and ultimately through the void between the supporting plates.
When six jacks were completely uninstalled, the soil displacement was equal throughout. e open cracks in Figure 4(b) decreased and closed owing to the balanced stabilization. Coupled with the expansion characteristics under the influence of water, serious cracks ultimately formed during the development of the automatic bridges, which evolved into multiple small cracks, as shown in Figure 4(c). e features of the mining-induced fractures can be observed from the clayey soil layer at the top. Figure 5 is analyzed to constrain the clay aquiclude characteristics when jacks 3, 4, and 6 were unloaded to monitor the jack fracture development characteristics. Figure 5 clearly shows a substantial decrease of the width of the new cracks that formed in the center upon unloading but that new cracks formed and continued to form below the bridging feature. is demonstrates that the cracks initially opened and then evolved into having bridge characteristics. e water insulation properties of the clay waterproof layer also accordingly changed during the experiment. In the initial static state prior to mining damage, the water permeability of the soil layer was good and the water inflow between the supporting plates at the bottom was very weak. During the mining disturbance, a certain number of cracks appeared in the soil layer along which water quickly penetrated. In the experiment, significantly more water flowed between the two supporting plates during the mining disturbance than prior to the mining disturbance. is indicates a weakened water permeability of the cohesive soil layer and enhancement of the permeability during unloading. However, after deformation and stabilization of the soil layer, the cracks in the soil layer healed and the water penetration into the soil layer slowed, thus also reducing the water inflow between the bottom supporting plates. is indicates that the water insulation of the cohesive soil layer can be restored to a certain extent after mining-induced deformation and stabilization. Figure 6, the permeability coefficient values of the two test groups were extremely small during the original static state prior to the mining disturbance. e increased amount of unloading continuously enhanced the extent of the mining disturbance in the clay soil and altered its mechanical properties, as reflected by an increased permeability coefficient with further mining disturbance. After unloading, the permeability coefficients of experimental groups 1 and 2 reached maximum values of 3.27 × 10 −6 and 2.82 × 10 −6 cm/s, respectively. Upon further unloading, the mining-induced fractures in the cohesive soil layer gradually decreased or even closed, leading to a gradual decrease of the permeability coefficient to 2.65 × 10 −6 and 2.02 × 10 −6 cm/s, respectively, once the deformation stabilized. However, the postmining permeability coefficient was still greater than the original value in the undisturbed state owing to the soil expansion. e crack widths and depths strongly decreased but did not completely heal. e soil infiltration path was shortened during deformation; thus the soil stability did not fully recover its waterproof ability. e results of the two different experimental schemes show a similar pattern but the increase of the soil permeability coefficient was less extreme for scheme 2 (16 cm thick soil layer) and decreased to a lower value after the peak than in scheme 1 (14 cm thick soil layer). is demonstrates that the overall deformation damage is more severe for thinner soil layers and the permeability enhancement is more extreme.

Recovery Mechanism Analysis
e experimental results show that water mining disturbances initially alter the properties of the water insulation clay layer but then somewhat restore its waterproof ability. Soil volume expansion is stronger under the influence of water expansion, which causes the original mining cracks to automatically heal and ultimately restores the soil water capacity.  e sticky soil mining water recovery process and associated mechanical isolation mechanism can thus be studied by analyzing the fracture evolution process.

eoretical Hypothesis of Clay Layer Spatial Movement.
Soil has reduced mobility under coal extraction conditions, which simultaneously subjects the lower and upper parts of the soil to tension and squeezing, respectively. Vertical shear stress is also produced under the action of gravity and loading. e combination of tension and shear stress produces vertical mining-induced fractures along the horizontal direction that develop from the bottom up. e water expansion characteristics of the viscous soil cause the mining cracks in the soil to gradually bridge, while the soil deformation stability remains essentially unchanged. e visible mining-induced opening/closing of fissures is closely related to soil movement. e variation of mininginduced fractures can be expressed by the spatial movement of the clay soil. e following model assumption is taken to study the viscous soil layer cracks and their mining mechanics mechanism. Soils that are affected by mining disturbances generate mining-induced fissures owing to water swelling of the clay layer, and mining cracks form during the process by which the soil moves to automatically bridge the cracks. Soil stability upon fracture closure is thus related to the soil movement.
Stochastic medium theory has become the most widely used in recent years to study the movement of subsided strata [33]. In the random medium theory calculation method, the spatial movement of clayey soil is analyzed using an infinitesimal element within a soil P (X, Y, Z) and its trajectory can be decomposed into horizontal and vertical directions, which are influenced by the dominant mining disturbance part of the soil in the vertical direction to approximate the degree of soil spatial movement. e calculation area selected here corresponds to the fluid-structure coupling mining water isolation experimental model in  Figure 7.
According to the theoretical hypothesis model, the generation of vertical displacement of an element body indicates that it is moving and that a certain mining-induced fracture will occur in this region. However, when the vertical displacement is reduced to zero, the element body stops moving and the fracture in this region automatically closes. e calculation of the vertical displacement of the element body is solved numerically using the Gauss-Legendre method and the model dimensional parameters are substituted into the formula. e superposition calculation is carried out using MATLAB software.

Evolution Process of Mining-Induced Fracture Opening-Healing.
e calculated evolution process of the mininginduced fractures during different stages of working face advancement is shown in Figure 8 based on the analysis of the relationship between the vertical subsidence displacement of the clay soil and mining-induced fractures. e soil layer near the working face is most strongly affected when the working face advances to 15 m, during which the soil layer is in the most active movement and deformation stage, and the vertical subsidence displacement strongly varies. Mining disturbance causes fractures to develop vertically in this area. When the working face advances to 30 m, the same soil range is affected and new cracks are generated. However, the rear part of the soil simultaneously gradually stabilizes and the vertical subsidence displacement essentially returns to the initial value owing to the viscous soil dilatability and crack bridging that occurs when the working face advances to 15 m. Similarly, when the working face advances to 45 m, new fractures occur in the soil layer near the working face that are severely affected by the mining disturbance. e generated fractures also undergo automatic healing at this point. e formed cracks thus ultimately close upon working face advancement and the mining-induced water isolation of the soil layer is weakened. After the movement and deformation are stabilized, the original cracks in the soil layer are automatically healed and the overall waterproof property of the soil layer is restored to a certain extent.

Water Conservation Mining in Mud-Capped Coal Seams
According to the literature [34], waterproof and protective coal pillars must be set up to ensure that no additional increase of mine water inflow occurs during the mining process. e coal pillar size should be greater than the sum of the maximum height of the overburden water-conducting fracture zone and protective layer thickness to avoid allowing the fracture zone to spread to water body. Experimental studies have shown that the clay layer will change after the first fracture opening process. Upon bridge repair, the clay layer also shows good waterproof performance by forming a mud-covered bridge belt. e bridge limits the height of further water crack development and can also serve as a waterproof coating. e coal seam thickness is used to calculate the protective coal pillar size using the traditional method, but the accuracy is limited and partially conservative owing to the low recovery rate and high waste of coal resources. e effect of the mud-cap damage rule in overlying rock mining should therefore be applied using the traditional method to optimize the coal pillar size and determine suitable new methods for designing mud-cap waterproof coal pillars.

Optimal Design of Waterproof Coal Pillar in Mud-Covered
Coal Seam. Using the conventional method to calculate a nearly mud-covered type of coal pillar can be problematic owing to damage changes of the overburden rock. e result is influenced by the viscous soil inhibition rock, which leads to a strongly reduced height of the developed cracks and large waterproof coating thickness. e mud coal pillar design must thus account for the development of the fracture zone height and protective layer thickness.
Mud-capped rock leads to crack heights that depend on their own unique destruction behavior ( Figure 9) and water loss. Borehole drilling, computed tomography (CT) scans, digital imaging physics of similar material simulations, direct or indirect testing, and a large amount of field measured data show that the mud-cap effect typically leads to the development of a fracture zone height that is 0.5-0.7 times the thickness of the fully developed bedrock layer. A reasonable thickness of the protective layer should therefore account for the water insulation of the mud cover sealing zone that forms after the automatic sealing of the mininginduced cracks on top of the fracture zone. ere is no need to leave a protective layer if the water insulation of the mud cover sealing zone is good and can exert an adequate water insulation effect. However, if the water insulation of the mud  Advances in Civil Engineering cover bonding zone is weak, a certain primary cohesive soil layer thickness should be left on the upper part of the bonding zone as a supplement. Because the water insulation properties of the primary cohesive soil layer are strong, the supplementary thickness left can also be appropriately reduced by 0-3 times the mining height. According to the water insulation effect of the mud-capped bonding zone, it is comprehensively determined that the protective layer thickness should be 0-3 times the mining thickness. e actual water insulation effect of the mud-capped bonding zone can be determined by mining water insulation tests of the reclaimed rock mass.
However, the key to the size design of a waterproof protective coal pillar when mining a mud-capped coal seam is to accurately assess whether or not the overburden failure produces a mud-cap effect. If the overburden failure does not produce a mud-cap effect, the calculation can be carried out using the traditional waterproof coal pillar formula. If the overlying rock failure causes the mud-cap effect, it is necessary to optimize the coal pillar size determination method.
e formation of a mud-cap effect owing to overburden failure is closely related to the rock structure, hydrogeological characteristics, as well as the physical mechanics and hydrologic characteristics of the rock and soil mass. e following three conditions for generating the mud-cap effect by overburden failure are summarized according to a large amount of field data measured in the early mining stage.
(1) e minerals of the mud cap have certain crack resistance, expansion, and water-resistance properties.
A method to optimize the size and design of the waterproof protective coal pillar for the mining of mud-capped coal seams is shown in Figure 10 and the specific steps are listed as follows.
(1) e lithology and thickness of each layer above the coal seam roof are determined according to surface exploration drilling and a stratum structure profile is drawn. e mineral composition of the rock and soil body is also determined, as well as the basic physical mechanics, geoengineering, and hydrological properties. (2) e production of a mud-cap effect owing to overburden strata failure is assessed according to the occurrence conditions of the overburden mud-cap effect combined with the mining parameters of the working face, strata structure, and hydrogeological characteristics. (3) If the damaged strata can produce a mud cover effect, the section water loss can be assessed by boring, drilling, drilling a CT scan, digital imaging physics of similar material simulations, and direct or indirect methods. e mud-cap damage rule of overlying rock can be detected to assess the development of the crack height. (4) e position and thickness of the overburden mudcap bonding zone are analyzed and the mining water isolation is tested through experiments to determine the reasonable protective layer thickness. (5) If the sum of the developed water conduction fracture zone height and protective layer thickness does not exceed the distance between the coal seam and aquifer, safe mining and water conservation mining can be operated without the threat of roof water damage. If the sum exceeds this value, the safety requirements cannot be met. e mining scheme design of limited thickness mining, wide strip mining, filling mining, and room and pillar mining and the technical measures for water prevention and control and water conservation are given.

Determination of the Waterproof Pillar Size in the Hao Jialiang Coal Mine.
e size of the waterproof protective coal pillar for a mud-covered coal seam is optimized taking the 2301 working face within coal seam no. 3 of the Hao Jialiang coal mine as an example. e average extraction length is 7 m, the oriented face is 220 m thick and 1700-m long, and the large mining height has broken owing to fully mechanized mining. e coal seam is buried to a depth of 110-160 m. e bedrock thickness is 12-45 m and deposited as a 38.4-62.4 m thick layer above the Neogene bedrock of JingLe red clay, Quaternary Lishi group, and the Quaternary Sarah WuSu aquitard, which is a wind-blown sand layer. e Holocene roof is a typical combination of thick "BoJi"-type soil strata lithology. e traditional linear regression formula projects a vertical height of the waterproof protection coal pillar of 100.37-104.45 m. We use the proposed method of mudcapped coal seams to check this value according to [35][36][37] through drilling, drill loss observations, and digital imaging methods such as physical analog simulations integrated to determine the water flowing through the fracture zone of the 2301 face at a height of 49.3 m. Soil regeneration tests have shown that the sticky soil of the Hao Jialiang coal mine has good water retention properties and the design of a   Advances in Civil Engineering waterproof protection coal pillar cannot serve as a protective layer. e fractured zone height is thus a minimum waterproof protection coal pillar size, yielding a value of approximately 50 m for the 2301 working face. e calculated height of the waterproof protective coal pillar is shown in Figure 11 compared with the bedrock section and laterite layers within the working face.
A comparative analysis of Figure 11 shows that the height of the waterproof protective coal pillar calculated using the traditional method will break through the Neogene red soil layer. e pore fracture division of the loess overlying the loose Quaternary rock and Salawusu Formation will thus pose a threat to safe production of the working face. It is therefore evaluated that the total mining height of the 2301 working face should be 7 m at a time. e optimization design method is used to determine the height of a waterproof protection coal pillar that does not break through the Neogene red soil and Quaternary unconsolidated rock. e coal pillar design is therefore reasonable and meets the requirements of water prevention and control of the working face.

Evaluation of Water Conservation Mining Effect of a Mud-Capped Coal Seam.
A low-lying area riverbed occurs above the 2301 working face. After mining, the surface of the basin subsided. e subsided basin has a water depth of 1-5 m with a water area of nearly 20,000 m 2 and volume of 60,000-70,000 m 3 . Ground fissure and basin water accumulation above the working face are shown in Figure 12. e water depth of the basin was observed synchronously during the mining of the working face. e results show that water depth of the basin changed little after mining. e water inflow changes during working face mining are shown in Figure 13. e maximum and minimum water inflow values are 70 and 35 m 3 /h, respectively. e face directly fills water only for the Mesozoic clastic rock with cranny pore water and confined water and Quaternary unconsolidated rock, but not the surface water through the  Advances in Civil Engineering 9 water channel into the working face. Special water pumps are also arranged as transport and return air channels for drainage during the mining process. e production practice shows that no seepage accidents occurred in the working face and the water source in the roof aquifer poses no threat to the normal production. is not only ensures safe coal mining under the water body, but also ensures that the fragile ecological environment on the surface remains unaffected, which achieves the purpose of water conservation mining.

Conclusion
(1) e physical and mechanical property test results of the clay soil in the Yushen mining area show a relatively high clay mineral content of the soil, which is dominated by a Yimeng mixed layer of mostly illite. Compared with sandy soil, the clay soil in this mining area has the basic characteristics of high shear strength, strong expansion, extremely weak water permeability, and good water insulation. (2) e results of fluid-solid coupling mining water isolation experiments of the clay layer show that mining-induced fractures in the soil layer initially open with increased unloading and then close. e overall water isolation of the soil layer initially weakens and then recovers and the corresponding permeability coefficient of the soil layer initially increases and then decreases. e variation of the mining-induced water insulation characteristics can be measured by the movement of the soil layer. e evolution of the mining-induced fractures is analyzed using stochastic medium theory and the recovery mechanism of the mining-induced water insulation properties in the clay layer is expounded. (3) e mud-capped strata form a mud-covered bridge belt during mining. e existence of the bridge can reduce the coal layer thick and affect the waterproof protective coal pillar design. is approach is applied to the 2301 working face of the Hao Jialiang coal mine and the optimization design method is verified based on the measured mine water inflow data, which improves the working face mining upper limit.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.