A Review of Stability of Dam Structures in Coal Mine Underground Reservoirs

Abstract

Coal has remained the primary component of China’s energy structure, and high-intensity extraction has continued in the central and western coal-producing regions of China. In contrast to the abundant coal resources, water resources have become extremely scarce in these regions, creating a conflict between coal resource extraction and water resource conservation. The coal mine underground reservoir (CMUR), as a typical technology for combined coal and water extraction and water-preserving coal mining, has been applied in numerous mines in central and western China. This effectively alleviates water resource shortages and achieves the goal of water resource conservation. The CMURs utilizes the goaf created by longwall mining as the water storage space. The reservoir dam structure comprises coal pillars, which serve as protective coal pillars in the mining area, and artificial dam structures that filled the gaps between these coal pillars. The stability of the dam structure under the complex stress effects of hydraulic coupling has been identified as the key to maintaining the safe operation of the CMUR. The mechanical properties, stress field, fracture field, and seepage field (“three fields”) change mechanisms, as well as the research results on size optimization of coal pillar dams and artificial dams in CMURs, were systematically reviewed. The core content included the instability and failure mechanisms of dam structures under the comprehensive coupling effects of factors such as dry–wet cycles of mine water, long-term immersion, chemical effects of high-salinity water, dynamic and static loads, and cyclic loads. This paper is considered to have certain reference value for the study of the stability of dam structures in CMURs and to provide some guidance for the safe operation of CMURs.

1. Introduction

China has been recognized as the world’s largest producer and consumer of coal. With the development of China’s social economy, the demand for coal remained high. The overall energy resource situation in China has been characterized by a scarcity of oil, a shortage of natural gas, and a relative abundance of coal. In recent years, however, the proportion of coal as a primary energy source in China’s energy structure has gradually decreased [1]. In Figure 1, the structure of China’s energy consumption from 2013 to 2022 is depicted [2,3]. In 2023, the proportion of coal in China’s total energy consumption remained as high as 55.3% [4]. For a considerable period in the future, coal is expected to remain the primary component of China’s energy structure [5]. High-intensity coal mining has been found to damage groundwater and cause significant water resource waste. Approximately two tons of mine water is generated per ton of coal, resulting in about 8 billion m³ of mine water annually in China, with a utilization rate of around 40%. Each year, about 5 billion m³ of mine water is not effectively utilized, equivalent to 50% of the annual industrial and domestic water shortage in China (10 billion m³) [6,7]. China’s primary coal-producing regions, located in the central and western provinces (Shanxi, Shaanxi, Inner Mongolia, Ningxia, and Xinjiang), hold over 80% of the nation’s coal reserves and contribute more than 70% to its coal production. However, water resources in these regions only account for 6.7% of the national total and 0.3% of the global total [8,9]. A reverse distribution trend between coal resources and water resources has been observed, highlighting the prominent conflict between coal resource development and water scarcity in western China. Figure 2 illustrates the distribution of coal production and water resources across provinces in 2023. The coal-rich regions in central and western China are predominantly located in arid and semi-arid ecologically fragile zones, characterized by water scarcity and vulnerable surface ecosystems. Major coal bases experience severe water shortages, with mine area water supply primarily relying on groundwater extraction and the comprehensive utilization of mine water. Challenges in these mining areas included water sourcing, integrated water resource utilization, and water rights allocation. Therefore, the question of how to achieve both safe and efficient coal mining and the protection of groundwater resources has been identified as a major challenge faced by coal mining operations in central and western mining areas.

Due to the shallow burial of coal seams, large seam thickness, and thin overlying bedrock in the central and western regions, the full-height mining of coal seams inevitably caused the overlying strata to fracture and connect with aquifers, resulting in significant underground water leakage and loss [10,11,12]. In response to this situation, traditional passive water-preserving mining techniques, characterized by “blocking” underground water leakage, were previously employed in coal mines. Techniques such as limited height mining, backfill mining, and partitioned mining were utilized to ensure that the overlying strata fractures caused by coal extraction did not connect with aquifers, thereby preventing damage to the aquifer structure [13,14,15]. Within the coal mine field, groundwater was found to converge towards the mining area (goaf). It is estimated that every 10,000 t of coal produces 1000 to 1600 m³ of void space. These spaces provided conditions for storing mine water and were considered ideal locations for mine water storage [16]. Based on the characteristics of the collapsed space in the goaf and drawing on traditional underground reservoir construction concepts [17], a new concept was proposed by Gu to address the phenomenon of water resource evaporation and loss during coal mining in western China’s mining areas. This concept involved constructing CMURs in the goaf caved zones to achieve underground storage and the utilization of mine water, representing water resource protection technology characterized by “conducting” [6,7,18,19]. This technology enabled the safe underground storage and utilization of mine water, preventing evaporation losses associated with surface discharge. The theoretical framework and technical system of CMURs were widely promoted and applied in mining areas such as Shendong. With a maximum storage capacity of 35 Mm³, these reservoirs supplied over 95% of the production, domestic, and ecological water needs of the mining areas, and also provided water for nearby power plants and coal-to-oil projects, achieving a transition from major water consumers to major water suppliers [20].

The concept of CMURs and the strategy of storing and using mine water underground adopted in China differed from the Western approach, which only used abandoned mine tunnels as water storage spaces. Underground reservoirs in Western countries primarily utilized abandoned mines. After these mines were abandoned, groundwater accumulated in the abandoned spaces through mining-induced fractures and seepage, ultimately forming flooded mines. Examples include the Martelange mine in Belgium [21], the Freiberg mine and Staßfurt mine in Germany [22,23], the Gaspé mine in Canada [24], the Barredo-Figaredo underground reservoir in Spain [25], and the abandoned mines in the Appalachian coal region of the United States [26]. Such abandoned flooded mines did not require consideration of mining impacts, and the storage of mine water did not necessitate pre-planned reservoir construction based on the mining sequence. Therefore, the artificial construction of dam structures for underground reservoirs was not necessary. In China, water was stored in the rock voids of the goaf formed after longwall coal mining. The reservoir dam structure consists of safety coal pillars supplemented by artificial dam structures. Water intake and extraction facilities were constructed simultaneously, making full use of the rock mass’s natural ability to purify mine water in the goaf. The construction of CMUR projects, as shown in Figure 3, was undertaken [27]. Coal pillar dams and artificial dams were identified as the most critical components of CMURs, with the stability of these dam structures being crucial for the safety of the underground reservoirs.

At the same time, the continued extraction of coal resources inevitably resulted in the creation of a substantial amount of abandoned coal mine underground spaces, primarily consisting of shafts, abandoned tunnels, and goafs. The effective utilization of these abandoned coal mine underground spaces was identified as an urgent issue that needed to be addressed. Currently, the primary approach for dealing with abandoned underground spaces in mines has been their conversion into pumped hydro storage (PHS) systems [28,29,30,31]. Due to the typical occurrence of coal seams in groups in China and the extremely high surface evaporation rates in central and western mining areas, effective water source replenishment could not be achieved. Therefore, PHS systems in China have primarily been implemented in the form of underground reservoirs. Figure 4 illustrates a PHS system utilizing the underground goaf in a coal mine [32]. The CMUR was identified as the most suitable technological approach for utilizing abandoned underground spaces in coal mines in central and western China. While addressing water resource conservation issues, this approach also provided an effective engineering reference for the efficient use of underground spaces.

The above research and analysis indicates that, to address the issue of water scarcity in central and western mining areas and resolve the conflict between coal mining and water conservation, the construction of CMURs within the abandoned underground spaces of coal mines was a highly scientific and rational choice. CMURs had already become an important hydraulic infrastructure for the utilization of mine water. Since the CMUR primarily utilized the goaf created by longwall mining as the water storage space, it had a larger storage capacity and longer dam structure compared to underground reservoirs primarily utilizing tunnels. However, it was inevitably affected by mining activities on adjacent working faces and mine-induced seismic events during its service life. Therefore, several key issues were encountered during the construction of CMURs, such as the stability, deformation patterns, and safety of the dam structure under conditions of immersion, dry–wet cycles, high-concentration saltwater erosion, and complex stress environments. This paper, in conjunction with the development history of CMURs in China, provides a comprehensive review of the stability of the key structures, coal pillar dams, and artificial dam structures. Section 2 first introduces the main technologies and key structures involved in the construction of CMURs. Section 3 and Section 4 primarily discuss the mechanical properties of coal pillar dam structures and artificial dam structures under complex hydraulic coupling and stress conditions (laboratory scale), as well as the changes in stress fields, fracture fields, and seepage fields (“three fields”) and size optimization (engineering scale). Section 5 provides an outlook on the construction and future development prospects of CMURs. This paper aimed to serve as a reference for the construction of CMURs.

2. Main Technologies and Key Structures of CMURs

The construction of CMURs has been recognized as a complex system engineering project, involving the organic integration of multiple disciplines, including mining engineering, hydraulic engineering, engineering geology, hydrogeology, and environmental engineering. The design, construction, and operation of CMURs involves key technologies such as reservoir site selection, storage capacity calculation, water quality purification, and environmental impact assessment. The main structures include coal pillar dams, artificial dams, reservoir roofs, reservoir floors, water storage spaces, and drainage facilities. From a macro perspective, reservoir site selection and storage capacity calculation have been identified as the primary technologies for CMURs. From a structural perspective, coal pillar dams and artificial dams are considered the most critical structures. The reservoir site selection, storage capacity calculation, coal pillar dams, and artificial dams are introduced in this section.

2.1. Reservoir Site Selection

Reservoir site selection and storage capacity calculation have been identified as important tasks during the planning and design stages of CMURs. These tasks are typically carried out in a coordinated and intersecting manner. Reservoir site selection requires comprehensive consideration of factors such as mining height, surrounding rock properties, water storage range and level. The main factors influencing the performance of underground reservoirs are identified as storage coefficient, mine self-flow recharge, floor impermeability, and the safe distance from working faces. Identifying a suitable site is acknowledged as essential for building CMURs. However, limited research exists on the location selection of CMURs.

The aforementioned factors affecting the site selection of CMURs are numerous, and the ways and angles in which these factors influence the reservoirs are varied. Building CMURs by injecting water into abandoned goafs to form storage spaces inevitably disrupts the original dynamic balance of the groundwater system. Chen et al. [33] developed a detailed hydrogeological model for both the mining area and the underground reservoir. The impact of underground reservoir construction on the flow field was analyzed using factors like mining height at the reservoir site, storage range, overlying rock lithology, and water level. Sensitivity coefficients for each factor were determined. Numerical simulation was employed as a common method for establishing and analyzing groundwater systems. Figure 5 shows the groundwater system model of the Daliuta coal mine in the Shendong mining area [33]. Another method for considering the site selection of CMURs is based on water storage capacity. Distributed CMURs are suitable for coal mines with simple geological conditions. Faults, karst, and other geological structures have a significant impact on stress changes, hydraulic fracture propagation, and permeability distribution, which may lead to water seepage and even dam collapse. Therefore, the construction of CMURs was recommended to avoid areas with well-developed geological structures of this type. The storage capacity of CMURs was found to be positively correlated with mining height, roof permeability, and storage coefficient. Conversely, it was negatively correlated with the level of the coal seam floor and the floor permeability. A method for selecting sites for CMURs based on entropy weight and a calculation model for the storage capacity were developed by Zhang et al. [34]. Furthermore, the significance of hydrochemistry for the selection of CMUR sites was discussed by Guo et al. [35]. The durability of artificial concrete dams could be reduced under the dry–wet cycling effects of high-sulfate water. CMURs were not suitable for locations with poor water quality and a high tendency for water-rock interactions. The concentrations of Ca²⁺, sodium adsorption ratio (SAR), and F^- in the water were found to be further increased under the influence of water-rock interactions, adversely affecting water quality improvement.

2.2. Storage Capacity Calculation

CMURs primarily utilize the free voids between the fractured and collapsed rock masses in the goaf for water storage. The reservoir capacity is defined as the total amount of free voids within the storage range of the goaf that can be used for water storage [36]. The free voids mainly involve the development height of the “three zones” (caved zone, fractured zone, and bending zone) of the goaf, which are central to the design and estimation of reservoir capacity. The formation of the three zones and hydraulic migration resulting from mining is illustrated in Figure 6 [37]. The evaluation metric for the water storage capacity of CMURs is the water storage coefficient, which was the unit volume water storage capacity of the goaf. It is defined as the total voidage of the broken rock mass in the caved zone [38]. A close relationship has been found between the water storage coefficient and the mining height, mining methods, overburden pressure, and surrounding rock properties. It primarily depends on the spatial structural characteristics of the storage space [27]. Regarding spatial structural characteristics, the formation patterns of the spatial structure after working face mining have been studied through various approaches, including theoretical analysis [39,40], numerical simulation [41,42], and similar simulation experiments [27,36,43]. Additionally, the compaction characteristics of the gangue accumulation process are influenced by mining activities on adjacent working faces and the coal seam inclination, causing variations. Detailed calculations also consider the shape and accumulation characteristics of the gangue blocks. However, due to the significant randomness of gangue breakage and caving, research in this area remains in the exploratory stage [44].

2.3. Coal Pillar Dam

The longwall mining method has typically been used in China’s underground coal mining. This method established protective coal pillars along the boundary of the entire mining area to maintain overall stability and reduce mutual interference between adjacent mining areas [45]. In CMURs, boundary protective coal pillars are designed and modified to convert them into coal pillar dams. These are connected with artificial dam structures to form an integrated dam structure, serving as a permanent hydraulic construction for the reservoir [46,47]. Figure 7 illustrates the relationship between working face mining and the formation of coal pillar dams [48]. Coal pillar dams are major components of the dam structure in CMURs and play a crucial role in ensuring the safe operation of CMURs and coal mining activities [49,50]. Additionally, within the goafs converted into storage spaces, residual coal pillars have been identified as remnants left behind from the mining of adjacent working faces. Although residual coal pillars are typically not considered part of the dam structure due to their location within the reservoir, they play an important role in maintaining the stability of the roof of the water storage space [51].

2.4. Artificial Dam

In addition to the reserved coal pillar dams, artificially constructed non-coal dam structures, known as artificial dams, have also been constructed as part of the dam structure of CMURs. Artificial dams are primarily used to connect the gaps between adjacent coal pillar dams, forming a complete and sealed water storage space maintenance structure together with the coal pillar dams. Together, they constitute the dam structure of the CMUR [52,53]. The artificial dam structure has played an important role in protecting the stability and durability of the underground reservoir. Unlike coal pillar dams, which can reach thicknesses of tens of meters, most artificial dams in CMURs have been designed with a thickness of approximately 1–3 m [54]. These structures are typically constructed as reinforced concrete monolithic beams [55], predominantly in flat or “T” shapes. Due to the relatively thin thickness of artificial dams, their waterproof performance is crucial. Therefore, high-grade concrete, such as C30, C40, C50, and engineered cementitious composites (ECC), have often been used when constructing artificial dams [45,56]. Figure 8 illustrates a typical artificial dam structure [56]. There are also cases where the waterproof performance has been improved by altering the structural form. For example, at the Daliuta coal mine, the artificial dam was designed as a combination of an impermeable wall and a bearing wall. The impermeable wall was constructed from brick and filled with impermeable materials such as loess, while the bearing wall used I-beams as the framework and concrete as the bonding filler [57].

3. Key Theories and Technologies for the Stability of Coal Pillar Dams

As the core component of the dam structure in CMURs, coal pillar dams are not only influenced by dynamic and static loads such as roof breakage, fracturing, and slipping during the mining process but also subject to long-term water erosion due to contact with mine water on the water-facing side. This is particularly significant during the repeated rising and falling of the storage water level caused by the water extraction process in CMURs [58,59]. After the CMUR was filled with water, the coal pillar dam is subjected to a complex environment of combined dynamic and static stress fields and water immersion. Due to water immersion, significant changes occurr in the mechanical parameters of the coal, such as strength, elastic modulus, and cohesion. Under hydraulic coupling, the coal pillar dam is more susceptible to damage and even instability [60].

Following the mining of a working face, the overlying rock strata will experience roof collapse and stress redistribution. The stress environment of the coal pillar dam changes from a stable triaxial stress environment to a complex stress environment, which includes stress from the overlying rock pressure, stope bearing pressure, lateral pressure from the fractured roof, and superimposed microseismic activity from the mine. Figure 9 illustrates the stress environment of the coal pillar dam [61]. Based on the types of loads acting on the coal pillar dam, the stress can be categorized into static load, mining-induced load, and mine seismic load, with the latter two summarized as vibrational loads. The impacts of water immersion on coal pillar dams were categorized into two main types: dynamic water level fluctuations causing dry–wet cycles and variations in immersion time (water content); and long-term creep effects under stable water level conditions, along with erosion induced by saline mine water. Therefore, clarifying the mechanical deformation characteristics, as well as the changes in stress field, fracture field, and seepage field of coal pillar dams under complex hydraulic coupling and stress conditions, is an important prerequisite for maintaining the long-term stability of coal pillar dams in CMURs and for determining their reasonable width and dimensions.

3.1. Mechanical Properties of Coal Pillar Dams

As shown in Figure 7, the coal pillar dams in CMURs are composed of multiple segments of boundary protective coal pillars. After the goaf has filled with water, one side of the coal pillar dam is exposed while the other side is subjected to water pressure, resulting in a very complex stress state. Laboratory tests are an important means of studying the mechanical properties of coal pillar dam samples. Through uniaxial compression, triaxial compression, and cyclic loading tests on coal pillar dam samples, researchers have replicated the mechanical response and failure patterns of coal pillar dams under load-bearing conditions on a laboratory scale. Since coal pillar dams in CMURs are significantly different from ordinary protective coal pillars due to the substantial impact of water, most studies have considered the effects of water immersion.

(1)

Mechanical properties of coal pillar dams affected by water immersion

During the use of underground reservoirs, the cyclic stages of water storage and drainage inevitably cause changes in the water level. Coal pillar dams periodically experience alternating states of drying and water loss, and water saturation. Research conducted on coal pillar dam samples under repeated water immersion more closely approximate the actual environmental changes of underground reservoirs. Currently, laboratory-scale experiments consider the mechanical properties of coal pillar dam samples under the influence of cyclic water storage and drainage primarily use the dry–wet cycle (repeated water immersion) mode. In these experiments, coal samples are repeatedly subjected to drying and water saturation treatments. An oven has typically been used as the dehydration equipment, with a non-destructive water immersion device used as the water saturation equipment, as shown in Figure 10 [62]. The number of dry–wet cycles has been shown to significantly affect the saturation water content of coal pillar dam samples. Overall, the water content of the coal samples initially increased rapidly and then gradually stabilized as the number of dry–wet cycles increased. The saturation water content of the coal samples showed a logarithmic increase with the dry–wet cycle numbers. Figure 11 presents the curve of coal sample water content w versus the number of dry–wet cycles, as summarized from various studies [59,62,63,64]. During the dry–wet cycle process, the complex physical and chemical interactions between water and coal promoted the further expansion of the primary fractures in the coal samples, as well as the formation of additional pores and voids, thereby increasing their saturation water content. The mechanical properties of coal pillar dam samples under dry–wet cycles have typically been tested using uniaxial compression tests and triaxial compression tests. Additionally, acoustic emission (AE) non-destructive testing has been used to analyze the mechanical damage mechanisms and energy evolution characteristics of the coal samples during failure. The conclusions indicated that under uniaxial compression, the uniaxial compressive strength gradually decreased with the increase in the number of dry–wet cycles. The peak stress of the coal samples exhibited an exponential decline with the increase in the number of cycles, while the peak strain gradually increased, and the elastic modulus also gradually decreased [63,65,66]. Under triaxial compression, the triaxial compressive strength of the coal samples decreased with the increase in the number of water immersion cycles at different confining pressures. The most significant decrease in triaxial compressive strength occurred after a single immersion, with the rate of decline gradually slowing thereafter. The AE cumulative ringing counts of the samples also decreased with the increase in the number of dry–wet cycles, indicating a reduction in the energy required for the coal samples to fracture [64].

During the water storage stage, as the water level rises, the immersion time of the coal pillar dam continually increases, and the water content will also change, gradually transitioning from a natural water content state to a saturated water content state. The uniaxial compression stress–strain curve of water-saturated coal samples can be segmented into five phases: (a) the fracture compaction phase, (b) the elastic deformation phase, (c) the stable fracture propagation phase, (d) the unstable fracture propagation phase, and (e) the unloading failure phase. Under uniaxial compression, as the water content increased, the compaction stage of the coal sample curve becomes more gradual. This is due to the softening of the coal sample and clay minerals upon water immersion, which prolongs the compaction process. The peak stress of the coal sample gradually decreases, while the peak strain gradually increases [58,67]. The elastic modulus also decreases with the increase in water content. The primary reason is that water has penetrated the coal sample particles and interacted with the soluble substances in the particle cement, leading to hydrolysis reactions that reduced the bonding force between particles. Additionally, the lubricating effect of water decreases the friction between particles, which macroscopically manifests as a reduction in the friction of the fracture surface [68,69,70,71,72]. Under the combined stress of compression and shear, the microcrack propagation in the coal sample is controlled by shear cracks. At a low water content, the macroscopic fracture mode of the coal sample gradually shifts from tensile failure to shear failure as the loading angle increases. At a moderate water content, the coal sample mostly exhibits a combination of tensile and shear failure. At a high water content, the coal sample predominantly exhibits shear failure [60,73,74]. Many coal pillar dams experienced multiple disturbances from the mining of adjacent working faces before the underground reservoir stored water. In experiments, coal samples have been subjected to multiple loading and unloading cycles, which accumulates stress damage, exhibiting characteristics of decreased peak stress, increased elastic modulus, and increased residual deformation with prolonged immersion time [75].

During the stable service stage, the coal pillars located at the lower part of the dam remain in a long-term water-saturated state. Coal pillar dams are susceptible to creep failure under the long-term influence of overburden load. Studying the creep fracture behavior of water-saturated coal samples can more accurately reveal the instability mechanism of coal pillars. The dissolution of coal’s microstructure by water has been identified as the primary cause of coal softening, leading to the propagation of macroscopic cracks during immersion [76,77]. Under axial load, the coal pillar samples have been shown to experience instantaneous elastic strain, followed by decelerating creep deformation, with the creep rate gradually decreasing as strain increases. Once the creep rate reaches a certain level, the coal pillar samples enter the steady creep stage. Finally, the samples transition into an accelerated creep phase, characterized by a rapid increase in creep strain that follows an approximate exponential relationship with time [78,79].

In some mines where CMURs were constructed, the mine water was of the Cl·S0₄-K·Na type, with high salinity, resulting in the coal pillar dams being immersed in a high-concentration saline water environment for extended periods [47]. In these studies, the process by which high salinity ions move within the coal pillar dam, and ions’ effect on the degradation of its structural integrity, was shown to be crucially important. During uniaxial compression tests, it has been shown that the peak strength and elastic modulus of the coal pillar samples exponentially decreases with increasing saline water concentration and immersion time, while the peak strain before failure increases significantly. Similar patterns have been observed during dry–wet cycling tests [47,80,81]. At the microstructural level, the quantity of mineral particles on the coal surface significantly decreases, along with an increase in the internal structural connectivity and porosity [82].

(2)

Mechanical properties of coal pillar dams affected by vibrational loads

Researchers have addressed the issue of dynamic load disturbance to coal pillar dams caused by mining activities and mine-induced seismic events. It has been shown that the impact loads generated by mining disturbances and mine seismic events propagate as stress waves through the rock layers, causing damage to the coal pillar dams. Additionally, the weakening effect of water erosion has made the study of the stability of coal pillar dams under dynamic loads more complex. The mechanical properties of coal pillar dams under the influence of vibrational loads have mainly been studied using split Hopkinson pressure bar (SHPB) tests to examine the dynamic mechanical characteristics at high strain rates. Here, the dynamic compressive strength and elastic modulus followed exponential and logarithmic functions of the strain rate, respectively, showing a significant positive correlation. As the water content increased, the dynamic elastic modulus increased almost linearly, while the compressive strength gradually decreased. Under the same impact load, samples with higher water content failed more quickly, and the expansion of parallel cracks into intersecting cracks exacerbated the damage [68,83]. However, some studies indicated that the presence of water pressure had a positive effect on the samples, increasing the dynamic strength of the coal pillar samples and reducing the failure strain [84]. Figure 12 shows an innovative SHPB experimental apparatus that considers water pressure [84].

3.2. Three-Field Patterns of Coal Pillar Dams

Based on the findings of laboratory mechanical testing, the mechanical properties of coal pillar dam samples subjected to mine water have been observed to undergo a certain degree of weakening. Therefore, their response on an engineering scale has warranted significant attention. The coal pillar dam is situated in a typical hydraulic coupling stress environment, and the stress distribution patterns, fracture development patterns, and seepage patterns of coal pillar dams under complex stress conditions have consistently been key research issues. The study of the “three-field” patterns (stress field, fracture field, and seepage field) of coal pillar dams in CMURs has primarily employed similar simulation experiments and numerical simulation methods, although some researchers have also conducted analyses using theoretical methods. Due to the highly complex stress environment of coal pillar dams, theoretical derivations have struggled to fully reflect hydraulic coupling conditions. Typically, the elastic mechanics equation and hydraulics equations (Darcy’s law, Biot equation) were coupled, often from the perspectives of permeability and hydraulic conductivity [85,86], and validated through numerical models.

Similar simulation experiments, which can directly reflect deformation and stress phenomena on an engineering scale, have been widely used to study the movement patterns of overlying rock in coal mining fields. In studies on the stability of coal pillar dams considering the water environment, some scholars have also adopted this approach. Similar simulation experiments conveniently simulate the disturbances to coal pillar dams caused by different coal seam mining activities. However, the impact of mine water is simplified to the horizontal water pressure applied to the coal pillar dam through the horizontal load of the experimental apparatus, and the hydrochemical effects of water on the coal pillar cannot be considered. Some scholars have adopted the direct water injection method to study the seepage patterns of the rock layers in the dam floor [87]. Figure 13 illustrates a similar simulation experiment system for a CMUR [86]. During the advancement of the lower working face towards the upper underground reservoir, the interlayer rock undergoes tensile and shear failure, with fractures gradually extending towards the coal pillar dam. The interlayer fracture network connectivity during multi-seam mining of underground reservoirs is progressive, differing from the sudden connectivity observed in laboratory rock mechanics experiments. During the water storage process in CMURs, as the water storage height increases, the horizontal stress borne by the coal pillar dam gradually increases. It has been shown that displacement changes do not occur before reaching the ultimate strength. However, after reaching the ultimate strength, displacement gradually increases. In these studies, repeated processes of water injection and extraction caused damage to the coal pillar dam. Even after the water level dropped (reducing horizontal stress), the water-blocking capacity of the coal pillar dam was diminished. To ensure the safety of the underground reservoir, it was considered not advisable for them to store water again [86,88].

An important cause of damage to coal pillar dams after water immersion is the softening effect of water on coal. To investigate the damage and degradation characteristics of coal pillar dams under water softening, it has been necessary to obtain real-time feedback the relationship between the mechanical parameters of the coal pillar and the water content and pore water pressure into the research model [89,90]. Therefore, many scholars have adopted numerical simulation methods [49,63,91,92]. Numerical simulation allows for the convenient, simultaneous output of the stress patterns, fracture patterns, and seepage patterns of coal pillar dams. Detailed comparative analysis of the stress distribution, pore water pressure distribution, and plastic zone expansion patterns of coal pillar dams, with and without the influence of water immersion, can intuitively reflect the impact of water immersion on coal pillar dams. In these works, the strength of the coal pillar dam decreased after water immersion, leading to a gradual increase in the plastic zone of the coal pillar. The development of the plastic zone caused stress redistribution and transfer to the deeper parts of the coal pillar. Damage and degradation allowed water to infiltrate the coal pillar dam, further reducing the coal pillar’s strength. During the immersion process following mining disturbances, the coal pillar dam exhibited progressive failure, with the depth of the plastic zone development corresponding to the location of the vertical peak stress. The coal pillar exhibited a “damage–seepage–accumulated damage” effect in a water immersion-stress coupling environment. As water immersion deepened, the permeability in the elastic zone of the coal pillar decreased, the weakening effect of water immersion gradually diminished, and the damage to the coal pillar stabilized [61,75]. Figure 14 provides a detailed illustration of the stress patterns and plastic zone expansion of coal pillar dams with and without the influence of water immersion [75]. Additionally, under the influence of dynamic loads, it has been shown that the stress concentration within the coal pillar dam increases and the range of the high-stress zone expands and gradually extends towards the elastic zone and the floor. The disturbance effect of near-field mine seismic loads on the coal pillar dam is significant. The stress concentration level increases after the dynamic load ended, leading to more severe damage to the dam [68].

3.3. Size Optimization of Coal Pillar Dams

Coal pillar dams in CMURs were created from sectional coal pillars, which often had an excessive width. In early coal mine designs, the conversion of goafs into underground reservoirs was not considered, resulting in sectional coal pillars with widths that were overly large for use as coal pillar dams in underground reservoirs. In current designs for new mines in central and western regions of China, the construction of CMURs has been preemptively planned. Therefore, reasonably determining the width of early sectional coal pillars, and thus the width of coal pillar dams in CMURs, is crucial for the efficient recovery of coal resources. After plastic deformation occurs on both sides of the coal pillar, there must be a sufficiently wide elastic zone in its internal area. From the perspective of the stress environment and structure, the coal pillar dam is divided into the reservoir side and the roadway side, as shown in Figure 15. Therefore, the width of the coal pillar dam is the sum of the widths of the plastic zone on the reservoir side L_Cp, the elastic zone L_Ce, and the plastic zone on the roadway side L_Hp [93]. Yao et al. [65] and Chen et al. [93] derived the calculation formula for the width L of coal pillar dams from the perspective of elastoplastic theory, considering the stress environment and structural characteristics of coal pillar dams. As shown in

Table 1. Theoretical calculation formula for the width of coal pillar dam.

	Yao [65]	Chen [93]
LCp	$L_{Cp}=\frac{M}{2\lambda f}\ln \left[\frac{K_1\gamma H}{\lambda p+{\sigma }_r}-\frac{S_m{S}_g\left(1-\eta \right)}{2\lambda f\left(\lambda p+{\sigma }_r\right)}\right]; \eta =\exp \left[\frac{2\lambda f}{S_m{S}_g}\left({\sigma }_c-{\sigma }_r\right)\right]$
LHp	$L_{Hp}=\frac{M}{2\lambda f}\ln \left\{\frac{K_2\gamma H}{\sigma r}-S_m{S}_g\left[1-\exp \left(\frac{2\lambda f}{S_m{S}_g}\left({\sigma }_c-{\sigma }_r\right)\right)\right]/2\lambda f{\sigma }_r\right\}$
LCe	$L_{Ce}=kM$	$L_{Ce}=0.5kM\sqrt{\frac{3p}{K_p}}$
L	L = LCp + LHp + LCe

Note(s): where M is the thickness of the coal seam, m; f is the friction factor of the interface between the coal seam and the roof and floor; K₁ is the stress concentration coefficient of the reservoir side; K₂ is the stress concentration factor of the roadway side; σ_c is the uniaxial compressive strength of coal, MPa; σ_r is the residual strength of coal, MPa; S_m is the softening modulus of the coal mass, MPa; S_g is the strain gradient of the coal mass in the plastic zone; γ is the volume force; p is the water pressure, MPa; k is the safety factor, usually 2–5; K_p is the tensile strength of coal, MPa; and λ = (1 + sin ϕ)/(1 − sin ϕ), ϕ is the effective internal friction angle of coal.

, the latter specifically provided a formula for determining the recommended width of the elastic zone. However, this derivation treated coal as a continuous and homogeneous medium, whereas in reality, coal is a type of typical sedimentary rock with distinct layered characteristics. Therefore, the calculation results have a certain degree of error. Zhang et al. [94] suggested that the ideal width of coal pillar dams could be defined by their deformation and stress state. Smaller coal pillar sizes were more prone to stress concentration, which was more likely to “erode” the coal pillar dam. A correlation was observed between the dimensions of the coal pillar dam and the extent of crack development. As the coal pillar dimensions increased, the stress concentration effect became less pronounced, and fractures were less likely to penetrate both the interior and exterior of the dam. Zhi et al. [95] identified the maximum extent of mine water infiltration into coal pillars at various water storage levels using seepage experiments.

4. Key Theories and Technologies for the Stability of Artificial Dams

Similar to coal pillar dams, artificial dams in CMURs are also subjected to the coupled effects of stress and seepage. In addition to the pressure from the overlying rock layer, mining-induced stress, and water pressure, artificial dams, used to connect the spaces between adjacent sectional coal pillars, are also subjected to friction and compression forces from the coal pillars on both sides (referred to as coal pillar forces) [96,97], as shown in Figure 16 [98]. Additionally, unlike the material properties of coal, the primary construction material for artificial dams is concrete. As a typical porous material, concrete is significantly affected by water seepage. This effect is particularly notable when mine water is highly saline, as the high concentration of salt ions in the water can react with the components of the concrete, further damaging the strength of the artificial dam.

4.1. Mechanical Properties of Artificial Dams

At the laboratory scale, the research methods for studying the mechanical properties of artificial dams have been similar to those for coal pillar dams. In addition to conventional dynamic and static mechanical tests, due to the particularity of the artificial dam material, many studies have borrowed methods from concrete durability research, focusing on the microscopic level of concrete samples. Nuclear magnetic resonance imaging (NMRI), scanning electron microscopy (SEM), CT scanning, and three-dimensional digital reconstruction have commonly been used to characterize the mechanical properties and structural changes of artificial dam samples after saline water erosion from a microscopic perspective [56,99].

(1)

Mechanical properties of artificial dams affected by water immersion

During the operation stage of CMURs, the stability of artificial dams is influenced by the sustained, repeated infiltration of mine water. Similar to the coal pillar dams, the water content under repeated dry–wet cycles has also been found to exhibit a logarithmic relationship with the cycle numbers, but the increase was significantly lower. Under the action of dry–wet cycles with saline water, the anisotropic concrete artificial dam samples exhibited complex mechanical properties and damage mechanisms. This altered the macro-micro crack evolution patterns and the geometric distribution characteristics of macro cracks in the samples. The percentage of micro shear cracks in the samples changed from initially decreasing and then increasing to a continuous increasing trend, and the complexity of the crack geometric distribution significantly increased. The physicochemical changes in the artificial dam samples led to alterations in the pore structure, a reduction in density, and the connection of weak structural surfaces. These were the reasons for the decrease in uniaxial compressive strength and the changes in the macro- and micro-crack evolution patterns in the concrete. The chemical damage caused by ion exchange was the primary reason for the changes in the mechanical properties and the macro- and micro-crack evolution patterns of the concrete under dry–wet cycles [98,99]. Figure 17 clearly illustrates the cracking changes in the samples through fractal dimension analysis [99].

Wang et al. [56] and Yang et al. [100] further determined the degradation mechanism of the mechanical properties of concrete samples from artificial dams from a micro-ion chemical reaction perspective. As the concentration of salt ions increased, the concentrated sulfate solution continuously transformed portlandite into gypsum and ettringite phases. The adhesion of concrete to Ca(OH)₂ gradually weakened and even detached from its original stable structure, which was the main reason for the degradation of concrete’s mechanical properties. Adding nanomaterials could change the failure mode, crack propagation mode, and load-bearing capacity of saturated concrete. Adding 3 wt% nano-Al₂O₃ to the concrete of artificial dams significantly increased its uniaxial compressive strength and effectively reduced crack formation [101].

(2)

Mechanical properties of artificial dams affected by vibrational loads

Since the size of the artificial dam is much smaller than that of the coal pillar dam and is constructed through post-mining artificial pouring, vibrations and mine seismic activities caused by mining operations frequently occur. These uncertain loads, along with geostress and water pressure, have become common load forms [102]. The study of the seismic performance of artificial dams is identified as a crucial aspect in ensuring safety throughout its lifecycle and forms the basis for monitoring and early warning in field engineering practice [103]. Unlike the study of the mechanical properties of coal pillar dams under vibrational loads, which commonly uses SHPB tests, the smaller size of artificial dams allows for the use of similar simulation combined with shaking table tests to study the mechanical properties of artificial dams under different vibrational loads. These findings can then be validated through numerical simulation tests. Figure 18 illustrates a reasonable test system combining similar simulation and shaking table tests [53].

Chi et al. [53] conducted vibration similar simulation tests on the seismic performance of flat-type artificial dams. The results indicated that, under the combined action of vibration and external loads on the top, the bottom of the artificial dam was the stress concentration area. Under the combined effects of impact and compression from the crushed stones in the goaf, the reservoir-side stress was approximately 10% higher than the roadway-side stress. Under increased upper loads, the artificial dam exerted a “supporting” effect, leading to uneven compressive deformation of the coal pillar dam. Particularly, due to the differences in materials between the artificial dam and the coal pillar dam, the acceleration generated by vibrational loads differed. Under the same seismic wave, the displacement peak of the artificial dam was greater than that of the coal pillar dam, with the displacement of the artificial dam being approximately 1.14 times that of the coal pillar dam. The difference in velocity caused stress concentration at the connection between the coal pillar dam and the artificial dam, making it a weak point in the reservoir dam structure during an earthquake. Comparing the seismic performance of flat-type artificial dams and arch-type artificial dams under the same geological and seismic intensity conditions, the seismic safety performance of arch-type artificial dams was significantly superior to that of flat-type artificial dams. The primary reason is that the arch effect generated partial compressive stress in the reaction force of the coal pillar on the artificial dam. Part of the vibrational load on the arch-type artificial dam was transferred to the coal pillar dam, reducing the shear stress on the artificial dam caused by mine seismic events and improving the seismic performance of the arched artificial dam [104]. However, from a practical engineering perspective, since the construction of an arch-type artificial dam requires slotting on the coal pillar dam structure before pouring, the existing seepage prevention methods are less applicable to the arch-type artificial dam structure compared to the flat-type artificial dam structure.

4.2. Three-Field Patterns of Artificial Dams

As previously mentioned, the thickness of the coal pillar dam reaches tens of meters, whereas the thickness of the artificial dam is only a few meters. Additionally, because the artificial dam is poured at a later stage, the connections between the concrete structure and the roof, floor, and adjacent coal pillar dams are highly susceptible to leakage under water pressure and require special attention. For example, a slight leakage occurred in the MB-1 artificial dam at the Wulanmulun mine, as shown in Figure 19 [54]. The coal pillar dam, with a thickness of approximately 80 m, exhibited good impermeability, and therefore, no leakage was observed. Preliminary analysis suggested that the leakage in the artificial dam stemmed from the incomplete collapse of the goaf, which led to weak connections between the dam and the surrounding rock in the roof. Additionally, the deformation and crack propagation patterns of the artificial dam under different water pressures directly affected the seepage patterns. Therefore, for artificial dams, much of the research has focused on their seepage patterns.

Kong et al. [54] and Fang [55] conducted similar experimental studies on the stress, fracture, and seepage patterns of artificial dams under the influence of water pressure and mining pressure using a self-developed nested multi-coupled test system. The test system is shown in Figure 20 [54]. The results indicated that an increase in water pressure and model top load both led to an increase in the deformation of the artificial dam surface. Under the influence of water pressure, seepage occurred sequentially at the following locations: the bottom slot of the artificial dam, the top slot of the artificial dam, the coal pillar dam, the reservoir floor, the reservoir roof, the roadway roof, the roadway floor, and the surface of the artificial dam. As the water pressure increased, the permeation boundary range on the model surface gradually expanded. The failure of the artificial dam was characterized by compression-shear failure, with cracks extending from the lower outer side of the artificial dam model to near the upper inner vertex, penetrating the entire artificial dam at an approximate 45° angle to the cross-section, as shown in Figure 21.

After the CMUR is filled with water, the original failure zone and plastic zone of the coal pillar dam expand under long-term water storage and repeated immersion. At this time, the original bearing pressure environment changes, and both the magnitude and location of the bearing pressure peak points change. The artificial dam located in the middle of the coal pillar dam acts as a local support body and cannot deform in coordination with the surrounding rock. The maximum displacement on the outer surface of the artificial dam has been shown to have a clear linear correlation with the water storage height under different roof and floor pressures [105]. Ma et al. [106] further studied the displacement and stress patterns of artificial dams at different water storage heights. The stress distribution caused by different water storage heights was basically the same, but the stress increased with the increase in water storage height. The strength of the artificial dam played a critical role.

4.3. Size Optimization of Artificial Dams

Unlike the primary goal of optimizing coal pillar dam size, which is to maximize the recovery of coal resources, the main purpose of optimizing the size of artificial dams is to enhance their impermeability and improve the stability of the slot structures between the artificial dam and the adjacent coal pillar dams. However, some studies, through theoretical derivation of mechanical models for specific CMUR cases, suggested that the vertical support wall of “T”-shaped artificial dams could be removed, and the dam wall thickness could be reduced to within the range of 0.6–0.7 m. This would save materials and improve the economic efficiency of dam construction [57]. The connection between the artificial dam and the coal pillar dam is a vulnerable point in the dam structure of a CMUR. Currently, slotting is used as the connection method between the coal pillar dam and the artificial dam. When lateral loads, such as water pressure, are applied to the artificial dam, they are transmitted from the artificial dam to the coal pillar dam at the slot position. The slot depth dictates the stress conveyed to the coal pillar, influencing the stability of the entire dam structure. Because of mining disturbances, a certain plastic zone had already formed in the adjacent sections of the coal pillar dam prior to the artificial dam’s construction. These plastic zones undoubtedly provided pathways for water seepage, so slotting operations should avoid the plastic zone as much as possible [52]. Kong et al. [107,108] comprehensively analyzed the stress, displacement, plastic zone distribution, and seepage patterns of the artificial dam and coal pillar dam at the slot connection through similar simulation and numerical simulation. They concluded that increasing the slot depth within a certain range could improve the stability of the underground reservoir dam. However, when the slot depth exceeded a critical value, further increases in slot depth were detrimental to the reservoir’s stability. They recommended a slot depth of 0.6 m. In addition to flat, “T”-shaped, and arch-shaped forms, artificial dams also come in wedge and inverted cone shapes, among other forms. However, there has been little research on specially shaped artificial dams. Zhao [109] used numerical simulation to demonstrate that inverted truncated cone-shaped dams had better stress conditions and smaller displacements when subjected to longitudinal and lateral stresses. These dams could achieve uniform pressure distribution on the coal pillar under high water pressure, thereby reducing the deformation of the coal pillar. Compared to flat-type artificial dams, wedge-shaped artificial dams also have certain advantages in terms of stress distribution and displacement deformation. Flat-type artificial dams are more suitable for use under smaller water pressures.

5. Conclusions and Prospects

With the depletion of coal resources in the eastern mining areas of China, large-scale coal mining has shifted to the central and western regions. The intensity of coal mining in these regions is substantial (>10 million t/a), which starkly contrasts with their characteristics of water scarcity and high surface evaporation. Government departments and coal companies have gradually recognized the importance of water-preserving coal mining. During the coal mining process, it is necessary not only to protect mine water from being damaged and wasted but also to make full use of mine water [7]. Since the completion of the first goaf water storage facility at the Daliuta Coal Mine in 1998, nearly 20 years of engineering practice have led to the development of CMUR water storage technology, which utilizes goafs from longwall mining as water storage spaces. This established the concept of storing mine water underground in coal mines [16]. According to statistics, more than 35 CMURs have been constructed in the central and western regions of China [1]. The total number of constructed, under-construction, and planned CMURs will exceed 50 [11]. A CMUR technology demonstration zone will be established in the Inner Mongolia–Shaanxi border area. Based on the development history of CMUR construction in China, this paper systematically reviewed the research progress and achievements in the foundational theories of the stability of coal pillar dams and artificial dams in CMURs. This includes the mechanical properties under hydraulic coupling, the evolution mechanisms and impact relationships of the “three fields” (stress field, fracture field, and seepage field), and the optimized design of dam dimensions. Clarifying the instability and failure patterns of CMUR dam structures is essential to maintaining dam structure stability. However, engineering practice in CMUR technology has outpaced technical research, and technology has advanced beyond theoretical exploration. To further apply this technology for the protection and utilization of mine water resources under different geological and working conditions, increased technical exploration and theoretical research are necessary. Based on a systematic review of the stability research of CMUR dam structures in China, the main conclusions and outlook are as follows:

(1)

Reservoir site selection and storage capacity calculation are prerequisites for the construction of CMURs. The water storage space in CMURs consists of the voids in the caved rock masses of the goaf, with the core element being the determination of the water storage coefficient. Further theoretical and predictive research on the development patterns of overburden fracture zones and groundwater migration mechanisms during coal seam mining should be conducted. A systematic understanding of the spatial and temporal evolution patterns of void spaces in caved rock masses within the goaf, as well as the seepage paths and migration patterns of mine water in the goaf and overburden, is necessary.

(2)

As core components of the dam structure in CMURs, coal pillar dams and artificial dams should address key issues during the design, construction, and operation stages. These include understanding the mechanical and hydrochemical mechanisms of coal pillar dams and artificial dams under complex hydraulic coupling stress conditions, such as overlying rock load, mining-induced load, and mine water pressure. It is essential to investigate the micro- to macro-scale failure, seepage, and instability mechanisms. Special attention should be given to the creep effects of coal pillar dams and the durability of artificial dams under long-term exposure to corrosive mine water. Additionally, the issue of steel reinforcement corrosion in artificial dams due to the influence of corrosive ions such as chloride and sulfate ions needs to be addressed.

(3)

Currently, the constructed CMURs are all near-horizontal reservoirs, with coal seams deposited at near-horizontal angles (0°–8°). There has been limited application and research under gently inclined coal seam conditions (8°–25°), with only the Lingxin Coal Mine having applications in gently inclined coal seams (10°–15°). At present, there are no engineering cases of CMURs constructed under complex steeply inclined coal seam conditions. The applicability and systematic technical standards for steeply inclined coal seams remain to be studied.

(4)

In eastern mining areas, due to resource depletion and capacity withdrawal, abandoned mines are increasing year by year. However, the water resources in these abandoned mines have not been effectively utilized. For abandoned mines, the impact of future coal mining on the mine reservoir does not need to be considered. Therefore, it is of great significance to fully utilize the underground space of coal mines for the construction of CMURs. The underground spaces of abandoned mines should be fully utilized to explore the application of CMUR technology, thereby achieving full utilization of mine water resources in abandoned mines.