Introduction

Many flooding hazards have occurred in the coal mines of China during the past few decades, causing serious economic loses and fatalities (Jin et al. 2011; Li et al. 2015; Wu et al. 2011, 2013), especially in North China (Wu et al. 2014) where most mine water disasters originated from the underlying formations (referred to as floor strata in the paper) due to the circumstances of ground stress, temperature, karst water pressure, and mining disturbance (He et al. 2005). The main consequences of mining disturbance include deformation or mining-induced fractures within the floor strata, which contributes to the mine flooding hazards from floor (the underlying aquifers). Therefore, it is important to recognize the characteristics of in situ and induced rock stress field not only in rock mechanics and engineering (Ge and Hou 2012), but also in coal exploitation. It is well known that the stress field within a rock mass surrounding a mine excavation changes during the course of mining (Wang et al. 2013; Singh et al. 2011; Suchowerska et al. 2013). This causes the occurrence under the coal seam floor of deformation and fractures that evolve as the length of the excavation increases. Studying the characteristics of stress evolution with time and space under mined coal seams is of great significance for safe production (Islam et al. 2009; Orlecka-Sikora et al. 2012; Griffith et al. 2014). The stress would increase the difficulty of tunnel excavation particularly close to the mined coal seams (Tan et al. 2010a). Understanding the stress is critical to determining the failure depth in the formation underlying a coal seam (Zhu et al. 2014), especially when the mining takes place above aquifers (Zhang and Shen 2004). Understanding the dynamic nature of the stress field also helps in determining the width of waterproof pillar in the presence of faults or karst collapse columns, and comparing stress with resistance/friction along a fault plane. In addition, a thorough understanding of the stress environment under the mine excavation floor is essential for describing the relief gas flow (Yang et al. 2011) and also for predicting potential failures of rock strata under the coal seam (Suchowerskan et al. 2014).

There are several methods and techniques to measure the in situ rock stress. They can be classified into the following (Ge and Hou 2012): hydraulic fracturing method (Haimson and Cornet, 2003; Amadei and Stephansson 1997; Cai et al. 1995; Fairhurst 2003; Ljunggrena et al. 2003), stress relief method (Sugawara and Obara 1999; Cai et al. 1995; Fairhurst 2003; Ljunggrena et al. 2003), flat jacking method (Amadei and Stephansson 1997; Cai et al. 1995), strain recovery method (Amadei and Stephansson 1997; Cai et al. 1995), borehole breakout method (Amadei and Stephansson 1997; Cai et al. 1995), differential strain curve analysis method (Amadei and Stephansson 1997;Cai et al. 1995; Oikawa et al. 1993), drilling-induced tensile fracture method (Brudy and Zoback 1999), acoustic emission method (Lavrov 2003), and geophysical method (Cai et al. 1995). All these methods are helpful to know the stress field in engineering. From coal mining perspective, the stress distribution underlying or overlying coal seams has been studied using (1) theoretical models such as the half-infinite body (Wang et al. 2013), semi-plane elasticity (Peng et al. 2005), boundary integral formula of semi-plane elastic (Zhu et al. 2007), elastic–plastic theory elastic (Zhang 2005), and ultimate bearing capacity formula with plastic slide models (Vesic 1972), (2) field observations (Zhu et al. 2013, 2014; Zhao et al. 2012; Zhang et al. 2006), (3) numerical modeling (Yang et al. 2011; Nemcik et al. 2014), and (4) physical analogs (Jiang et al. 2011). The most effective method to measure the evolution of stress and strain in the mine excavation floor rocks is the in situ dynamic monitoring. Because coal mining is a dynamic process (Yang et al. 2011), the stress evolves with mining activities. Although field observations have been employed by many researchers, most studies do not acknowledge the dynamic redistribution of horizontal stress increments, and there is still a lack of understanding and quantifying of the changing stress patterns with distance and depth.

The present study is the first to discuss the quantitative relationship between stress increment on vertical and horizontal directions and depth under coal seam floor using in situ dynamic monitoring and to further analyze the different distance and depth influenced by stress on horizontal and vertical directions, respectively. The study contributes to the prevention of flooding disasters in coal mines with the help in determining the maximum depth of the fracture zone under the coal seam floor, developing dynamic monitoring technologies within the stress environment, and determining the width of waterproof pillars when faults or karst collapse columns are present.

Study area

Nantun Coal Mine is located in the southern Yanzhou coal field, which is the main coal production area in Shandong province of China (Fig. 1). The main minable coal seams belong to the formations of Carboniferous–Permian age (Fig. 1d). The main minable coal seams are the no. 16 and no. 17 coal seams from the Taiyuan formation (Pennsylvanian) and the no. 3 coal seam from the Shanxi formation (Cisuralian), which has been extensively mined. The no. 16 coal seam will likely become the object of the main mining activities in the future. The no. 1604 working face of no. 16 coal seam is the first to be opened in the western area of no. 11 distinct of Nantun Coal Mine (Fig. 1c). The working face elevation ranges from −387.1 to −405.1 m below mean sea level, and the ground elevation ranges from 40.5 to 41.8 m above mean sea level. The area of the working face is approximately 178,600 m2 with the strike length of 940 m and the dip width of 190 m. The thickness of the no. 16 coal seam ranges from 0.70 to 1.20 m, with an average of 0.90 m, and its dip angle is between 2° and 5°. Four faults with drop heights of 0.5–1.5 m were found by exploration at the middle roadway, and the minimum distance between the test point and the no. 16F38 fault is approximately 80 m. One fault with drop height of 6–10 m was identified beside the open-off cut of no. 1604 working face. The mining method consists of retreat longwall mining.

Fig. 1
figure 1

a Location of Nantun Coal Mine and the no. 1604 working face. b The figure with geological map at local scale; c figure with geological map in Yanzhou coalfield; d Geological cross sections representing formations, lithostratigraphy (age), tectonics, and hydrogeology

Full size image

The main roof of the no. 1604 working face consists of 10 m thick fine grained sandstone, and the immediate roof consists of limestone with a thickness of 4 m. The immediate floor of the working face consists of mudstone with a thickness of 1.2 m, and the main floor consists of fine and medium grained sandstone with a thickness of 5.5 m. The mechanics parameters of the rock mass underlying the coal seam are presented in Table 1.

Table 1 Rock mass mechanics parameters
Full size table

The main confined aquifers under the no. 16 coal seam are the Carboniferous no. 14 limestone and the Ordovician limestone. The two aquifers are separated from the no. 16 coal seam by 52 and 82 m thick aquicludes composed of mudstone and siltstone. The no. 14 limestone is 5.5 m thick. The in situ measured specific capacity of the no. 14 limestone was 0.000392–0.0138 l/s/m with a water pressure of 3.62 MPa. The massive Ordovician limestone is approximately 800 m thick and has a measured specific capacity of 0.0006–0.506 l/s/m and a water pressure of 5.6 MPa. The Ordovician aquifer is a confined karst aquifer containing an abundant supply of water and with a very high water pressure (Zhang and Shen 2004) and is the one posing a serious threat to safe mining by providing the flood water in Nantun Coal Mine.

The magnitude and orientation of the primary stresses including both vertical and horizontal stresses have tested with the stress relief method at depths between 443 and 456 m below ground surface at no. 1604 working face. The vertical stress are 10.3–11.7 MPa, whereas the horizontal stresses are 16.6–19.2 MPa and the orientations are 96°–101°.

Research methodology

Principle

In situ measurements of strain and stress have been widely used in coal mines by employing a multi-wire borehole instrumentation system consisting of mechanical sensors installed at different depths in boreholes drilled in longwall panels at mining working faces (Tan et al. 2010 b). As coal mining proceeds, the observed vertical and horizontal strains vary as a function of the changing underground pressure. Strain variations are caused directly by the floor strata deforming under stress, and the magnitude reflects the degree of deformation and failure to which the floor is subject. With the working face advancing, its floor experiences a series of deformations, and the degree and distribution of deformation can be measured out by a group of strain sensors installed at different depths in the boreholes distributed on different floor positions (Tan et al. 2010a). All strain sensors are installed within the same rock mass in order to take the material as an isotropic, homogeneous, and linear elastic rock material (Ge and Hou 2012), so the elastic theory is applicable to describing the relationship between strain and stress. Also, all the measured strain data are able to reflect stress features and distribution zones under coal seam floor with longwall mining.

Borehole drilling and sensors installation

A monitoring borehole was drilled in the headgate at a distance of 107 m distance from no. 1604 working face (Fig. 2). We calculated the value for failure depth of coal seam floor (FDCSF) roughly with an empirical formula (National Coal Industrial Bureau 2000), which is a function of the depth of mining, dip angle of coal seam, thickness of coal seam, and width of the working face. Equation (1) presents the empirical formula:

$$ h_{1} = w_{1} H + w_{2} A + w_{3} M + w_{4} L + w_{0} $$
(1)

where, h 1 is the failure depth of a coal seam floor in m, w i is a weight coefficient attributed to each factor, H is the depth of mining in m, A is the dip angle of coal seam in degree (°), M is the thickness of coal seam in m, L is the longwall panels width in m, and w 0 is the intercept.

Fig. 2
figure 2

Position of monitoring borehole at the no. 1604 working face

Full size image

The value for FDCSF was 20.04 m, which guided the depth of the borehole. The borehole had a diameter of 110 mm and was drilled at an 85° angle to a depth of 21.81 m, resulting in a total length of 22.58 m. Four ANZI (Australia and New Zealand Inflatable)-cells sensors (McQueen 2004), JCYL1, JCYL2, JCYL3, and JCYL4 were installed in the borehole at depths of 6.97, 9.69, 11.84, and 21.81 m, and in the medium grained sandstone, siltstone, mudstone, and aluminous mudstone, respectively (Fig. 3). Each testing sensor was made of 12 strain gauges produced by SCT Operations Pty Ltd., Australia (Sun 2003), that measured both the axial and radial strains. There were 6 groups of strain gauges placed on the cylinder surface with an interval of 60°. Then, the monitoring borehole was fully sealed with cement. The strain measurements started after 48 h of cement curing.

Fig. 3
figure 3

Lithological column and location and depth of sensors in monitoring borehole under coal seam floor. a The brief lithological column of coal bearing strata; b Lithological column of formation Taiyuan and depth of sensors in monitoring borehole

Full size image

Measurements

The strain below the excavation was monitored for 66 days of active mining activities. Strain data were collected every 3 m until the working face was 35 m away from the borehole after which the data were recorded every 1 m until the face passed the position of the borehole and continued for 16 m (Fig. 2). Greater distance between the monitoring borehole and the approaching face suggests weaker effect of mining activities. The 3-D strain data were then converted into vertical and horizontal stress data.

Results

Strain increment

Figures 4, 5, 6 and 7 show the strain increment curves based on the borehole data recorded by the JCYL1, JCYL2, JCYL3, and JCYL4 sensors at depths of 6.97, 9.69, 11.84, and 21.81 m, respectively. The horizontal axis represents the distance between the pre-mining and post-mining position of the working face relative to the location of the test borehole starting at 107 m and ending at −16 m. The number 0 represents the testing point, while positive and negative numbers represent pre-mining and post-mining positions, respectively. The vertical axis represents the microstrain which means the ratio of deformation and original size with the unit of με (one percent of one million of strain) in the 12 strain gauges. All four sensors worked well with the exception of the second gauge of sensor JCYL2, which was damaged during installation. Plotting the data and drawing the strain increment curves provided by the strain monitoring equipment showed that:

Fig. 4
figure 4

Strain increment curve of the no. JCYL1 sensor (the horizontal axis represents distance from working face to testing borehole, the number 0 represent the testing point, and positive and negative number represent pre-mining and post-mining position, respectively.)

Full size image
Fig. 5
figure 5

Strain increment curve of the no. JCYL2 sensor

Full size image
Fig. 6
figure 6

Strain increment curve of the no. JCYL3 sensor

Full size image
Fig. 7
figure 7

Strain increment curve of the no. JCYL4 sensor

Full size image
  1. 1.

    In the case of sensors JCYL1, JCYL2, and JCYL3 and for the curves located above the 0 increment line, the strain increment increased at different rates until the working face reached the position of the monitoring borehole and then decreased. The value of the strain increment was peaked at 422, 987, 1655, and 3665 με in JCYL4, JCYL3, JCYL2, and JCYL1 sensors, respectively. The opposite trend is present for the curves located below the 0 increment line. The location of the point at which the increase/decrease rates became significantly higher was a function of the sensor depth with the shallowest sensor being the first to record the rate change.

  2. 2.

    The data coming from the deepest sensor, JCYL4, showed a continuous increase (above the 0 increment line) or decrease (below the 0 increment line) of strain increments at slightly different rates irrespective of the position of the working face relative to the monitoring borehole with the sensors.

Vertical and horizontal stress increment

The vertical and horizontal stress increment curves of the four sensors (Fig. 8a, b) show that:

Fig. 8
figure 8

Vertical (a) and horizontal (b) stress increment curve of four sensors

Full size image
  1. 1.

    The vertical stress increment calculated for all sensors increased until the working face advanced to within 10 m from the position of the monitoring borehole and then decreased to 0 until past the borehole. The maximum stress values were 5.8, 5.0, 2.0, and 0.4 MPa for JCYL1, JCYL2, JCYL3, and JCYL4, respectively.

  2. 2.

    The horizontal stress increment followed a similar trend to the vertical one for JCYL3 and JCYL4 until 10 m past the borehole after which it increased slightly. The trends calculated for the other two sensors (JCLY1 and JCYL2) are similar, but the decrease and the final increase are steeper. All of horizontal tress directions are parallel to or less than 10 degrees with the mining direction during mining. However, there were two horizontal stress directions in the JCYL1 and JCYL2 sensors that had changed perpendicular to or approximately 80 degrees with tunnel after the mining action advanced the testing borehole position. The maximum horizontal stress increment values were 4.1, 3.5, 2.5, and 1.3 MPa for JCYL1, JCYL2, JCYL3, and JCYL4, respectively. The minimum values in JCYL1 and LCYL2 were −3.6 and −2.4 MPa, respectively.

Orientation of horizontal stress

Orientation and state of horizontal stress increment variation with mining at different depth under coal seam floor are shown in Fig. 9. Most of the orientation and of horizontal stresses are the same as the primary direction, except for the right parts of Fig. 9a, b, where the orientation had changed as the properties of stress transferred from compressive into tensile. Figure 9a shows the orientation of horizontal stress that keeps the primary direction for pre-mining 30 m. All the magnitude and orientation and properties of the horizontal stress had changed from pre-mining 20 m to post-mining in JCYL1. The direction changed from 96°–101° to 26°–170°, and the properties shifted from compressive into tensile after mining through the test point. Similarly, in JCYL2, all the orientation and properties of the horizontal stress just transferred at the mining position and post-mining, as showed in Fig. 9b. However, all the orientation and properties of horizontal stress shown in Fig. 9c, d had not changed, but the magnitude increased with decreasing distance of mining position to test point. The magnitudes in JCYL4 were smaller than those in JCYL3, and the orientation increased 14° clockwise in JCYL3.

Fig. 9
figure 9

Plan map of orientation and status of horizontal stress of four sensors (a, b, c and d represents the JCYL1, JCYL2, JCYL3, and JCYL4 sensor, respectively.)

Full size image

Discussion

The in situ dynamic monitoring of strain and stress under a coal seam floor as the working face advances during longwall mining recorded the spatial and temporal evolution of their increments.

The strain and both the vertical and horizontal stress increments under the coal seam floor were clearly affected by mining activities with their values and patterns being a function of depth and the distance between the location of the monitoring borehole and the advancing working face. Greater increments, indicative of greater deformation of the underlying strata, occurred at shallower depth and closer distance. Also, the transition from positive to negative horizontal stress values recorded in JCYL1 and JCYL2 marks the change from a compressive stress field to a tensile one. Therefore, it is expected that areas that are the most deformed and fractured under the coal seam floor where they are subjected first to compressive then to tensile stress.

Relationship between the maximum vertical and horizontal stresses with depth under coal seam

In general, the absolute values of the recorded strain increments decrease with increasing sensor depth in the monitoring borehole. The stress increment changed with depth both vertically and horizontally. The maximum values of both the horizontal and vertical stresses are functions of the sensor depth (Fig. 10a, b). The maximum vertical stress increment decreased exponentially with depth (Eq. 2), which is in general similar to what was described by Jiang (1993), but the coefficients were different.

$$ \Delta \sigma_{\text{z}} = 26.521 \times e^{ - 0.188z} $$
(2)

where \( \Delta \sigma_{\text{z}} \) is the maximum vertical stress increment in MPa; z is the depth under the coal seam floor in m. Of course, this formula is affected by the specific conditions, such as the geological condition, mining method, mechanics parameters, and types of rock masses under the floor.

Fig. 10
figure 10

Relation curve of Max.vertical (a) and horizontal (b) stress increment with different depth under coal seam floor

Full size image

The relationship between the maximum horizontal stress increment and depth was inverse as well, but it is linear and is expressed as follows (Eq. 3):

$$ \Delta \sigma_{\text{x}} = - 0.189z + 5.381 $$
(3)

where \( \Delta \sigma_{x} \) is the maximum horizontal stress increment in MPa; z is the depth under the coal seam floor in m. This formula indicates that there is no horizontal stress increment when the depth (z) is less than 28.48 m under the geological and mining conditions of the no. 1604 working face in Nantun Coal Mine.

Division of stress stages at horizontal direction under coal seam floor

According to the mining pressure theory, the floor strata responded at different degrees under the mining pressure. Four zones are divided at the horizontal direction from the viewpoint of stress condition (Jiang 1993), and they are: primary stress zone, stress-concentrated zone, stress relief zone, and stress recovery zone. The length of each zone depends on the geological condition, mining method, characteristics of coal seam, primary stresses field, mechanics parameters, and types of rock masses under the floor. Based on detailed analysis of the in situ data of strain and stress increments, the current study identified the following four stress stages of the floor strata during horizontal mining (Fig. 11):

Fig. 11
figure 11

Sketch of stress distribution under coal seam floor

Full size image
  1. 1.

    Weak response stage The vertical and horizontal stress increments increased at a low rate when the working face advanced from 107 to 20 m relative to the position of the monitoring borehole.

  2. 2.

    Significant increase stage From 20 to 2 m, the vertical and horizontal stress increments increased at a high rate.

  3. 3.

    Dramatic decrease stage From a distance of 2 m to the borehole to a distance of 10 m past the borehole, the vertical and horizontal stress increments decrease steeply and the horizontal stress switches from compressive to tensile.

  4. 4.

    Slow restoration stage After the working face advanced to more than 10 m past the borehole, the subfloor pressure started gradually the restoration process.

These different stages with different width are significant to design the width of waterproof pillars at faults or at karst collapse column in coal mine. Take the fault as the example, to avoid the “fault activation”, enough wide waterproof pillar at fault is the key method to prevent water flow through fault. “Regulations for coal mine water prevention and control” point out that the width of waterproof pillar at fault is bigger than 20 m at least, which is coincide with the length of significant increase stage in this article.

Division of stress stages at vertical direction under coal seam floor

In the vertical direction under coal seam floor, both the vertical and horizontal stresses and strains presented a different regularity, which helps to recognize the characteristics of the floor rock mass, especially on the fractures. Based on the underground pressure theory (Qian et al. 2010) and comparison of the different strain increment and stress status under the coal seam floor, some new fractures were reduced by the underground pressure. As shown in Fig. 9a, b, both orientation and properties of horizontal stress had changed, which suggests that the rock strata are undertaking a transition process from compressive stress to tensile stress. Meanwhile, the orientation changed from approximately EW to SN. Therefore, the rock strata were damaged easily where the conditions change from compressive deformation to tensile deformation until there is no tensile stress within the rock mass. We define the sign of new fractures induced by underground pressure as the existence of the tensile strain and stress. In this study, the JCYL2 sensor was installed in the siltstone, and the JCYL3 sensor was installed in mudstone. Siltstone is easier to crack than mudstone because siltstone is mechanically more fragile. On the contrary, because the mudstone is more flexible than siltstone, the new fractures are hard to expand and connect to each other. Therefore, no new fractures were developed or broadened, and connected in JCYL3 and JCYL4, as shown in Fig. 9c, d. Also, the maximum strain and stress increments were helpful to discriminate the fractures.

According to the fracture degree of underneath strata affected by coal mining, several failure zones can be identified (Tan et al. 2010b). The strata under the coal seam were divided vertically into three zones according to the maximum strain and stress values as follows (Fig. 11):

  1. 1.

    Fracture zone Greater increments occur at shallower depth, which is indicative of greater deformation of underlying strata. As shown in Figs. 4, 5, 6, 7, 8 and 9, the maximum depth of new fractures is close to but does not exceed the depth of sensor JCYL3 at 14.84 m. So, the fracture zone extended from the bottom of coal seam to about 13 m depth, which is similar to the results of Li (2008), who used water injection test in a borehole drilled into the roadway of the no. 1601 working face where the geological and hydrogeological conditions were similar to those present at the no. 1604 working face in the same coal mine (see Fig. 1c). Abundant fractures have developed under stress leading to higher hydraulic conductivity compared to the original undisturbed strata. Therefore, the depth of the fracture zone is a key parameter to prevent and control ground water inrushes from coal seam floor.

  2. 2.

    Deformation zone The rock strata were deformed without any significant fracture systems developing. The hydraulic conductivity of this zone tends to be less than that of the undisturbed strata because of the compressive stress. The deformation zone extends from 13 to 22 m below the bottom of the coal seam. This zone have an effect on resisting confined water from underlying aquifer.

  3. 3.

    Lightly affected zone Only minor or no deformations are present in the rock, and the hydraulic conductivity is similar to that of the undisturbed strata. This zone extends below 22 m under the coal seam. This zone is the most effective to obstruct the confined water from underlying aquifer.

Conclusions

The rocks below the coal seam floor are subjected to stress during longwall mining and undergo periodic compression, relief, and restoration. The maximum damage to the floor is caused by shear deformation and failure and occurs during the transition from compression to stress relief when the stress increments fluctuate widely. In addition, rock fracturing through delamination can occur during the stress relief period when the rock expands.

The characteristics and distribution of deformation and stress of the coal seam floor during longwall mining with backward extraction were studied in situ in Nantun Coal Mine, China. The spatial and temporal trends of strain and stress were monitored in the coal seam floor using four sensors installed at different depths in a borehole drilled into a mine roadway next to an advancing working face. The recording started when the working face was 107 m from the borehole and stopped when the face had advanced 16 m past the borehole. Four stress stages were obtained in the floor strata along the direction of mining: (1) weak response stage, (2) significant increase stage, (3) dramatic decrease stage, and (4) slow restoration stage. Based on the deformation characteristics, the floor also was divided into three vertical structural zones: (1) fractures zone, (2) deformation zone, and (3) lightly affected zone.

The in situ dynamic monitoring study revealed the inverse correlations between the horizontal and vertical maximum stress increments and depth. The correlations together with strain and stress pattern change during mining will allow mine specialists to predict when the damage to the coal seam floor is substantial and the safe operation of the mine is compromised.

The in situ dynamic monitoring data indicated that the maximum depth of mining-induced fracture zone under the coal seam floor at the no. 1604 working face was up to 13 m during the longwall mining. And the width of significant stress increase stage was 20 m ahead of the working face. It is helpful to get a clear understanding of the deformation and failure of the aquicludes between the coal seam and underlying confined aquifers.