Floor Failure Characteristics in Deep Island Longwall Panel: Theoretical Analysis and Field Verification

Beijing Key Laboratory for Precise Mining of Intergrown Energy and Resources, China University of Mining and Technology (Beijing), Beijing 100083, China School of Energy and Mining Engineering, China University of Mining and Technology (Beijing), Beijing 100083, China State Key Laboratory of Coal Resources and Safe Mining, China University of Mining and Technology (Beijing), Beijing 100083, China School of Mechanics and Civil Engineering, China University of Mining and Technology (Beijing), Beijing 100083, China


Introduction
With the increase of operational coal mining depth in China, the surrounding rock in deep mining is prone to large-scale destabilization or dynamic response damage [1]. Increased range of floor failure zone induced by the coupling of high-water pressure and ground stress as well as strong mining disturbances in deep coal mining [2][3][4][5][6]. Water inrushes occur when floor failure zones are connected with high water pressure aquifers in the overlying strata under the floor [7]. Therefore, it is essential to investigate the floor fail-ure characteristics induced by mining and confined aquifers to manage floor water inrush.
Three methods of theoretical analysis, numerical calculation and on-site monitoring are generally applied to investigate the floor failure characteristics. In terms of theoretical analysis, according to the abutment pressure distribution around the working face, a mechanical model of coal seam floor failure under the abutment pressure induced by coal seam mining was established, and the stress distribution and the maximum failure depth of the coal seam floor during mining were also calculated [8][9][10][11]. Xue et al. [12] established a mechanical calculation model of the floor failure patterns in the strike and dip directions along the working face after mining to study the floor failure characteristics of roof cutting and gob side entry retaining in confined water. The floor rock mass was regarded as a transversely isotropic body and an analytical solution for the floor stress was derived [13][14][15][16]. A mechanical model for the floor failure depth in inclined coal seam mining was established to obtain the floor failure characteristics in inclined coal seams as "upper-small-lower-large" [17][18][19]. Song [19] and Liang et al. [20,21] established a mechanical model under the combined action of water pressure and mining by considering the actions of high-water pressure and analyzed the stress field and failure characteristics. Ma et al. [5] proposed an improved theoretical model that considered the whole confined aquifer in the floor to analyze the stress distribution and failure zones.
On-site monitoring is a direct and effective method for understanding the extent of damage to the surrounding rock. Currently, the most commonly used field monitoring methods are the borehole water injection, borehole acoustic wave, and borehole imaging as well as borehole strain method [22][23][24][25]. In recent years, microseismic monitoring technology has been applied by many scholars to monitor floor damage and water inrush problems. Cheng et al. [26] adopted a microseismic monitoring system to identify the spatial location and formation process of water flow channels. Zhao et al. [27] proposed a set of rock mass seepage channel inversion methods based on microseismic data. Zhou et al. [28] presented a microseismic monitoring analysis method for evaluating potential seepage channels. Ma et al. [5] investigated the evolution characteristics and development pattern of floor failure during high thickness coal seam mining using microseismic monitoring. Cheng et al. [29] examined and calculated the floor disturbance depth through microseismic virtual reality visualization.
Scholars have conducted many investigations and yielded abundant findings in research on the failure characteristics of coal seam floor. Due to the inclination of the coal seam, the failure characteristics of the surrounding rock of the working face are obviously different from those of the horizontal and near horizontal coal seams [2,30]. In addition, due to the superposition of high stress caused by nearby mined longwall panels, the high-stress concentration zone is formed in the surrounding rock [31]. The large deformation and large-scale failure of floor rock strata are induced by high-stress concentration. Water-conducting channels tend to be formed due to the large-scale and high-intensity stress disturbance induced by repeated mining [24,32]. The risk of water inrush from coal seam floor in deep island longwall panel increases under deep complex hydrogeological conditions [16]. However, most previous scholars have investigated the floor failure characteristics of longwall mining faces along strikes, and few have addressed the floor failure characteristics of island longwall mining panels in up-dip mining. Therefore, the study of floor failure characteristics of the island longwall panel in up-dip mining above confined water are essential to propose a reasonable method to prevent and control floor water inrush in deep mining.
Therefore, a mechanical model of the island longwall panel in up-dip mining under the combined effect of highwater and mining pressure was established in this paper. The stress distribution in the mining floor rock mass was calculated, and then the floor failure characteristics were obtained based on the Mohr-Coulomb criterion. In addition, the floor failure characteristics before and after RHF were comparatively analyzed, and the results of on-site measurements verified the reasonableness of the theoretical model. The research results could provide a reference for studying the mechanism and prevention of floor water inrush under complex mining geological conditions.

Site Descriptions
The Xingdong coal mine is located in the northeastern Hanxing coalfield, Hebei Province, China. Figure 1 shows the general mining layout and the location of the study area. The mining depth of the longwall panel NO. 2129 at Xingdong coal mine is 1027-1125 m and 92.5 m wide, and the lengths of the ventilation and haulage roadways are 416 m and 367 m, respectively. The longwall panel NO. 2129 is an island longwall panel with a goaf on both sides. The ventilation roadway is adjacent to the goafs of longwall panels NOs. 2125-2127, and water inrush occurred in these panels. The ventilation roadway is also adjacent to the longwall panels NO. 2123 and NO. 2124 that do not have floor water inrushes. The average width of coal pillar between ventilation roadway and the goafs is 30 m, which is called narrow The water pressure of the Ordovician limestone aquifer is 10-15 MPa. After regional grouting treatment, water inrush disasters from the coal seam floor in deep mining continue to occur frequently, which seriously restricts the safe production of the Xingdong coal mine.

Theoretical Analysis
3.1. Basic Principles. Based on the spatial semi-infinite body theory in elasticity [33] for the semi-infinite body bearing normal load or tangential load (Figure 3), the stress compo-nent of any point M (x, y) can be expressed by Equations (1) and (2), respectively.
Assuming that By substituting Equation (3) into Equations (1) and (2), the polar coordinate expression of vertical stress, horizontal stress, and shear stress can be obtained as shown in Equations (4) and (5).

Model Establishment.
According to statistical analysis of floor water inrush disaster cases in deep mining, floor water inrush is frequently induced by strong mining pressure during periodic weighting [34,35]. Therefore, it is of significant practical engineering to explore the redistribution of the floor stress and the floor failure range during periodic weighting [36]. Taking the longwall panel NO. 2129 at Xingdong coal mine as an example, the mining stress and the original rock stress are superimposed during the advancement of the working face. Figure 4 shows the general state of the floor rock mass of the island longwall panel in up-dip mining during periodic weighting based on the theory of ground pressure and strata control [37]. α is the dip angle of coal seam in Figure 4(a). Notably, compared with previous work, the dip angle of the mining direction and the goaf on both sides of the working face are considered, thereby enriching the investigation of floor stress and failure characteristics under different mining conditions. Therefore, a mechanical model for the combination of the hydraulic pressure and mining pressure was constructed, as shown in Figure 5 and according to Figure 4. Both the mining pressure and the water pressure are simplified to a linear distribution in Figure 5. The water pressure is regarded as an uneven external force applied to the bottom of the aquifuge. In addition, the gravity of the floor strata is also considered. As shown in Figures 1 and 4 In addition, the loads caused by the abutment pressure in Figure 5 are simplified to two triangular loads and decomposed accordingly into normal loads (q 1 and q 2 ) vertical to the floor and tangential loads (q 1 ′ and q 2 ′) parallel to the floor. The maximum value of them is determined by ðk 1 − 1ÞγHcosα. k 1 represents stress concentration coefficient, γ is the average bulk density of the overlying strata, and H denotes burial depth of coal seam. The floor stress in the mined zone is considered a uniformly distributed rectangular load q 0 , which is released to the free surface with the magnitude of −γHcosα. The stress recovery zone is gradually compacted by the overlying strata of the collapse zone, which is regarded as a triangular distribution load q 3 with the minimum value of −γHcosα and the maximum value of 0. Since floor failure is mainly affected by the stress increment, the component of the stress increment zone parallel to the floor is considered, ignoring the component of the stress in the unloading zone.
In Figure 5, L 1 is the length of the elastic zone caused by advanced abutment pressure, L 2 is the length of the plastic zone caused by advanced abutment pressure, L 3 is the length of the crushing zone induced by advanced abutment pressure, L 4 is the length of the stress recovery zone, L 5

Geofluids
The rock strata are assumed to be a continuous, intact, homogeneous, and an isotropic medium in the developed mechanical model, and the rock mass is assumed to apply the linear elastic constitutive model. According to the above assumptions, the stress components of M (x, z) can be derived by substituting the equations for all loads in Figure 5 into the corresponding Equations (4) and (5).
The calculation of stress components resulting from the triangular load q 1 are shown in Equation (6).
As regards the stress components induced by the triangular load q 2 , θ 1 and θ 2 in Equation (6) should be replaced by θ 3 and θ 2 , respectively, and the corresponding distance by L 2 . Similarly, the corresponding stress components for the triangular loads q 3 , q 4, q 6, q 7, q 9, q 10, and q 11 can be obtained.
The calculation of stress components resulting from the triangular load q 0 are shown in Equation (7).
For the stress component induced by the rectangular load q 8 , θ 4 and θ 5 in Equation (7) should be replaced by θ 11 and θ 12 , respectively. Similarly, the corresponding stress component for the rectangular load q 12 can be obtained.
Equation (8) calculate the stress components caused by the load q 5 .
The stress components caused by the load q 1 ′ can be obtained using Equation (9).
For the stress component induced by the load q 2 ′ , θ 1 and θ 2 in Equation (9) should be replaced by θ 3 and θ 2 , respectively.
In Equations (6)- (9), H is the burial depth of the coal seam, α is the dip angle of the coal seam, k 1 is the concentration coefficient of advance abutment pressure, k 2 and k 3 are the concentration coefficient of abutment pressure on coal pillar at both ends of working face along the strike, γ indicates bulk density of the rock mass, ρ w is density of confined water, P 0 denotes the hydraulic pressure of the confined aquifer, g is gravitational acceleration, and H z is the distance between the coal seam floor and confined aquifer.
The angle in the above Equation can be calculated according to Equation (10). ð11Þ Where Equation (11) is the total stress component of the floor stress along the inclination and Equation (12) is the total stress component of the floor stress along the strike.
According to the site investigation, it is found that the roof of the coal seam #2 at Xingdong coal mine is difficult to cave in time, with a larger overhanging roof distance, which leads to a larger weighting interval, and the concentration degree of nearby abutment pressure will increase,      The vertical stress concentration is observed in the floor underlying the advanced abutment pressure zone along the inclination (Figure 6(a)). The maximum value of stress concentration is near the elastic-plastic bond of the coal seam, that is, near the peak of the advanced abutment pressure. The stress unloading zone is mainly located in the mined area, as well as the maximum degree of stress relief within the weighting interval (L 5 ). Along the strike direction, the vertical stress concentration in the floor below the isolated coal pillar due to the overhanging roof of the coal seam after mining is greater than that in other areas (Figure 6(b)). Figure 7 shows that the variation of the horizontal stress is consistent with the distribution of the vertical stress. The degree of concentration and unloading and influence range of horizontal stress are significantly smaller than those of vertical stress. The influence range of horizontal stress     9 Geofluids decreases rapidly as the depth increases. However, tensile stress is generated in the unloading zone along the inclination and along the strike.
The shear stress in the mined zone and the coal wall approximately presents a positive and a negative shear force couple along the inclination, forming a sharp shear stress variation zone (Figure 8(a)). The floor strata of this area are prone to compression shear or tension shear deformation damage, which is consistent with the vulnerable location around the stope on site. The maximum shear stress appears at the coal wall zones affected by advanced abutment pressure. The shear stress along the strike is also a positive and negative shear couple in the floor strata under the mined zones and the coal pillars, while the maximum shear stress appears on the side of the isolated coal pillar (Figure 8(b)).

Failure Characteristics of Floor Strata. Based on the
Mohr-Coulomb failure criterion [38], the rock mass will be damaged when the maximum shear stress at a point in the floor rock strata equals or exceeds its shear strength. Equation (13) calculates the condition for shear failure to occur in the floor [17].
where the maximum shear stress τ max is calculated as follows: Substituting Equation (15) into Equation (13), we can construct the floor failure criterion Equation (15).
where c and φ denote the cohesion and internal friction angle of the floor rock mass, respectively.
When F (x, z) ≤ 0, the floor strata fails. According to the laboratory test results, the weighted average method is used to determine the average friction angle φ = 35:4°and cohesive force c = 4 MPa and substituting into Equation (15). The F (x, z) contour is obtained by using Python programming to calculate the theoretical floor failure characteristics of the island longwall panel in up-dip mining above confined aquifer during periodic weighting, as shown in Figure 9.
As seen in Figure 9(a), the floor failure characteristic along the inclination is spoon-shaped. Under the effect of advanced abutment pressure and high-water pressure, the maximum depth of floor failure reaches 43.2 m. The location of the maximum failure depth is in the floor of the stress concentration area of the coal wall in front of the working face. Tensile failure occurs within 10 m below the mined zone, resulting in floor heave. Shear failure occurs in the range of 10 m-43.2 m below the mined zone. With the increase of the floor failure depth, high-pressure confined water can easily rise from the shear failure zone to the bottom of the aquifuge, resulting in water inrush accidents. Figure 9(b) shows that the floor failure range along the strike has an "asymmetric saddle shape." The maximum failure depth of the floor is 45.1 m, which is located on the side of the isolated coal pillar. Due to the abutment pressure of the two working faces on the side of the isolated coal pillar, the stress concentration factor increases, resulting in a large floor failure depth, and floor water inrush is likely to occur.

In Situ Measurement
To verify the feasibility of theoretical analysis, the borehole televiewer (BTV) and microseismic monitoring were used to detect the floor failure characteristics of the island longwall panel NO. 2129 before and after RHF.

Results of Borehole
Televiewer Detection before RHF. The BTV is an effective means to visually analyze the failure range of the surrounding rock [39]. The evolution of internal fractures in the surrounding rock can be observed by a digital borehole television [40], which helps to determine the depth and range of fracture development induced by mining [41]. Therefore, to investigate the floor borehole, we examined the distribution of fractures in the sidewalls of the borehole by using a BTV ( Figure 10). A CXK12(A) of BTV is equipped with a wide-viewing panoramic color camera with 1.34 million pixels, which can resolve cracks of 0.1 mm. According to the detection results, the floor fracture development demarcation point can be observed. The BTV image shows that the cracks are densely developed 0-45.7 m away from the borehole wall ( Figure 10), but no cracks exceed this

Results of Microseismic
Monitoring after RHF. The formation process of the water conducting channels and the damage degree of the surrounding rock can be determined by using microseismic events [5,[26][27][28]. To monitor the development range of mining fractures in the floor after RHF, a microseismic monitoring system was installed in the roadway of the longwall panel NO. 2129 ( Figure 11). A total of 10 microseismic sensors were installed in the ventilation roadway and haulage roadway, and the microseismic events were transmitted to the data processing center in real time through the collection substation. The sensors are ESG (Engineering Seismology Group) detectors. Three collection substations were installed with a sampling frequency of 5 kHz.     Figure 12 shows the distribution of microseismic monitoring events after roof fracturing. The abscissa is the position of the working face, and the ordinate is the depth of the coal seam floor. Figure 12(a) shows that the microseismic events in the floor are concentrated within 15 m of the floor, indicating that the shallow floor rock mass is first damaged under the action of mining stress. With increasing floor depth, the density of floor microseismic events decreases gradually. The distribution shape of the microseismic event density in the floor is a "spoon-shaped" along the advancing direction of the working face. According to the microseismic monitoring results, the maximum mining failure depth is 28.6 m (Figure 12(a)). Microseismic events are mainly concentrated on the floor of the goaf at the side of the isolated coal pillar along the strike. The floor failure range is also an "asymmetric inverted saddle-shaped," and the maximum failure depth is 29.1 m, located on the side of the isolated coal pillar (Figure 12

Conclusion
A two-dimensional hydraulic calculation model was proposed to analyze the floor failure characteristics in deep island longwall panel using the up-dip mining method. The effects of mining stress and confined water pressure in the calculation model was considered in this paper. The floor failure depth and failure characteristics before and after RHF were discussed in association with the BTV and microseismic monitoring system, verifying the theoretical model. The main conclusions in this paper are as follows: (1) The vertical stress in the floor is principally concentrated in the rock strata advanced the working face and under the coal pillar. The degree of concentration and unloading of the vertical stress are signifi-cantly greater than those of the horizontal stress. The shear stress distribution is approximately the coupling of positive and negative. Along the working face advance direction, the local stress concentration on the side of the coal pillar, and stress release on the side of the mined zone are observed. The stress concentration under the coal pillar is greater than that in other zones along the strike due to the stress superposition on the coal pillar

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.