Numerical investigation of failure evolution for the surrounding rock of a super‐large section chamber group in a deep coal mine

The stability control of surrounding rock for a large or super‐large section chamber group is a difficult technical problem in deep mining conditions, and this stability control has become one of the most important factors restricting the safety of a large‐scale coal mine. Based on the in‐site geological conditions of a super‐large section chamber group that functions for a coal gangues separation system in the Longgu Coal Mine, we first researched the stress, deformation, and failure characteristics of the in‐site chamber group surrounding rocks by using the FLAC3D software. Simulation results showed that the maximum vertical stress, deformation, and failure range of the surrounding rock for a super‐large section chamber group are larger than those of an ordinary section chamber. In addition, the roof subsidence and the plastic zone radius at the intersection are obviously larger than that of other parts, which increase by approximately 50.8% and 44.4%, respectively. Therefore, the chamber group intersection should be taken as the key area for surrounding rock control. Then, the influences of two key parameters of the chamber group, spacing and angle, were researched in detail to help in the design of the chamber group. Results show that when the chamber spacing is 80 m, the interaction begins to occur. When the angle is 70°, the stress of the surrounding rock and roof subsidence reach the minimum. Thus, the optimum chamber group parameters are determined to be a spacing of 80 m and an angle of 70°. Finally, from the perspective of chamber group stability, the optimal chamber group parameters can better meet the normal use for the super‐large section chamber group. This research provides a reference for the design of a super‐large section chamber group under the same or similar conditions and provides a strategy for controlling the surrounding rock.


| INTRODUCTION
Coal resources account for 70% and 60% of the primary energy production and consumption structure, respectively, [1][2][3][4][5][6] and they have always been the main energy source of China. In recent years, the mining scale and intensity of coal mines have gradually increased with the prosperity of coal production each year, and many large and super-large section chambers and chamber groups have emerged. [7][8][9][10][11][12] The cluster distribution of super-large section chambers makes the interaction between the chambers fierce, and it is very easy to cause dynamic accidents such as roof fall and rib spalling. This interaction has become a major technical problem restricting the safe mining for coal mines. Currently, some scholars have performed research on the failure evolution and control methods for the surrounding rock of chamber group and have revealed the deformation and failure characteristics of the surrounding rock under various conditions. [13][14][15][16][17][18][19] Unlike the single chamber, the deformation and failure evolution of chamber group are affected by many factors, including internal and external factor. In terms of internal factor, Qi et al 20 studied two different chamber layouts by using FLAC3D and obtained a stable chamber group structure. He et al 21 revealed the influence of excavation sequence on surrounding rock stability of soft rock intersection chamber group. For the external factor, Lin et al 22 explored the causes of deformation and control measures for the large section chamber in complex geological structural area by analyzing numerical simulation and field measurement results. Cheng et al 23 revealed the rock displacement law considering the lithological characteristics in the process of chamber group excavation. Sun et al 24 revealed that dynamic load disturbance makes the stability of surrounding rock of chamber group decline sharply and intensifies the stress concentration. Under the combined action of superimposed stress generated by the excavation and external factors, chamber groups interact with each other, causing local destruction and even chain instability in the whole chamber group.
However, the shallow coal resources have gradually become exhausted with the long-term continuous high-intensity mining of Chinese coal resources, and coal mining is gradually transferring to the depths. [25][26][27][28][29][30][31][32][33] In deep mining conditions, the mechanical properties and engineering response of coal rock have changed under the combined influence of high geostress and strong dynamic disturbance, [34][35][36][37][38][39][40][41][42][43][44] the surrounding rock deformed and fractured more seriously which makes it more difficult to control. Li et al 45,46 used numerical simulation and field measurement to analyze the instability mechanism of the surrounding rock in a soft rock roadway under high-stress conditions and proposed the surrounding rock stability control method. Qian et al 47 determined the influence of creep and time-dependent behavior on the stability of deep underground chamber through large fault zones in argillaceous rocks by in-site monitoring. Wang et al 48 deduced the critical load of initial buckling and continuous buckling failure of a thick and hard floor by using thin plate theory and revealed the floor failure mechanism of a deep large chamber. Sun et al 49 studied the mining-induced impacts on the stability, proposed a control strategy of overlying multi-tunnels with backfill mining by using a numerical simulation method, revealed that the panel width and advancing distance have an important influence on overlying multi-tunnel stability in a deep coal mine.
The above research studies have determined the deformation and failure characteristics of chamber group surrounding rock under some specific geological and mining conditions, which provided some support for stability control. However, there are a few studies on the deformation and failure evolution of the surrounding rock for a deep super-large section chamber group. In addition, the key parameters affecting its chain instability have not been thoroughly studied. Therefore, this paper takes coal gangue separation system of Longgu Coal Mine as the research background. FLAC3D numerical simulation software is used to reveal the failure evolution and characteristics of the surrounding rock for a deep super-large section chamber group. On this basis, different chamber group spacings and angles were simulated separately, and the optimal spacing and angle were determined through comparative analysis of the simulation results.

F I G U R E 2
The stratigraphic sequence of coal seam and its roof and floor strata F I G U R E 3 Layout and working process of the coal gangue separation system | 3127 of Heze City. The administrative division is subordinate to Longgu Town, Juye County. The location of Longgu Coal Mine is shown in Figure 1. The geological reserves of the Longgu Coal Mine are approximately 1.683 billion tons, the designed production capacity is approximately 6 million tons per year, and the designed service life is 82 years. Longgu Coal Mine has good development prospects.
The coal gangue separation system chamber group in Longgu Coal Mine mainly consists of four super-large section chambers: screening crushing chamber (SCC), screening product transferring chamber (SPTC), dense medium shallow groove separation chamber (DSSC) and slime water and medium chamber (SWMC). The coal gangue separation system chamber group is situated in the triangular area surrounded by the north main haulage roadway, No. 1 district rise and No. 1301 mining area. The mining area is situated in No. 3 coal, the buried depth is 778.6 ~ 848.6 m with an average of 813.6 m. The geological structure is simple which contains 0-3 layers of intercalated rocks, and the fissures are relatively developed. In addition, fat coal and coking coal are the main coal types in No. 3 coal of Longgu coal field, which are high-quality coal with low ash, sulfur, phosphorus, and high calorific value. The stratigraphic sequence of the coal seam and its roof and floor strata are shown in Figure 2.
The layout and working process of the coal gangue separation system are shown in Figure 3. The run-of-mine coal is transported from the north main haulage roadway to the SC chamber with a height of 9.0 m, width of 8.0 m, and length of 80 m. After screening and crushing operations, the raw coal is transported to the SPT chamber with a height of 9.0 m, width of 8.0 m, and length of 80 m. Moreover, the SPT chamber and the SC chamber are arranged vertically on the plane. Then, the raw coal is transported to the DSS chamber with a height of 9.0 m, width of 8.0 m, and length of 60 m. This chamber is perpendicular to the SPTC in plane and parallel to the SC chamber in the strike direction. Then, the clean coal is transported by belt conveyor to the north main haulage roadway, and the gangue is removed through the medium draining screen and discharged to the No. 1 district rise. To facilitate the addition of dense medium and the fluency of the scraping suction tank, the angle between the SWM chamber and the SC chamber is 74°, and the SWM chamber width is 8.0 m, height is 9.0 m, and length is 83 m.

| Rock mass properties
The mechanical properties of the rock mass have an important influence on the stability of the surrounding rock, and different rocks have different mechanical properties. According to the suggested method of the International Society for Rock Mechanics and Rock Engineering (ISRM), 50-55 the RLJW-2000 microcomputer-controlled rock servo test machine was used to carry out uniaxial compression tests on roof and floor rocks of the chamber group in the Longgu Coal Mine to obtain the mechanical parameters, as shown in Figure  4. The RLJW-2000 microcomputer-controlled rock servo test machine adopts a German DOLI full digital servo controller, with the advantages of high control accuracy, full protection function and strong reliability, which effectively ensures the reliability of the test results. In the test, the experimental materials used were 50 × 100 mm standard rock specimens, the axial load of the test machine was controlled by displacement loading, and the loading rate was 0. 25  However, the macroscopic strength of the rock mass is far less than the macroscopic strength of the rock samples due to the joints, cracks and different mineral compositions in engineering rock mass. Therefore, we use the strength reduction method [56][57][58][59] to properly modify and adjust the mechanical parameters of the rock samples obtained in the experiment to adapt to the in-site geological conditions. The mechanical properties used in the numerical simulation model are shown in Table 2.

| Model development
According to the in-site geological conditions of the coal gangue separation chamber group in the Longgu Coal Mine, the FLAC3D numerical simulation software was adopted for establishing the numerical model, as shown in Figure 5. The simulation model size is 320 m × 320 m × 200 m, and the model is divided into 3 169 580 zones and has 3 270 141 grid points. The model has a fixed bottom boundary and a stress boundary at the top. The stress value is the self-weight of the upper rock strata, and the horizontal displacement constraint is applied to the rest of the model. In addition, Mohr-coulomb was selected as the constitutive model.

| In-site chamber group
According to the in-site geological condition in the Longgu Coal Mine, the super-large section coal gangue separation chamber group is simulated, as shown in Figure 6. The length of the SC chamber is 80 m. The length of the SPT chamber is 80 m, and the SPT chamber is perpendicular to the SC chamber. The length of the DSS chamber is 60 m, and the DSS chamber is perpendicular to the SPT chamber and parallel to the SC chamber. The length of the SWM chamber is 83 m, and the angle between the SWM chamber and the SC chamber is 74°. In addition, the height and width of each chamber in the coal gangue separation chamber group are 9 m × 8 m. Then, displacement monitoring points are arranged at the center of the rock-coal pillar surrounded by the chamber group to monitor the vertical displacement of the rock-coal pillar. Five stress monitoring points are arranged 5 m apart along the diagonal line to monitor the magnitude and location of the maximum vertical stress.
The detailed simulation process was as follows (

Chamber group spacing
By changing the spacing between the SC chamber, the DSS chamber, the SPT chamber and the SWM chamber, 9 different simulation schemes of chamber group spacing were designed. The chamber group spacings were chosen as 100 m, 90 m, 80 m, 70 m, 60 m, 50 m, 40 m, 30 m, and 20 m. To control the research variables, the horizontal direction between chambers was set to be mutually perpendicular. At the same time, different simulation schemes of chamber group spacing were monitored, including the vertical displacement at the center of the rock-coal pillar surrounded by the chamber group and the vertical stress at the intersection of the chamber group. The simulation schemes of chamber group spacing and the measuring point arrangement are shown in Figure 8.

Chamber group angle
By changing the angle between the SC chamber and the SWM chamber in the horizontal direction, 8 different simulation schemes of chamber group angles were designed. The chamber group angles were chosen as 85°, 80°, 75°, 70°, 65°, 60°, 55°, and 50°. To control the variables, the distance between the SC chamber and the DSS chamber was set as 80 m, and the SPT chamber is perpendicular to the SC chamber and the DSS chamber. At the same time, displacement monitoring points are arranged at the center of the rock-coal pillar to monitor its vertical displacement, and stress monitoring points are arranged at the intersection of the chamber group to monitor the maximum vertical stress. The simulation schemes of the chamber group angle and the measuring point arrangement are shown in Figure 9.

Simulation procedure
The detailed simulation process was as follows ( The plastic zone area, roof subsidence and vertical stress of the in-site chamber group are shown in Figure 11. The vertical stress monitoring curve and the vertical displacement monitoring curve of the rock-coal pillar are shown in Figure 12. As shown in Figure 11A, the plastic zone of the chamber group is relatively developed due to the deep mining conditions and the super-large sectional of the chamber. The maximum failure depth of two sides of the SWM chamber is to 10 m, and the maximum failure depths of two sides of the SC chamber, the SPT chamber and the DSS chamber are 9 m. The maximum failure depth of the chamber group roof and floor is 11.0 m and 13.0 m, respectively. In addition, there are also some differences in the surrounding rock at the chamber group intersections. Marking the 4 chamber group intersections as A, B, C and D, the maximum failure radius at intersection A and C is 12.0 m, that at intersection B is 14.2 m, and at intersection D is 12.4 m. As shown in Figure 11B, the roof subsidence reaches the maximum at the chamber group intersection, and the maximum roof subsidence, which is 75.4 mm, appears at intersection C. As shown in Figure 11C, the stress concentration area appears at the chamber group intersection for the stress redistribution caused by chamber group excavation, and the maximum vertical stress is 48.4 MPa. The stress monitoring curve ( Figure 12A) shows that the stress concentration area is approximately 10 m away from intersection C, and the maximum vertical displacement at the rock-coal pillar is 181.0 mm, as shown in Figure 12B.
According to the figures, the maximum vertical stress of the coal gangue separation chamber group in the Longgu Coal Mine is approximately 10 m away from the intersection, which is 48.6 MPa. The vertical displacement of the rock-coal pillar surrounded by the chamber group is 181.0 mm. The maximum deformation of the surrounding rock is located at the roof of the chamber group intersection, which is 75.4 mm; the maximum failure radius of the surrounding rock is located at intersection B, which is 14.2 m. Therefore, the surrounding rock at the chamber group intersection is affected by high ground stress and repeated excavation disturbance, which makes stress concentration more complex and the maximum vertical stress larger. In addition, stress concentration causes the surrounding rock to exceed its ultimate strength, which intensifies the degree of failure, enlarges the scope of the failure and makes the surrounding rock deformation more intense. The roof subsidence and the plastic zone radius at the chamber group intersection are obviously larger than the roof subsidence and the plastic zone radius of other parts of the chamber groups, which increased by approximately 50.8% and 44.4%, respectively. In addition, the interaction between the chambers is more severe, and the radius of the plastic zone is the largest at the intersection B for the smaller angle of the SC chamber and the SWM chambers. Therefore, the chamber group intersection should be the key area of control of the surrounding rock, and the integrity of the intersection affects the overall stability of the chamber group.
Statistical analysis of the existing medium and large section chamber/roadway (Table 3) shows that the chamber/ roadway depth is positively correlated with the maximum vertical stress, deformation, and failure of the surrounding rock under the similar cross section. When the buried depth is similar, the cross section determines the stress and deformation of the surrounding rock, which means the larger the cross section, the more severe deformation and failure of the surrounding rock. Comparing the simulation results of the in-site chamber group with the existing medium and large section chamber/roadway, the surrounding rock failure range of the deep super-large section chamber is larger, the surrounding rock deformation is more severe, and the maximum vertical stress is higher during excavation.

| Chamber group spacing
The numerical simulation of the plastic zone and the vertical displacement for different chamber group spacings are shown in Figures 13-15. The rock-coal pillar vertical displacement monitoring curve and maximum vertical stress monitoring curve are shown in Figure 16. With the chamber group spacing decrease, the stress, displacement, and plastic zone area of the surrounding rock increased to varying degrees.
When the chamber group spacing is 100-80 m, the surrounding rock of the chamber group develops a large range of the plastic zone due to the large burial depth and cross section. The maximum failure depth of two sides is 9 m, and the maximum failure depth of the roof and the floor are 12.0 m and 13.0 m, respectively. Moreover, the maximum failure radius at the intersection of the chamber group is 11 m. With the decrease in the chamber spacing, the maximum failure radius of the surrounding rock remains basically unchanged, and the plastic zone area increases slightly. The maximum roof deformation is located at the chamber intersection, and the subsidence increases slightly from 69.9 mm to 72.7 mm. The deformation of the rock-coal pillar increases slightly, from 130.9 mm to 142.3 mm. The stress of the surrounding rock of the chamber group is redistributed, and the stress concentration area is located approximately 10 m from the chamber intersection. As the spacing decreases, the maximum vertical stress increases slightly from 44.2 MPa to 44.9 MPa. Therefore, there appears to be no interaction in the surrounding rock of the chamber group when the spacing is larger than 80 m.
When the chamber group spacing decreases from 80 m to 50 m, the maximum failure depth of the chamber two sides increases slightly, but the maximum failure depth of the roof and floor obviously increases, which is 14 m and 15 m, respectively. The failure range of the surrounding rock is obviously larger, and there is a tendency for mutual penetration at the floor. The maximum failure radius at the intersection increases to 13 m. The subsidence and deformation range of the roof obviously increased, and the maximum subsidence increased from 72.7 mm to 84.6 mm. In addition, the deformation of the rock-coal pillar is more obvious. According to the monitoring curve, the vertical displacement of the rock-coal pillar increases from 142.3 mm to 240.0 mm with the decrease in the chamber spacing. The maximum vertical stress of the surrounding rock also increases greatly, from 44.9 MPa to 49.0 MPa. At that point, there is an interaction between the chamber group. The stresses begin to superimpose on each other, leading to the maximum roof subsidence and deformation range at the intersection increasing significantly. The plastic zone area of the surrounding rock increases greatly, and there is a trend of mutual penetration. The first impact position is the middle floor surrounding rock.
As chamber group spacing continues to decrease, the interaction is further enhanced, the plastic zone area continues to increase, and the degree of failure is gradually intensified. When the spacing between chamber groups decreases from 50 m to 20 m, the plastic penetration zone of the surrounding rock transfers from the floor to the roof and then penetrates completely. Serious deformation and failure occur at the chamber group roof, and the maximum plastic zone radius reaches 17 m. The maximum roof subsidence range gradually covers the entire chamber group roof from the intersection, and the maximum subsidence increases from 84.6 mm to 138.7 mm. Furthermore, when the chamber group spacing is 20 m, the chamber group surrounding rock loses its bearing capacity and cannot continue to bear the overlying strata load due to the complete penetration of the plastic zone. The plastic zone transfers to the deep intact rock mass of the chamber group. Therefore, the maximum vertical stress of the surrounding rock of the chamber group suddenly decreases, and the vertical displacement of the rock-coal pillar first increases and then decreases before gradually stabilizing at a spacing of 20 m.
The relationship of the plastic zone area, the maximum roof subsidence, the rock-coal pillar vertical displacement, and the maximum vertical stress with chamber group spacing F I G U R E 1 0 Key parameters simulation procedure is calculated, as shown in Figure 17. The figures show that when the spacing is 80 m, the plastic zone area increases by 3.1%, the maximum roof subsidence increases by 4.0%, the rock-coal pillar vertical displacement increases by 8.7% and the maximum vertical stress increases by 1.6% compared with the spacing of 100 m, which shows that there is hardly any mutual interaction when the chamber group spacing decreases from 100 m to 80 m, and the surrounding rock remains intact and stable. When the spacing decreases from 80 m to 70 m, the plastic zone area increases by 7.1%, the maximum roof subsidence increases by 7.7%, the rock-coal pillar vertical displacement increases by 7.7%, and the maximum vertical stress increases by 3.6%. Thus, when the spacing is 70 m, the increase range is obviously increased, and chamber groups begin to interact with each other. When the chamber group spacing decreases to 30 m, the plastic zone area increases by 55.9%, the maximum roof subsidence increases by 45.9%, the rock-coal pillar vertical displacement increases by 139.0%, and the maximum vertical stress increases by 32.1% compared with the spacing of 80 m. At this point, the interaction of the chamber groups is intense, leading to the instability failure in the surrounding rock. When chamber group spacing is 20 m, the plastic zone area and the maximum roof subsidence continue to increase, 18.0% and 30.7% higher than the plastic zone area and the maximum roof subsidence of 30 m. However, due to the complete penetration of the plastic zone into the surrounding rock, the coal pillar loses its bearing capacity and cannot support the overlying strata load. Therefore, the vertical displacement and maximum vertical stress of the rock-coal pillar decrease. Therefore, the optimal spacing of the chamber group can be determined to be 80 m.
The plastic zone area and the maximum roof subsidence are fitted by function, respectively. With the chamber group spacing decrease, the plastic zone area increases parabolically, conforming to the quadratic polynomial relationship. The maximum roof subsidence increases exponentially, conforming to the exponential relationship. The correlation coefficient squares (R 2 ) are both greater than .99. Then, the rock-coal pillar vertical displacement and the maximum vertical stress without spacing 20 m are fitted. The rock-coal pillar vertical displacement and the maximum vertical stress increase exponentially with the decrease in the chamber spacing, both in accordance with the exponential relationship, and R 2 is greater than .99.

| Chamber group angle
The numerical simulation of the plastic zone and the vertical displacement for different chamber group angles are shown in Figures 18 and 19. The rock-coal pillar vertical displacement monitoring curve and maximum vertical stress monitoring curve are shown in Figure 20.
The figures show that when the chamber group angle is 85°-70°, the plastic zone radius at the intersection is larger than other parts of the chamber group and shows a growth trend. To be precise, the maximum plastic zone radius at intersection A, C, and D is slightly increased from 12.0 m to 13.2 m. The failure at intersection B is more obvious, where the maximum plastic zone radius increases from 12.7 m to 15.3 m. Roof deformation decreases with chamber group angle reduction. The maximum roof subsidence occurs at chamber intersection C, from 78.0 mm to 74.6 mm. The minimum roof deformation occurs at chamber intersection B, from 75.0 mm to 70.0 mm. The monitoring result of the rock-coal pillar vertical displacement shows that the deformation decreases from 213.4 mm to 168.4 mm. Moreover, the stress monitoring result of the surrounding rock shows that the stress concentration area is approximately 10 m away from intersection A. The maximum vertical stress decreases from 50.8 MPa to 47.9 MPa with the decrease of the chamber group angle. Therefore, when the chamber group angle is 85°-70°, the rock-coal vertical displacement, the maximum vertical stress, and the maximum roof subsidence are reduced to different extents, and the maximum plastic zone radius is gradually increased.
When the chamber group angle is 70°-50°, the plastic zone radius at the intersection continues to increase, and the maximum plastic zone radius at intersection D increases slightly from 12.7 mm to 15.0 mm. The increase at intersection B is most pronounced, from 15.3 m to 22.4 m. The deformation of the rock-coal pillar is further reduced, from 168.4 mm to 141.7 mm. Roof deformation shows different trends with chamber group angle decrease. The roof subsidence at intersection C increases from 74.6 mm to 78.6 mm, while the roof subsidence at intersection B continues to decrease from 70.0 mm to 60.0 mm. The stress concentration area is still located approximately 10 m away from intersection A, and the maximum vertical stress increases with chamber group angle decrease, from 47.9 MPa to 50.5 MPa. Therefore, when the chamber group angle is 70°-50°, the maximum plastic zone radius, the maximum vertical stress and the maximum roof subsidence increase in varying degrees, and the rock-coal pillar vertical displacement is further reduced.
The relationship of the maximum plastic zone radius at the intersection, the rock-coal pillar vertical displacement, the maximum roof subsidence and the maximum vertical stress with the chamber group angle are counted, as shown in Figure 21. With the chamber group angle decrease, the maximum plastic zone radius increases continuously. Compared with the angle of 85°, the maximum plastic zone radius is increased by 76.4%, and the rock-coal pillar vertical displacement decreases by 33.6% at an angle of 50°. The maximum plastic zone radius and the rock-coal pillar vertical displacement are fitted by function, respectively ( Figure  21A). The maximum plastic zone radius at the intersection increases approximately parabolically with the decrease in the chamber group angle, which conforms to the quadratic polynomial relationship. In addition, the rock-coal pillar vertical displacement decreases approximately parabolically, which conforms to the quadratic polynomial relationship as well. The R 2 are both greater than .98. As shown in Figure  21B, the maximum vertical stress and the maximum roof subsidence first decrease and then increase. Compared with the angle of 85°, the maximum vertical stress decreases by 5.7% and the maximum roof subsidence decreases by 4.6% at the angle of 70°. When the angle is reduced to 50°, the maximum vertical stress increases by 5.4% and the maximum roof subsidence increases by 5.5% compared with the angle of 70°. Based on the above analysis, the optimal angle of the chamber group can be determined to be 70°.

| Verification and optimization
According to the chamber group parameters determined by the above research, the optimal chamber group parameters are compared with the in-site parameters, as shown in Figure  22. The maximum roof subsidence is reduced from 75.4 mm to 74.6 mm, the maximum vertical stress is reduced from T A B L E 3 Statistics of existing medium and large section chamber/roadway 46,48,[60][61][62][63][64][65][66] No.  48.4 MPa to 47.9 MPa, and the rock-coal pillar vertical displacement is reduced from 181.0 mm to 168.4 mm. In addition, the maximum plastic zone radius is slightly increased, from 14.2 m to 15.3 m because the area of the rock-coal pillar increases with the decrease in the chamber group angle, which plays an effective supporting role on the overlying strata. In addition, the interaction between two adjacent chambers intensifies, which causes the surrounding rock at the intersection B to exceed its ultimate strength and leads eventually to large-scale failure. The field application effect of a super-large section chamber group for coal gangue separation system in the Longgu Coal Mine is shown in Figure  23. The roof deformation is small, and the surrounding rock of the chamber group is relatively intact, which can meet the normal production needs.

Location
The research results show that the in-site chamber group spacing is the optimum, and the chamber group angle can continue to be optimized, not only further reducing the peak stress and deformation of the surrounding rock of the chamber group but also effectively reducing the excavation quantity and postmaintenance workload. This study provides guidance and reference for the layout design of the chamber group under the same or similar geological conditions.

| DISCUSSION
In this paper, the authors revealed the failure evolution for a super-large section chamber group in a deep coal mine and determined the optimal spacing and angle. Currently, some scholars focus on medium and large section chambers under shallow or medium burial depth [13][14][15][16][17][18] to obtain the failure evolution and influencing factors of the surrounding rock of a chamber group under specific geological and mining conditions, which provides some support for the stability control. However, there are a few studies on deep super-large section chamber groups, and the research on the spacing and angle of the chamber group is also scarce. In this paper, the authors reveal the failure evolution and characteristics of the surrounding rock for a deep super-large section chamber group and find that the intersection area of the chamber group is the key area for controlling the overall chamber group, providing a solid foundation for the support design. Moreover, chamber groups with different spacings and angles were simulated separately, and the influences on the stress, deformation, and failure evolution of the surrounding rocks were obtained. On this basis, the optimal spacing and angle were determined. This study provides guidance and reference for the design of a super-large section chamber group construction under the same or similar conditions and its surrounding rock control strategy.
We adopt the controlling variables method in the optimal chamber group parameter determination, which can make the simulation schemes more pertinent, but it also has limitations to a certain extent. Moreover, we obtained the optimal spacing and angle of the chamber group by numerical simulation but did not obtain a quantitative analytical solution theoretically. In future work, we will further establish the corresponding mechanical model to reveal the instability mechanism of surrounding rock to make it more widely acceptable.

| CONCLUSIONS
The aim of this research was to improve the stability of superlarge section chamber group in a deep coal mine by studying the failure evolution of surrounding rocks and determining its optimal key parameters. Compared with the current published works, this work has at least two original aspects: 1. The surrounding rock stress, deformation, and failure evolution for an in-site chamber group were revealed, and the critical area for controlling the stability of the chamber group was found to be the intersection area.
2. The influences of two key parameters, spacing and angle, on the surrounding rock stability of a deep super-large section chamber group were studied, and the optimal spacing and angle were determined.
Numerical simulations showed that the stress concentration, plastic zone and deformation of the surrounding rocks of coal gangue separation system are much larger than those of other medium or large section chambers/roadways. Moreover, the roof subsidence and plastic zone radius at the intersection are approximately 50.8% and 44.4% larger than those of other parts for the superimposed stress and repeated excavation disturbance. Additionally, with the spacing decreasing from 100 m to 20 m, the plastic zone area, roof subsidence, and vertical stress remain constant first and increase significantly since the spacing is smaller than 80 m. With the angle decreasing from 85° to 50°, the vertical stress and roof subsidence first decrease and then increase, and they all reach the minimum at the angle of 70°. Compared with the actual parameters, the spacing is the optimum, but the angle can be further optimized. Field application observations verifies this conclusion.
It should be noted that we did not obtain a quantitative analytical solution from the theoretical perspective in this work. In future work, we will further establish the corresponding mechanical model to study the instability mechanism of chamber group surrounding rock.