Mechanism and Control of Rockburst Induced by Draining Spatial Islands and Squaring

National Engineering Laboratory of Coalmine Backfilling Mining, Shandong University of Science and Technology, Taian 271000, China State Key Laboratory of Mining-Induced Response and Disaster Prevention and Control in Deep Coal Mines, Anhui University of Science and Technology, Huainan 232000, China School of Mining Engineering, Anhui University of Science and Technology, Huainan 232000, China Coal Burst Prevention and Control Research Centre, Yankuang Group Co., Ltd., Zoucheng 273500, China Chahasu Coal Mine, Inner Mongolia Energy Co., Ltd., Inner Mongolia, Ordos 017000, China

In this study, rockburst manifestations were primarily related to working face squaring and spatial island working faces, so we primarily collected and sorted through existing data related to both spatial island working faces and squaring. Cao et al. [15] theoretically examined how the spatial structure and rupture movement of thick and hard overburden of island working faces would affect mine earthquake activities. ey held that the large-scale rupture movement of the key strata of a working faces' "T"-shaped overburden space structure was the primary force source of mine earthquake activities as well as provided densified borehole pressure relief in the primary squaring, second squaring, and third squaring regions of working faces. By combining the overburden space structure and mine pressure, Jiang et al. [16] investigated the overburden movement laws of the "103 upper 02 asymmetric island fully mechanized working face" of the Baodian Coal Mine, explored the prediction method for "hard rock fracture-type" mine earthquakes, and identified the "squaring" region of a working face as its primary rockburst hazard region.
Island working faces are generally classified as either plane island working faces (formed by goafs on the two sides of the same coal seam) or spatial island working faces (formed by goafs on the two sides of different coal seams, formed by faults and goafs, or formed by drainage regions and goafs (under investigation in this study)). Many experts and scholars have extensively studied spatial island working faces (formed by two goafs of different coal seams or formed by faults and goafs). Yang et al. [17] proposed putting the rockbursts of island working faces under "classified control" and classified island working faces into six types, that is, island working faces with critical extraction on both sides, island working faces with subcritical extraction on both sides, island working faces with critical extraction on one side, and subcritical extraction on the other side, stereoscopic island working faces formed by goafs of different coal seams, hidden island working faces formed by geological structures, and composite island working faces formed by a geological structure-goaf. Cao et al. [18] examined the dynamic load mechanism of fault slip in the rockburst appearance of an island working face adjacent to a fault on one side and a goaf on the other side and performed a CT-based back analysis on the rockburst hazard of a working face with dynamic early warning. Zhu et al. [19] studied the overall instability type of rockburst in a quasi-island working face formed by fault cutting. Wang et al. [20] investigated the mine pressure manifestations of the 220 quasi-island working face formed by mining of the 3-2 and 4-2 diverged/ converged working section of Xiashijie Coal Mine in Tongchuan, finding that the dynamic pressure manifestation of the working face was closely related to periodic weighting and roof activities. e above scholars have extensively studied island working faces formed by faults and goafs as well as those formed by goafs of different coal seams. is study primarily investigated a type of novel static-stress spatial islands formed by roof drainage regions and goafs. ese spatial islands were formed by drainage stress transfer and goaf lateral stress transfer in the sandstone aquifers of the Yan'an Formation and the Zhiluo Formation. e erratic stress anomaly regions formed inside working faces inevitably affect production safety. In particular, the coupling of stress anomaly regions and working face squaring regions produced abnormal pressure manifestations on the gob side of the coupling region. is is an issue worthy of further research. With regard to the obvious dynamic phenomena occurring on the gob side of a spatial island working face, the formation and stress evolution process of the spatial island working face were explored using theoretical analysis as well as on-site monitoring. e force sources behind its rockburst appearance were also analyzed. e mechanism of rockburst induced by high static loads and by dynamic and static loads under the combined action of spatial islands and working face squaring was clarified, and control measures were prepared.

Engineering Background
When the 2202 gob-side working face of a coal mine in the Shaanxi-Inner Mongolia mining area was advanced to 200 m-300 m, the gob-side gate road of the working face experienced frequent coal burst and ejection phenomena as well as abnormal pressure manifestations in the yellow area in Figure 1. e 2202 working face had a burial depth of 731.4 m, a dip angle of 0-4, and a coal mass strength of 13.3 MPa and was characterized by simple coal seam structure and small relief. e effects of geological structures (such as faults and folds) were not detected. Table 1 provides the composite columnar section of the 2202 working face. e space above the 2-2 coal seam contains alternating sandy mudstone and medium sandstone. e relative water-rich anomaly regions of the Yan'an Formation and the Zhiluo Formation occur in the roof. e area filled by the red transverse lines is the relative water-rich anomaly region of the Yan'an Formation, 6.87 m away from the 2-2 coal seam. e area filled by the green transverse lines is the relative water-rich anomaly region of the Zhiluo Formation, 58.34 m away from the 2-2 coal seam.

Mechanism of Rockburst Induced by Spatial Islands
Island working faces are generally formed by goafs on the two sides caused by production continuity problems. Pillartype islands are also present and are the result of pillar recycling. As for the spatial islands formed in the 2202 working face in this study, on the one side, there were goafs formed after mining of the 2201 working face. On the other side, there were goaf-like stress transfer regions formed by the relative water-rich anomaly regions of the roof after dewatering, such as 1# and 3# spatial islands in Figure 2. In addition, some spatial islands were formed by drainage regions alone, such as 2# spatial island in Figure 2. 1# spatial island was formed by drainage regions (primarily the drainage region of the Yan'an Formation) and goafs. After dewatering of the sandstone water-rich region of the Yan'an Formation, the support of the upper strata was weakened, and the upper strata settled and deformed. e stress transferred towards the drainage region boundary and superposed with goaf lateral stress, resulting in a unimodal stress distribution. Depending on the effect of the drainage region of the Zhiluo Formation on 1# spatial island, 1# spatial island was divided into two zones, i.e., zone 1 and zone 2. Zone 1 was located below the drainage region of the Zhiluo Formation. e decline in coal seam stress in the drainage region of the Zhiluo Formation lowered the stress level of the spatial island, as illustrated by the profile in Figure 3. Zone 2 was located outside the drainage region of the Zhiluo Formation and raised the stress level of the island working face, as illustrated by the profile in Figure 4.
2# spatial island was formed by the sandstone water-rich region of the Yan'an Formation after dewatering; however, the entire spatial island working face was located below the drainage region of the Zhiluo Formation. As shown in Figure 5, the mine pressure manifestations of the spatial island, affected by interactions between the sandstone drainage regions of the Yan'an and Zhiluo Formations, were weak.
3# spatial island was formed by the drainage region of the Zhiluo Formation and goafs. e drainage stress transfer of the sandstone water-rich region of the Zhiluo Formation superposed with goaf lateral stress, resulting in bimodal stress distribution as shown in Figure 6. e stress concentration level of 3# spatial island was lower than that of 1# spatial island. 3# spatial island was formed by the drainage region of the Zhiluo Formation and goafs. e drainage stress transfer of the sandstone water-rich region of the Zhiluo Formation superposed with goaf lateral stress, resulting in bimodal stress distribution. e stress concentration level of 3# spatial island was lower than that of 1# spatial island. e breakage degree of coal-rock mass is related to stress level, and coal-rock mass breakage produces microseismic phenomena. us, microseismic energy and frequency can satisfactorily reflect the stress concentration level of coal-rock    at is, the higher the energy and frequency of microseismic events, the higher the stress concentration level within a coal-rock mass [21]. In this study, the presence of spatial island working faces was validated using "SOS" microseismic monitoring system software monitoring data. All microseismic events from the 2202 working face occurring over cumulative footage of 487 m were collected. Events with an energy level of above 10E3 were screened and projected onto the layout plan of the working face. Figure 7 shows the distribution of microseismic events. e distribution of microseismic events satisfactorily reflects the stress concentration levels of different regions of the working face. eoretical analysis in this study revealed that 1# spatial island region had a high-stress concentration level as well as a significant concentration of microseismic events.
1# spatial island region was divided into three zones, that is, zone 1, zone 2, and zone 3. Zone 2 was located below the drainage region of the Zhiluo Formation. Considering the wide scope of 3# spatial island, a zone in 3# spatial island with the same width as 1# spatial island was selected as zone 4 to avoid the effect of goaf distance. e microseismic data (see Figure 8 e average frequency of a zone was defined as the ratio of the total microseismic frequency of the zone to its total area, and the average energy of a zone was defined as the ratio of the total microseismic energy of the zone to its total area (the two special events with an energy level of 10E6 were excluded in statistics), as provided in Table 2.
e average frequencies of zones 2 and 3 were 2.04-2.41 times of those of other zones. e average energies of zones 2 and 3 were 1.37-2.04 times of those of other zones. ese results validated the presence of 1# spatial island. Zones 4 and 5 had similar average frequencies and average energies, which verified the conclusion that the stress transferred from the drainage region of the Zhiluo Formation was not intensified by superposition with 2201 goaf lateral stress. e average frequency and average energy of zone 2 were 0.92 and 0.78 times of those of zone 3, respectively, which confirmed the protective effect of the drainage region of the Zhiluo Formation for its underlying spatial islands.

Mechanism of Rockburst Induced by Squaring Effect
Starting with the open-off cut, goaf area increased linearly as the working face advanced. e squaring position was the dividing line marking the transition from short edge change to long edge change for the goaf. Before squaring, the two long edges of the goaf remained constant, while its two short edges continuously increased. Upon squaring, the short and long edges became equal in length. After squaring, the two short edges of the goaf remained constant, while its two long edges continuously increased. e roof stratum of the coal seam can be regarded as a constantly changing rectangular plate. In this case, identifying the distribution pattern of a coal mass stress under the action of the rectangular plate is of vital significance for clarifying the essence of the dynamic pressure of working face squaring as well as understanding  in depth the effect of overburden space structure on the evolution of coal mass stress. For a rectangular stope, the advance distance and length of the working face are assumed to be a and b, respectively. According to existing studies [22][23][24], before squaring (a < b), the bending moment distribution, coal mass plastic zone, and coal mass stress of the roof rectangular plate all indicate that the long edges of the rectangular plate have a relatively large moment load bearing capacity. e lead abutment stress of the working face increases with its advance.
When advance distance is equal to the length of the working face (a � b) [25], the working face was in the squaring stage, and σ Z reaches its maximum value as shown in Figure 9. eoretically, the abutment pressure of the stope is highest when the working face advances to "squaring." Clearly, working face squaring commonly produces stress anomaly regions, which can easily trigger the occurrence of rockbursts. e squaring stage is the formation stage of stope overburden space structure, and the gob-side working face produces an "S"-shaped stope overburden space structure. e initial fracture of high-level strata produces dynamic loads, which commonly causes dynamic disasters in coal seams [27,28].
When the 2202 working face advanced to 210 m without any significant change in mining velocity, microseismic event count and maximum energy increased rapidly, so the 2202 working face produced a significant squaring effect, using "origin" analysis software to make Figure 10.

Mechanism of Rockburst Induced by Static Stress.
High static load stress is the stress path that leads to the occurrence of rockburst. When the static load reaches the critical rockburst load (i.e., when the static load exceeds 1.5 times of the comprehensive compressive strength of the coal mass), a high rockburst hazard occurs [29,30].
where σ z is the gravity stress; c is the stress gradient, set as 2.5 MPa/100 m; and H is the burial depth of the working face. e 1# spatial island region demarcated in the 2202 working face is affected by lead abutment pressure, 2201 goaf lateral pressure, drainage boost region, and squaring-induced static load increase. e static load σ j of the coal mass can be calculated using the following formula: where k is the static load concentration coefficient of 1# spatial island; α is the lead abutment pressure concentration coefficient; ß is the lateral pressure concentration coefficient of the 2201 goaf; λ is the stress concentration coefficient of the drainage boost region; and υ is the squaring-induced static load concentration coefficient. e average comprehensive compressive coefficient μ of the coal mass ranges from 1 to 5 and is set as 1 in the vicinity of roadway surface as well as 3-5 in three-dimensional stress 1#spatial island 3# spatial island  Geofluids state. In this study, 1# spatial island region was 75 m in width, and the width of pressure relief was 20 m. μ was set as 3.
When the static load is equal to or greater than 1.5 times the comprehensive compressive strength of the coal mass, 1# spatial island working face may experience a static load rockburst; that is, and when the superposed sum of the stress concentration coefficients of the lead abutment pressure, 2201 goaf lateral pressure, drainage boost region, and squaring-induced static load increase is equal to or greater than 3.27, spatial island regions may experience static load rockburst.

Mechanism of Rockburst Induced by Dynamic and Static
Loads. When the superposed sum is less than 3.27, the critical static load for rockburst is not reached. When the dynamic load produced by the initial fracture of high-level overburden is transferred to the coal mass, a rockburst may be triggered.
Related studies have revealed that rockbursts are triggered by mine earthquakes or blasting under highly concentrated loads, that is, by superposition of dynamic and static loads [31][32][33][34][35], as expressed by the following formula: where σ j is the static load on the coal-rock mass; σ d is the dynamic load produced by mine earthquakes or blasting; and σ b is the ultimate load in case of a rockburst. e deep coal mass of the working face exists in a triaxial stress state. e comprehensive compressive strength of deep coal mass is far greater than that of roadway surface coal mass, and the critical rockburst load of deep coal mass is also far greater than that of roadway surface coal mass (see Figure 11). 6 Geofluids e dynamic load in the abnormal mine pressure manifestation region of 1# spatial island working face is primarily the dynamic load produced during spatial structure formation in the squaring period. Spatial structures are formed by fracturing of high-level rock beams, which produces a dynamic load greater than the first-weighting and periodic-weighting dynamic loads of the working face [28]. In this study, the periodic weighting dynamic load coefficient of the 2202 working face fell within the range of 1.00-1.80. When the dynamic load coefficient in the squaring period was set to 1.80, the static load stress concentration coefficient required by rockburst induced by superposition of dynamic and static loads would be �3.27/ 1.80≈1.81. Under the combined action of lead abutment pressure, 2201 goaf lateral pressure, drainage boost region, and squaring-induced static load increase, the squaring region was exposed to rockburst induced by the superposition of dynamic and static loads.

Rockburst Control Measures
According to the force sources for rockbursts identified in the above analysis, a control scheme was prepared, including overall hydraulic fracturing of the external roof of the drainage region, reduction of stoping speed, and pressure relief of large-diameter boreholes. 2202 working face adopts pressure relief measures such as large-diameter drilling and hydraulic fracturing along the goaf. e number and energy of microseismic events increase obviously during the square period. 4 events with energy above 10E5J were detected by microseismic monitoring system (coal stress, drilling cuttings method does not monitor the data exceeding the standard).

Overall Hydraulic Fracturing of the External Roof of the Drainage Region.
In roof hydraulic fracturing, a directional initial crack was made in the roof, and high-pressure water was continuously injected into the crack, in order to propagate the roof crack until it connected with adjacent cracks and damaged roof integrity [36]. e fracture characteristics and 3D fractal fracture of solid-fluid coupling have been studied [37,38]. Rockburst in the drainage region was primarily induced by drainage stress transfer, but the transferred stress remained in the working face and affected the safe stoping of the working face. 1# spatial island working face formed because the drainage region transferred stress to a more dangerous area. As a result of roof hydraulic fracturing outside the drainage region, the drainage region and the hydraulic fracturing zone formed an overall regular pressure relief protection layer (see Figure 12), which could reduce static loads in the fracturing zone as well as at the drainage boundary and weaken the fracturing dynamic loads of highlevel strata.

Pressure Relief of Large-Diameter Boreholes (Triaxial Transfer of Stress and Reduction of Static Loads on Roadway
Surrounding Rock). In pressure relief of large-diameter boreholes, boreholes distributed at small spacings connected the cracks in coal mass to relieve and transfer coal mass stress. In this way, they modified the mechanical properties of the coal mass and weakened the ability of coal seams to store elastic energy, mitigating rockburst within the coal mass [39].
Large-diameter boreholes formed a pressure relief belt around the roadway and kept coal mass stress away from the roadway (see Figure 13). e comprehensive compressive strength and the critical rockburst load of the deep coal mass were far greater than those of the coal mass near the roadway, mitigating the chance of a rockburst.

Reduction of Stoping Speed.
e aim of reducing stoping speed is to lower lead abutment pressure and reduce the effect of stoping on overburden. Studies have shown that a lower stoping speed leads to smaller peak lead abutment stress and longer distance of peak position from the working face. A low stoping speed can reduce the movement of roof masonry beam structures, lower the level of energy released by their fractures, and weaken the intensity of mine earthquakes. ere is a critical stoping speed under specific stoping conditions. After exceeding the critical stoping speed, microseismic energy and major event frequency both increase significantly with increasing stoping speed. In contrast, before reaching the critical stoping speed, microseismic monitoring data are not responsive to stoping speed [40].

Geofluids
In regions facing the rockburst induced by high static loads and dynamic and static loads, properly reducing stoping speed helps to mitigate lead stress concentration on the working face and weaken the intensity of the fracturing dynamic loads of roof strata, thus achieving the goal of rockburst control. In the square period of the working face, the mining speed and the microseismic maximum energy are shown in Figure 14. e microseismic monitoring system begins to detect large energy events when the mining speed exceeds 3.8 m.
is phenomenon verifies the relationship between the mining speed and the intensity of roof movement.

Conclusions
With regard to the specific conditions of the 2202 working face, this study explored the static-stress spatial islands formed by roof drainage regions and goafs as well as the squaring effect of the working face and analyzed the dynamic and static load mechanism of the rockburst appearance of the gob-side roadway of the 2202 working face from stoping to squaring stage. e static load evolution process of drainage regions and goafs under different spatial relationships was analyzed to obtain the stress states of various spatial islands. e drainage region of the Yan'an Formation and 2201 goaf formed 1# spatial island (after initial mining). e average frequency and average energy of 1# spatial island were 2.04-2.41 times and 1.37-2.04 times of those of other zones, respectively.
is result validated the presence of spatial islands in the working face.
Based on the evolution of the roof rectangular plate as well as existing studies on numerical simulation, it was found that the lead abutment stress of the squaring region was 1.2 times of that of the nonsquaring region, and high static loads existed in the coal wall in front of the working face during the squaring period. e microseismic frequency and maximum energy of the squaring region of the 2202 working face were both more than three times of those of the nonsquaring region, so there was a significant squaringinduced dynamic loading effect.
Following the theory of rockburst induced by superposition of dynamic and static loads, this study set the critical stress concentration coefficient for rockburst induced by spatial islands and squaring under high static loads as 3.27 and the critical static-stress concentration coefficient for rockburst induced by dynamic and static loads as 1.81. A control scheme was prepared, including overall hydraulic fracturing of the external roof of the drainage region, reduction of stoping speed, and pressure relief of large-diameter boreholes. e first squaring region was advanced safely on-site. ere were no spatial islands in the second squaring region, nor were any rockburst observed.

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

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.