CDEM simulation of top‐coal caving characteristics under synchronous multiwindow drawing in longwall top‐coal caving mining of extra‐thick coal seam

Abstract

Traditionally, longwall top coal caving (LTCC) panels adopt the single‐window drawing mode and their drawing efficiency is relatively low. The synchronous multi‐window drawing (SMWD) mode is a remarkable approach to resolve the incompatibility between shearer coal mining and top‐coal drawing. The present paper analyzes the caving characteristics of top coal under SMWD in LTCC mining of extra‐thick coal seam. The caving and migration of top‐coal fragments were simulated by using the continuum–discontinuum element method (CDEM). The results show that the SMWD mode can effectively improve the drawing smoothness with larger caving channels compared to the traditional single‐window drawing mode. The main caving channel for top‐coal fragments was directly above the assembled multi‐window (AMW), and the size of the AMW had a great influence on the distribution of the velocity field. Additionally, the simulated draw body of the initial drawing adopted an elliptical shape, which changed into a variation ellipse with the increasing size of the AMW. The boundary of the simulated draw body of the initial drawing presented an irregular shape that deviated from the results predicted by the Bergmark–Roos model. Furthermore, the drawing efficiency increased with the expansion of the AMW, whereas the recovery ratio first increased and then decreased with the expansion of the AMW. Finally, field tests revealed that the SMWD mode can greatly improve the drawing efficiency. Thus, the recovery ratio under A2 and A3 SMWD modes was higher than that of the traditional single‐window drawing mode. Based on the results of field tests and CDEM simulation, it is recommended to adopt A2 and A3 SMWD modes to enhance the top‐coal recovery ratio and drawing efficiency. These research results could constitute a technical reference for automated highly efficient top‐coal drawing and high‐yield LTCC panel extracting extra‐thick coal seam.

INTRODUCTION

Longwall top coal caving (LTCC) mining is the most efficient method for thick and extra-thick coal seam extraction. It is extensively applied in western China.^1–5 As illustrated in Figure 1, the bottom coal is cut by the shearer, while the thick top coal is fractured and crushed by the advanced abutment pressure or artificial weakening method.^6–9 Then, a rear scraper conveyor is utilized to recover the caved top coal from the drawing window. The first prototype of LTCC mining dates back to the 1960s and 1970s, when some European countries, especially France and the former Yugoslavia, used the total soutirage method to extract thick and irregular coal seams.^2,10 In the early 1980s, LTCC mining was first introduced in China.^11–14 After 40 years of development and innovation, the LTCC mining method has experienced rapid development and has succeeded in China.^11,15 Nowadays, the LTCC technology is the main mining method for the extraction of extra-thick coal seams in high-yield coal mines with an annual output exceeding 10 million tons. There are more than 20 coal mines that apply the LTCC mining method to extract extra-thick coal seams, yielding more than 10 million tons of raw coal annually in western China. The typical engineering case is the Tashan Coal Mine in the Datong Coalfield.^16,17 In this mine, the LTCC method is used to mine carboniferous 3–5 coal seams with a cutting height of 3.8 m and a caving height of approximately 16.0 m. The maximum annual output of a single LTCC face could reach 10 million tons.

Figure 1. Sketch map of the longwall top coal caving mining method

In the top-coal drawing process, there is a conflicting situation between increasing the drawing volume of the top coal and reducing the amount of gangue mixing.¹⁸ Owing to the coal loss in the drawing process, the recovery ratio of the LTCC mining method is not satisfactory. To improve the top-coal recovery ratio and reduce the gangue content, many scholars studied cavability,^1,2,7,19,20 deformational characteristics,^9,13,21–23 caving mechanisms,^3,24–26 drawing laws or flow characteristics^5,27–30 of top coal through theoretical analysis, numerical simulation, and laboratory tests. Top-coal caving characteristics and the migration law are closely related to the recovery ratio and gangue mixing ratio.^3,6,29 Besides, the caving characteristics and migration law of the top coal constitute core contents of the fundamental theory of LTCC mining.

To study the deformation and movement law of top coal and roof stratum, multipoint extensometer and tracking instruments were installed in inclined boreholes drilled into the top coal and roof.¹³ This field observation provides a direct method for top-coal deformation analysis by monitoring the vertical and horizontal displacements of the top coal.^13,21 Furthermore, the fragmentation and caving process of the top coal and roof strata were studied by physical modeling and numerical simulation.^9,11,23,24 Concerning the drawing characteristics, based on the similarity principle, Wu and Yu found that the draw body presents a deflection ellipsoid shape according to a laboratory physical simulation test, and they derived a mathematical model¹⁸ to describe the draw body. Assuming that the top coal above the support is a loose granular medium, Wang et al. proposed a granular flow model for top-coal caving and a new research system for top-coal drawing.^3,5,27,31 This new research system can take into account the boundary of top coal, draw body, recovery ratio, and rock mixed ratio. Based on the aforementioned granular flow model and research system, the morphology of the draw body, flowing law of top coal, and characteristics of the coal–gangue interface were systematically studied. In recent years, Wang and his team have studied the influence of the size distribution on the flow characteristics of granular top coal in LTCC mining and developed a prediction model for top-coal recovery.^28,32,33 In the drawing process, the flow and migration laws are essential for optimizing the drawing parameters and improving the recovery ratio.³⁰ Using the particle flow code, flow and caving characteristics such as velocity field, migration trajectory, and coal–gangue interface of the granular top coal have been simulated by many researchers.^6,29 The results show that the unevenly distributed velocity field could lead to the formation of arches above the drawing window, which affects the top-coal drawing efficiency and recovery ratio.³⁴ Recently, the continuum–discontinuum element method (CDEM) and the finite discrete element method have been applied in the analysis of the drawing process of LTCC mining,^35,36 and the interaction between the top coal and the roof strata has been considered by coupling the particle and block elements.³⁶ Moreover, the particle model of the CDEM has been used to simulate the drawing process of synergetic multiwindow caving in LTCC mining.³⁷

For LTCC face extracting extra-thick coal seam, the top-coal fragments have a longer flow path owing to the relatively greater thickness of the top coal. In addition, the fragmentation of the upper part lags behind the lower part of the thick top coal, and the top-coal fragments exhibit a relatively large block size.³⁸ Based on physical modeling and field tests, the thick top coal could be classified into three layers according to its fragmentation and caving features.³⁹ Furthermore, the stone bands in the extra-thick-coal seams always lead to insufficient fragmentation of the top coal, which has adverse effects on top-coal drawing.¹¹ Therefore, the arch in top coal is generated more easily in LTCC mining of extra-thick coal seam compared with conventional thick coal seam.⁴⁰

The above research results assume that the top coal is sufficiently fragmented, that is, the top coal above the shield could be regarded as a granular media. However, in LTCC mining of extra-thick coal seam, top-coal fragmentation is insufficient and heterogeneous owing to the ultra caving height of the top coal (ranging from three to five times the cutting height). Consequently, it is difficult to completely capture the nonuniform crushing state of the ultra-thick top coal using the theory of granular media. Besides, most of the existing top-coal drawing characteristics and migration laws were obtained under the single-window drawing mode. Given that the caving area of a single drawing window is relatively small, the top-coal fragments easily form an arch, which makes the drawing process insufficient and extremely limits further improvement of the coal drawing efficiency. In the near future, intelligent LTCC mining technology will constitute the development direction and target for extra-thick coal seam extraction.⁴¹ Indeed, coal mining is in the process of being transformed from automation to intelligent mining. Nowadays, high-yield, high-efficiency automatic LTCC mining of extra-thick-coal seam requires further improvement of the drawing efficiency and recovery ratio, which puts forward the need for research on synchronous multi-window drawing (SMWD). In recent years, SMWD has attracted much research attention; however, the relevant drawing theory and migration law are still under experimentation.

There is a pressing need to study the caving characteristics of top-coal fragments under SMWD in LTCC mining of extra-thick coal seam. Taking the nonuniform fragmentation of thick top coal into consideration, in the present study, we adopted a block model to study the top-coal drawing process using the CDEM code. The caving characteristics of the crushed top-coal fragments under SMWD were comprehensively studied. The dynamic development of the draw body, coal–gangue interface, and distribution of the top-coal velocity field was obtained through CDEM simulation. In addition, the top-coal drawing efficiency and recovery ratio under SMWD were analyzed. Furthermore, a field test of SMWD was conducted in the 8222 automatic LTCC panel of the Tashan Coal Mine. It is recommended to adopt A2 and A3 SMWD modes to further improve the top-coal recovery ratio and drawing efficiency according to the conducted field test and simulation results. These research results could set up a reference for automatic top-coal drawing in LTCC mining of extra-thick coal seam under similar conditions.

BASIC THEORY OF CDEM AND TOP-COAL DRAWING Profile of CDEM

CDEM is a typical hybrid simulation method that couples the finite element and distinct element methods. It has been applied for the simulation of the progressive failure process of geotechnical materials.^42–44 The CDEM code is a dynamic explicit simulation program widely used in both mining engineering and oil and gas engineering.^43,45 In a CDEM model, illustrated in Figure 2, the computational domain is represented by a series of discrete blocks, and each block is discretized with a mesh consisting of finite elements. The interfaces, which act as a bond, are embedded between the edges of all adjacent blocks. On the one hand, the blocks are used to model the continuous properties of the material, such as elasticity and plasticity, and the deformation and stress state of each block element is solved by dynamic relaxation technology. On the other hand, the interface is adopted to characterize the discontinuous properties, such as joints and fractures of the materials. CDEM simulation comprises a large number of interacting blocks. When the stress exceeds the material strength, the transition from a continuum to a discontinuum is accomplished by fracturing and fragmenting processes. CDEM does not only capture the elastic–plastic behavior in small deformation problems but also enables simulation of large deformation problems associated with fracturing and caving. In view of this, this method has been widely used in both geotechnical and mining engineering in recent years.^36,43,45,46

Figure 2. Schematic diagram of the continuum–discontinuum element method model

Fundamental principles of CDEM

For each finite element in a CDEM numerical model, the governing equation is established based on the Lagrange system as follows: [Image Omitted. See PDF]where $${\bf{M}}$$ denotes the concentrated mass matrix, $${\bf{C}}$$ is the damping matrix, $${\bf{K}}$$ represents the element stiffness matrix, $${\bf{u}}$$ is the displacement vector, and $${\bf{F}}$$ is the vector of external forces.

In this governing equation, numerical damping is utilized to model quasi-static problems by dynamic relaxation. At each time step, the internal forces are calculated through the deformation of the block and interface. The relative displacement of the contact interface of adjacent elements is proportional to the spring forces and can be calculated by Hooke's law as follows: [Image Omitted. See PDF]where $${\rm{\Delta }}{u}_{{\rm{n}}}$$ and $${\rm{\Delta }}{u}_{{\rm{\tau }}}$$ denote the normal and tangential relative displacements, respectively; $${\rm{\Delta }}{F}_{{\rm{n}}}$$ and $${\rm{\Delta }}{F}_{{\rm{\tau }}}$$ represent the increment of the normal and tangential forces, respectively; $${\rm{\Delta }}{\sigma }_{{\rm{n}}}$$ and $${\rm{\Delta }}{\sigma }_{{\rm{\tau }}}$$ are the normal and tangential stresses of the contact pair, respectively; $${K}_{{\rm{n}}}$$ and $${K}_{{\rm{\tau }}}$$ are spring stiffness in normal and tangential directions, respectively; and $$A$$ is the area of the contact interface.

An explicit finite-difference integration scheme is applied to solve the equations of motion for each element separately. The difference form is expressed as follows: [Image Omitted. See PDF]where $$n$$ denotes the current time step, $$n+1$$ is the next time step, and $${\rm{\Delta }}t$$ represents the time increment.

Failure criterion Failure criterion for blocks

The combined criterion of maximum tensile stress and Mohr–Coulomb criterion was implemented in the CDEM code. This combined criterion was set to be more consistent with the actual conditions achieved by the simulation of changes in the shear strength with mounting confining pressure: [Image Omitted. See PDF]where $${c}_{{\rm{e}}}$$ and $${\phi }_{{\rm{e}}}$$ represent the cohesion and friction angle of the element, respectively, and $${\sigma }_{\text{et}}$$ denotes the tensile strength of the element.

Failure criterion for interfaces

The interface between elements can be viewed as springs in the CDEM model, and all the interfaces are set as contact elements, including a normal spring and a tangential spring, through initialization. The state of the springs is judged by the tension-shear composite criterion. For example, if the forces pulling the two jointed elements apart exceed the ultimate tensile strength, a crack is created by breaking the bond: [Image Omitted. See PDF]where $${\sigma }_{{\rm{n}}}$$ denotes the normal stress on the interface and $${\sigma }_{\mathrm{ft}}$$ is the tensile strength. After tension failure, the normal contact force of the contact pair is set to 0.

According to the Mohr–Coulomb criterion, when the tangential stress of the contact interface reaches its shear strength, the block will experience a shear failure: [Image Omitted. See PDF]where $${\sigma }_{{\rm{t}}}$$ denotes the tangential stress on the contact interfacem and $${c}_{{\rm{f}}}$$ and $${\phi }_{{\rm{f}}}$$ are the cohesion and friction angles of the interface, respectively. After shear failure, the tangential contact force of the contact pair is modified as $${F}_{{\rm{\tau }}}={F}_{{\rm{n}}}\times \tan \,{\phi }_{{\rm{f}}}$$.

Theory of top-coal drawing

In research on top-coal drawing in LTCC mining, the crushed top coal is postulated to be loose granular media,^3,47 and the drawing process is modeled by the gravity flow of top-coal fragments. The outline of the original location of the top coal that has been drawn from the draw opening is termed as draw body, and the outline of the original location of top coal or gangue that has moved from its original location is called ellipsoid of movement.^48,49 Several theories and mathematical models were posed to predict the drawing law and flow characteristics of the caved coal and rock.^18,50–53 The Bergmark–Roos model, shown in Figure 3, is one of the acceptable theories for calculating the draw body of the granular top coal.^36,47,54 This model postulates that the top-coal fragment moves along the line connecting its initial position and the center of the drawing opening with constant acceleration.^48,50 In addition, it assumes that the fragments are only affected by gravity and friction caused by the interaction with adjacent particles, and all the fragments are at rest before the opening of the draw window.

Figure 3. Schematic diagram of the Bergmark–Roos model

On any migration trajectory, the acceleration of the fragment can be calculated as follows: [Image Omitted. See PDF]where $$g$$ is the acceleration of gravity, $$\theta $$ represents the migration angle, that is, the angle between the direction of migration and the vertical direction, and $${\theta }_{{\rm{m}}}$$ denotes the maximum migration angle. At the $${\theta }_{{\rm{m}}}$$ angle, the projection of gravity in the migration direction is equal to the friction force, that is, $$mg\,\cos \,{\theta }_{{\rm{m}}}={F}_{{\rm{f}}}$$.

It is assumed that the top-coal fragments reach the coal drawing opening after a drawing time $$t$$; the moving distance of fragments is expressed as follows: [Image Omitted. See PDF]

The moving distance of fragments directly above the drawing window, that is, $$\theta =0$$, is the longest, with $${S}_{0}=\frac{1}{2}g(1-\,\cos \,{\theta }_{{\rm{m}}}){t}^{2}$$. The moving distance of fragments in any inclination direction $$\theta $$ can be expressed as follows: [Image Omitted. See PDF]

Therefore, using the polar coordinate system, the boundary of the draw body can be expressed by the following equation: [Image Omitted. See PDF]where is the width of the drawing window and $${r}_{\theta }$$ represents the distance from the virtual draw point to the drawing window in the angle of $$\theta $$.

CDEM MODEL AND SIMULATION SCHEME Numerical model

The conducted numerical study is based on the 8222 LTCC panel of Tashan Coal Mine as the engineering background. In this panel, the coal seam is a typical extra-thick seam with an average thickness of 19.93 m. The numerical simulation model (shown in Figure 4) extends 105 m in length and 32 m in height. It was developed according to the field geological condition of the 8222 panels. Research results show that the fragments of the crushed top coal exhibit different block sizes and irregular shapes in the vertical direction, and that the fragmentation degree of the top coal decreases with increasing block size.³⁹ Therefore, a block model, rather than a particle model, can better characterize the nonuniform crushing state of top coal and truly simulate the interaction between fragments. The block model was utilized to simulate the caving and migration characteristics of the crushed top-coal fragments. In the CDEM model, the top coal was divided into six layers for different block meshing; they are listed in Table 1. The block shape adopted mixed polygons, including triangles, quadrilaterals, and pentagons. The block size ranged from 0.05 to 0.35 m and followed a normal distribution.

Figure 4. Numerical model of top coal drawing for the 8222 longwall top coal caving panel

Table 1 Block meshing of the top coal

Layer number	Range of block size (m)	Layer thickness (m)
Ⅰ	0.05–0.15	3
Ⅱ	0.05–0.25	2.5
Ⅲ	0.10–0.20	2.5
Ⅳ	0.15–0.25	2.5
Ⅴ	0.20–0.30	2.5
Ⅵ	0.25–0.35	2.93

The boundary conditions were set as follows. The horizontal displacement of the left and right boundaries of the model was fixed. A small vertical stress of 0.83 MPa was applied to the top boundary of the model, while the vertical displacement was fixed at the bottom with the wall boundary. To eliminate the boundary effect, approximately 21 m were reserved at the left and right sides of the model in the boundary portion, where the top-coal drawing was not implemented. There were 36 drawing windows in the middle of the model, and the width of each single window was 1.75 m. Before the drawing process, gravity was applied and an initial equilibrium was obtained to simulate the initial state of the top-coal body. Then, the caving and migration of the top-coal fragments were simulated with the drawing window opening. The physical and mechanical parameters of the block model used in the initial equilibrium iteration are listed in Table 2. In the drawing process, the mechanical parameters of the model were obtained by the inversion method to ensure that the top coal could be drawn in time.

Table 2 Physical and mechanical parameters of the block model

Material	Density (kg/m3)	Elastic modulus (GPa)	Passion's ratio	Cohesion (MPa)	Tensile strength (MPa)	Friction angle (°)	Thickness (m)
3–5 coal seams	1360	2.43	0.27	8.72	1.22	32.6	19.93
Mudstone	2420	8.27	0.13	12.05	1.71	33	4
2# coal seam	1360	2.70	0.22	8.96	1.27	30	2
Sandy mudstone	2530	11.42	0.21	19.08	2.23	26	2
Sandstone	2580	17.2	0.11	25.0	4.10	34	8

CDEM simulation scheme

To further stimulate the high-yield, high-efficiency mining potential of LTCC mining for extra-thick coal seams, a new SMWD technique is proposed. In this drawing technique, several continuously distributed single drawing windows are combined as an AMW, and act as a combined single drawing window, that is, the AMW becomes open or close synchronously during the drawing process.

We considered five drawing modes listed in Table 3. They were designed through CDEM simulation. The last number of the scheme code indicates the number of single drawing windows combined in the AMW. The A1 scheme corresponds to the traditional single-window drawing mode. The size of the coal drawing window was 1.75 m. In the A3 scheme, three single drawing windows were combined in the AMW, with an opening size of 5.25 m. During the simulation of the drawing process, all combined single window of the AMW was opened and closed simultaneously, that is, the top coal was caving synchronously. When the gangue appeared at any single window, the AMW was closed and the adjacent AMW was opened. Therefore, a single round of sequential drawing technology was performed in the SMWD numerical simulation to compare with the traditional single-window drawing.

Table 3 Top coal drawing scheme in the CDEM simulation

Scheme	Drawing mode	Combination of the drawing window	Width of the AMW (m)
A1	Traditional single-window drawing	1, 2, 3, …, 36	1.75
A2	Synchronous double-window drawing	1–2, 3–4, 5–6, …, 35–36	3.50
A3	Synchronous triple-window drawing	1–3, 4–6, 7–9, …, 34–36	5.25
A4	Synchronous quadri-window drawing	1–4, 5–8, 9–12, …, 33–36	7.00
A5	Synchronous penta-window drawing	1–5, 6–10, 11–15, …, 31–35	8.75

Abbreviations: AMW, assembled multiwindow; CDEM, continuum–discontinuum element method. RESULTS AND DISCUSSION Velocity field of the top-coal drawing Velocity field of the initial drawing

Concerning the initial drawing, the distribution of velocity fields for the different drawing modes is illustrated in Figure 5. Generally, the velocity field is approximately symmetrical along the central line of the drawing window and unevenly distributed in the vertical direction. The migration speed of the top-coal fragments in the area directly above the drawing window was significantly greater than that on both sides. According to the speed magnitude, the migration velocity can be classified into four levels, that is, low, medium, medium-fast, and high speeds. The velocity vector of each speed level is marked in red, orange, yellow, and green, respectively. For different drawing modes, the migration velocity field of the initial drawing is distributed differently. With the increase of width in AMW, the moving range and maximum migration speed of the fragments kept increasing.

Figure 5. Distribution of the velocity field for the initial drawing

Under the A1 drawing mode, that is, traditional single-window drawing, the caving velocity field of the top coal was symmetrical and the migration speed of the fragments near the drawing window was larger than that of the fragments far away from the drawing window. In addition, the migration speed of the fragments directly above the drawing window was greater than that of the fragments on both sides of the drawing window. Owing to the caving of the immediate roof, some blocks exhibited large caving velocity, as shown on the top left in Figure 5A. Furthermore, owing to the small caving channel, only 1.75 m in width, the top-coal fragments were clamped by the squeeze and friction of the fragments, and the top-coal drawing process was greatly limited. Therefore, the migration speed of the top-coal fragments and gangue was slow under single-window drawing. The top-coal migration was insufficient. It easily gave rise to a top-coal arch in the drawing process under A1 drawing mode as the fragments squeezed against each other.

As shown in Figure 5B, under the A2 drawing mode, that is, the synchronous assembled double-window drawing scheme, the drawing process was still limited by the clamping effect caused by squeezing between fragments. However, the clamping effect in the A2 drawing mode was much smaller than in the A1 mode, and both the moving range and maximum migration speed of the top-coal fragments greatly improved. When the number of assembled single drawing windows was greater than two, that is, for the A3, A4, and A5 drawing modes, illustrated in Figure 5C, Figure 5D and Figure 5E, the distribution of the velocity field was clearly different from those of the A1 and A2 drawing modes. Owing to the increased caving channel, the clamping effect was greatly reduced in the top-coal caving process. The caving and migration of the top-coal fragments were sufficient, and the distribution of the velocity field approximately adopted a Y-shape. The migration speed of the top-coal fragments directly above the coal drawing opening reached medium-fast and high speeds. The velocity on both sides of the coal drawing opening was relatively low, and the velocity direction was inclined to the center line of the coal drawing opening.

Velocity field of the subsequent drawing

The distribution of the velocity field of the top-coal fragments under A2 drawing mode during the subsequent drawing process is illustrated in Figure 6. Generally, the migration velocity field of the top-coal fragments was S-shaped and asymmetrical from left to right, and uneven from top to bottom. The migration magnitude in most areas of the upper and middle top-coal fragments was sustained at low speed, while the bottom part of the top-coal fragments caved in at high speed. Owing to the squeezing from the caved gangue in the drawn area, the migration velocity of the fragments on the left side of the drawing opening was relatively small. In these areas, the velocity vector was sparse, and the migration direction was complex. On the right side of the drawing opening, the migration speed was faster compared to the left side, and the velocity direction on the right side was approximately pointing to the central line of the drawing opening. The maximum migration velocity of fragments was located near the drawing opening. However, the velocity direction of the top-coal fragments near the drawing mouth was divergent and not perpendicular to the drawing opening.

Figure 6. Distribution of the velocity field of top-coal fragments under A2 synchronous multiwindow drawing mode

The distribution of the velocity field of top-coal fragments under A3 drawing mode at different drawing sequence are illustrated in Figure 7. Note that the drawing smoothness is significantly improved. In the top-coal drawing space, the migration velocity of 60% of the fragments ranged from medium to fast speed at least. Owing to the increase of the coal caving channel with AMW, the moving areas expanded. The main caving path for top coal fragments was directly above the AMW. The distribution of the velocity field of the top-coal fragments under A4 and A5 drawing modes is illustrated in Figures 8 and 9, respectively. The distribution of the velocity field was asymmetrical and approximately presented a saddle shape revealing clear differences in velocity vector distribution on both sides of the AMW. In the top-coal caving space, the third- and fourth-order velocities of the top-coal fragments dominated and concentrated directly above the AMW. The velocity vector distribution was sparse in the left side of the AMW (the caved and drawn areas) owing to void space in the caving cavity. On the right side of the AMW, the velocity vector of the top-coal fragments was dense, and the flow direction was variable due to the high-speed collision between the fragments. Compared with the velocity field of Scheme A3, the top-coal migration range continued to expand, and the drawing smoothness improved remarkably.

Figure 7. Distribution of the velocity field of top-coal fragments under A3 synchronous multiwindow drawing mode

Figure 8. Distribution of the velocity field of top-coal fragments under A4 synchronous multiwindow drawing mode

Figure 9. Distribution of the velocity field of top-coal fragments under A5 synchronous multiwindow drawing mode

Characteristics of the coal–gangue interface 

The coal–gangue interface can reflect the migration characteristics of both top-coal fragments and the gangue. When the gangue penetrated the coal–gangue interface, it arrived at the draw window earlier than some of the top-coal fragments. Under this condition, the recovered top coal diminished according to the drawing principle of “close the drawing window when gangue appears,” which could subsequently reduce the recovery ratio of top coal. Therefore, the coal–gangue interface is an important caving feature in the top-coal drawing process.

Coal–gangue interface of the initial drawing

The distribution of the coal–gangue interface of the initial drawing under A1 drawing mode is illustrated in Figure 10A. The coal–gangue interface was funnel-shaped, and the center line of the draw cone was not strictly consistent with the center line of the drawing window. The coal–gangue interface was rough and irregular due to the mutual squeeze and restriction of the top-coal fragments. Note also that the movement zone is approximately an ellipsoid body, similar to the shape of an egg. Given that the migration of the fragments was restricted by the squeeze effect in the A1-mode drawing process (see Figure 5A), the caving process of the top-coal fragments propagated slowly upward. Hence, the subsidence of the coal–gangue interface was gentle and uniform. However, some caved rock fragments penetrated the coal–gangue interface and arrived at the drawing window early, leading to the closure of the drawing window under the principle of “stop drawing when gangue appears.” The phenomenon of rock fragment penetration can cause some loss of coal as a result of a drawing process stopped earlier than necessary.

Figure 10. Distribution of coal–gangue interface for the initial drawing

The coal–gangue interfaces of the initial drawing under SMWD mode are shown in Figure 10B–E. Generally, all drawing cones under SMWD mode have a sharp bottom. Along the height direction, the inclination of the coal–gangue interface varied. The variation of the inclination exhibited two stages with the inflexion point located at two-third of the top-coal thickness. The inclination of the interface in the lower and medium parts of the top coal was greater than that of the upper part. In addition, the funnel angle enlarged with the increase in the size of the AMWs. Furthermore, some large caved blocks of the roof significantly influenced the morphology of the coal–gangue interface. Owing to the compression of such large caved blocks, the coal–gangue interface adopted a clear stair shape shown in Figure 10D.

Final coal–gangue interface

The final coal–gangue interfaces under different drawing modes are shown in Figure 11. Note that the final coal–gangue interface was distributed in a wavy shape, and the caved roof blocks are deposited in layers. Given that the principle of “stop drawing when gangue appears” was utilized to control the drawing process, all the SMWD mode witnessed a coal loss mainly due to gangue penetration, as illustrated in Figure 11A to Figure 11D. The lost coal was mainly located in the middle of two adjacent assembled windows. Under the A2 drawing mode, the coal–gangue interface is distributed in complex morphology and the thickness of the lost coal is relatively small. With an increase in the width of the AMW, the coal–gangue interface adopted a tongue shape, and the thickness of the lost coal was relatively greater due to the gangue penetration. In addition, the thickness of the lost coal for the A4 and A5 drawing modes was much larger than that of the A2 and A3 modes. Therefore, the coal loss increased with the size of the AMW.

Figure 11. Distribution of the final coal–gangue interface

Draw body shape Draw body shape of the initial drawing

After the top-coal fragments were drawn, the draw body was obtained by tracking the spatial position of the drawn block at the initial state. The top coal draw bodies under SMWD mode are shown in Figure 12. The draw body of the initial drawing adopted an elliptical shape. With the increase of the width of the AMW, the shape of the draw body changed into a variable ellipse. Given that an irregular block model was adopted in the numerical model, the caving and migration processes of the top-coal fragments became more similar to the usual engineering practice compared with the particle method, and the boundary of the draw body simulated by CDEM was sharp and rough. It is worth noting that the migration of the top-coal fragments was not sufficient because of the limited coal flow channel of the A1 drawing mode, which could lead to an easier arch formation in the drawing process. In Figure 12A, the blank void in the draw body represents the fragments that formed the arch structure in the drawing process. Note that those fragments are mainly distributed in the middle and upper parts of the top coal and presented a relatively large block size. With the increase in the width of the AMW, the smoothness of the top-coal drawing process increased and the arching phenomenon notably decreased. Note also that when the number of single windows in the AMW exceeded three, almost no arch effect took place in the top-coal migration, as shown in Figure 12C to Figure 12E.

Figure 12. Distribution of the draw body shape for the initial drawing

Draw body of the subsequent drawing

To further analyze the dynamic evolution process of the draw body for the subsequent drawing, the morphological distribution of subsequent draw bodies, shown in Figure 13, under different SMWD modes was simulated by CDEM. Generally, the shape of the subsequent draw body adopts a half-crescent shape, the blank part inside the draw body is the deleted stuck top-coal fragment, and the blank part between each draw body is the back coal loss. Note that the drawn volume in the lower part of the top coal was the largest, the drawn volume in the middle part of top coal was moderate, and the upper part of the top coal presented the least drawn amount. The back coal loss became mainly distributed in the middle and upper parts of the top coal, and it increased with the width of the AMW. In addition, the subsequent draw body presented a “large-small” spacing distribution under A2 and A3 drawing modes while the width and height of the draw body kept changing, as shown in Figure 13A and Figure 13B. The increased drawn volume of the previous drawing opening reduced the coal drawing volume of the adjacent drawing opening. With the increase in width of the AMW, the width of the draw body increased and the morphological difference of the caving body decreased. For the A4 and A5 scheme shown in Figure 13C Figure 13D respectively, the development range of the draw body was the largest, whereas the morphology of each draw body was approximately uniform. In addition, with the increase in width of the AMW, the proportion of drawn volume from the middle and lower parts was relatively reduced, while the proportion of drawn volume from the upper part relatively increased. Furthermore, it was observed that the back coal loss mainly developed in the middle and upper parts of the top coal.

Figure 13. Draw body evolution under synchronous assembled multiwindow drawing

Discussion Draw-body boundary analysis

The initial draw-body boundary obtained by CDEM simulation and from the Bergmark–Roos model is presented in Figure 14. Note that the boundary of the draw body simulated by CDEM is not completely consistent with the results predicted by the Bergmark–Roos model. There is a certain deviation between the simulation and predicted results, as illustrated in Figure 14A to Figure 14E. Note also that the lower and middle boundaries of the simulated draw body are wider than the prediction from the Bergmark–Roos model in all drawing schemes, except for the A5 drawing mode, as shown in Figure 14E. The lower and middle boundaries of the simulated draw body of Scheme A5 are consistent with the prediction from the Bergmark–Roos model, while the upper simulation boundary is smaller than the results predicted by the Bergmark–Roos model. The reasons for this deviation can be attributed to the following two aspects. First, the block model was adopted to represent the interaction among top-coal fragments, and both the shape and size of this block were unevenly distributed; when the size of the caving channel of the drawing opening was small, the squeezing effect of the top-coal fragments could easily lead to the arch formation, which could, in turn, hinder the caving of the above top coal and increased the caving tendency for fragments in the lower part of the top coal. The frequent arch formation could result in more drawn volume in the lower moving zone, which caused the horizontal dimensions of the lower part of the draw body to increase. Second, the Bergmark–Roos model is based on the gravity flow theory of granular media. In practice, there is a gap between top-coal fragments and granular media. Given that the block interaction and block size effects cannot be considered properly, the Bergmark–Roos model cannot fully describe the complex migration law of top-coal fragments. Generally, with the increase in the width of the AMW, the CDEM simulation results of the draw-body boundary will become closer to the results predicted by the Bergmark–Roos model.

Figure 14. Comparison between theory prediction and continuum–discontinuum element method simulation of the draw body

Top-coal drawing efficiency and recovery ratio

High-yield, high-efficiency coal production of an LTCC panel depends on a high top-coal recovery ratio and drawing efficiency. Therefore, the recovery ratio and drawing efficiency are crucial for a successful LTCC panel extracting extra-thick coal seam. In the conducted CDEM simulations, the recovery ratio was calculated by dividing the mass of the drawn volume from each AMW by the amount of top coal located directly above the drawn opening. The drawing efficiency was measured by analyzing the top-coal drawn volume per unit iteration time. The recovery ratio and drawing efficiency under different SMWD modes are illustrated in Figure 15. The recovery ratio and drawing efficiency of the A1 drawing mode, that is, the traditional single-window drawing, were 88.12% and 4.50 t/Mstep, respectively. Given that the caving channel was narrow under A1 drawing mode, the squeezing effect of the top coal fragments caused frequent arch formation during the drawing process. On the one hand, the arch structure decreases the drawing efficiency by prolonging the drawing time. On the other hand, the stable arch structure leads to failure to recover the top coal above, which in turn seriously reduces the top-coal recovery ratio. In the CDEM simulation of Scheme A1, arch formation occurred 18 times during the drawing process, which seriously restricted further improvement of the drawing efficiency and top-coal recovery ratio. Therefore, the traditional single-window drawing mode is not an ideal choice for an automated LTCC panel to extract extra-thick coal seam.

Figure 15. Simulated top-coal recovery ratio and drawing efficiency under different drawing modes

Under A2 and A3 drawing modes, the top-coal recovery ratios were 92.43% and 91.20%, respectively, and the drawing efficiency was also improved. The drawing efficiency of the A2 scheme was 33.24 t/Mstep, which is approximately 6.0 times higher than that of the A1 drawing mode. The drawing efficiency of Scheme A3 was 38.32 t/Mstep, which is 7.5 times higher than that of A1 mode. With a further increase in the width of the AMW, the drawing efficiency dramatically increased owing to the increased caving channel. Thus, a large amount of top-coal fragments could be caving synchronously. The simulated drawing efficiency of the A4 and A5 drawing modes was much higher than that of the A3 mode. However, the top-coal recovery ratio dramatically decreased owing to a large amount of coal loss in the back coal located in the upper part of the top coal. The simulated top-coal recovery ratio of the A4 and A5 schemes was approximately 85.27% and 82.32%, which are much lower than those of other drawing modes considered in this study.

Compared with the conventional thick-coal seam extraction, LTCC mining of extra-thick coal seam is characterized by large caving space and long migration distance. The drawing process is more complex, and it is difficult to control the top-coal recovery ratio and drawing efficiency. Owing to the low drawing efficiency, the traditional single-window drawing mode, that is, the A1 drawing mode, can hardly meet the requirements of the high-yield, high-efficiency LTCC panel in the extraction of an extra-thick coal seam. Generally speaking, the drawing efficiency increases with the expanding size of the AMW, while the recovery ratio first increases and then decreases with the increase in the size of the AMW. According to the simulation results, the drawing mode of Schemes A2 or A3 is the optimal choice for automated LTCC panel extracting extra-thick-coal seam by keeping the balance between top-coal recovery ratio and drawing efficiency.

FIELD APPLICATION OF THE SMWD TECHNIQUE Engineering background

The 82222 LTCC panel of the Tashan Coal Mine was chosen for a field test of SMWD. This panel extracts 3–5 coal seams in the Taiyuan Formation of Carboniferous, and the total thickness of the coal seam ranges from 14.47 to 29.21 m. With a standard average thickness of 19.93 m, the 3–5 coal seams is a typical extra-thick coal seam. The average buried depth of the 8222 LTCC panel is approximately 479 m, whereas the designed inclined length and advancing length of the LTCC panel are 230.5 m and 2644.5 m, respectively. The cutting height of the panel is 3.8 m, and hence, the residual top coal is approximately 16 m in thickness. Owing to the large thickness of the top coal, improving the recovery rate of the top coal and caving efficiency is crucial to ensure high-yield, high-efficiency coal production for the 8222 LTCC panel. From the core logging data, the lithology of the LTCC panel is displayed in Figure 16. The immediate roof is mudstone, thin coal seam, and sand-mudstone with a total thickness of approximately 7.41 m. The main roof is sandstone with a thickness of approximately 8.3 m.

Figure 16. Lithology of the longwall top coal caving panel

Field test of the SMWD technique

In the conducted field test of the SMWD technique, an automatic top-coal drawing control system, shown in Figure 17, was developed based on an existing electrohydraulic control system. The software of this system can independently realize automatic control of the opening and closing of the drawing window for hydraulic support. To facilitate the collection of field statistics of the top-coal drawn volume and drawing time, 30 hydraulic supports, numbered from 21 to 50, near the main gate were selected as the testing area, as shown in Figure 18. The single hydraulic supports in the testing area were united and grouped to form an assembled synchronous multidrawing window. In the field test, each SMWD mode was tested three times with the advancement of the LTCC working face. When the top-coal drawing operation arrived at the test area, the coal cutting by the shearer was suspended, and the drawing operation in other areas was also stopped.

Figure 17. Automatic top-coal drawing system

Figure 18. Field test of the synchronous multiwindow drawing mode

Then, the SMWD mode was implemented and tested in sequence, and the AMW was closed under the principle of “stop drawing when gangue appears.” During the test, the weight of the draw volume was measured by the belt weigher, and the drawing time was recorded automatically by the automatic top-coal drawing control software. The thickness of top coal in the test area was predetected by geological radar, and the total amount of top coal in the test area was then calculated. Therefore, the recovery ratio in the test area was the actual amount of drawn coal divided by the total amount of top coal. The coal drawing efficiency is represented by the ratio of the total amount of drawn coal in the test area to the total drawing time. Owing to the limited transportation capacity, the rear scraper conveyor was shut down as a result of overload under the A4 and A5 drawing modes. Therefore, only the A1, A2, and A3 drawing modes were successfully tested.

The top-coal recovery ratio and drawing efficiency resulting from the field test are listed in Table 4. Under A1 drawing mode, that is, the traditional single-window drawing, the total amount of drawn coal was 791.21 t, and the average top coal recovery ratio was 81.2%. The total coal drawing time was 4257 s, with an average of 141.9 s, as illustrated in Figure 19. Hence, the actual draw efficiency was 11.15 t/min. Under A1 drawing mode, the coal caving was frequently interrupted and the drawing window was easily blocked by large blocks from stone bands in the coal seam during the drawing process. Under A2 drawing mode, the total amount of recovered top coal was 863.32 t, and the average top coal recovery ratio was 88.6%. Compared with the A1 drawing mode, the recovery ratio was improved by 7.4 percentage points. The average drawing time was 162.6 s for the A2 mode, and the draw efficiency was 21.24 t/min, which is approximately 1.9 times that of the A1 drawing mode. The average top-coal recovery ratio was 86.3% under the A3 drawing mode, and the draw efficiency was 27.50 t/min. The top-coal recovery ratio slightly decreased with respect to the A2 drawing mode, while the draw efficiency improved notably.

Table 4 Top-coal recovery ratio and drawing efficiency under different drawing modes

Drawing mode	Size of assembled multiwindows (m)	Average draw weight (t)	Average drawing time (s)	Recovery ratio (%)	Draw efficiency (t/min)
A1	1.75	26.37	141.9	81.2	11.15
A2	3.5	28.78	162.6	88.6	21.24
A3	5.25	28.03	183.5	86.3	27.50
A4	7	/	/	/	/
A5	8.75	/	/	/	/

Figure 19. Drawing time under different drawing mode in the testing zone

CONCLUSIONS

In the present study, the characteristics of caving and migration for top-coal fragments under SMWD mode were simulated using the CDEM. The variation of top-coal recovery ratio and drawing efficiency in different drawing modes was analyzed. The conclusions are as follows:

• (1).

  The size of the AMW has a remarkable influence on the distribution of the velocity field of the top-coal fragments under SMWD mode. The area directly above the AMW is the main caving channel for top-coal fragments, and the drawing smoothness increases with the width of the AMW. Compared with the traditional single-window drawing mode, the SMWD mode can effectively improve the smoothness of top-coal caving, given that the squeezing effect is significantly mitigated for a larger caving channel.

• (2).

  In the initial drawing, the angle of the draw cone increases with the width of the AMW under the SMWD mode. The draw body adopts an elliptic shape, which changes into a variation ellipse with the increase in width of the AMW. Given that an irregular block model was adopted in the numerical simulation, the boundary of the draw body simulated by CDEM was sharp and rough and presented some deviation from the results predicted by the Bergmark–Roos model. The subsequent draw body was a half-crescent shape. With the increase in the size of the AMW, the proportion of the drawn volume from the middle and lower parts relatively reduces, while the proportion of the drawn volume from the upper part relatively increases. The back coal loss mainly develops in the middle and upper parts of the top coal.

• (3).

  The simulated recovery ratio under different drawing modes ranged from 82.3% to 92.4%, and the drawing efficiency ranged from 4.49 to 84.32 t/million iteration steps. The drawing efficiency increases with the expansion of the width of the AMW, while the recovery ratio first increases and then decreases with the increase in width of the AMW.

• (4).

  The conducted field test shows that it is easier to form an arch structure and has a low drawing efficiency under the traditional single-window drawing mode. By contrast, the SMWD mode can greatly improve the drawing efficiency. Based on the field test and simulation results, it is recommended to adopt A2 and A3 SMWD modes to further improve the top-coal recovery ratio and caving efficiency, thereby ensuring high-yield, high-efficiency production for an LTCC panel extracting extra-thick coal seam.

ACKNOWLEDGMENTS

This study was supported by the National Key R&D Program of China (Grant Nos. 2018YFC0604501 and 2018YFC0604506).

CONFLICT OF INTEREST

The authors declare no conflict of interest.