Numerical Study of the Respirable Coal Dust Removal Performance of a Vortex Ventilation System at an Excavation Face

Coal dust is one of the most serious issues in coal mines. The miners at an excavation face have a high potential for exposure to high concentrations of coal dust. The main method for removing coal dust is by ventilation. To better control the coal dust at the excavation face, it is important to understand the coal dust movement patterns and distribution characteristics. This paper used the computational fluid dynamics (CFD) approach to simulate coal dust movement under two different ventilation systems. The Eulerian-Lagrange method was used to study the gas-solid two-phase fluid behavior. The coal dust concentration distributions under the two ventilation systems are presented and further compared. From the comparison results, it is found that a vortex ventilation system has better coal dust removal performance. This ventilation system is recommended for excavation for better coal dust control and miners’ health protection.


Introduction
Coal is an important energy source around the world. With the development of total mechanization in the mining industry, coal production has increased rapidly in recent years. Large amounts of coal dust have also been generated during mining activities. Coal dust not only has the potential to affect the performance of mining equipment, but also creates safety issues for the mining industry. In addition, most of the dust is respirable, so it could penetrate the human respiratory system and enter the lungs. Studies have shown that long-term exposure to high concentrations of coal dust can lead to pneumoconiosis and lung cancers [1,2]. Such lung diseases contribute to a high mortality rate in miners [3][4][5]. It has been reported that more than 80% of all coal dust is generated at the excavation face during the excavation activity [6]. The miners who work in this area have a high potential for exposure to high concentrations of coal dust. In order to provide a safe working environment for miners, it is necessary to control the coal dust effectively at the excavation face.
Currently, there are three main methods to control coal dust at the excavation face. The first one is ventilation, which is also the most popular method. The other two methods are water spraying and foam technology [7,8]. However, the control performance is limited by the single usage of each method. For example, water spraying requires a large amount of water and the high pressure may system. Both the distribution of total dust size and respirable size are analyzed in this work. The obtained simulation data under two ventilation systems are further compared. The results showed that the vortex ventilation system performed better in term of coal dust removal.

Physical Model
The physical model built in this paper is based on an underground excavation face in Dongtan Coal Mine, located in Shandong Province, China. The length of the working face is about 40 m and the cross-sectional dimensions are 4 m (width) × 3 m (height), as shown in Figure 1. For the long blowing and short suction ventilation system, the long blowing duct is close to the right-side wall and around 0.2 m away from the roof; the short suction duct is located in the middle of the working face, 2 m and 1.5 m away from the right side wall and roof, respectively. The diameter of both the blowing duct and suction is 0.5 m. The meshes for the model are generated by tetrahedral meshes, as shown in Figure 2. The number of volumes for the mesh is around 0.2 million. The mesh independence study has been conducted in previous research. The same size of mesh is used in this study for the simulation [18]. The total element quality for the mesh is above 0.4; all the mesh has passed the Gambit grid check utility, which is acceptable for most of the solvers in Fluent.

Mathematical Model
The Eulerian-Lagrangian approach is used to simulate the gas-solid coupled flows in this paper. In this method, the airflow is first simulated to steady-state by the Eulerian method, then the airflow field is frozen before emitting the particles. After tis the Lagrangian approach is applied to calculate the motion of dust particles.
The gas phase is considered incompressible and is governed by Navier-Stokes Equation [19]. The continuity and momentum equations are given by: where is the air velocity, is the air density, is the static pressure, is the gravity acceleration, is the molecular viscosity, is the turbulent viscosity, k is the turbulent kinetic energy and is the unit tensor.
The motion of dust particles is governed by the Lagrangian method [20], which is calculated by Newton's second law. The one-way coupling method (only considering the effect of air flow on the particles' motion) is selected as the particles' volume fraction is less than 10 −6 [21]. With a low particle volume fraction, the interaction between particles is less significant compared to other forces, and thus could be ignored in this study. In the airflow field, many forces act on the motion of particles, like gravity, lift force, buoyancy force, drag force, pressure gradient force, etc. However, only gravity, drag force, and buoyancy force are considered because other forces do not affect the particles significantly. The equation for the particles' motion is given by: where is the particle mass, is the particle velocity, is the buoyancy force, and is the drag force. The drag force is given by: where is the particle diameter and is the drag force coefficient, which is given by: Meshes for the physical model.

Mathematical Model
The Eulerian-Lagrangian approach is used to simulate the gas-solid coupled flows in this paper. In this method, the airflow is first simulated to steady-state by the Eulerian method, then the airflow field is frozen before emitting the particles. After tis the Lagrangian approach is applied to calculate the motion of dust particles.
The gas phase is considered incompressible and is governed by Navier-Stokes Equation [19]. The continuity and momentum equations are given by: where u g is the air velocity, ρ g is the air density, p is the static pressure, g is the gravity acceleration, µ is the molecular viscosity, µ t is the turbulent viscosity, k is the turbulent kinetic energy and I is the unit tensor. The motion of dust particles is governed by the Lagrangian method [20], which is calculated by Newton's second law. The one-way coupling method (only considering the effect of air flow on the particles' motion) is selected as the particles' volume fraction is less than 10 −6 [21]. With a low particle volume fraction, the interaction between particles is less significant compared to other forces, and thus could be ignored in this study. In the airflow field, many forces act on the motion of particles, like gravity, lift force, buoyancy force, drag force, pressure gradient force, etc. However, only gravity, drag force, and buoyancy force are considered because other forces do not affect the particles significantly. The equation for the particles' motion is given by: where m p is the particle mass, u p is the particle velocity, F b is the buoyancy force, and is the F d drag force. The drag force is given by: where d p is the particle diameter and C d is the drag force coefficient, which is given by: where Re is the Reynolds number. In addition, the particle diffusion effect of turbulence is also accounted for.

Boundary Conditions and Numerical Schemes
ANSYS Fluent was used for all the simulations in this paper. The density and viscosity of the air flow are 1.225 kg/m 3 and 1.7894 × 10 −5 Pa·s, respectively. The gravity acceleration is 9.8 m/s 2 . No heat transfer considered during the simulation. The air flow field was first simulated to steady-state by a standard k-ε turbulent model. The pressure velocity coupled equations were solved by a SIMPLEC algorithm. The numerical schemes for the solver and the simulation boundary conditions are given in Tables 1-3. It is assumed that the particles are stored at the heading face. The size distribution obeys Rosin-Rammler distribution. The dust size distributions for total dust and respirable dust are presented in Table 4. Rosin-Rammler distribution assumes a relationship between the particle's diameter d and the particle's mass fraction Y d , which is given by [22]: where d is the mean diameter for all particles, n is the spread parameter. In this study, the d total dust and d respirable dust is 59 µm and 4.4 µm, respectively. The spread parameter for total dust and respirable dust is 2.42 and 3.47, respectively. The mass flow rate for total dust and respirable dust is 0.0044 kg/s and 0.0034 kg/s, respectively. The density of coal dust used in this study is 1000 kg/m 3 .  The wall collision boundary conditions are summarized in Table 5. Three wall collision boundary conditions are used in this simulation. Reflecting boundary condition is used for the tunnel wall, roof and the duct wall. Trap boundary condition is used for the tunnel floor, which means the calculation would stop after the particle was trapped on the floor. Escape boundary condition is used for the duct inlet and outlet, and the roadway inlet outlet.

Airflow Field Distribution
The coal dust distributions are significantly affected by the airflow behaviors, thus it is important to understand the airflow field velocity and vortex distributions of different ventilation systems. be seen in Figures 3 and 4, the rebounded airflow still had a relatively high velocity and most of the airflow moved to the non-duct side of the wall. Then the airflow velocity gradually decreased. It is noted that a large vortex area was generated between the heading face and suction duct due to the combined effects of the fresh air from the blowing ducts and the reversed airflow. The air velocity in this area was quite low, about 0.3 m/s. The center of the vortex is about 2 m, 1 m, and 1.2 m away from the heading face, floor, and suction duct head, respectively. Due to the recirculating airflow in the vortex, coal dust may accumulate in this area. Due to the negative pressure near the suction duct head, a high-velocity zone was generated in this area. The velocity around this area decreased rapidly far from the duct head, as shown in Figures 5 and 6.          Figure 9. The velocity near the vortex center was around 0.28 m/s. With such a velocity, the airflow was not able to blow the dropped dust to the bottom again. As shown in Figures 10 and 11, the airflow velocity direction changed about 5 m from the heading face. When the distance was less than 5 m, most of the airflow went towards the suction duct head. When the distance was greater than 5 m, the airflow went towards the outlet of the working face. This helps to control the coal dust in the 5 m from the heading face and protects the operator of the excavator from the coal dust. On the other hand, such airflow distribution also contributes to the high efficiency of coal dust collection.   Figure 9. The velocity near the vortex center was around 0.28 m/s. With such a velocity, the airflow was not able to blow the dropped dust to the bottom again. As shown in Figures 10 and 11, the airflow velocity direction changed about 5 m from the heading face. When the distance was less than 5 m, most of the airflow went towards the suction duct head. When the distance was greater than 5 m, the airflow went towards the outlet of the working face. This helps to control the coal dust in the 5 m from the heading face and protects the operator of the excavator from the coal dust. On the other hand, such airflow distribution also contributes to the high efficiency of coal dust collection.  Figure 9. The velocity near the vortex center was around 0.28 m/s. With such a velocity, the airflow was not able to blow the dropped dust to the bottom again. As shown in Figures 10 and 11, the airflow velocity direction changed about 5 m from the heading face. When the distance was less than 5 m, most of the airflow went towards the suction duct head. When the distance was greater than 5 m, the airflow went towards the outlet of the working face. This helps to control the coal dust in the 5 m from the heading face and protects the operator of the excavator from the coal dust. On the other hand, such airflow distribution also contributes to the high efficiency of coal dust collection.          Figures 12 and 13 show the coal dust residence time under two different ventilation systems. It is obvious from the figures that most of the particles accumulated at the front of the heading face. Under the long blowing and short suction ventilation system, although most particles were carried to the suction duct and captured by the dust collector, there were still a great number of particles transported to the middle section of the heading face with the airflow. This phenomenon is more significant for the respirable dust because the respirable dust has a relatively small diameter. Such small particles are more easily suspended in the air and thus be affected by the airflow. Thus, the miners who work in this area should be given protective gear to avoid exposing them to high concentrations of coal dust. However, this situation does not exist under the vortex ventilation. Most of the particles were sucked into the suction duct and few particles moved to the middle section of the working face. It seems the vortex ventilation system has better coal dust removal performance.     Figures 12 and 13 show the coal dust residence time under two different ventilation systems. It is obvious from the figures that most of the particles accumulated at the front of the heading face. Under the long blowing and short suction ventilation system, although most particles were carried to the suction duct and captured by the dust collector, there were still a great number of particles transported to the middle section of the heading face with the airflow. This phenomenon is more significant for the respirable dust because the respirable dust has a relatively small diameter. Such small particles are more easily suspended in the air and thus be affected by the airflow. Thus, the miners who work in this area should be given protective gear to avoid exposing them to high concentrations of coal dust. However, this situation does not exist under the vortex ventilation. Most of the particles were sucked into the suction duct and few particles moved to the middle section of the working face. It seems the vortex ventilation system has better coal dust removal performance.  Figures 12 and 13 show the coal dust residence time under two different ventilation systems. It is obvious from the figures that most of the particles accumulated at the front of the heading face. Under the long blowing and short suction ventilation system, although most particles were carried to the suction duct and captured by the dust collector, there were still a great number of particles transported to the middle section of the heading face with the airflow. This phenomenon is more significant for the respirable dust because the respirable dust has a relatively small diameter. Such small particles are more easily suspended in the air and thus be affected by the airflow. Thus, the miners who work in this area should be given protective gear to avoid exposing them to high concentrations of coal dust. However, this situation does not exist under the vortex ventilation. Most of the particles were sucked into the suction duct and few particles moved to the middle section of the working face. It seems the vortex ventilation system has better coal dust removal performance.

Coal Dust Distribution under Two Ventilation Systems
The total coal dust and respirable coal dust concentration distribution are illustrated in Figures 14  and 15. As can be seen, the high coal dust concentration area formed at the front of the heading face due to the existing vortex area under the long blowing and short suction ventilation system. This area expanded to the further distance with airflow towards the downstream. This phenomenon is even more obvious for the respirable dust because a respirable high dust concentration zone was generated behind the suction duct head. It is worth noting that almost no respirable dust accumulated at the center of the vortex. This is due to the large air velocity at the front of the duct head; most of the particles at this area went into the duct with the airflow. Because of the high dust concentration area at the back of the suction duct head, it is believed that the small particles (respirable dust) are more easily affected by the airflow. In addition, the highest dust concentration zone occurred at the right side of the heading face. This is caused by two things. Firstly, a low-velocity zone was generated at this area. Secondly, after hitting the heading face, much of the dust dropped in this area due to gravity. The high dust concentration area under the vortex concentration is much smaller than that of the long blowing and short suction ventilation system. However, the average concentration in this area is higher than that of the long-short ventilation system. For vortex ventilation, most particles accumulated at the front of the suction duct head due to the vortex airflow. This generated a higher dust concentration area. It is noticed that the dust concentrations behind the suction head were quite low. The total coal dust and respirable coal dust concentration distribution are illustrated in Figures  14 and 15. As can be seen, the high coal dust concentration area formed at the front of the heading face due to the existing vortex area under the long blowing and short suction ventilation system. This area expanded to the further distance with airflow towards the downstream. This phenomenon is even more obvious for the respirable dust because a respirable high dust concentration zone was generated behind the suction duct head. It is worth noting that almost no respirable dust accumulated at the center of the vortex. This is due to the large air velocity at the front of the duct head; most of the particles at this area went into the duct with the airflow. Because of the high dust concentration area at the back of the suction duct head, it is believed that the small particles (respirable dust) are more easily affected by the airflow. In addition, the highest dust concentration zone occurred at the right side of the heading face. This is caused by two things. Firstly, a low-velocity zone was generated at this area. Secondly, after hitting the heading face, much of the dust dropped in this area due to gravity. The high dust concentration area under the vortex concentration is much smaller than that of the long blowing and short suction ventilation system. However, the average concentration in this area is higher than that of the long-short ventilation system. For vortex ventilation, most particles accumulated at the front of the suction duct head due to the vortex airflow. This generated a higher dust concentration area. It is noticed that the dust concentrations behind the suction head were quite low.

Dust Concentration Comparison between Two Ventilation Systems
The average dust concentration at different cross sections for two ventilation systems are presented in Figures 16 and 17. As can be seen, the dust concentration at different positions of the working face is close for total dust and respirable dust. For a long blowing and short suction ventilation system, the highest dust concentration occurs near the heading face; the concentration dropped sharply to around 580 mg/m 3 and 240 mg/m 3 , respectively, for total dust and respirable dust. Then the concentration arrived at another peak point 10 m away from the heading face for the total dust and 13 m for the respirable dust. After that, the dust concentration dropped gradually with the particles moving downstream. It is worth noting that the total dust concentration is almost double that of the respirable dust concentration under this ventilation system, as shown in Figure 16. For a vortex ventilation system, the highest dust concentration for both total dust and respirable dust also occurs at the front of the heading face, then the concentration decreases rapidly to less than 50 mg/m 3 at 5 m, where the suction duct inlet is also located. After that, the concentration keeps stable after 6

Dust Concentration Comparison between Two Ventilation Systems
The average dust concentration at different cross sections for two ventilation systems are presented in Figures 16 and 17. As can be seen, the dust concentration at different positions of the working face is close for total dust and respirable dust. For a long blowing and short suction ventilation system, the highest dust concentration occurs near the heading face; the concentration dropped sharply to around 580 mg/m 3 and 240 mg/m 3 , respectively, for total dust and respirable dust. Then the concentration arrived at another peak point 10 m away from the heading face for the total dust and 13 m for the respirable dust. After that, the dust concentration dropped gradually with the particles moving downstream. It is worth noting that the total dust concentration is almost double that of the respirable dust concentration under this ventilation system, as shown in Figure 16. For a vortex ventilation system, the highest dust concentration for both total dust and respirable dust also occurs at the front of the heading face, then the concentration decreases rapidly to less than 50 mg/m 3 at 5 m, where the suction duct inlet is also located. After that, the concentration keeps stable after 6 m. It is clearly seen that in the 5 m away from the heading face, the dust concentration of a vortex ventilation system is higher than that of a long-short ventilation system, but this concentration is much lower after 5 m for vortex ventilation system. This means that the vortex ventilation system could better keep the particles at the front of the suction duct and moved them to the suction duct.
To better analyze the dust removal efficiency of two ventilation systems, the histograms of dust concentration distribution at the working face outlet for the two systems are compared in Figure 18. As can be seen, under the long-short ventilation system, most of the total dust concentration is allocated between 10 and 30 mg/m 3 . The concentrations of respirable dust mainly fall in the range of 30 to 130 mg/m 3 . By comparison, most of the concentrations of the total dust and respirable dust are below 10 mg/m 3 under the vortex ventilation system. Based on the onsite data, the total dust concentration and respirable concentration at the outlet of the working face under the long-short ventilation system are 16.4 mg/m 3 and 58.8 mg/m 3 , respectively. However, these values for the vortex ventilation system are 4.1 mg/m 3 and 4.1 mg/m 3 , respectively. This indicates that most particles went into the suction duct and were collected by the dust collector for the vortex ventilation system. Thus, the vortex ventilation system has better dust control performance. m. It is clearly seen that in the 5 m away from the heading face, the dust concentration of a vortex ventilation system is higher than that of a long-short ventilation system, but this concentration is much lower after 5 m for vortex ventilation system. This means that the vortex ventilation system could better keep the particles at the front of the suction duct and moved them to the suction duct.
To better analyze the dust removal efficiency of two ventilation systems, the histograms of dust concentration distribution at the working face outlet for the two systems are compared in Figure 18. As can be seen, under the long-short ventilation system, most of the total dust concentration is allocated between 10 and 30 mg/m 3 . The concentrations of respirable dust mainly fall in the range of 30 to 130 mg/m 3 . By comparison, most of the concentrations of the total dust and respirable dust are below 10 mg/m 3 under the vortex ventilation system. Based on the onsite data, the total dust concentration and respirable concentration at the outlet of the working face under the long-short ventilation system are 16.4 mg/m 3 and 58.8 mg/m 3 , respectively. However, these values for the vortex ventilation system are 4.1 mg/m 3 and 4.1 mg/m 3 , respectively. This indicates that most particles went into the suction duct and were collected by the dust collector for the vortex ventilation system. Thus, the vortex ventilation system has better dust control performance.   (c) (d) Figure 18. Dust concentration distribution histogram at the working face outlet: (a) total dust for longshort ventilation system; (b) respirable dust for long-short ventilation system; (c) total dust for vortex ventilation system; (d) respirable dust for vortex ventilation system. Figure 18. Dust concentration distribution histogram at the working face outlet: (a) total dust for long-short ventilation system; (b) respirable dust for long-short ventilation system; (c) total dust for vortex ventilation system; (d) respirable dust for vortex ventilation system.

Limitations and Perspectives
This study intends to evaluate a new ventilation system's dust removal performance. However, there are still some limitations. The numerical simulation model in the manuscript is based on a low-gas coal mine and the gas parameters are not considered in the simulation calculation. However, in a field situation, there are many high-gas coal seams and coal and gas outburst seams in China. In these mines, the balance between gas concentration and ventilation parameters should be considered in the design of ventilation and dust control systems. Numerical simulation coupled with dust and gas is an important part of our future research work.
In addition, the design of a rotating annular disk is the key to vortex ventilation. A rotating annular disk with different shapes and design parameters can have varied dust control performance, especially for respirable dust. It is also necessary to study the optimum design of a rotating annular disk by modeling and computing the dust control effect and dust distribution law with different design schemes. A further study will be conducted to resolve all of the above limitations and provide a more comprehensive ventilation system.

Conclusions
A CFD simulation study has been conducted based on an on-site excavation in this paper. Two ventilation systems, a long blowing and short suction ventilation system and a vortex ventilation system, have been modeled to evaluate their performance on coal dust removal. The Eulerian-Lagrange approach was used to analyze the coal dust movement pattern in the airflow field. The coal dust distribution and concentration data were obtained from the simulation results under two ventilation systems. For both ventilation systems, dust accumulated between the heading face and suction duct head due to the combined effects of the excavation activity and the large negative pressure near the suction duct head. In this area, it was noted that both the total dust and respirable dust concentrations under the vortex ventilation system were higher than those of the long-short ventilation system. For the long-short ventilation system the high dust concentration area expanded to the middle area of the working face, whereas this did not occur under the vortex ventilation system. It is believed that the vortex ventilation system has a stronger dust capture ability. Under this ventilation system, most of the dust was captured into the suction duct by the generated vortex airflow before traveling downstream. The dust concentration at the outlets for the vortex ventilation system were also much lower than those of the long-short ventilation system, which indicates that the vortex ventilation system has the capacity to remove coal dust efficiently. To better control coal dust and protect miners' health, the vortex ventilation system is recommended for excavation.