Design optimization of a coal seam annular air reverse circulation bit

Reverse circulation (RC) drilling technology enables the rapid acquisition of coal seam samples, which is crucial in accurately measuring gas content. In addition, the design of the drill bit plays a key role in RC sampling. In this study, the turbulence model and simulation method applied during the RC drilling process is analyzed, and an annular jet flow bit is developed. An experimental laboratory setup is used to simulate the RC performance of the proposed drill bit, based on which an orthogonal design simulation of the structural parameters of the drilling bit studies was performed to investigate the effects of the RC drilling process. The experimental results show that annular jet flow moving up the central channel wall of the bit is conducive to the suction of annular airflow and results in the strong flow field created by the center channel of the bit generating negative pressure in the center channel of the bit, a comparison of the computational fluid dynamics data on negative pressure and suction rate shows that the annular jet flow RC bit is suitable for use in RC drilling.


| INTRODUCTION
Coal seam gas content is an important index for mine gas control, for which get coal samples must be obtained quickly. 1 Air reverse technology is the most common sampling method and uses uninterrupted drilling for sampling. 2,3 The effectiveness of air reverse circulation (RC) sampling technology depends on whether air flushing can exhaust the coal sample into the center channel of the drill. Here, the structure of the RC drill bit plays a crucial role in RC sampling, especially. 4,5 In recent years, studies have highlighted the potential of bottom structures and spatial layouts of RC bits in improving the effectiveness of RC sampling. For instance, in one study, a flushing nozzle and pressurerestoring groove was added to an RC drill bit systematically to reduce its energy consumption and improve RC drilling performance. 6 Another study introduced the principle of the ejector from the spatial layout of an RC bit, which contributed to the improvement of RC performance. [7][8][9] Another reported that negative pressure is produced at the bottom of the bit when compressed air is ejected from the suction nozzles at high speed, suctioning the air from the bottom into the center channel of the double-wall drill pipe. 10 Zhao et al. 11 and Yin et al. 12 analyzed the effects of the number, diameter, and inclination angles of suction nozzles on RC drill bits using numerical simulation methods. Ren et al. 13 studied the influence of suction nozzles on the ability of an RC drill bit, experimentally assessed the influence of suction nozzles on the ability of an RC drill bit to obtain sampling. Cao et al. 6 observed that the suction effect of a drill bit could be improved by designing a swirling generator. Han et al. 14,15 found that a reasonable air velocity was conducive to discharge coal cuttings in the study of coal seam RC. Finally, Yang et al. 16 studied the effect of bit structure on RC drilling performance.
Despite these efforts, air RC bits continue to be far from optimal, owing to their low sampling efficiency in coal seams. In this study, a bit with an annular slit inside was developed using computational fluid dynamics (CFD), which was employed to simulate the flow phenomena of the bit. In addition, a series of tests was performed to determine the optimal parameter configuration.

| DESCRIPTION OF THE AIR RC TECHNOLOGY
An air RC sampling system comprises a rotary drilling rig, dual-wall drill pipes, flexible air input tube, doublepassage swivel, flexible discharging tube, gooseneck tube, specially designed bit, and other drilling tools in Figure 1. In the conventional drilling, the compressed gas flows downward through the bit along the center of the drill pipe, while coal samples and drill cuttings are delivered to the surface in the annulus between the drill pipe and the borehole. However, the mixing of coal samples is unavoidable, especially in coal seams; thus, the samples collected by this drilling method are always mixed with nontarget coal samples. For an air RC sampling system, the compressed air flows through the flexible air input tube into the annular gap of the double-wall drill pipe. When it reaches the bottom of the borehole, the coal sample at the bottom of the hole is sucked upwards. The compressed air and coal sample enters the central channel instead of the annulus between the drill pipe and the borehole, which returns to the surface along the central channel of the double-wall drill pipe. A crosssection of the RC drill bit is shown in Figure 2, where the arrows represent the circulating medium flow direction. The red arrows represent the flow direction of the circulating medium that flows down the annular channel of the double-wall drill pipe and carries suction gas into the inner pipe on reaching the bottom; the blue arrows represent the flow direction of the circulation medium and suction gas mixture; and the pink arrows represent the suction gas flow direction. The advantage of the air RC drilling in coal seam sampling is that the compressed air returns to the ground along the center channel of the dual-wall drill pipe, 17,18 which improves the suction capacity of the drilling process.
According to the multi-nozzle ejector principle, 19 the bit designed for RC drilling systems comprises a built-in ejector and bit body, as shown in Figure 2. The high-pressure gas flows into an annular gap, which is formed by the inner wall of the bit body and the outer wall of the built-in ejector, and enters the central channel through suction nozzles at high speed, which leads to a pressure reduction at the bottom of the bit and forms a negative pressure zone. In this manner, the gas at the bottom of the borehole is sucked into the central channel of the bit; therefore, the negative pressure at the bottom of the borehole is one of the core indices for evaluating RC bit performance. Other  streams of air that enter the flushing nozzles are ejected through the pressure-restoring groove, purging the gas into the central channel of the bit. Eventually, the air return to the ground after mixing the gas inside the bit.
Generally, the perturbative effect of the working fluid of the built-in ejector on the migration path of the particles is minimized to reduce the number of collisions and the amount of fragmentation between particles. In this manner, the integrity of the coal sample particles can be maintained during sampling. The index for evaluating the performance of the bit is the value of the negative pressure, which is directly related to the structure of the built-in ejector. Therefore, in this study, the RC drilling bit was designed as illustrated in Figure 2; the built-in ejector comprises a nozzle, suction section (gas receiving section), contractile segment, and throat (air and gas mixing section). The air flowing through the suction nozzles is known as the working fluid, the gas entering the contractile segment is called the ejected fluid, and the gas flowing out of the throat is called the mixed fluid.
The main structural parameters of the bit for RC coal seam sampling are shown in Figure 2. The diameter of the bit is 212 mm, and five flushing nozzles with diameters (denoted by d 2 ) of 10 mm are evenly distributed within it. The suction section diameter (D 1 ) is 50 mm, and its length (L 1 ) is 145 mm. The contractile segment length (L 2 ) is 42 mm, and the expansion angle (α) is 23°. Suction nozzles with diameters of d 1 are designed to be located along the circumference of the built-in ejector. The tilt angle (β) is set to 30°. The length of the throat (L 3 ) is 125 mm.

| Computational domain and grid
For a drill bit diameter of 212 mm, the borehole formed had an aperture of approximately. Therefore, to simplify the proposed analysis of the simulation, the borehole was represented by a cylinder with 222 mm aperture. In addition, the influence of cutter teeth and bit rotation on the flow field were ignored for the sake of simplicity.
As the working fluid and ejected fluid mix and exchange energy through the suction nozzles, a large turbulent shear force is produced. Here, the bit and borehole walls significantly influence the turbulent calculations. Therefore, an expansion boundary layer was set in the computational fluid domain grid, and the suction nozzle grids were encrypted to improve calculation accuracy.

| Computational domain
The structure of the RC bit designed in the SOLID-WORKS software was imported into the ANSYS Geometry Design Modeler to generate the computational fluid domain, which was constructed by creating the fluid edge surface and filling. The fluid domain is shown in Figure 3. To simplify the simulation analysis, the face was defined and divided into different types of working faces, including walls, inlet, central outlet, annular outlet, suction nozzles, and flushing nozzles.

| Surface meshing
The structural model of the fluid domain shows that the mesh size was divided into four local dimensions: walls, inlet and outlet, flushing nozzles, and suction nozzles; the corresponding local mesh sizes were 5, 4, 2, and 1 mm, respectively. Considering that the fluid domain comprises nozzles and holes, the growth rate was set to the value of 1.2. For the structure of the bit, the curvature and proximity grid generation method for a complex structure model was adopted. The curvature was used to determine the mesh size of a curved surface, and proximity was used to control the adjacent mesh size. This approach was adopted because combining curvature and proximity can provide a better fit for a complex structure.
For hexahedral, triangular, and quadrilateral meshes, the threshold value of the slope deviation improved by the collapse method was lower than 0.8, and the maximum slope deviation of the surface mesh for the drilling system designed in this study was 0.795. The entry was defined as a mass flow inlet and the outlets were defined as pressure outlets. For the entrance, the mass flow rate was set as the parameter instead of the volume flow rate because the mass flow F I G U R E 4 Computational grids for the reverse circulation bit. rate cannot be influenced by certain pressure and temperature changes at the bottom of the wall. In addition, the inlet boundary type that uses the mass flow rate is more reliable for assessing the gravitational field.

| Volume meshing
The volume mesh was divided into a structured mesh and an unstructured mesh, which comprised a series of quadrilateral and hexahedral meshes. The unstructured mesh was composed of triangular and tetrahedral meshes, which compensated for the mesh overlap caused by the irregular structural shape and met the requirements of the complex solution domain.
Owing to the complex structure of the RC bit fluid field, the poly-Hexcore volume mesh generation method was selected, which can produce a hexahedron and polyhedron co-node connection (no overlapping surfaces) and does not require any additional manual mesh setting. The number of hexahedrons was increased in the mesh using this method, which improves the efficiency and accuracy of the solution.
The boundary layer was set as the aspect ratio mode for volume mesh generation with five layers, and the initial aspect ratio was set to 10. Finally, as shown in Figure 4, the poly-hexcore volume mesh was formed by combining the previously generated surface meshes, which has the advantage of calculation accuracy for forming quickly and dividing finely.

| Governing equations
Air with a constant density was used as the research object in this experiment, and its governing equations are the standard k-epsilon model (k-ε) governing equation system 20 : where ϕ denotes the dependent variable, υ denotes the velocity vector, Γ ϕ denotes the diffusion coefficient, and S ϕ is a general source term.

| Boundary conditions
In the fluid domain simulation, the circulating medium was gas with a density of 1.29 kg/m 3 , the inlet mass flow rate was 0.14 kg/s, the two outlets were defined as pressure outlet boundary conditions, and the gauge pressures were set to zero. The mass-flow-inlet boundary condition was used to describe the entry condition of the calculation field, and a pressure outlet boundary condition with atmospheric pressure was set for the outlet. In addition, no-slip wall boundary conditions were used at the boundaries of the calculation field.

| Slover settings
The compressible high-speed airflow in the drilling bit is complex for a structure with an annular jet flow, and the air streams ejected from the nozzles collide and interfere with each other. Moreover, collisions between the air and the inner wall of the bit are highly probable, which indicates that the flow field in the bit is pertaining to typically turbulent. Although the average velocity and static pressure distributions predicted by the k-ɛ model were not completely accurate, they were in fair agreement with the measured values. Moreover, the prediction of the sidewall effect in the sidewall jet was quite accurate. Therefore, the k-ɛ model, which can accurately predict three-dimensional complex flow, was selected for the simulations.
To simulate a high-speed compressible gas, the second-order upwind scheme was used for the convective terms, and the central difference scheme was used to discretize the diffusion term. A semi-implicit method for pressure-linked equations was used to solve the pressure-velocity coupling equations. Spatial discretization parameters, including the gradient (least squaresbased), pressure (standard), momentum (second-order upwind), turbulent kinetic energy (second-order upwind), and turbulent dissipation (second-order upwind), were also set. In addition, the subrelaxation factor was set as the control.

| Grid independence experiment
A mesh independence test was performed to select a reasonable fluid domain calculation, and the density of the grid was changed with the condition that the structure, simulation model, fluid type, boundary conditions, and calculation methods remain constant. The computational domains comprised five primary parts: inlet, outlet, suction nozzles, flushing nozzles, annulus space between the inner and outer walls of the drill bit, and annulus space formed by the drill bit and borehole wall. As shown in Table 1, wall, inlet, and outlet, suction nozzles, and flushing nozzles had different mesh sizes. For the case of an inlet with the same mass flow rate, finite element analyses of different grid conditions (I-III) produced a maximum mass flow rate deviation of 0.006 kg/s in the center channel of the three grid cells, representing a 4.05% deviation, which is lower than the 5% grid independence requirement. Figure 5A-C show that the maximum velocity value is between 179 and 182 m/s which is a velocity deviation of 3 m/s that amounts to a maximum deviation percentage of merely 1.68%. In addition, Figure 6 shows that the T A B L E 2 Orthogonal design experiment.

| Orthogonal experimental method
An orthogonal experimental method was used to study multiple levels and mutually interacting factors affecting the RC drill bit. According to the parameters of the RC bit, the L27 (3 7 ) orthogonal design table, shown in Table 2, was selected. The mass flow rate at Outlet 2 (outlet of the annulus) was greater than zero, indicating that the annular space could suck air into it. In contrast, the mass flow rate at Outlet 1 (outlet of the center channel) was less than zero, which means that the air returns to the ground through the center channel. As shown in Table 2, the mass flow rate of T A B L E 3 Range analysis of the simulation results.

| 2005
Outlet 2 is usually greater than zero; therefore, the RC drill bit in this study possesses suction capability. The RC sampling abilities of the three groups (Exp. 1 and Exp. 13, and exp. better) of the structures was further analyzed because the mass flow rates of Outlets 1 and 2 were higher than those of the others. The results of the range analysis of the RC bit simulation are presented in Table 3, where the sequence priorities of the parameters are as follows: flushing deviation angle (α F ), flushing diameter (d 2 ), suction diameter (d 1 ), suction number (N s ), elevation angle of the pressure-restoring groove (γ E ), suction deviation angle (β), and depth of the pressure-restoring groove (H E ).
Therefore, an RC drill bit without a flushing nozzle was introduced for comparative analysis.

| Comparative analysis
In the fluid domain simulation, the inlet mass flow rate was 0.14 kg/s, the two outlets were defined as the pressure outlet boundary condition, and the gauge pressures were set to zero. Figures 7-9 show the velocity contours for the different structures. The velocity of the  RC drilling bit without the flushing nozzle was higher than its velocity without the flushing nozzle, and the better group was faster than the others, with a maximum speed of 327 and 1490 m/s with and without the flushing nozzle, respectively.
Negative pressure is the primary criterion for evaluating RC ability. To investigate the internal negative pressure generation in the three drill bits, the variations in the pressure performance over the structures were recorded during the experiment and are shown in    the results are illustrated in Figure 14. As demonstrated in Figure 14A, the increase in the inlet air mass flow rate and the ratio between the sucked and input for the three bits are 0.030, 0.033, and 0.038, respectively. The corresponding maximum suction rate fluctuations are approximately 7.7%, 9.3%, and 4.5%, respectively. Figure 14B shows that the suction capacities of the bits are 0.172, 0.213, and 0.336, respectively. Finally, the maximum suction rate fluctuation are approximately 1.7%, 3.9%, and 4.3%, respectively. Furthermore, for the same inlet air mass flow rate, the suction capacity of the drill bits without flushing nozzles is higher than that of drill bits with flushing nozzles. In addition, the suction capacity of the RC bit is shown to remains unchanged as the inlet air mass flow rate increases. The results indicate that the RC bit designed in this study results in a stable suction rate. F I G U R E 15 Reverse circulation bit testing platform. 1, Test stand support; 2, vacuum pressure gauge; 3, a casing; 4, bit body; 5, built-in ejector; 6, outer pipe; 7, inner pipe; 8, pressure gauge; 9, gas inlet pipe; 10, mass flow rate sensor; 11, vortex flowmeter.
To verify the simulation results, a prototype of the drill bit was developed based on the optimal parameter configuration of the drilling bit. A schematic of the test platform is shown in Figure 15, where the casing was used to represent the drilling hole, and the injected fluid pipe was set on the casing, and the parameters of the outer and inner pipes on the test platform were identical to those of the dual-wall drill pipes used in coal seam RC sampling. The amount of gas flowing into the ejected fluid channel is one of the indices that reflects the RC performance of the bit. Therefore, a gas mass flowmeter, with upper limits and the measuring accuracies are 25 L/min and 1.5%FS, respectively, was installed horizontally in the ejected fluid channel. A pressure calibrator was employed to measure the negative pressure at the bottom of the bit in the range −0.1 to −2.5 MPa. A vortex flowmeter was used to measure the mass flow rate of inlet with the error ±1.0%. Figure 16 shows the three ejectors used in the RC drilling bit. Three ejectors were installed in the RC drill bit for verification experiments.
To investigate the impacts of input air mass flow rate on the suction capacity, six different input air mass flow rates of 0.04, 0.06, 0.08, 0.10, 0.12, and 0.14 kg/s were selected. The results of the parameter of this investigation for RC drilling with a flushing nozzle is shown in Figure 17A, which shows that the maximum mass flow rate is 0.018. Furthermore, although a discrepancy exists in the suction performance between the different structures, the results clearly show that the optimal parameter configuration of RC results in a higher ratio between the sucked and input air mass, confirming the results presented in Section 4.3.2. However, the result shown in Figure 17B is a direct contradiction. Although, the maximum suction mass flow rate of the optimal parameter configuration of RC is 0.167 without a flushing nozzle, the suction ratio limits the improvement in the inlet mass flow rate because the annular suction does not leave sufficient space for inlet airflow. However, this phenomenon can be attributed to the pipe used to measure the mass flow rate of annular suction air. Therefore, we can conclude that the optimal parameter configuration of the RC performs better than others, and that the experimental results are consistent with the simulation results.

| CONCLUSIONS
To achieve rapid sampling and accurate measurement of gas content, an RC drilling bit with an annular jet structure was proposed. The structural parameters of the RC were analyzed to determine their effect on suction capacity using CFD simulation. The following conclusions were drawn: 1. According to the numerical simulation and range analysis of the CFD simulation results, the optimal parameter configuration of the RC drill bit is as follows: a flushing deviation angle α F = 0°, flushing diameter d 2 = 8 mm, ejector diameter d 1 = 6 mm, number of ejectors N s = 10, suction deviation angle β = 20°, depth of the pressure restoring groove H E = 5 mm, pressure restoring groove deviation angle γ E = 10°. The maximum negative pressure of intake air in the RC drilling system is 2.51 kPa, and maximum mass flow rate is 0.038, and the RC drilling performance was superior to that of the other configurations. 2. RC drilling without the flushing nozzle is better at rapid sampling with a maximum suction mass flow rate of 0.167, which is 10 times that obtained with the flushing nozzle.
This study is limited to the simulation and experiment of single-phase airflow and lacks the multiphase analysis of coal particles and airflow in coal seam. Therefore, the scope for further research is the flow characteristics of coal particles inside the drilling bit.