Optimization design and experimental study of axial-swirling bubble generator

On the basis of the venturi tube bubble generator, a new axial swirl bubble generator is designed, and the numerical simulation of the bubble generator with different number of three baffles is carried out by using Fluent software, and it is found that the three-lobed type is significantly better than the two-leaf type and four-leaf type in the swirl velocity and core swirl area, and the response surface optimization method is used to optimize the key structure and analyze the interaction influence. The results show that the parameter that can most affect the bubble diameter is the throat diameter, followed by the expansion angle, the contraction angle has the least effect, the optimal parameters are 21.9°, the throat diameter is 5mm, and the expansion angle is 13.8°, and the accuracy of the numerical model is verified by experimental comparative analysis, and finally the influence of gas-liquid flow on bubble size is studied with the help of high-speed camera.


Introduction
In recent years, the problem of industrial pollution has become increasingly prominent, a large amount of industrial sewage is discharged into the river, and the chemical and toxic components in it have led to serious pollution of water quality, and even threatened the life and health of residents, so the purification of polluted water is imperative [1].
Micro-nano bubble generator is currently the most effective technology to deal with water pollution, its structure design is simple, and it has the characteristics of low production cost and high bubble generation rate.The micro-nano bubbles produced by it have unique properties that ordinary bubbles do not have, such as high interface potential, large specific surface area, hydroxyl radicals and long retention time [2], in addition to sewage treatment, it is also widely used in aquaculture, ship drag reduction, mineral flotation, biomedicine and other fields [3][4][5][6], so it has broad market development prospects.Ohnari [7] studied the distribution and breaking process of bubbles in the designed tangential swirl bubble generator.Li Zhekun [8] et al. optimized the structural parameters of the designed swirling bubble generator; Ding Guodong [9] studied the position of the air injection hole of the venturi bubble generator and found that the size of the air injection hole located in the larynx was smaller than that of the air injection hole located at the water inlet; A. Kaneko [10] used a high-speed camera to film the brokening process of the Venturi tube bubble generator, pointing out that the speed of the mixed phase in the throat suddenly increased and the bubbles deformed in the diffusion section and broke into fine bubbles; The literature [11][12] studies the effects of water flow, gas injection rate and geometric parameters on the number, size and distribution of bubbles.In order to effectively break the bubbles into micron bubbles, a new axial swirl bubble generator is designed on the basis of the venturi bubble generator, adding a fan-shaped baffle can improve the swirl strength, strengthen the bubble breaking and shear effect, and analyze the influence of the number of baffles on the swirl intensity of the flow field, and then use the response surface method to optimize and analyze the key structural parameters to obtain the optimal structure design of the bubble generator, and finally use the optimized bubble generator to study the factors affecting the size and distribution of bubbles.

Bubble generator structure design
The structure of the swirl bubble generator is shown in Figure 1.The overall structure consists of water inlet, fan baffle, shrinkage section, hose, air inlet and diffusion section.During the working process, the water phase and the gas phase enter the bubble generator from the water inlet and the air inlet respectively, and the initial bubbles generated by the gas phase in the throat and the water phase collide and break and shear at high speed in the larynx, and then form micron-level bubbles.Cassidy [13] et al. studied the characteristics of gas-liquid two-phase flow in the swirl field, and believed that the higher the swirl intensity, the greater the breaking and shear effect of water relative to the bubble, and the smaller the diameter of the generated microbubbles, combined with the conventional venturi structure, a new axial swirl bubble generator was designed, which is characterized by a fan-shaped baffle at the water inlet section, and the swirl formed by the fan-shaped baffle can effectively improve the turbulence intensity of the flow field, thereby further reducing the bubble diameter.

Experimental process
The experimental circuit in this article is shown in Figure 2, and the experimental device includes a water tank, a booster pump, a ball valve, a flow meter, an axial swirl bubble generator, an air tank, a light source, and a high-speed camera.During the experiment, the booster pump runs the water flow through the entire circuit, after the pressure is stable, the ball valve is used to adjust the water flow and air flow to the set value required for the experiment, at this time, the water phase and the gas phase in the throat for turbulent collision and crushing and shearing, and finally form a large number of microbubbles, and finally turn on the LED and use the high-speed camera to shoot continuously in the observation area, and store the data in the computer, and obtain the particle size of the bubble under different working conditions through image processing.
The experiment uses Photron's fastcam mini wx100 high-speed camera with a resolution of 2048 2048@1080fps.The image processing software adopts Image-Pro Plus 6, and its processing steps are as follows: 1) set the reference ruler; 2) Select the measurement area; 3) Grayscale processing; 4) Calculate the equivalent diameter of the bubble, and then explore the law of the influence of water flow rate on the particle size of the bubble through the Sauter average diameter of the bubble, and the formula is as follows: where  is the equivalent diameter of the bubble.The initial main parameters of the bubble generator are shown in Table 1, among which the effects of shrinkage angle, throat diameter and diffusion angle on bubble breaking efficiency and bubble particle size are critical, and the response surface method is used to optimize the design of this key structural parameters.
where  ⃗ m is the average velocity of mass;  is the mixing density.The momentum equation is: where  ⃗ , is the drift velocity of the kth phase, n is the number of phases,  ⃗ is the volume force, and  is the density of the mixture.

Turbulence model.
The RNG k-ε turbulence model used in this simulation not only has higher calculation accuracy than the standard k-ε turbulence model, but also is more suitable for swirl flow and local complex shear flow.
2.3.3.Boundary conditions are mesh independent validation.The working conditions of this paper are normal temperature and pressure, the main phase is water, the secondary phase is air, the gas and liquid inlet is the speed inlet, the speed size of the aqueous phase is 1m/s, the gas-liquid ratio is 10%, the outlet is the pressure outlet, and the wall is set as the boundary condition of the non-slip wall.In order to avoid the divergence of calculation results, the first-order-upwind is adopted first, and the simplec algorithm is used for computational simulation, and the time step of transient calculation is 0.001s.Using the ANSYS software Meshing for meshing, in order to improve the calculation accuracy, the grid local encryption of the strongly flowing areas such as sector baffles, throats and air injection holes was obtained, and four different sets of mesh numbers were obtained for grid independence verification, namely 184390, 256176, 380709, 452531.It is found that with the increase of the number of meshes, the radial velocity of the throat gradually increases, and the velocity curve of the grid number of 380709 and 452531 tends to flatten, indicating that the continuous increase of the number of meshes has little effect on the calculation results, so the grid number can be used 380709, and the meshing is shown in Figure 3.

The effect of the number of sector bezels
Figure 4 shows the velocity vector distribution cloud of three new bubble generators with different numbers of baffles (X=-30mm).The water flow velocity of the corresponding area of the three baffles is much lower than that of the corresponding area of the baffle sidewall, which is because when the water phase passes through the fan-shaped baffle of the water inlet, it will be subject to a certain flow resistance, so that the water flow velocity of the corresponding area of the baffle decreases, while the water flow velocity in the corresponding area of the baffle sidewall gradually increases and forms a swirl in the contraction section.As can be seen from Figure 4, the number of baffles has a significant influence on the velocity and swirl intensity, which in turn affects the production efficiency of microbubbles.When the number of baffles is 2, a weak swirl is generated between the two baffles, and the distribution of the entire swirl area in the plane is relatively uniform, but the maximum water velocity is only 1.8m/s, which is much lower than that of the three-lobed type and the four-lobed type.Compared with the two-lobed structure, the maximum water velocity of the four-leaf type is 2.2m/s and the swirl intensity is higher than that of the two-lobed structure, but the swirl core area along the side wall of the flow channel is too small, because the blockage area of the four-lobed structure is too large, resulting in fluid flowing mainly in the void between the four lobes.The maximum water velocity of the three-blade structure is 2.4m/s, and the swirl core area is evenly distributed in the plane, which can provide higher swirl intensity for subsequent bubble breaking and shearing.

Optimized design of critical structures
Kress [14] et al. found that the liquid-phase Reynolds number Re is inversely proportional to the particle size d of the microbubble by studying the breakage characteristics of bubbles in venturi, and obtained a theoretical formula: where  is the Weber number,  is the surface tension, g is the coefficient,  is the water flow density, and  is the water flow viscosity.
In the design process of the new axial swirl bubble generator, it is found that the key structural parameters at the throat not only determine the turbulence intensity of the expansion section, but also affect the water flow velocity of the expansion section, thereby further affecting the particle size of the microbubbles.Therefore, the response surface method (RSM) was used to optimize the contraction angle, throat diameter and expansion angle of the three key structural parameters in the new axial swirl bubble generator.The design-expert design software was used to simulate the key structural parameters, and the factors and levels were shown in Table 2, and the second-order Box-Behnken (BBD) was selected for the design scheme, and the maximum water flow velocity of the expansion section (X=25mm) was used as the response target, and the 17 selected schemes were simulated by CFD numerical simulation.As shown in Figure 5(a), the parameter that has the greatest influence on the maximum water flow velocity of the expansion section (X=25mm) is the diameter of the pipe, and when the contraction angle remains unchanged, the maximum flow velocity of the expansion section increases with the decrease of the diameter of the throat.The reduction of the diameter of the throat of the new axial swirl bubble generator increases the pressure at the throat, greatly increases the fluid velocity of the expansion section, and at the same time increases the turbulence energy sharply and reduces the surface tension of the bubble, which makes the bubble more fragile.Secondly, when the contraction angle remains unchanged, the maximum water flow velocity of the expansion section increases with the increase of the expansion angle, but the influence of the expansion angle on the water flow velocity is significantly smaller than that of the throat diameter.This is due to the increase of the expansion angle, which increases the cross-sectional area of the flow channel of the expansion section, and increases the vortex strength of the wall of the expansion section, thereby strengthening the bubble breaking process.Based on the response surface optimization model, the optimal design of the key structural parameters of the new axial swirl bubble generator is sought by taking the maximum water flow velocity of the expansion section (X=25mm) as the response goal.The results showed that when the contraction angle was 21.9 0 , the diameter of the throat pipe was 5mm, and the expansion angle was 13.8 0 , the maximum water flow velocity was 13.8m/s.Compared with the initial design scheme, the maximum water flow velocity is increased by 16%, that is, the bubble particle size is reduced by 16%, and the optimization effect is obvious.

Bubble effect comparison
As shown in Figure 6-7, during the experiment, it was found that the inlet section was ordinary transparent color of the water body, while the diffusion section became milky white.When the gasliquid flow rate is  =0.02m 3 /h and  =0.8m 3 /h is unchanged, it can be found that the average particle size of the microbubbles generated by the initial bubble generator is 701m, while the average particle size of the optimized bubble generator is 613 m, and the bubble particle size is reduced by 12.5%, which is within the allowable error range, so the accuracy of the above numerical calculation can be expressed.

Effect of gas-liquid flow on bubble diameter
As shown in Figure 8, when the gas flow is constant, the average diameter of the bubble Sauter decreases with the increase of the liquid flow.This is due to the increase of liquid flow leading to a high velocity of the bubble generator throat, resulting in a larger crushing and shear capacity of the MATMA-2023 Journal of Physics: Conference Series 2691 (2024) 012036 liquid in the expansion section, which can break and merge large bubbles for many times, so that the average diameter of bubbles tends to decrease.When the liquid flow rate is unchanged, the larger the gas flow, the gas volume increases, and the water flow cannot efficiently break the bubbles, thereby increasing the average diameter of the bubbles.

Conclusions
In this paper, a new axial swirl bubble generator is designed, which uses the commercial software ANSYS Fluent to simulate the baffle and key structures numerically, and uses visual experiments to study the relationship between water flow and bubble size, and obtains the following conclusions: (1) Comparative analysis of the new bubble generator with different number of three baffles, the results show that the three-leaf type is higher than the two-lobed and four-lobed types in the swirl velocity and core swirl area.(2) The key structural parameters of bubble occurrence were optimized by the response surface method, and the results showed that the parameter that could most affect the bubble diameter was the throat diameter, followed by the expansion angle, the contraction angle had the least influence, the optimal parameters were 21.90, the throat diameter was 5mm, the expansion angle was 13.80, and the optimization rate was 12.5% through experiments.(3) The results show that the average diameter of bubble Sauter decreases with the increase of water flow, and increases with the increase of air flow, and the average diameter of bubble decreases.

Figure 2 .
Figure 2. Schematic diagram of the experimental process.

Figure 8 .
Figure 8.The average diameter of the bubble varies with the flow rate.

Table 1 .
Structural parameters of new axial swirl bubble generator.In this paper, ANSYS Fluent software is used for numerical simulation, and its multiphase flow model includes VOF model, Mixture model and Eulerian model, among which the Mixture model, as a simplified multiphase flow model, has the advantages of high simulation accuracy and fast calculation speed, which has good applicability to this calculation.The continuity control equation for the Mixture model is:

Table 2 .
Factors and levels of the response surface method.