Experimental Study on Mixing Uniformity of Injection On-Line Mixer of Crop Protection Equipment

At present, the traditional pre-mixing is usually used in the plant protection operation in agricultural production, and a few large plant protection machinery uses a simple premixed box to complete the pre-mixing. This kind of mixing method is wasteful for the use of pesticides, and the concentration of liquid can not be controlled accurately. With the rise of precision agriculture, it has become a development tendency of the crop protection equipment to realize online mixing technology of pesticide solution separation by separating the pesticide tank and water tank. And it is inevitable to develop an online mixing system with high level automation and better mixing effect [1-3].

The research content involved in the online mixing technology mainly includes three major directions, namely the design and optimization of the mixing device, namely mixer, the detection and evaluation of the mixing effect, and the research on fluid mixing theory and related control algorithm. Dai et al. [4] And the online mixing devices are mainly divided into two types: jet-type and injection-type. Zhou et al. [5] Many scholars have carried out a lot of research on the mixing device. In 1991, Ghate and Phatak [6] proposed compressed-air mixing device, which adjusts the injection dose of pesticides according to the speed of the machine. Mickaily-Huber et al. [7] proposed that placing enough obstacles inside the static mixer can induce disordered flow to promote mixing, but the pressure drop increases as the mixing unit increases. Sarvanan et al. [8] analyzed the structural parameters of the jet-type mixer based on the fluid dynamics method and found that the mixing effect of the jet-type mixer was greatly influenced by the nozzle diameter. Chen and Xu [9] used highspeed camera to capture the distribution of tracer particles in the mixture of pesticides and processed these images to analyze the mixing effect. Zangina et al. [10] Optimized the operation path of plant protection robot in greenhouse, achieving target spray and variable spray to reduce pesticide application. Gillis et al. [11] combines machine vision system with DIS to develop a variable sprayer for controlling roadside weeds. The experimental results show that the system can meet the flow demand of different numbers of nozzle opening and ensure pressure fluctuation within 10% range, but because the injection point is far away from the sprinkler, the concentration response delay is longer. Downey et al. [12] changed the above-mentioned system from hydraulic power to pneumatic power, and used it as injection power source and current carrying power source respectively. The results show that: the delay time from the response of drug supply to the appearance of the solution in the nozzle is in the range of 20 ~ 30 ms, and the time to reach the target concentration is more than 100 ms. Zhu et al. [13] studied the delay influencing factors of DNIS, and found that diaphragm pump is more suitable for drug delivery than aerodynamic and gear pump; for the mixing characteristics and delay relationship under the condition of continuous current carrying, pneumatic on-off valve was used. Zhu [14] obtained the relationship between the length diameter ratio of static mixers with different structures and the coefficient of variation (CV) of mixing uniformity based on experiments. Vondricka and Lammers [15]; Vondricka and Lammers [16] tried to add a mixing section in front of the nozzle to promote the uniform mixing of liquid medicine, and achieved good results. Luck et al. [17] evaluated the application of red rhodamine pigment as tracer in the experiment of measuring uniformity based on absorbance. The image processing method is another feasible method for uniformity evaluation, the above studies are relatively single test methods. Realpe and Velázquez [18] distinguished the concentration of suspension agent by comparing the gray value, and compared with the detection results obtained by spectrophotometer, it was found that the error was very small; Berthiaux et al. [19] studied the mixing uniformity of two different color powders by principal component analysis (PCA); Muerza et al. [20] used autocorrelation function to dynamically study the static mixing effect of two different color powder particles. Satjaritanun et al. [21] Used Taguchi method to optimize the internal structure of various mixers, such as blades, impellers, and wrong baffle slots. Through the mixing test of solid particles, the rotary impeller mixer was optimized. Adusei-Bonsu et al. [22] Carried out parameter optimization and mixing uniformity analysis of mechanical feed mixer for feed, in which the coefficient of variation was mainly used to judge the mixing uniformity. Alnakeeb et al. [23] Optimized and simulated the heat transfer performance of fin and tube heat exchanger by CFD method. The first factor considered was the pressure drop loss in the heat transfer process, which was the main factor in the fluid exchange process. Fatalla et al. [24] Studied the mass transfer and heat transfer rate of the inner wall of a simulated stirrer with a turbulence promoter, mainly considering the influence of the collective shape of the turbulence promoter on the heat transfer and diffusion. At present, most researchers use a single test method to study the mixing process. This paper proposes that the structure of the mixer should be optimized by simulation, and the image method should be combined with spectrophotometry. The qualitative and quantitative analysis of the mixing process is carried out.

In order to further improve the mixing accuracy and mixing performance of the online mixing system, this paper designs an injection mixer through the preliminary numerical simulation based on the high clearance crop protection equipment spray operation platform. The numerical simulation on structural parameters and fluid parameters of the mixer is performed by the CFD method to determine the best structural parameters of the mixer. And then the experiment is carried out to analyze the mixing performance of these mixers. Then qualitative and quantitative experiments were carried out to analyze the mixing performance of these mixers by high-speed photography and spectrophotometry. Through the closed loop control technology of precision mixing, the zero residual liquid after each operation is reduced, the mixing uniformity of the liquid is improved, the spraying effect is guaranteed, and the maximum efficiency of pesticide utilization is realized. At the same time, the technology reduces the pollution to the environment, effectively improves the working environment of the workers, and reduces the harm of pesticide volatilization to human health. it can effectively reduce the use of pesticides and reduce pesticide residues and secondary pollution.

3.1 CFD Model building of injection mixer

According to the design principle of the mixer and the installation position of the crop protection equipment, a built-in spoiler static mixer model was determined to be built with a main pipe with a diameter of 100 mm and a length of 700 mm, a 10 mm diameter injection port and a 32 mm filling pipe end, and 7 spoilers with a height of 45 mm and an angle of 75°, besides, the spoiler was staggered in the pipe in helical equidistance and the first spoiler faced the injection port with an angle of 75°. Firstly, the injection mixer was constructed with SolidWorks software in a 1:1 3D entity model. According to the requirements of actual operation parameters, it is known that the spray quantity is 140 L/min and the ratio of pesticide solution is 1:300 ~ 1:3000, the agronomic requirements of pesticide mixing ratio, The pesticide mix ratio is usually set to an integer multiple of 1:300. According to the Reynolds Number Formula (1) and turbulence intensity Formula (2), the Reynolds Number of the flow field in the mixer is 29533, which is in the complete turbulence interval. Therefore, a standard k-ɛ turbulence model can be applied in the flow field.

$\text{Re}=\frac{\text{D}\rho }{\mu }\times \frac{{{\text{V}}_{1}}+{{\text{V}}_{2}}}{{}^{1}/{}_{4}\pi {{\text{D}}^{2}}\times 3600}$              (1)

$\mathrm{I}=0.16(\mathrm{Re})^{-1 / 8}$           (2)

3.2 Pre-simulation processing and calculation

According to the potion ratio required by the experiment, the mixer was set with the following boundary conditions: The filling pipe end and injection port were set as the speed inlet to achieve rapid convergence, and the outlet was set as the pressure outlet. The interface was where the injection port and the injection port intersect with the inside of the mixer. The tube wall of the mixer was set as solid surface, etc. The detailed parameters are shown in Table 1.

According to the ICEM module in ANSYS 16.0, the mixer model was divided into blocks, and the mesh of the mixer injection port, filling pipe end and outlet were refined. And the whole grid area was partitioned by combining structure and unstructured grid. The mixer meshing generates 7,872,079 Total elements, 1,404,712 Nodes, Faces, 162,568 individual Cells. The meshing quality is excellent.

Table 1. Basic parameters and boundary condition settings

Since the mixing of the crop protection equipment can be approximated as a relatively stable liquid mixing process, in this simulation, the steady-state model was selected for simulation, the simple algorithm was adopted as the solution equation, and the Second Order Upwind equation was set as the energy equation for discretization. Also, the residual and the liquid volume fraction at the exit were set as monitors, and the accuracy of convergence was set to 10^-5, and convergence was calculated within 8,000 steps.

3.3 Analysis of simulation results

The software Fluent was used to simulate the mixer. Take the simulated data when the pesticide and water pressure are both 0.1 MPa and the potion ratio is 1:300 as an example, CFD-Post was used to process the data to obtain the velocity streamline distribution, as shown in Figure 3(a). And the cloud chart of solution volume fraction is shown in Figure 3(b).

It can be seen from Figure 3(a) that after the water solution is injected from the injection port, the water velocity in the mixer decreases gradually due to the obstruction of the spoiler, while the speed of pesticides injected into the mixer gradually increases with the flow of water. As the rate of the water flow is much faster than that of the pesticide flow, the pesticide and water were merged after the pesticide and water were injected into the mixer, then the mixture was cut by the spoiler to produce turbulence, which makes the mixture continuously divide and recombine, so that the pesticide and water can be gradually mixed. Moreover, when the two fluids enter the mixer, due to the large differences between the speed and density, so in the fluid mixing process, the mixer structure (spoiler) and the interaction between their own fluid. The two types of cutting can significantly accelerate the formation of turbulence and accelerate the fluid mixing.

3a.jpg

(a) Water velocity flow diagram

3b.jpg

(b) Pesticide velocity flow diagram

Figure 3. Speed streamline diagram

4a.png

4b.png

Figure 4. Pesticide volume fraction cloud

5.png

Figure 5. Internal pressure along the mixer

It can be seen from Figure 4(a) that after the pesticide and water are injected into the mixer, there is nearly no mixing between pesticide and water before flowing through the spoiler, and the stratification between the two is obvious. After the pesticide and water passing through the spoiler, the fluid before and after the spoiler generates turbulence, resulting in continuous dividing and recombination of the fluid, and the two can be seen gradually mixing in the cloud chart. The maximum volume fraction of pesticide at the injection port in the mixer is 1. As the pesticide is injected into the mixer, turbulence is generated by the spoiler and mixed with water, the volume fraction of pesticide drops sharply and the mixture uniformity of solution also improves continuously. Figure 4 (b~h) are all the cloud charts of the pesticide volume fraction at the cross section taken from the mixer to each spoiler. According to the figure, with the increase of the number of spoilers, the variation coefficient of the volume fraction of pesticides on the cross section decreases continuously, and the mixing uniformity improves continuously. After the mixed solution passed through the fourth spoiler, the solution was basically mixed evenly, and the subsequent concentration and volume fraction tended to be stable.

It is known that the first spoiler of the mixer is at 120 mm, and the last spoiler is at 680 mm. The point on the center line in the mixer is selected for pressure drop analysis, and the pressure drop is plotted as shown in Figure 5. It can be seen from Figure 5 that the pressure in the mixer decreases gradually as the distance between the spoiler and the inlet increases. In the mixer, before the solution flows through the spoiler, the pressure increases briefly due to the obstruction of the spoiler, but the pressure of the solution drops sharply behind the spoiler. The overall pressure loss in the mixer is less than 400 PA, and the pressure drop is mainly concentrated in the first half of the mixer. Because the solution at the back end has been basically mixed evenly, the smaller the pressure drop is, the greater the pressure drop is, the better the mixing uniformity is, which is consistent with the cloud image of pesticide volume fraction. With the increase of pressure drop, the faster the mixing process, the better the mixing uniformity.

In order to accurately judge the mixing performance of the mixer, a quantitative analysis was carried out on the numerical simulation results of the mixer. Taking into account the experimental subjects of this study, the feasibility of the software simulation and experimental verification, the variation coefficient CV was used as an indicator to evaluate the mixing effect. The coefficient of variation is the most important parameter to consider the mixing uniformity when the material and fluid are mixed. Many references cited in this paper use the coefficient of variation as the main parameter to evaluate the mixing uniformity. The value range of the variation coefficient CV was 0 to 1; when CV=1, it is considered that there is no mixing at all, and when CV≤5%, it can be regarded as uniformly mixed. $C V=\sigma / \bar{c}$. Where, σ: the standard deviation of the pesticide volume fraction; $\bar{C}$: the average volume fraction of the pesticide solution in the entire cross section [25]. The variation coefficient solution formula is shown in formula 3:

$\text{cv}=\frac{\sigma }{{\bar{c}}}=\frac{\sqrt{\frac{1}{N}\sum {{\left( ci-\bar{c} \right)}^{2}}}}{\frac{1}{N}\sum ci}$           (3)

where,

ci ——— the pesticide volume fraction in the i area;

σ ——— the standard deviation of the pesticide volume fraction;

$\bar{C}$ ——— the average volume fraction of the pesticide solution in the entire cross section.

For the optimized mixer model, when the drug pressure and water pressure were all 0.1 MPa, and the mixing ratios were 1:300, 1:600, 1:900, 1:1,000, 1:1,500, 1:2,000, 1:2,500, 1:3,000 respectively, taking the cross section at 120 mm, 220 mm, 320 mm, 400 mm, 4,800 mm, 560 mm, 680 mm away from the spoiler in the mixer. FLUENT (a commercial CFD software package) software was used to derive the volume fraction of pesticides in each cross section, and these volume fractions were calculated using formulas to obtain the variation coefficient of mixing uniformity at each cross section. See Table 2.

Table 2. Simulated values of variation coefficient of mixed uniformity at different distances at 0.1 MPa

Table 3. Simulated values of variation coefficient of mixed uniformity at different distances at 1 MPa

It can be seen from Table 2 that as the number of spoilers increases, the variation coefficient of mixing uniformity decreases gradually, which reveals that adding spoiler inside the mixer can contribute to solution mixing. When the mixing ratio of the pesticide solution was 1:300, the mixing uniformity coefficient of the solution was less than 0.05 after the fourth spoiler, that is, at 400 mm inside the mixer, which means that the pesticide solution has been mixed evenly at 400 mm. When the mixing ratio of pesticide solution is in the range of 1:300 to 1:3000, and the variation coefficient of mixing uniformity at the outlet of the mixer is less than 0.05, which indicates that the optimized mixer has an excellent mixing effect.

In order to verify the effect of pesticide pressure on the mixing uniformity of the mixer, the mixer was numerically simulated as follows. The ratio of the pesticide solution is 1:300 to 1:3000, the inlet pressure of pesticide was set at 1 MPa and the water injection outlet pressure was set at 0.1 MPa. The variation coefficient of the mixing uniformity of the mixer under the pesticide pressure is obtained by calculation processing, as shown in Table 3.

It can be seen from Table 3 that the mixing uniformity of the mixer under pressuring is slightly better than that in normal pressure. When the mixing ratio was 1:300, the variation coefficient of the mixing uniformity of the mixer reached the minimum value 0.0235, and the mixing worked the best. At a distance of 400 mm from the water inlet, that is, after the 4th spoiler, after mixing the potions evenly, the variation coefficient could be obtained as 0.0414 and 0.0235 at the outlet of the mixer. Then, the mixing ratio was increased and the mixer model was simulated and analyzed. When the ratio of pesticide to water was 1 to 3,000, the variation coefficient of mixing uniformity at the outlet of the mixer was 0.0262, which met the requirement of less than 5%. Therefore, it is concluded that all the mixing effect is better during the proportion change of mixing 1:300-1:3000, and the mixer can meet the mixing demand.

5.1 Experiment method

Before the experiment, a carmine standard solution curve was established. And 5 copies of carmine solution with a mass concentration of 3 g/L were prepared to replace water-soluble pesticides, the sample solution was diluted to 1.0, 1.5, 2.0, 2.5, 3.0 mg/L standard samples, and pure water was taken as a blank control. The experiment equipment used Shimadzu Corporation UVmini-1240 ultraviolet-visible spectrophotometer to set the absorption wavelength to 507 nm, and the absorbance of the mixture was measured. MATLAB software and the least square method were used to process the absorbance data of the sample solution. According to the requirement of the ratio of the pesticides, the ratio coefficient was input and the flow data collected by the water flow sensor was read by the control unit. According to the actual ratio requirements, the speed of the peristaltic pump was controlled through the RS-485 bus, and the actual pesticide flow rate was adjusted according to the required ratio. In order to meet the requirements of this experiment, measuring cups were placed under the nozzle and samplings were performed to detect the mixing stability and uniformity of the online mixing system under different injection pressures and solution ratios. The quantitative detection of spectrophotometry shows that the drug liquid has good time and space distribution uniformity after passing through the mixer.

5.2 Experimental study on mixing uniformity of mixing system under different mixing ratios

Mixing uniformity is one of the important indicators to evaluate the mixing effect of mixer. In this paper, the actual concentration under different ratios is obtained according to the absorbance of the carmine mixed solution under different proportions of the solution. The mixture uniformity of the pesticide solution in the injection mixer was tested on the experimental platform of pressurization and normal pressure. The water flow rate was set to 140 L/min, the mixing ratios of the potion and water were changed within the range of 1:300 to 1:3000. The samplings were carried out under the same nozzle for 2 min in each group. Samples were taken for 2 min in each group and randomly sample 10 times in each sampling period, and the single sampling time was 2s. Spectrophotometry was used to detect the samples, and the system mixing uniformity experiment was carried out on the pharmaceutical normal pressure and pressure experiment platform, that is, when the pesticide pressure was 0.1 MPa and 1 MPa respectively. After collating and calculating the medicine taking data, the variation coefficient of mixing uniformity was obtained under different injection pressures and mixing ratios, as shown in Table 4.

It can be seen from Table 4 that when the mixing ratio is 1:300-1:3000 and the pesticide is under normal pressure 0.1MPa, the maximum variation coefficient of pesticide mixing uniformity is 6.87%, at this time, the mixing ratio is 1:3000. And when the mixing ratio is 1:300, the minimum mixing uniformity coefficient becomes 3.33%. Thus, it can be proved that when pesticides are mixed under normal pressure, the overall trend of the variation coefficient of mixing uniformity increases with the increase of the mixing ratio. The reason may be that the ratio of pesticide solution increases, the flow of pesticide is too small, and the internal pressure of the mixer is greater than the pressure of the drug taking pipe, so the pesticide cannot be injected into the mixer as needed. However, when mixed under pressure and within the range of the experimental syrup ratio, the variation coefficient of mixing uniformity is less than 5%, and the maximum is 3.903%, which shows that the mixer has well mixing uniformity during pressure mixing.

Table 4. Coefficient of mixing at mixer outlet with different pressure and mixing ratio

5.3 Experimental study on mixer uniformity under different water flow

It is known that changes in water flow will affect the mixing effect of the online mixing system. To ensure that the designed mixing system has a well mixing effect, change the water amount of diaphragm pump at each mixing ratio, and conduct the experiment on the mixing performance of the system. When the mixing ratio was 1:300, 1:500, 1:1,000, 1:1,500, 1:2,000, 1:2,500, 1:3,000, the opening of variable ratio solenoid valve was changed by control unit to adjust the water intake. Under different mixing ratios and water flow rates, samples with measuring cups were taken in the same nozzle, the time was set for each sampling as 2 min, and randomly sample 10 times in each sampling period. The single sampling time was 2 s, and the absorbance of each sample solution was recorded, the variation coefficient of uniformity was calculated by formula (3). As shown in Figure 12:

12.png

Figure 12. Online mixing system variable working condition test curve

It can be seen from the Figure 12 that the mixing uniformity coefficient of the online mixer decreases with the increase of the water flow rate, which is consistent with the fact that the turbulent kinetic energy increases and the liquid mixing effect improves with the increase of the water flow rate. In the range of 300:1-3000:1 of pesticide solution ratio, when the water flow changes between 35-140 L/min, the mixing uniformity coefficient of the online mixing system is less than 5%, which proves that the designed injection mixer has a well mixing uniformity.