Optimization Design of Online Mixing Apparatus and Mixing Performance Experiment for Crop Protection Equipment

The research of on-line mixing technology is mainly divided into three parts: on-line mixing device, mixing effect, mixing theory and control system [1]. The on-line mixing device is mainly divided into two types, jet type and injection typey [2, 3]. Injection variable spray system usually has some problems such as delay [4, 5]. The delay problem of the system can be improved by using the nozzle direct injection system, that is, the pesticide is directly injected into the carrier stream from the front of the nozzle [6, 7], which shortens the conveying distance of pesticide. However, the system has the defect of insufficient mixing of pesticide and carrier stream. In the aspect of analysis and evaluation of fluid mixing uniformity effect, Luck et al. [8] evaluated the application of red rhodamine pigment as tracer in the experiment of measuring uniformity based on absorbance; Based on the iodine decolorization reaction, Vondricka et al. [9] measured the light transmittance of the mixture and obtained the mixing uniformity at each time of on-line mixing, which is different from the method of evaluating the uniformity by measuring the pesticide concentration on the time series [10, 11]. The image processing method is another feasible method for uniformity evaluation. Realpe and Velázquez [12] 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. [13] studied the mixing uniformity of two different color powders by principal component analysis (PCA); Muerza et al. [14] used autocorrelation function to dynamically study the static mixing effect of two different color powder particles; Xu et al. [15-17] used fluid image processing technology to study the influence of pump speed and system pressure on mixing uniformity. Efendiev et al. [18], Yang et al. [19], Chen et al. [20], Xu et al. [21] and others have carried out some research on online drug mixing and related control technology. Yin [22] found that by staggering the blades in the mixer, the mixing uniformity was significantly improved with the dislocation angle and the pressure loss in the mixer was significantly reduced. He et al. [23] found that the periodic arrangement of baffle structures inside the flow channel can increase the fluid contact area and induce chaotic advection. A number of domestic scholars have conducted research on the online mixing technology, mainly focusing on the precise control of real-time dosage of low-flow pesticides. Zhang and Liu [24] and Li et al. [25] and Yang et al. [26]. In order to further improve the mixing accuracy and mixing performance of the online mixing system, this paper designs an online mixing system based on the high clearance crop protection equipment spray operation platform, and adopts CFD (Computational Fluid Dynamics) method to carry out numerical simulation for multiple mixers, determine the optimal structure parameters of the mixers, and then carry out experiments to analyze the mixing performance of these mixers, thus providing reference for the design and optimization of online mixing system for crop protection equipment and the later variable spray experiment.

In order to further improve the accuracy and mixing performance of the online mixing system, an online mixed drug system was designed based on the highland gap plant protection machine spray platform, using CFD (Computational Fluid). Dynamics) numerical simulation of a number of mixers, determining the best mixer structure parameters, and conducting experiments, analyzing the mixing performance, providing reference for the design, optimization and variable rate spray test of the online mixing system of the plant protection machine.

3.1 CFD mathematical modeling

SolidWorks software was used to conduct 1:1 3D solid modeling for the mixer. The model is shown in Figure 3. It is known that the mixing effect of the staggered blades is better than that of the in-line arrangement. The number of spoilers were increased in the Model d in sequence, and the spoiler was staggered in the pipe in helical equidistance. The models d1, d2, d3 and d4 were established in sequence, and the number of internal spoiler was 3, 5, 7 and 9 respectively. It is known that the spray volume is 140L/min, and the pesticide solution ratio is 1:300 to 1:3000. According to the Reynolds Number and turbulence intensity formula (2), (3), it was calculated that the Reynolds Number of the flow field in the mixer was 29533, which was in the complete turbulence interval, so the standard turbulence model k-ε was adopted for the flow field. As the model was water-pesticide mixture, Eulerian model and mixing model in Fluent could be selected. However, the calculation stability of Eulerian model is relatively poor, existing a problem without convergence. Therefore, the mixing model was chosen in this paper. The mass conservation equation and the mixed flow equation [24] of the mixing model are as follows:

$\operatorname{Re}_{D H}=\frac{D \rho u}{\mu}, u=\frac{v_{1}+v_{2}}{\frac{1}{4} \pi D^{2} \times 3600}$     (2)

$I=0.16\left(\mathrm{Re}_{D H}\right)^{-1 / 8}$     (3)

$\partial p_{m}+\nabla \times\left(p_{m} v_{m}\right)=S_{m}, v_{m}=\frac{\sum_{i=1}^{n} a_{i} p_{i} v_{i}}{p_{m}}$     (4)

where,

p_m-the average density of the pesticide solution mixture phase, kg/m³;

v_m-the average flow rate of the pesticide solution phase, m/s;

a_i-the volume fraction of the i-th phase;

pi- the density of the i-th phase;

v_i-the average flow velocity of the i-th phase (i =1 or 2, where the first phase is the water phase and the second phase is the pharmaceutical phase);

v₁, v₂-the volume flow rate of pesticide and water, m³/h.

3.2 Boundary conditions and parameter settings

The boundary conditions of the mixer were set according to the actual mixing ratio requirements. The main parameters are shown in Table 2. Among them, the filling pipe end and the pesticide injection port are speed inlets, respectively set as water phase and pharmaceutical phase, pressure outlet for outlet, the interface surface for the intersection of the filling pipe end and the pesticide injection port with the interior of the mixer, and the wall for the tube wall of the mixer tube. Simple algorithm and Second Order Upwind were adopted to solve the equations. The volume fraction and residual of liquid at outlet were used as monitors, the convergence accuracy was set to 10^-5, and the convergence within 5000 steps was calculated.

Table 2. Basic parameters and boundary condition settings

3.3 Meshing

ICEM module in ANSYS16.0 was used for the division of the fluid domain mesh of the mixer. Due to the complex structure of the mixer, block division method was adopted in grid division to refine the mesh of pesticide injection port, filling pipe end and outlet of the mixer. And the whole grid area was divided by a method of combination with structured grid and unstructured grid. The grid division of the mixer is shown in Figure 4. The grid numbers of (a), (b), (c), (d₁), (d₂), (d₃), and (d₄) are 2473347, 2280608, 2731353, 3442783, 3685568 and 3702420 respectively, showing good quality of grid division.

qqtu_pian_20201229114836.png

Figure 4. Built-in spoiler mixer meshing diagram

3.4 Analysis of simulation results

The current detection indicators for the mixing effect of the static mixer include relative standard deviation, the variation coefficient, distribution of residence time, and shear rate and so on. Considering the feasibility of the experimental subject, software simulation and experimental examination, the variation coefficient CV is proposed to be the index to evaluate the mixing effect. The value range of the variation coefficient CV is from 0 to 1. When the value is 1, it is considered to be completely unmixed. Theoretically, when the CV is less than 5%, it can be considered to be completely mixed. The formula of the variation coefficient is as follows:

$C V=\frac{\sigma}{\bar{c}}=\frac{\sqrt{\frac{1}{N} \sum\left(c_{i}-\bar{c}\right)}}{\frac{1}{N} \sum c_{i}}$     (5)

where,

c_i-the concentration of the pesticide in the i-th area, mg/L;

σ-he standard deviation of the agent concentration;

$\begin{equation}

\bar{c}

\end{equation}$ -the average concentration of the pesticide solution in the entire cross section, mg/L.

According to the actual conditions, the concentration is replaced by the volume fraction in the simulation calculation. Different mixer models take cross sections from 0.1 m to the outlet along the axis direction when the pesticide pressure and water pressure are all 0.1 MPa, and calculate the CV value of each cross section. As shown in Table 3, the increase in the number of spoilers inside the mixer can help liquid mixing at normal pressure of the pesticide solution. When the number of spoilers are 7, the pesticide solution in the mixer mixes evenly at 220 mm from the filling pipe end and the variation coefficient is 0.031, the variation coefficient at the exit is 0.022, which meets the design requirements.

Figure 5 shows the pesticide liquid volume fraction cloud chart of the four optimization models of model d on the axial center section. As shown in Figure 5, when the density and viscosity of the simulated pesticides are exactly the same, the increase in the number of spoilers inside the mixer helps liquid mixing, but there is distortion. When the spoiler is increased to a certain number, the mixing performance becomes worse, which is consistent with the simulation data. Through numerical simulation, it is found that the mixing effect is the best when seven spoilers are arranged inside the mixer.

Figure 6 is a pesticide solution flow diagram of the axial direction of mixer. It is found that there is a slight pesticide liquid reflux in the pesticide injection port. Therefore, an online mixing system for the mixer that pesticide is injected at high pressure is proposed. As the pressure of the nozzle was 0.3 Mpa, and there was pressure loss in the entire online mixing system loop, so the pressure of the injection port was set to 1Mpa, and numerical simulation was carried out. The simulation results are shown in Figure 7. From which, we can see that the mixing uniformity of the mixer pressurized state is better than that of normal pressure; under the condition that the mixture ratio of pesticide is 300:1, among the seven designed models, when seven spoilers are built in model d₄, the variation coefficient is the lowest and the mixing effect is the best. At 250 mm from the filling pipe end, that is, under the third spoiler, the pesticide solution was mixed evenly, the variation coefficient was 0.038, and the variation coefficient at the outlet of the mixer was 0.02, which met the standard of the mixing uniformity coefficient.

Table 3. Simulated values of coefficient of variation for different cross-section models of different mixer model

5.png

Figure 5. Mixer axial pesticide liquid integral number cloud

6.png

Figure 6. Axial velocity streamline diagram

7.png

Figure 7. Mixer coefficient of variation analog value curve

A simulation analysis was conducted on the model with seven built-in spoilers by increasing the pesticide solution mixing ratio. When the pesticide and water ratio was 3000:1, the coefficient of variation at the outlet of the mixer was 0.018, which met the requirement of CV < 5%. Therefore, this paper selected the model d4 with seven spoilers for experiment.

(1) This paper designed a set of accurate online mixing system, which can improve the effective utilization rate of pesticides in spray operation. Numerical simulation of the mixer used in the mixing system shows that adding spoiler inside the mixer can improve the mixing uniformity. And the mixing effect is the best when there are 7 spoilers inside the mixer and the pesticide solution flows are facing the first spoiler.

(2) When an experiment of the online mixing system was conducted for the mixing uniformity and stability, the pressure was 1 MPa, the relative error of the mixing concentration of the pesticide solution was 4.301% at the maximum, which is significantly better than the 6% of the normal pressure experiment, and the mixing performance is better. And the variation coefficient of the pressurized pesticide solution is less than 3.989%, which proves that the design system has a good mixing performance.

(3) The off-design condition experiment of the online mixing system proves that when the water flow changes, the online mixing system still have a good mixing effect in the pesticide mixing ratio of 300:1 to 3000:1, and the system can be used in the variable spray operation mode, which provides certain reference for the variable spray experiment of crop protection equipment later.

(4) By comparing the simulation value of the mixing uniformity coefficient with the experimental value, it is found that the simulation result is basically consistent with the experiment result. The established CFD model can accurately describe the flow field change in the mixer and the mixing effect of the online mixture system. When conditions and apparatus of the experiment are limited, CFD can be used for virtual experiment to provide more theoretical basis for the design and optimization of online mixing system.