Optimization of Structural Parameters of Piezoelectric Injection Valve Based on Model Experiment and CFD Simulation

The miniaturization of electronic devices requires a decreasing quality of adhesive points. In order to obtain the minimum defect free adhesive point for injection, this paper uses a combination of Taguchi experiment and CFD method to systematically study the influence of structural changes in the needle nozzle on the mass size of the adhesive point of the piezoelectric injection valve. The results indicate that the structural changes of the needle nozzle have a significant impact on the quality of the sprayed adhesive point. For the five types of needle and nozzle structures studied, when the nozzle size is 0.05mm and the needle size is 1.0mm, the adhesive produced by the injection valve is the smallest, and the minimum adhesive mass is 13.3% of the original injection valve (APJ1500) injection mass. Through CFD simulation, it was found that due to the smaller nozzle diameter of the optimal parameter combination during the valve opening stage, and the higher reflux ratio of the nozzle combination during the needle lowering stage, it can produce smaller adhesive mass.


Introduction
With the constantly changing demand for electronic products, many electronic components and devices are developing towards miniaturization and lightweight [1,2], which has led to higher requirements for micro dispensing technology used in the field of electronic packaging [3,4].Piezoelectric injection valves are widely used due to their high efficiency and minimal damage to the substrate [5,6].Piezoelectric injection dispensing technology is a non-contact dispensing technology that uses a piezoelectric injection valve to spray small adhesive droplets onto a bonded substrate.The structural parameters of the dispensing valve have a significant impact on the quality of the sprayed dispensing [7].
Based on a combination of numerical calculation and Taguchi experiment, Lin X L [8] explored the impact of system specific process parameter changes on the droplet volume of the dispensing valve, and provided process parameters for the dispensing valve with the minimum droplet volume.Zheng H [9] used numerical simulation methods to systematically study the effects of single variable changes (driving pressure, liquid viscosity, needle displacement, nozzle diameter) on droplet detachment and distribution volume.It was found that the mass of the adhesive point is positively correlated with driving pressure, needle displacement, and nozzle diameter.Qiu Z Q [10] used CFD method to obtain the spray speed of adhesive at the nozzle mouth and the hydraulic state of adhesive in the sealed inner cavity during the dispensing process.It was concluded that the geometric shape of the ejector pin, the movement stroke of the ejector pin, and the nozzle opening angle are the main factors affecting adhesive injection.Zhou C [11] used numerical simulation methods to systematically study the relationship between nozzle cone angle, nozzle outlet diameter, needle nozzle gap, and jet fluid volume.The results showed that the needle nozzle gap, nozzle outlet diameter, and jet fluid volume were positively correlated, while the nozzle cone angle was negatively correlated with jet fluid volume.Shu [12] used model experimental methods to systematically study the relationship between process parameters such as stroke length, temperature, back pressure, pulse intensity, and adhesive amount, and concluded that the optimized injection system can achieve a minimum droplet diameter of 0.35mm.
In summary, changes in the structure of the needle or nozzle have a significant impact on the amount of glue applied to the dispensing valve.However, most of the above literature only studied the impact of a single variable change on the dispensing performance of injection valves, and the influence of the coupling between the needle and nozzle structure on the dispensing performance of injection valves is still unclear.In order to obtain a smaller adhesive mass for piezoelectric injection valves, this paper uses a combination of model experiments and numerical calculations to systematically study the effect of the coupling between the needle and nozzle structure on the adhesive performance of piezoelectric injection valves.

Structure and Working Principle of Piezoelectric Injection Valve
Figure 1 shows the physical diagram and internal structure diagram of the piezoelectric injection valve APJ1500 used in this article.The valve is mainly composed of PZT piezoelectric ceramics, displacement amplification hinges, springs, flow channels, impact pins, and nozzles.When piezoelectric ceramics are electrified, their shape will change and generate small displacement.This small displacement is magnified by the displacement of the hinge acting on the needle, and the reciprocating motion of the needle is achieved by controlling the positive and negative voltage.During the reciprocating motion of the needle, the fluid is separated and sprayed through the switch of the valve, and the entire process is non-contact droplet spraying.According to existing literature, the nozzle and needle structure of the injection valve play a crucial role in fluid mass distribution.In order to obtain a smaller adhesive mass, this paper uses a combination of model experiments and CFD to systematically study the effects of five different needle and nozzle structures on the adhesive performance of the piezoelectric injection valve, as shown in figure 2.   Considering the mechanical strength requirements and coordination requirements of the firing pin nozzle, based on the existing firing pin nozzle, the top and nozzle dimensions of the firing pin remain unchanged.The diameter of the firing pin and the nozzle outlet diameter are changed respectively, and the bottom of the firing pin is rounded to the side of the firing pin with an arc transition.For the convenience of expression, it is later referred to as the 1.5mm diameter needle designed based on a 1.5mm diameter circle.The nozzle with a 0.05mm outlet diameter in figure 2 is referred to as the 0.05mm nozzle, and the names of other needle nozzles are similar.Figure 3 shows the physical image of the needle used in the model experiment of this article.The number of the needle nozzle is from 1 to 5 from left to right, and table 1 shows the parameters of the needle structure for each nozzle.

Establishment of Experimental Platform
Figure 4 shows the experimental platform for piezoelectric injection valves, which consists of a piezoelectric valve, a feeding pressure pipe, a condensing pipe, a displacement control system, a high-precision weighing platform, and a microscopic vision system.Piezoelectric valves mainly use the piezoelectric effect to drive the needle to separate the liquid, achieving the spraying of liquid droplets.
The feeding pressure pipe provides the feeding pressure, and provides glue for the piezoelectric valve.
The cooling tube controls the temperature of the piezoelectric valve to prevent the temperature of the valve body from rising during the droplet injection process, which affects the liquid separation effect.The displacement control system is used to achieve precise injection of piezoelectric valve positions.The high-precision weighing platform can accurately measure the fluid mass measured by the piezoelectric valve, with a display accuracy of 0.0002mg.The microscope vision system can observe the state of the droplets when they spray onto the substrate, thereby determining whether the quality of the adhesive point is qualified.Table 2 shows the physical parameters of the glue used, which is a polymer compound and a non-Newtonian fluid.Its fluid dynamic viscosity changes with shear force.
The amount of extruded glue and the final shape of glue points are important parameters for evaluating the performance of dispensing [13].The experimental process is as follows: first, record the set parameters and perform a glue dot array.Use a microscopic vision system to observe whether there are defects.If there are no defects, perform a droplet spray quality inspection and record the data.Among them, the quality detection method involves measuring the mass of 50 adhesive points using a high-precision weighing platform, averaging the measurement results, and ultimately obtaining the weight of a single droplet.Repeat this step 10 times and average the data by removing the maximum and minimum values.The final value obtained is the droplet injection mass under this parameter; If there are defects, skip the quality inspection stage and record them.

Taguchi Experimental Design
Taguchi experimental design is a statistical analysis method, whose core idea is to conduct experiments in multiple combinations of different parameter values to determine the degree of influence of each parameter on the output variable and the optimal parameter combination [14][15][16].In the Taguchi method, signal-to-noise ratio (S/N) is the core indicator for analyzing the advantages and disadvantages of the experimental scheme.For different optimization objectives, there are three models with high signal-to-noise ratio, low signal-to-noise ratio, and low signal-to-noise ratio.Since the optimization objective used in this article is to minimize the amount of adhesive sprayed, in order to minimize the target response, the signal-to-noise ratio model is chosen to be small, and its definition formula is: Figure 5 shows the average signal-to-noise ratio of each factor at different levels.From the graph, it can be observed that the average signal-to-noise ratio of the other firing pin diameters decreases with the increase of the horizontal value, except for the firing pin diameter of 0.5mm.The average signal-to-noise ratio of the nozzle diameter decreases monotonically with the increase of the horizontal value.Through the analysis of table 5, the influence of the diameter of the needle and nozzle on the amount of adhesive sprayed by the piezoelectric injection valve can be obtained.The degree of influence can be determined by the parameter p Z , The larger the value of p Z , the deeper the impact, which is specifically defined as: In the formula:L pi represents the average signal-to-noise ratio of each factor p at levels i=1, 2, 3, 4, and 5.
Figure 6 shows the p Z corresponding to each factor_ From the figure, it can be seen that the influence of structural parameters on the amount of adhesive sprayed by the piezoelectric injection valve is d>D, indicating that the nozzle diameter has a greater impact on the amount of adhesive than the needle diameter.Figure 7 shows the variation of the injection quality of the injection valve with different firing pins and nozzle structures.From figure 7 (a), it can be seen that the dispensing quality of the injection valve first increases with the increase of the firing pin diameter.When the firing pin diameter is greater than 2.0mm, the dispensing quality tends to stabilize.However, from figure 7 (b), it can be seen that when the firing pin is 0.5mm, the dispensing quality remains 0, indicating that the droplet cannot be sprayed.In other firing pin situations, as the nozzle diameter increases, the trend of injection mass increase is roughly the same and positively correlated with the nozzle diameter.When the firing pin diameter is 1.0mm, 1.5mm, and 2.0mm, the trend line spacing is almost equidistant, while when the firing pin diameter is 2.5mm to 2.0mm, the trend line is almost coincident.It can be found that the injection mass of 2.0mm diameter firing pins and 2.5mm diameter firing pins is not significantly different in different nozzle combinations.In order to analyze the reasons, this article further analyzes the differences in the structure of the needle.Based on experience, the difference in the fit between different firing pins and nozzles lies in the length of the contact line between the firing pin and nozzle.The relationship between the contact line length of firing pins with different diameters and nozzles in this article is shown in table 6.It can be observed that the length of the contact wire of the impact pins with diameters of 0.5mm, 1.0mm, 1.5mm, and 2.0mm increases with the increase of the impact pin diameter.According to Figure 7 (a), the length of the contact wire of the impact pin is positively correlated with the amount of adhesive sprayed.The length of the contact line between the 2.0mm diameter needle and the 2.5mm diameter needle is very close, so the amount of adhesive sprayed is equivalent when the 2.0mm diameter needle and the 2.5mm diameter needle are matched with different nozzles.

Analysis of the Reasons for Different Dispensing Performance Caused by Different Firing Needle Nozzle Structures
In order to investigate the reasons for the different dispensing performance of the dispensing valve using different firing needle nozzle structures, experimental group 6 with the best dispensing performance and experimental group 20 with the worst dispensing performance were selected as the research objects.
CFD method was used to analyze the flow field structure of the selected dispensing valve.The flow field of the dispensing valve is solved using the fluid solver Fluent2020, and the two-phase flow is solved using the VOF method.The coupled equations of pressure and velocity are solved using the SIMPLE algorithm.The time discretization format is a first-order implicit format, and the discretization gradient is selected based on the least squares element gradient dispersion.The pressure discretization is done using PRESTO!The momentum discretization adopts the QUICK format, while the volume fraction discretization adopts the compressive format.The accuracy of absolute residual convergence is, and the motion of the driving needle is controlled by an embedded UDF program and achieved using dynamic grid technology.
Figure 8 shows the calculation domain and boundary conditions.The pressure inlet is the feed pressure, the pressure outlet is atmospheric pressure, and the lower boundary uses Axis boundary conditions.The other boundary conditions are all wall.
Figure 9 is a schematic diagram of the grid after discretization of the computational domain, dividing the fluid domain into a moving region and a stationary region.Fluid domain 1 is a moving region, and a triangular mesh is used for mesh reconstruction.Fluid domain 2 and fluid domain 3 are stationary domains, keeping the quadrilateral mesh unchanged.The initial condition is that the glue is filled with fluid domain 1.     Figure 11 shows the time dependent curve of the adhesive amount at the outlet of the two injection valves studied.From figure 11, it can be seen that the amount of glue increases with time before closing the valve, and tends to stabilize after closing the valve.Therefore, the amount of glue is related to the opening stage and the falling stage of the needle before closing the valve.In addition, the mass increment increases during the descent stage of the firing pin from 1.2ms to 1.7ms, indicating that as the distance between the firing pin and the nozzle mouth decreases, the mass flow rate at the nozzle outlet increases.
In order to explore its internal mechanism, figure 12 shows the streamline diagrams of the attachment of the needle nozzle during the injection process of the two studied injection valves.It can be clearly seen that the nozzle diameter plays a major role in the valve opening stage, as the nozzle diameter in Experiment 6 is smaller and there is less high-speed fluid at the outlet, resulting in less glue output during this stage.In the descending stage of the needle, the fluid domain pushed down by the needle is divided into two parts: one is squeezed by the needle and flows towards the nozzle, and the other is flowing back towards the feeding port along the gap between the valve body and the needle.The proportion of glue in both directions determines the amount of glue produced during the needle's descent process.The nozzle needle combination in Experiment 6 has a higher reflux ratio, combined with less glue output during the valve opening stage, so this structural parameter can spray smaller droplets.By comparing the streamline diagrams of the 1.35ms and 1.65ms injection processes, it can be seen that when the needle is close to the nozzle, its reflux ratio is greatly reduced, which is the reason for the negative correlation between the mass flow rate at the nozzle outlet and the distance between the needle and the nozzle.

Conclusion
In order to reduce the adhesive mass produced by existing piezoelectric injection valves, this paper adopts a combination of model experiments and numerical calculations to systematically study the effect of the coupling between the needle and nozzle structure on the adhesive performance of piezoelectric injection valves.The main conclusions obtained through analysis are as follows: (1) The structural changes of the needle and nozzle have a significant impact on the dispensing performance of the dispensing valve.For the five nozzle structures studied, the dispensing quality of the dispensing valve first increases with the increase of the diameter of the needle.When the diameter of the needle is greater than 2.0mm, the dispensing quality tends to stabilize, while when the diameter of the needle is 0.5mm, the dispensing valve cannot complete dispensing; On the other hand, for the five needle structures studied, the dispensing quality of the dispensing valve increases with the increase of nozzle diameter.
(2) The coupling parameters of the nozzle needle for ejecting the minimum droplet mass are 1.0mm needle and 0.05mm nozzle, and the minimum droplet mass that can be ejected is 0.0039mg, which is 13.3% of the existing APJ1500 injection mass.
(3) Through CFD simulation, it was found that due to the smaller diameter of the nozzle port in Experiment 6 during the valve opening stage and the higher reflux ratio of the nozzle needle combination in Experiment 6 during the needle descent stage, this structural parameter has a smaller adhesive output.Because the impact force exerted by the needle on the fluid and the gravity of the droplet itself under this structural parameter are sufficient to overcome the viscous force and surface tension of the fluid and detach from the nozzle, this structural parameter can spray smaller droplets.

Figure 1 .
Figure 1.Physical diagram and internal structure diagram of piezoelectric injection valve.

Figure 3 .
Figure 3. Physical image of the needle.Figure 4. Experimental platform.

Figure 4 .
Figure 3. Physical image of the needle.Figure 4. Experimental platform.

Figure 5 .
Figure 5. Average signal-to-noise ratio of various factors.

Figure 6 .
Figure 6.The degree of influence of various factors on the amount of adhesive sprayed.

Figure 7 .
Figure 7. Relationship between structural parameters and mass of sprayed adhesive points.

Figure 8 .
Figure 8. Calculation domain and boundary conditions.

Figure 9 .
Figure 9. Schematic diagram of computational domain grid.The simulation time from 0ms to 1.2ms is the valve opening stage, from 1.2ms to 1.7ms is the needle descent stage, and after 1.7ms is the valve closing stage.The reason why droplets can successfully achieve spraying is because the impact force of the needle on the fluid and the gravity of the droplets themselves make the fluid enough to overcome the viscous force and indicate tension and detach from the nozzle.From the simulation diagram of the colloidal spraying process in figure10, it can be seen that both Experiment 6 and Experiment 20 meet this condition.

Figure 10 .
Figure 10.Fluid Injection Process.Figure 11. shows the time dependent curve of the adhesive content at the outlet of the two injection valves studied.

Figure 11 .
Figure 10.Fluid Injection Process.Figure 11. shows the time dependent curve of the adhesive content at the outlet of the two injection valves studied.

Table 1 .
Structural parameters of nozzle needle.

Table 3 .
Parameters of Taguchi Experiment.In order to explore the relationship between fluid injection structure and droplet mass, and find the minimum injection mass, this article designed a two factor and five level Taguchi experimental table, combining each type of needle with each type of nozzle.Table3shows the combined experimental parameters table.During the experiment, the fixed and unchanged parameters are shown in table4.

Table 4 .
Experimental Process Parameters.Analysis of Taguchi Experiment ResultsTable5shows the injection quality of injection valves with different needle nozzle structures obtained through experiments.From table 5, it can be seen that the amount of adhesive sprayed by the original needle in combination with the nozzle (existing APJ1500) is 0.029325mg.The combination with the highest amount of adhesive sprayed is Experiment 20, and the combination of needle diameter 2.0 and nozzle diameter 0.15 can achieve a maximum amount of adhesive sprayed of 0.0735mg.The combination with the smallest amount of spray glue is Experiment 6.The combination of a needle diameter of 1mm and a nozzle diameter of 0.5mm has a minimum spray glue amount of 0.0039mg, which is 13.3% of the original needle and nozzle combined spray glue amount.

Table 5 .
Results of Taguchi Experiment.

Table 6 .
Relationship between the diameter of the needle and the length of the contact wire.