  +86-15658151051                          sales@xingultrasonic.com 
Articles Detail
Home / Articles / ultrasonic impact treatment / How to use FEM ANSYS parameter optimization and probability design of ultrasonic welding horn

How to use FEM ANSYS parameter optimization and probability design of ultrasonic welding horn

Views:140     Author:Site Editor     Publish Time: 2019-07-17      Origin:Site

How to use FEM ANSYS parameter optimization and probability design of ultrasonic welding horn




With the development of ultrasonic technology, its application is more and more extensive, it can be used to clean tiny dirt particles, and it can also be used for welding metal or plastic. Especially in today's plastic products, ultrasonic welding is mostly used, because the screw structure is omitted, the appearance can be more perfect, and the function of waterproofing and dustproofing is also provided. The design of the plastic welding horn has an important impact on the final welding quality and production capacity. In the production of new electric meters, ultrasonic waves are used to fuse the upper and lower faces together. However, during use, it is found that some horns are installed on the machine and cracked and other failures occur in a short period of time. Some welding horn The defect rate is high. Various faults have had a considerable impact on production. According to the understanding, equipment suppliers have limited design capabilities for horn, and often through repeated repairs to achieve design indicators. Therefore, it is necessary to use our own technological advantages to develop durable horn and a reasonable design method.

2 Ultrasonic plastic welding principle

Ultrasonic plastic welding is a processing method that utilizes the combination of thermoplastics in the high-frequency forced vibration, and the welding surfaces rub against each other to produce local high-temperature melting. In order to achieve good ultrasonic welding results, equipment, materials and process parameters are required. The following is a brief introduction to its principle.

2.1 Ultrasonic plastic welding system

Figure 1 is a schematic view of a welding system. The electrical energy is passed through the signal generator and the power amplifier to produce an alternating electrical signal of ultrasonic frequency (> 20 kHz) that is applied to the transducer (piezoelectric ceramic). Through the transducer, the electrical energy becomes the energy of the mechanical vibration, and the amplitude of the mechanical vibration is adjusted by the horn to the appropriate working amplitude, and then uniformly transmitted to the material in contact with it through the horn. The contact surfaces of the two welding materials are subjected to high-frequency forced vibration, and the friction heat generates local high temperature melting. After cooling, the materials are combined to achieve welding.


In a welding system, the signal source is a circuit part that contains a power amplifier circuit whose frequency stability and drive capability affect the performance of the machine. The material is a thermoplastic, and the design of the joint surface needs to consider how to quickly generate heat and dock. Transducers, horns and horns can all be considered mechanical structures for easy analysis of the coupling of their vibrations. In plastic welding, mechanical vibration is transmitted in the form of longitudinal waves. How to effectively transfer energy and adjust the amplitude is the main point of design.


The horn serves as the contact interface between the ultrasonic welding machine and the material. Its main function is to transmit the longitudinal mechanical vibration outputted by the variator evenly and efficiently to the material. The material used is usually high quality aluminum alloy or even titanium alloy. Because the design of plastic materials changes a lot, the appearance is very different, and the horn has to change accordingly. The shape of the working surface should be well matched with the material, so as not to damage the plastic when vibrating; at the same time, the first-order longitudinal vibration solid frequency should be coordinated with the output frequency of the welding machine, otherwise the vibration energy will be consumed internally. When the horn vibrates, local stress concentration occurs. How to optimize these local structures is also a design consideration. This article explores how to apply ANSYS design horn to optimize design parameters and manufacturing tolerances.

3 welding horn design

As mentioned earlier, the design of the welding horn is quite important. There are many ultrasonic equipment suppliers in China that produce their own welding horns, but a considerable part of them are imitations, and then they are constantly trimming and testing. Through this repeated adjustment method, the coordination of horn and equipment frequency is achieved. In this paper, the finite element method can be used to determine the frequency when designing the horn. The horn test result and the design frequency error are only 1%. At the same time, this paper introduces the concept of DFSS (Design For Six Sigma) to optimize and robust design of horn. The concept of 6-Sigma design is to fully collect the customer's voice in the design process for targeted design; and pre-consideration of possible deviations in the production process to ensure that the quality of the final product is distributed within a reasonable level. The design process is shown in Figure 2. Starting from the development of the design indicators, the structure and dimensions of the horn are initially designed according to the existing experience. The parametric model is established in ANSYS, and then the model is determined by the simulation experiment design (DOE) method. Important parameters, according to the robust requirements, determine the value, and then use the sub-problem method to optimize other parameters. Taking into account the influence of materials and environmental parameters during the manufacture and use of the horn, it has also been designed with tolerances to meet the requirements of manufacturing costs. Finally, the manufacturing, test and test theory design and actual error, to meet the design indicators that are delivered. The following step-by-step detailed introduction.


3.1 Geometric shape design (establishing a parametric model)

Designing the welding horn first determines its approximate geometric shape and structure and establishes a parametric model for subsequent analysis. Figure 3 a) is the design of the most common welding horn, in which a number of U-shaped grooves are opened in the direction of vibration on a material of approximately cuboid. The overall dimensions are the lengths of the X, Y, and Z directions, and the lateral dimensions X and Y are generally comparable to the size of the workpiece being welded. The length of Z is equal to the half wavelength of the ultrasonic wave, because in the classical vibration theory, the first-order axial frequency of the elongated object is determined by its length, and the half-wave length is exactly matched with the acoustic wave frequency. This design has been extended. Use, is beneficial to the spread of sound waves. The purpose of the U-shaped groove is to reduce the loss of lateral vibration of the horn. The position, size and number are determined according to the overall size of the horn. It can be seen that in this design, there are fewer parameters that can be freely regulated, so we have made improvements on this basis. Figure 3 b) is a newly designed horn that has one more size parameter than the traditional design: the outer arc radius R. In addition, the groove is engraved on the working surface of the horn to cooperate with the surface of the plastic workpiece, which is beneficial to transmit vibration energy and protect the workpiece from damage. This model is routinely parametrically modeled in ANSYS, and then the next experimental design.

3.2 DOE experimental design (determination of important parameters)

DFSS is created to solve practical engineering problems. It does not pursue perfection, but is effective and robust. It embodies the idea of 6-Sigma, captures the main contradiction, and abandons "99.97%", while requiring the design to be quite resistant to environmental variability. Therefore, before making the target parameter optimization, it should be screened first, and the size that has an important influence on the structure should be selected, and their values should be determined according to the robustness principle.

3.2.1 DOE parameter setting and DOE

The design parameters are the horn shape and the size position of the U-shaped groove, etc., a total of eight. The target parameter is the first-order axial vibration frequency because it has the greatest influence on the weld, and the maximum concentrated stress and the difference in the working surface amplitude are limited as state variables. Based on experience, it is assumed that the effect of the parameters on the results is linear, so each factor is only set to two levels, high and low. The list of parameters and corresponding names is as follows.

DOE is performed in ANSYS using the previously established parametric model. Due to software limitations, full-factor DOE can only use up to 7 parameters, while the model has 8 parameters, and ANSYS's analysis of DOE results is not as comprehensive as professional 6-sigma software, and can't handle interaction. Therefore, we use APDL to write a DOE loop to calculate and extract the results of the program, and then put the data into Minitab for analysis.

3.2.2 Analysis of DOE results

Minitab's DOE analysis is shown in Figure 4 and includes the main influencing factors analysis and interaction analysis. The main influencing factor analysis is used to determine which design variable changes have a greater impact on the target variable, thereby indicating which are important design variables. The interaction between the factors is then analyzed to determine the level of the factors and to reduce the degree of coupling between the design variables. Compare the degree of change of other factors when a design factor is high or low. According to the independent axiom, the optimal design is not coupled to each other, so choose the level that is less variable.

The analysis results of the welding horn in this paper are: the important design parameters are the outer arc radius and the slot width of the horn. The level of both parameters is "high", that is, the radius takes a larger value in the DOE, and the groove width also takes a larger value. The important parameters and their values were determined, and then several other parameters were used to optimize the design in ANSYS to adjust the horn frequency to match the operating frequency of the welding machine. The optimization process is as follows.

3.3 Target parameter optimization (horn frequency)

The parameter settings of the design optimization are similar to those of the DOE. The difference is that the values of two important parameters have been determined, and the other three parameters are related to the material properties, which are regarded as noise and cannot be optimized. The remaining three parameters that can be adjusted are the axial position of the slot, the length and the horn width. The optimization uses the subproblem approximation method in ANSYS, which is a widely used method in engineering problems, and the specific process is omitted.

It is worth noting that using frequency as the target variable requires a little skill in operation. Because there are many design parameters and a wide range of variation, the vibration modes of the horn are many in the frequency range of interest. If the result of modal analysis is directly used, it is difficult to find the first-order axial mode, because the modal sequence interleaving may occur when the parameters change, that is, the natural frequency ordinal corresponding to the original mode changes. Therefore, this paper adopts the modal analysis first, and then uses the modal superposition method to obtain the frequency response curve. By finding the peak value of the frequency response curve, it can ensure the corresponding modal frequency. This is very important in the automatic optimization process, eliminating the need to manually determine the modality.

After the optimization is completed, the design working frequency of the horn can be very close to the target frequency, and the error is less than the tolerance value specified in the optimization. At this point, the horn design is basically determined, followed by manufacturing tolerances for production design.


3.4 Tolerance design

The general structural design is completed after all design parameters have been determined, but for engineering problems, especially when considering the cost of mass production, tolerance design is essential. The cost of low precision is also reduced, but the ability to meet design metrics requires statistical calculations for quantitative calculations. The PDS Probability Design System in ANSYS can better analyze the relationship between design parameter tolerance and target parameter tolerance, and can generate complete related report files.

3.4.1 PDS parameter settings and calculations

According to the DFSS idea, tolerance analysis should be performed on important design parameters, and other general tolerances can be determined empirically. The situation in this paper is quite special, because according to the ability of machining, the manufacturing tolerance of geometric design parameters is very small, and has little effect on the final horn frequency; while the parameters of raw materials are greatly different due to suppliers, and the price of raw materials accounts for More than 80% of horn processing costs. Therefore, it is necessary to set a reasonable tolerance range for the material properties. The relevant material properties here are density, modulus of elasticity and speed of sound wave propagation.

Tolerance analysis uses random Monte Carlo simulation in ANSYS to sample the Latin Hypercube method because it can make the distribution of sampling points more uniform and reasonable, and obtain better correlation by fewer points. This paper sets 30 points. Assume that the tolerances of the three material parameters are distributed according to Gauss, initially given an upper and lower limit, and then calculated in ANSYS.

3.4.2 Analysis of PDS results

Through the calculation of PDS, the target variable values corresponding to 30 sampling points are given. The distribution of the target variables is unknown. The parameters are fitted again using Minitab software, and the frequency is basically distributed according to the normal distribution. This ensures the statistical theory of tolerance analysis.

The PDS calculation gives a fitting formula from the design variable to the tolerance expansion of the target variable: where y is the target variable, x is the design variable, c is the correlation coefficient, and i is the variable number.


According to this, the target tolerance can be assigned to each design variable to complete the task of tolerance design.

3.5 Experimental verification

The front part is the design process of the entire welding horn. After the completion, the raw materials are purchased according to the material tolerances allowed by the design, and then delivered to the manufacturing. Frequency and modal testing are performed after manufacturing is completed, and the test method used is the simplest and most effective sniper test method. Because the most concerned index is the first-order axial modal frequency, the acceleration sensor is attached to the working surface, and the other end is struck along the axial direction, and the actual frequency of the horn can be obtained by spectral analysis. The simulation result of the design is 14925 Hz, the test result is 14954 Hz, the frequency resolution is 16 Hz, and the maximum error is less than 1%. It can be seen that the accuracy of the finite element simulation in the modal calculation is very high.

After passing the experimental test, the horn is put into production and assembly on the ultrasonic welding machine. The reaction condition is good. The work has been stable for more than half a year, and the welding qualification rate is high, which has exceeded the three-month service life promised by the general equipment manufacturer. This shows that the design is successful, and the manufacturing process has not been repeatedly modified and adjusted, saving time and manpower.

4 Conclusion

This paper starts with the principle of ultrasonic plastic welding, deeply grasps the technical focus of welding, and proposes the design concept of new horn. Then use the powerful simulation function of finite element to analyze the design concretely, and introduce the 6-Sigma design idea of DFSS, and control the important design parameters through ANSYS DOE experimental design and PDS tolerance analysis to achieve robust design. Finally, the horn was successfully manufactured once, and the design was reasonable by the experimental frequency test and the actual production verification. It also proves that this set of design methods is feasible and effective.



 Ms. Yvonne
  sales@xingultrasonic.com   
   Room 1103B, Nature business building ,  NO.1160 GongWang Road ,FuYang, Hangzhou,Zhejiang,China