Font Size: a A A

Study On Optimization Of Process Parameters For Laser Ball Planting Welding Based On BP Neural Network

Posted on:2023-08-17Degree:MasterType:Thesis
Country:ChinaCandidate:K ShiFull Text:PDF
GTID:2531306836462464Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
As a new micro bump manufacturing technology of electronic packaging,laser ball planting technology can heat and melt welding balls with diameters of tens to hundreds of microns through accurate control of laser,and then micro connect with the substrate to form micro bumps.Compared with traditional welding technology,laser ball planting technology has many advantages,such as high efficiency,local heating,fast cooling and fast heating.However,laser ball planting also has some problems,such as unstable welding quality and poor consistency of welding joints.To solve this problem,this paper compares the two optimization methods of orthogonal test and BP neural network in order to find out the combination of process parameters that can improve the welding quality.In addition,the effects of different Shear conditions on the shear strength and failure mode of solder joints are analyzed,and the stress behavior of solder joints in the shear process is analyzed by simulation.The main research contents and results of this paper are as follows:(1)Exploration on the relationship between laser ball planting process parameters and solder joint shear strengthIn this paper,four process parameters affecting the shear strength of solder joints are selected:laser power,laser action time,nitrogen pressure and welding height.According to the BGA solder joint International Push ball test standard(JESD22-B117A)and the actual use of equipment,select shear height of 50μm,shear speed of 200μm/s to test solder joint shear strength,test equipment selected MFM series multi-functional shear force tester.The relationship between the four process parameters of laser ball planting process and the shear strength of solder joint was studied by orthogonal test.It is found that the four process parameters have a certain influence on the shear strength of solder joints,and the most influential factor is the laser power and the second is nitrogen pressure,laser action time and welding height.In addition,the shear strength of solder joints is not only affected by a single factor,but also restricted by many factors at the same time.It is found that the interaction between laser action time,nitrogen pressure and welding height is stronger than that between laser power and any of its factors.The best combination of process parameters is also deduced through orthogonal test: laser power 85%,laser action time 70 ms,nitrogen pressure1000 Pa and welding height 1mm.(2)Analysis of shear strength change and failure mode of solder joints under different shear conditionsIn the actual working condition,the solder joint is inevitably affected by different forces from all directions.Therefore,the effects of different shear conditions on the shear strength and failure mode of solder joints are studied,and the stress behavior of solder joints in the shear process is analyzed based on ANSYS finite element.The results show that reducing the shear height will increase the shear strength of the solder joint,and increasing the shear speed will reduce the shear strength of the solder joint.Regardless of reducing the shear height or increasing the shear speed,the failure mode of the solder joint will change from the inside of the solder to the IMC layer.The specific performance is as follows: 1)fix the shear speed to 200μm/s,the shear height is from 50μm reduced to 30μm.The failure model of solder joints changed from ductile fracture to brittle fracture,and the shear height continued to be reduced to 10μm.This change is more obvious.The shear strength of solder joints increases with the decrease of shear height;2)When the fixed shear height is 50μm.Change the shear speed from 200μm/s to 700μm/s,the failure model of solder joint completely changes from ductile fracture to brittle fracture,and the shear strength of solder joint decreases.In addition,through the simulation,it is found that in the shear process,the shear degree of each area of the solder joint is different,and the shear force at the bottom is the largest,resulting in the fracture of the solder joint from the IMC layer.(3)Establishment of prediction model of solder joint shear strength and optimization of process parametersAccording to the data of orthogonal test,BP neural network is used to construct the prediction model of solder joint shear strength,and the network structure is optimized by genetic algorithm(GA-BP model is constructed).The model determination coefficient and prediction accuracy before and after optimization are compared.In addition,the process parameters of laser ball planting are optimized by GA-BP model,and compared with the analysis results of orthogonal test,so as to obtain the optimal combination of process parameters.The specific performance is as follows: 1)the determination coefficient R ^ 2 of the BP neural network prediction model before optimization is 0.9196,the maximum percentage of prediction error is 0.1147,and the prediction ability is general.Compared with the former,GA-BP model requires less iterations and higher efficiency in training,and its determination coefficient R ^ 2 is 0.99217,which is increased by 7.9% compared with the former,indicating that GA-BP model has better network performance;2)Comparing the prediction error percentage of the two models,it is found that the maximum prediction error percentage of GA-BP model is reduced by 0.0648,the average error percentage is reduced by 0.0232,and the range value of the error percentage of GA-BP model is smaller,indicating that GA-BP model has higher prediction stability,the amplitude change of the prediction value is smaller and closer to the actual value;3)According to the GA-BP model,the combination of process parameters shear strength is predicted and optimized.The optimal combination of process parameters: laser power,laser action time,nitrogen pressure and welding height are 63.75 W,68ms,1037 Pa and 1mm respectively.Compared with the optimal combination of process parameters obtained by orthogonal test,it is found that they have the same laser power and welding height,and the laser power has the greatest influence on the shear strength.Therefore,it is considered that the prediction ability of GA-BP model is very accurate.The parameter combination optimization of orthogonal test is only selected in the given levels,which has limitations.At the same time,it is found that the corresponding shear strength of the optimal combination obtained by GA-BP model is 58.27 MPa,which is higher than the results of all orthogonal tests.Therefore,the optimal combination of process parameters: laser power,laser action time,nitrogen pressure and welding height are 63.75 W,68ms,1037 Pa and 1mm respectively.Through the above research,this paper obtains the influence degree of four process parameters on the shear strength of solder joints.The effects of different shear conditions on the shear strength and fracture path of solder joints are studied,which provides a reference for the selection and use of devices under actual working conditions.The finite element model of solder joint shear process is established,and the stress behavior of solder joint in shear process is analyzed.The prediction model of solder joint shear strength is established,and compared with the results of orthogonal test,the optimal combination of process parameters is obtained,which has a certain reference significance for improving the quality of laser ball planting welding.
Keywords/Search Tags:Laser ball planting, shear strength, orthogonal test, failure mode, force analysis, neural network
PDF Full Text Request
Related items