Study On Process And Related Prediction Algorithm Of Ultrasonic Welding Of Dissimilar Metals

Ultrasonic welding technology has the advantages of fast welding speed,wide range of weldable materials and simple operation,and is widely used in aviation,automobile,electronics and other fields.In recent years,ultrasonic welding technology has developed rapidly,but there are still many problems,such as a large number of experiments to obtain high-quality welding head;The temperature during welding is difficult to measure;The research on ultrasonic welding mechanism of some dissimilar metal materials is not enough.In view of the above problems,this paper adopts the research means of combining experiment and intelligent algorithm to carry out the research work of ultrasonic welding.The concrete research contents are as follows:(1)The piezoelectric transducer is designed theoretically.For TC4 horn,through COMSOL modal analysis and harmonic response analysis,the amplification factor,displacement node and stress of the horn are obtained,and the design of the tool head is completed.(2)Considering the heat generated by plastic deformation and friction in the process of ultrasonic welding,the finite element model of ultrasonic welding is established by ABAQUS,and the ultrasonic welding temperature field is analyzed.It is found that the heat distribution shows the trend of diffusion from the welding center to the surrounding,and the variation law of welding temperature field with process parameters and weldment thickness is obtained.(3)The influence of process parameters on the tensile shear force of welded joints is studied by single factor variable method.The orthogonal test is designed for welding time,welding pressure and welding amplitude.The primary and secondary order of the influence of various factors on the tensile properties of welded joints is obtained by range analysis and variance analysis,and the best process parameters are welding time0.8s,welding amplitude 8μm,welding pressure 0.3MPa.(4)The prediction model of ultrasonic welding temperature based on BP neural network optimized by sparrow search algorithm and the prediction model of ultrasonic welding tensile strength based on BP neural network optimized by gray wolf algorithm were established respectively.The results show that the prediction accuracy of the two models is high and the generalization effect is good.(5)Aiming at the problems of slow convergence speed and easy to fall into local optimal solution in sparrow search algorithm,this paper proposes an improved hybrid sparrow search algorithm based on t-distribution,logistic mapping and multiuniverse optimization algorithm to change the defects of the original algorithm.Through simulation comparison,Wilcoxon rank sum test and MAE analysis,it can be found that ISSA has a great improvement in both convergence accuracy and convergence speed.