Research On Digital Twin System Of Welding Robot

Automobile body-in-white welding is one of the common welding work in industry.At present,the welding robot workbench is mainly composed of welding robot,welding clamp,welding fixture and weldment.The use of digital twin technology is an effective method to realize intelligent welding.In order to improve the welding quality and work efficiency of body-in-white welding,the digital twin system of welding robot is studied.Based on the welding requirements of the rear assembly of the left inner panel of the car,the overall design of the welding robot workbench is completed.Aiming at the design goal of the welding robot workbench,the design scheme is proposed.The welding robot workbench is divided into robot module,welding clamp module and welding fixture module.The selection of welding robot is completed,the principle of spot welding is analyzed and the suitable welding clamp is selected according to the welding requirements.The welding fixture was designed for the rear assembly of the left inner panel of the body in white,and the mechanical simulation was carried out.Based on the industrial digital twin tool of Visual Components,the digital twin of welding robot workbench is built by using various modeling methods.According to the principle of digital twin modeling in Visual Components,three construction methods of digital twin are proposed.The digital twins of welding robot and welding clamp are built,and the theoretical analysis of the twins is carried out based on the kinematics of the robot,which provides the basis for the correctness of the digital twin position change of the welding robot.Based on Python script,the definition of welding clamp and welding fixture is realized,and the precise modeling of welding robot workbench is completed.Based on the communication method of serial port server,the virtual and real communication between the welding robot workbench and its digital twin is realized.Based on the communication requirements of digital twin system,the experimental framework of virtual and real linkage of digital twin body of welding robot workbench is built.Aiming at Visual Components industrial digital twin software,two methods for realizing virtual-real communication are proposed.Based on Python script,a TCP server is developed for virtual-real communication.Experiments show that the communication method based on serial port server can realize the virtual-real linkage between digital twins and physical entities.The welding quality prediction model of welding robot workbench is built by BP neural network training method.Aiming at the welding requirements of the rear assembly of the left inner panel of the automobile,the welding parameter data set is obtained based on the orthogonal experimental method.The neural network method is used to train the data,and the model between welding current,motor pressure,welding time,preloading time,maintenance time and tensile shear stress is established.The accuracy rate is 94.81%,which can be used for daily welding production reference.In this thesis,a digital twin system of welding robot workbench is built based on digital twin technology,which realizes the virtual and real linkage between digital twin and physical entity.The neural network method is used to train the data extracted from communication,and the mathematical model of welding parameters and tensile shear stress of weldment is obtained,which provides a reference for this type of welding quality prediction.