Research Of Seam Tracking For Robot Arc Welding

Welding is an indispensable technology in modern industrial production,which has been widely applied in automobile manufacturing,shipbuilding and other industries.With the advancement of welding robot technology,more and more welding processes are completed by robots.However,at present,most welding robots finish the welding process by the mode of teaching programming,which means that welding robots can only operate along the instructed path.But in the actual welding process,due to processing error occurred in the production process of welding parts,assembly error occurred when the welding parts were assembled to the welding table and error caused by heat deformation,there will be errors between the actual welding seam and instructed path.Therefore,it is necessary to research real-time welding seam tracking technology for welding robots,so that those robots can detect errors of welding seams and timely adjust the movement trails of the robots.In this way,the welding gun can align the welding seams all the time during the welding process.The robot welding seam tracking system is composed of welding robot control module and welding seam tracking module.Although there are robot welding systems in the market,their path planning ports are not opened to users.As a result,it is impossible to revise movement trails in the operation process of the welding robots.In this context,to design a robot arc welding seam tracking system,firstly,a complete set of welding robot control platform was designed.At joint spaces,D-H model was utilized to build a kinematical model of industrial robots with six degrees of freedom and figure out forward and inverse kinematics solutions of the robots' kinematic model.At Cartesian spaces,acceleration S curve was adopted to conduct trace interpolation and complete the track planning of the robots at Cartesian spaces.And Quaternion method was used to design the postures of the robots.In this dissertation,several commonly used sensors used to detect welding seam errors are compared.In view of the characteristic of swing arc welding that it will do periodic motions on both sides of welding seams,Hall current sensor was used to detect welding seam errors.Then several common methods to extract welding seam errors are analyzed,so as to design a kind of method which can acquire welding seam errors through extracting characteristic frequency.Then simulation and experiment were carried out to verify the accuracy of the method.In addition,features of several welding seam error compensation algorithms were explored and then a new error compensation algorithm was designed to make good use of interpolation to compensate for welding seam errors into movement trails of the robots,so that the robots can timely adjust target points of their movement.Moreover,the robots won't swing because of the frequent changes of target points during the operation process.Massive amount of experiments under different welding conditions such as V-type groove straight seam,L-type groove straight seam and L-type groove circular seam were carried out,with a view to verifying the accuracy and stability of this set of robot arc welding seam tracking system.The experiment results show that: In terms of DC swing welding of carbon steel and stainless steel thick plates,under normal welding conditions,when the welding current,swing amplitude and frequency are changed,this set of robot arc welding seam tracking system can achieve favorable tracking effect.