Research On Narrow Weld Seam Tracking Method Based On Coaxial Vision Sensor

With the wide application of laser welding technology,the automation degree of welding production becomes higher and higher.In order to ensure the precision and efficiency of laser welding and improve the quality of welding workpiece,it is necessary to develop an automatic seam tracking system.Therefore,based on the coaxial monitoring technology and the gantry 3D movement platform,this paper builds the laser welding process monitoring system from several aspects,such as image acquisition module,control module,executive mechanism module and humancomputer interaction module.This system adopts the master-slave control system architecture of "PC+ motion controller and PA numerical control system".PC as the upper computer,which is responsible for the signal collection and processing,parameter setting,sending commands and graphical display.The motion controller and PA numerical control system as the child controller,which are responsible for completing the execution and feedback of various orders.Among them,the motion controller communicates with the master control system PC through the PCI interface,and interacts with the PA numerical control system through the signal adapter,so as to realize the synchronous movement of each axis of the gantry 3D motion platform.How to identify and correct the deviation information of weld seam in real time is the key to realize the tracking of weld seam.At present,linear structural light is widely used for weld identification,which is strongly dependent on the type of weld groove and is not applicable to narrow welding seams with no obvious structural characteristics.In this paper,the laser point light source is used as the auxiliary light source to directly illuminate the weld area,the coaxial monitoring method based on visual sensing is adopted to obtain the weld image information.The image processing method based on threshold segmentation and centroid finding was used to extract the coordinates of the weld center.Through mathematical modeling,the transverse deviation,tangential angle deviation and azimuth angle deviation of the motion state parameters of welded joints are analyzed quantitatively,and found that lateral deviation error can well reflect the relative relationship between the seam and welding head.At the same time,using the image processing algorithm proposed in this paper can easily obtain lateral deviation,meet the real-time demand in seam tracking.Then a digital filtering algorithm based on mean limiting filtering is proposed,which can significantly improve the accuracy of the collected signals.In order to improve the precision of seam tracking,a fuzzy adaptive PID control algorithm based on integral separation is proposed in this paper.With transverse deviation e and deviation rate ec as the input,in the inner loop,PID controller parameter change delta Kp,delta Ki,delta Kd as the output of fuzzy controller,establish two input three output fuzzy control system model;in the outer loop,the PID control system model is established by taking the pulse quantity U of the control servo motor as the output quantity.At the same time,according to the real-time collection and statistical analysis of the size of the welding seam deviation,we proposed the idea of integral separation,in order to meet the requirements of response speed and tracking accuracy of the system.By this way,the seam tracking process is divided into fuzzy adaptive PD control,a half of fuzzy adaptive PID control and fuzzy adaptive PID control,which can reduce the amount of overshoot,eliminate the steady-state error and improve the response speed.It can also avoid larger oscillation when the system from PD control convert to PID control,so that the system can smooth transition and improve the control precision.Finally,we used Simulink to simulate the fuzzy PID controller and the welding seam tracking system,and design experiments.The results show that the mean and limiting filter algorithm and the fuzzy adaptive PID control algorithm based on integral separation can effectively improve the welding seam tracking accuracy.