Welding Seam Tracking System For Narrow Gap Laser Welding Based On Robot

Robot welding technology has great advantages over traditional manual welding in terms of welding efficiency and welding quality.With the rapid development of technology,robot welding technology is more and more widely used in the manufacturing industry,and robot welding technology has become an inevitable trend in the transformation of enterprises in the industrial field.As the key technology of intelligent welding,seam tracking plays a vital role in the welding process.The development and application of visual sensing technology has laid a foundation for obtaining accurate welding seam position information.The intelligent combination of vision sensing and welding robot improves the stability and accuracy of welding seam tracking technology.In this thesis,research on the seam tracking system for narrow gap laser welding based on robot has been carried out.Due to the characteristics of narrow-gap plates,this thesis uses passive vision for image acquisition,and installs the line array camera coaxially with the welding head to avoid pre-errors.This thesis first introduces the research background and significance of the seam tracking system for narrow gap laser welding based on robots,summarizes the research status of seam tracking technology,and introduces the main research content of the subject.Secondly,it introduces the system design and principle,including vision system design,upper computer system and software design,robot welding system design,and explains the purpose of coaxial design.For the collected plate image,the gray value curve image is selected to represent it,and the corresponding feature analysis is carried out,and preprocessing work such as obtaining the ROI region of interest and image filtering is carried out on the image.Aiming at the gray value curve image,a method of finding the undetermined area of the weld is improved,and a variety of feature screening conditions are improved to extract the straight and curved weld positions.A general method for effectively identifying straight and curved welds within the field of view using a line scan camera is derived.By optimizing the weld seam recognition algorithm and selecting appropriate parameter thresholds,the accuracy of characteristic condition recognition is improved.Using a simplified system calibration method,the conversion relationship between pixel coordinates and robot space coordinates is obtained,and the least square method is used to perform polynomial fitting on the straight line and curved weld path,and the welding process is tracked through KRL programming,and the acceleration and deceleration control method is used to plan the path speed.The software interface is designed,the system is visualized,and the system communication design is completed,the communication between each device is configured,and the data transmission channel is established.Finally,this thesis builds the welding experiment platform,selects the appropriate camera acquisition speed to collect images,and introduces the welding experiment preparation and the required process of the experiment.Weld recognition experiments are carried out on plates,and the experimental results show that the recognition algorithm can accurately extract the position information of welds under different conditions.The tracking welding experiment was carried out,and the metallographic structure of the cross-section of straight and curved welds was analyzed.The experiments show that the seam tracking system in this paper has a good tracking welding effect,and the straight and curved seam tracking welding path can completely cover the position of the welding seam,and there is no obvious deviation between the welding area and the seam position,which proves that the system is feasible.