Research on weld pool control of welding robot with computer vision

Purpose – The control of weld penetration in gas tungsten arc welding (GTAW) is required for a “teach and playback” robot to overcome the gap variation in the welding process. This paper aims to investigate this subject. Design/methodology/approach – This paper presents a robotic system based on the real‐time vision measurement. The primary objective has been to demonstrate the feasibility of using vision‐based image processing to measure the seam gap in real‐time and adjust welding current and wire‐feed rate to realize the penetration control during the robot‐welding process. Findings – The paper finds that vision‐based measurement of the seam gap can be used in the welding robot, in real‐time, to control weld penetration. It helps the “teach and playback” robot to adjust the welding procedures according to the gap variation. Research limitations/implications – The system requires that the seam edges can be accurately identified using a correlation method. Practical implications – The system is applicable to storage tank welding of a rocket. Originality/value – The control algorithm based on the knowledge base has been set up for continuous GTAW. A novel visual image analysis method has been developed in the study for a welding robot.