Research On Robotic Welding System And Multipass Planning Based On Laser Vision Sensor

Thick plate workpiece is widely used in construction machinery, shipbuilding industries and so on. At present, such workpieces usually adopt manual or semi-automatic welding. Due to the harsh working environment, it is urgent to replace the welders with welding robots. Unlike thin plate, it is difficult for robot to automatically adjust path-planning in multi-pass welding, especially when the work piece has misalignment. Thus, seam tracking and adaptive path-planning are the key technologies to realize robot automatic for thick plate workpiece.This paper designed an intelligentized robotic welding system based on laser vision sensor, industry robot, welding robot, welding machine, wire feeder and so on. An image processing method is developed to recognize feature points of weld seam in the captured image. To minimize the interference from arc light and spatter, this method extracted a region of interest and calculated the mass center of laser stripe in this area. Due to the projection relationship, the second derivative of laser stripe peaks at the edges of seam. This method recognize this peak and creates the 3D reconstruction of the weld seam edges. An experiment is conducted and the result shows that the precision of seam tracking is less than 0.8mm.In this study, the author also developed an adaptive path planning algorithm to change the welding parameters according to the real situation of weld seam. The main aim of this algorithm is to control the formation of the seam when the weld joint gap or the width of groove root changed. The algorithm changes the weld speed and weaving amplitude according to the width of seam gap. This algorithm also calculate the position and attitude adjustment of weld torch. A variable gap multipass welding experiment is conducted and weld appearance of each layer shows that the developed method meets the practical welding requirements.