Robotic seam tracking system based on vision sensing and human-machine interaction for multi-pass MAG welding

Abstract

Abstract In the seam tracking for the metal active gas (MAG) welding of heavy plate, there are many problems, such as the strong arc light due to the large current, spatter and fumes, the losses of centerline information and complex conditions in groove induced by previous passes, etc. The problems often make the characteristic in the images needed for seam tracking indistinct and make it difficult for industrial robots to correct the deviation of multi-pass welding. Regarding this issue, a robotic seam tracking system based on vision sensing and human-machine interaction for the multi-pass MAG welding was proposed. The groove edges were used as reference to tell the deviation and to control it according to a simple, fast and effective control strategy based on human-machine interaction. The deviation could be judged by comparing the left or right groove edge alternatively with the corresponding specified positioning ruler from the real-time images obtained by industrial camera. The operators could adjust the torch position in real time by the hint on the positioning ruler, which followed the control strategy. The accuracy, response rate and smoothness of the system were verified by reasonable path planning in multi-pass MAG welding. By comparing the deviation correction of robot with the actual deviation of weld seam, the accuracy of system was rated, which had an average error less than 0.5 mm and was acceptable. The comparison between the actual weld height and width after correction and the designed height and width showed that the system worked smoothly and steadily and the system could track the weld in real time.