Contribution to Bead Width Using Welding Torch Feedback Control with Real-time AI Discrimination

TIG arc welding is often used because this welding has high quality and strength. However, when the distance between electrodes is changed, the arc length and arc voltage are displaced, and the heat transfer rises or falls, causing welding defects. Therefore, it is necessary to adjust the interelectrode distance, although there is a limit to how much an artisan can adjust and control it. The welding torch was controlled longitudinally using feedback control based on the results of the classification. The bead width generated at this time was also calculated. Specifically, the image of the arc captured by the camera was classified by supervised learning, and the image was classified by a predetermined labeling. Based on this, the welding torch was driven and controlled to keep the interelectrode distance constant. The measured arc voltage confirmed the displacement of the interelectrode distance, and the resulting bead width confirmed the change in heat input to the base metal with AI control of the welding torch. As a result, the AI discrimination controlled the interelectrode distance to be constant and suppressed the variation of bead width.