Development of an automatic orbital welding system with robust weaving width control and a seam-tracking function for narrow grooves

From a broad point of view, seam tracking has been one of the main challenges with respect to welding automation. Several attempts at automation have been successful in seam tracking with a fixed weaving width. As a solution for seam-tracking methods for varying groove widths, visual sensors such as CCD cameras have been adopted. Although these vision-sensing techniques can achieve high accuracy, their weakness is that a well-prepared vision sensor environment is needed to obtain high-quality visual measurements and that these measurements can be easily affected by the significant noises present in industrial areas. This paper presents an alternative approach using an arc sensor to measure weld widths and positions for weld seam tracking. Arc sensing is less sensitive than vision sensors, which are susceptible to environmental noises, and requires no additional costs except for the measurement apparatus for detecting the arc voltage or welding current. In this study, a special measurement device for the arc voltage is also developed to enhance the reliability of the seam-tracking sensor. Based on a well-known seam-tracking algorithm, an automatic weaving width control algorithm is proposed which can predict weld position more accurately. The usefulness of the weaving width control is verified through an example in which it is applied to gas tungsten arc welding (GTAW).