Fuzzy control of pulsed GTA welds by using real-time root bead image feedback

Abstract

Abstract The soundness of steel pipes welds is primarily determined by the quality and penetration of the root pass. In order to ensure quality welds with full penetration, the gas tungsten arc welding (GTAW) process is widely used in industry. Pulsed current GTAW can significantly reduce total welding heat input, weld heat-affected zone (HAZ), residual stresses, and welding distortion, but the welding parameters are more complicated. Usually, the quality of manual GTA welding almost solely depends on the experience of individual welders, but the quality of automatic orbital welding is greatly influenced by proper selection and control of welding parameters to overcome the influences from gravity, arc force, surface tension, heat accumulation, joint preparation, and fit up precision, etc. Consequently, an automatic-welding system without feedback control may easily result in poor welding quality, especially in field welding when the joint preparation and fit up precision are far from ideal. In this paper, an automatic pulsed GTAW system with fuzzy control technique was proposed by using a real-time root bead image feedback, and the width of root pass was used as control indication for full penetration. Bead width of the root pass was determined from CCD images signals using the edge detection technique. The rule of fuzzy controller is based on experiences acquired from welding of single-V grooved plates. Experiments show that the welding quality of steel pipes was very successfully by using this nonlinear, multivariable, on-line controlled pulsed GTAW process.