Traditional welding methods with a sequentially moving, localized heat input, offer limited simultaneous control over the multiple features of weld quality and productivity. This research will focus on the invention of a novel parallel technique, scan welding, implemented by fast repetitive cycling of the torch along the weld surface to provide a regulated continuous heat distribution. This new method will produce the full bead length simultaneously, and yield longitudinally uniform thermal properties of the weld. The dynamics of scan welding will be studied by analytical and numerical modeling of its temperature, phase and stress-strain field. On this basis, a distributed adaptive control system with in-process parameter identification will be designed to modulate the torch power and motion, so as to obtain the specified thermal field. Scan welding will also be implemented on a robotic plasma welding system with thermal feedback from an infrared pyrometer, and experiments will be performed for model verification and comparison of scan welding to conventional techniques. The expected advantages include high process speed and efficiency, elimination of preheat and annealing requirements, and close simultaneous control of weld bead geometry, material microstructure and properties, and residual stresses and distortions of the joint. These benefits provide the potential for optimization of the welding productivity and quality, especially in critical and highly demanding applications.
