| One of the recent and most spectacular advances in agile manufacturing technology is the development of Rapid Prototyping (RP) or Desktop Manufacturing methods. The limitation with the current RP techniques is their utilization of custom materials. Metals are more suitable for prototypes or customized tools. In this direction, a new RP technique was studied and developed, based on robotic Plasma Arc Welding material deposition by cold wire feeding. Proper guidance of this spot material addition in adjacent 1-D beads and overlaying 2-D layers generates the desired solid 3-D geometry. In such a thermal material deposition method, the solid part is developed by combined heat and mass transfer mechanisms, determining the composite prototype quality. The resulting metallurgical microstructure is of paramount significance for functional metal prototypes, where material properties comparable to those of cast or molded parts are desired. To succeed in achieving such a favorable material properties distribution, each level of material should be heat treated after its deposition. The heat treatment is done in a closed-loop fashion, using the recently developed scanned thermal process where a welding torch sweeps in a fast, repetitive motion the whole area of interest on the workpiece and provides at each location the heat needed, dictated by the control algorithm. The necessary temperature feedback is given from selected surface locations using non-conduct infrared sensing. The inherent nonlinearity of heat transfer mechanisms and the limitations of infrared thermal sensing lead to the establishment of a linearized multiple-input, multiple-output model, with in-process identification of its parameters. The thermal regulation system adjusts the power and guides the motion of the torch to the part region with the largest deviation from the desired temperature distribution, using two different in-process thermal optimization methods, the complex optimization and the simulated annealing optimization. Both those methods were successfully implemented in computer simulations and real-time experiments, using a Robotic Plasma Arc Welding experimental workstation. |
