Multivariable Robust Control of the Scan Welding Process

Scan welding is a new joining technique with distributed-parameter thermal actuation on the part surface by a sweeping heat source, employing thermal monitoring and feedback to implement a desired temperature distribution in the weld. However, since such infinite-dimensional sensing, actuation and control techniques are not yet perfected, an alternative multivariable formulation of the thermal distribution is proposed in this article. An efficient quasilinear simulation model of the temperature field is developed by Green’s function analysis, in which the process nonlinearities, disturbances and alterations are reflected as time-variation of its parameters. Thus, a design methodology for a robust stabilizing controller is developed and validated both by computational simulation and experiments on a plasma processing setup with infrared thermometry feedback. The obtained thermal field ensures favorable material structure and properties of the scan welds.