Optimizing geometrical control of surface distributions in rapid prototyping deposition

Rapid prototyping techniques employing thermal material deposition are intended for direct layered fabrication of functional products. This article introduces an analytical geometric model of such material transfer processes, using superposition of unit deposition distributions, composed of elementary ellipsoidal primitives. The real-time surface geometry model, with its deposition features identified by in-process height measurements, is used for Smith-prediction of the material shape in the deposition region that is inaccessible to sensing. Thus, a distributed-parameter geometry control scheme is established to obtain a desired surface topology, by modulating the feed and motion of a moving mass source. This is obtained by dynamic surface shaping based on minimization of the geometric error-distribution by robust real-time optimization techniques.,A moving constrained polytope and a simulated annealing method at random sampled Surface locations are employed for this purpose. These methodologies are compared experimentally in geometric command following a robotic fused wire deposition plasma-arc welding station, using laser profilometric scanner sensing.