Morphological control in desktop manufacturing by distributed attraction optimization methods

Rapid prototyping techniques employing thermal material deposition are intended for direct fabrication of functional metal products and tooling. This article introduces an analytical model of such material transfer processes, using superposition of unit deposition distributions, composed of elementary ellipsoid primitives. The real-time surface geometry model, with its parameters identified by in-process profile measurements, is used for Smith-prediction of the material shape in the inaccessible deposition region. 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 based on dynamic geometry shaping by a real-time complex optimization technique, using distributed attraction of the mass source. The methodology is tested in geometric command following on a robotic fused wire deposition PAW station, using optical sensing by a scanning laser stripe.