Neuro-fuzzy control of the process of feeding wire material in additive technologies of the FDM family

Abstract

FDM-based 3D-printing is one of the most common kinds of additive technologies. A typical printing unit consisting of a material feeding device and a printing head has a number of significant drawbacks associated with the lack of control over the basic parameters of extrusion. Basically, the only controlled parameter is the heating temperature. The article considers an alternative point of view on the problem, based on the multiagent approach and on elements of fuzzy and neural-fuzzy logic, which allows, through the introduction of additional feedbacks and simultaneous consideration of several basic parameters, to improve the quality of material extrusion, and to reduce the influence of the wire-material quality on the final 3D-printing result.