A design for additive manufacturing tool for parts obtained through a material extrusion process

Additive Manufacturing (AM) is the most disruptive technology of Industry 4.0 to date, capable of overturning traditional production paradigms. It is a real revolution since production no longer involves the removal of material, but rather starting with a (virtual) 3D model and then 'printing' it layer by layer. To design parts that will be manufactured by AM processes is needed to think additively by changing completely the point of view compared to the design of the past. Some numerical tools to support design are making themselves available, but they hardly are connected with the tools already existing. Moreover, the commercial numerical tools to simulate the additive manufacturing processes can give as output the deviations of the obtained part from the nominal CAD model, but they are not able to translate this into the geometric dimensioning and tolerancing deviations of the part’s surfaces. The present work overcomes this limitation by showing a design for an Additive Manufacturing tool, i.e., a virtual model that makes talking new and old tools to predict the accuracy of parts manufactured by AM techniques in terms of dimensional and geometric deviations of parts’ surfaces. The developed virtual tool assists the designer during all the steps of product development, from design to production and inspection. The accuracy was evaluated by considering different geometric deviations, not just the form deviations, together with the dimensional ones. The virtual model estimates the accuracy by simulating the AM process. Once virtually printed, the model of a manufactured part is converted into a skin model shape representation to which least squares algorithms are applied to define dimensional and geometrical deviations of the boundary surfaces of the part. A material extrusion process and ABS material were considered to experimentally validate the virtual model. The numerical results are in good agreement with the experimental ones, the difference between numerical and experimental results is under 3%.