Metal Additive Manufacturing Conference - MAMC 2020

Over the last years, additive manufacturing (AM) has transitioned from a system and materials to a complete end-to-end solution business. Printing cost becomes then a major factor to reduce in order to achieve the business goals. There are continuous attempts from additive manufacturing players to reduce material costs and increase process productivity such that AM becomes a key driver of digital manufacturing on a cost per part level. In this context, our paper describes a methodology to optimally choose the most economical support structure design for metal printing in terms of material by guarantying a minimal and acceptable thermal deformations of the printed parts. This methodology consists of three steps: (1) design support structures with different parameters using an available design tool, (2) simulate the thermal deformations and residual stresses using these different support structures , (3) derive a lumped model that generates a pareto front highlighting the optimal design. In order to demonstrate and validate this methodology, two designs of metal parts are made with an increasing complexity of the design. The first design is relatively simple in geometry and allows to cover large thermal variations due to a broad variations of the printing process parameters. The second design is more complex (derived from an industrial part) and allows to cover thermal variations due to complex geometry while maintaining the printing process parameters close to nominal values. Both designs are printed many times using a selective laser printing process where layer per layer variations on printing parameters are performed. The thermal deformations of these different prints are experimentally measured using a 3d scanner and compared to the simulation results. The validation proves a very good predictive power of the simulation to accurately predict the thermal deformations of the printed parts using more economical support structures. This methodology allows the designers to rely into the simulation in order to choose the optimal support structure design without having to perform expensive trial and errors experiments. The method has the potential to predict layer per layer thermal deformations of printed parts allowing thus to adapt the process parameters during the print and avoid scraps due to thermal stresses.