Interest of process monitoring and numerical simulation to improve metallic Additive Manufacturing processes

Additive Manufacturing (AM) is a promising manufacturing technology compared to subtractive processes, in terms of cost and freedom of manufacturing, depending on the type of the component. Among all the AM techniques, Direct Energy Deposition (DED) processes are dedicated to functional metallic parts manufacturing. As DED processes are complex, there is a real need to predict acceptable operating strategies prior to manufacturing and to control the process in real-time. In this regard, numerical simulation and in-situ monitoring are the most widely used solutions. Concerning simulation, some numerical models focused on specific physical aspects of DED processes in order to improve the understanding of the occurring phenomena during manufacturing, but they hardly enabled to simulate a manufacturing of real parts. Thermo-mechanical Finite Element (FE) models have been applied to DED processes, but their predictions are currently limited to few layers manufacturing; model complexity and calculation time being major obstacles to simulate more complex parts. Concerning monitoring, most developed strategies have focused on only one aspect of manufacturing, mainly thermal or geometrical, using single-sensor control systems. Nevertheless, as DED processes involve coupled phenomena, single monitoring systems cannot provide a comprehensive control over a wide range of manufacturing conditions. Consequently, few multi-sensor monitoring strategies have also been proposed, but were hardly developed and adapted to several cases and processes. The present works propose to use both monitoring and numerical simulation solutions to enhance DED processes performances. To this end, a novel multiple monitoring methodology dedicated to DED processes is under developments at the GeM institute, currently coupling thermal and geometrical in-situ controls. In this paper, a simplified 3D FE thermal simulation of the WAAM process is presented and compared to in-situ monitored data. The corresponding results show that even if the numerical model effectively captures experimental trends, it is not sufficient to fully master the manufacturing. Thus, there is a need to perform in-situ monitoring alongside numerical simulation in order to guarantee acceptable final parts.