Optimizing thermo-mechanical processing and material coupling parameters in numerical modeling for additive manufacturing

Additive manufacturing (AM) is gaining increasing industrial interest. Initially conceived to facilitate the pre-production, to manufacture efficiently and cheaply unique parts such as prototypes, it is now able to deliver parts that meet industrial production needs. In addition, AM is an effective way to achieve parts designed using topological optimization. Laser beam melting is an AM technique able to produce specific parts with mechanical properties matching industrial expectations. However, efficient production remains complex because of distortion, cracking and other failures linked to the process and to the machine parameters. The prime material being generally expensive and in order to achieve the required manufacturing quality from the very first attempt while reducing the processing time to market necessary for low cost mass production, a high-quality digital simulation is mandatory. We thus study the relation between the process and the material parameters with the final mechanical state of the part using a numerical model, which is developed in parallel. This work focuses on the macroscopic scale. In order to carry out the thermo-mechanical study we use a finite element resolution method on the whole domain defined by the part, its supports and the baseplate. At this scale, one can neglect the packing of the powder as well as the hydrodynamic behavior of the melt pool in the laser beam melting process. We consider a Gaussian energy distribution for the heat source, imposed on several layers of powders below the deposited layer. Starting from a previous study [1], which considers the temperature dependence of the Young modulus, we improve, in this work, the model by considering the temperature dependency of other physical parameters pertaining to material properties or to elastoplastic and thermal laws coefficients. The aim is to bridge the macroscopic scale study of AM to the mesoscopic scale from results of this study. In a first step, simulations are carried out with simple models like walls, cubes, beams and the popular cantilever used for calibrations. The results show the gain in precision with the contribution of the temperature dependence of various parameters as well as by considering the phase-transition during the printing process. The computing time is compatible with laptops and can iterate the simulations easily. The accuracy of the model is validated by comparing the distortion to the experimental results.