Thermo-mechanical simulations of powder bed fusion processes: accuracy and efficiency

In this contribution, the accuracy and efficiency of various modeling assumptions and numerical settings in thermo-mechanical simulations of powder bed fusion (PBF) processes are analyzed. Thermo-mechanical simulations are used to develop a better understanding of the process and to determine residual stresses and distortions based on the temperature history. In these numerically very complex simulations, modeling assumptions are often made that reduce computational effort but lead to inaccuracies. These assumptions include the omission of the surrounding powder or the use of geometrically linearized material models. The numerical setting, in particular the temporal and spatial discretizations, can further lead to discretization errors. Here, a highly parallelized and adaptive finite element method based on the open source C++ library deal.II is validated and utilized, to investigate some of these modeling assumptions and to identify the required temporal and spatial discretizations for the simulation of PBF of Ti-6Al-4V. The insights initially gained on a simple wall-like geometry are transferred to a larger open rectangular profile where the results of a detailed simulation are compared with those of a more efficient one. The results for the efficient approach show a maximum deviation of approximate to 8% in the displacements and approximate to 3.5% in the residual stresses while significantly reducing the computational time.