Fast Computational Model for Consistent Thermo-Mechanical Simulations in Laser Powder Bed Fusion Additive Manufacturing on Part-Scale

• Applicant: Professor Dr.-Ing. Christoph Meier
• Subject Area: Mechanics
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 572590752

Metal additive manufacturing (AM) via laser powder bed fusion (LPBF) offers highest production flexibility, almost unlimited freedom of design, and even the potential for point-wise control of microstructure and material properties, thus representing a paradigm shift in production technologies. However, to fully exploit the potential of LPBF, more emphasis must be put on simulation-informed workflows for process, geometry, and alloy design, requiring more elaborate computational modeling approaches to predict thermal residual stresses on part-scale. While there exist various types of heuristic approaches, only scan-resolved models, i.e., thermo-mechanical models consistently representing the laser-scan path, account for the critial correlation between part geometry, scan pattern, temperature evolution and residual stresses. However, due to the multi-scale nature of LPBF, scan-resolved models lead to very high computational costs. Even the most efficient state-of-the-art models have only allowed for simulations on the millimeter-scale so far, which is far below practically relevant part sizes. In the applicant’s preliminary work, a high-performance computing (HPC) model for the thermal problem in LPBF has been developed, which enabled for the first time scan-resolved simulations of the full AM Bench 2022 cantilever benchmark geometry, demonstrating an increase in computational throughput by more than one order of magnitude compared to state-of-the-art models. In contrast to this purely thermal model, the computational costs of thermo-mechanical models are still by orders of magnitude too high, which can be traced back to the implicit nature of the mechanical subproblem, requiring a very large number of time steps to be solved in an iterative manner. The main objective of the FastSAM project is the development of a scan-resolved thermo-mechanical computational model for LPBF, accurate enough for high-resolution predictions of residual stresses and efficient enough for simulations on the scale of realistic part sizes. For this purpose, it aims at novel numerical discretization and solution schemes to tackle the two main performance bottlenecks of existing models by reducing (i) the large number of time steps and (ii) the high evaluation costs per time step caused by the iterative solution procedure of the mechanical problem. These fully adaptive schemes combine the advantages of classical scan-resolved and existing heuristic models in terms of predictive accuracy and computational efficiency. Due to the critical role of residual stresses in LPBF, the developed FastSAM simulation tool will address one of the major obstacles towards resource-efficient, simulation-based workflows in AM industries. In AM research, the FastSAM tool will allow for physical insights at unprecedented time and length scales, which will support process understanding and the simulation-inspired development of new process strategies.

• DFG Programme: Research Grants