Process simulation for laser-assisted additive manufacturing

Abstract

Physics-based simulation models for laser-based additive manufacturing processes become increasingly important due to the high complexity of the underlying coupled physical phenomena and the large (and increasing) number of tunable process parameters and materials of interest for applications. To gain understanding of the underlying coupled multiphysical problem, the onset of defects, and the consequence of process parameter settings, material choice and their interaction, simulation models with high predictive capabilities are essential. For this purpose, a high fidelity, physics-based simulation model for the prediction of laser-based manufacturing processes was developed, including process-specific model additions that enable the simulation of the most widespread metal additive manufacturing process, laser-based powder bed fusion. Starting from the basic principles of conservation of mass, momentum and energy, a multiphysical framework for the simulation of laser-based material processing at the scale of the melt pool is developed and validated against experimental results for different processing conditions and materials. The simulation framework is then further expanded by incorporating process-relevant features such as non-equilibrium solidification, multi-layer recoating in powder bed fusion, weakly coupled solid mechanics and grain envelope tracking. Furthermore, the model’s ability to predict liquid film cavitation within the mushy zone during solidification is turned into a new metric for hot crack susceptibility and validated. The simulation model is applied to reduce crack susceptibility of non-weldable nickel-based alloys in an application-oriented study. This is achieved via introduction of a secondary, low-intensity laser beam aiming at a cooling rate reduction. The predictive capabilities of the here-developed model are shown to outperform the previous state of the art due to increased coupling between condensed and gaseous matter, especially during laser-induced evaporation and condensation events. Building and extensively validating a high-fidelity simulation framework that does not need empirical calibration, and incorporating multiphysical and multiscale aspects inherent to AM processes, this thesis presents important groundwork on the way towards digital process twins for laser-based additive manufacturing.