An adaptive smoothed particle hydrodynamics (SPH) scheme for efficient melt pool simulations in additive manufacturing

Smoothed particle hydrodynamics (SPH) is a well-developed mesh-free method that has proven highly efficient for mesoscopic additive manufacturing (AM) simulations. Nevertheless, all current three-dimensional SPH models in this ever-growing field of application are based on a uniform discretization size (i.e., single-resolution domain) and cannot exploit the computational efficiency of this method from an algorithmic point of view. In this work, we present a spatially fully-adaptive SPH scheme and apply it for the first time to simulate the melt pool behavior in laser powder bed fusion (LPBF) additive manufacturing. The implementation contains state-of-the-art numerical stabilization techniques to ensure the SPH stability and robustness, as well as all crucial physical phenomena, such as the recoil pressure, wetting, and Marangoni effects, to capture the melt pool dynamics in detail. Full spatial adaptivity is enabled by particle splitting and merging, where the spatial resolution can be refined and coarsened concurrently multiple times. As a result of this and a novel sorting algorithm, the code performs about 5x faster in powder-based AM applications, making it feasible to simulate multi-track LPBF processes within reasonable times without parallel computing for the first time.