3D multi-layer grain structure simulation of powder bed fusion additive manufacturing

TLDR

In powder bed fusion additive manufacturing, laser or electron beam melting of powder layers, together with hatching strategies and beam parameters, determines the grain structure, yet the precise influence of individual parameters on microstructure remains largely unknown, making accurate prediction of mechanical properties challenging. The study adapts a 3‑D dendritic growth model for predicting grain structure in powder bed fusion. The model computes heat input via an analytical solution of transient heat conduction and leverages massively parallel HPC processing to simulate grain structure of small parts efficiently. The model accurately reproduces experimental grain structures of Inconel 718 specimens, captures grain selection zones and microstructure over several millimeters with unprecedented detail, and provides a cutting‑edge tool for reliably predicting beam parameters for arbitrary applications.

Abstract

In powder bed fusion (PBF) additive manufacturing, powder layers are locally melted with a laser or an electron beam to build a component. Hatching strategies and beam parameters as beam power, scan velocity and line offset significantly affect the grain structure of the manufactured part. While experiments reveal the result of specific parameter combinations, the precise impact of distinct parameters on the resulting grain structure is widely unknown. This knowledge is necessary for a reliable prediction of the microstructure and consequently the mechanical properties of the manufactured part. We introduce the adaption of a three-dimensional model for the prediction of dendritic growth for use with PBF. The heat input is calculated using an analytical solution of the transient heat conduction equation. Massively parallel processing on a high-performance cluster computer allows the computation of the grain structure on the scale of small parts within reasonable times. The model is validated by accurately reproducing experimental grain structures of Inconel 718 test specimens manufactured by selective electron beam melting. The grain selection zone within the first layers as well as the subsequent microstructure in several millimeters build height is modeled in unprecedented level of detail. This model represents the cutting-edge of grain structure simulation in PBF and enables a reliable numerical prediction of appropriate beam parameters for arbitrary applications.