Predictive microstructure distribution and printability maps in laser powder bed fusion for a Ni–Cu alloy

Abstract

The solidification microstructure of a melt pool under additive manufacturing conditions is highly heterogeneous due to the heterogeneity in the thermal spatio-temporal fields. This work combines a finite element (FE)-based thermal model with a phase field model (PFM) to predict microstructure distribution among the process parameter span in LPBF, which is strongly controlled by local thermal histories. The segregation distribution across the parameter space can be classified into four different microstructure distribution types: (i) fully planar, (ii) bottom dendritic, (iii) top dendritic, and (iv) fully dendritic. Also, the relationship between the thermal histories (the temperature gradient (G) and the growth rate (R)) variation induced by P and V→ and the microstructure distribution is clearly analyzed in the paper. For a Ni-20 at.%Cu alloy, the predicted microstructural distribution is verified experimentally. The parameter space is further divided into homogeneous and heterogeneous regions using the predicted area fraction of cellular–dendritic segregation across the melt pools. The process map is then used to build AM parts with homogeneous microstructures, where only planar microstructure is found experimentally. This methodology will aid in the exploitation of the alloy and processing space to identify alloy-process combinations that yield microstructurally-homogeneous, defect-free parts, provided an unconditionally printable regime can be identified.