Numerical 3D simulation of the preheating step during electron beam melting

Within the context of additive manufacturing techniques, the powder bed fusion with electron beam (PBF-EB) process is characterized by its distinctive preheating phase. The diffusive phenomena at the atomic level triggered by the heat generated within the preheating step result in solid connections between powder particles. These material bridges, better known as sintering necks, assume a central role in the feasibility of the PBF-EB process, since they prevent the unsafe smoke phenomenon and control thermal and electrical conductivities. The neck growth, and in general the overall preheating step, is commonly optimized by a manual “trial-and-error” procedure although this is known to be wasteful in time, energy and raw material. The focus of scientific research is increasingly shifting towards modeling the preheating phase, thus enabling new scenarios for studying the phenomenon of necking. Among the various modeling techniques such as Monte Carlo and Finite Elements (FE), the Phase-Field (PF) method is preferred for its capability to solve free boundary problems such as the solid-state sintering one. Most of the current literature about sintering at the typical time and temperature evolution of the PBF-EB, consider sintering in 2D. The computational limit and the absence of a clear and robust formulation for the neck measurement make the tridimensional scenario appealing but at the same time apparently unsolved This master thesis addresses the problems related to a 3D PF simulation of sintering at the PBF-EB conditions, focusing on the 3D neck growth. Initially, the neck growth at each PF simulation timestep was traced looking at the evolution of the field variable c with a manual images-based approach. Then, neck growth detections were automated with the introduction of the novel NeckValue function. This approach was validated against the two-dimensional literature with a maximum percentage relative error below 4%. Later, it was extended to the 3D case. The challenge of three-dimensional computational effort was accepted by modelling portions of the powder bed with 1/32 particles of Ti6Al4V . The effect of the third dimension was first evaluated by comparing the 3D necking phenomenon with the 2D neck growth. Subsequently, the 3D neck growth was investigated by performing an analysis by reference/tilted planes with the aim to observe possible directional inhomogeneities. This entirely novel methodology was tested on multiple 3D models and applied to a case study which replicates the powder bed. The results of this work show the feasibility of simulating the sintering at PBF-EB conditions in 3D, opening new perspectives on a more realistic model to describe the preheating phase.