Mid Sweden University

Additive Manufacturing (AM) uses thin layers of material to form a 3D object, and during recent years, the development has rapidly increased with more studies on the subject being made. One AM branch is Powder Bed Fusion (PBF) which uses an electron beam or laser to melt and combine powder material into a solid part. Common materials are metals or polymers. One PBF technique is Electron Beam Melting (EBM). EBM uses a high-power electron beam to melt a metal powder, and the process is made in a vacuum so a high temperature and controlled manufacturing can be used. The research project, SELEKT-AM, uses EBM to control material properties by varying the microstructure in a material. This thesis continues the SELEKT-AM project and aims to analyse the microstructure in the melt pools z-axis and, from the results, create a new experiment to research the subject further. The study uses nine sections of a start plate made with 304 stainless steel and melts ten lines on each section with EBM Arcam A2. The lines have different settings of beam speed, beam current, focus offset, multibeam and overlap. The samples were divided with a cross-section orthogonal to the beam, prepared and then characterised with SEM Maia3 to measure all the values for smelt depth and width of the melt pool. A numerical analysis of the values from the characterisation was made with Minitab. The result shows that when beam speed is increased from 200 to 2000 mm/s, and the other variables were constant, the smelt pool depth decreased linearly by −18% / (mm/s). The width was quadratically increased between 600−1000 mm/s with the linear variation below and above that. When changing the beam current between 2 to 20 mA and keeping the other variable constant, the smelt pool depth increased linearly by 21% / mA, and width increased linearly by 36 % /mA. Focus offset increased the width linearly by 7 % / mA for 600 mm/s and 3% / mA for 1000 mm/s and decreased the smelt depth by −13% / mA for 600 mm/s and −9 % / mA for 1000 mm/s. The result also shows that overlap and multibeam offsets the graph. From the result, a new experiment was designed. The number of repetitions and factorial levels to perform a 23-factorial experiment was set. The factor levels were set to a high and low value for beam speed, beam current and focus offset, and the data will be analysed using ANOVA and Pareto or a normal probability plot. The scan path will be a single line smelt with Ti6Al4V and use seven, eight, nine, ten or eleven repetitions, depending on the wanted statistical significance. The thesis shows that beam speed and current are the parameters that affected the result the most; however, since only one parameter is changed per section, the interaction cannot be analysed. Further studies on that subject are therefore needed. The aim and the study's goals are met, and an experiment design is developed to analyse the microstructural properties in EBM machining so further control of the material properties can be made.