2022 年 63 巻 2 号 p. 176-184
This work examined the behaviors of densification and microstructural formation of the Al–10 mass%Si–0.35 mass%Mg alloy fabricated by Selective Laser Melting (SLM) method on the basis of experimental work and machine learning. Additionally, the effect of scanning repeated twice in each layer (double scanning) in the SLM process was also investigated. The SLM-ed Al–10 mass%Si–0.35 mass%Mg alloy exhibited the columnar grained microstructure with a (α-Al–Si) eutectic cell structure. Refined microstructures were produced at an increasing scanning speed with a decreasing the energy density (J/mm3). Relative density tended to increase with an increasing of energy density for scan pitch conditions of 0.1 mm and 0.05 mm. And a scattering was obviously exhibited at a higher relative density more than 95%. The analysis based on machine learning revealed that a scanning pitch of 0.2 mm was just a condition to achieve a high relative density. Except for the condition at a scanning pitch of 0.2 mm, a scan speed was the most important factor in affecting the relative density. Thus, a machine learning approach enabled to identify the important processing factor for affecting the behavior quantitatively. Additionally, compared to a conventional single scanning process, it was found in this work that the double scanning resulted in a higher relative density with keeping the fine microstructural formation.
This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 84 (2020) 365–373. Figure 2 and Fig. 6 were slightly modified. Caption of Fig. 7 was slightly modified.