Page 25 - MSAM-1-4

Materials Science in Additive Manufacturing Process optimization of SEBM IN718 via ML

To accelerate the application of the SEBM, the defects. The beam current and scan speed were used as
generation of defects during the print processing must the input of GPR and support vector regression (SVR)
be solved in the first place since macroscopical cracks to establish the relationship between different processing
preferentially nucleate from defects region, which affect parameters and relative density, and the optimized
the overall property of materials . The types of internal processing window is determined. Then, four high-density
[13]
defects reported generally in SEBM process include Inconel 718 samples with different microstructures and
[14]
entrapped gas porosity, lack of fusion porosity , shrinkage properties were built according to the processing window.
porosity, and hot cracking . In Ni-based superalloys, The relationship among SEBM processing parameters,
[16]
[15]
the formation of detrimental intermetallic phases, such defect, microstructure, and property of Inconel 718 was
as laves phases, δ phases, carbides, and nitrides, is also a studied.
kind of defect, because they reduce the number of elements
used for solid solution and γ′/γ′′ precipitation [17-20] . 2. Methodology
Although some post-processing treatment methods, such
as hot isostatic pressing (HIP), could reduce the defect 2.1. Inconel 718 powder and SEBM process
density, not all defects can be repaired [21,22] . Therefore, it Pre-alloyed powders were supplied by Guangzhou
is of great significance to control defect generation during Sailong Additives Manufacturing. Co., Ltd., Guangzhou,
AM process by optimizing processing parameters. Various China. The chemical composition of powder is listed
researchers have optimized processing parameters through in Table 1. Figure 1 shows the surface morphology and
experiment, such as taguchi method , energy density , microstructure of the Inconel 718 alloy powder. The
[24]
[23]
processing window and dimensionless number , and powders mostly exhibit a spherical shape (>90%), only few
[25]
[26]
performed physical calculation simulation [27-29] to build satellites (red arrow) and non-spherical powders (white
defect-free parts with a flat surface successfully. However, arrow) were observed, as indicated in Figure 1A. On the
the processing parameters optimization is still a big smooth surface of the powder, dendrite structure can be
challenge due to the high-dimensionality and complex seen, as shown in Figure 1B. Inconel 718 powder consists
combination of parameters, during which complex physical of fine equiaxed grains with average grain size of 5.8 μm
processes and interactions also need to be considered.
Recently, with the development of materials A B
informatics, machine learning method has been broadly
adopted to facilitate composition and process optimization
for complex alloys [30-36] . Liu et al. developed a machine
[37]
learning approach based on Gaussian process regression
(GPR) to identify the processing window for AlSi10Mg
alloy by laser powder bed fusion, wherein the fully dense
alloy with high strength and ductility was manufactured.
Aoyagi et al. proposed a simple method that combines C
[38]
uniform design and support vector machine to correlate
processing parameters with surface conditions and
generated a processing map that can obtain the best
densification for SEBM CoCr alloy by fewer samples.
Thereafter, Lei et al. [39,40] optimized multiple processing
parameters of SEBM for superalloy Alloy713ELC. Sah
et al. trained multiple machine learning algorithms to
[41]
predict the density and defect formation in LBPF sample.
Figure 1. Surface morphologies and microstructures of pre-alloyed
In this study, the most commonly used superalloy Inconel 718 powder. (A) SEM in low magnifications. (B) SEM high
Inconel 718 was selected to investigate the effects of different magnifications. (C) IPF map are based on EBSD measurements with a
parameters on the surface morphology and internal step size of 0.3 μm.

Table 1. Chemical composition of Inconel 718 powder.
Ni Cr Fe Nb Mo Ti Al Co Cu Si C O N
Bal. 21 17.2 5.12 3.21 0.85 0.45 0.2 0.089 0.039 0.032 0.0149 0.0146

Volume 1 Issue 4 (2022) 2 https://doi.org/10.18063/msam.v1i4.23