Page 75 - MSAM-1-1

Materials Science in Additive Manufacturing A ML model for AM PSP of Ti64

reduces its direct application for metal AM. To better to challenging high-dimensional data representation
represent the AM microstructure, a combination of with limited quantification of grain morphology metrics
statistical methods needs to be developed. Lu and and statistical evidence . However, effort focused on
[17]
Torquato (1992) provided lineal-path function and cluster homogeneous materials; additional investigation is needed
correlation function, which are widely used to extract for establishing S-P linkages based on the AM materials.
microstructure information of materials and generate Manogharan et al. (2015) illustrated that post-
statistical data that can represent the heterogeneity of processing steps, such as hot isostatic pressing, machining,
[11]
a multiphase structure . Hence, in the present study, and/or surface finishing operations, are required to create
multiple types of statistic models have been used to capture functional metal AM surfaces . Hence, establishing and
[18]
information from AM microstructure. validating novel S-P linkages that are able to capture AM
In the case of P-S linkage, Gan et al. (2019) reported metal part machining behavior accurately are imperative
the challenges in establishing a validated relationship for metal AM development. Zhang et al. (2019) indicated
based on the experimental measurements of multiple that in machining of wrought and AM Ti-6Al-4V, cutting
AM process phenomena across multiple spatial-temporal conditions including cooling strategy, cutting tool
scales . Hence, numerical simulation methods become a geometry, tool coating, and machining parameters strongly
[12]
tool to build connections between microstructure and the affect the machined surface and related mechanical
AM process parameters. Thijs et al. (2010) showed that in properties of metal AM parts . Due to the presence of
[19]
the PBF process, the AM process (e.g., EB-PBF vs. L-PBF) fine microstructure and martensitic phase distribution,
and the corresponding AM processing conditions, such PBF Ti-6Al-4V samples exhibit a higher hardness when
as beam power density and scan strategies, significantly compared to the wrought parts. Hence, a higher wear rate
influence both the overall grain morphology and local was observed during the machining of AM Ti-6Al-4V.
microstructure . This can be attributed to the effects Liu and Shin (2019) found that due to a higher thermal
[13]
of input energy density, which affects the growth gradient and cooling rate when compared with traditional
direction of the elongated grains as a function of build parts, PBF Ti-6Al4V usually shows higher ultimate tensile
[9]
height and the resulting cooling rate that influences the stress, yield stress, and lower elongation rate , indicating
phase transformation of the material. For instance, Li that the conventional machining parameters might need
et al. (2017) noted that the temperature fields on DED to be optimized for AM parts, and a valid S-P linkage
processing created by the moving heat power source and would be critical for AM material research. In addition,
material absorption conditions can directly affect the Edwards and Ramulu (2014) found that inherently large
microstructural evolution due to cyclic thermal processing, temperature gradients in Ti-6Al-4V AM parts result in
which results in complex solidification and “in-build” higher residual stress, which increases with an increase
thermal cycling of previous layers . Due to the significant in AM processing time, that is, number of AM layers .
[20]
[14]
differences between PBF and DED processes, the present Studies have shown that residual stresses are larger along
study focused on Ti-6Al-4V PBF material and machining the scan direction than perpendicular direction due to
properties. the larger thermal gradient, which creates an anisotropic
For structure-property (S-P) linkages, Paulson et al. residual stress distribution, ultimately affecting the
(2019) explored the correlation between grain morphology mechanical and machining behavior of AM parts.
and mechanical properties, such as microhardness, In summary, a reduced-order computational and
tensile, fatigue behavior, and elastic localization, based analytical method is required to thoroughly explore the
on the correlation function . However, considering mechanical and machining behavior for various metallic
[15]
the inherent layer-by-layer characteristic of metal AM AM materials. Such a high-throughput computational data
processing, the resulting heterogeneous microstructure science-based analysis of material structure and resulting
will need to be considered to achieve high accuracy in the material behavior will connect a massive dataset that
S-P correlation function. Hence, Fernandez-Zelaia et al. stores microstructural characteristics to the properties of
(2019) established S-P linkages based on spatial statistical materials to gain critical insights into this complex PSP
metric feature descriptors . However, during the AM linkage. To this end, Matouš et al. (2017) provided several
[16]
processes, different fabrication processes (e.g., L-PBF and ML and deep learning based predictive models on material
EB-PBF) and parameters would lead to drastic differences databases for building PSP linkages and estimating the
in different parts, so a large dataset could challenge the material properties where no experimental data may
conventional methods. Yang et al. (2018) provided S-P be available . In the case of machining research, Leo
[21]
linkages using ML methods with limited success due (2001) provided ML tools that could be further developed

Volume 1 Issue 1 (2022) 3 https://doi.org/10.18063/msam.v1i1.6