Page 94 - MSAM-2-3
P. 94

Materials Science in Additive Manufacturing                        Validation of a novel ML model for AM-PSP



            typical beam  sizes  range  from  0.1  – 0.5  mm,  with layer   scattering and dissipation of electron beams that occur in
            thicknesses between 20 and 75 μm when employing a laser   the atmospheric environment, electron beaming systems
            source in an inert atmosphere, and 0.2 – 1 mm with layer   are operated in a vacuum environment (10  Mbar in the
                                                                                                  −3
            thicknesses of approximately 75 – 125 μm for an electron   chamber and 10  Mbar in the gun column). Compared to
                                                                            −6
            beam source within a vacuum environment. In the DED   laser-based AM processes, the EB-PBF process is operated
            process,  the  typical  beam  sizes  range  from  2  –  4  mm,   at higher working temperatures without the risk of oxygen
            with layer thicknesses between 500 and 1000 μm within   uptake due to excellent thermal isolation and the desired
            a shielding gas or vacuum environment. The parameters   vacuum environment. Higher chamber temperatures
            employed in the DED process result in a coarse surface   lead to lower residual stress compared to laser-based AM
            finish [2-6] . In recent years, several studies have explored   systems, and parts processed using EB-PBF can be used
            the application of machining learning and deep learning   without secondary stress-relieving. Furthermore, the
            methods to create the process-structure-property (PSP)   EB-PBF process involves a preheating step before melting,
            linkages [7-11] . Figure 1 shows the workflow of the Ti-6Al-4V   which partially sinters the powder layer to increase
            process-structure-property linkage study.          thermal conductivity during a melting pass, reducing
                                                               the temperature gradient and preventing the formation
            1.1. PBF AM methods for metals                     of thermal cracks [4,13] . The preheating process also holds

            The electron beam melting (EB-PBF) process is a PBF   powder particles together, which can act as supports.
            AM process known for its ability to produce full-density   Therefore, in the EB-PBF process, support structures are
            functional parts. This is achieved using a high-energy   often employed for heat conduction. After the melting pass,
            electron beam and “hot-bed” metal powder. Throughout   the entire build plate is heated again with an electron beam
            the manufacturing process, the entire build tank is   (higher velocity and lower current) before the next power
            maintained at an elevated temperature under a vacuum   layer is spread. The EB-PBF process includes an “in-built”
            environment. The electron beam is powered by a high-  heat treatment step, where the machine is allowed to cool
            voltage electron beam (30 – 60 kV), as shown in Figure 2.  down to room temperature in an inert atmosphere before
              The EB-PBF can process multiple conductive material   the build chamber is accessed.
            systems such as stainless steel (316), tool steel (H13), Ti-based   Laser PBF (L-PBF) process, also known as selective
            superalloys (Ti-6Al-4V), Ni-based superalloys (Inconel 718),   laser sintering, selective laser melting (SLM), direct
            Co-based superalloys (Stellite 21), low-expansion alloys   metal laser sintering, direct metal laser melting, uses a
            (Invar), hard metal, intermetallic compounds, aluminum,   high-power laser beam to selectively fuse powder within
            copper, niobium, and beryllium [4,12,13] . To eliminate the   an inert atmosphere, as shown in Figure 3. This process




























            Figure 1. This research presents a validated machine learning model to predict process-structure-property linkage in additive manufacturing of Ti-6Al-4V.
            Abbreviations: CNC: Computer numerical control; DED: Directed energy deposition; EB-PBF: Electron beam-powder bed fusion; EBSD:  Electron
            backscattered diffraction analysis; L-PBF: Laser powder bed fusion; PBF: Powder bed fusion; PCA: Principal component analysis; PSP: Process-structure-
            property; SEM: Scanning electron microscopy; XRD: X-ray diffraction analysis.


            Volume 2 Issue 3 (2023)                         2                       https://doi.org/10.36922/msam.0999
   89   90   91   92   93   94   95   96   97   98   99