Light refreshments will be served at 4:15pm
Author: A.D. Rollett
Collaborators: E.A. Holm, B. DeCost, H. Jain, L. Scime†, J. Beuth†, T. Ozturk, R. Cunningham, C. Kantzos, J. Pauza, J. Aroh, P.-J. Chiang, R. Jiang, Z. Wu, N. Parab*, T. Sun*, C. Zhao*
Dept. Materials Sci. & Eng., Carnegie Mellon Univ., Pittsburgh PA
†Mechanical Eng. Dept., Carnegie Mellon Univ., Pittsburgh PA
*Advanced Photon Source, Argonne Natl. Lab., Chicago, IL
3D Printing, Porosity, Synchrotron Experiments and Machine Learning
3D printing of metals has advanced rapidly in the past decade and is used across a wide range of industry. Many aspects of the technology are considered to be well understood in the sense that machines make parts and temperature histories with residual stress can be simulated. At the microscopic scale, however, more work is required to quantify, understand and predict defect- and micro-structures, which affect properties such as fatigue resistance. The most dramatic synchrotron-based technique is dynamic x-ray radiography (DXR) which provides ultra-high speed imaging of laser melting of metals and their powders. This has, e.g., enabled the keyhole phenomenon to be quantified, which in turn has demonstrated the importance of power density, as opposed to energy density. This latter quantity, while informative, also fails to capture the crucial boundary between full density and lack-of fusion porosity because it does not take account of melt pool overlap. Synchrotron-based 3D X-ray computed microtomography (µXCT) showed that essentially all metal powders exhibit porosity that partially persists into the printed metal. This explanation is reinforced by evidence both DXR and simulation. To illustrate the power of machine learning, Computer vision (CV) has successfully classified different microstructures, including powders. The power of CV is further demonstrated by its ability to detect and classify defects in the spreading of powder. High speed synchrotron x-ray diffraction is beginning to provide new information on solidification and phase transformation in, e.g., IN718, Ti-6Al-4V and stainless steel. High Energy (x-ray) Diffraction Microscopy (HEDM) experiments also is also providing data on 3D microstructure and local elastic strain in 3D printed materials such as Ti-6Al-4V and stainless steel.