About this Abstract |
Meeting |
2022 TMS Annual Meeting & Exhibition
|
Symposium
|
Materials Design and Processing Optimization for Advanced Manufacturing: From Fundamentals to Application
|
Presentation Title |
A Machine Learning Based Methodology to Predict the Build Quality of Metallic Alloys Additively Manufactured by Laser Powder Bed Fusion |
Author(s) |
Jeongmin Woo, Kevin Graydon, Yongho Sohn |
On-Site Speaker (Planned) |
Jeongmin Woo |
Abstract Scope |
Metallic alloys additively manufactured by laser powder bed fusion (LPBF) can suffer from flaw formation either due to lack-of-fusion or keyhole porosity, corresponding to insufficient or excessive transfer, respectively, of energy generated by scanning laser. Relationships among volumetric energy density of LPBF, density of samples produced and thermophysical properties of alloys were examined using the experimental results from Mg (WE43), Al (Al10SiMg), Ti (Ti6Al4V), Fe (316SS, 15-5PH), Ni (IN718, IN625) alloys. A constitutive function based on relevant thermally-activated processes were employed to model the variation in sample density as a function of LPBF volumetric energy density. Coefficients of the functions were determined using a genetic algorithm, and their relations to thermophysical properties of metallic alloys were examined using machine learning. Findings are presented with respect to progress towards prediction methodologies for identifying optimal LPBF parameters for metallic alloys, both commercially available and to be designed in the future. |
Proceedings Inclusion? |
Planned: |
Keywords |
Additive Manufacturing, Modeling and Simulation, Computational Materials Science & Engineering |