Graduate Seminar Series

March 03, 2023

10:00 a.m. ET

7500 Wean Hall

On Physics-informed Neural Networks for Grain Boundary Dynamics

Nearly all structural and functional materials are polycrystalline systems; they are composed of crystalline grains that are internally joined at grain boundaries (GBs). Even minute amounts of dopants at GBs result in profound changes to GB dynamics, specifically grain growth during processing treatments or under high-temperature operating conditions. While GB solute segregation has been the subject of active research, most existing studies focus on the thermodynamics of GB segregation and the kinetic role (i.e., dynamic solute drag) remains unexplored. The challenge here is that GB solute drag depends on several properties, such as alloy thermodynamics (e.g., heat of mixing), and kinetic processes including GB solute diffusion and boundary migration; solute drag is a hypersurface. In this talk, we present physics-informed neural networks (PINNs) to investigate GB solute drag in regular solution metallic alloys. The PINN model not only predicts the solute drag hypersurface, but also honors the nonlinear governing equations describing GB segregation and solute transport. Representative PINN results are presented to reveal new insights about the roles of asymmetric GB segregation and solute-solute interactions in drag effects. In broad terms, our modeling approach provides avenues to detangle the thermodynamic and kinetic roles of GB segregation in grain coarsening of a wide range of metallic alloys. 

Fadi Abdeljawad, Ph.D, Assistant Professor, Mechanical Engineering/Materials Science and Engineering, Clemson University

Fadi AbeljawadFadi Abdeljawad is the Bob and Kay Stanzione Assistant Professor and Dean’s Faculty Fellow in the Department of Mechanical Engineering with a joint appointment in the Department of Materials Science and Engineering at Clemson University. Prior to joining Clemson in 2018, he was a Staff Scientist (2014- 2018) in the Computational Materials and Data Science Department at Sandia National Laboratories. Abdeljawad obtained his M.A. (2010) and Ph.D. (2014) from Princeton University with a primary focus on theoretical and computational materials science. He is the recipient of the 2022 TMS Early Career Faculty Fellow Award, 2019 Ralph E. Powe ORAU Award, 2022 Clemson Junior Researcher of the Year Award, and 2020 Byars Prize for Excellence in Teaching Award. Abdeljawad’s research is funded by DOD, DOE, NSF, and Sandia National Laboratories.