|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Computational Method and Experimental Approaches for Model Development and Validation, Uncertainty Quantification, and Stochastic Predictions
||Errors of Molecular Dynamics Simulations, and Development of “Accurate” Analytical Bond Order Potentials for Al-Cu-H and Mg-H Systems
||Xiaowang Zhou, Brandon C. Wood, Foster E. Michael, Mark. D. Allendorf, Tae Wook Heo, Shinyoung Kang
|On-Site Speaker (Planned)
Modern computer clusters have enabled the uncertainty error of thermal noises and time/length scale approximations of molecular dynamics simulations to be reduced to near zero for some problems including calculations of diffusivities, dislocation core energies, and finite-temperature elastic constants. Current methods, however, still cannot rigorously quantify the error of interatomic potentials, let along systematically reducing such an error. In our potential development, we emphasize on capturing the crystalline growth of the equilibrium phase during molecular dynamics vapor deposition simulations. This ensures the lowest energy for the equilibrium phase as compared to any other configurations including infinite amorphous structures. Such a rigor cannot be achieved by fitting any finite number of phases. With the crystalline growth captured, we then fit the property trends of many phases as normally done in literature. As examples, we demonstrate our recent development of the analytical bond order potentials for Al-Cu-H and Mg-H systems.
||Planned: Supplemental Proceedings volume