Conference Logo ProgramMaster Logo
Conference Tools for MS&T25: Materials Science & Technology
Login
Register as a New User
Help
Submit An Abstract
Propose A Symposium
Presenter/Author Tools
Organizer/Editor Tools

About this Abstract

Meeting MS&T25: Materials Science & Technology
Symposium Applications of Uncertainty Quantification (UQ) in Science and Engineering
Presentation Title Sparse grids for magneto-hydrodynamics
Author(s) Miroslav Stoyanov, Hoang Tran
On-Site Speaker (Planned) Miroslav Stoyanov
Abstract Scope Tasmanian is a library for Uncertainty Quantification developed at Oak Ridge National Laboratory, specializing in sparse grid surrogate modeling, Bayesian inference, and optimization. This library has been applied to a variety of fields, including creating surrogates for melt-pool shape geometry, studying plasma-material interactions and magneto-hydrodynamics (MHD) in plasma physics, as well as analyzing fracture mechanics in solids. Tasmanian provides a diverse array of tools that cater to problems with varying levels of smoothness in model outputs concerning model inputs, ranging from very smooth to discontinuous, and accommodating different computational demands, from a few minutes to several days on cluster computers. In addition, Tasmanian works on all current exascale supercomputers, utilizing extreme concurrency and GPU acceleration. We present a cross study of different surrogate techniques for MHD models using very limited number of samples, where Tasmanian sparse grids approach shows excellent accuracy in capturing the full model dynamics.

OTHER PAPERS PLANNED FOR THIS SYMPOSIUM

A case study of Bayesian parameter estimation for thermal property inference and uncertainty quantification
Leveraging Archival Additively Manufacturing Fatigue Data to Investigate the Role of Processing Porosity with Greater Precision
Representative microstructure for macro-scale property prediction using multi-scale models
Sparse grids for magneto-hydrodynamics
Uncertainty Quantification via Deep Kernel Learning for Predicting Multimodal β-phase Volume Fraction from SXRD Patterns

Questions about ProgramMaster? Contact programming@programmaster.org | TMS Privacy Policy | Accessibility Statement