"Stellar Turbulence Modeling with Adaptive Mesh Refinement: Assessment of Performance of Select Refinement Strategies"
Wednesday, Mar 19, 2025
Abstract:
Turbulent flows are ubiquitous in Earth-bound and astrophysical physical systems, and numerical simulations of such flows features prominently in such varied applications as weather forecasting, studies of convective plumes within the Earth's mantle, and modeling of supernova progenitors. Such studies require significant amounts of computational resources and benefit greatly from advances in computational methods, specifically in the case of terrestrial combustion studies, defense applications, problems in planetary sciences and astrophysics, the use of Adaptive Mesh Refinement (AMR), has achieved particular success.
AMR leverages the fact that computational effort to obtain certain solution accuracy may not be uniform throughout the computational domain. Thus, AMR schemes selectively place finer-resolution grids in regions corresponding to high solution error, and coarser grids in regions corresponding to lower solution error. Such discontinuous refinement seeks to compute solutions on the same order of accuracy as uniform meshes, but much more efficiently. In the case of realistic fluid flows, computation of exact solution error is unachievable, and thus various proxies for solution error must be used, resulting in a wide variety of mesh refinement schemes, using solution quality estimators to selectively refine and coarsen the grid.
Despite the widespread implementation of AMR methods using various refinement schemes in computational fluid dynamic codes, the effect of these discontinuous discretizations on the statical properties of turbulent flows has largely been unexamined, potential having a deleterious effect on the accuracy of numerical solution studied using such codes. In this talk, we present our findings on the effect of various AMR refinement schemes on select key statistical metrics of turbulence in a variety of weakly-to-moderately compressible flows under conditions typical of select astrophysical systems.
| File | Description | File size |
|---|---|---|
 RyanLearnResearch.jpg | 1732 kB |
