Abstract Details
Turbulent Mixing, Transport and Subgrid Models
Author: James Glimm
Requested Type: Oral Only
Submitted: 2009-04-20 12:04:12
Co-authors:
Contact Info:
Stony Brook University
Nichols Road
Stony Brook, NY 11794-3
US
Abstract Text:
We combine two conventional methods of turbulent mixing
modeling. More precisely, we combine two distinct themes and add additional
ideas. The results allow accurate simulation of macroscopic mixing
as well as microscopic or molecular level mixing, with
feasible grid resolution, well short of full DNS resolution.
We combine the dynamic subgrid models proposed by Moin and
coworkers with the high resolution methods favored by the capturing
community, and to this we add our front tracking approach, which carries
the high resolution futher. We improve the front tracking, to allow
non zero mass diffusion across an interface, but only as given by the
physical transport parameters, without numerical diffusion.
These ideas have been verified in an extensive 2D numerical study
of a Richtmyer-Meshkov instability including reshock. For this
problem, the macroscopic variables such as the
mixing zone edges are shown to be insensitive to physical and
numerical modeling of laminar and turbulent transport, but the
molecular level mixing observables such as the joint pdf of temperature
and concentration, or the chemical reaction rate of a temperature sensitive
process, is very sensitive to these effects. For the problem considered,
the chemical reaction rate is subject to statistical fluctuations even
after a spatial average, a fact which obscures and interfers with a
mesh convergence study.
At least in a form without the subgrid models, these ideas
have been validated by the simulation of 3D Rayliegh-Taylor instabiities,
in agreement with experiment. In this case both the micro and the macro
observables (including the famous mixing rate parameter alpha) are
sensitive to laminar and turbulent physical and numerical modeling issues.
We are pleased to acknowledge the joint efforts of
a team of collaborators from Stony Brook University, Los Alamos
National Laboroatory, and Brookhaven National Laboratory, and
especially of many very talented students who carried out much of the
reported work.
Comments: