Time-dependent subgrid scales in residual-based large eddy simulation of turbulent channel flow

The present study investigates the effect of taking into account the time dependency of the fine (subgrid) scales in a residual-based variational multiscale approach for large eddy simulation. The residual-based variational multiscale method with time-dependent (dynamic) subgrid scales is presented,...

Full description

Saved in:
Bibliographic Details
Published in:Computer methods in applied mechanics and engineering 2010-02, Vol.199 (13), p.819-827
Main Authors: Gamnitzer, Peter, Gravemeier, Volker, Wall, Wolfgang A.
Format: Article
Language:English
Subjects:
Computational techniques
Exact sciences and technology
Flows in ducts, channels, nozzles, and conduits
Fluid dynamics
Fundamental areas of phenomenology (including applications)
Large eddy simulation
Mathematical methods in physics
Physics
Residual-based variational multiscale method
Time-dependent subscales
Turbulence simulation and modeling
Turbulent channel flow
Turbulent flows, convection, and heat transfer
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The present study investigates the effect of taking into account the time dependency of the fine (subgrid) scales in a residual-based variational multiscale approach for large eddy simulation. The residual-based variational multiscale method with time-dependent (dynamic) subgrid scales is presented, and the impact of the time dependency is studied for the well-known test case of turbulent channel flow. Results are presented from computations for various values of the Reynolds number, namely Re τ = 180 , Re τ = 395 and Re τ = 590 , and several time-step sizes. A generalized- α time-integration scheme is employed. Results from our numerical experiments with dynamic subgrid scales are compared to results obtained with an approximation not explicitly taking the time-dependency of the subgrid scales into account. For all Re τ values and time-step sizes under consideration, results for resolved quantities computed by both approaches are very similar. This statement applies to both, mean streamwise velocity and root-mean-square velocity fluctuations. However, it provides a model for the subgrid-scales not depending on the time-step size and enabling a more robust representation of unresolved scales. Thus we conclude that the time-dependent subgrid-scale approximation is not capable of producing more accurate results for this type of flow if the time-step size is chosen within an optimal range. However, we expect it to be more advantageous for more complex problems, since our results indicate that it provides a more reliable representation of unresolved scales.
ISSN:0045-7825
1879-2138
DOI:10.1016/j.cma.2009.07.009