Multigrid methods for large-eddy simulation

A. Gordner, S. Nägele, Gabriel Wittum

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The Large-Eddy simulation (LES) method can be used in order to break the multi-scale complexity within turbulent flow simulations, since not all turbulent length scales have to be resolved, but will be given by an appropriate subgrid model. Beside this filtering in space, a filtering in time allows for larger time steps as well and gives rise to implicit methods, where an algebraic system of equations has to be solved. Multigrid as a numerical multi-scale approach matches LES quite well with that respect and will be applied. It is essential to control the numerical error introduced by the discretisation and numerical solver in order to minimize the influence on the turbulent solution and hence, being able to identify the model error of the subgrid model. Two different stabilization methods, that are used within the collocated Finite Volume dicretisation for unstructured grids, are investigated with respect to their mass conservation error. The obtained solutions will be compared with benchmark solutions found in literature. The used subgrid model takes advantage of mesh dependent parameters. A practical solution within the multigrid procedure is to derive the model parameter on the finest grid level and inject it successively to the coarser grid levels. By this strategy good convergence rates result.

Original languageEnglish (US)
Title of host publicationReactive Flows, Diffusion and Transport
Subtitle of host publicationFrom Experiments via Mathematical Modeling to Numerical Simulation and Optimization Final Report of SFB (Collaborative Research Center) 359
PublisherSpringer Berlin Heidelberg
Pages375-396
Number of pages22
ISBN (Print)354028379X, 9783540283799
DOIs
StatePublished - Dec 1 2007

ASJC Scopus subject areas

  • Mathematics(all)

Fingerprint Dive into the research topics of 'Multigrid methods for large-eddy simulation'. Together they form a unique fingerprint.

Cite this