Recent progress on reliability assessment of large-eddy simulation

Bernard J. Geurts (Corresponding author), Amirreza Rouhi, Ugo Piomelli

Research output: Contribution to journalArticleAcademicpeer-review

1 Citation (Scopus)

Abstract

Reliability assessment of large-eddy simulation (LES) of turbulent flows requires consideration of errors due to shortcomings in the modeling of sub-filter scale dynamics and due to discretization of the governing filtered Navier–Stokes equations. The Integral Length-Scale Approximation (ILSA) model is a pioneering sub-filter parameterization that incorporates both these contributions to the total simulation error, and provides user control over the desired accuracy of a simulation. It combines an imposed target for the ‘sub-filter activity’ and a flow-specific length-scale definition to achieve LES predictions with pre-defined fidelity level. The performance of the ‘global’ and the ‘local’ formulations of ILSA, implemented as eddy-viscosity models, for turbulent channel flow and for separated turbulent flow over a backward-facing step are investigated here. We show excellent agreement with reference direct numerical simulations, with experimental data and with predictions based on other, well-established sub-filter models. The computational overhead is found to be close to that of a basic Smagorinsky sub-filter model.

Original languageEnglish
Article number102615
Number of pages13
JournalJournal of Fluids and Structures
Volume91
Early online date3 May 2019
DOIs
Publication statusPublished - Nov 2019

Fingerprint

Large eddy simulation
Turbulent flow
Direct numerical simulation
Channel flow
Parameterization
Viscosity

Cite this

@article{79b416c0595e418a9ce11dc55a60e9a9,
title = "Recent progress on reliability assessment of large-eddy simulation",
abstract = "Reliability assessment of large-eddy simulation (LES) of turbulent flows requires consideration of errors due to shortcomings in the modeling of sub-filter scale dynamics and due to discretization of the governing filtered Navier–Stokes equations. The Integral Length-Scale Approximation (ILSA) model is a pioneering sub-filter parameterization that incorporates both these contributions to the total simulation error, and provides user control over the desired accuracy of a simulation. It combines an imposed target for the ‘sub-filter activity’ and a flow-specific length-scale definition to achieve LES predictions with pre-defined fidelity level. The performance of the ‘global’ and the ‘local’ formulations of ILSA, implemented as eddy-viscosity models, for turbulent channel flow and for separated turbulent flow over a backward-facing step are investigated here. We show excellent agreement with reference direct numerical simulations, with experimental data and with predictions based on other, well-established sub-filter models. The computational overhead is found to be close to that of a basic Smagorinsky sub-filter model.",
author = "Geurts, {Bernard J.} and Amirreza Rouhi and Ugo Piomelli",
year = "2019",
month = "11",
doi = "10.1016/j.jfluidstructs.2019.03.008",
language = "English",
volume = "91",
journal = "Journal of Fluids and Structures",
issn = "0889-9746",
publisher = "Elsevier",

}

Recent progress on reliability assessment of large-eddy simulation. / Geurts, Bernard J. (Corresponding author); Rouhi, Amirreza; Piomelli, Ugo.

In: Journal of Fluids and Structures, Vol. 91, 102615, 11.2019.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Recent progress on reliability assessment of large-eddy simulation

AU - Geurts, Bernard J.

AU - Rouhi, Amirreza

AU - Piomelli, Ugo

PY - 2019/11

Y1 - 2019/11

N2 - Reliability assessment of large-eddy simulation (LES) of turbulent flows requires consideration of errors due to shortcomings in the modeling of sub-filter scale dynamics and due to discretization of the governing filtered Navier–Stokes equations. The Integral Length-Scale Approximation (ILSA) model is a pioneering sub-filter parameterization that incorporates both these contributions to the total simulation error, and provides user control over the desired accuracy of a simulation. It combines an imposed target for the ‘sub-filter activity’ and a flow-specific length-scale definition to achieve LES predictions with pre-defined fidelity level. The performance of the ‘global’ and the ‘local’ formulations of ILSA, implemented as eddy-viscosity models, for turbulent channel flow and for separated turbulent flow over a backward-facing step are investigated here. We show excellent agreement with reference direct numerical simulations, with experimental data and with predictions based on other, well-established sub-filter models. The computational overhead is found to be close to that of a basic Smagorinsky sub-filter model.

AB - Reliability assessment of large-eddy simulation (LES) of turbulent flows requires consideration of errors due to shortcomings in the modeling of sub-filter scale dynamics and due to discretization of the governing filtered Navier–Stokes equations. The Integral Length-Scale Approximation (ILSA) model is a pioneering sub-filter parameterization that incorporates both these contributions to the total simulation error, and provides user control over the desired accuracy of a simulation. It combines an imposed target for the ‘sub-filter activity’ and a flow-specific length-scale definition to achieve LES predictions with pre-defined fidelity level. The performance of the ‘global’ and the ‘local’ formulations of ILSA, implemented as eddy-viscosity models, for turbulent channel flow and for separated turbulent flow over a backward-facing step are investigated here. We show excellent agreement with reference direct numerical simulations, with experimental data and with predictions based on other, well-established sub-filter models. The computational overhead is found to be close to that of a basic Smagorinsky sub-filter model.

UR - http://www.scopus.com/inward/record.url?scp=85065037610&partnerID=8YFLogxK

U2 - 10.1016/j.jfluidstructs.2019.03.008

DO - 10.1016/j.jfluidstructs.2019.03.008

M3 - Article

AN - SCOPUS:85065037610

VL - 91

JO - Journal of Fluids and Structures

JF - Journal of Fluids and Structures

SN - 0889-9746

M1 - 102615

ER -