skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

This content will become publicly available on May 14, 2020

Title: Sub-grid scale model classification and blending through deep learning

Abstract

In this article we detail the use of machine learning for spatio-temporally dynamic turbulence model classification and hybridization for large eddy simulations (LES) of turbulence. Here, our predictive framework is devised around the determination of local conditional probabilities for turbulence models that have varying underlying hypotheses. As a first deployment of this learning, we classify a point on our computational grid as that which requires the functional hypothesis, the structural hypothesis or no modelling at all. This ensures that the appropriate model is specified from a priori knowledge and an efficient balance of model characteristics is obtained in a particular flow computation. In addition, we also utilize the conditional-probability predictions of the same machine learning to blend turbulence models for another hybrid closure. Our test case for the demonstration of this concept is given by Kraichnan turbulence, which exhibits a strong interplay of enstrophy and energy cascades in the wavenumber domain. Our results indicate that the proposed methods lead to robust and stable closure and may potentially be used to combine the strengths of various models for complex flow phenomena prediction.

Authors:
 [1]; ORCiD logo [1];  [1];  [1]
  1. Oklahoma State Univ., Stillwater, OK (United States)
Publication Date:
Research Org.:
Oklahoma State Univ., Stillwater, OK (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR) (SC-21)
OSTI Identifier:
1593562
Grant/Contract Number:  
SC0019290
Resource Type:
Accepted Manuscript
Journal Name:
Journal of Fluid Mechanics
Additional Journal Information:
Journal Volume: 870; Journal ID: ISSN 0022-1120
Publisher:
Cambridge University Press
Country of Publication:
United States
Language:
English
Subject:
42 ENGINEERING; 97 MATHEMATICS AND COMPUTING; Mathematical Foundations: Computational methods; Turbulent Flows: Turbulence modelling

Citation Formats

Maulik, Romit, San, Omer, Jacob, Jamey D., and Crick, Christopher. Sub-grid scale model classification and blending through deep learning. United States: N. p., 2019. Web. doi:10.1017/jfm.2019.254.
Maulik, Romit, San, Omer, Jacob, Jamey D., & Crick, Christopher. Sub-grid scale model classification and blending through deep learning. United States. doi:10.1017/jfm.2019.254.
Maulik, Romit, San, Omer, Jacob, Jamey D., and Crick, Christopher. Tue . "Sub-grid scale model classification and blending through deep learning". United States. doi:10.1017/jfm.2019.254.
@article{osti_1593562,
title = {Sub-grid scale model classification and blending through deep learning},
author = {Maulik, Romit and San, Omer and Jacob, Jamey D. and Crick, Christopher},
abstractNote = {In this article we detail the use of machine learning for spatio-temporally dynamic turbulence model classification and hybridization for large eddy simulations (LES) of turbulence. Here, our predictive framework is devised around the determination of local conditional probabilities for turbulence models that have varying underlying hypotheses. As a first deployment of this learning, we classify a point on our computational grid as that which requires the functional hypothesis, the structural hypothesis or no modelling at all. This ensures that the appropriate model is specified from a priori knowledge and an efficient balance of model characteristics is obtained in a particular flow computation. In addition, we also utilize the conditional-probability predictions of the same machine learning to blend turbulence models for another hybrid closure. Our test case for the demonstration of this concept is given by Kraichnan turbulence, which exhibits a strong interplay of enstrophy and energy cascades in the wavenumber domain. Our results indicate that the proposed methods lead to robust and stable closure and may potentially be used to combine the strengths of various models for complex flow phenomena prediction.},
doi = {10.1017/jfm.2019.254},
journal = {Journal of Fluid Mechanics},
number = ,
volume = 870,
place = {United States},
year = {2019},
month = {5}
}

Journal Article:
Free Publicly Available Full Text
This content will become publicly available on May 14, 2020
Publisher's Version of Record

Citation Metrics:
Cited by: 2 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

A Machine Learning Strategy to Assist Turbulence Model Development
conference, January 2015

  • Tracey, Brendan D.; Duraisamy, Karthikeyan; Alonso, Juan J.
  • 53rd AIAA Aerospace Sciences Meeting
  • DOI: 10.2514/6.2015-1287

Diffusion Approximation for Two-Dimensional Turbulence
journal, January 1968


Training convolutional neural networks to estimate turbulent sub-grid scale reaction rates
journal, May 2019


An eddy-viscosity subgrid-scale model for turbulent shear flow: Algebraic theory and applications
journal, October 2004


Improved subgrid-scale models for large-eddy simulation
conference, February 2013

  • Bardina, J.; Ferziger, J.; Reynolds, W.
  • 13th Fluid and PlasmaDynamics Conference
  • DOI: 10.2514/6.1980-1357

A proposed modification of the Germano subgrid‐scale closure method
journal, March 1992

  • Lilly, D. K.
  • Physics of Fluids A: Fluid Dynamics, Vol. 4, Issue 3
  • DOI: 10.1063/1.858280

Two-Dimensional Turbulence
journal, January 2012


A neural network approach for the blind deconvolution of turbulent flows
journal, October 2017


Determination of the Smagorinsky–Lilly constant CS
journal, May 1997

  • Canuto, V. M.; Cheng, Y.
  • Physics of Fluids, Vol. 9, Issue 5
  • DOI: 10.1063/1.869251

Subgrid-scale modeling for implicit large eddy simulation of compressible flows and shock-turbulence interaction
journal, October 2014

  • Hickel, Stefan; Egerer, Christian P.; Larsson, Johan
  • Physics of Fluids, Vol. 26, Issue 10
  • DOI: 10.1063/1.4898641

The vortices of two-dimensional turbulence
journal, October 1990


A dynamic subgrid‐scale eddy viscosity model
journal, July 1991

  • Germano, Massimo; Piomelli, Ugo; Moin, Parviz
  • Physics of Fluids A: Fluid Dynamics, Vol. 3, Issue 7
  • DOI: 10.1063/1.857955

Neural networks based subgrid scale modeling in large eddy simulations
journal, January 2003


Turbulence Modeling in the Age of Data
journal, January 2019


Smagorinsky constant in LES modeling of anisotropic MHD turbulence
journal, September 2007

  • Vorobev, Anatoliy; Zikanov, Oleg
  • Theoretical and Computational Fluid Dynamics, Vol. 22, Issue 3-4
  • DOI: 10.1007/s00162-007-0064-z

Generalized Deconvolution Procedure for Structural Modeling of Turbulence
journal, October 2017


High-order methods for decaying two-dimensional homogeneous isotropic turbulence
journal, June 2012


Subgrid-scale scalar flux modelling based on optimal estimation theory and machine-learning procedures
journal, June 2017


Evaluation of scale-aware subgrid mesoscale eddy models in a global eddy-rich model
journal, July 2017


Investigations of data-driven closure for subgrid-scale stress in large-eddy simulation
journal, December 2018

  • Wang, Zhuo; Luo, Kun; Li, Dong
  • Physics of Fluids, Vol. 30, Issue 12
  • DOI: 10.1063/1.5054835

Data-driven Adaptive Physics Modeling for Turbulence Simulations
conference, June 2017

  • Ling, Julia; Kurzawski, Andrew
  • 23rd AIAA Computational Fluid Dynamics Conference
  • DOI: 10.2514/6.2017-3627

Optimal LES formulations for isotropic turbulence
journal, November 1999


A framework for large eddy simulation of Burgers turbulence based upon spatial and temporal statistical information
journal, March 2015

  • LaBryer, A.; Attar, P. J.; Vedula, P.
  • Physics of Fluids, Vol. 27, Issue 3
  • DOI: 10.1063/1.4916132

A simple and stable scale-similarity model for large Eddy simulation: Energy balance and existence of weak solutions
journal, November 2003


Machine-Learning-Augmented Predictive Modeling of Turbulent Separated Flows over Airfoils
journal, July 2017

  • Singh, Anand Pratap; Medida, Shivaji; Duraisamy, Karthik
  • AIAA Journal, Vol. 55, Issue 7
  • DOI: 10.2514/1.J055595

Deep neural networks for data-driven LES closure models
journal, December 2019


The development of algebraic stress models using a novel evolutionary algorithm
journal, December 2017


Structure functions in two-dimensional turbulence
journal, February 1992

  • Grossmann, Siegfried; Mertens, Peter
  • Zeitschrift für Physik B Condensed Matter, Vol. 88, Issue 1
  • DOI: 10.1007/BF01573844

A coupled approximate deconvolution and dynamic mixed scale model for large-eddy simulation
journal, May 2007

  • Habisreutinger, Marc A.; Bouffanais, Roland; Leriche, Emmanuel
  • Journal of Computational Physics, Vol. 224, Issue 1
  • DOI: 10.1016/j.jcp.2007.02.010

Two-dimensional turbulence: a physicist approach
journal, May 2002


Using field inversion to quantify functional errors in turbulence closures
journal, April 2016

  • Singh, Anand Pratap; Duraisamy, Karthik
  • Physics of Fluids, Vol. 28, Issue 4
  • DOI: 10.1063/1.4947045

Ten questions concerning the large-eddy simulation of turbulent flows
journal, January 2004


A dynamic framework for functional parameterisations of the eddy viscosity coefficient in two-dimensional turbulence
journal, February 2017


Multilayer feedforward networks are universal approximators
journal, January 1989


A dynamic eddy-viscosity closure model for large eddy simulations of two-dimensional decaying turbulence
journal, June 2014


Subgrid modelling for two-dimensional turbulence using neural networks
journal, November 2018


An explicit filtering method for large eddy simulation of compressible flows
journal, August 2003

  • Mathew, Joseph; Lechner, Richard; Foysi, Holger
  • Physics of Fluids, Vol. 15, Issue 8
  • DOI: 10.1063/1.1586271

Data-driven deconvolution for large eddy simulations of Kraichnan turbulence
journal, December 2018

  • Maulik, R.; San, O.; Rasheed, A.
  • Physics of Fluids, Vol. 30, Issue 12
  • DOI: 10.1063/1.5079582

Mathematical Perspectives on Large Eddy Simulation Models for Turbulent Flows
journal, June 2004

  • Guermond, J. -L.; Oden, J. T.; Prudhomme, S.
  • Journal of Mathematical Fluid Mechanics, Vol. 6, Issue 2
  • DOI: 10.1007/s00021-003-0091-5

Machine learning strategies for systems with invariance properties
journal, August 2016


Inertial Ranges in Two-Dimensional Turbulence
journal, January 1967


Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
journal, October 2016

  • Ling, Julia; Kurzawski, Andrew; Templeton, Jeremy
  • Journal of Fluid Mechanics, Vol. 807
  • DOI: 10.1017/jfm.2016.615

Dynamic optimization methodology based on subgrid-scale dissipation for large eddy simulation
journal, January 2016

  • Yu, Changping; Xiao, Zuoli; Li, Xinliang
  • Physics of Fluids, Vol. 28, Issue 1
  • DOI: 10.1063/1.4940044

Neural Network Modeling for Near Wall Turbulent Flow
journal, October 2002

  • Milano, Michele; Koumoutsakos, Petros
  • Journal of Computational Physics, Vol. 182, Issue 1
  • DOI: 10.1006/jcph.2002.7146

A novel evolutionary algorithm applied to algebraic modifications of the RANS stress–strain relationship
journal, November 2016


Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data
journal, March 2017


Log-Normal Turbulence Dissipation in Global Ocean Models
journal, February 2018


A turbulent heat flux prediction framework based on tensor representation theory and machine learning
journal, July 2018


Theoretically based optimal large-eddy simulation
journal, October 2009

  • Moser, Robert D.; Malaya, Nicholas P.; Chang, Henry
  • Physics of Fluids, Vol. 21, Issue 10
  • DOI: 10.1063/1.3249754

Approximate deconvolution large eddy simulation of a barotropic ocean circulation model
journal, January 2011


An approximate deconvolution procedure for large-eddy simulation
journal, July 1999

  • Stolz, S.; Adams, N. A.
  • Physics of Fluids, Vol. 11, Issue 7
  • DOI: 10.1063/1.869867

The similarity subgrid stresses associated to the approximate Van Cittert deconvolutions
journal, March 2015


Machine learning methods for turbulence modeling in subsonic flows around airfoils
journal, January 2019

  • Zhu, Linyang; Zhang, Weiwei; Kou, Jiaqing
  • Physics of Fluids, Vol. 31, Issue 1
  • DOI: 10.1063/1.5061693

Super-resolution reconstruction of turbulent flows with machine learning
journal, May 2019

  • Fukami, Kai; Fukagata, Koji; Taira, Kunihiko
  • Journal of Fluid Mechanics, Vol. 870
  • DOI: 10.1017/jfm.2019.238