Optimal estimator and artificial neural network as efficient tools for the subgrid-scale scalar flux modeling

This work is devoted to exploring a new procedure to develop subgrid-scale (SGS) models in the context of large-eddy simulation (LES) of a passive scalar. Starting from the Noll's formula (Noll 1967), the concept of an optimal estimator is first used to de- termine an accurate set of parameters to derive a SGS model. The SGS model is then defined as a surrogate model built from this set of parameters by training an artificial neural network (ANN) on a filtered DNS database. This ANN model is next compared with the dynamic nonlinear tensorial diffusivity (DNTD) model proposed by Wang et al. (2007). The DNTD model is also based on Noll's formula, and can be seen as a nonlinear extension of the dynamic eddy-diffusivity (DED) model proposed by Moin et al. (1991). The a priori and a posteriori tests performed on the ANN model demonstrate the abil- ity of this new model to well reproduce the behavior of the exact SGS term, and show an improvement in comparison with DED and DNTD models. The concept of optimal estimator associated with machine learning procedure thus appears as a useful tool for SGS model development

[1]  J. Ferziger,et al.  Evaluation of subgrid-scale models using an accurately simulated turbulent flow , 1979, Journal of Fluid Mechanics.

[2]  K. Alvelius,et al.  RANDOM FORCING OF THREE-DIMENSIONAL HOMOGENEOUS TURBULENCE , 1999 .

[3]  Charles Meneveau,et al.  Universality of large eddy simulation model parameters across a turbulent wake behind a heated cylinder , 2002 .

[4]  Olivier Teytaud,et al.  Optimal estimation for Large-Eddy Simulation of turbulence and application to the analysis of subgrid models , 2006, ArXiv.

[5]  P. Moin,et al.  A dynamic subgrid‐scale model for compressible turbulence and scalar transport , 1991 .

[6]  Guillaume Balarac,et al.  A dynamic regularized gradient model of the subgrid-scale scalar flux for large eddy simulations , 2012 .

[7]  Walter Noll,et al.  Representations of certain isotropic tensor functions , 1970 .

[8]  S. Pope,et al.  Direct numerical simulations of the turbulent mixing of a passive scalar , 1988 .

[9]  F. Ducros,et al.  Subgrid scale variance and dissipation of a scalar field in large eddy simulations , 2001 .

[10]  T. S. Lund,et al.  Parameterization of subgrid-scalestress by the velocity gradient tensorBy , 1992 .

[11]  F. Sarghini,et al.  Neural networks based subgrid scale modeling in large eddy simulations , 2003 .

[12]  R. Moser,et al.  Optimal LES formulations for isotropic turbulence , 1999, Journal of Fluid Mechanics.

[13]  Thomas M. Breuel,et al.  Evaluation of robustness and performance of Early Stopping Rules with Multi Layer Perceptrons , 2009, 2009 International Joint Conference on Neural Networks.

[14]  Q. Zheng.  Theory of Representations for Tensor Functions—A Unified Invariant Approach to Constitutive Equations , 1994 .

[15]  Jose C. F. Pereira,et al.  Analysis of the gradient-diffusion hypothesis in large-eddy simulations based on transport equations , 2007 .

[16]  Parviz Moin,et al.  Erratum: ‘‘A dynamic subgrid‐scale eddy viscosity model’’ [Phys. Fluids A 3, 1760 (1991)] , 1991 .

[17]  Eugene Yee,et al.  A complete and irreducible dynamic SGS heat-flux modelling based on the strain rate tensor for large-eddy simulation of thermal convection , 2007 .

[18]  John G. Proakis,et al.  Probability, random variables and stochastic processes , 1985, IEEE Trans. Acoust. Speech Signal Process..

[19]  P. Sagaut.  BOOK REVIEW: Large Eddy Simulation for Incompressible Flows. An Introduction , 2001 .