Machine learning-augmented turbulence modeling for RANS simulations of massively separated flows

We combine data assimilation and machine learning to correct the RANS Spalart-Allmaras turbulence model. The final neural-network contribution is a Boussinesq-correction, rather than a turbulent eddy-viscosity adjustment. Flows over periodic hills at distinct Reynolds numbers and geometries were selected to demonstrate the potential gain of machine learning-augmented turbulence models.

[1]  Karthik Duraisamy,et al.  Machine Learning-augmented Predictive Modeling of Turbulent Separated Flows over Airfoils , 2016, ArXiv.

[2]  T. Hughes,et al.  Streamline upwind/Petrov-Galerkin formulations for convection dominated flows with particular emphasis on the incompressible Navier-Stokes equations , 1990 .

[3]  Paola Cinnella,et al.  Quantification of model uncertainty in RANS simulations: A review , 2018, Progress in Aerospace Sciences.

[4]  T. Hughes,et al.  Stabilized finite element methods. I: Application to the advective-diffusive model , 1992 .

[5]  Arne V. Johansson,et al.  An explicit algebraic Reynolds stress model for incompressible and compressible turbulent flows , 2000, Journal of Fluid Mechanics.

[6]  Heng Xiao,et al.  A Priori Assessment of Prediction Confidence for Data-Driven Turbulence Modeling , 2017, Flow, Turbulence and Combustion.

[7]  Karthik Duraisamy,et al.  Turbulence Modeling in the Age of Data , 2018, Annual Review of Fluid Mechanics.

[8]  Christopher L. Rumsey,et al.  Verification and Validation of a Second-Moment-Closure Model , 2016 .

[9]  Michael A. Leschziner,et al.  Approximate near-wall treatments based on zonal and hybrid RANS–LES methods for LES at high Reynolds numbers , 2006 .

[10]  P. Cinnella,et al.  Large Eddy Simulation Requirements for the Flow over Periodic Hills , 2019, Flow, Turbulence and Combustion.

[11]  J. Templeton.  Evaluation of machine learning algorithms for prediction of regions of high Reynolds averaged Navier Stokes uncertainty , 2015 .

[12]  Karthik Duraisamy,et al.  A paradigm for data-driven predictive modeling using field inversion and machine learning , 2016, J. Comput. Phys..

[13]  G. Tryggvason,et al.  Using statistical learning to close two-fluid multiphase flow equations for bubbly flows in vertical channels , 2016 .

[14]  Jean Rabault,et al.  Artificial neural networks trained through deep reinforcement learning discover control strategies for active flow control , 2018, Journal of Fluid Mechanics.

[15]  Sanjay Mittal,et al.  On the performance of high aspect ratio elements for incompressible flows , 2000 .

[16]  Peter J. Schmid,et al.  A data-assimilation method for Reynolds-averaged Navier–Stokes-driven mean flow reconstruction , 2014, Journal of Fluid Mechanics.

[17]  J. Templeton,et al.  Reynolds averaged turbulence modelling using deep neural networks with embedded invariance , 2016, Journal of Fluid Mechanics.

[18]  Christophe Eric Corre,et al.  Subgrid-scale scalar flux modelling based on optimal estimation theory and machine-learning procedures , 2017 .

[19]  J. Fröhlich,et al.  Investigation of wall-function approximations and subgrid-scale models in large eddy simulation of separated flow in a channel with streamwise periodic constrictions , 2003 .

[20]  Peter J. Schmid,et al.  Data assimilation of mean velocity from 2D PIV measurements of flow over an idealized airfoil , 2017 .

[21]  Yan Bao,et al.  Numerical prediction of aerodynamic characteristics of prismatic cylinder by finite element method with Spalart-Allmaras turbulence model , 2011 .

[22]  F. Schmitt.  About Boussinesq's turbulent viscosity hypothesis: historical remarks and a direct evaluation of its validity , 2007 .

[23]  Data-Driven Integral Boundary-Layer Modeling for Airfoil Performance Prediction in Laminar Regime , 2017 .

[24]  Steven L. Brunton,et al.  Deep model predictive flow control with limited sensor data and online learning , 2020, Theoretical and Computational Fluid Dynamics.

[25]  L. Temmerman,et al.  A hybrid two-layer URANS–LES approach for large eddy simulation at high Reynolds numbers , 2005 .

[26]  Vincent Couaillier,et al.  On the use of a high-order discontinuous Galerkin method for DNS and LES of wall-bounded turbulence , 2017, Computers & Fluids.

[27]  Julia Ling,et al.  Machine learning strategies for systems with invariance properties , 2016, J. Comput. Phys..

[28]  Jochen Fröhlich,et al.  Highly resolved large-eddy simulation of separated flow in a channel with streamwise periodic constrictions , 2005, Journal of Fluid Mechanics.

[29]  A. Stuart,et al.  Ensemble Kalman methods for inverse problems , 2012, 1209.2736.

[30]  Jinlong Wu,et al.  Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data , 2016, 1606.07987.

[31]  L. Franca,et al.  Stabilized finite element methods. II: The incompressible Navier-Stokes equations , 1992 .

[32]  Frédéric Hecht,et al.  New development in freefem++ , 2012, J. Num. Math..

[33]  J. Alonso,et al.  SU2: An Open-Source Suite for Multiphysics Simulation and Design , 2016 .

[34]  Heng Xiao,et al.  Flows over periodic hills of parameterized geometries: A dataset for data-driven turbulence modeling from direct simulations , 2019, Computers & Fluids.

[35]  Heng Xiao,et al.  Representation of stress tensor perturbations with application in machine-learning-assisted turbulence modeling , 2019, Computer Methods in Applied Mechanics and Engineering.

[36]  G. Tryggvason,et al.  Using statistical learning to close two-fluid multiphase flow equations for a simple bubbly system , 2015 .

[37]  Michael Manhart,et al.  Flow over periodic hills – Numerical and experimental study in a wide range of Reynolds numbers , 2009 .

[39]  Heng Xiao,et al.  Regularized ensemble Kalman methods for inverse problems , 2020, J. Comput. Phys..

[40]  G. Balarac,et al.  A wall-layer model for large-eddy simulations of turbulent flows with/out pressure gradient , 2011 .

[41]  T. Poinsot,et al.  Training convolutional neural networks to estimate turbulent sub-grid scale reaction rates , 2018, Combustion and Flame.

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

[43]  Koji Fukagata,et al.  Synthetic turbulent inflow generator using machine learning , 2018, Physical Review Fluids.

[44]  Stephen J. Wright,et al.  Numerical Optimization , 2018, Fundamental Statistical Inference.

[45]  Steven L. Brunton,et al.  Deep Model Predictive Control with Online Learning for Complex Physical Systems , 2019, ArXiv.

[46]  S. Pope.  A more general effective-viscosity hypothesis , 1975, Journal of Fluid Mechanics.

[47]  Shia-Hui Peng,et al.  Hybrid LES‐RANS modelling: a one‐equation SGS model combined with a k–ω model for predicting recirculating flows , 2003 .

[48]  H. Aluie,et al.  The direct enstrophy cascade of two-dimensional soap film flows , 2013, 1309.4894.