Despite well-known limitations of Reynolds-averaged Navier–Stokes (RANS) simulations, this methodology remains the most widely used tool for predicting many turbulent flows due to computational efficiency. Machine learning is a promising approach to improve the accuracy of RANS simulations. One major area of improvement is using machine learning models to represent the complex relationship between the mean flow field gradients and the Reynolds stress tensor. In the present work, modifications to improve the stability of previous optimal eddy viscosity approaches for RANS simulations are presented and evaluated. The optimal eddy viscosity is reformulated with a non-negativity constraint, which promotes numerical stability. We demonstrate that the new formulation of the optimal eddy viscosity improves conditioning of RANS equations for a periodic hills test case. To demonstrate the suitability of this proportional/orthogonal tensor decomposition for use in a physics-informed data-driven turbulence closure, we use two neural networks (structured on this specific tensor decomposition, which is incorporated as an inductive bias into the network design) to predict the newly reformulated linear and non-linear parts of the Reynolds stress tensor. Injecting these network model predictions for the Reynolds stresses into RANS simulation improves predictions of the velocity field, even when compared to a sophisticated (state of the art) physics-based turbulence closure model. Finally, we apply shapley additive explanations values to obtain insights from the learned representation for inner workings of the neural network used to predict the optimal eddy viscosity from the input feature data.

1.
F. D.
Witherden
and
A.
Jameson
, “
Future directions of computational fluid dynamics
,” in
23rd AIAA Computational Fluid Dynamics Conference
(
AIAA
,
2017
), pp.
1
16
.
2.
J.
Slotnick
,
A.
Khodadoust
,
J.
Alonso
, and
D.
Darmofal
, “
CFD vision 2030 study: A path to revolutionary computational aerosciences
,”
Report No. NF1676L-18332
(
2014
).
3.
K.
Duraisamy
, “
Perspectives on machine learning-augmented Reynolds-averaged and large eddy simulation models of turbulence
,”
Phys. Rev. Fluids
6
,
050504
(
2021
).
4.
K.
Duraisamy
,
G.
Iaccarino
, and
H.
Xiao
, “
Turbulence modeling in the age of data
,”
Annu. Rev. Fluid Mech.
51
,
357
377
(
2019
).
5.
S. L.
Brunton
,
B. R.
Noack
, and
P.
Koumoutsakos
, “
Machine learning for fluid mechanics
,”
Annu. Rev. Fluid Mech.
52
,
477
508
(
2020
).
6.
J. N.
Kutz
, “
Deep learning in fluid dynamics
,”
J. Fluid Mech.
814
,
1
4
(
2017
).
7.
J. X.
Wang
,
J. L.
Wu
, and
H.
Xiao
, “
Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data
,”
Phys. Rev. Fluids
2
,
034603
(
2017
).
8.
Y.
Zhu
and
N.
Dinh
, “
A data-driven approach for turbulence modeling
,” arXiv:2005.00426 (
2020
).
9.
C.-W.
Chang
and
N. T.
Dinh
, “
Reynolds-averaged turbulence modeling using type I and type II machine learning frameworks with deep learning
,” arXiv:1804.01065 (
2018
).
10.
C.
Jiang
,
R.
Vinuesa
,
R.
Chen
,
J.
Mi
,
S.
Laima
, and
H.
Li
, “
An interpretable framework of data-driven turbulence modeling using deep neural networks
,”
Phys. Fluids
33
,
055133
(
2021
).
11.
S.
Bhushan
,
G. W.
Burgreen
,
W.
Brewer
, and
I. D.
Dettwiller
, “
Development and validation of a machine learned turbulence model
,”
Energies
14
,
1465
(
2021
).
12.
W.
Liu
,
J.
Fang
,
S.
Rolfo
,
C.
Moulinec
, and
D. R.
Emerson
, “
An iterative machine-learning framework for RANS turbulence modeling
,”
Int. J. Heat Fluid Flow
90
,
108822
(
2021
).
13.
P. A.
Srinivasan
,
L.
Guastoni
,
H.
Azizpour
,
P.
Schlatter
, and
R.
Vinuesa
, “
Predictions of turbulent shear flows using deep neural networks
,”
Phys. Rev. Fluids
4
,
054603
(
2019
).
14.
Y.
Yin
,
P.
Yang
,
Y.
Zhang
,
H.
Chen
, and
S.
Fu
, “
Feature selection and processing of turbulence modeling based on an artificial neural network
,”
Phys. Fluids
32
,
105117
(
2020
).
15.
R.
Fang
,
D.
Sondak
,
P.
Protopapas
, and
S.
Succi
, “
Neural network models for the anisotropic Reynolds stress tensor in turbulent channel flow
,”
J. Turbul.
21
,
525
543
(
2020
).
16.
X.
Zhang
,
J.
Wu
,
O.
Coutier-Delgosha
, and
H.
Xiao
, “
Recent progress in augmenting turbulence models with physics-informed machine learning
,”
J. Hydrodyn.
31
,
1153
1158
(
2019
).
17.
L.
Zhu
,
W.
Zhang
,
J.
Kou
, and
Y.
Liu
, “
Machine learning methods for turbulence modeling in subsonic flows around airfoils
,”
Phys. Fluids
31
,
015105
(
2019
).
18.
R.
Matai
and
P. A.
Durbin
, “
Zonal eddy viscosity models based on machine learning
,”
Flow, Turbul. Combust.
103
,
93
109
(
2019
).
19.
Z. M.
Nikolaou
,
C.
Chrysostomou
,
Y.
Minamoto
, and
L.
Vervisch
, “
Evaluation of a neural network-based closure for the unresolved stresses in turbulent premixed V-flames
,”
Flow, Turbul. Combust.
106
,
331
(
2021
).
20.
J.
Tan
,
X.
He
,
G.
Rigas
, and
M.
Vahdati
, “
Towards explainable machine-learning-assisted turbulence modeling for transonic flows
,” in
14th European Conference on Turbomachinery Fluid dynamics & Thermodynamics
,
2021
.
21.
M.
Kaandorp
,
Machine Learning for Data-Driven RANS Turbulence Modelling
(
Delft University of Technology
,
2018
).
22.
M. L.
Kaandorp
and
R. P.
Dwight
, “
Data-driven modelling of the Reynolds stress tensor using random forests with invariance
,”
Comput. Fluids
202
,
104497
(
2020
).
23.
X. D.
Song
,
Z.
Zhang
,
Y. W.
Wang
,
S. R.
Ye
, and
C. G.
Huang
, “
Reconstruction of RANS model and cross-validation of flow field based on tensor basis neural network
,” in
Proceedings of the ASME-JSME-KSME 2019 8th Joint Fluids Engineering Conference
,
2019
.
24.
Z.
Zhang
,
X. D.
Song
,
S. R.
Ye
,
Y. W.
Wang
,
C. G.
Huang
,
Y. R.
An
, and
Y. S.
Chen
, “
Application of deep learning method to Reynolds stress models of channel flow based on reduced-order modeling of DNS data
,”
J. Hydrodyn.
31
,
58
65
(
2019
).
25.
J.
Ling
,
A.
Kurzawski
, and
J.
Templeton
, “
Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
,”
J. Fluid Mech.
807
,
155
166
(
2016
).
26.
J. L.
Wu
,
H.
Xiao
, and
E.
Paterson
, “
Physics-informed machine learning approach for augmenting turbulence models: A comprehensive framework
,”
Phys. Rev. Fluids
3
,
074602
(
2018
).
27.
M. A.
Cruz
,
R. L.
Thompson
,
L. E.
Sampaio
, and
R. D.
Bacchi
, “
The use of the Reynolds force vector in a physics informed machine learning approach for predictive turbulence modeling
,”
Comput. Fluids
192
,
104258
(
2019
).
28.
B. P.
Brener
,
M. A.
Cruz
,
R. L.
Thompson
, and
R. P.
Anjos
, “
Conditioning and accurate solutions of Reynolds average Navier-Stokes equations with data-driven turbulence closures
,”
J. Fluid Mech.
915
,
A110
(
2021
).
29.
J. L.
Wu
,
R.
Sun
,
S.
Laizet
, and
H.
Xiao
, “
Representation of stress tensor perturbations with application in machine-learning-assisted turbulence modeling
,”
Comput. Methods Appl. Mech. Eng.
346
,
707
726
(
2019
).
30.
K.
Hanjalić
,
M.
Popovac
, and
M.
Hadžiabdić
, “
A robust near-wall elliptic-relaxation eddy-viscosity turbulence model for CFD
,”
Int. J. Heat Fluid Flow
25
,
1047
1051
(
2004
).
31.
P. A.
Durbin
, “
Near-wall turbulence closure modeling without ‘damping functions
,’”
Theor. Comput. Fluid Dyn.
3
,
1
13
(
1991
); available at https://link.springer.com/article/10.1007/BF00271513.
32.
R.
McConkey
,
E.
Yee
, and
F. S.
Lien
, “
A curated dataset for data-driven turbulence modelling
,”
Sci. Data
8
,
255
(
2021
).
33.
H.
Xiao
,
J. L.
Wu
,
S.
Laizet
, and
L.
Duan
, “
Flows over periodic hills of parameterized geometries: A dataset for data-driven turbulence modeling from direct simulations
,”
Comput. Fluids
200
,
104431
(
2020
).
34.
R. L.
Thompson
,
G.
Mompean
, and
T.
Laurent
, “
A methodology to quantify the nonlinearity of the Reynolds stress tensor
,”
J. Turbul.
11
,
N33
(
2010
).
35.
S. B.
Pope
,
Turbulent Flows
(
Cambridge University Press
,
Cambridge, UK
,
2000
).
36.
J.
Ho
and
A.
West
, “
Field Inversion and Machine Learning for turbulence modelling applied to three-dimensional separated flows
,” in
AIAA Aviation 2021 Forum
,
2021
.
37.
D. R.
Laurence
,
J. C.
Uribe
, and
S. V.
Utyuzhnikov
, “
A robust formulation of the v2–f model
,”
Flow, Turbul. Combust.
73
,
169
185
(
2005
).
38.
OpenCFD Ltd.
,
OpenFOAM: User Guide v2006
(
OpenCFD Ltd
.,
2019
).
39.
F. S.
Lien
and
G.
Kalitzin
, “
Computations of transonic flow with the v2–f turbulence model
,”
Int. J. Heat Fluid Flow
22
,
53
61
(
2001
).
40.
G.
Klambauer
,
T.
Unterthiner
,
A.
Mayr
, and
S.
Hochreiter
, “
Self-normalizing neural networks
,” in
Advances in Neural Information Processing Systems
(
NIPS
,
2017
), pp.
972
981
.
41.
A.
Géron
,
Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow
, 2nd ed. (
O'Reilly Media Inc
.,
Sebastopol, CA
,
2019
), p.
1150
.
42.
See
F.
Chollet
, https://keras.io for “
Keras
” (
2015
).
43.
A. J.
Spencer
and
R. S.
Rivlin
, “
Isotropic integrity bases for vectors and second-order tensors—Part I
,”
Arch. Ration. Mech. Anal.
9
,
45
63
(
1962
).
44.
F.
Pedregosa
,
G.
Varoquaux
,
A.
Gramfort
,
V.
Michel
,
B.
Thirion
,
O.
Grisel
,
M.
Blondel
,
P.
Prettenhofer
,
R.
Weiss
,
V.
Dubourg
,
J.
Vanderplas
,
A.
Passos
,
D.
Cournapeau
,
M.
Brucher
,
M.
Perrot
, and
E.
Duchesnay
, “
scikit-learn: Machine learning in Python
,”
J. Mach. Learn. Res.
12
,
2825
2830
(
2011
); available at https://jmlr.csail.mit.edu/papers/v12/pedregosa11a.html.
45.
S. M.
Lundberg
and
S.-I.
Lee
, “
A unified approach to interpreting model predictions
,” in
Advances in Neural Information Processing Systems
, edited by
I.
Guyon
,
U. V.
Luxburg
,
S.
Bengio
,
H.
Wallach
,
R.
Fergus
,
S.
Vishwanathan
, and
R.
Garnett
(
Curran Associates, Inc
.,
2017
), Vol.
30
.
46.
S.
Lardeau
and
F.
Billard
, “
Development of an elliptic-blending lag model for industrial applications
,” AIAA Paper No. 2005-522,
2016
.
47.
R.
McConkey
,
E.
Yee
, and
F. S.
Lien
(
2021
). “A curated dataset for data-driven turbulence modeling,”
kaggle
.
You do not currently have access to this content.