Abstract

The generalizability of a convolutional encoder–decoder based model in predicting aerodynamic flow field across various flow regimes and geometric variation is assessed. A rich master dataset consisting of 11,000+ simulations including cambered, uncambered, thin, and thick airfoils simulated at varying angles of attack is generated. The various Mach and Reynolds number (Re) chosen allows analysis across compressible, incompressible, low, and high Re flow regimes. Multiple studies are carried out with the model trained on datasets that are categorized based on the aforementioned parameters. In each study, the loss of prediction accuracy by training the model on a larger dataset (generalizability), versus a smaller categorically sorted dataset, is evaluated. Largely disparate flow features across the Re range lead to a 25.56% loss, while the generalization across Mach range led to an average of 23.95% loss. However, flow-field changes induced due to geometric variation exhibited a better generalization potential, through an increased accuracy of 12.4%. The encoder–decoder architecture allows extraction of relevant geometric features from largely different geometries (geometric generalization) providing a better out-of-sample prediction accuracy in comparison to physics-based generalization. It is shown that, through user-informed choice of training data (removal of geometrically similar samples), computational costs incurred in generating training data can be reduced. This is important for the application of such methods in the design optimization of platforms and components that require the analysis of the fluid flows.

References

References
1.
Brunton
,
S. L.
,
Noack
,
B. R.
, and
Koumoutsakos
,
P.
,
2020
, “
Machine Learning for Fluid Mechanics
,”
Annu. Rev. Fluid Mech.
,
52
(
1
), pp.
1
31
. 10.1146/annurev-fluid-010719-060214
2.
Yumer
,
M. E.
,
Asente
,
P.
,
Mech
,
R.
, and
Kara
,
L. B.
,
2015
, “
Procedural Modeling Using Autoencoder Networks
,”
Proceedings of the 28th Annual ACM Symposium on User Interface Software Technology, UIST’15
,
Charlotte, NC
, ACM, pp.
109
118
.
3.
Chen
,
W.
,
Fuge
,
M.
, and
Chazan
,
N.
,
2017
, “
Design Manifolds Capture the Intrinsic Complexity and Dimension of Design Spaces
,”
ASME J. Mech. Des.
,
139
(
5
), p.
051102
. 10.1115/1.4036134
4.
Zhang
,
W.
,
Yang
,
Z.
,
Jiang
,
H.
,
Nigam
,
S.
,
Yamakawa
,
S.
,
Furuhata
,
T.
,
Shimada
,
K.
, and
Kara
,
L. B.
,
2019
, “
3d Shape Synthesis for Conceptual Design and Optimization Using Variational Autoencoders
,”
ASME 2019 International Design Engineering Technical Conferences and Computers and Information in Engineering and Conference
,
Anaheim, CA
,
Aug. 18–21
, American Society of Mechanical Engineers.
5.
Tracey
,
B.
,
Duraisamy
,
K.
, and
Alonso
,
J. J.
,
2015
, “
A Machine Learning Strategy to Assist Turbulence Model Development
,”
53rd AIAA Aerospace Sciences Meeting (January)
,
Kissimmee
,
FL, Jan. 5–9
.
6.
Ling
,
J.
,
Kurzawski
,
A.
, and
Templeton
,
J.
,
2016
, “
Reynolds Averaged Turbulence Modelling Using Deep Neural Networks With Embedded Invariance
,”
J. Fluid Mech.
,
807
, pp.
155
166
. 10.1017/jfm.2016.615
7.
Singh
,
A. P.
,
Medida
,
S.
, and
Duraisamy
,
K.
,
2017
, “
Machine-Learning-Augmented Predictive Modeling of Turbulent Separated Flows Over Airfoils
,”
AIAA J.
,
55
(
7
), pp.
2215
2227
. 10.2514/1.J055595
8.
Bar-Sinai
,
Y.
,
Hoyer
,
S.
,
Hickey
,
J.
, and
Brenner
,
M. P.
,
2019
, “
Learning Data-Driven Discretizations for Partial Differential Equations
,”
Proc. Natl. Acad. Sci. USA
,
116
(
31
), pp.
15344
15349
. 10.1073/pnas.1814058116
9.
Loiseau
,
J. C.
, and
Brunton
,
S. L.
,
2018
, “
Constrained Sparse Galerkin Regression
,”
J. Fluid Mech.
,
838
, pp.
42
67
. 10.1017/jfm.2017.823
10.
Raissi
,
M.
, and
Karniadakis
,
G. E.
,
2018
, “
Hidden Physics Models: Machine Learning of Nonlinear Partial Differential Equations
,”
J. Comput. Phys.
,
357
, pp.
125
141
. 10.1016/j.jcp.2017.11.039
11.
Raissi
,
M.
,
Perdikaris
,
P.
, and
Karniadakis
,
G. E.
,
2019
, “
Physics-Informed Neural Networks: A Deep Learning Framework for Solving Forward and Inverse Problems Involving Nonlinear Partial Differential Equations
,”
J. Comput. Phys.
,
378
, pp.
686
707
. 10.1016/j.jcp.2018.10.045
12.
Lecun
,
Y.
,
Bengio
,
Y.
, and
Hinton
,
G.
,
2015
, “
Deep Learning
,”
Nature
,
521
(
7553
), pp.
436
444
. 10.1038/nature14539
13.
Kutz
,
J. N.
,
2017
, “
Deep Learning in Fluid Dynamics
,”
J. Fluid Mech.
,
814
, pp.
1
4
. 10.1017/jfm.2016.803
14.
Guo
,
X.
,
Li
,
W.
, and
Iorio
,
F.
,
2016
, “
Convolutional Neural Networks for Steady Flow Approximation
,”
Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Aug. 13–17
,
San Francisco, CA
,
Aug. 13–17
, pp.
481
490
.
15.
Bhatnagar
,
S.
,
Afshar
,
Y.
,
Pan
,
S.
,
Duraisamy
,
K.
, and
Kaushik
,
S.
,
2019
, “
Prediction of Aerodynamic Flow Fields Using Convolutional Neural Networks
,”
Comput. Mech.
,
64
(
2
), pp.
525
545
. 10.1007/s00466-019-01740-0
16.
Sekar
,
V.
,
Jiang
,
Q.
,
Shu
,
C.
, and
Khoo
,
B. C.
,
2019
, “
Fast Flow Field Prediction Over Airfoils Using Deep Learning Approach
,”
Phys. Fluids
,
31
(
5
), p.
057103
. 10.1063/1.5094943
17.
Nabian
,
M. A.
, and
Meidani
,
H.
,
2018
, “
Physics-Informed Regularization of Deep Neural Networks
,”
arXiv preprint
. https://arxiv.org/abs/1810.05547
18.
Liu
,
Y.
,
Lu
,
Y.
,
Wang
,
Y.
,
Sun
,
D.
,
Deng
,
L.
,
Wang
,
F.
, and
Lei
,
Y.
,
2019
, “
A CNN-Based Shock Detection Method in Flow Visualization
,”
Comput. Fluids
,
184
, pp.
1
9
. 10.1016/j.compfluid.2019.03.022
19.
Chen
,
Q.
,
Zeng
,
Z.
, and
Lian
,
L.
,
2018
, “
Accelerating Underwater Vehicle Hydrodynamic Analysis Using Convolutional Networks
,”
2018 OCEANS – MTS/IEEE Kobe Techno-Oceans, OCEANS – Kobe 2018
,
Kobe, Japan
, pp.
1
5
.
20.
Hennigh
,
O.
,
2017
, “
Lat-Net: Compressing Lattice Boltzmann Flow Simulations Using Deep Neural Networks
,”
arXiv preprint
. https://arxiv.org/abs/1705.09036
21.
Viquerat
,
J.
, and
Hachem
,
E.
,
2019
, “
A Supervised Neural Network for Drag Prediction of Arbitrary 2D Shapes in Low Reynolds Number Flows
,”
Comp. Fluids.
,
210
, p.
104645
. https://doi.org/10.1016/j.compfluid.2020.104645
22.
Zhang
,
Y.
,
Sung
,
W. J.
, and
Mavris
,
D.
,
2018
, “
Application of Convolutional Neural Network to Predict Airfoil Lift Coefficient
,”
AIAA/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference (210049)
,
Kissimmee, FL
, pp.
1
9
.
23.
Miyanawala
,
T. P.
, and
Jaiman
,
R. K.
,
2017
, “
An Efficient Deep Learning Technique for the Navier–Stokes Equations: Application to Unsteady Wake Flow Dynamics
,”
arXiv preprint
. https://arxiv.org/abs/1710.09099v3
24.
Hajgató
,
G.
,
Gyires-Tóth
,
B.
, and
Paál
,
G.
,
2019
, “
Accelerating Convergence of Fluid Dynamics Simulations With Convolutional Neural Networks
,”
Periodica Polytechnica Mech. Eng.
,
63
(
3
), pp.
230
239
. 10.3311/PPme.14134
25.
Stoecklein
,
D.
,
Lore
,
K. G.
,
Davies
,
M.
,
Sarkar
,
S.
, and
Ganapathysubramanian
,
B.
,
2017
, “
Deep Learning for Flow Sculpting: Insights Into Efficient Learning Using Scientific Simulation Data
,”
Sci. Rep.
,
7
, pp.
1
11
. 10.1038/srep46368
26.
Sekar
,
V.
,
Zhang
,
M.
,
Shu
,
C.
, and
Khoo
,
B. C.
,
2019
, “
Inverse Design of Airfoil Using a Deep Convolutional Neural Network
,”
AIAA J.
,
57
(
3
), pp.
993
1003
. 10.2514/1.J057894
27.
Mukherjee
,
R.
,
Li
,
Q.
,
Chen
,
Z.
,
Chu
,
S.
, and
Wang
,
H.
,
2018
, “
NeuralDrop: DNN-Based Simulation of Small-Scale Liquid Flows on Solids
,”
arXiv preprint
. https://arxiv.org/abs/1811.02517v1
28.
Kim
,
B.
,
Azevedo
,
V. C.
,
Thuerey
,
N.
,
Kim
,
T.
,
Gross
,
M.
, and
Solenthaler
,
B.
,
2019
, “
Deep Fluids: A Generative Network for Parameterized Fluid Simulations
,”
Comput. Graph. Forum
,
38
(
2
), pp.
59
70
. 10.1111/cgf.13619
29.
Georgiou
,
T.
,
Schmitt
,
S.
,
Olhofer
,
M.
,
Liu
,
Y.
,
Back
,
T.
, and
Lew
,
M.
,
2018
, “
Learning Fluid Flows
,”
Proceedings of the International Joint Conference on Neural Networks
,
Rio de Janeiro, Brazil
,
July 8–13
.
30.
Lee
,
S.
, and
You
,
D.
,
2019
, “
Data-Driven Prediction of Unsteady Flow Over a Circular Cylinder Using Deep Learning
,”
J. Fluid Mech.
,
879
, pp.
217
254
. 10.1017/jfm.2019.700
31.
Orchard
,
A.
,
Vogel
,
J.
, and
Coffin
,
E. K.
,
2020
, “
A Practical Approach to Sizing Thermal Protection for Spaceshiptwo
,”
AIAA Scitech 2020 Forum
,
Orlando, FL
, p.
0258
.
32.
Medic
,
G.
,
Kalitzin
,
G.
, and
You
,
D.
,
2006
, “
Integrated RANS/LES Computations of Turbulent Flow Through a Turbofan Jet Engine
,”
Annual Research Briefs
, pp.
275
285
.
33.
Min
,
A. T. W.
,
Sagarna
,
R.
,
Gupta
,
A.
,
Ong
,
Y.-S.
, and
Goh
,
C. K.
,
2017
, “
Knowledge Transfer Through Machine Learning in Aircraft Design
,”
IEEE Comput. Intell. Mag.
,
12
(
4
), pp.
48
60
. 10.1109/MCI.2017.2742781
34.
Neyshabur
,
B.
,
Bhojanapalli
,
S.
,
McAllester
,
D.
, and
Srebro
,
N.
,
2017
, “
Exploring Generalization in Deep Learning
,”
Advances in Neural Information Processing Systems
,
Long Beach, CA
.
35.
Hoffer
,
E.
,
Hubara
,
I.
, and
Soudry
,
D.
,
2017
, “
Train Longer, Generalize Better: Closing the Generalization Gap in Large Batch Training of Neural Networks
,”
Advances in Neural Information Processing Systems
,
Long Beach, CA
, pp.
1731
1741
.
36.
Zhang
,
C.
,
Bengio
,
S.
,
Hardt
,
M.
,
Recht
,
B.
, and
Vinyals
,
O.
,
2016
, “
Understanding Deep Learning Requires Rethinking Generalization
,”
arXiv preprint
. https://arxiv.org/abs/1611.03530
37.
Wang
,
H.
,
Ren
,
K.
, and
Song
,
J.
,
2017
, “
A Closer Look at Batch Size in Mini-Batch Training of Deep Auto-Encoders
,”
2017 3rd IEEE International Conference on Computer and Communications (ICCC)
,
Chengdu, China
, IEEE, pp.
2756
2761
.
38.
Keskar
,
N. S.
,
Mudigere
,
D.
,
Nocedal
,
J.
,
Smelyanskiy
,
M.
, and
Tang
,
P. T. P.
,
2016
, “
On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima
,”
arXiv preprint
. https://arxiv.org/abs/1609.04836v2
39.
Wu
,
L.
,
Zhu
,
Z.
, and
Weinan
,
E.
,
2017
, “
Towards Understanding Generalization of Deep Learning: Perspective of Loss Landscapes
,”
arXiv preprint
. https://arxiv.org/abs/1706.10239
40.
Dinh
,
L.
,
Pascanu
,
R.
,
Bengio
,
S.
, and
Bengio
,
Y.
,
2017
, “
Sharp Minima Can Generalize for Deep Nets
,”
arXiv preprint
. https://arxiv.org/abs/1703.04933
41.
Le
,
L.
,
Patterson
,
A.
, and
White
,
M.
,
2018
, “
Supervised Autoencoders: Improving Generalization Performance With Unsupervised Regularizers
,”
Advances in Neural Information Processing Systems
,
Montréal, Canada
.
42.
Maurer
,
A.
,
Pontil
,
M.
, and
Romera-Paredes
,
B.
,
2016
, “
The Benefit of Multitask Representation Learning
,”
J. Mach. Learn. Res.
,
17
(
1
), pp.
2853
2884
.
43.
Caruana
,
R.
,
1997
, “
Multitask Learning
,”
Mach. Learn.
,
28
(
1
), pp.
41
75
. 10.1023/A:1007379606734
44.
Rigol
,
J. P.
,
Jarvis
,
C. H.
, and
Stuart
,
N.
,
2001
, “
Artificial Neural Networks as a Tool for Spatial Interpolation
,”
Int. J. Geograph. Inf. Sci.
,
15
(
4
), pp.
323
343
. 10.1080/13658810110038951
45.
Coelho
,
L. P.
, and
Richert
,
W.
,
2015
,
Building Machine Learning Systems With Python
,
Packt Publishing Ltd.
,
Birmingham, UK
.
46.
Zhang
,
Y.
,
2018
,
A Better Autoencoder for Image, ICONIP17-DCEC, http://users.cecs.anu.edu.au/~Tom.Gedeon/conf/ABCs2018/paper/ABCs2018_paper_58.pdf, Accessed March 23, 2017
.
47.
Ioffe
,
S.
, and
Szegedy
,
C.
,
2015
, “
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift.
arXiv preprint
. https://arxiv.org/abs/1502.03167
48.
Sethian
,
J. A.
,
1996
, “
A Fast Marching Level Set Method for Monotonically Advancing Fronts
,”
Proc. Natl. Acad. Sci. USA
,
93
(
4
), pp.
1591
1595
. 10.1073/pnas.93.4.1591
You do not currently have access to this content.