Three-dimensional super-resolution reconstruction of turbulent flow using 3D-ESRGAN with random sampling strategy

This study introduces a deep learning framework that uses an enhanced three-dimensional super-resolution generative adversarial network (3D-ESRGAN) to reconstruct high-resolution turbulent flow fields from low-resolution data. To minimize the reliance on complete datasets during training, a random s...

Full description

Saved in:
Bibliographic Details
Published in:Computers & fluids Vol. 305; p. 106890
Main Authors: Yu, Linqi, Chen, Yanyun, Yousif, Mustafa Z., Lim, Hee-Chang
Format: Journal Article
Language:English
Published: Elsevier Ltd 30.01.2026
Subjects:
ISSN:0045-7930
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract This study introduces a deep learning framework that uses an enhanced three-dimensional super-resolution generative adversarial network (3D-ESRGAN) to reconstruct high-resolution turbulent flow fields from low-resolution data. To minimize the reliance on complete datasets during training, a random sampling strategy is used. In each training epoch, approximately 1 % of spatial points are randomly chosen from both the predicted and ground-truth fields for loss computation. This method mimics sparse sensor measurements in real-world experiments, allowing the model to learn accurate mappings based on limited spatial observations. Furthermore, physics-guided loss functions are incorporated to ensure consistency with the underlying physical laws, thereby improving the reliability of the reconstructed flow fields. Subsequently, the framework is tested on two typical flow scenarios: (1) flow over a finite wall-mounted square cylinder at Red=500, and (2) fully developed turbulent channel flow at Reτ=180 and 500. Both scenarios are generated using direct numerical simulation (DNS). Finally, the results are presented and analyzed both qualitatively and quantitatively. The qualitative findings show that the reconstructed fields effectively restore the key vortex structures and turbulence features absent in the coarse data. Quantitative comparisons with the ground-truth data confirm high accuracy in terms of velocity profiles, Reynolds stresses, probability density functions (PDFs), and energy spectra. Additionally, the relative error regarding the streamwise velocity magnitude is calculated, and all cases show a low error rate of approximately 5 %. In summary, the findings confirm the efficacy of combining GAN-based super-resolution with a random sampling strategy for accurate and data-efficient 3D turbulence reconstruction. The results suggest that the proposed framework can be successfully applied in real-world scenarios where only sparse measurements are accessible.
AbstractList This study introduces a deep learning framework that uses an enhanced three-dimensional super-resolution generative adversarial network (3D-ESRGAN) to reconstruct high-resolution turbulent flow fields from low-resolution data. To minimize the reliance on complete datasets during training, a random sampling strategy is used. In each training epoch, approximately 1 % of spatial points are randomly chosen from both the predicted and ground-truth fields for loss computation. This method mimics sparse sensor measurements in real-world experiments, allowing the model to learn accurate mappings based on limited spatial observations. Furthermore, physics-guided loss functions are incorporated to ensure consistency with the underlying physical laws, thereby improving the reliability of the reconstructed flow fields. Subsequently, the framework is tested on two typical flow scenarios: (1) flow over a finite wall-mounted square cylinder at Red=500, and (2) fully developed turbulent channel flow at Reτ=180 and 500. Both scenarios are generated using direct numerical simulation (DNS). Finally, the results are presented and analyzed both qualitatively and quantitatively. The qualitative findings show that the reconstructed fields effectively restore the key vortex structures and turbulence features absent in the coarse data. Quantitative comparisons with the ground-truth data confirm high accuracy in terms of velocity profiles, Reynolds stresses, probability density functions (PDFs), and energy spectra. Additionally, the relative error regarding the streamwise velocity magnitude is calculated, and all cases show a low error rate of approximately 5 %. In summary, the findings confirm the efficacy of combining GAN-based super-resolution with a random sampling strategy for accurate and data-efficient 3D turbulence reconstruction. The results suggest that the proposed framework can be successfully applied in real-world scenarios where only sparse measurements are accessible.
ArticleNumber 106890
Author Lim, Hee-Chang
Yu, Linqi
Yousif, Mustafa Z.
Chen, Yanyun
Author_xml – sequence: 1
  givenname: Linqi
  orcidid: 0000-0002-5674-6261
  surname: Yu
  fullname: Yu, Linqi
– sequence: 2
  givenname: Yanyun
  orcidid: 0009-0009-7003-2116
  surname: Chen
  fullname: Chen, Yanyun
– sequence: 3
  givenname: Mustafa Z.
  orcidid: 0000-0002-5542-5474
  surname: Yousif
  fullname: Yousif, Mustafa Z.
– sequence: 4
  givenname: Hee-Chang
  orcidid: 0000-0001-8504-0797
  surname: Lim
  fullname: Lim, Hee-Chang
  email: hclim@pusan.ac.kr
BookMark eNqFkM1OAjEUhbvARECfwb7AYKfz05klQUQTooniuum0t1Ay02I7I-HtLWLcuro55-Sc3HwTNLLOAkJ3KZmlJC3v9zPpuoNuB6NmlNAiumVVkxEaE5IXCaszco0mIexJ1BnNx-hzs_MAiTId2GCcFS0OwwF84iG4duijhT1IZ0PvB_kjncb94JuhBdtj3bojHoKxW5w9JMv3t9X8BR9Nv8NeWOU6HER3aM9xHBA9bE836EqLNsDt752ij8flZvGUrF9Xz4v5OpG0zPqkKhRhmSp1LSjVdVlGQ9ZZ2ehaCVZQ2TAGBRCSkrpoJMuB5ULJqqKCAWtYNkXssiu9C8GD5gdvOuFPPCX8TIvv-R8tfqbFL7Ric35pQnzvy4DnQRqwEpSJKHqunPl34xsFkn6I
Cites_doi 10.1063/1.5140772
10.1016/j.jcp.2013.12.042
10.1017/S0022112087000892
10.1063/5.0129203
10.1103/PhysRevLett.128.024502
10.1109/MM.2008.57
10.1017/jfm.2024.1136
10.1063/5.0039845
10.1016/S0167-6105(02)00395-1
10.1007/s11280-018-0636-4
10.1017/jfm.2021.231
10.1063/5.0074724
10.1146/annurev-fluid-010518-040547
10.1146/annurev.fluid.30.1.539
10.1038/s41598-023-29525-9
10.1103/RevModPhys.71.S383
10.1103/PhysRevE.101.061201
10.1016/0021-9991(92)90324-R
10.1063/1.1509751
10.1063/1.1404847
10.1017/jfm.2022.61
10.1038/s42256-022-00572-7
10.1017/S0022112001005821
10.1088/1367-2630/9/4/094
10.1063/5.0020698
10.1017/jfm.2019.238
10.1063/1.869966
10.1007/s00162-020-00518-y
10.1016/0021-9991(91)90238-G
10.1063/1.5127031
10.1017/jfm.2016.803
10.1146/annurev-fluid-010719-060214
10.1146/annurev.fluid.30.1.31
10.1017/jfm.2024.1202
10.1145/3478513.3480492
10.1017/jfm.2019.700
10.1007/s00348-007-0319-x
10.1063/5.0066077
10.1146/annurev.fluid.010908.165203
10.1063/1.2162185
10.1038/s41416-019-0451-4
10.1063/5.0058346
10.1038/s43588-022-00264-7
10.1016/j.ces.2020.116133
ContentType Journal Article
Copyright 2025 Elsevier Ltd
Copyright_xml – notice: 2025 Elsevier Ltd
DBID AAYXX
CITATION
DOI 10.1016/j.compfluid.2025.106890
DatabaseName CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
ExternalDocumentID 10_1016_j_compfluid_2025_106890
S0045793025003500
GroupedDBID --K
--M
-~X
.DC
.~1
0R~
1B1
1~.
1~5
4.4
457
4G.
5GY
5VS
7-5
71M
8P~
9JN
AAEDT
AAEDW
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AATTM
AAXKI
AAXUO
AAYWO
ABAOU
ABJNI
ABMAC
ACDAQ
ACGFS
ACIWK
ACLOT
ACRLP
ACVFH
ADBBV
ADCNI
ADEZE
ADGUI
ADTZH
AEBSH
AECPX
AEIPS
AEKER
AENEX
AEUPX
AFJKZ
AFPUW
AFTJW
AGHFR
AGUBO
AGYEJ
AHHHB
AHJVU
AIEXJ
AIGII
AIGVJ
AIIUN
AIKHN
AITUG
AKBMS
AKRWK
AKYEP
ALMA_UNASSIGNED_HOLDINGS
AMRAJ
ANKPU
APXCP
ARUGR
AXJTR
BJAXD
BKOJK
BLXMC
CS3
DU5
EBS
EFJIC
EFKBS
EFLBG
EO8
EO9
EP2
EP3
FDB
FIRID
FNPLU
FYGXN
G-Q
GBLVA
IHE
J1W
JJJVA
KOM
MHUIS
MO0
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
PQQKQ
Q38
ROL
RPZ
SDF
SDG
SDP
SES
SEW
SPC
SPCBC
SPD
SST
SSW
SSZ
T5K
TN5
XPP
ZMT
~G-
~HD
29F
6TJ
9DU
AAQXK
AAYXX
ABDPE
ABEFU
ABFNM
ABWVN
ABXDB
ACKIV
ACNNM
ACRPL
ADIYS
ADMUD
ADNMO
AFFNX
AGQPQ
AI.
ASPBG
AVWKF
AZFZN
CITATION
EJD
FEDTE
FGOYB
G-2
HLZ
HVGLF
HZ~
LG9
LY7
M41
R2-
SBC
SET
T9H
VH1
WUQ
ID FETCH-LOGICAL-c263t-85d073d6f9a22f96685dc936bf9da752cb77e5e001095bc74e74adc882a7e7b73
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001620450600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0045-7930
IngestDate Thu Nov 27 01:07:34 EST 2025
Wed Dec 10 14:23:42 EST 2025
IsPeerReviewed true
IsScholarly true
Keywords Deep learning
Random sampling strategy
Super-resolution reconstruction
Three-dimensional turbulent flow
3D-ESRGAN
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c263t-85d073d6f9a22f96685dc936bf9da752cb77e5e001095bc74e74adc882a7e7b73
ORCID 0000-0002-5542-5474
0000-0002-5674-6261
0009-0009-7003-2116
0000-0001-8504-0797
ParticipantIDs crossref_primary_10_1016_j_compfluid_2025_106890
elsevier_sciencedirect_doi_10_1016_j_compfluid_2025_106890
PublicationCentury 2000
PublicationDate 2026-01-30
PublicationDateYYYYMMDD 2026-01-30
PublicationDate_xml – month: 01
  year: 2026
  text: 2026-01-30
  day: 30
PublicationDecade 2020
PublicationTitle Computers & fluids
PublicationYear 2026
Publisher Elsevier Ltd
Publisher_xml – name: Elsevier Ltd
References Brunton, Noack, Koumoutsakos (bib0022) 2020; 52
Vinuesa, Brunton (bib0031) 2022; 2
Duraisamy, Iaccarino, Xiao (bib0024) 2019; 51
Güemes, Vila, Discetti (bib0038) 2022; 4
Hoyas, Jiménez (bib0048) 2006; 18
Kaneda, Ishihara (bib0012) 2006
Hoyas, Oberlack, Alcántara-Ávila, Kraheberger, Laux (bib0051) 2022; 7
Chang, Collis, Ramakrishnan (bib0015) 2002; 14
Fukami, Fukagata, Taira (bib0047) 2020; 34
Lee, You (bib0027) 2019; 879
Güemes, Discetti, Ianiro, Sirmacek, Azizpour, Vinuesa (bib0037) 2021; 33
Yousif, Yu, Lim (bib0026) 2022; 34
Hfyngaard (bib0005) 1992; 205
Munir, Lisec, Swinnen, Zaidi (bib0019) 2019; 120
Ishihara, Gotoh, Kaneda (bib0014) 2009; 41
Fischer P., Merzari E., Min M., Kerkemeier S., Lan Y.-H., Phillips M., et al. Highly optimized full-core reactor simulations on summit.2021.
Bai, Wang, Desbrun, Liu (bib0010) 2021; 40
Moser, Kim, Mansour (bib0046) 1999; 11
Yu, Yousif, Zhang, Hoyas, Vinuesa, Lim (bib0039) 2022; 34
Han, Huang (bib0030) 2020; 32
Deng, He, Liu, Kim (bib0035) 2019; 31
Yousif, Yu, Hoyas, Vinuesa, Lim (bib0032) 2023; 13
.
Sreenivasan (bib0001) 1999; 71
Jiang, Gu, Lee, Smith, Linden (bib0023) 2021; 914
Spalart, Moser, Rogers (bib0054) 1991; 96
Deng, Yu (bib0017) 2014; 7
Nathan (bib0021) 2017; 814
Wang, Yang, Wu, Ma, Peng, Liu (bib0006) 2021; 229
Page (bib0041) 2025; 1002
Yousif, Yu, Lim (bib0028) 2022; 936
Kim, Moin, Moser (bib0045) 1987; 177
Pullin, Saffman (bib0008) 1998; 30
Heinonen, Diamond (bib0016) 2020; 101
Garland, Grand, Nickolls, Anderson, Hardwick, Morton (bib0020) 2008; 28
Alcántara-Ávila, Hoyas, Pérez-Quiles (bib0050) 2021; 916
Oberlack, Hoyas, Kraheberger, Alcántara-Ávila, Laux (bib0052) 2022; 128
Nielsen (bib0009) 1993; 63
Lluesma-Rodríguez, González, Hoyas (bib0049) 2021; 9
Bewley, Moin, Temam (bib0013) 2001; 447
Dong, Karniadakis, Chryssostomidis (bib0042) 2014; 261
Wang, Kashinath, Mustafa, Albert, Yu (bib0029) 2020
Gopaul, Andrews (bib0004) 2007; 9
Deville, Fischer, Mund (bib0044) 2002; 9
Yousif, Yu, Lim (bib0036) 2021; 33
Kim, Baik (bib0003) 2003; 91
Fukami, Fukagata, Taira (bib0033) 2019; 870
Moin, Mahesh (bib0011) 1998; 30
Lavoie, Avallone, Gregorio, Romano, Antonia (bib0007) 2007; 43
Pouyanfar, Tao, Tian, Chen, Shyu (bib0018) 2019; 22
Nakamura, Fukami, Hasegawa, Nabae, Fukagata (bib0025) 2021; 33
Fukami, Taira (bib0040) 2024; 1001
Liu, Tang, Huang, Lu (bib0034) 2020; 32
Goodfellow, Pouget-Abadie, Mirza, Xu, Warde-Farley, Ozair (bib0055) 2014; 27
Lele (bib0053) 1992; 103
Kadanoff (bib0002) 2001; 54
Pullin (10.1016/j.compfluid.2025.106890_bib0008) 1998; 30
10.1016/j.compfluid.2025.106890_bib0043
Munir (10.1016/j.compfluid.2025.106890_bib0019) 2019; 120
Garland (10.1016/j.compfluid.2025.106890_bib0020) 2008; 28
Yu (10.1016/j.compfluid.2025.106890_bib0039) 2022; 34
Fukami (10.1016/j.compfluid.2025.106890_bib0033) 2019; 870
Yousif (10.1016/j.compfluid.2025.106890_bib0032) 2023; 13
Moin (10.1016/j.compfluid.2025.106890_bib0011) 1998; 30
Goodfellow (10.1016/j.compfluid.2025.106890_bib0055) 2014; 27
Heinonen (10.1016/j.compfluid.2025.106890_bib0016) 2020; 101
Pouyanfar (10.1016/j.compfluid.2025.106890_bib0018) 2019; 22
Nakamura (10.1016/j.compfluid.2025.106890_bib0025) 2021; 33
Liu (10.1016/j.compfluid.2025.106890_bib0034) 2020; 32
Sreenivasan (10.1016/j.compfluid.2025.106890_bib0001) 1999; 71
Ishihara (10.1016/j.compfluid.2025.106890_bib0014) 2009; 41
Moser (10.1016/j.compfluid.2025.106890_bib0046) 1999; 11
Kadanoff (10.1016/j.compfluid.2025.106890_bib0002) 2001; 54
Gopaul (10.1016/j.compfluid.2025.106890_bib0004) 2007; 9
Yousif (10.1016/j.compfluid.2025.106890_bib0028) 2022; 936
Güemes (10.1016/j.compfluid.2025.106890_bib0038) 2022; 4
Fukami (10.1016/j.compfluid.2025.106890_bib0047) 2020; 34
Han (10.1016/j.compfluid.2025.106890_bib0030) 2020; 32
Kim (10.1016/j.compfluid.2025.106890_bib0045) 1987; 177
Jiang (10.1016/j.compfluid.2025.106890_bib0023) 2021; 914
Deng (10.1016/j.compfluid.2025.106890_bib0017) 2014; 7
Bewley (10.1016/j.compfluid.2025.106890_bib0013) 2001; 447
Nathan (10.1016/j.compfluid.2025.106890_bib0021) 2017; 814
Kim (10.1016/j.compfluid.2025.106890_bib0003) 2003; 91
Wang (10.1016/j.compfluid.2025.106890_bib0029) 2020
Lluesma-Rodríguez (10.1016/j.compfluid.2025.106890_bib0049) 2021; 9
Deville (10.1016/j.compfluid.2025.106890_bib0044) 2002; 9
Page (10.1016/j.compfluid.2025.106890_bib0041) 2025; 1002
Nielsen (10.1016/j.compfluid.2025.106890_bib0009) 1993; 63
Duraisamy (10.1016/j.compfluid.2025.106890_bib0024) 2019; 51
Bai (10.1016/j.compfluid.2025.106890_bib0010) 2021; 40
Spalart (10.1016/j.compfluid.2025.106890_bib0054) 1991; 96
Lele (10.1016/j.compfluid.2025.106890_bib0053) 1992; 103
Wang (10.1016/j.compfluid.2025.106890_bib0006) 2021; 229
Güemes (10.1016/j.compfluid.2025.106890_bib0037) 2021; 33
Oberlack (10.1016/j.compfluid.2025.106890_bib0052) 2022; 128
Chang (10.1016/j.compfluid.2025.106890_bib0015) 2002; 14
Vinuesa (10.1016/j.compfluid.2025.106890_bib0031) 2022; 2
Fukami (10.1016/j.compfluid.2025.106890_bib0040) 2024; 1001
Kaneda (10.1016/j.compfluid.2025.106890_bib0012) 2006
Dong (10.1016/j.compfluid.2025.106890_bib0042) 2014; 261
Lee (10.1016/j.compfluid.2025.106890_bib0027) 2019; 879
Alcántara-Ávila (10.1016/j.compfluid.2025.106890_bib0050) 2021; 916
Brunton (10.1016/j.compfluid.2025.106890_bib0022) 2020; 52
Lavoie (10.1016/j.compfluid.2025.106890_bib0007) 2007; 43
Yousif (10.1016/j.compfluid.2025.106890_bib0026) 2022; 34
Yousif (10.1016/j.compfluid.2025.106890_bib0036) 2021; 33
Hfyngaard (10.1016/j.compfluid.2025.106890_bib0005) 1992; 205
Hoyas (10.1016/j.compfluid.2025.106890_bib0048) 2006; 18
Hoyas (10.1016/j.compfluid.2025.106890_bib0051) 2022; 7
Deng (10.1016/j.compfluid.2025.106890_bib0035) 2019; 31
References_xml – volume: 879
  start-page: 217
  year: 2019
  end-page: 254
  ident: bib0027
  article-title: Data-driven prediction of unsteady flow over a circular cylinder using deep learning
  publication-title: J Fluid Mech
– volume: 28
  start-page: 13
  year: 2008
  end-page: 27
  ident: bib0020
  article-title: Parallel computing experiences with cuda
  publication-title: IEEE Micro
– volume: 34
  year: 2022
  ident: bib0026
  article-title: Super-resolution reconstruction of turbulent flow fields at various reynolds numbers based on generative adversarial networks
  publication-title: Phys Fluid
– volume: 9
  year: 2002
  ident: bib0044
  article-title: High-order methods for incompressible fluid flow
– volume: 33
  year: 2021
  ident: bib0036
  article-title: High-fidelity reconstruction of turbulent flow from spatially limited data using enhanced super-resolution generative adversarial network
  publication-title: Phys Fluid
– volume: 261
  start-page: 83
  year: 2014
  end-page: 105
  ident: bib0042
  article-title: A robust and accurate outflow boundary condition for incompressible flow simulations on severely-truncated unbounded domains
  publication-title: J Comput Phys
– volume: 43
  start-page: 39
  year: 2007
  end-page: 51
  ident: bib0007
  article-title: Spatial resolution of piv for the measurement of turbulence
  publication-title: Exp Fluid
– volume: 1001
  year: 2024
  ident: bib0040
  article-title: Single-snapshot machine learning for super-resolution of turbulence
  publication-title: J Fluid Mech
– volume: 32
  year: 2020
  ident: bib0030
  article-title: Active control for drag reduction of turbulent channel flow based on convolutional neural networks
  publication-title: Phys Fluid
– volume: 870
  start-page: 106
  year: 2019
  end-page: 120
  ident: bib0033
  article-title: Super-resolution reconstruction of turbulent flows with machine learning
  publication-title: J Fluid Mech
– volume: 33
  year: 2021
  ident: bib0025
  article-title: Convolutional neural network and long short-term memory based reduced order surrogate for minimal turbulent channel flow
  publication-title: Phys Fluid
– volume: 96
  start-page: 297
  year: 1991
  end-page: 324
  ident: bib0054
  article-title: Spectral methods for the navier-stokes equations with one infinite and two periodic directions
  publication-title: J Comput Phys
– volume: 34
  year: 2022
  ident: bib0039
  article-title: Three-dimensional esrgan for super-resolution reconstruction of turbulent flows with tricubic interpolation-based transfer learning
  publication-title: Phys Fluid
– volume: 13
  year: 2023
  ident: bib0032
  article-title: A deep-learning approach for reconstructing 3d turbulent flows from 2d observation data
  publication-title: Sci Rep
– volume: 33
  year: 2021
  ident: bib0037
  article-title: From coarse wall measurements to turbulent velocity fields through deep learning
  publication-title: Phys Fluid
– volume: 229
  year: 2021
  ident: bib0006
  article-title: Estimation of the dissipation rate of turbulent kinetic energy: a review
  publication-title: Chem Eng Sci
– volume: 9
  start-page: 1
  year: 2021
  end-page: 3
  ident: bib0049
  article-title: Cfd simulation of a hyperloop capsule inside a closed environment
  publication-title: Result Eng
– volume: 32
  year: 2020
  ident: bib0034
  article-title: Deep learning methods for super-resolution reconstruction of turbulent flows
  publication-title: Phys Fluid
– volume: 4
  start-page: 1165
  year: 2022
  end-page: 1173
  ident: bib0038
  article-title: Super-resolution generative adversarial networks of randomly-seeded fields
  publication-title: Nat Mach Intell
– volume: 11
  start-page: 943
  year: 1999
  end-page: 945
  ident: bib0046
  article-title: Direct numerical simulation of turbulent channel flow up to
  publication-title: Phys Fluid
– volume: 30
  start-page: 539
  year: 1998
  end-page: 578
  ident: bib0011
  article-title: Direct numerical simulation: a tool in turbulence research
  publication-title: Annu Rev Fluid Mech
– volume: 447
  start-page: 179
  year: 2001
  end-page: 225
  ident: bib0013
  article-title: DNS-Based predictive control of turbulence: an optimal benchmark for feedback algorithms
  publication-title: J Fluid Mech
– volume: 1002
  year: 2025
  ident: bib0041
  article-title: Super-resolution of turbulence with dynamics in the loss
  publication-title: J Fluid Mech
– volume: 91
  start-page: 309
  year: 2003
  end-page: 329
  ident: bib0003
  article-title: Effects of inflow turbulence intensity on flow and pollutant dispersion in an urban street canyon
  publication-title: J Wind Eng Ind Aerodyn
– volume: 103
  start-page: 16
  year: 1992
  end-page: 42
  ident: bib0053
  article-title: Compact finite difference schemes with spectral-like resolution
  publication-title: J Comput Phys
– volume: 14
  start-page: 4069
  year: 2002
  end-page: 4080
  ident: bib0015
  article-title: Viscous effects in control of near-wall turbulence
  publication-title: Phys Fluid
– volume: 916
  year: 2021
  ident: bib0050
  article-title: Direct numerical simulation of thermal channel flow for
  publication-title: J Fluid Mech
– volume: 30
  start-page: 31
  year: 1998
  end-page: 51
  ident: bib0008
  article-title: Vortex dynamics in turbulence
  publication-title: Annu Rev Fluid Mech
– volume: 31
  year: 2019
  ident: bib0035
  article-title: Super-resolution reconstruction of turbulent velocity fields using a generative adversarial network-based artificial intelligence framework
  publication-title: Phys Fluid
– volume: 22
  start-page: 1893
  year: 2019
  end-page: 1911
  ident: bib0018
  article-title: Multimodal deep learning based on multiple correspondence analysis for disaster management
  publication-title: World Wide Web
– volume: 40
  start-page: 1
  year: 2021
  end-page: 16
  ident: bib0010
  article-title: Predicting high-resolution turbulence details in space and time
  publication-title: ACM Transact Graph (TOG)
– volume: 936
  year: 2022
  ident: bib0028
  article-title: Physics-guided deep learning for generating turbulent inflow conditions
  publication-title: J Fluid Mech
– volume: 177
  start-page: 133
  year: 1987
  end-page: 166
  ident: bib0045
  article-title: Turbulence statistics in fully developed channel flow at low reynolds number
  publication-title: J Fluid Mech
– volume: 101
  year: 2020
  ident: bib0016
  article-title: Turbulence model reduction by deep learning
  publication-title: Phys Rev E
– volume: 7
  year: 2022
  ident: bib0051
  article-title: Wall turbulence at high friction reynolds numbers
  publication-title: Phys Rev Fluid
– volume: 34
  start-page: 497
  year: 2020
  end-page: 519
  ident: bib0047
  article-title: Assessment of supervised machine learning methods for fluid flows
  publication-title: Theor Comput Fluid Dyn
– volume: 914
  year: 2021
  ident: bib0023
  article-title: A metamorphosis of three-dimensional wave structure in transitional and turbulent boundary layers
  publication-title: J Fluid Mech
– start-page: 1457
  year: 2020
  end-page: 1466
  ident: bib0029
  article-title: Towards physics-informed deep learning for turbulent flow prediction
  publication-title: Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining
– volume: 71
  year: 1999
  ident: bib0001
  article-title: Fluid turbulence
  publication-title: Rev Mod Phys
– volume: 27
  year: 2014
  ident: bib0055
  article-title: Generative adversarial nets
  publication-title: Adv Neural Inf Process Syst
– volume: 18
  year: 2006
  ident: bib0048
  article-title: Scaling of the velocity fluctuations in turbulent channels up to re
  publication-title: Phys Fluid
– volume: 120
  start-page: 1090
  year: 2019
  end-page: 1098
  ident: bib0019
  article-title: Lipid metabolism in cancer cells under metabolic stress
  publication-title: Br J Cancer
– volume: 814
  start-page: 1
  year: 2017
  end-page: 4
  ident: bib0021
  article-title: Deep learning in fluid dynamics
  publication-title: J Fluid Mech
– volume: 205
  start-page: 33
  year: 1992
  ident: bib0005
  article-title: Atmospheric turbulence
  publication-title: Annu Rev Fluid Mech
– volume: 9
  start-page: 94
  year: 2007
  ident: bib0004
  article-title: The effect of atmospheric turbulence on entangled orbital angular momentum states
  publication-title: New J Phys
– volume: 63
  start-page: 835
  year: 1993
  end-page: 838
  ident: bib0009
  article-title: Turbulence effects on the settling of suspended particles
  publication-title: J Sediment Res
– volume: 51
  start-page: 357
  year: 2019
  end-page: 377
  ident: bib0024
  article-title: Turbulence modeling in the age of data
  publication-title: Annu Rev Fluid Mech
– reference: Fischer P., Merzari E., Min M., Kerkemeier S., Lan Y.-H., Phillips M., et al. Highly optimized full-core reactor simulations on summit.2021.
– reference: .
– start-page: 20
  year: 2006
  ident: bib0012
  article-title: High-resolution direct numerical simulation of turbulence
– volume: 41
  start-page: 165
  year: 2009
  end-page: 180
  ident: bib0014
  article-title: Study of high-reynolds number isotropic turbulence by direct numerical simulation
  publication-title: Annu Rev Fluid Mech
– volume: 54
  start-page: 34
  year: 2001
  end-page: 39
  ident: bib0002
  article-title: Turbulent heat flow: structures and scaling
  publication-title: Phys Today
– volume: 7
  year: 2014
  ident: bib0017
  article-title: Deep learning: methods and applications
  publication-title: Foundat Trends® Signal Process
– volume: 52
  start-page: 477
  year: 2020
  end-page: 508
  ident: bib0022
  article-title: Machine learning for fluid mechanics
  publication-title: Annu Rev Fluid Mech
– volume: 128
  year: 2022
  ident: bib0052
  article-title: Turbulence statistics of arbitrary moments of wall-bounded shear flows: a symmetry approach
  publication-title: Phys Rev Lett
– volume: 2
  start-page: 358
  year: 2022
  end-page: 366
  ident: bib0031
  article-title: Enhancing computational fluid dynamics with machine learning
  publication-title: Nat Comput Sci
– volume: 32
  issue: 2
  year: 2020
  ident: 10.1016/j.compfluid.2025.106890_bib0034
  article-title: Deep learning methods for super-resolution reconstruction of turbulent flows
  publication-title: Phys Fluid
  doi: 10.1063/1.5140772
– volume: 261
  start-page: 83
  year: 2014
  ident: 10.1016/j.compfluid.2025.106890_bib0042
  article-title: A robust and accurate outflow boundary condition for incompressible flow simulations on severely-truncated unbounded domains
  publication-title: J Comput Phys
  doi: 10.1016/j.jcp.2013.12.042
– volume: 63
  start-page: 835
  issue: 5
  year: 1993
  ident: 10.1016/j.compfluid.2025.106890_bib0009
  article-title: Turbulence effects on the settling of suspended particles
  publication-title: J Sediment Res
– start-page: 1457
  year: 2020
  ident: 10.1016/j.compfluid.2025.106890_bib0029
  article-title: Towards physics-informed deep learning for turbulent flow prediction
– volume: 914
  issue: 4
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0023
  article-title: A metamorphosis of three-dimensional wave structure in transitional and turbulent boundary layers
  publication-title: J Fluid Mech
– volume: 177
  start-page: 133
  year: 1987
  ident: 10.1016/j.compfluid.2025.106890_bib0045
  article-title: Turbulence statistics in fully developed channel flow at low reynolds number
  publication-title: J Fluid Mech
  doi: 10.1017/S0022112087000892
– volume: 34
  issue: 12
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0039
  article-title: Three-dimensional esrgan for super-resolution reconstruction of turbulent flows with tricubic interpolation-based transfer learning
  publication-title: Phys Fluid
  doi: 10.1063/5.0129203
– volume: 7
  issue: 1
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0051
  article-title: Wall turbulence at high friction reynolds numbers
  publication-title: Phys Rev Fluid
– volume: 128
  issue: 2
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0052
  article-title: Turbulence statistics of arbitrary moments of wall-bounded shear flows: a symmetry approach
  publication-title: Phys Rev Lett
  doi: 10.1103/PhysRevLett.128.024502
– volume: 28
  start-page: 13
  issue: 4
  year: 2008
  ident: 10.1016/j.compfluid.2025.106890_bib0020
  article-title: Parallel computing experiences with cuda
  publication-title: IEEE Micro
  doi: 10.1109/MM.2008.57
– volume: 1001
  year: 2024
  ident: 10.1016/j.compfluid.2025.106890_bib0040
  article-title: Single-snapshot machine learning for super-resolution of turbulence
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2024.1136
– volume: 205
  start-page: 33
  year: 1992
  ident: 10.1016/j.compfluid.2025.106890_bib0005
  article-title: Atmospheric turbulence
  publication-title: Annu Rev Fluid Mech
– volume: 33
  issue: 2
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0025
  article-title: Convolutional neural network and long short-term memory based reduced order surrogate for minimal turbulent channel flow
  publication-title: Phys Fluid
  doi: 10.1063/5.0039845
– volume: 27
  year: 2014
  ident: 10.1016/j.compfluid.2025.106890_bib0055
  article-title: Generative adversarial nets
  publication-title: Adv Neural Inf Process Syst
– volume: 91
  start-page: 309
  issue: 3
  year: 2003
  ident: 10.1016/j.compfluid.2025.106890_bib0003
  article-title: Effects of inflow turbulence intensity on flow and pollutant dispersion in an urban street canyon
  publication-title: J Wind Eng Ind Aerodyn
  doi: 10.1016/S0167-6105(02)00395-1
– start-page: 20
  year: 2006
  ident: 10.1016/j.compfluid.2025.106890_bib0012
– volume: 22
  start-page: 1893
  year: 2019
  ident: 10.1016/j.compfluid.2025.106890_bib0018
  article-title: Multimodal deep learning based on multiple correspondence analysis for disaster management
  publication-title: World Wide Web
  doi: 10.1007/s11280-018-0636-4
– volume: 916
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0050
  article-title: Direct numerical simulation of thermal channel flow for
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2021.231
– volume: 34
  issue: 1
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0026
  article-title: Super-resolution reconstruction of turbulent flow fields at various reynolds numbers based on generative adversarial networks
  publication-title: Phys Fluid
  doi: 10.1063/5.0074724
– volume: 51
  start-page: 357
  issue: 1
  year: 2019
  ident: 10.1016/j.compfluid.2025.106890_bib0024
  article-title: Turbulence modeling in the age of data
  publication-title: Annu Rev Fluid Mech
  doi: 10.1146/annurev-fluid-010518-040547
– volume: 30
  start-page: 539
  issue: 1
  year: 1998
  ident: 10.1016/j.compfluid.2025.106890_bib0011
  article-title: Direct numerical simulation: a tool in turbulence research
  publication-title: Annu Rev Fluid Mech
  doi: 10.1146/annurev.fluid.30.1.539
– volume: 13
  issue: 1
  year: 2023
  ident: 10.1016/j.compfluid.2025.106890_bib0032
  article-title: A deep-learning approach for reconstructing 3d turbulent flows from 2d observation data
  publication-title: Sci Rep
  doi: 10.1038/s41598-023-29525-9
– volume: 71
  year: 1999
  ident: 10.1016/j.compfluid.2025.106890_bib0001
  article-title: Fluid turbulence
  publication-title: Rev Mod Phys
  doi: 10.1103/RevModPhys.71.S383
– volume: 101
  issue: 6
  year: 2020
  ident: 10.1016/j.compfluid.2025.106890_bib0016
  article-title: Turbulence model reduction by deep learning
  publication-title: Phys Rev E
  doi: 10.1103/PhysRevE.101.061201
– volume: 103
  start-page: 16
  issue: 1
  year: 1992
  ident: 10.1016/j.compfluid.2025.106890_bib0053
  article-title: Compact finite difference schemes with spectral-like resolution
  publication-title: J Comput Phys
  doi: 10.1016/0021-9991(92)90324-R
– volume: 14
  start-page: 4069
  issue: 11
  year: 2002
  ident: 10.1016/j.compfluid.2025.106890_bib0015
  article-title: Viscous effects in control of near-wall turbulence
  publication-title: Phys Fluid
  doi: 10.1063/1.1509751
– volume: 54
  start-page: 34
  issue: 8
  year: 2001
  ident: 10.1016/j.compfluid.2025.106890_bib0002
  article-title: Turbulent heat flow: structures and scaling
  publication-title: Phys Today
  doi: 10.1063/1.1404847
– volume: 936
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0028
  article-title: Physics-guided deep learning for generating turbulent inflow conditions
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2022.61
– volume: 4
  start-page: 1165
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0038
  article-title: Super-resolution generative adversarial networks of randomly-seeded fields
  publication-title: Nat Mach Intell
  doi: 10.1038/s42256-022-00572-7
– volume: 447
  start-page: 179
  issue: 7
  year: 2001
  ident: 10.1016/j.compfluid.2025.106890_bib0013
  article-title: DNS-Based predictive control of turbulence: an optimal benchmark for feedback algorithms
  publication-title: J Fluid Mech
  doi: 10.1017/S0022112001005821
– volume: 9
  start-page: 94
  issue: 4
  year: 2007
  ident: 10.1016/j.compfluid.2025.106890_bib0004
  article-title: The effect of atmospheric turbulence on entangled orbital angular momentum states
  publication-title: New J Phys
  doi: 10.1088/1367-2630/9/4/094
– volume: 32
  issue: 9
  year: 2020
  ident: 10.1016/j.compfluid.2025.106890_bib0030
  article-title: Active control for drag reduction of turbulent channel flow based on convolutional neural networks
  publication-title: Phys Fluid
  doi: 10.1063/5.0020698
– volume: 870
  start-page: 106
  year: 2019
  ident: 10.1016/j.compfluid.2025.106890_bib0033
  article-title: Super-resolution reconstruction of turbulent flows with machine learning
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2019.238
– volume: 11
  start-page: 943
  issue: 4
  year: 1999
  ident: 10.1016/j.compfluid.2025.106890_bib0046
  article-title: Direct numerical simulation of turbulent channel flow up to reτ= 590
  publication-title: Phys Fluid
  doi: 10.1063/1.869966
– volume: 34
  start-page: 497
  issue: 4
  year: 2020
  ident: 10.1016/j.compfluid.2025.106890_bib0047
  article-title: Assessment of supervised machine learning methods for fluid flows
  publication-title: Theor Comput Fluid Dyn
  doi: 10.1007/s00162-020-00518-y
– volume: 96
  start-page: 297
  issue: 2
  year: 1991
  ident: 10.1016/j.compfluid.2025.106890_bib0054
  article-title: Spectral methods for the navier-stokes equations with one infinite and two periodic directions
  publication-title: J Comput Phys
  doi: 10.1016/0021-9991(91)90238-G
– volume: 31
  issue: 12
  year: 2019
  ident: 10.1016/j.compfluid.2025.106890_bib0035
  article-title: Super-resolution reconstruction of turbulent velocity fields using a generative adversarial network-based artificial intelligence framework
  publication-title: Phys Fluid
  doi: 10.1063/1.5127031
– volume: 814
  start-page: 1
  year: 2017
  ident: 10.1016/j.compfluid.2025.106890_bib0021
  article-title: Deep learning in fluid dynamics
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2016.803
– volume: 52
  start-page: 477
  issue: 1
  year: 2020
  ident: 10.1016/j.compfluid.2025.106890_bib0022
  article-title: Machine learning for fluid mechanics
  publication-title: Annu Rev Fluid Mech
  doi: 10.1146/annurev-fluid-010719-060214
– volume: 30
  start-page: 31
  issue: 1
  year: 1998
  ident: 10.1016/j.compfluid.2025.106890_bib0008
  article-title: Vortex dynamics in turbulence
  publication-title: Annu Rev Fluid Mech
  doi: 10.1146/annurev.fluid.30.1.31
– volume: 1002
  year: 2025
  ident: 10.1016/j.compfluid.2025.106890_bib0041
  article-title: Super-resolution of turbulence with dynamics in the loss
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2024.1202
– volume: 40
  start-page: 1
  issue: 6
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0010
  article-title: Predicting high-resolution turbulence details in space and time
  publication-title: ACM Transact Graph (TOG)
  doi: 10.1145/3478513.3480492
– volume: 879
  start-page: 217
  year: 2019
  ident: 10.1016/j.compfluid.2025.106890_bib0027
  article-title: Data-driven prediction of unsteady flow over a circular cylinder using deep learning
  publication-title: J Fluid Mech
  doi: 10.1017/jfm.2019.700
– volume: 43
  start-page: 39
  year: 2007
  ident: 10.1016/j.compfluid.2025.106890_bib0007
  article-title: Spatial resolution of piv for the measurement of turbulence
  publication-title: Exp Fluid
  doi: 10.1007/s00348-007-0319-x
– volume: 33
  issue: 12
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0036
  article-title: High-fidelity reconstruction of turbulent flow from spatially limited data using enhanced super-resolution generative adversarial network
  publication-title: Phys Fluid
  doi: 10.1063/5.0066077
– volume: 41
  start-page: 165
  issue: 1
  year: 2009
  ident: 10.1016/j.compfluid.2025.106890_bib0014
  article-title: Study of high-reynolds number isotropic turbulence by direct numerical simulation
  publication-title: Annu Rev Fluid Mech
  doi: 10.1146/annurev.fluid.010908.165203
– volume: 18
  issue: 1
  year: 2006
  ident: 10.1016/j.compfluid.2025.106890_bib0048
  article-title: Scaling of the velocity fluctuations in turbulent channels up to reτ= 2003
  publication-title: Phys Fluid
  doi: 10.1063/1.2162185
– volume: 7
  year: 2014
  ident: 10.1016/j.compfluid.2025.106890_bib0017
  article-title: Deep learning: methods and applications
  publication-title: Foundat Trends® Signal Process
– volume: 120
  start-page: 1090
  issue: 12
  year: 2019
  ident: 10.1016/j.compfluid.2025.106890_bib0019
  article-title: Lipid metabolism in cancer cells under metabolic stress
  publication-title: Br J Cancer
  doi: 10.1038/s41416-019-0451-4
– volume: 33
  issue: 7
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0037
  article-title: From coarse wall measurements to turbulent velocity fields through deep learning
  publication-title: Phys Fluid
  doi: 10.1063/5.0058346
– volume: 9
  start-page: 1
  issue: 1
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0049
  article-title: Cfd simulation of a hyperloop capsule inside a closed environment
  publication-title: Result Eng
– volume: 2
  start-page: 358
  issue: 6
  year: 2022
  ident: 10.1016/j.compfluid.2025.106890_bib0031
  article-title: Enhancing computational fluid dynamics with machine learning
  publication-title: Nat Comput Sci
  doi: 10.1038/s43588-022-00264-7
– volume: 229
  year: 2021
  ident: 10.1016/j.compfluid.2025.106890_bib0006
  article-title: Estimation of the dissipation rate of turbulent kinetic energy: a review
  publication-title: Chem Eng Sci
  doi: 10.1016/j.ces.2020.116133
– ident: 10.1016/j.compfluid.2025.106890_bib0043
– volume: 9
  year: 2002
  ident: 10.1016/j.compfluid.2025.106890_bib0044
SSID ssj0004324
Score 2.4508326
Snippet This study introduces a deep learning framework that uses an enhanced three-dimensional super-resolution generative adversarial network (3D-ESRGAN) to...
SourceID crossref
elsevier
SourceType Index Database
Publisher
StartPage 106890
SubjectTerms 3D-ESRGAN
Deep learning
Random sampling strategy
Super-resolution reconstruction
Three-dimensional turbulent flow
Title Three-dimensional super-resolution reconstruction of turbulent flow using 3D-ESRGAN with random sampling strategy
URI https://dx.doi.org/10.1016/j.compfluid.2025.106890
Volume 305
WOSCitedRecordID wos001620450600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVESC
  databaseName: ScienceDirect Freedom Collection
  issn: 0045-7930
  databaseCode: AIEXJ
  dateStart: 19950101
  customDbUrl:
  isFulltext: true
  dateEnd: 99991231
  titleUrlDefault: https://www.sciencedirect.com
  omitProxy: false
  ssIdentifier: ssj0004324
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3Nb9MwFLdKxwEOaHxpA4Z84BZ5iuw4TrhV0G2gUSEYUsclshNH6lTSrlnG_hD-YJ4_kqbbpLEDlyhyFaf2--U9v-fn30PonaQSHC5LfZlIEokoIZJrRhgXUklq_CBbteRYTCbJdJp-HQz-tGdhLueiqpKrq3T5X0UNbSBsc3T2HuLuOoUGuAehwxXEDtd_FPxKa1IY1n7HuBHUzVKvCPjV_r2B9YI75libJtDA7M4tUdN88TtobASBfSTj798ORxMXrQWrVix-BbU0Seg2CmFpJjb2hdsiEbWFVDlvZkW3aD9tfBDgfLZOKnBa7xRUUlP1NFA9s2SRX8zxrlIGP_fb345d9ecjGKI9GNGPWlAbtfAbMF4TR4Yq0zd5TcxC3tOl4KwmrpToDTXvIg5nRkpLOxRw9CnfXz-xSax9zeB1aYhthttZ1nWUmY4y19EDtEUFT5Mh2hp9Gk8_r8_bMurYvf0YNvIGb_1Pt696eiuZk230xLsgeOSg8xQNdPUMPe4RUz5H5zdAhK-DCG-CCC9K3IEIGxBhCyLcgQgbEGEHItyCCLcgeoF-HIxPPhwRX5yD5DRmFyThBViHIi5T-KJLcJqhIU9ZrMq0kILTXAmhubZbr1zlItIikkUODp0UWijBXqJhtaj0DsIyLAWYBRkyKqNCKcXAMPAwLhXNY0XFLgrbycuWjoMlu0N0u-h9O8mZX0q6JWIGELrr4Vf3f99r9GiN8jdoCJOv99DD_PJiVq_eevz8BVPQnV8
linkProvider Elsevier
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Three-dimensional+super-resolution+reconstruction+of+turbulent+flow+using+3D-ESRGAN+with+random+sampling+strategy&rft.jtitle=Computers+%26+fluids&rft.au=Yu%2C+Linqi&rft.au=Chen%2C+Yanyun&rft.au=Yousif%2C+Mustafa+Z.&rft.au=Lim%2C+Hee-Chang&rft.date=2026-01-30&rft.issn=0045-7930&rft.volume=305&rft.spage=106890&rft_id=info:doi/10.1016%2Fj.compfluid.2025.106890&rft.externalDBID=n%2Fa&rft.externalDocID=10_1016_j_compfluid_2025_106890
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0045-7930&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0045-7930&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0045-7930&client=summon