A deep learning framework for predicting and optimizing flow fields in reactive flows

Computational Fluid Dynamics (CFD) is widely used for solving and optimizing the flow fields of different systems and applications. However, running CFD simulations, especially for reactive flow systems, can be very time consuming and memory intensive, which limits e.g. design space exploration in t...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Chemical engineering journal advances Ročník 25; s. 100966
Hlavní autoři: Gharib, Mohsen, Maleki, Farideh Hoseinian, Rößger, Philip, Gräbner, Martin, Richter, Andreas
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 01.03.2026
Elsevier
Témata:
ISSN:2666-8211, 2666-8211
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Computational Fluid Dynamics (CFD) is widely used for solving and optimizing the flow fields of different systems and applications. However, running CFD simulations, especially for reactive flow systems, can be very time consuming and memory intensive, which limits e.g. design space exploration in the optimization tasks. In this work, a data-driven modeling methodology has been developed to predict 2D flow distributions in a chemical reactor. The primary objective was establishing correlations between global boundary conditions (including process and geometrical parameters), and the resulting CFD flow field distributions. A convolutional autoencoder was used to compress and reduce the data dimensions efficiently. Simultaneously, a multilayer perceptron served as the mapping mechanism that linked the global boundary conditions to the compressed data. The methodology developed in this work provides a very successful demonstration of its ability to map both geometric and process parameters to flow fields. The results showed a prediction accuracy of approximately 94%–97% for the CFD cases, indicating a very high prediction quality. Besides this, the prediction time was less than a second, which is significantly lower compared to the computational effort required for CFD simulations. To demonstrate the practical applicability of this approach, an interactive tool was developed to enable real-time visualization of predicted flow fields. This tool represents a foundational step toward applying digital twins and integrating such models into industrial practice. [Display omitted] •Developed a data-driven surrogate model for predicting 2D reactive flow fields in chemical reactors.•Integrated convolutional autoencoders and multilayer perceptrons to map global boundary conditions to CFD-based flow distributions.•Achieved high computational efficiency in estimating key flow variables within reactive flow systems.•Created an interactive visualization tool enabling real-time design exploration and optimization of reactor performance.
AbstractList Computational Fluid Dynamics (CFD) is widely used for solving and optimizing the flow fields of different systems and applications. However, running CFD simulations, especially for reactive flow systems, can be very time consuming and memory intensive, which limits e.g. design space exploration in the optimization tasks. In this work, a data-driven modeling methodology has been developed to predict 2D flow distributions in a chemical reactor. The primary objective was establishing correlations between global boundary conditions (including process and geometrical parameters), and the resulting CFD flow field distributions. A convolutional autoencoder was used to compress and reduce the data dimensions efficiently. Simultaneously, a multilayer perceptron served as the mapping mechanism that linked the global boundary conditions to the compressed data. The methodology developed in this work provides a very successful demonstration of its ability to map both geometric and process parameters to flow fields. The results showed a prediction accuracy of approximately 94%–97% for the CFD cases, indicating a very high prediction quality. Besides this, the prediction time was less than a second, which is significantly lower compared to the computational effort required for CFD simulations. To demonstrate the practical applicability of this approach, an interactive tool was developed to enable real-time visualization of predicted flow fields. This tool represents a foundational step toward applying digital twins and integrating such models into industrial practice. [Display omitted] •Developed a data-driven surrogate model for predicting 2D reactive flow fields in chemical reactors.•Integrated convolutional autoencoders and multilayer perceptrons to map global boundary conditions to CFD-based flow distributions.•Achieved high computational efficiency in estimating key flow variables within reactive flow systems.•Created an interactive visualization tool enabling real-time design exploration and optimization of reactor performance.
Computational Fluid Dynamics (CFD) is widely used for solving and optimizing the flow fields of different systems and applications. However, running CFD simulations, especially for reactive flow systems, can be very time consuming and memory intensive, which limits e.g. design space exploration in the optimization tasks. In this work, a data-driven modeling methodology has been developed to predict 2D flow distributions in a chemical reactor. The primary objective was establishing correlations between global boundary conditions (including process and geometrical parameters), and the resulting CFD flow field distributions. A convolutional autoencoder was used to compress and reduce the data dimensions efficiently. Simultaneously, a multilayer perceptron served as the mapping mechanism that linked the global boundary conditions to the compressed data. The methodology developed in this work provides a very successful demonstration of its ability to map both geometric and process parameters to flow fields. The results showed a prediction accuracy of approximately 94%–97% for the CFD cases, indicating a very high prediction quality. Besides this, the prediction time was less than a second, which is significantly lower compared to the computational effort required for CFD simulations. To demonstrate the practical applicability of this approach, an interactive tool was developed to enable real-time visualization of predicted flow fields. This tool represents a foundational step towards applying digital twins and integrating such models into industrial practice.
ArticleNumber 100966
Author Gharib, Mohsen
Maleki, Farideh Hoseinian
Gräbner, Martin
Rößger, Philip
Richter, Andreas
Author_xml – sequence: 1
  givenname: Mohsen
  surname: Gharib
  fullname: Gharib, Mohsen
  email: Mohsen.Gharib@ikts.fraunhofer.de
  organization: Fraunhofer Institute for Ceramic Technologies and Systems IKTS, Winterbergstr. 28, 01277 Dresden, Germany
– sequence: 2
  givenname: Farideh Hoseinian
  surname: Maleki
  fullname: Maleki, Farideh Hoseinian
  organization: Institute of Energy Process Engineering and Chemical Engineering IEC, TU Bergakademie Freiberg, Fuchsmühlenweg 9, 09599 Freiberg, Germany
– sequence: 3
  givenname: Philip
  surname: Rößger
  fullname: Rößger, Philip
  organization: Institute of Energy Process Engineering and Chemical Engineering IEC, TU Bergakademie Freiberg, Fuchsmühlenweg 9, 09599 Freiberg, Germany
– sequence: 4
  givenname: Martin
  orcidid: 0000-0002-6474-4012
  surname: Gräbner
  fullname: Gräbner, Martin
  organization: Fraunhofer Institute for Ceramic Technologies and Systems IKTS, Winterbergstr. 28, 01277 Dresden, Germany
– sequence: 5
  givenname: Andreas
  surname: Richter
  fullname: Richter, Andreas
  organization: Fraunhofer Institute for Ceramic Technologies and Systems IKTS, Winterbergstr. 28, 01277 Dresden, Germany
BookMark eNp9kE1qwzAQhUVJoWmaC3SlCyTVyLZsQzch9CcQ6KZZC1kaBbmOZeSQ0J6-clxKV13N8GbeY-a7JZPWt0jIPbAlMBAP9VJjrZac8SwKrBTiiky5EGJRcIDJn_6GzPu-ZozxAiBJYUp2K2oQO9qgCq1r99QGdcCzDx_U-kC7gMbp4zBQraG-O7qD-7rsNf5MrcPG9NS1NKCKaye86P0dubaq6XH-U2dk9_z0vn5dbN9eNuvVdqE5iyflOklZkkOeiBRRqLwsK1SgBMttIgoNhmU5cMwUaF2VRgDYzGZ5atGIlGXJjGzGXONVLbvgDip8Sq-cvAg-7KUKR6cblFUmTFUUeZGWOhXAlch0IkwRG24qrWIWH7N08H0f0P7mAZMDZ1nLgbMcOMuRczQ9jiaMX54cBtlrh62O1ALqYzzD_Wf_BqCNh18
Cites_doi 10.1007/s00466-019-01740-0
10.1126/science.1127647
10.1016/j.cej.2021.132442
10.1016/j.powtec.2024.119836
10.1016/j.jbiomech.2019.109544
10.2516/ogst/2016011
10.1016/j.biombioe.2017.01.029
10.1016/j.compfluid.2019.104393
10.1016/j.fuel.2014.12.004
10.1038/s41598-021-03651-8
10.1016/j.ijheatmasstransfer.2020.120417
10.1016/j.cej.2017.06.061
10.1016/j.combustflame.2023.113215
10.1016/j.fuproc.2016.02.022
10.1016/j.joei.2022.05.003
10.1016/j.fuel.2023.128971
10.1016/j.wasman.2016.08.023
10.1063/1.5094943
10.1016/j.csite.2021.101651
10.1016/j.compchemeng.2017.09.008
10.1016/j.energy.2023.128138
10.1016/j.cej.2022.140367
10.1016/j.combustflame.2023.113094
10.1016/j.fuproc.2016.04.008
10.1016/j.ces.2022.117841
10.1016/j.fuel.2017.03.089
10.1016/j.joei.2022.08.012
10.3390/en15239204
10.1016/j.eswa.2021.116087
10.1016/j.ces.2021.116886
10.1080/08982112.2023.2231064
ContentType Journal Article
Copyright 2025
Copyright_xml – notice: 2025
DBID 6I.
AAFTH
AAYXX
CITATION
DOA
DOI 10.1016/j.ceja.2025.100966
DatabaseName ScienceDirect Open Access Titles
Elsevier:ScienceDirect:Open Access
CrossRef
DOAJ (Directory of Open Access Journals)
DatabaseTitle CrossRef
DatabaseTitleList

Database_xml – sequence: 1
  dbid: DOA
  name: DOAJ Directory of Open Access Journals
  url: https://www.doaj.org/
  sourceTypes: Open Website
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 2666-8211
ExternalDocumentID oai_doaj_org_article_b56db887849c4612a65c36d82a62dbca
10_1016_j_ceja_2025_100966
S2666821125002601
GroupedDBID 0R~
6I.
AAEDW
AAFTH
AALRI
AAXUO
AAYWO
ACVFH
ADCNI
ADVLN
AEUPX
AEXQZ
AFJKZ
AFPUW
AIGII
AITUG
AKBMS
AKYEP
ALMA_UNASSIGNED_HOLDINGS
AMRAJ
APXCP
EBS
FDB
GROUPED_DOAJ
M~E
OK1
AAYXX
CITATION
ID FETCH-LOGICAL-c2066-7c3403717364ee6a799bea1a607f368c1d05712e5a1ccb9d611f5f574fed64053
IEDL.DBID DOA
ISSN 2666-8211
IngestDate Mon Dec 01 19:30:58 EST 2025
Thu Nov 27 01:02:10 EST 2025
Sat Nov 29 17:03:56 EST 2025
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Keywords CFD
Multilayer perceptron
Convolutional autoencoders
Chemical reactors
Neural networks
Language English
License This is an open access article under the CC BY license.
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c2066-7c3403717364ee6a799bea1a607f368c1d05712e5a1ccb9d611f5f574fed64053
ORCID 0000-0002-6474-4012
OpenAccessLink https://doaj.org/article/b56db887849c4612a65c36d82a62dbca
ParticipantIDs doaj_primary_oai_doaj_org_article_b56db887849c4612a65c36d82a62dbca
crossref_primary_10_1016_j_ceja_2025_100966
elsevier_sciencedirect_doi_10_1016_j_ceja_2025_100966
PublicationCentury 2000
PublicationDate 2026-03-01
PublicationDateYYYYMMDD 2026-03-01
PublicationDate_xml – month: 03
  year: 2026
  text: 2026-03-01
  day: 01
PublicationDecade 2020
PublicationTitle Chemical engineering journal advances
PublicationYear 2026
Publisher Elsevier B.V
Elsevier
Publisher_xml – name: Elsevier B.V
– name: Elsevier
References Nadda, Shah, Roy, Yadav (b10) 2023; 8
Ouyang, Vandewalle, Chen, Plehiers, Dobbelaere, Heynderickx, Marin, Van Geem (b27) 2022; 429
Baldi (b37) 2012
Bazai, Kargar, Mehrabi (b9) 2021; 246
Li, Song (b13) 2022; 105
Nadda, Singh, Roy, Yadav (b28) 2024; 441
Gharib, Tischer, Schulze, Gräbner, Richter (b35) 2024; 260
Seo, Yoon, Kim (b22) 2022; 15
Chollet (b39) 2021
Peng, Sun, Zhou, Xie, Han, Xiao (b17) 2022; 260
Förster, Voloshchuk, Richter, Meyer (b6) 2017; 203
Ajorloo, Ghodrat, Scott, Strezov (b3) 2022; 102
Marcato, Santos, Boccardo, Viswanathan, Marchisio, Prodanović (b23) 2023; 455
Zhang (b33) 2018
Cacciarelli, Kulahci (b43) 2023; 35
Guo, Liu, Zhu, Yin (b34) 2017
Hinton, Salakhutdinov (b36) 2006; 313
Samadi, Ghobadian, Nosrati, Rezaei (b14) 2023; 333
Pandey, Das, Pan, Leahy, Kwapinski (b1) 2016; 58
Laubscher, Rousseau (b29) 2020; 163
Peng, Liu, Aubry, Chen, Wu (b32) 2021; 28
Masci, Meier, Cireşan, Schmidhuber (b38) 2011
Rößger, Richter (b25) 2018; 108
Uebel, Rößger, Prüfert, Richter, Meyer (b30) 2016; 148
Voloshchuk, Vascellari, Hasse, Meyer, Richter (b5) 2017; 327
Nemati, Jahangirian (b24) 2023; 17
de Oliveira, Hudebine, Guillaume, Verstraete (b2) 2016; 71
Pandey, Raza, Bhattacharyya (b15) 2023; 351
Bakurov, Buzzelli, Schettini, Castelli, Vanneschi (b42) 2022; 189
Poobathy, Chezian (b41) 2014; 10
Bhatnagar, Afshar, Pan, Duraisamy, Kaushik (b21) 2019; 64
K. Deng, H. Chen, Y. Zhang, Flow structure oriented optimization aided by deep neural network, in: Proceedings of the Tenth International Conference on Computational Fluid Dynamics, ICCFD10, Barcelona, Spain, 2018, pp. 9–13.
Ozaki, Aoyagi (b7) 2022; 12
Ansari, Boosari, Mohaghegh (b8) 2020; 101
Richter, Seifert, Compart, Tischer, Meyer (b4) 2015; 152
Sekar, Jiang, Shu, Khoo (b18) 2019; 31
Baruah, Baruah, Hazarika (b12) 2017; 98
Liu, Luo, Cheng, Liu, Li, Fan, Balachandar (b26) 2023; 258
Guo, Li, Iorio (b11) 2016
Uebel, Rößger, Prüfert, Richter, Meyer (b31) 2016; 149
Kim, Shin, Kim (b16) 2023; 280
Wu, Liu, An, Chen, Lyu (b19) 2020; 198
Liang, Mao, Sun (b40) 2020; 99
Kim (10.1016/j.ceja.2025.100966_b16) 2023; 280
Guo (10.1016/j.ceja.2025.100966_b11) 2016
Sekar (10.1016/j.ceja.2025.100966_b18) 2019; 31
Nadda (10.1016/j.ceja.2025.100966_b10) 2023; 8
Bazai (10.1016/j.ceja.2025.100966_b9) 2021; 246
Laubscher (10.1016/j.ceja.2025.100966_b29) 2020; 163
Liang (10.1016/j.ceja.2025.100966_b40) 2020; 99
Chollet (10.1016/j.ceja.2025.100966_b39) 2021
Ozaki (10.1016/j.ceja.2025.100966_b7) 2022; 12
Bakurov (10.1016/j.ceja.2025.100966_b42) 2022; 189
Wu (10.1016/j.ceja.2025.100966_b19) 2020; 198
Baldi (10.1016/j.ceja.2025.100966_b37) 2012
Gharib (10.1016/j.ceja.2025.100966_b35) 2024; 260
Poobathy (10.1016/j.ceja.2025.100966_b41) 2014; 10
Voloshchuk (10.1016/j.ceja.2025.100966_b5) 2017; 327
de Oliveira (10.1016/j.ceja.2025.100966_b2) 2016; 71
Marcato (10.1016/j.ceja.2025.100966_b23) 2023; 455
Peng (10.1016/j.ceja.2025.100966_b32) 2021; 28
10.1016/j.ceja.2025.100966_b20
Ajorloo (10.1016/j.ceja.2025.100966_b3) 2022; 102
Peng (10.1016/j.ceja.2025.100966_b17) 2022; 260
Förster (10.1016/j.ceja.2025.100966_b6) 2017; 203
Samadi (10.1016/j.ceja.2025.100966_b14) 2023; 333
Pandey (10.1016/j.ceja.2025.100966_b1) 2016; 58
Uebel (10.1016/j.ceja.2025.100966_b31) 2016; 149
Zhang (10.1016/j.ceja.2025.100966_b33) 2018
Li (10.1016/j.ceja.2025.100966_b13) 2022; 105
Hinton (10.1016/j.ceja.2025.100966_b36) 2006; 313
Cacciarelli (10.1016/j.ceja.2025.100966_b43) 2023; 35
Pandey (10.1016/j.ceja.2025.100966_b15) 2023; 351
Baruah (10.1016/j.ceja.2025.100966_b12) 2017; 98
Nemati (10.1016/j.ceja.2025.100966_b24) 2023; 17
Richter (10.1016/j.ceja.2025.100966_b4) 2015; 152
Ansari (10.1016/j.ceja.2025.100966_b8) 2020; 101
Nadda (10.1016/j.ceja.2025.100966_b28) 2024; 441
Ouyang (10.1016/j.ceja.2025.100966_b27) 2022; 429
Masci (10.1016/j.ceja.2025.100966_b38) 2011
Rößger (10.1016/j.ceja.2025.100966_b25) 2018; 108
Liu (10.1016/j.ceja.2025.100966_b26) 2023; 258
Seo (10.1016/j.ceja.2025.100966_b22) 2022; 15
Uebel (10.1016/j.ceja.2025.100966_b30) 2016; 148
Bhatnagar (10.1016/j.ceja.2025.100966_b21) 2019; 64
Guo (10.1016/j.ceja.2025.100966_b34) 2017
References_xml – start-page: 373
  year: 2017
  end-page: 382
  ident: b34
  article-title: Deep clustering with convolutional autoencoders
  publication-title: Neural Information Processing: 24th International Conference, ICONIP 2017, Guangzhou, China, November 14-18, 2017, Proceedings, Part II 24
– volume: 101
  start-page: 1
  year: 2020
  end-page: 8
  ident: b8
  article-title: Successful implementation of artificial intelligence and machine learning in multiphase flow smart proxy modeling: two case studies of gas-liquid and gas-solid CFD models
  publication-title: J. Pet. Environ. Biotechnol.
– volume: 189
  year: 2022
  ident: b42
  article-title: Structural similarity index (SSIM) revisited: A data-driven approach
  publication-title: Expert Syst. Appl.
– volume: 163
  year: 2020
  ident: b29
  article-title: Application of generative deep learning to predict temperature, flow and species distributions using simulation data of a methane combustor
  publication-title: Int. J. Heat Mass Transfer
– volume: 327
  start-page: 307
  year: 2017
  end-page: 319
  ident: b5
  article-title: Numerical study of natural gas reforming by non-catalytic partial oxidation based on the virtuhcon benchmark
  publication-title: Chem. Eng. J.
– volume: 64
  start-page: 525
  year: 2019
  end-page: 545
  ident: b21
  article-title: Prediction of aerodynamic flow fields using convolutional neural networks
  publication-title: Comput. Mech.
– volume: 258
  year: 2023
  ident: b26
  article-title: Surrogate modeling of parameterized multi-dimensional premixed combustion with physics-informed neural networks for rapid exploration of design space
  publication-title: Combust. Flame
– volume: 313
  start-page: 504
  year: 2006
  end-page: 507
  ident: b36
  article-title: Reducing the dimensionality of data with neural networks
  publication-title: Science
– volume: 10
  start-page: 55
  year: 2014
  end-page: 61
  ident: b41
  article-title: Edge detection operators: Peak signal to noise ratio based comparison
  publication-title: IJ Image Graph. Signal Process.
– volume: 260
  year: 2022
  ident: b17
  article-title: The accurate prediction and analysis of bed expansion characteristics in liquid–solid fluidized bed based on machine learning methods
  publication-title: Chem. Eng. Sci.
– volume: 429
  year: 2022
  ident: b27
  article-title: Speeding up turbulent reactive flow simulation via a deep artificial neural network: A methodology study
  publication-title: Chem. Eng. J.
– volume: 441
  year: 2024
  ident: b28
  article-title: A comparative assessment of CFD based LSTM and GRU for hydrodynamic predictions of gas-solid fluidized bed
  publication-title: Powder Technol.
– volume: 71
  start-page: 45
  year: 2016
  ident: b2
  article-title: A review of kinetic modeling methodologies for complex processes
  publication-title: Oil Gas Sci. Technol.–Rev. D’IFP Energies Nouv.
– volume: 12
  start-page: 447
  year: 2022
  ident: b7
  article-title: Prediction of steady flows passing fixed cylinders using deep learning
  publication-title: Sci. Rep.
– volume: 35
  start-page: 741
  year: 2023
  end-page: 750
  ident: b43
  article-title: Hidden dimensions of the data: PCA vs autoencoders
  publication-title: Qual. Eng.
– volume: 58
  start-page: 202
  year: 2016
  end-page: 213
  ident: b1
  article-title: Artificial neural network based modelling approach for municipal solid waste gasification in a fluidized bed reactor
  publication-title: Waste Manage.
– volume: 246
  year: 2021
  ident: b9
  article-title: Using an encoder-decoder convolutional neural network to predict the solid holdup patterns in a pseudo-2d fluidized bed
  publication-title: Chem. Eng. Sci.
– start-page: 37
  year: 2012
  end-page: 49
  ident: b37
  article-title: Autoencoders, unsupervised learning, and deep architectures
  publication-title: Proceedings of ICML Workshop on Unsupervised and Transfer Learning
– volume: 98
  start-page: 264
  year: 2017
  end-page: 271
  ident: b12
  article-title: Artificial neural network based modeling of biomass gasification in fixed bed downdraft gasifiers
  publication-title: Biomass Bioenergy
– volume: 102
  start-page: 395
  year: 2022
  end-page: 419
  ident: b3
  article-title: Recent advances in thermodynamic analysis of biomass gasification: A review on numerical modelling and simulation
  publication-title: J. Energy Inst.
– volume: 15
  start-page: 9204
  year: 2022
  ident: b22
  article-title: Establishment of CNN and encoder–decoder models for the prediction of characteristics of flow and heat transfer around NACA sections
  publication-title: Energies
– year: 2018
  ident: b33
  article-title: A better autoencoder for image: Convolutional autoencoder
  publication-title: ICONIP17-DCEC
– volume: 198
  year: 2020
  ident: b19
  article-title: A deep learning approach for efficiently and accurately evaluating the flow field of supercritical airfoils
  publication-title: Comput. & Fluids
– volume: 108
  start-page: 232
  year: 2018
  end-page: 239
  ident: b25
  article-title: Performance of different optimization concepts for reactive flow systems based on combined CFD and response surface methods
  publication-title: Comput. Chem. Eng.
– volume: 99
  year: 2020
  ident: b40
  article-title: A feasibility study of deep learning for predicting hemodynamics of human thoracic aorta
  publication-title: J. Biomech.
– start-page: 52
  year: 2011
  end-page: 59
  ident: b38
  article-title: Stacked convolutional auto-encoders for hierarchical feature extraction
  publication-title: International Conference on Artificial Neural Networks
– volume: 280
  year: 2023
  ident: b16
  article-title: Predicting biomass composition and operating conditions in fluidized bed biomass gasifiers: An automated machine learning approach combined with cooperative game theory
  publication-title: Energy
– start-page: 481
  year: 2016
  end-page: 490
  ident: b11
  article-title: Convolutional Neural Networks for Steady Flow Approximation
– volume: 105
  start-page: 176
  year: 2022
  end-page: 183
  ident: b13
  article-title: Artificial neural network model of catalytic coal gasification in fixed bed
  publication-title: J. Energy Inst.
– volume: 31
  year: 2019
  ident: b18
  article-title: Fast flow field prediction over airfoils using deep learning approach
  publication-title: Phys. Fluids
– volume: 149
  start-page: 290
  year: 2016
  end-page: 304
  ident: b31
  article-title: CFD-based multi-objective optimization of a quench reactor design
  publication-title: Fuel Process. Technol.
– volume: 260
  year: 2024
  ident: b35
  article-title: Flame lift-off detector based on deep learning neural networks
  publication-title: Combust. Flame
– volume: 333
  year: 2023
  ident: b14
  article-title: Investigation of factors affecting performance of a downdraft fixed bed gasifier using optimized MLP neural networks approach
  publication-title: Fuel
– volume: 203
  start-page: 954
  year: 2017
  end-page: 963
  ident: b6
  article-title: 3D numerical study of the performance of different burner concepts for the high-pressure non-catalytic natural gas reforming based on the Freiberg semi-industrial test facility HP POX
  publication-title: Fuel
– reference: K. Deng, H. Chen, Y. Zhang, Flow structure oriented optimization aided by deep neural network, in: Proceedings of the Tenth International Conference on Computational Fluid Dynamics, ICCFD10, Barcelona, Spain, 2018, pp. 9–13.
– volume: 17
  start-page: 60
  year: 2023
  end-page: 74
  ident: b24
  article-title: A data-driven machine learning approach for turbulent flow field prediction based on direct computational fluid dynamics database
  publication-title: J. Appl. Fluid Mech.
– volume: 455
  year: 2023
  ident: b23
  article-title: Prediction of local concentration fields in porous media with chemical reaction using a multi scale convolutional neural network
  publication-title: Chem. Eng. J.
– volume: 152
  start-page: 110
  year: 2015
  end-page: 121
  ident: b4
  article-title: A large-scale benchmark for the CFD modeling of non-catalytic reforming of natural gas based on the Freiberg test plant HP POX
  publication-title: Fuel
– volume: 28
  year: 2021
  ident: b32
  article-title: Data-driven modeling of geometry-adaptive steady heat conduction based on convolutional neural networks
  publication-title: Case Stud. Therm. Eng.
– volume: 351
  year: 2023
  ident: b15
  article-title: Development of explainable AI-based predictive models for bubbling fluidised bed gasification process
  publication-title: Fuel
– volume: 148
  start-page: 198
  year: 2016
  end-page: 208
  ident: b30
  article-title: A new CO conversion quench reactor design
  publication-title: Fuel Process. Technol.
– year: 2021
  ident: b39
  article-title: Deep learning with python
– volume: 8
  year: 2023
  ident: b10
  article-title: CFD-based deep neural networks (DNN) model for predicting the hydrodynamics of fluidized beds
  publication-title: Digit. Chem. Eng.
– volume: 64
  start-page: 525
  year: 2019
  ident: 10.1016/j.ceja.2025.100966_b21
  article-title: Prediction of aerodynamic flow fields using convolutional neural networks
  publication-title: Comput. Mech.
  doi: 10.1007/s00466-019-01740-0
– volume: 17
  start-page: 60
  issue: 1
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b24
  article-title: A data-driven machine learning approach for turbulent flow field prediction based on direct computational fluid dynamics database
  publication-title: J. Appl. Fluid Mech.
– volume: 313
  start-page: 504
  issue: 5786
  year: 2006
  ident: 10.1016/j.ceja.2025.100966_b36
  article-title: Reducing the dimensionality of data with neural networks
  publication-title: Science
  doi: 10.1126/science.1127647
– volume: 429
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b27
  article-title: Speeding up turbulent reactive flow simulation via a deep artificial neural network: A methodology study
  publication-title: Chem. Eng. J.
  doi: 10.1016/j.cej.2021.132442
– volume: 441
  year: 2024
  ident: 10.1016/j.ceja.2025.100966_b28
  article-title: A comparative assessment of CFD based LSTM and GRU for hydrodynamic predictions of gas-solid fluidized bed
  publication-title: Powder Technol.
  doi: 10.1016/j.powtec.2024.119836
– volume: 99
  year: 2020
  ident: 10.1016/j.ceja.2025.100966_b40
  article-title: A feasibility study of deep learning for predicting hemodynamics of human thoracic aorta
  publication-title: J. Biomech.
  doi: 10.1016/j.jbiomech.2019.109544
– volume: 71
  start-page: 45
  issue: 3
  year: 2016
  ident: 10.1016/j.ceja.2025.100966_b2
  article-title: A review of kinetic modeling methodologies for complex processes
  publication-title: Oil Gas Sci. Technol.–Rev. D’IFP Energies Nouv.
  doi: 10.2516/ogst/2016011
– volume: 98
  start-page: 264
  year: 2017
  ident: 10.1016/j.ceja.2025.100966_b12
  article-title: Artificial neural network based modeling of biomass gasification in fixed bed downdraft gasifiers
  publication-title: Biomass Bioenergy
  doi: 10.1016/j.biombioe.2017.01.029
– volume: 198
  year: 2020
  ident: 10.1016/j.ceja.2025.100966_b19
  article-title: A deep learning approach for efficiently and accurately evaluating the flow field of supercritical airfoils
  publication-title: Comput. & Fluids
  doi: 10.1016/j.compfluid.2019.104393
– volume: 152
  start-page: 110
  year: 2015
  ident: 10.1016/j.ceja.2025.100966_b4
  article-title: A large-scale benchmark for the CFD modeling of non-catalytic reforming of natural gas based on the Freiberg test plant HP POX
  publication-title: Fuel
  doi: 10.1016/j.fuel.2014.12.004
– volume: 10
  start-page: 55
  year: 2014
  ident: 10.1016/j.ceja.2025.100966_b41
  article-title: Edge detection operators: Peak signal to noise ratio based comparison
  publication-title: IJ Image Graph. Signal Process.
– start-page: 373
  year: 2017
  ident: 10.1016/j.ceja.2025.100966_b34
  article-title: Deep clustering with convolutional autoencoders
– volume: 12
  start-page: 447
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b7
  article-title: Prediction of steady flows passing fixed cylinders using deep learning
  publication-title: Sci. Rep.
  doi: 10.1038/s41598-021-03651-8
– start-page: 481
  year: 2016
  ident: 10.1016/j.ceja.2025.100966_b11
– volume: 163
  year: 2020
  ident: 10.1016/j.ceja.2025.100966_b29
  article-title: Application of generative deep learning to predict temperature, flow and species distributions using simulation data of a methane combustor
  publication-title: Int. J. Heat Mass Transfer
  doi: 10.1016/j.ijheatmasstransfer.2020.120417
– volume: 327
  start-page: 307
  year: 2017
  ident: 10.1016/j.ceja.2025.100966_b5
  article-title: Numerical study of natural gas reforming by non-catalytic partial oxidation based on the virtuhcon benchmark
  publication-title: Chem. Eng. J.
  doi: 10.1016/j.cej.2017.06.061
– volume: 260
  year: 2024
  ident: 10.1016/j.ceja.2025.100966_b35
  article-title: Flame lift-off detector based on deep learning neural networks
  publication-title: Combust. Flame
  doi: 10.1016/j.combustflame.2023.113215
– volume: 148
  start-page: 198
  year: 2016
  ident: 10.1016/j.ceja.2025.100966_b30
  article-title: A new CO conversion quench reactor design
  publication-title: Fuel Process. Technol.
  doi: 10.1016/j.fuproc.2016.02.022
– volume: 102
  start-page: 395
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b3
  article-title: Recent advances in thermodynamic analysis of biomass gasification: A review on numerical modelling and simulation
  publication-title: J. Energy Inst.
  doi: 10.1016/j.joei.2022.05.003
– volume: 351
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b15
  article-title: Development of explainable AI-based predictive models for bubbling fluidised bed gasification process
  publication-title: Fuel
  doi: 10.1016/j.fuel.2023.128971
– start-page: 52
  year: 2011
  ident: 10.1016/j.ceja.2025.100966_b38
  article-title: Stacked convolutional auto-encoders for hierarchical feature extraction
– volume: 58
  start-page: 202
  year: 2016
  ident: 10.1016/j.ceja.2025.100966_b1
  article-title: Artificial neural network based modelling approach for municipal solid waste gasification in a fluidized bed reactor
  publication-title: Waste Manage.
  doi: 10.1016/j.wasman.2016.08.023
– volume: 333
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b14
  article-title: Investigation of factors affecting performance of a downdraft fixed bed gasifier using optimized MLP neural networks approach
  publication-title: Fuel
– volume: 31
  issue: 5
  year: 2019
  ident: 10.1016/j.ceja.2025.100966_b18
  article-title: Fast flow field prediction over airfoils using deep learning approach
  publication-title: Phys. Fluids
  doi: 10.1063/1.5094943
– start-page: 37
  year: 2012
  ident: 10.1016/j.ceja.2025.100966_b37
  article-title: Autoencoders, unsupervised learning, and deep architectures
– year: 2021
  ident: 10.1016/j.ceja.2025.100966_b39
– volume: 8
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b10
  article-title: CFD-based deep neural networks (DNN) model for predicting the hydrodynamics of fluidized beds
  publication-title: Digit. Chem. Eng.
– volume: 28
  year: 2021
  ident: 10.1016/j.ceja.2025.100966_b32
  article-title: Data-driven modeling of geometry-adaptive steady heat conduction based on convolutional neural networks
  publication-title: Case Stud. Therm. Eng.
  doi: 10.1016/j.csite.2021.101651
– volume: 108
  start-page: 232
  year: 2018
  ident: 10.1016/j.ceja.2025.100966_b25
  article-title: Performance of different optimization concepts for reactive flow systems based on combined CFD and response surface methods
  publication-title: Comput. Chem. Eng.
  doi: 10.1016/j.compchemeng.2017.09.008
– volume: 101
  start-page: 1
  issue: 1
  year: 2020
  ident: 10.1016/j.ceja.2025.100966_b8
  article-title: Successful implementation of artificial intelligence and machine learning in multiphase flow smart proxy modeling: two case studies of gas-liquid and gas-solid CFD models
  publication-title: J. Pet. Environ. Biotechnol.
– volume: 280
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b16
  article-title: Predicting biomass composition and operating conditions in fluidized bed biomass gasifiers: An automated machine learning approach combined with cooperative game theory
  publication-title: Energy
  doi: 10.1016/j.energy.2023.128138
– volume: 455
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b23
  article-title: Prediction of local concentration fields in porous media with chemical reaction using a multi scale convolutional neural network
  publication-title: Chem. Eng. J.
  doi: 10.1016/j.cej.2022.140367
– volume: 258
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b26
  article-title: Surrogate modeling of parameterized multi-dimensional premixed combustion with physics-informed neural networks for rapid exploration of design space
  publication-title: Combust. Flame
  doi: 10.1016/j.combustflame.2023.113094
– volume: 149
  start-page: 290
  year: 2016
  ident: 10.1016/j.ceja.2025.100966_b31
  article-title: CFD-based multi-objective optimization of a quench reactor design
  publication-title: Fuel Process. Technol.
  doi: 10.1016/j.fuproc.2016.04.008
– ident: 10.1016/j.ceja.2025.100966_b20
– volume: 260
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b17
  article-title: The accurate prediction and analysis of bed expansion characteristics in liquid–solid fluidized bed based on machine learning methods
  publication-title: Chem. Eng. Sci.
  doi: 10.1016/j.ces.2022.117841
– volume: 203
  start-page: 954
  year: 2017
  ident: 10.1016/j.ceja.2025.100966_b6
  article-title: 3D numerical study of the performance of different burner concepts for the high-pressure non-catalytic natural gas reforming based on the Freiberg semi-industrial test facility HP POX
  publication-title: Fuel
  doi: 10.1016/j.fuel.2017.03.089
– volume: 105
  start-page: 176
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b13
  article-title: Artificial neural network model of catalytic coal gasification in fixed bed
  publication-title: J. Energy Inst.
  doi: 10.1016/j.joei.2022.08.012
– volume: 15
  start-page: 9204
  issue: 23
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b22
  article-title: Establishment of CNN and encoder–decoder models for the prediction of characteristics of flow and heat transfer around NACA sections
  publication-title: Energies
  doi: 10.3390/en15239204
– volume: 189
  year: 2022
  ident: 10.1016/j.ceja.2025.100966_b42
  article-title: Structural similarity index (SSIM) revisited: A data-driven approach
  publication-title: Expert Syst. Appl.
  doi: 10.1016/j.eswa.2021.116087
– year: 2018
  ident: 10.1016/j.ceja.2025.100966_b33
  article-title: A better autoencoder for image: Convolutional autoencoder
– volume: 246
  year: 2021
  ident: 10.1016/j.ceja.2025.100966_b9
  article-title: Using an encoder-decoder convolutional neural network to predict the solid holdup patterns in a pseudo-2d fluidized bed
  publication-title: Chem. Eng. Sci.
  doi: 10.1016/j.ces.2021.116886
– volume: 35
  start-page: 741
  issue: 4
  year: 2023
  ident: 10.1016/j.ceja.2025.100966_b43
  article-title: Hidden dimensions of the data: PCA vs autoencoders
  publication-title: Qual. Eng.
  doi: 10.1080/08982112.2023.2231064
SSID ssj0002811341
Score 2.3196874
Snippet Computational Fluid Dynamics (CFD) is widely used for solving and optimizing the flow fields of different systems and applications. However, running CFD...
SourceID doaj
crossref
elsevier
SourceType Open Website
Index Database
Publisher
StartPage 100966
SubjectTerms CFD
Chemical reactors
Convolutional autoencoders
Multilayer perceptron
Neural networks
Title A deep learning framework for predicting and optimizing flow fields in reactive flows
URI https://dx.doi.org/10.1016/j.ceja.2025.100966
https://doaj.org/article/b56db887849c4612a65c36d82a62dbca
Volume 25
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVAON
  databaseName: DOAJ Directory of Open Access Journals
  customDbUrl:
  eissn: 2666-8211
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0002811341
  issn: 2666-8211
  databaseCode: DOA
  dateStart: 20200101
  isFulltext: true
  titleUrlDefault: https://www.doaj.org/
  providerName: Directory of Open Access Journals
– providerCode: PRVHPJ
  databaseName: ROAD: Directory of Open Access Scholarly Resources
  customDbUrl:
  eissn: 2666-8211
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0002811341
  issn: 2666-8211
  databaseCode: M~E
  dateStart: 20200101
  isFulltext: true
  titleUrlDefault: https://road.issn.org
  providerName: ISSN International Centre
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV05T8MwGLVQxQAD4hTlkgc2FFE3PuKxoFYsVAxU6mb5RKkgrdoCEgO_HR8JZIKFJYosy47el_h7iV_eB8Bl7ojrK2QzjWjYZpQ240z3soI67MmcpMalYhNsPC6mU_7QKvUVNGHJHjgBd60INco_CQXmGvt0LCnROTWFP-kbpSM18qyn9TI1i5-MUHAqq_-SSYIubWfBaKhPgjKAR1vEn0wUDftbCamVZEa7YKdmh3CQrmoPbNhqH2y3PAMPwGQAjbULWJd7eIKukVdBzz_hYhl2XoKWGcrKwLlfEV7Kj9jvef4Oo2BtBcsKerIYl7rYvjoEk9Hw8fYuq2sjZDoYsGdM5zi47bGcYmupZJwrK5GkPeZyWmhkPBFDfUsk0lpxQxHyUSEMO2uoJ2n5EehU88oeA0i4xRhLK4NbWS6pdNQxp6hTjgR3tC64anASi2SBIRpt2EwEVEVAVSRUu-AmQPndM9hXxwYfVFEHVfwV1C4gTSBEzQRShvdDlb9MfvIfk5-CLT8kTUKzM9BZL1_tOdjUb-tytbyI95k_3n8OvwAeUNh1
linkProvider Directory of Open Access Journals
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=A+deep+learning+framework+for+predicting+and+optimizing+flow+fields+in+reactive+flows&rft.jtitle=Chemical+engineering+journal+advances&rft.au=Gharib%2C+Mohsen&rft.au=Maleki%2C+Farideh+Hoseinian&rft.au=R%C3%B6%C3%9Fger%2C+Philip&rft.au=Gr%C3%A4bner%2C+Martin&rft.date=2026-03-01&rft.pub=Elsevier+B.V&rft.issn=2666-8211&rft.eissn=2666-8211&rft_id=info:doi/10.1016%2Fj.ceja.2025.100966&rft.externalDocID=S2666821125002601
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2666-8211&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2666-8211&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2666-8211&client=summon