Machine Learning Enhanced Collision Operator for the Lattice Boltzmann Method Based on Invariant Networks

Integrating machine learning techniques in established numerical solvers represents a modern approach to enhancing computational fluid dynamics simulations. Within the lattice Boltzmann method (LBM), the collision operator serves as an ideal entry point to incorporate machine learning techniques to...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Bedrunka, Mario Christopher, Horstmann, Tobias, Picard, Ben, Reith, Dirk, Foysi, Holger
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Bedrunka, Mario Christopher
Horstmann, Tobias
Picard, Ben
Reith, Dirk
Foysi, Holger
description Integrating machine learning techniques in established numerical solvers represents a modern approach to enhancing computational fluid dynamics simulations. Within the lattice Boltzmann method (LBM), the collision operator serves as an ideal entry point to incorporate machine learning techniques to enhance its accuracy and stability. In this work, an invariant neural network is constructed, acting on an equivariant collision operator, optimizing the relaxation rates of non-physical moments. This optimization enhances robustness to symmetry transformations and ensures consistent behavior across geometric operations. The proposed neural collision operator (NCO) is trained using forced isotropic turbulence simulations driven by spectral forcing, ensuring stable turbulence statistics. The desired performance is achieved by minimizing the energy spectrum discrepancy between direct numerical simulations and underresolved simulations over a specified wave number range. The loss function is further extended to tailor numerical dissipation at high wave numbers, ensuring robustness without compromising accuracy at low and intermediate wave numbers. The NCO's performance is demonstrated using three-dimensional Taylor-Green vortex (TGV) flows, where it accurately predicts the dynamics even in highly underresolved simulations. Compared to other LBM models, such as the BGK and KBC operators, the NCO exhibits superior accuracy while maintaining stability. In addition, the operator shows robust performance in alternative configurations, including turbulent three-dimensional cylinder flow. Finally, an alternative training procedure using time-dependent quantities is introduced. It is based on a reduced TGV model along with newly proposed symmetry boundary conditions. The reduction in memory consumption enables training at higher Reynolds numbers, successfully leading to stable yet accurate simulations.
doi_str_mv 10.48550/arxiv.2412.08229
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2412_08229</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2412_08229</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2412_082293</originalsourceid><addsrcrecordid>eNqFjrsOgkAQRbexMOoHWDk_IAJCgi0Eo4loY08msLgTYZYsG3x8vUjsLW5uc05yhFh6rhNEYehu0Dypd_zA8x038v3dVFCGhSKWcJJomPgGKSvkQpaQ6LqmjjTDpZUGrTZQDbNqgNFaKiTEurbvBpkhk1bpEmLsBnNQjtyjIWQLZ2kf2ty7uZhUWHdy8fuZWO3Ta3JYj1F5a6hB88q_cfkYt_1PfADr80Wz</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Machine Learning Enhanced Collision Operator for the Lattice Boltzmann Method Based on Invariant Networks</title><source>arXiv.org</source><creator>Bedrunka, Mario Christopher ; Horstmann, Tobias ; Picard, Ben ; Reith, Dirk ; Foysi, Holger</creator><creatorcontrib>Bedrunka, Mario Christopher ; Horstmann, Tobias ; Picard, Ben ; Reith, Dirk ; Foysi, Holger</creatorcontrib><description>Integrating machine learning techniques in established numerical solvers represents a modern approach to enhancing computational fluid dynamics simulations. Within the lattice Boltzmann method (LBM), the collision operator serves as an ideal entry point to incorporate machine learning techniques to enhance its accuracy and stability. In this work, an invariant neural network is constructed, acting on an equivariant collision operator, optimizing the relaxation rates of non-physical moments. This optimization enhances robustness to symmetry transformations and ensures consistent behavior across geometric operations. The proposed neural collision operator (NCO) is trained using forced isotropic turbulence simulations driven by spectral forcing, ensuring stable turbulence statistics. The desired performance is achieved by minimizing the energy spectrum discrepancy between direct numerical simulations and underresolved simulations over a specified wave number range. The loss function is further extended to tailor numerical dissipation at high wave numbers, ensuring robustness without compromising accuracy at low and intermediate wave numbers. The NCO's performance is demonstrated using three-dimensional Taylor-Green vortex (TGV) flows, where it accurately predicts the dynamics even in highly underresolved simulations. Compared to other LBM models, such as the BGK and KBC operators, the NCO exhibits superior accuracy while maintaining stability. In addition, the operator shows robust performance in alternative configurations, including turbulent three-dimensional cylinder flow. Finally, an alternative training procedure using time-dependent quantities is introduced. It is based on a reduced TGV model along with newly proposed symmetry boundary conditions. The reduction in memory consumption enables training at higher Reynolds numbers, successfully leading to stable yet accurate simulations.</description><identifier>DOI: 10.48550/arxiv.2412.08229</identifier><language>eng</language><subject>Physics - Computational Physics</subject><creationdate>2024-12</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2412.08229$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2412.08229$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Bedrunka, Mario Christopher</creatorcontrib><creatorcontrib>Horstmann, Tobias</creatorcontrib><creatorcontrib>Picard, Ben</creatorcontrib><creatorcontrib>Reith, Dirk</creatorcontrib><creatorcontrib>Foysi, Holger</creatorcontrib><title>Machine Learning Enhanced Collision Operator for the Lattice Boltzmann Method Based on Invariant Networks</title><description>Integrating machine learning techniques in established numerical solvers represents a modern approach to enhancing computational fluid dynamics simulations. Within the lattice Boltzmann method (LBM), the collision operator serves as an ideal entry point to incorporate machine learning techniques to enhance its accuracy and stability. In this work, an invariant neural network is constructed, acting on an equivariant collision operator, optimizing the relaxation rates of non-physical moments. This optimization enhances robustness to symmetry transformations and ensures consistent behavior across geometric operations. The proposed neural collision operator (NCO) is trained using forced isotropic turbulence simulations driven by spectral forcing, ensuring stable turbulence statistics. The desired performance is achieved by minimizing the energy spectrum discrepancy between direct numerical simulations and underresolved simulations over a specified wave number range. The loss function is further extended to tailor numerical dissipation at high wave numbers, ensuring robustness without compromising accuracy at low and intermediate wave numbers. The NCO's performance is demonstrated using three-dimensional Taylor-Green vortex (TGV) flows, where it accurately predicts the dynamics even in highly underresolved simulations. Compared to other LBM models, such as the BGK and KBC operators, the NCO exhibits superior accuracy while maintaining stability. In addition, the operator shows robust performance in alternative configurations, including turbulent three-dimensional cylinder flow. Finally, an alternative training procedure using time-dependent quantities is introduced. It is based on a reduced TGV model along with newly proposed symmetry boundary conditions. The reduction in memory consumption enables training at higher Reynolds numbers, successfully leading to stable yet accurate simulations.</description><subject>Physics - Computational Physics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNqFjrsOgkAQRbexMOoHWDk_IAJCgi0Eo4loY08msLgTYZYsG3x8vUjsLW5uc05yhFh6rhNEYehu0Dypd_zA8x038v3dVFCGhSKWcJJomPgGKSvkQpaQ6LqmjjTDpZUGrTZQDbNqgNFaKiTEurbvBpkhk1bpEmLsBnNQjtyjIWQLZ2kf2ty7uZhUWHdy8fuZWO3Ta3JYj1F5a6hB88q_cfkYt_1PfADr80Wz</recordid><startdate>20241211</startdate><enddate>20241211</enddate><creator>Bedrunka, Mario Christopher</creator><creator>Horstmann, Tobias</creator><creator>Picard, Ben</creator><creator>Reith, Dirk</creator><creator>Foysi, Holger</creator><scope>GOX</scope></search><sort><creationdate>20241211</creationdate><title>Machine Learning Enhanced Collision Operator for the Lattice Boltzmann Method Based on Invariant Networks</title><author>Bedrunka, Mario Christopher ; Horstmann, Tobias ; Picard, Ben ; Reith, Dirk ; Foysi, Holger</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2412_082293</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Physics - Computational Physics</topic><toplevel>online_resources</toplevel><creatorcontrib>Bedrunka, Mario Christopher</creatorcontrib><creatorcontrib>Horstmann, Tobias</creatorcontrib><creatorcontrib>Picard, Ben</creatorcontrib><creatorcontrib>Reith, Dirk</creatorcontrib><creatorcontrib>Foysi, Holger</creatorcontrib><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Bedrunka, Mario Christopher</au><au>Horstmann, Tobias</au><au>Picard, Ben</au><au>Reith, Dirk</au><au>Foysi, Holger</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Machine Learning Enhanced Collision Operator for the Lattice Boltzmann Method Based on Invariant Networks</atitle><date>2024-12-11</date><risdate>2024</risdate><abstract>Integrating machine learning techniques in established numerical solvers represents a modern approach to enhancing computational fluid dynamics simulations. Within the lattice Boltzmann method (LBM), the collision operator serves as an ideal entry point to incorporate machine learning techniques to enhance its accuracy and stability. In this work, an invariant neural network is constructed, acting on an equivariant collision operator, optimizing the relaxation rates of non-physical moments. This optimization enhances robustness to symmetry transformations and ensures consistent behavior across geometric operations. The proposed neural collision operator (NCO) is trained using forced isotropic turbulence simulations driven by spectral forcing, ensuring stable turbulence statistics. The desired performance is achieved by minimizing the energy spectrum discrepancy between direct numerical simulations and underresolved simulations over a specified wave number range. The loss function is further extended to tailor numerical dissipation at high wave numbers, ensuring robustness without compromising accuracy at low and intermediate wave numbers. The NCO's performance is demonstrated using three-dimensional Taylor-Green vortex (TGV) flows, where it accurately predicts the dynamics even in highly underresolved simulations. Compared to other LBM models, such as the BGK and KBC operators, the NCO exhibits superior accuracy while maintaining stability. In addition, the operator shows robust performance in alternative configurations, including turbulent three-dimensional cylinder flow. Finally, an alternative training procedure using time-dependent quantities is introduced. It is based on a reduced TGV model along with newly proposed symmetry boundary conditions. The reduction in memory consumption enables training at higher Reynolds numbers, successfully leading to stable yet accurate simulations.</abstract><doi>10.48550/arxiv.2412.08229</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2412.08229
ispartof
issn
language eng
recordid cdi_arxiv_primary_2412_08229
source arXiv.org
subjects Physics - Computational Physics
title Machine Learning Enhanced Collision Operator for the Lattice Boltzmann Method Based on Invariant Networks
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-06T13%3A03%3A48IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Machine%20Learning%20Enhanced%20Collision%20Operator%20for%20the%20Lattice%20Boltzmann%20Method%20Based%20on%20Invariant%20Networks&rft.au=Bedrunka,%20Mario%20Christopher&rft.date=2024-12-11&rft_id=info:doi/10.48550/arxiv.2412.08229&rft_dat=%3Carxiv_GOX%3E2412_08229%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true