No Cover Image

Journal article 700 views

Model identification of reduced order fluid dynamics systems using deep learning

Z. Wang, Dunhui Xiao Orcid Logo, F. Fang, R. Govindan, C. C. Pain, Y. Guo

International Journal for Numerical Methods in Fluids, Volume: 86, Issue: 4, Pages: 255 - 268

Swansea University Author: Dunhui Xiao Orcid Logo

Full text not available from this repository: check for access using links below.

Check full text

DOI (Published version): 10.1002/fld.4416

Abstract

This paper presents a novel model reduction method: deep learning reduced order model, which is based on proper orthogonal decomposition and deep learning methods. The deep learning approach is a recent technological advancement in the field of artificial neural networks. It has the advantage of lea...

Full description

Published in: International Journal for Numerical Methods in Fluids
ISSN: 0271-2091
Published: Wiley 2018
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa46449
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2018-12-06T20:27:44Z
last_indexed 2023-01-11T14:23:16Z
id cronfa46449
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2022-09-27T16:56:07.1907903</datestamp><bib-version>v2</bib-version><id>46449</id><entry>2018-12-06</entry><title>Model identification of reduced order fluid dynamics systems using deep learning</title><swanseaauthors><author><sid>62c69b98cbcdc9142622d4f398fdab97</sid><ORCID>0000-0003-2461-523X</ORCID><firstname>Dunhui</firstname><surname>Xiao</surname><name>Dunhui Xiao</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2018-12-06</date><deptcode>AERO</deptcode><abstract>This paper presents a novel model reduction method: deep learning reduced order model, which is based on proper orthogonal decomposition and deep learning methods. The deep learning approach is a recent technological advancement in the field of artificial neural networks. It has the advantage of learning the nonlinear system with multiple levels of representation and predicting data. In this work, the training data are obtained from high fidelity model solutions at selected time levels. The long short&#x2010;term memory network is used to construct a set of hypersurfaces representing the reduced fluid dynamic system. The model reduction method developed here is independent of the source code of the full physical system.The reduced order model based on deep learning has been implemented within an unstructured mesh finite element fluid model. The performance of the new reduced order model is evaluated using 2 numerical examples: an ocean gyre and flow past a cylinder. These results illustrate that the CPU cost is reduced by several orders of magnitude whilst providing reasonable accuracy in predictive numerical modelling.</abstract><type>Journal Article</type><journal>International Journal for Numerical Methods in Fluids</journal><volume>86</volume><journalNumber>4</journalNumber><paginationStart>255</paginationStart><paginationEnd>268</paginationEnd><publisher>Wiley</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>0271-2091</issnPrint><issnElectronic/><keywords>deep learning, LSTM, POD, ROM</keywords><publishedDay>11</publishedDay><publishedMonth>1</publishedMonth><publishedYear>2018</publishedYear><publishedDate>2018-01-11</publishedDate><doi>10.1002/fld.4416</doi><url/><notes/><college>COLLEGE NANME</college><department>Aerospace Engineering</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>AERO</DepartmentCode><institution>Swansea University</institution><apcterm/><funders/><projectreference/><lastEdited>2022-09-27T16:56:07.1907903</lastEdited><Created>2018-12-06T14:51:57.8117552</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering</level></path><authors><author><firstname>Z.</firstname><surname>Wang</surname><order>1</order></author><author><firstname>Dunhui</firstname><surname>Xiao</surname><orcid>0000-0003-2461-523X</orcid><order>2</order></author><author><firstname>F.</firstname><surname>Fang</surname><order>3</order></author><author><firstname>R.</firstname><surname>Govindan</surname><order>4</order></author><author><firstname>C. C.</firstname><surname>Pain</surname><order>5</order></author><author><firstname>Y.</firstname><surname>Guo</surname><order>6</order></author></authors><documents/><OutputDurs/></rfc1807>
spelling 2022-09-27T16:56:07.1907903 v2 46449 2018-12-06 Model identification of reduced order fluid dynamics systems using deep learning 62c69b98cbcdc9142622d4f398fdab97 0000-0003-2461-523X Dunhui Xiao Dunhui Xiao true false 2018-12-06 AERO This paper presents a novel model reduction method: deep learning reduced order model, which is based on proper orthogonal decomposition and deep learning methods. The deep learning approach is a recent technological advancement in the field of artificial neural networks. It has the advantage of learning the nonlinear system with multiple levels of representation and predicting data. In this work, the training data are obtained from high fidelity model solutions at selected time levels. The long short‐term memory network is used to construct a set of hypersurfaces representing the reduced fluid dynamic system. The model reduction method developed here is independent of the source code of the full physical system.The reduced order model based on deep learning has been implemented within an unstructured mesh finite element fluid model. The performance of the new reduced order model is evaluated using 2 numerical examples: an ocean gyre and flow past a cylinder. These results illustrate that the CPU cost is reduced by several orders of magnitude whilst providing reasonable accuracy in predictive numerical modelling. Journal Article International Journal for Numerical Methods in Fluids 86 4 255 268 Wiley 0271-2091 deep learning, LSTM, POD, ROM 11 1 2018 2018-01-11 10.1002/fld.4416 COLLEGE NANME Aerospace Engineering COLLEGE CODE AERO Swansea University 2022-09-27T16:56:07.1907903 2018-12-06T14:51:57.8117552 Faculty of Science and Engineering School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering Z. Wang 1 Dunhui Xiao 0000-0003-2461-523X 2 F. Fang 3 R. Govindan 4 C. C. Pain 5 Y. Guo 6
title Model identification of reduced order fluid dynamics systems using deep learning
spellingShingle Model identification of reduced order fluid dynamics systems using deep learning
Dunhui Xiao
title_short Model identification of reduced order fluid dynamics systems using deep learning
title_full Model identification of reduced order fluid dynamics systems using deep learning
title_fullStr Model identification of reduced order fluid dynamics systems using deep learning
title_full_unstemmed Model identification of reduced order fluid dynamics systems using deep learning
title_sort Model identification of reduced order fluid dynamics systems using deep learning
author_id_str_mv 62c69b98cbcdc9142622d4f398fdab97
author_id_fullname_str_mv 62c69b98cbcdc9142622d4f398fdab97_***_Dunhui Xiao
author Dunhui Xiao
author2 Z. Wang
Dunhui Xiao
F. Fang
R. Govindan
C. C. Pain
Y. Guo
format Journal article
container_title International Journal for Numerical Methods in Fluids
container_volume 86
container_issue 4
container_start_page 255
publishDate 2018
institution Swansea University
issn 0271-2091
doi_str_mv 10.1002/fld.4416
publisher Wiley
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering
document_store_str 0
active_str 0
description This paper presents a novel model reduction method: deep learning reduced order model, which is based on proper orthogonal decomposition and deep learning methods. The deep learning approach is a recent technological advancement in the field of artificial neural networks. It has the advantage of learning the nonlinear system with multiple levels of representation and predicting data. In this work, the training data are obtained from high fidelity model solutions at selected time levels. The long short‐term memory network is used to construct a set of hypersurfaces representing the reduced fluid dynamic system. The model reduction method developed here is independent of the source code of the full physical system.The reduced order model based on deep learning has been implemented within an unstructured mesh finite element fluid model. The performance of the new reduced order model is evaluated using 2 numerical examples: an ocean gyre and flow past a cylinder. These results illustrate that the CPU cost is reduced by several orders of magnitude whilst providing reasonable accuracy in predictive numerical modelling.
published_date 2018-01-11T03:58:01Z
_version_ 1763752942627717120
score 11.013731

Similar Items