No Cover Image

Journal article 484 views

Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network

Jinlong Fu Orcid Logo, Min Wang Orcid Logo, Dunhui Xiao, Shan Zhong Orcid Logo, Xiangyun Ge Orcid Logo, Minglu Wu, Ben Evans Orcid Logo

Computer Methods in Applied Mechanics and Engineering, Volume: 410, Start page: 116049

Swansea University Authors: Jinlong Fu Orcid Logo, Ben Evans Orcid Logo

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

Check full text

DOI (Published version): 10.1016/j.cma.2023.116049

Abstract

The relationships between porous microstructures and transport properties are of fundamental importance in various scientific and engineering applications. Due to the intricacy, stochasticity and heterogeneity of porous media, reliable characterization and modeling of transport properties often requ...

Full description

Published in: Computer Methods in Applied Mechanics and Engineering
ISSN: 0045-7825
Published: Elsevier BV 2023
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa63170
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2023-04-17T09:48:01Z
last_indexed 2023-04-20T03:23:42Z
id cronfa63170
recordtype SURis
fullrecord <?xml version="1.0" encoding="utf-8"?><rfc1807 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><bib-version>v2</bib-version><id>63170</id><entry>2023-04-17</entry><title>Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network</title><swanseaauthors><author><sid>e870d228a5035d2ef500eacbfc9f0302</sid><ORCID>0000-0003-2964-4777</ORCID><firstname>Jinlong</firstname><surname>Fu</surname><name>Jinlong Fu</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>3d273fecc8121fe6b53b8fe5281b9c97</sid><ORCID>0000-0003-3662-9583</ORCID><firstname>Ben</firstname><surname>Evans</surname><name>Ben Evans</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-04-17</date><deptcode>AERO</deptcode><abstract>The relationships between porous microstructures and transport properties are of fundamental importance in various scientific and engineering applications. Due to the intricacy, stochasticity and heterogeneity of porous media, reliable characterization and modeling of transport properties often require a complete dataset of internal microstructure samples. However, it is often an unbearable cost to acquire sufficient 3D digital microstructures by purely using microscopic imaging systems. This paper presents a machine learning-based technique to hierarchically reconstruct 3D well-connected porous microstructures from one isotropic or several anisotropic low-cost 2D exemplar(s). To compactly characterize the large-scale microstructural features, a Gaussian image pyramid is built for each 2D exemplar. Local morphology patterns are collected from the Gaussian image pyramids, and then they serve as the training data to embed the 2D morphological statistics into feed-forward neural networks at multiple length levels. By using a specially-developed morphology integration scheme, the 3D morphological statistics at different levels can be inferred from the statistics-informed neural networks. Gibbs sampling is adopted to hierarchically reconstruct 3D microstructures by using multi-level 3D morphological statistics, where the large-scale, regional and local morphological patterns are statistically generated and successively added to the same 3D random field. The proposed method is tested on a series of porous media with distinct morphologies, and the statistical equivalence between the reconstructed and the real microstructures is systematically evaluated by comparing morphological descriptors and transport properties. The results demonstrate that the proposed 2D-to-3D microstructure reconstruction method is a universal and efficient approach to generating morphologically and physically realistic samples of porous media.</abstract><type>Journal Article</type><journal>Computer Methods in Applied Mechanics and Engineering</journal><volume>410</volume><journalNumber/><paginationStart>116049</paginationStart><paginationEnd/><publisher>Elsevier BV</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>0045-7825</issnPrint><issnElectronic/><keywords/><publishedDay>1</publishedDay><publishedMonth>5</publishedMonth><publishedYear>2023</publishedYear><publishedDate>2023-05-01</publishedDate><doi>10.1016/j.cma.2023.116049</doi><url>http://dx.doi.org/10.1016/j.cma.2023.116049</url><notes/><college>COLLEGE NANME</college><department>Aerospace Engineering</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>AERO</DepartmentCode><institution>Swansea University</institution><apcterm/><funders>EPSRC, United Kingdom grant: PURIFY (E P/V 000756/1) and Swansea University Impact Fund, United Kingdom, and we also would like to acknowledge the support of the Supercomputing Wales project, which is part-funded by the European Regional Development Fund (ERDF) via Welsh Government.</funders><projectreference/><lastEdited>2023-05-11T13:15:44.7478614</lastEdited><Created>2023-04-17T10:45:18.5963788</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>Jinlong</firstname><surname>Fu</surname><orcid>0000-0003-2964-4777</orcid><order>1</order></author><author><firstname>Min</firstname><surname>Wang</surname><orcid>0000-0002-4454-2480</orcid><order>2</order></author><author><firstname>Dunhui</firstname><surname>Xiao</surname><order>3</order></author><author><firstname>Shan</firstname><surname>Zhong</surname><orcid>0000-0002-2016-1288</orcid><order>4</order></author><author><firstname>Xiangyun</firstname><surname>Ge</surname><orcid>0000-0002-6743-6306</orcid><order>5</order></author><author><firstname>Minglu</firstname><surname>Wu</surname><order>6</order></author><author><firstname>Ben</firstname><surname>Evans</surname><orcid>0000-0003-3662-9583</orcid><order>7</order></author></authors><documents><document><filename>Under embargo</filename><originalFilename>Under embargo</originalFilename><uploaded>2023-04-19T15:28:43.3685632</uploaded><type>Output</type><contentLength>52982222</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2024-04-14T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling v2 63170 2023-04-17 Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network e870d228a5035d2ef500eacbfc9f0302 0000-0003-2964-4777 Jinlong Fu Jinlong Fu true false 3d273fecc8121fe6b53b8fe5281b9c97 0000-0003-3662-9583 Ben Evans Ben Evans true false 2023-04-17 AERO The relationships between porous microstructures and transport properties are of fundamental importance in various scientific and engineering applications. Due to the intricacy, stochasticity and heterogeneity of porous media, reliable characterization and modeling of transport properties often require a complete dataset of internal microstructure samples. However, it is often an unbearable cost to acquire sufficient 3D digital microstructures by purely using microscopic imaging systems. This paper presents a machine learning-based technique to hierarchically reconstruct 3D well-connected porous microstructures from one isotropic or several anisotropic low-cost 2D exemplar(s). To compactly characterize the large-scale microstructural features, a Gaussian image pyramid is built for each 2D exemplar. Local morphology patterns are collected from the Gaussian image pyramids, and then they serve as the training data to embed the 2D morphological statistics into feed-forward neural networks at multiple length levels. By using a specially-developed morphology integration scheme, the 3D morphological statistics at different levels can be inferred from the statistics-informed neural networks. Gibbs sampling is adopted to hierarchically reconstruct 3D microstructures by using multi-level 3D morphological statistics, where the large-scale, regional and local morphological patterns are statistically generated and successively added to the same 3D random field. The proposed method is tested on a series of porous media with distinct morphologies, and the statistical equivalence between the reconstructed and the real microstructures is systematically evaluated by comparing morphological descriptors and transport properties. The results demonstrate that the proposed 2D-to-3D microstructure reconstruction method is a universal and efficient approach to generating morphologically and physically realistic samples of porous media. Journal Article Computer Methods in Applied Mechanics and Engineering 410 116049 Elsevier BV 0045-7825 1 5 2023 2023-05-01 10.1016/j.cma.2023.116049 http://dx.doi.org/10.1016/j.cma.2023.116049 COLLEGE NANME Aerospace Engineering COLLEGE CODE AERO Swansea University EPSRC, United Kingdom grant: PURIFY (E P/V 000756/1) and Swansea University Impact Fund, United Kingdom, and we also would like to acknowledge the support of the Supercomputing Wales project, which is part-funded by the European Regional Development Fund (ERDF) via Welsh Government. 2023-05-11T13:15:44.7478614 2023-04-17T10:45:18.5963788 Faculty of Science and Engineering School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Aerospace Engineering Jinlong Fu 0000-0003-2964-4777 1 Min Wang 0000-0002-4454-2480 2 Dunhui Xiao 3 Shan Zhong 0000-0002-2016-1288 4 Xiangyun Ge 0000-0002-6743-6306 5 Minglu Wu 6 Ben Evans 0000-0003-3662-9583 7 Under embargo Under embargo 2023-04-19T15:28:43.3685632 Output 52982222 application/pdf Accepted Manuscript true 2024-04-14T00:00:00.0000000 true eng
title Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
spellingShingle Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
Jinlong Fu
Ben Evans
title_short Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
title_full Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
title_fullStr Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
title_full_unstemmed Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
title_sort Hierarchical reconstruction of 3D well-connected porous media from 2D exemplars using statistics-informed neural network
author_id_str_mv e870d228a5035d2ef500eacbfc9f0302
3d273fecc8121fe6b53b8fe5281b9c97
author_id_fullname_str_mv e870d228a5035d2ef500eacbfc9f0302_***_Jinlong Fu
3d273fecc8121fe6b53b8fe5281b9c97_***_Ben Evans
author Jinlong Fu
Ben Evans
author2 Jinlong Fu
Min Wang
Dunhui Xiao
Shan Zhong
Xiangyun Ge
Minglu Wu
Ben Evans
format Journal article
container_title Computer Methods in Applied Mechanics and Engineering
container_volume 410
container_start_page 116049
publishDate 2023
institution Swansea University
issn 0045-7825
doi_str_mv 10.1016/j.cma.2023.116049
publisher Elsevier BV
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
url http://dx.doi.org/10.1016/j.cma.2023.116049
document_store_str 0
active_str 0
description The relationships between porous microstructures and transport properties are of fundamental importance in various scientific and engineering applications. Due to the intricacy, stochasticity and heterogeneity of porous media, reliable characterization and modeling of transport properties often require a complete dataset of internal microstructure samples. However, it is often an unbearable cost to acquire sufficient 3D digital microstructures by purely using microscopic imaging systems. This paper presents a machine learning-based technique to hierarchically reconstruct 3D well-connected porous microstructures from one isotropic or several anisotropic low-cost 2D exemplar(s). To compactly characterize the large-scale microstructural features, a Gaussian image pyramid is built for each 2D exemplar. Local morphology patterns are collected from the Gaussian image pyramids, and then they serve as the training data to embed the 2D morphological statistics into feed-forward neural networks at multiple length levels. By using a specially-developed morphology integration scheme, the 3D morphological statistics at different levels can be inferred from the statistics-informed neural networks. Gibbs sampling is adopted to hierarchically reconstruct 3D microstructures by using multi-level 3D morphological statistics, where the large-scale, regional and local morphological patterns are statistically generated and successively added to the same 3D random field. The proposed method is tested on a series of porous media with distinct morphologies, and the statistical equivalence between the reconstructed and the real microstructures is systematically evaluated by comparing morphological descriptors and transport properties. The results demonstrate that the proposed 2D-to-3D microstructure reconstruction method is a universal and efficient approach to generating morphologically and physically realistic samples of porous media.
published_date 2023-05-01T13:15:43Z
_version_ 1765599968961232896
score 11.016235

Similar Items