Bayesian Learning of Gas Transport in Three-Dimensional Fracture Networks
Modeling gas flow through fractures of subsurface rock is a particularly challenging problem because of the heterogeneous nature of the material. High-fidelity simulations using discrete fracture network (DFN) models are one methodology for predicting gas particle breakthrough times at the surface,...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2023-06 |
---|---|
Hauptverfasser: | , , , , , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Shi, Yingqi Berry, Donald J Kath, John Shams Lodhy Ly, An Percus, Allon G Hyman, Jeffrey D Moran, Kelly Strait, Justin Sweeney, Matthew R Viswanathan, Hari S Stauffer, Philip H |
description | Modeling gas flow through fractures of subsurface rock is a particularly challenging problem because of the heterogeneous nature of the material. High-fidelity simulations using discrete fracture network (DFN) models are one methodology for predicting gas particle breakthrough times at the surface, but are computationally demanding. We propose a Bayesian machine learning method that serves as an efficient surrogate model, or emulator, for these three-dimensional DFN simulations. Our model trains on a small quantity of simulation data and, using a graph/path-based decomposition of the fracture network, rapidly predicts quantiles of the breakthrough time distribution. The approach, based on Gaussian Process Regression (GPR), outputs predictions that are within 20-30% of high-fidelity DFN simulation results. Unlike previously proposed methods, it also provides uncertainty quantification, outputting confidence intervals that are essential given the uncertainty inherent in subsurface modeling. Our trained model runs within a fraction of a second, which is considerably faster than other methods with comparable accuracy and multiple orders of magnitude faster than high-fidelity simulations. |
doi_str_mv | 10.48550/arxiv.2306.03416 |
format | Article |
fullrecord | <record><control><sourceid>proquest_arxiv</sourceid><recordid>TN_cdi_arxiv_primary_2306_03416</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2823311792</sourcerecordid><originalsourceid>FETCH-LOGICAL-a956-83cce656accbffc1530a6c3271deb7dcf014863a19329b92d1d05af1823d82733</originalsourceid><addsrcrecordid>eNotkMtOwzAURC0kJKrSD2CFJdYp9r2x4yyh0FKpgk320a3jgEvrFDsF-vf0wWo2R6M5w9iNFOPcKCXuKf767zGg0GOBudQXbACIMjM5wBUbpbQSQoAuQCkcsPkj7V3yFPjCUQw-vPOu5TNKvIoU0raLPfeBVx_RuezJb1xIvgu05tNItt9Fx19d_9PFz3TNLltaJzf6zyGrps_V5CVbvM3mk4dFRqXSmUFrnVaarF22rZUKBWmLUMjGLYvGtkLmRiPJEqFcltDIRihqpQFsDBSIQ3Z7rj151tvoNxT39dG3PvkeiLszsY3d186lvl51u3jYnGowxytkUQL-AXMHV-o</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2823311792</pqid></control><display><type>article</type><title>Bayesian Learning of Gas Transport in Three-Dimensional Fracture Networks</title><source>arXiv.org</source><source>Free E- Journals</source><creator>Shi, Yingqi ; Berry, Donald J ; Kath, John ; Shams Lodhy ; Ly, An ; Percus, Allon G ; Hyman, Jeffrey D ; Moran, Kelly ; Strait, Justin ; Sweeney, Matthew R ; Viswanathan, Hari S ; Stauffer, Philip H</creator><creatorcontrib>Shi, Yingqi ; Berry, Donald J ; Kath, John ; Shams Lodhy ; Ly, An ; Percus, Allon G ; Hyman, Jeffrey D ; Moran, Kelly ; Strait, Justin ; Sweeney, Matthew R ; Viswanathan, Hari S ; Stauffer, Philip H</creatorcontrib><description>Modeling gas flow through fractures of subsurface rock is a particularly challenging problem because of the heterogeneous nature of the material. High-fidelity simulations using discrete fracture network (DFN) models are one methodology for predicting gas particle breakthrough times at the surface, but are computationally demanding. We propose a Bayesian machine learning method that serves as an efficient surrogate model, or emulator, for these three-dimensional DFN simulations. Our model trains on a small quantity of simulation data and, using a graph/path-based decomposition of the fracture network, rapidly predicts quantiles of the breakthrough time distribution. The approach, based on Gaussian Process Regression (GPR), outputs predictions that are within 20-30% of high-fidelity DFN simulation results. Unlike previously proposed methods, it also provides uncertainty quantification, outputting confidence intervals that are essential given the uncertainty inherent in subsurface modeling. Our trained model runs within a fraction of a second, which is considerably faster than other methods with comparable accuracy and multiple orders of magnitude faster than high-fidelity simulations.</description><identifier>EISSN: 2331-8422</identifier><identifier>DOI: 10.48550/arxiv.2306.03416</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Accuracy ; Bayesian analysis ; Confidence intervals ; Fractures ; Gas flow ; Gas transport ; Gaussian process ; Machine learning ; Modelling ; Physics - Data Analysis, Statistics and Probability ; Physics - Geophysics ; Quantiles ; Simulation ; Statistical analysis ; Three dimensional models ; Uncertainty</subject><ispartof>arXiv.org, 2023-06</ispartof><rights>2023. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,784,885,27925</link.rule.ids><backlink>$$Uhttps://doi.org/10.1016/j.cageo.2024.105700$$DView published paper (Access to full text may be restricted)$$Hfree_for_read</backlink><backlink>$$Uhttps://doi.org/10.48550/arXiv.2306.03416$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Shi, Yingqi</creatorcontrib><creatorcontrib>Berry, Donald J</creatorcontrib><creatorcontrib>Kath, John</creatorcontrib><creatorcontrib>Shams Lodhy</creatorcontrib><creatorcontrib>Ly, An</creatorcontrib><creatorcontrib>Percus, Allon G</creatorcontrib><creatorcontrib>Hyman, Jeffrey D</creatorcontrib><creatorcontrib>Moran, Kelly</creatorcontrib><creatorcontrib>Strait, Justin</creatorcontrib><creatorcontrib>Sweeney, Matthew R</creatorcontrib><creatorcontrib>Viswanathan, Hari S</creatorcontrib><creatorcontrib>Stauffer, Philip H</creatorcontrib><title>Bayesian Learning of Gas Transport in Three-Dimensional Fracture Networks</title><title>arXiv.org</title><description>Modeling gas flow through fractures of subsurface rock is a particularly challenging problem because of the heterogeneous nature of the material. High-fidelity simulations using discrete fracture network (DFN) models are one methodology for predicting gas particle breakthrough times at the surface, but are computationally demanding. We propose a Bayesian machine learning method that serves as an efficient surrogate model, or emulator, for these three-dimensional DFN simulations. Our model trains on a small quantity of simulation data and, using a graph/path-based decomposition of the fracture network, rapidly predicts quantiles of the breakthrough time distribution. The approach, based on Gaussian Process Regression (GPR), outputs predictions that are within 20-30% of high-fidelity DFN simulation results. Unlike previously proposed methods, it also provides uncertainty quantification, outputting confidence intervals that are essential given the uncertainty inherent in subsurface modeling. Our trained model runs within a fraction of a second, which is considerably faster than other methods with comparable accuracy and multiple orders of magnitude faster than high-fidelity simulations.</description><subject>Accuracy</subject><subject>Bayesian analysis</subject><subject>Confidence intervals</subject><subject>Fractures</subject><subject>Gas flow</subject><subject>Gas transport</subject><subject>Gaussian process</subject><subject>Machine learning</subject><subject>Modelling</subject><subject>Physics - Data Analysis, Statistics and Probability</subject><subject>Physics - Geophysics</subject><subject>Quantiles</subject><subject>Simulation</subject><subject>Statistical analysis</subject><subject>Three dimensional models</subject><subject>Uncertainty</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><sourceid>GOX</sourceid><recordid>eNotkMtOwzAURC0kJKrSD2CFJdYp9r2x4yyh0FKpgk320a3jgEvrFDsF-vf0wWo2R6M5w9iNFOPcKCXuKf767zGg0GOBudQXbACIMjM5wBUbpbQSQoAuQCkcsPkj7V3yFPjCUQw-vPOu5TNKvIoU0raLPfeBVx_RuezJb1xIvgu05tNItt9Fx19d_9PFz3TNLltaJzf6zyGrps_V5CVbvM3mk4dFRqXSmUFrnVaarF22rZUKBWmLUMjGLYvGtkLmRiPJEqFcltDIRihqpQFsDBSIQ3Z7rj151tvoNxT39dG3PvkeiLszsY3d186lvl51u3jYnGowxytkUQL-AXMHV-o</recordid><startdate>20230606</startdate><enddate>20230606</enddate><creator>Shi, Yingqi</creator><creator>Berry, Donald J</creator><creator>Kath, John</creator><creator>Shams Lodhy</creator><creator>Ly, An</creator><creator>Percus, Allon G</creator><creator>Hyman, Jeffrey D</creator><creator>Moran, Kelly</creator><creator>Strait, Justin</creator><creator>Sweeney, Matthew R</creator><creator>Viswanathan, Hari S</creator><creator>Stauffer, Philip H</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope><scope>GOX</scope></search><sort><creationdate>20230606</creationdate><title>Bayesian Learning of Gas Transport in Three-Dimensional Fracture Networks</title><author>Shi, Yingqi ; Berry, Donald J ; Kath, John ; Shams Lodhy ; Ly, An ; Percus, Allon G ; Hyman, Jeffrey D ; Moran, Kelly ; Strait, Justin ; Sweeney, Matthew R ; Viswanathan, Hari S ; Stauffer, Philip H</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a956-83cce656accbffc1530a6c3271deb7dcf014863a19329b92d1d05af1823d82733</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Accuracy</topic><topic>Bayesian analysis</topic><topic>Confidence intervals</topic><topic>Fractures</topic><topic>Gas flow</topic><topic>Gas transport</topic><topic>Gaussian process</topic><topic>Machine learning</topic><topic>Modelling</topic><topic>Physics - Data Analysis, Statistics and Probability</topic><topic>Physics - Geophysics</topic><topic>Quantiles</topic><topic>Simulation</topic><topic>Statistical analysis</topic><topic>Three dimensional models</topic><topic>Uncertainty</topic><toplevel>online_resources</toplevel><creatorcontrib>Shi, Yingqi</creatorcontrib><creatorcontrib>Berry, Donald J</creatorcontrib><creatorcontrib>Kath, John</creatorcontrib><creatorcontrib>Shams Lodhy</creatorcontrib><creatorcontrib>Ly, An</creatorcontrib><creatorcontrib>Percus, Allon G</creatorcontrib><creatorcontrib>Hyman, Jeffrey D</creatorcontrib><creatorcontrib>Moran, Kelly</creatorcontrib><creatorcontrib>Strait, Justin</creatorcontrib><creatorcontrib>Sweeney, Matthew R</creatorcontrib><creatorcontrib>Viswanathan, Hari S</creatorcontrib><creatorcontrib>Stauffer, Philip H</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Access via ProQuest (Open Access)</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection><collection>arXiv.org</collection><jtitle>arXiv.org</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Shi, Yingqi</au><au>Berry, Donald J</au><au>Kath, John</au><au>Shams Lodhy</au><au>Ly, An</au><au>Percus, Allon G</au><au>Hyman, Jeffrey D</au><au>Moran, Kelly</au><au>Strait, Justin</au><au>Sweeney, Matthew R</au><au>Viswanathan, Hari S</au><au>Stauffer, Philip H</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Bayesian Learning of Gas Transport in Three-Dimensional Fracture Networks</atitle><jtitle>arXiv.org</jtitle><date>2023-06-06</date><risdate>2023</risdate><eissn>2331-8422</eissn><abstract>Modeling gas flow through fractures of subsurface rock is a particularly challenging problem because of the heterogeneous nature of the material. High-fidelity simulations using discrete fracture network (DFN) models are one methodology for predicting gas particle breakthrough times at the surface, but are computationally demanding. We propose a Bayesian machine learning method that serves as an efficient surrogate model, or emulator, for these three-dimensional DFN simulations. Our model trains on a small quantity of simulation data and, using a graph/path-based decomposition of the fracture network, rapidly predicts quantiles of the breakthrough time distribution. The approach, based on Gaussian Process Regression (GPR), outputs predictions that are within 20-30% of high-fidelity DFN simulation results. Unlike previously proposed methods, it also provides uncertainty quantification, outputting confidence intervals that are essential given the uncertainty inherent in subsurface modeling. Our trained model runs within a fraction of a second, which is considerably faster than other methods with comparable accuracy and multiple orders of magnitude faster than high-fidelity simulations.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><doi>10.48550/arxiv.2306.03416</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2023-06 |
issn | 2331-8422 |
language | eng |
recordid | cdi_arxiv_primary_2306_03416 |
source | arXiv.org; Free E- Journals |
subjects | Accuracy Bayesian analysis Confidence intervals Fractures Gas flow Gas transport Gaussian process Machine learning Modelling Physics - Data Analysis, Statistics and Probability Physics - Geophysics Quantiles Simulation Statistical analysis Three dimensional models Uncertainty |
title | Bayesian Learning of Gas Transport in Three-Dimensional Fracture Networks |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T14%3A15%3A33IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_arxiv&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Bayesian%20Learning%20of%20Gas%20Transport%20in%20Three-Dimensional%20Fracture%20Networks&rft.jtitle=arXiv.org&rft.au=Shi,%20Yingqi&rft.date=2023-06-06&rft.eissn=2331-8422&rft_id=info:doi/10.48550/arxiv.2306.03416&rft_dat=%3Cproquest_arxiv%3E2823311792%3C/proquest_arxiv%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2823311792&rft_id=info:pmid/&rfr_iscdi=true |