T. Kadeethum, V. L. S. Silva, P. Salinas, C. C. Pain, H. Yoon
{"title":"Boosting Barlow Twins Reduced Order Modeling for Machine Learning-Based Surrogate Models in Multiphase Flow Problems","authors":"T. Kadeethum, V. L. S. Silva, P. Salinas, C. C. Pain, H. Yoon","doi":"10.1029/2023wr035778","DOIUrl":null,"url":null,"abstract":"We present an innovative approach called boosting Barlow Twins reduced order modeling (BBT-ROM) to enhance the reliability of machine learning surrogate models for multiphase flow problems. BBT-ROM builds upon Barlow Twins reduced order modeling that leverages self-supervised learning to effectively handle linear and nonlinear manifolds by constructing well-structured latent spaces of input parameters and output quantities. To address the challenge of high contrast data in multiphase flow problems due to injection wells and faults, we employ a boosting algorithm within BBT-ROM. This algorithm sequentially trains a set of weak models (i.e., inaccurate models), improving prediction accuracy through ensemble learning. To evaluate the performance of BBT-ROM, we conduct three three-dimensional multiphase flow problems, including waterflooding and geologic carbon storage (GCS), with varying numbers of input parameter cases and model domain features. The results demonstrate that BBT-ROM excels at predicting non-wetting phase saturation (e.g., oil or <span data-altimg=\"/cms/asset/5c0cf0e9-c30f-45e8-b121-258a56d3eb6f/wrcr27508-math-0001.png\"></span><mjx-container ctxtmenu_counter=\"391\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27508-math-0001.png\"><mjx-semantics><mjx-mrow data-semantic-annotation=\"clearspeak:unit\" data-semantic-children=\"0,3\" data-semantic-content=\"4\" data-semantic- data-semantic-role=\"implicit\" data-semantic-speech=\"normal upper C normal upper O 2\" data-semantic-type=\"infixop\"><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic- data-semantic-parent=\"5\" data-semantic-role=\"latinletter\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi><mjx-mo data-semantic-added=\"true\" data-semantic- data-semantic-operator=\"infixop,\" data-semantic-parent=\"5\" data-semantic-role=\"multiplication\" data-semantic-type=\"operator\" style=\"margin-left: 0.056em; margin-right: 0.056em;\"><mjx-c></mjx-c></mjx-mo><mjx-msub data-semantic-children=\"1,2\" data-semantic- data-semantic-parent=\"5\" data-semantic-role=\"latinletter\" data-semantic-type=\"subscript\"><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic- data-semantic-parent=\"3\" data-semantic-role=\"latinletter\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi><mjx-script style=\"vertical-align: -0.15em;\"><mjx-mn data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic- data-semantic-parent=\"3\" data-semantic-role=\"integer\" data-semantic-type=\"number\" size=\"s\"><mjx-c></mjx-c></mjx-mn></mjx-script></mjx-msub></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27508:wrcr27508-math-0001\" display=\"inline\" location=\"graphic/wrcr27508-math-0001.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow data-semantic-=\"\" data-semantic-annotation=\"clearspeak:unit\" data-semantic-children=\"0,3\" data-semantic-content=\"4\" data-semantic-role=\"implicit\" data-semantic-speech=\"normal upper C normal upper O 2\" data-semantic-type=\"infixop\"><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic-parent=\"5\" data-semantic-role=\"latinletter\" data-semantic-type=\"identifier\" mathvariant=\"normal\">C</mi><mo data-semantic-=\"\" data-semantic-added=\"true\" data-semantic-operator=\"infixop,\" data-semantic-parent=\"5\" data-semantic-role=\"multiplication\" data-semantic-type=\"operator\"></mo><msub data-semantic-=\"\" data-semantic-children=\"1,2\" data-semantic-parent=\"5\" data-semantic-role=\"latinletter\" data-semantic-type=\"subscript\"><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic-parent=\"3\" data-semantic-role=\"latinletter\" data-semantic-type=\"identifier\" mathvariant=\"normal\">O</mi><mn data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic-parent=\"3\" data-semantic-role=\"integer\" data-semantic-type=\"number\" mathvariant=\"normal\">2</mn></msub></mrow>$\\mathrm{C}{\\mathrm{O}}_{\\mathrm{2}}$</annotation></semantics></math></mjx-assistive-mml></mjx-container> saturation) and fluid pressure, with average relative errors ranging from 0.5% to 3%. Importantly, BBT-ROM showcases robustness when faced with limited input parameter space during GCS testing.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"44 1","pages":""},"PeriodicalIF":4.6000,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Water Resources Research","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1029/2023wr035778","RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
We present an innovative approach called boosting Barlow Twins reduced order modeling (BBT-ROM) to enhance the reliability of machine learning surrogate models for multiphase flow problems. BBT-ROM builds upon Barlow Twins reduced order modeling that leverages self-supervised learning to effectively handle linear and nonlinear manifolds by constructing well-structured latent spaces of input parameters and output quantities. To address the challenge of high contrast data in multiphase flow problems due to injection wells and faults, we employ a boosting algorithm within BBT-ROM. This algorithm sequentially trains a set of weak models (i.e., inaccurate models), improving prediction accuracy through ensemble learning. To evaluate the performance of BBT-ROM, we conduct three three-dimensional multiphase flow problems, including waterflooding and geologic carbon storage (GCS), with varying numbers of input parameter cases and model domain features. The results demonstrate that BBT-ROM excels at predicting non-wetting phase saturation (e.g., oil or saturation) and fluid pressure, with average relative errors ranging from 0.5% to 3%. Importantly, BBT-ROM showcases robustness when faced with limited input parameter space during GCS testing.
期刊介绍:
Water Resources Research (WRR) is an interdisciplinary journal that focuses on hydrology and water resources. It publishes original research in the natural and social sciences of water. It emphasizes the role of water in the Earth system, including physical, chemical, biological, and ecological processes in water resources research and management, including social, policy, and public health implications. It encompasses observational, experimental, theoretical, analytical, numerical, and data-driven approaches that advance the science of water and its management. Submissions are evaluated for their novelty, accuracy, significance, and broader implications of the findings.