[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-290631-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-290631-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","rapid-modelling-of-reactive-transport-in-porous-media-using-machine-learning-limitations-and-solutions","Rapid modelling of reactive transport in porous media using machine learning: limitations and solutions","","Machine learning is used to replace geochemical simulations in reactive transport through porous media, aiming to accelerate computations while preserving predictive accuracy. The study addresses the heavy cost of coupling fluid flow, transport, and geochemical reactions, where geochemical calculations are required for every grid cell and time step. Models perform well for isolated predictions but degrade during rollout across successive time steps, and the work proposes physics-based constraints and tailored dataset generation to stabilize performance on a cation exchange benchmark.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/rapid-modelling-of-reactive-transport-in-porous-media-using-machine-learning-limitations-and-solutions/290631/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/rapid-modelling-of-reactive-transport-in-porous-media-using-machine-learning-limitations-and-solutions/290631.png","ImageObject",300,407,{"name":42,"@type":43},"Levi","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-18","2026-09-17",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",9,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"Why are geochemical simulations computationally intensive in reactive transport?","Question",{"text":63,"@type":64},"Reactive transport requires geochemical reactions to be computed at each grid cell and at every time step, which is costly when simulations contain hundreds to millions of elements.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"What problem occurs when using machine learning surrogates for reactive transport?",{"text":68,"@type":64},"Surrogates can be accurate for isolated predictions, but they fail to maintain accuracy during rollout over successive time steps, leading to accumulating errors.",{"name":70,"@type":61,"acceptedAnswer":71},"How does the chapter improve machine learning surrogate performance for successive time steps?",{"text":72,"@type":64},"It introduces modifications such as physics-based constraints and tailored dataset generation strategies to enhance accuracy during successive time-step predictions.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},290631,1789642234,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,115,119,122,126],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":113,"slug":114},8,30,"research-report",{"id":55,"doc_module":4,"doc_module_name":25,"category_name":116,"show_sort_weight":117,"slug":118},"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":25,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":25,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":97,"slug":129},19,"General","general",{"code":4,"msg":82,"data":131},{"doc_id":79,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":112,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":117,"language":139,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":12,"update_tm":80,"read_time":104,"preview_page_count":4},7971461740909,"https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d","arXiv :2405 . 14548v1 [ cs .CE] 23 May 2024  \nRapid modelling of reactive transport in porous media using machine  \nlearning: limitations and solutions  \nVinicius L S Silva∗,a,b,c Geraldine Regnierb,f Pablo Salinasb,d Claire E Heaneya,e Matthew Jacksonb Christopher C Paina,e  \nMay 24, 2024  \naApplied Modelling & Computation Group, Imperial College London, UK b Novel Reservoir Modelling and Simulation Group, Imperial College London, UK  \nc Petroleo Brasileiro S.A. (Petrobras), Rio de Janeiro, Brazil  \ndOpenGoSim, Leicester, UK  \ne Centre for AI Physics Modelling, Imperial-X, Imperial College London, UK  \nf Suez International, Tashkent, Uzbekistan  \n∗ Corresponding author email: [v.santos-silva19@imperial.ac.uk](v.santos-silva19@imperial.ac.uk)  \nKey points:  \n• Machine learning is explored as a replacement for geochemical simulators in reactive transport for faster and accurate computations.  \n• Machine learning surrogates excel in isolated predictions but require modifications to avoid failure in successive time-step predictions.  \n• Physics-based constraints & tailored datasets are key to enhancing machine learning surrogate accuracy in successive time-step predictions.  \nAbstract  \nReactive transport in porous media plays a pivotal role in subsurface reservoir processes, influencing fluid properties and geochemical characteristics. However, coupling fluid flow and transport with geochemical reactions is computationally intensive, requiring geochemical calculations at each grid cell and each time step within a discretized simulation domain. Although recent advancements have integrated machine learning techniques as surrogates for geochemical simulations, ensuring computational efficiency and accuracy remains a challenge. This chapter investigates machine learning models as replacements for a geochemical module in a reactive transport in porous media simulation. We test this approach on a well-documented cation exchange problem. While the surrogate models excel in isolated predictions, they fall short in rollout predictions over successive time steps. By introducing modifications, including physics-based constraints and tailored dataset generation strategies, we show that machine learning surrogatescan achieve accurate rollout predictions. Our findings emphasize that, when judiciously designed, machine learning surrogates can substantially expedite the cation exchange problem without compromising accuracy, offering significant potential for a range of reactive transport applications.  \n1 Introduction  \nReactive transport within porous media occurs in a range of subsurface processes including geothermal energy extraction, hydrocarbon recovery, CO2 storage, and groundwater flow. These phenomena  \ncan significantly affect the fluid dynamics, as well as the petrophysical and geochemical properties of the reservoirs [Appelo and Postma, 2004 , Fan et al. , 2012 , Kala and Voskov, 2020 , Molins et al. , 2012 , Nogues et al. , 2013 , Oliveira et al. , 2019 , Soulaine et al. , 2021 , Yekta et al. , 2021] . Nonetheless, simulating fluid flow and transport coupled with chemical reactions is highly computationally intensive, due to the necessity of computing chemical reactions within each discrete element of the simulation grid, which can range from hundreds to millions of elements in size.  \nTwo predominant methodologies exist for solving transport and reaction processes together: the sequential method [Oliveira et al. , 2019 , Soulaine et al. , 2021 , Yekta et al. , 2021] and the fully coupled (or simultaneous) method [Fan et al. , 2012 , Kala and Voskov, 2020 , Seigneur et al. , 2023] . In the sequential method, transport and chemical reactions are solved in a stepwise manner. The fully coupled method, on the other hand, simultaneously resolves both transport and reaction equations, iterating until convergence is achieved. Both methodologies necessitate executing geochemical calculations on a per-grid-block basis, which remains the pri","cbCaidvQvVfTh5Yz","https://ap.wps.com/l/cbCaidvQvVfTh5Yz","pdf",9062707,"English","# Key points\n# Abstract\n# 1 Introduction","[{\"question\":\"Why are geochemical simulations computationally intensive in reactive transport?\",\"answer\":\"Reactive transport requires geochemical reactions to be computed at each grid cell and at every time step, which is costly when simulations contain hundreds to millions of elements.\"},{\"question\":\"What problem occurs when using machine learning surrogates for reactive transport?\",\"answer\":\"Surrogates can be accurate for isolated predictions, but they fail to maintain accuracy during rollout over successive time steps, leading to accumulating errors.\"},{\"question\":\"How does the chapter improve machine learning surrogate performance for successive time steps?\",\"answer\":\"It introduces modifications such as physics-based constraints and tailored dataset generation strategies to enhance accuracy during successive time-step predictions.\"}]","Rapid modelling of reactive transport in porous media using machine learning: limitations and solutions | PDF"]