[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-123657-en":3,"doc-seo-123657-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},123657,549758146520,"Patrick","https://ap-avatar.wpscdn.com/avatar/80002397d8c0411e94?_k=1775819394049821470",8,"Research & Report","Applied Machine Learning for Prediction and Control of Fluid Flows - Thesis","Modern aerodynamic platforms such as unmanned aerial systems and horizontal-axis wind turbines face highly stochastic, turbulent gust forces. Traditional modeling and control approaches struggle to mitigate these effects in real time. This thesis applies machine learning experimentally using two complementary tracks: model-free reinforcement learning for gusty flow control on a sensor-equipped testbed, and Fourier neural operators for forecasting turbulent wake dynamics from measured velocity fields, including imperfect data and transfer learning.","Applied machine learning for prediction and control of  \nfluid flows  \nThesis by  \nPeter Ian James Renn  \nIn Partial Fulfillment of the Requirements for the Degree of  \nDoctor of Philosophy  \nCALIFORNIA INSTITUTE OF TECHNOLOGY Pasadena, California  \n2023  \nDefended January 11, 2023  \nii  \n© 2023  \nPeter Ian James Renn ORCID: 0000-0002-5735-3873  \nAll rights reserved  \niii  \nACKNOWLEDGEMENTS  \nFirst, I would like to thank my advisor Dr. Mory Gharib, who has provided me support and guidance throughout my time at Caltech. I deeply appreciate the opportunities which he has given me over the past several years as well as the manythings that he has taught me.  \nI also acknowledge and thank the members of my committee for their insightful counsel: Dr. Anima Anandkumar, Dr. Jane Bae, and Dr. John Dabiri. I have particularly benefited through collaborations with Dr. Anandkumar and her students which have deeply influenced the latter half of the research presented in this thesis.  \nI would like to thank both past and present members of the Gharib group, especially my first mentors in the lab: Chris Dougherty and Marcel Veismann. My early experiences in the lab would not have been nearly as fruitful or fulfilling without their guidance. Special thanks also to Chris Roh and Cong Wang, who have both given me excellent counsel on my research and life pursuits throughout this whole process.  \nI also must thank CAST managers Noel and Reza for their assistance and advice throughout my time here. Additionally, thank you to Martha Salcedo, Jamie Meighen-Sei, and Sarah Pontes for helping solve countless logistical issues.  \nI also would like to thank my parents and siblings. This work would not have been possible without their unwavering support and encouragement. I deeply appreciate everything they have given me.  \nFinally, I would like to thank my partner, Emily. From our first year problem sets to writing this thesis, she has been a great source of joy in my life and I can’t imagine having done any of it without her. Being able to share this time of my life with her, and our two dogs Margot and Charles, has been truly a gift.  \nFinancial support for this work came from the National Science Foundation Graduate Research Fellowship under Grant No. DGE-1745301, as well as the Caltech Center for Autonomous Systems and Technologies.  \niv  \nABSTRACT  \nModern aerodynamic technologies such as unmanned aerial systems and horizontal axis wind turbines must regularly contend with forces from highly stochastic and turbulent atmospheric gusts. Conventional methods for modeling and controlling fluid flows are limited in their ability to mitigate these aerodynamic forces in realtime. By applying modern machine learning techniques in an experimental setting, this thesis demonstrates the utility of machine learning in addressing these important problems. We follow two complementary approaches towards this goal.  \nFirst, we find an end-to-end solution for control in a gusty environment with modelfree reinforcement learning. We deploy state-of-the-art reinforcement learning algorithms on a generalized aerodynamic test-bed consisting of an airfoil with motorized trailing edge flaps. The system features embedded flow sensors, enabling the inclusion of flow measurements in state observations. We place this system ina highly irregular wake behind a bluff-body, dynamically mounted on elastic bands and therefore free to oscillate, and train reinforcement learning agents to minimize the net lifting force on the system by controlling the position of the trailing edge flaps. We find that model-free reinforcement learning agents can outperform basic linear controllers in this gusty, turbulent environment. We also show that augmenting state observations with flow measurements can lead to more consistent learning of the system dynamics.  \nNext, we explore Fourier neural operators (FNOs) as a method for forecasting the time evolution of turbulent fluid flows. FNOs are capable","cbCaiiFAL0EfBt3L","https://ap.wps.com/l/cbCaiiFAL0EfBt3L","pdf",20577292,1,114,"English","en",105,"# Acknowledgements\n# Abstract\n# Published Content and Contributions\n# Table of Contents\n# List of Illustrations\n# List of Tables\n# Chapter I: Introduction\n## Machine learning applications for modern fluid mechanics\n## Summary of work\n# Bibliography","[{\"question\":\"How does the thesis approach real-time control of gusty fluid flows?\",\"answer\":\"It uses model-free reinforcement learning to control motorized trailing-edge flaps on a generalized aerodynamic testbed, with embedded flow sensors optionally included in the observations.\"},{\"question\":\"What forecasting method is used for turbulent fluid flows in this thesis?\",\"answer\":\"The thesis uses Fourier neural operators (FNOs) to learn operator solutions governing families of partial differential equations and to predict time evolution of experimentally measured velocity fields.\"},{\"question\":\"How does the thesis address the challenge of imperfect experimental measurements?\",\"answer\":\"It shows that FNOs can still accurately predict turbulent wake evolution even when trained with imperfect measurements and can adapt to unseen conditions using transfer learning with minimal data.\"}]","Applied Machine Learning for Prediction and Control of Fluid Flows - Thesis | PDF",1785817865,287,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"applied-machine-learning-for-prediction-and-control-of-fluid-flows-thesis","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/applied-machine-learning-for-prediction-and-control-of-fluid-flows-thesis/123657/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"How does the thesis approach real-time control of gusty fluid flows?","Question",{"text":75,"@type":76},"It uses model-free reinforcement learning to control motorized trailing-edge flaps on a generalized aerodynamic testbed, with embedded flow sensors optionally included in the observations.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What forecasting method is used for turbulent fluid flows in this thesis?",{"text":80,"@type":76},"The thesis uses Fourier neural operators (FNOs) to learn operator solutions governing families of partial differential equations and to predict time evolution of experimentally measured velocity fields.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the thesis address the challenge of imperfect experimental measurements?",{"text":84,"@type":76},"It shows that FNOs can still accurately predict turbulent wake evolution even when trained with imperfect measurements and can adapt to unseen conditions using transfer learning with minimal data.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]