Bayesian optimization for active flow control
Abstract A key question in flow control is that of the design of optimal controllers when the control space is high-dimensional and the experimental or computational budget is limited. We address this formidable challenge using a particular flavor of machine learning and present the...
| Main Authors: | , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
The Chinese Society of Theoretical and Applied Mechanics; Institute of Mechanics, Chinese Academy of Sciences
2022
|
| Online Access: | https://hdl.handle.net/1721.1/142072 |
| _version_ | 1826209211924611072 |
|---|---|
| author | Blanchard, Antoine B. Cornejo Maceda, Guy Y. Fan, Dewei Li, Yiqing Zhou, Yu Noack, Bernd R. Sapsis, Themistoklis P. |
| author2 | Massachusetts Institute of Technology. Department of Mechanical Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Mechanical Engineering Blanchard, Antoine B. Cornejo Maceda, Guy Y. Fan, Dewei Li, Yiqing Zhou, Yu Noack, Bernd R. Sapsis, Themistoklis P. |
| author_sort | Blanchard, Antoine B. |
| collection | MIT |
| description | Abstract
A key question in flow control is that of the design of optimal controllers when the control space is high-dimensional and the experimental or computational budget is limited. We address this formidable challenge using a particular flavor of machine learning and present the first application of Bayesian optimization to the design of open-loop controllers for fluid flows. We consider a range of acquisition functions, including the recently introduced output-informed criteria of Blanchard and Sapsis (2021), and evaluate performance of the Bayesian algorithm in two iconic configurations for active flow control: computationally, with drag reduction in the fluidic pinball; and experimentally, with mixing enhancement in a turbulent jet. For these flows, we find that Bayesian optimization identifies optimal controllers at a fraction of the cost of other optimization strategies considered in previous studies. Bayesian optimization also provides, as a by-product of the optimization, a surrogate model for the latent cost function, which can be leveraged to paint a complete picture of the control landscape. The proposed methodology can be used to design open-loop controllers for virtually any complex flow and, therefore, has significant implications for active flow control at an industrial scale.
Graphic Abstract |
| first_indexed | 2024-09-23T14:19:16Z |
| format | Article |
| id | mit-1721.1/142072 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T14:19:16Z |
| publishDate | 2022 |
| publisher | The Chinese Society of Theoretical and Applied Mechanics; Institute of Mechanics, Chinese Academy of Sciences |
| record_format | dspace |
| spelling | mit-1721.1/1420722023-01-30T16:02:24Z Bayesian optimization for active flow control Blanchard, Antoine B. Cornejo Maceda, Guy Y. Fan, Dewei Li, Yiqing Zhou, Yu Noack, Bernd R. Sapsis, Themistoklis P. Massachusetts Institute of Technology. Department of Mechanical Engineering Abstract A key question in flow control is that of the design of optimal controllers when the control space is high-dimensional and the experimental or computational budget is limited. We address this formidable challenge using a particular flavor of machine learning and present the first application of Bayesian optimization to the design of open-loop controllers for fluid flows. We consider a range of acquisition functions, including the recently introduced output-informed criteria of Blanchard and Sapsis (2021), and evaluate performance of the Bayesian algorithm in two iconic configurations for active flow control: computationally, with drag reduction in the fluidic pinball; and experimentally, with mixing enhancement in a turbulent jet. For these flows, we find that Bayesian optimization identifies optimal controllers at a fraction of the cost of other optimization strategies considered in previous studies. Bayesian optimization also provides, as a by-product of the optimization, a surrogate model for the latent cost function, which can be leveraged to paint a complete picture of the control landscape. The proposed methodology can be used to design open-loop controllers for virtually any complex flow and, therefore, has significant implications for active flow control at an industrial scale. Graphic Abstract 2022-04-26T11:55:30Z 2022-04-26T11:55:30Z 2022-01-10 2022-04-26T03:55:49Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/142072 Blanchard, Antoine B., Cornejo Maceda, Guy Y., Fan, Dewei, Li, Yiqing, Zhou, Yu et al. 2022. "Bayesian optimization for active flow control." en https://doi.org/10.1007/s10409-021-01149-0 Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ The Chinese Society of Theoretical and Applied Mechanics and Springer-Verlag GmbH Germany, part of Springer Nature application/pdf The Chinese Society of Theoretical and Applied Mechanics; Institute of Mechanics, Chinese Academy of Sciences The Chinese Society of Theoretical and Applied Mechanics; Institute of Mechanics, Chinese Academy of Sciences |
| spellingShingle | Blanchard, Antoine B. Cornejo Maceda, Guy Y. Fan, Dewei Li, Yiqing Zhou, Yu Noack, Bernd R. Sapsis, Themistoklis P. Bayesian optimization for active flow control |
| title | Bayesian optimization for active flow control |
| title_full | Bayesian optimization for active flow control |
| title_fullStr | Bayesian optimization for active flow control |
| title_full_unstemmed | Bayesian optimization for active flow control |
| title_short | Bayesian optimization for active flow control |
| title_sort | bayesian optimization for active flow control |
| url | https://hdl.handle.net/1721.1/142072 |
| work_keys_str_mv | AT blanchardantoineb bayesianoptimizationforactiveflowcontrol AT cornejomacedaguyy bayesianoptimizationforactiveflowcontrol AT fandewei bayesianoptimizationforactiveflowcontrol AT liyiqing bayesianoptimizationforactiveflowcontrol AT zhouyu bayesianoptimizationforactiveflowcontrol AT noackberndr bayesianoptimizationforactiveflowcontrol AT sapsisthemistoklisp bayesianoptimizationforactiveflowcontrol |