A statistical learning strategy for closed-loop control of fluid flows

Guéniat, Florimond and Mathelin, Lionel and Yousuff Hussaini, M (2016) A statistical learning strategy for closed-loop control of fluid flows. Theoretical and Computational Fluid Dynamics, 30 (6). pp. 497-510. ISSN 0935-4964

[img]
Preview
Text (Published Version)
Hash_Control.pdf - Published Version

Download (4MB)

Abstract

This work discusses a closed-loop control strategy for complex systems utilizing scarce and streaming data. A discrete embedding space is first built using hash functions applied to the sensor measurements from which a Markov process model is derived, approximating the complex system’s dynamics. A control strategy is then learned using reinforcement learning once rewards relevant with respect to the control objective are identified. This method is designed for experimental configurations, requiring no computations nor prior knowledge of the system, and enjoys intrinsic robustness. It is illustrated on two systems: the control of the transitions of a Lorenz’63 dynamical system, and the control of the drag of a cylinder flow. The method is shown to perform well.

Item Type: Article
Additional Information: This is a post-peer-review, pre-copyedit version of an article published in Theoretical and Computational Fluid Dynamics. The final authenticated version is available online https://doi.org/10.1007/s00162-016-0392-y
Uncontrolled Keywords: Closed-loop control, Reinforcement learning, Machine learning
Subjects: F300 Physics
H100 General Engineering
Divisions: Faculty of Computing, Engineering and the Built Environment
Faculty of Computing, Engineering and the Built Environment > School of Engineering and the Built Environment
REF UoA Output Collections > REF 2021 UoA12: Engineering
Depositing User: Euan Scott
Date Deposited: 18 Jan 2019 15:57
Last Modified: 25 Jan 2019 12:37
URI: http://www.open-access.bcu.ac.uk/id/eprint/6871

Actions (login required)

View Item View Item

Research

In this section...