Uncovering Nonlinear Flow Physics with Machine Learning Control and Sparse Modeling
Abstract
The overall objective of this work is to use machine learning control (MLC) to explore new flow regimes and behaviors and then use model identification techniques, to identify parsimonious and interpretable models that characterize the underlying flow physics. Machine learning constitutes a growing set of data-driven optimization techniques that are ideal for the modeling and control of high-dimensional, nonlinear, and multi-scale systems, such as are found in fluid dynamics. Further, sparse regression techniques have the potential to identify models that are both physically interpretable and generalize beyond the training data. This work will provide new computational methods to analyze data from fluid simulations and experiments, and will also result in a better understanding of the fundamental structure and interaction physics of unsteady fluid flows. The modeling and control of fluid flows remains a grand challenge problem of the modern era, with potentially transformative scientific, technological, and industrial impact. Indeed, better understanding of complex flow physics may enable drag reduction, lift increase, mixing enhancement, and noise reduction in domains as diverse as transportation, energy, security and medicine. Fluid dynamics is a canonically difficult problem because of strong nonlinearity, high-dimensionality, and multi-scale physics; both modeling and control may be thought of as extremely challenging optimization problems. Recent advances in machine learning and sparse optimization arerevolutionizing how we approach these traditionally intractable problems. We envision that these methods will enable the discovery of novel flow physics as well as practical new control strategies to achieve improved performance in engineering flows. At the end of this work, we will have learned a tremendous deal about important canonical flows.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 30, 2019
- Accession Number
- AD1098059
Entities
People
- Steven Brunton
Organizations
- University of Washington