Data-Driven Fluid Mechanics: Combining First Principles and Machine Learning
Machine Learning for Fluid Mechanics: Analysis, Modeling, Control and Closures

About the lecture Series
This is the first lecture series from the von Karman Institute for Fluid Dynamics dedicated to machine learning for fluid mechanics. Machine learning is driving comprehensive economic and social transformations and is rapidly shaping fluid mechanics’ research. This revolution is driven by the unprecedented amount of high-quality data provided by our ever-increasing experimental and numerical capabilities. Machine learning extracts knowledge from data without the need of first principles and introduces a new paradigm to applied scientists: use data to discover, rather than validate, new hypotheses and models. This new paradigm enlarges the available methodological portfolio of fluid dynamicists but challenges undergraduate students and practitioners alike in gathering the required know-how from all the disciplines that are intersected by its methods. These include computer science, statistics, optimization, signal processing, pattern recognition, nonlinear dynamics, and control.
The Lecture Series took place at the Université libre de Bruxelles from 24 to 28 february 2020.
Content of the lecture series
Part I: Coherent Structures
Lecture 1 - Prof. B.R. Noack Analysis, Modeling and Control of the Cylinder Wake
Lecture 2 - Prof. J. Jiménez Coherent Structures in Turbulent Flows
Lecture 3 - Prof. S.T.M. Dawson The Proper Orthogonal Decomposition
Lecture 4 - Prof. P.J. Schmid The Dynamic Mode Decomposition: From Koopman Theory to Applications
Part II: Mathematical Analysis
Lecture 5 - Prof. Mendez Mathematical Tools, Part I: Continuous and Discrete LTI Systems
Lecture 6 - Prof. Discetti Mathematical Tools, Part II: Time-Frequency Analysis
Lecture 7 - Prof. Mendez Generalized Modal Analysis and Multiscale POD
Lecture 8 - Prof. Ianiro Applications and Good Practice
Part III: Dynamical Systems
Lecture 9 - Prof. P.J. Schmid Modern Tools for the Stability Analysis of Fluid Flows
Lecture 10 - Prof. S.T.M. Dawson Linear Dynamical Systems and Control
Lecture 11 - Prof. S.L. Brunton Nonlinear Dynamical Systems
Lecture 12 - Prof. S.L. Brunton Methods for System Identification
Part IV: Reduced-Order Modeling
Lecture 13 - Prof. S.L. Brunton Introduction to Machine Learning Methods
Lecture 14 - Prof. S.L. Brunton Machine Learning in Fluids: Pairing Methods with Problems
Lecture 15 - Prof. B.R. Noack Machine Learning for Reduced-Order Modeling
Lecture 16 - Prof. A. Parente Advancing Reacting Flow Simulations with Data-Driven Models: Chemistry Accelerations and Reduced-Order Modeling
Part V: Control, Closure and Perspectives
Lecture 17 - Prof. S. Görtz Reduced-Order Modeling for Aerodynamic Applications and MDO
Lecture 18 - Prof. B.R. Noack Machine Learning for Turbulence Control
Lecture 19 - Prof. J. Jiménez The Computer as Turbulence Researcher