von Karman Institute Lecture Series and Events
Machine Learning for Fluid Mechanics: Analysis, Modeling, Control and Closures
Monday 24 February 2020 - Friday 28 February 2020VKI secretariat, This email address is being protected from spambots. You need JavaScript enabled to view it.; Phone: +32 2 359 96 04
Introduction
NEWS:
1) REGISTRATION CLOSED!
2) This lecture will be hosted by the Free University of Brussels (ULB), located on the Erasme Campus, Madeleine De Genst , building W - W.1.302, rue Meylemeersch 20 (entrance 6), 1070 Anderlecht.
Big data and machine learning are driving extensive economic and social changes and are permeating every area of applied science. Face or voice recognition, real-time language translators, self-driving cars, advanced social media, and customer analytics are just some of the products that the data revolution has provided in the last decade. The success of these products relies on the ability of the underlying algorithms to recognize (and classify) the relevant information out of an unwieldy large amount of data and learn from it, by building simple models that enable fast and accurate predictions. Every area of applied science is increasingly benefitting from such powerful tools.
Fluid mechanics is historically a field of big data as experimental and numerical methods provide datasets of ever-growing size and resolution. The ongoing big data revolution, which has its roots in computer science, statistics, pattern recognition, and artificial intelligence fields, is now entering the fluid mechanic community and is extensively improving the way we analyze data and extract knowledge from it.
This new course aims at providing a unified treatment of the machine learning tools that are now paving the way towards advanced methods for model order reduction, system identification, and flow control. The course will gather ideas and notions from various fields, starting from the data decompositions that were pioneered in fluid mechanics and moving towards machine learning methods that were initially developed in machine vision, pattern recognition, and artificial intelligence. This material will be supported with a comprehensive review of the mathematical background and the theory of dynamical systems, including a review on stability analysis for fluid flows and system identification. Furthermore, the lectures will be complemented with a practical exercise and coding sessions that will provide hands-on experience and a reference/starting point to develop a computational proficiency on the subject.
The covered spectra of topics will range from introductory to state of the art research methods, to make the participants capable of exploiting the enormous opportunity offered by the current big data revolution, and able to keep track of the rapid evolution of the field. At the end of the course, the attendees will be capable of designing advanced tools to analyze numerical and experimental data, perform model order reduction, data-driven system identification, and flow control. Whilst the course is intended primarily for the use by fluid dynamics practitioners, it is believed that most of its content will flow through the technological pipeline into a broad spectrum of applications that could include automotive, aeronautical, wind energy, ship designers, and process engineers.
The lecture series will host a poster session, which will allow the participants to further exchange and interact with the lecturers. All the participants are encouraged to submit a 1-page abstract before December 1st, 2019. Please note that the number of participants is limited and admission will be granted on a first come, first served basis.
The Lecture Series codirectors are Miguel A. Mendez from the von Karman Institute (Belgium), Alessandro Parente from the Université libre de Bruxelles (Belgium), Andrea Ianiro from Universidad Carlos III de Madrid (Spain), Bernd R. Noack from Harbin Institute of Technology, Shenzhen and TU Berlin (GERMANY) and Steven L. Brunton from University of Washington (USA).
Preliminary programme
Monday 24 February 2020: Coherent Structures
09:00 Registration
09:30 Welcome address
10:00 Analysis, Modeling and Control of the Cylinder Wake
Prof. B.R. Noack, Harbin Institute of Technology, Shenzhen, China and TU Berlin, Germany
11:15 Coffee Break
11:45 Coherent Structures in Turbulent Flows
Prof. J. Jiménez, Universidad Politecnica de Madrid, Spain
13:00 Lunch
14:30 The Proper Orthogonal Decomposition
Prof. S.T.M. Dawson, Illinois Institute of Technology, USA
15:45 Coffee Break
16:15 The Dynamic Mode Decomposition: From Koopman Theory to Applications
Prof. P.J. Schmid, Imperial College London, UK
17:30 Reception
Tuesday 25 February 2020: mathematical Analysis
09:00 Mathematical Tools, Part I: Continuous and Discrete LTI Systems
Prof. M.A. Mendez, von Karman Institute for Fluid Dynamics, Belgium
10:15 Coffee Break
10:45 Mathematical Tools, Part II: Time-Frequency Analysis
Prof. S. Discetti, Universidad Carlos III de Madrid, Spain
12:00 Poster Session – posters will be on display during the entire Lecture Series
12:45 Lunch
14:00 Generalized and Multiscale Data-Driven Modal Analysis
Prof. M.A. Mendez, von Karman Institute for Fluid Dynamics, Belgium
15:15 Coffee Break
15:45 Applications and Good Practice
Prof. A. Ianiro, Universidad Carlos III de Madrid, Spain
17:00 End of day
Wednesday 26 February 2020: Dynamical Systems
09:00 Modern Tools for the Stability Analysis of Fluid Flows
Prof. P.J. Schmid, Imperial College London, UK
10:30 Coffee Break
11:00 Linear Dynamical Systems and Control
Prof. S.T.M. Dawson, Illinois Institute of Technology, USA
12:15 Lunch
14:00 Nonlinear Dynamical Systems
Prof. S.L. Brunton, University of Washington, USA
15:15 Coffee Break
15:45 Methods for System Identification
Prof. S.L. Brunton, University of Washington, USA
17:00 End of day
Thursday 27 February 2020: Reduced Order Modeling
09:00 Introduction to Machine Learning Methods
Prof. S.L. Brunton, University of Washington, USA
10:30 Coffee Break
11:00 Machine Learning in Fluids: Pairing Methods with Problems
Prof. S.L. Brunton, University of Washington, USA
12:30 Lunch
14:00 Machine Learning for Reduced-Order Modeling
Prof. B.R. Noack, Harbin Institute of Technology, Shenzhen, China and TU Berlin, Germany
15:15 Coffee Break
15:45 Advancing Reacting Flow Simulations with Data-Driven Models: Chemistry Accelerations and Reduced-Order Modelling
Prof. A. Parente, ULB, Belgium
17:00 Posters and General Discussion
19:00 End of day
Friday 28 February 2020: Control, Closures and Perspectives
09:00 Reduced-Order Modeling for Aerodynamic Applications and MDO
Dr. S. Görtz, German Aerospace Center (DLR), Germany
10:15 Coffee Break
10:45 Machine Learning for Turbulence Control
Prof. B.R. Noack, Harbin Institute of Technology, Shenzhen, China and TU Berlin, Germany
12:00 Lunch
13:45 The Computer as Turbulence Researcher
Prof. J. Jiménez, Universidad Politecnica de Madrid, Spain
15:00 Coffee Break
15:30 Round Table
17:00 End of day
Abstract Submission
All the participants to the Lecture Series "Machine Learning for Fluid Mechanics" are invited to present their work at the poster session, which will represent the central forum of discussion between the participants and lecturers.
To present a poster, the participants should submit a one-page abstract electronically by email to This email address is being protected from spambots. You need JavaScript enabled to view it. before December 1st, 2019.
Acceptance of the poster will be announced to the corresponding author by December 20nd, 2019.
Please note that the number of participants is limited and admission will be granted on a first come, first served basis.
Abstracts should summarize the main concepts of the work, key results, and future perspectives. The presented work might not necessarily include the latest innovative developments and contributions: we also encourage submissions by early career researchers.
The abstract must be written in English following the template provided. Times Roman font 12pt must be used, with the exceptions of the title (14pt), affiliations, and references (11pt). The text width is 160mm. The file must be converted into PDF format; other formats are not accepted.
The abstract must contain the full name(s) and affiliation(s) of the author(s). In the case of joint authorship, the name of the author who will present the poster should be indicated with an asterisk.
The final acceptance of the poster publication and presentation requires the reception of the payment of the presenting author's registration fee until 1 December, 2019. Poster boards (size A0, 119 cm high by 84 cm wide) will be available, including mounting material. All the posters will be on display for four days starting on Tuesday 25 February 2020.
In case of doubts, please contact the organizers at This email address is being protected from spambots. You need JavaScript enabled to view it. or This email address is being protected from spambots. You need JavaScript enabled to view it.
Location and Map
Campus Erasme
Auditorium Madeleine De Genst (Building W - W.1.302)
rue Meylemeersch 20(entrance 6), 1070 Anderlecht
The subway stops at the entrance 2, Route de Lennik 808, 1070 Anderlecht
See the local map and how to reach the campus here: https://www.ulb.be/en/maps-directions/erasme
Location : Campus Erasme, ULB, Auditorium Madeleine De Genst, , building W - W.1.302