Motivation

Big data and machine learning are driving comprehensive economic and social transformations and are rapidly re-shaping the toolbox and the methodologies of applied scientists. Machine learning tools are designed to learn functions from data with little to no need for prior knowledge. As continuous developments in experimental and numerical methods improve our ability to collect high-quality data, machine learning tools become increasingly viable and promising in disciplines rooted in physical principles. Fluid dynamics is one of them.

Objectives

This course gives an overview and practical hands-on experience in integrating machine learning in fluid dynamics. The course originated as a compressed version of the course Machine Learning for Fluid Dynamics, given at the Research Master program at the von Karman Institute. After a brief review of the machine learning landscape, we show how to frame problems in fluid mechanics as machine learning problems, and we explore challenges and opportunities. Attendees will be guided through a series of tutorial sessions in Python and will tackle several relevant applications: aeroacoustics' noise prediction, turbulence modelling, reduced-order modelling and forecasting, meshless integration of (partial) differential equations, super-resolution and flow control.

Approach

All lectures consist of a short theoretical session and a set of practical exercises using Python. Moreover, several group exercises will be provided. These are working sessions in which the role of the instructor is marginal, and participants should be able to complete their assignment independently. These will provide hands on experience and consolidate the understanding of the provided tools.

Pre-requisites

The course pre-requisite is a basic understanding of Python programming and basic knowledge of Calculus, Linear Algebra and Fluid dynamics. The course is pitched for undergraduate and graduate students alike, as well as practitioners in the fields.

Certificates

1. Certificates of Participation. This will certify that the participant attended all lectures. No grading is involved.

2. Graded Course Certificate. This will be a graded certificate, which participants can use to obtain ECTS in their home Universities. The grading is based on the results of an online exam, which will be organized within one month after the course. The online exam will include both theoretical questions and exercises. The evaluation criteria for the activities in Python are based on the clear and bug-free implementation of the various algorithms developed during the course. The final grade will be the average between the quiz results and the exercises. The estimated workload of the course consists of 22.5 h of lectures and 22.5 h of self-study (of which 7.5 during the course).

A discount of 50% is applied for an online participation

Registration deadline for on site participation: 13/05/2024

Registration deadline for online participation: 20/05/2024