Master Thesis on Deep Learning for Fluid Mechanics
Starting date: January 2018
The project is aimed at using machine learning techniques, in particular deep learning, to tackle several problems of great relevance in the analysis of wall-bounded flows. A comprehensive database of high-fidelity data, including instantaneous fields of turbulent channels and boundary layers, will be used to assess the performance of various deep learning algorithms. The goal is to develop robust near-wall models and predictions techniques, and to obtain optimum initial conditions, in order to improve the efficiency of currently available techniques of turbulent flow simulation.
The details of the project are to be decided in collaboration between the student and supervisors.
- Supervisors at department of Mechanics: Philipp Schlatter (email@example.com) and Ricardo Vinuesa (firstname.lastname@example.org)
- Supervisor at RPL department: Hossein Azizpour (email@example.com)
The main supervisor will be decided depending on the student's school.
Requirements: Prior experience with deep learning models and a framework (e.g. tensorflow, torch, etc.), basic understanding of fluid mechanics will be a plus.
Contact: Show your interest in this project by sending an email to the supervisors above, it may include a brief proposal, ideas or thoughts
Show earlier events (1)