Welcome! This is the Data-Informed Computational Fluid Dynamics (DI-CFD) research group led by Dr. Bernat Font (he/his), Assistant Professor at TU Delft. Our main goal is to improve classic CFD methods using data-driven techniques, focusing on computing, modelling, scalability, and flow control.
Some of our current projects involve:
- Scale-resolving simulations (SRS) of flows at high Reynolds numbers
- Acceleration of SRS using numerical methods enhanced with physics-based machine learning
- Data-driven wall modelling of non-equilibrium boundary layers
- Deep reinforcement learning for active flow control
- Discovery of partial differential equations from raw data using ML and sparse regression
- WaterLily.jl, a Julia CFD solver using finite-volume and immersed-boundary methods with backend-agnostic execution (CPU/GPU)
We advocate for open science. Most of the papers have an e-print version you can download for free. The code behind the papers is also open source, and you can find them in GitHub. Please feel free to get in touch for questions, suggestions, or collaborations.
Funding acknowledgements
B34.B-1-270, Mekelweg 2, 2628 CD Delft, Netherlands
Faculty of Mechanical Engineering
Delft University of Technology
Faculty of Mechanical Engineering
Delft University of Technology