Data science in CFD
 

Hello,

recently read some papers about data science in CFD.
Mainly extracting flow features of complex simulations like DNS and LES in a reduced order model (ROM) using deep learning. But it seems to be only scratching on the surface with simple test cases.

Do you think this will be a thing in the future, or just some playing around?
Anyone already applying data science methods for CFD?
I would be interested in what's your take on the subject.

Dear Mango,
here is a pretty heated discussion going on about machine learning methods in CFD: https://www.cfd-online.com/Forums/ma...gence-cfd.html


Maybe this is of interest to you!
Best
Vesparian

Hi Philipp Wiedemer,

I know the thread is 2 years old. I am starting to work on analysis of flow features from large LES data. If possible, could you share the literature on extracting/analysis flow features from DNS/LES data?
Thanks,
Karthick