2D flow over airfoil with aoa as parameter

Hi, I am using modulus 22.09 to train a model that provide flow field prediction for a 2D airfoil with changing angle of attack from 0deg to 10deg. However, the results is bad compared to the fixed aoa of 0deg and 10deg. The main error happens on the pressure on y diction. I tried to add the batch sizes of airfoil boundaries and interior points. it don’t work.
can someone give some advice to solve it?

Hi @uniquezedpg

I can’t promise that I know what I’m doing, but could you post some more information about the problem? What type of error are you seeing? Did you parameterize the geometry? Or are you using transfer learning and retraining?

Turbulence models each have strengths and weaknesses. When used to simulate flow around an airfoil some will predict separation early, some late, and some never at all. For an airfoil in an otherwise open flow I’d start with k-w:

k-w and k-e examples in Gitlab

Hi @uniquezedpg

Have a look at the recommended practices section of the documentation for some idea on how to improve your convergence.