Closed petebachant closed 9 years ago
Could follow a strategy similar to what is outlined here.
paramsweep.read_force_coeffs
will need to be changed to read the latest file and compute means rather than take the last value. Might be a good idea to log standard deviation of the coeffs as well.
I started doing some work on this on the pimpleFoam
branch. So far, it looks like it may be better to run pimpleFoam
from the start, or maybe run potentialFoam
first rather than simpleFoam
.
I think it may be useful to simply switch to pimpleFoam
altogether. Here's a NACA 0020 at 20 degrees angle of attack solved with pimpleFoam
only:
There are convergence issues at high angle of attack that may require unsteady simulations?