[Phill_G]

Hi, I'm trying to simulate a single wind turbine blade rotating in a wind tunnel, I have gotten it to work with using openFoam tools to mesh around an stl. and using a rotating mesh.

The general mesh is a big rectangle for the wind tunnel, a disk stl for the AMI that rotates, a blade stl that is located inside the disk stl for rotating.


Using a refinement level of 7 (standard square is 2m) around the blade, I have gotten results close but not yet acceptable to experimental, so I wanted to add layers to the blade to get into the y+ domain where the results would be closer. When I add the layers, if the refinement level is 8 it covers most of the blade stl, if its 7, it covers around 85% and doesn't form in the attack border of large parts of the blade, so I modified the mesh so the surrounding area of the blade was 8, followed by 7 in a bit larger area then before (using distance refinement)


This is the snappyhexmesh code I am using, snapping has been turned to false to see if it runs without it. currently that is the case, but will update my problem if it does not

snappyHexMeshDict.zip


The layers was added on 91.6% so in parts the mesh looks like this
20220510_125236.jpg


I will also add the schemes that worked on the mesh and the 0 file

files.zip


The last thing I tried was add more orthogonal corrector since the layers did bring the non orthogonality from around 49 to 52 max that resulted in a crash in more iterations
The crash and residuals log in the last attempt was this one

Screenshot from 2022-05-10 15-56-02.jpglog.zip


Any help would be appreciated, I have been struggling with this simulation.



I'm using pimple foam, kEpsilon and wall functions

[Fouch]

Hi,

I had this problem with kEpsilon. May you can try with kOmega ?

[Phill_G]

Originally Posted by Fouch

Hi,

I had this problem with kEpsilon. May you can try with kOmega ?

Currently trying it out and will post the results. It has not exploded yet, been running for a week