CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Meshing & Mesh Conversion (https://www.cfd-online.com/Forums/openfoam-meshing/)
-   -   [snappyHexMesh] snappyHexMesh cell alignment (https://www.cfd-online.com/Forums/openfoam-meshing/241526-snappyhexmesh-cell-alignment.html)

Miguel Hernandez March 3, 2022 18:12

snappyHexMesh cell alignment
 
2 Attachment(s)
Hello everyone,
with the goal of simulating a free-flowing channel (using interFoam as solver ) I am trying to set up the mesh, but the results are not satisfactory.

The channel has a maximum slope around 10% and is not straight.
For the mesh setup I followed the following steps:

Created the background mesh (blockMesh), without the boundary definitions, large enough to include all the stl geometries.

Created the stl for the definition of the channel (gray stl), inlet, outlet (pink stl) atmosphere (cyan stl), right and green boundary (boundary for air outlet - green and red stl).

The stl files were also used for the definition of the boundary conditions.

Using snappyHexMesh I created the mesh, defining the mesh itself as the internal volume of stl files.


The mesh created allows the simulation for 20 seconds of the 40 sec I've estimated, then it becomes unstable.
I noticed that the cells, due to the curve, are not aligned with the direction of the flow - as seen in the attached image which correspond in the castellatedMesh phase only (is not the "final" mesh I used in my tests simulations) but I think it makes the point.

Is it possible that the instability is due to this non-alignment?
Is possible to align the cell with the flow? If yes, how?

Thanks in advance...

Attachment 88643

Attachment 88644

AtoHM March 4, 2022 05:50

Have you looked at important quality metrics for the mesh? What does the skewness, non-orthogonality look like? Also cellVolumeRatio might be important for mulitphase flows.


You can create a blockMesh that follows roughly the curvature of the channel, this will align the cells and simultaneously improve performance of snappy when snapping the mesh. I would definitely recommend that here.


All times are GMT -4. The time now is 02:27.