BlockMeshDict geometry resolvement Problem
 

Hi all,
i am trying to mesh a simple domain. I tried different mesh sizes in my Blockmeshdict. But strangely with some parameters althoung the total number of meshes are much more higher then the others geometry is not resolved at all... its keeping the wrong cells... (Note: i am always giving the same location(lover left corner) in "LocationInMesh")

Does anyone have any idea?

blockmesh (5 5 5)
Attachment 8998
blockmesh (8 8 8)
Attachment 8999
blockmesh (20 20 20)
Attachment 9000

Hi

Seems like your geometry is allowing cell leak. Is it water tight?

Sometimes, better refinement is a good way to find leaks in geometry.

Goncalo

Hi,
thanks for reply.

I am pretty sure that its water tight but ofcourse worth to check. But if its not water tight doesnt openfoam supposed to give the same result for each blockmesh options?

Thanks
Levent

Hi Levent

snappyHexMesh will subdivide the object's surface by whatever factor you prescribe. If you start off with a finer blockmesh, your surface mesh will end up being smaller.

Is this a free-floating bluff body, or a wall mounted structure?

Goncalo

to be honest i have no idea what are free-floating bluff body, or a wall mounted structure.

But if my mesh is finer in (20 20 20 ) than (8 8 8). finer mesh supposed to give better results because surface mesh is finer, no?

Hi Levent

I think your problem is not so much the blockmesh as the snappyHexMesh step.

Want kind of geometry are you trying to mesh?

Of the 3 images you showed, it seems to me that the second one is closest to what you would like to see.

Goncalo

thats right.. second one is what i want to see.. but i want to understand the fact there... why i can get the geometry with (8 8 8) mesh but not with a finer one like (20 20 20).

Hi Levent

snappy basically first refines the cells at surfaces and then removes internal cells. If it detects a way into a solid body, it won't remove that internal mesh.

Is the structure attached to a boundary created by blockMesh?

Hi Goncalo,
actually they intersect... to see the mesh clearly i defined the domain in z direction narrower then the geometry itself... there is still no leak on the geometry surface but this intersection may cause the problem... i will check and inform u guys about the result.

Hi again,
i tried to enlarge the domain so that nothing would intersect with the geometry. It works... thanks a lot!

i am mistaken..... it does not work... still works for (8 8 8) but not (10 10 10)