[pankos]

I use Gambit software.
If i want to have more dense mesh (in pave method at 2d-face) at a region in one domain, i must make "hard edges" at this region. Right? After Can i delete all hard edges and to remain the face and its mesh?

With word "hard edges", we mean the split of face by an edge? This is Right?

[-mAx-]

an hard edge is an edge which splits partially a surface.
That means it doesn't divide you surface into 2 separated surfaces and it can give you meshing trouble.
Sans titre.png
If you want to refine your surface (homogeneous), you can decrease the value of element size in the meshing panel.
Or you can use split and size fonctions combination.

[gascortado]

Regarding hard edges. I'm trying to model a a plate inside a turbulent BL using LES and Gambit. Take a look at the attahced mesh. The simulation gets screwed up at the intersection of the structured and unstructered mesh. There are artificial eddies formed and the flow gets disturbed. Is there any way to solve this type of problem? I tried using Hex elements only, but my plate in a void inside the middle volume and I'm forced to use tet elements. Any sugestions?

[-mAx-]

I don't understand why you can't use hexa instead of tetra.
Can you post a picture of your result at the intersection between structured and unstructured mesh.

[gascortado]

Thanks for your reply Max and sorry for my late answer. take a look at the picture. I'm basically exporting my mesh to OpenFoam. LEt em know what you think. Thanks

[-mAx-]

Regarding your picture your are working with unsteady solver, right?
Did you try to run your model with steady solver?

[gascortado]

yes im using an unsteady solver because this is an LES simulation. Would a steady solver work in conbination with LES?

[-mAx-]

Indepedantly from your physics, you may switch to steady solver and standard turbulent model, just for be sure that your problem comes from your mesh (or not).