cancel
Showing results for 
Search instead for 
Did you mean: 

Boundary layer and 3D swept mesh for NX Flow?

Experimenter
Experimenter

Hello all,

 

Is there any way to make a combination of 3D swept mesh with boundary layer for fluid meshing, in NX11?

In the doc (https://docs.plm.automation.siemens.com/tdoc/nx/11/nx_help/#uid:xid950327) they mention only 3D tet to mesh the rest of the solid outside of the boundary layer.

 

Ideally, I would like to have swept mesh so I have exact same mesh on the upper and lower part of my solid and I can impose periodic boundary layer. My goal is to make a pure 2D simulation.

 

See picture here-below to understand the issue.

 

Thanks for the help,

 

Dimitri

image.png

 

13 REPLIES

Re: Boundary layer and 3D swept mesh for NX Flow?

Legend
Legend

Can't you accomplish that using split solids & local element size:

 

mesh.jpg

 

Re: Boundary layer and 3D swept mesh for NX Flow?

Experimenter
Experimenter

Hello,

 

Thanks for the reply.

What you suggest is a nice work around but still I cannot use the true Boundary Layer option that will create a gradient in the elements thickness.

 

I splitted the body using an "offset curve" of 15mm around my foil and then I created a "mesh control\boundary layer" on the inner surface of 15mm but still to actually generate the mesh I have no other option than using "3D tethahedral mesh" or "3D hybrid" mesh.. and the result is a nice boundary layer as expected except the very last portion which includes thetahedral elements. Even if I select a boundary layer thickness lartger than my inside split body, like 20mm for instance. See the result here-under..

boundary_layer.png

Any idea how to overcome this? I find this is a quite weird limitation of NX..

 

Regards,

 

Dimitri

Re: Boundary layer and 3D swept mesh for NX Flow?

Legend
Legend

Why are you callling these "not true boundary layer"?  The boundary layer icon is simply a packaged version of what you call a "work around", with less options and a fancy name...  You are aware that "boundary layer" here is not implying anything wrt CFD properties, right?  It's just the name of a meshing function.  One can use the boudary layer tool to create a brick inflation, regardless of the type of solution you want to work with.  It's just a meshing tool.

 

The second problem you have is resolved the same way I solved the first one, it takes less than a minute to get what you want :-)

Re: Boundary layer and 3D swept mesh for NX Flow?

Experimenter
Experimenter

What I call a true boundary layer meshing is the capability of having very thin elements at the wall and then growing thickness as it gets farther. This allows to capture adequately the strong velocity gradient at the wall while not using very thin elements everywhere.

bl.png

 

Maybe I didn't understand your idea, but I don't see how your work around could help to get the same result as shown here-below.

boundary_layer2.png

 

If I want a boundary layer meshing like in the following picture, it would mean that I have to make a 3 or 4 split bodies, and it will not be very clean solution to haver so many fluid interfaces with no matching nodes.

 

I don't see any way to create the brick inflation without using the fancy boundary layer option, which forces me to include some ugly tetrathedral elements at the interface with the rest of the fluid domain.

 

Thanks for your help. Regards,

 

Dimitri

 

 

Re: Boundary layer and 3D swept mesh for NX Flow?

Legend
Legend

You can accomplish what you call a true boundary layer with the classic tools, without using the boundary layer tool.  I don't know the dimensions or characteristics of your geometry, seems to me that even the boundary layer tool should be able to do this...  The mesh you're showing there is attrocious, I can't imagine that would be the only answer from the boundary layer tool in NX...  There are plenty of defaults to tweak to make things happen, you should experiment a little.

 

That being said, 3/4 slices in the geometry isn't the end of the world.  To create the brick inflation, you need to use bias on edge...  If you know what you're doing with NX, it will be very clean and you can certainly have the fluid lining up (although you don't need to).  We make models all day like that...

 

Good luck :-)

Re: Boundary layer and 3D swept mesh for NX Flow?

Creator
Creator

Correct me if I got it wrong.

 

As per your recommendation:

1] split the solid body into multiple solid parts

2] use a denser grid for required parts 

3] connect disjoint meshes

4] run (may be, only using serial solver)

 

If I need to use "Parallel Processing", the "Connect Disjoint Fluid Meshes" tool cannot be used, right? Would there be any alternate approach to this?

Re: Boundary layer and 3D swept mesh for NX Flow?

Siemens Legend Siemens Legend
Siemens Legend

If you use mating conditions with type "glue coincident" you will not have the need of connecting disjoint meshes as you have a consistent mesh.

Using the biasing function on the Boundary solids will give you the possibility of generating a boundary layer element mesh.

 

 

Re: Boundary layer and 3D swept mesh for NX Flow?

Legend
Legend

Exactly my point!

Re: Boundary layer and 3D swept mesh for NX Flow?

Creator
Creator

Hi,

 

I used the mesh mating (glue coincident) condition to connect the fluid mesh (tried automatic as well as manual pairing). But the solver (parallel solver) treats the fluid bodies as separate flow enclosures. This is not desired, I want to connect the fluid bodies and simulate them as one entire fluid domain.

 

Does it have anything to do with the way we define mesh mating i.e. the source and target bodies? Please let me know any suggestions.

 

Thanks in advance.