Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Problem with Boundary Layer Meshes: Plane with rectangular gap as flow inlet boundary

Please login with a confirmed email address before reporting spam

Hello Comsol Community

I have some major problems with my model, especially treating boundary layer meshes.
As you can see in the appended example-model, I've created 2 studies. one with BM meshes, one without.
Both cases are calculated with a extremely coase grid, I've also calculated both cases with finer grids. the difference in the results was just too big.

Example: Look at the 3D Cut Line, the pressure over the inlet boundary has to be the same. Thats the quantity I want to observe.

I think the problem is the geomtry of my model.
As you can see my flow inlet is the boundary in the middle of the geo, which is a box with a gap in the middle, where the fluid is coming out. since I wasn't able to create a rectangle in order to set the inlet, I just added the box in the middle, selecting only the lower boundary as inlet. But if I do so, I have big problem creating the BM meshes.

1. Somehow I have to select the inlet boundary also in the boundary mesh selection, otherwise I dont get the mesh lines. But BM meshes should also be inserted at the wall, so I am having a big problem.
The boundary 21 cant be deselected...

somehow if I have a plane with a gap in between which is the flow inlet, you cant use BM meshes!

What I am thinking of:
Increase the size of the rectangular box in the middle so that it goes a little bit into the wall (steel part).
then I can select the lower boundary as inlet. But then I would not get the same results as I have a pressure drop across the ''small inlet region.''

I would really appreciate your help.


4 Replies Last Post 25 juin 2012, 03:39 UTC−4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 juin 2012, 10:27 UTC−4
Hi

first of all you have, for me, a far too coarse mesh for a turbulent analysis, so the test of the boundary layer is only valid once the main mesh is about right.
Then I would not use a boundary layer over the inlet region (but this would perhaps not change anything) normally we use them on no-slip walls

By the way you are sure the inlet is 21 and not 22 ?

This said, I would start at a lower speed ensuring a closer to laminar flow (and analysis) first and use that as a starting point to ramp up the inlet flow in a turbulent model

--
Good luck
Ivar
Hi first of all you have, for me, a far too coarse mesh for a turbulent analysis, so the test of the boundary layer is only valid once the main mesh is about right. Then I would not use a boundary layer over the inlet region (but this would perhaps not change anything) normally we use them on no-slip walls By the way you are sure the inlet is 21 and not 22 ? This said, I would start at a lower speed ensuring a closer to laminar flow (and analysis) first and use that as a starting point to ramp up the inlet flow in a turbulent model -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 24 juin 2012, 11:59 UTC−4
Dear Ivar,

thanks for the help.
As I've metioned, I've already used finer mesh. I've just uploaded the corse mesh because of the file size.

I know that the boundary layer mesh is only interesting for the wall boundaries, but I somehow can't deselct the inlet () in my model. otherwise COMSOL gives me an error, saying EDGE...
If I also insert boundary layer meshes on the inlet, I can create the mesh.

And the inlet is 21, not 22. The Fluid is coming out of the wall.

Your hint means:
I use a smaller velocity for the inlet region, after calculation I use a higher inlet velocity + the existing solver solution in order to turn the laminar flow results into my turbulent flow results? am I right?
Dear Ivar, thanks for the help. As I've metioned, I've already used finer mesh. I've just uploaded the corse mesh because of the file size. I know that the boundary layer mesh is only interesting for the wall boundaries, but I somehow can't deselct the inlet () in my model. otherwise COMSOL gives me an error, saying EDGE... If I also insert boundary layer meshes on the inlet, I can create the mesh. And the inlet is 21, not 22. The Fluid is coming out of the wall. Your hint means: I use a smaller velocity for the inlet region, after calculation I use a higher inlet velocity + the existing solver solution in order to turn the laminar flow results into my turbulent flow results? am I right?

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 24 juin 2012, 16:21 UTC−4
Hi

its often easier to Edit - Clear the solutions, and the Mesh, File - Reset the history then save into a new name and uppload the file, at least we get a represetative model we can rerun ;)

Indeed, I always suggest to start laminar (with low inlet speed) to get the model running, then increase the speed (and often reduce the mesh density), and then switch to turbulent flow once the model seems to behave.

But when you are 100% sure about what you are doing I assume the CFD specialist out here jump straight into the stream and go Turbulent immediatly, I'm just an amateur for CFD so I suggest to go the slow way ;)

--
Good luck
Ivar
Hi its often easier to Edit - Clear the solutions, and the Mesh, File - Reset the history then save into a new name and uppload the file, at least we get a represetative model we can rerun ;) Indeed, I always suggest to start laminar (with low inlet speed) to get the model running, then increase the speed (and often reduce the mesh density), and then switch to turbulent flow once the model seems to behave. But when you are 100% sure about what you are doing I assume the CFD specialist out here jump straight into the stream and go Turbulent immediatly, I'm just an amateur for CFD so I suggest to go the slow way ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2012, 03:39 UTC−4
Hi,

Your geometry is a bit strange: you have several overlapping blocks, in particular lusft duese oben and luft duese seite. It may be better to use some boolean operations to avoid any overlapping.

Cheers
Hi, Your geometry is a bit strange: you have several overlapping blocks, in particular lusft duese oben and luft duese seite. It may be better to use some boolean operations to avoid any overlapping. Cheers

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.