[green iran]

Hi
I'm trying to create a structure mesh for the cyclon but with no any progress. I divided the model into several subdivisions. This leads to many internal surfaces which then I defined them as "Interior". When I tried to import the mesh file to Fluent it was showed me a lot of errors with something like this:
Error:Cannot change zone-7 to interior because there is only one adjacent cell thread. Error Object:
Actually, this message was seen for half of the internal surfaces. Afterwards, Those surfaces which had the error turned into wall surface.
I'm really confused and I need some help. I have a few questions:
1-How important it is to create a structure mesh for cyclon?
2-In case we have to use a structure mesh, how can it be generated for a cyclon?
3- should I use multiblock? If so, real or virtual?
Any help will be really appreciated

[masoodina]

hi .

once in fluent select the inner faces and change their b.c's to interface .
then define grid interface as coupled . and it will make some shadow walls (in fact no surface) . then chek your grid . it will help .

because of highly swirling flows and les or rsm model u have to have a good structured mesh .

[-mAx-]

Basically you don't need to define internal surfaces as "interior".
If your splits are well done, then volumes are connected, and those surfaces are automatically set as interior (but not visible as bc).
You will have problem, if those surfaces will be automatically set as wall, which means, than volumes aren't connected
You are not enforced to mesh your geometry with hexa. (I can suggest you tetra-hexcore)
Real and/or virtual is only a topology stuff, no influence on mesh
Multiblock is a must for hexa mesh. For tetra, you may do some split for a better mesh control on a mesh, or for example, you want to mesh a volume with hexa, but not the other (typically complexd geometry with inlet-outlet pipe system >> you can isolate both pipes and mesh them with hexa)

[Far]

leads to many internal surfaces which then I defined them as "Interior"

There is no need to define them internal, by default they are internal if you not specify the boundary condition and they have fluid on both sides.


The video tutorial given below presents the meshing of cyclone in ICEM CFD. But still you can get an idea about the blocking that how to split your geometry in gambit to get the similar mesh.

What I can suggest you is to extend the upper geometry topology where you have created very nice splits for the o-block meshing inside the pipe down to the bottom. Also edges outside and inside the pipe should be connected at the same vertices.

For tangential inlet, you can divide into into two volumes. For triangular you can use Tri Primitive http://202.118.250.111:8080/fluent/G...ide/mg0303.htm(equivalent to Quarter O-grid/Yblock in ICEM) and for rectangular part you can simply use mapped hexa.

use these keywords in Google to find the tutorial on cyclone meshing in ICEM CFD

structured mesh generation using ICEM CFD

[green iran]

Hello again guys
Eventually, I solved my problem. The problem was regarding to the way of splitting, Something I never thought of it. Actually, I had forgotten to check connect box option. It was a good experience but took about one week.
Anyway, I really appreciate you for your precious and constructive comments.