|
[Sponsors] |
O-grid for cylinder with varying diameter and several blocks |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
June 16, 2009, 07:25 |
O-grid for cylinder with varying diameter and several blocks
|
#1 |
New Member
Patrick Dems
Join Date: May 2009
Posts: 4
Rep Power: 17 |
Hi all,
I have the following problem with O-grid generation in icem: My geometry is a cylinder with variable diameter (consisting of several frustums). The block structure contains several blocks, one for each frustum. For generating an o-grid I select all the blocks and the two faces at the ends of the cylinder, as usual. Icem divides all blocks correctly each into 5 blocks.The o-grid at the faces at the two ends of the cylinder show a nice o-grid, but in between the grid at the central face of the "O" is like the pattern as I would mesh a circle without o-grid (highly skewed cells at the edges). The pattern of the o-grid changes continuously along the height of the cylinder from the nice o-grid at the ends to the bad grid in the middle. This case occurs for at least 3 frustums and changing diameters. Does anyone know this problem and found an solution? Thanks for answers! |
|
June 17, 2009, 16:13 |
Link edge or face shape...
|
#2 |
Senior Member
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47 |
It is hard to understand without an image, but if you want to force a certain shape, you could try link edge shape, or if you have 12.0, like face shape.
Post an image and I will try to help more. |
|
June 18, 2009, 06:18 |
|
#3 |
New Member
Patrick Dems
Join Date: May 2009
Posts: 4
Rep Power: 17 |
Here is a picture to clarify the problem (hopefully). It shows the mesh at one end of the cylinder and the transition to the bad mesh in the middle of the cylinder.
|
|
June 18, 2009, 14:54 |
Nested Ogrids?
|
#4 |
Senior Member
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47 |
Ok, since you are showing me the image in Fluent instead of ICEM CFD (where I could see the blocking) I can’t see why it is doing that in the back layer, but there are lots of ways to fix it…
1) If it is projecting (associated) to curves, then you have projected the inside of the OGrid (a H block) to a circle and this is what you get. If you need to capture that circle, put another ogrid inside this first one… It is easy, just restrict your index control so you only see the central blocks of the ogrid, then do an ogrid on that with faces at the ends. It will give you another layer to transition away from the curve. If you don’t need to project to the curve, then associate those edges with the surface and procede with 2) 2) If it is not projecting to curves and it is curving for some other reason, then just use move vertex to move the corner verts out along that surface. You can shape the edges or what ever it takes to get the right result. Best regards, |
|
December 21, 2009, 09:57 |
|
#5 |
New Member
Orcun Kozaka
Join Date: Dec 2009
Posts: 6
Rep Power: 17 |
Hi,
I had the same problem sometime ago. The central block was not projected onto any curves or any other geometric entity. I even do a disassociation for the edges of the central block but it did not help. Then I simply generated a cylinder geometry put an O-Grid inside it and started splitting the blocks in the axial direction. After the second split the central block lost its orthogonality and projected its edges onto a non-existent circle. Anyways I guess this is a weird bug but here is the work-around. You should do a "link edge" operation under the Edit Edge menu under Blocking tab. Basically you need to link the shape of a correct edge to the one which is messed up. It will be easier to see the messed up edges if you turn on "Projected Mesh Shape" after right clicking the Edges in the tree menu. |
|
Tags |
o-grid, several blocks, varying diameter |
|
|