# [ICEM] O-Grid meshing help

 Register Blogs Members List Search Today's Posts Mark Forums Read

March 17, 2021, 13:06
O-Grid meshing help
#1
New Member

Rhys Powell
Join Date: Feb 2021
Posts: 3
Rep Power: 2
Hi guys,

As part of my final year project at university, I'm modelling the flow around a cricket ball. This is my first time using ICEM and have generated the following mesh surrounding the ball. As you can see, at the seam position (marked in red) the mesh does not fall directly downwards with the actual geometry. In an ideal world, the elements on the seam would face directly downwards - to capture the seam edges. Any tips or tricks would be greatly appreciated.

Thanks
Attached Images
 cricket mesh.PNG (163.8 KB, 18 views)

 March 30, 2021, 06:38 #2 Senior Member   Sebastian Engel Join Date: Jun 2011 Location: Germany Posts: 557 Rep Power: 18 You need to associate the edges respective edges to the respective curves. What you see is the automatic association "face to nearest surface". Since you do not constrict an edge to follow a curve some faces span over a ridge. That will result in unaligned elements which are going to be patched as you can see in your screen shot. It seems you already have introduced splits left and right of the center to deal with the ridge. You now need to associate the circumferential edges to the respective curves (red and blue curves marked by you in the screen) There might be a few steps involved to further improve the boundary layer. You could introduce sublayers which have only the thickness of the ridges and map flush with it. That way you would not introduce sharp angle elements in the ridges proximity. However, try the simple approach first. p.s. i don't know the cricket-specific terminology to describe the ridge/ledge, i hope you understood aero_head likes this.

March 30, 2021, 08:12
#3
New Member

Rhys Powell
Join Date: Feb 2021
Posts: 3
Rep Power: 2
Quote:
 Originally Posted by bluebase You need to associate the edges respective edges to the respective curves. What you see is the automatic association "face to nearest surface". Since you do not constrict an edge to follow a curve some faces span over a ridge. That will result in unaligned elements which are going to be patched as you can see in your screen shot. It seems you already have introduced splits left and right of the center to deal with the ridge. You now need to associate the circumferential edges to the respective curves (red and blue curves marked by you in the screen) There might be a few steps involved to further improve the boundary layer. You could introduce sublayers which have only the thickness of the ridges and map flush with it. That way you would not introduce sharp angle elements in the ridges proximity. However, try the simple approach first. p.s. i don't know the cricket-specific terminology to describe the ridge/ledge, i hope you understood
Hi Bluebase,

Thanks for your reply, I've managed to fix the problem I was initially having. Please could elaborate on what you mean by subplayers? A sketch would be incredibly helpful, if that's not too much trouble.

Thanks

 April 3, 2021, 15:56 #4 Senior Member   Sebastian Engel Join Date: Jun 2011 Location: Germany Posts: 557 Rep Power: 18 I meant introducing sub-layers like this, a meridian cross-sectioon: refined-ogrid.jpg You'd split the initial o-grid close to the spherical surface a few times to flush off the two lowest layer with the ridges. the third layer from the surface(black lines in the sketch), is optional. Anyhow. As you can see, the sketch draws a much more complex blocking. If your current mesh is good enough, i wouldn't suggest to go for a more complex structure. aero_head and PowellRhys like this.

April 5, 2021, 04:21
#5
New Member

Rhys Powell
Join Date: Feb 2021
Posts: 3
Rep Power: 2
Quote:
 Originally Posted by bluebase I meant introducing sub-layers like this, a meridian cross-sectioon: Attachment 83689 You'd split the initial o-grid close to the spherical surface a few times to flush off the two lowest layer with the ridges. the third layer from the surface(black lines in the sketch), is optional. Anyhow. As you can see, the sketch draws a much more complex blocking. If your current mesh is good enough, i wouldn't suggest to go for a more complex structure.
Thanks a lot for the sketch, it's really helped. I've managed to recreate it for a 2D model however am struggling to replicate the same for my 3D model. Any tips for associating in the 3D domain. The border of the sphere does not appear in ICEM for me to associate the edges with.

 April 5, 2021, 14:27 #6 Senior Member   Sebastian Engel Join Date: Jun 2011 Location: Germany Posts: 557 Rep Power: 18 Have a look into the geometry>curves menu. Here you'll find a feature which can extract the boundary curves from surfaces, and if necessary an other feature which creates iso-curves on surfaces. And if this doesn't work, you might need to create curves by intersecting two surfaces---the ball's surface and a plane at the right position.