CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   STAR-CCM+ (https://www.cfd-online.com/Forums/star-ccm/)
-   -   Questions about Polyhedral and Prism layer meshing (https://www.cfd-online.com/Forums/star-ccm/177259-questions-about-polyhedral-prism-layer-meshing.html)

niablis September 9, 2016 02:36

Questions about Polyhedral and Prism layer meshing
 
5 Attachment(s)
I have created a polyhedral volume mesh that grows off of a prism layer attached to a surface. I believe have three issues currently. For those of you interested, here is the Star-CCM+ case file in google drive. Just FYI, its about 4 GB.

First, my polyhedral volume mesh has an ordered series of low quality cells. It almost seems as if the low quality cells were generated when the mesh was subdividing, but that is really just a guess. I have the polyhedral mesher selected with 1 Optimization cycle, and a quality threshold of 1. I'm not sure if this makes a difference, but I also am using a volumetric control the cell size in the entire volume, just like I'm doing to refine the wake region. Maybe the using the volumetric control leads to these low quality cells?


http://imgur.com/a/8YBym

In this next image, I have set my viewing threshold to a quality less that 0.2.

http://imgur.com/a/JRBXX

Here is a top down shot. You can see that there is the same structure throughout the entire volume. The darker sections are where volumetric control is active, reducing the cell size.

http://imgur.com/a/cH3nn



Second, in some locations, my volume mesh and my BL prism mesh seem to occupy the same space. It seems to occur when my cell size is under a certain threshold.

http://imgur.com/a/OxHJg

Here is a view showing the prism layer without the merging volume mesh.

http://imgur.com/a/qVWZx


Third, despite my best attempts, my BL prism layer quality is hit and miss. In some spots, it seems to regard curves and in others it doesn't. Some times it seems to randomly shrink in thickness. As it approaches the trailing edge of my flaps, the prism layer ends and then reappears on the trailing edge.

Prism layer projected on the cut plane and also onto the side of the step.

http://imgur.com/a/YKZVj

This image shows areas where the prism layer shrinks.

http://imgur.com/a/PTGym

Prism layer at the trailing edge

http://imgur.com/a/TsRFK


I really appreciate any and all advice you may have on any of the three topics.

Cheers!

niablis September 9, 2016 02:43

3 Attachment(s)
Attaching more images

me3840 September 22, 2016 21:29

Your prisms have trouble growing over tight corners because the mesh shrinks in size far too rapidly and your prism height is too large. Growing large prisms over small features like this is not robust, you need to either lower the prism size with a custom control or modify the prism layer min thickness percentage and the layer reduction percentage. Look for them in the user guide.

I would leave the quality settings at the default for a simple geometry like this.

Those partitions you see are probably there because you're meshing in parallel? It looks like a bug, I would disable it.

You also need to modify the volumetric blending factor, your volume control is too harsh of a stencil.

niablis September 23, 2016 01:32

me3840, thanks for your insight and advice!! Is the parallel meshing bug a known problem?


All times are GMT -4. The time now is 18:44.