CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   ICEM stopping the prism mesh (http://www.cfd-online.com/Forums/cfx/24452-icem-stopping-prism-mesh.html)

Wooster August 26, 2007 12:58

ICEM stopping the prism mesh
 
I have a problem with ICEM stopping the prism mesh early and failing to finish adding the layers required. The mesh isn't particularly large (refined) compared to some others (that work). Everything seems to be working, but once it reaches the 8th layer (out of 10) it just stops the prism application. I can't find any error messages that tell me what is going on. Anyone have any ideas what is causing this?

-W

Wooster August 26, 2007 14:53

Re: ICEM stopping the prism mesh
 
One of the most frustrating things is finding the obvious after hours of checking. For anyone experiencing this, examine the sizing of your prism mesh. Sometimes a 0.001 will look like a 0.0001.


Pankaj August 28, 2007 05:50

Re: ICEM stopping the prism mesh
 
I was facing this problem for prisms.

But Whenever I did it before any smoothing operation on tet,it worked.

And prisms always downs the quality of mesh hence I switched to prism warpage of 0.01. Never gone beyond this.

If u got any solution,convey it.

Thanks and regards, Pankaj.

Wooster August 29, 2007 06:17

Re: ICEM stopping the prism mesh
 
I'll try that, but I found in my particular case that the prism height began to overrun itself and the geometry. So for instance a prism height of 0.001 was actually larger (when the layers came out) than my geometry. I changed the tetra scheme and inflation ratio (to test different ideas) and found that it had a lot to do with the sizing. Sometimes I could get another layer in before it stopped. That's when I noticed that my 0.001 should be a 0.0001 (a typo on my part). I'm not sure why it didn't warn me that the mesh would not fit in the geometry. Right now that's my chief suspect and if the mesh is built with geometry in mind, this shouldn't happen (as I found out by a mere typing mistake).

-W

Dr. FLow Squad August 30, 2007 04:32

Re: ICEM stopping the prism mesh
 
I often use a different approach when creating prisms in my existing volume mesh. First I only create one layer of prisms setting the "Initial Height" to the size of my "Global Element Scale Factor" or higher and the "Prism Height Limit Factor" to 0.95 (Important!). After running the prisms mesher I smooth everything but "Penta_6" which are "freezed". Finally I split the prisms to 5 to 20 Number of layers using a prism ratio of 1.2. That gives you a good prism mesh... -Dr. Flow squad http://k43.pbase.com/g6/33/263333/3/...2.O4ipnxLx.jpg

Wooster August 30, 2007 14:56

Re: ICEM stopping the prism mesh
 
I never thought about making my layers like that. Makes sense. I'll try that on a couple of my models that have nasty mesh issues. Thanks

-W

Wooster September 4, 2007 11:28

Re: ICEM stopping the prism mesh
 
I tried this on some of my simpler meshes and I did get very good results. However, the thickness of the prism mesh is key to good answers and I've been playing around with this setup to identify an easy way of controlling the thickness. Basically, I need an initial height of 0.00075(") and I'm wondering how the setting would work since in your example you set it to the global Element Factor. I'm not 100% sure about if I should change the prism height limitation factor (0.95 as prescribed). Any thoughts on this?

-W

Dr. FLow Squad September 5, 2007 04:16

Re: ICEM stopping the prism mesh
 
No. I rarely set the initial height, since I often work with small cylinders (ducts) that opens up into very large cylinders. Therefore I have different cell size depending on the diameter of the cylinder. Thats why the Prism height limitation factor works well for me. Maybe you could elaborate with an image?

Wooster September 5, 2007 11:24

Re: ICEM stopping the prism mesh
 
Not sure I can send an image yet (patent stuff, etc.). I asked the question earlier for another issue but the gods haven't answered yet. Basically I have the same idea where I have a tube that's about 0.055" going into a huge annulus (2"). A mesh study revealed that the thickness of the prism layer can really determine results, hence I have to be careful about how thick the prism layers can be. I'm still tinkering with the setup and hopefully I can get a measure on it. Thanks a bunch!

-W


All times are GMT -4. The time now is 15:43.