CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   ICEM block mesh with a ball in a cylinder tube (http://www.cfd-online.com/Forums/cfx/26118-icem-block-mesh-ball-cylinder-tube.html)

Li July 18, 2008 08:50

ICEM block mesh with a ball in a cylinder tube
 
Hi everyone, I am trying to generate structure mesh in this geometry that has a ball in a cylinder tube in ICEM. I am quite new to ICEM and I just completed the tutorials. I am confused in the following problems: (1) Is it complicated to generate structure mesh in this case that has a ball in a cylinder tube. I only want to mesh the area between the tube and the ball (2) Can I just associcate the edges with the curves without assocaite vertexes with points first? If I cannot, why I should associate vertexes with points first? (3) What is the mean of material point exactly? It is said that the material point is not necessary in Hex mesh but it is quite important in generating tetra mesh. I cannot understand it.

They are probably stupid questions, but I do appreciate if you could give me some hints. Thank you very much!

Charles July 20, 2008 12:41

Re: ICEM block mesh with a ball in a cylinder tube
 
For hexa you can ignore the material point. It is used to tell the tetra mesher which part of the domain to mesh into which family. You can get away without associating the vertices explicitly, but you need to either move or snap them into appropriate positions. They will not be projected automagically when you create the pre-mesh. Ball in a tube sounds pretty simple, you should be able to do it with a cylindrical O-split running the length of the cylinder. Then do a plain split either side of the ball, and O-split that interior block.

Jerry Tanner July 20, 2008 22:48

Re: ICEM block mesh with a ball in a cylinder tube
 
Bingo!I totally agree with Charles.He's a pro.

Li July 21, 2008 04:43

Re: ICEM block mesh with a ball in a cylinder tube
 
Thank you very much! I will try again. I only need to mesh the area between the cylinder and the ball, so I nedd to delete the block around the ball, and then slect the block left to perform the process of O-grid, right?

Jerry Tanner July 21, 2008 11:41

Re: ICEM block mesh with a ball in a cylinder tube
 
Sorry to say no,Li.Please try the following steps: step1,make a big block around the cylinder. step2,divide the big block into 3 small blocks.one is on the left side of the ball,one is on the right side of the ball,and the left one just keeps the ball inside. Do O-blocking for the one which keeps the ball inside.Delete the mesh inside the ball and simply keep other meshes.

There's no way uploading pictures in this forum or we can make this easier by pictures :)

Li July 22, 2008 03:37

Re: ICEM block mesh with a ball in a cylinder tube
 
Jerry, thank you very much! It seems it's not the same thing to delete the block and delete the relatvie part mesh.(I though if I add the block to the part, let's say the ball surface, and then I delete the block which will result in deleting the mesh of the ball).

Would you please give me some hints to block this one: there is a connection of a ball and a cone inside a cylinder tube and I only need the mesh between the cylinder and the ball-cone. I don't know how to deal with the vetex and the acme of the cone.

Here is what I did, but it didn't succeed. I initialize the cylinder first with the block, and I split the block to four small blocks from left to right. And then I split the blcok around the ball to three parts from top to bottom to make the block is more fittable to the ball. After that, I split the block around the cone to three parts from top to bottom, then I encounter the problem. After I associate the four vextexes to the acme, I cannot get the mesh I wanted. It seems twisted.

Sorry for the long post, thank you very much for any suggestions!

Regards, Li


Gary Bristow August 11, 2008 05:44

Re: ICEM block mesh with a ball in a cylinder tube
 
Jerry Tanner, the steps which you have posted are excellent. I think this the right way to divide into meshes. Methods were appreciable, intelligent. Divide 0-bblocking to mesh the ball. ------------------ Greg Bristow WideCircles



All times are GMT -4. The time now is 05:43.