CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   definition of "node" / "element" in CFX and ICEM CFD (http://www.cfd-online.com/Forums/cfx/88843-definition-node-element-cfx-icem-cfd.html)

 murx May 27, 2011 07:40

definition of "node" / "element" in CFX and ICEM CFD

Hi,

I'm new to CFD and have a question to which I could not find an answer in the User Guide concerning the definition of node and element in CFX and ICEM CFD.

First of all, I have a block structured mesh that was built in ICEM CFD. ICEM gives me the following information about the mesh: Total nodes: 242340 Total Elements: 262571. The mesh report in CFD Post states the same number of nodes but only 223200 elements. Why the difference? And afaik, the nodes in finite volume discretization are located in the center of the control volume (CV), so shouldn't there always be the same number of nodes and elements? I guess the reason for the difference could be the CVs on domain boundaries.

And that is where my second question comes up. In the CFX Reference Guide, chapter 15.1, there is a figure picturing the mesh structure at boundaries. According to this picture, there are nodes ON the boundary and the corresponding CVs are only half as big as the others (while having a constant node density on the edge). Is this really the case?

I want to export the CV values of an outlet and need to calculate a mean value, that's why the volume of the CVs is important to me.

 kambireddy May 28, 2011 03:01

Hiiii

http://www.cfd-online.com/Forums/cfx...des-faces.html

Let me know still if you had any Q`s?

 ghorrocks May 28, 2011 06:43

First question: ICEM puts 2D elements on boundary faces so it knows these are boundary conditions. CFX ignores these elements and instead puts the boundary conditions on the 3D elements. But this means ICEM counts the surface elements as additional elements but CFX does not.

Second question: Yes, it really is the case. Why would they put it in the documentation if it was not the case?

 baserinia May 29, 2011 12:08

Quote:
 Originally Posted by murx (Post 309497) Hi, In the CFX Reference Guide, chapter 15.1, there is a figure picturing the mesh structure at boundaries. According to this picture, there are nodes ON the boundary and the corresponding CVs are only half as big as the others (while having a constant node density on the edge). Is this really the case?
Hi,
You may want to look at the difference between cell-centred vs. cell-vertex finite volume methods. CFX uses a cell-vertex approach; in other words, the mesh vertices are the CV nodes. In cell-centred approach, however, the CV node is at the centroid of each cell.

 All times are GMT -4. The time now is 01:22.