CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   meshing in CFX 5.5 (

henry January 17, 2002 10:04

meshing in CFX 5.5
Hi all

All are aware that CFX5.5 has a automated mesh generated,giving us no control over the mesh quality,

What I am interested is that is there a way we can control the mesh quality using something like expert parameters in build where we can say that the mximinum skewness should not exceed 0.6


is there a better way to do it other than mesh controls. As mesh adaption does not necessarily improve the mesh control.

Looking forward to your contributions............



Neale January 17, 2002 21:08

Re: meshing in CFX 5.5
Unfortunately there is no direct way to do this in CFX-Build. The first thing is that you have to make sure beforehand that you have a B-rep solid with a decent surface representation. See the meshing and CAD tutorial for some pointers on this.

One general problem with doing what you say is, depending on your geometry, is that it might be impossible for the mesher to generate a mesh with whatever "indices" you define. Furthermore, metrics like "skewness" and "aspect ratio", especially on a tetrahedral or hybrid mesh may not have anything to do with how well the solver performs an a particular grid, or with how accurate your solution will be.


Bill January 18, 2002 14:04

Re: meshing in CFX 5.5
In response to Henry, you use the mesh controls to avoid large aspect ratios. To say that they cause this problem isn't true.

If you put them in sensible places and use the proximity function, then there wont be a problem. I find the mesh controls a very powerful tool to enable me create a mesh quickly and easily.

henry January 19, 2002 02:50

thanks NEALE and BILL

thanks for your valuable suggestions,

I would like to clarify further things :

This is the first time someone has told me that aspect ratio has no effect on the solver.

please brief me on this.

thanks once again


Neale January 26, 2002 02:08

Re: thanks NEALE and BILL

I'm sorry if I came across as trying to suggest that high aspect ratio cells "never" have an effect. Let me clarify. I was just saying that some of the "standard" metrics, such as aspect ratio and skewness, are not necessarily the right metric to be looking at when deciding upon the quality of the mesh. Unfortunately the proper metrics are not available in CFX-5.

The quality of the mesh is really decided when you see how the solver behaves on it for a particular problem. The solver's behaviour on a particular mesh can be generally determined by the quality of linear coefficients which are assembled and sent to the linear solver. A good metric, which is not available, would be to look at the diagonal dominance of each control volume equation for a diffusion equaiton, and highlight control volumes which have a diagonal/offdiagonal ration >> 1. The solver will completely blow if this is the case.

Hopefully CFX will eventually add this and some other mesh quality indices along these lines.


All times are GMT -4. The time now is 17:46.