CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   asymmetric solution generated by symmetric problem (http://www.cfd-online.com/Forums/cfx/20648-asymmetric-solution-generated-symmetric-problem.html)

Twiti September 16, 2004 10:45

asymmetric solution generated by symmetric problem
 
Hi, everyone

Now, I am doing with a geometrically symmetric problem in CFX-5.7, but got asymmetric solutions either with k-e or SST turbulence model. The problem involves only fluid flow and heat transfer. It's sure that boundary conditions and mesh are both symmetric. Physically, a symmetric solution is expected.

Is it a problem due to poor mesh generation or some thing else?


Robin September 16, 2004 17:40

Re: asymmetric solution generated by symmetric pro
 
It is also possible that asymmetry in your mesh could cause what you are seeing. If you make your mesh symmetric or further refine it, a symmetric solution may be realized.

Keep in mind, however, that symmetric boundary conditions and geometry do not necessarily imply a symmetric flow. For instance, flow in a diffuser may separate from one side only, remaining attached on the other. In such a flow, the 'symmetric' solution is unstable and therefore is not realized. It's like standing a pencil on it's tip.

Regards, Robin

HK September 16, 2004 23:02

Re: asymmetric solution generated by symmetric pro
 
I think the only practical way to use axi-symmetry geometry for cfx 5.7 is to use mesh build/build in 4 if the problem is very complex.

I am just wondering what did you use to generate the mesh?

HK

Twiti September 16, 2004 23:24

Re: asymmetric solution generated by symmetric pro
 
hi, Robin

Thank you for you reply.

The problem I am dealing with is a relative simple one without any separation and other unstable factors.

Just now, I temporarily changed the two lateral walls to "Wall" type boundary, which both used to be "Symmetry" type, and got a expected symmetric solution. I wonder maybe that's the problem of the "Symmetry" bounday type?

Twiti September 16, 2004 23:28

Re: asymmetric solution generated by symmetric pro
 
Hi, HK

Thank you for you kindly answering.

Right now, I use ICEM5 for mesh generation.

Do you mean that, CFX-5 get use to the mesh built up by CFX-build, especially when the geometry is very complex?


HK September 16, 2004 23:44

Re: asymmetric solution generated by symmetric pro
 
In ICEM, how difficult is to make 2 sets of identical mesh which will be used as symmetry for a? I think it is very difficult.

So yes, still my earlier comments hold.

HK


Glenn Horrocks September 19, 2004 19:31

Re: asymmetric solution generated by symmetric pro
 
Hi HK,

It is very easy to make 2 sets of identicle mesh in ICEM. You can transform an existing mesh group to a new location and it's done.

Alternately if you mean create periodic boundary conditions with matching meshes on both faces, that can be done easily in ICEM too.

And for the record, you can create 2D axisymmetric or planar meshes in ICEM. Basically you create a solid with one face being your 2D geometry, then you delete all the volume elements and the shell elements except the shell elements describing the 2D geometry, then you recreate the 2D volumes by sweeping the mesh along a line or an axis of revolution. Finally you have to manually sort the edge elements out into the correct parts. It sounds complex but it is quite easy.

Regards, Glenn

SAM October 16, 2004 20:12

Re: asymmetric solution generated by symmetric pro
 
how can be the two walls be symetry!!!!!!.

it is wrong thats why u were getting wrong solution

regards


All times are GMT -4. The time now is 12:40.