CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Periodic grid-interfaces in FLUENT 5 (http://www.cfd-online.com/Forums/main/785-periodic-grid-interfaces-fluent-5-a.html)

tom May 4, 1999 12:46

Periodic grid-interfaces in FLUENT 5
 
Hy,

i try to make a periodic (rotational) grid interface in FLUENT 5, because i want to do some sliding mesh calculations. but the "building" of the grid-interface doesn't work. when i press "create", there appears an error message. if i have walls instead of the periodics, then it works. the two interfaces are at the same position, but have different meshes. is this perhaps a problem ?

thank you for your help

tom

Alberto Tamm May 5, 1999 03:50

Re: Periodic grid-interfaces in FLUENT 5
 
Hy Tom,

If you will make periodic boundary condition with Fluent the mesh in the interface must be the same. The best way to do this or sometimes the only way is (for example in Gambit) to link both faces. So you will be sure that the meshes are the same. Donīt assign periodic boundary condition in Gambit. Declare it as wall and assign periodic rotational B.C. in Fluent.

Iīm sure you will solve your problem if you follow this steps.

Alberto

Jonas Larsson May 5, 1999 07:45

Re: Periodic grid-interfaces in FLUENT 5
 
You have to have the same grids on the two periodic interfaces, that is probably your primary problem.

However, there is also a bug in the handling of periodic interfaces/sliding meshes in the latest version of Fluent. Some cases seem to work and others don't work. Fluent are aware of it and last I heard the rumor said that it now was fixed in the 5.1 beta. Perhaps someone from Fluent can confirm this? I'm involved in a European research project in the turbomachinery field and several of the partners have had similar problems with this.


tom May 5, 1999 13:41

Re: Periodic grid-interfaces in FLUENT 5
 
Thank you for your help. The mesh has to be the same on both sides. It moves now !!. But there are other problems. 1.) If I calculate a timestep, I can save the cas and dat file, but I cannot read the cas file anymore. If I delete the interface and save then the 2 files, I can read them. 2.) If I partition the mesh for a calculation on a multiprocessor-machine with the serial solver and try to read it into the general solver of the multiprocessor-machine, there appears an error message.

Are this also failures of the actual fluent-version ? I think, the people at fluent are less informed than cfd-online-readers.

Thank you again very much for your help.


Alberto Tamm May 6, 1999 05:49

Re: Periodic grid-interfaces in FLUENT 5
 
Hello tom,

You must save always the case and dat file for every time step. If you have problems with reading the case file try to delete the grid interface before saving, and then create it again after you read the case file.

I donīt think it is a problem of the fluent version.

As Jonas said, the first Fluent5 version have some problems with periodic boundary condition. In Fluent 4.2 it works good. The Fluent version 5.0.4 solved this problem. Ask them.

Alberto

Jonas Larsson May 6, 1999 06:51

Re: Periodic grid-interfaces in FLUENT 5
 
Fluent 5.0.4 did solve some problems, but we still have problems with several cases.

Alberto Tamm May 6, 1999 07:00

Re: Periodic grid-interfaces in FLUENT 5
 
Hello Jonas,

Now Iīm working with the Fluent5.0.4 version and I didnīt find errors. First I solved only one impeller channel with periodic boundary condition. Now I simulate the complete pump. It can be possible that my solutions have problems, so I will be very greatfull if you said me what kind of problems didnīt solve good fluent. Did you find some problems with MRF or Sliding Mesh without periodic B.C.

Thanks

Alberto

Jonas Larsson May 6, 1999 10:10

Re: Periodic grid-interfaces in FLUENT 5
 
If the case runs you don't have the problems I'm talking about - we get "floating point errors" when defining certain interfaces, negative volumes when sliding meshes etc. I don't know what causes it. The case is a quite complex 1.5 stage axial turbine including cavities and rim-seals, but we've also reproduced the problem in a simple "sliding sector" case that we've sent to Fluent. As I said, if your case runs you don't have the same problem.

Sung-Eun Kim May 13, 1999 20:27

Re: Periodic grid-interfaces in FLUENT 5
 
Dear client,

It's a known problem that sliding mesh + periodic boundary sometimes doesn't work properly.

The problem is fixed in FLUENT 5.1, which will be released in a matter of days. Please contact your support engineer to obtain v5.1.

Sorry for the inconvenience.


All times are GMT -4. The time now is 03:22.