CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   slidding grid (http://www.cfd-online.com/Forums/cfx/19596-slidding-grid.html)

nyatoto April 29, 2003 03:41

slidding grid
 
Hi All,

I use both CFX4.4 and CFX5.5.1. In the manual, I have seen that it is written somewhere that when you have SLIDDING GRIDS/UNMATCHED BOUNDARY then the REBLOCK in the VOLMSH should be OFF. However, in the Application example5 (CFX4.4) given in the manual, (in which unmatched boundary is used) the intruction says that " Make sure REBOLCK IN VOLMSH is ON!!

I have been doing my simulation with the re block on with unmatched boundary set, could some one tell me whether I am doing the correct thing.

Note that I have compared my problems with experimental data and other people simulations, and I can see that there is good agreement.

Further, VOLMSH warns me that, when the reblock is on the names of the block will change. However, when i read the mx.geo file for both the ON and OFF approaches, I cant see any difference!!

Could there be something I am missing?

thanks for taking your time to read this.

thanks nyatoto

John May 9, 2003 15:59

Re: slidding grid
 
Nyatoto,

I use CFX4.4 with silding grids and always reblock in volmesh. I have not encountered any difficulty resulting from reblocking, but it's a good idea to review the new block structure if you use the create patch command, output monitor block, or one of the user fortran routines.

Typically volmesh significantly reduces the number of blocks, so it is unlikely that the geo file for both the on and off approaches would be identical unless the geometry is already optimized.

John

Nyatoto May 11, 2003 10:34

Re: slidding grid
 
Hi John, Thanks for taking your time to read this!

Just try two problems; one with reblock ON and the other one with OFF. I tried it and it appeared to me that they were identical when I read the geo file. However, I do agree with you,if you look at them in Postprocess, they are different.

I have just started using SG. Something amazes me! In that system I had a rotating 3D patch for the inner block (call it USER3_ROTOR). When I set the rotating parameters for MFR, the USER3D_ROTOR name was recognized, however SG does not recognise this. If a give it the 'nick mane' that VOLMSH gave it with the reblock ON (say, BLOCK-NUMBER-1) it accepts that.

Now the question comes; Why cant SG recogise the name I gave and yet MFR does? In fact, I could have expected both SG and MFR not to accept the names I gave because VOLMSH renamed my blocks anyway. Thus, I would expect the name VOLMESH gave to override the names I gave in Build.

Why the difference??

Regards nyatoto

John May 12, 2003 10:35

Re: slidding grid
 
Nyatoto,

I would like to hear more about how and why you name the blocks in build. I can't say that I do that.

If you use volmesh to optimize the block structure, the number of blocks in the geo file will normally decrease. The number of blocks is the first number on the second line of the geo file. Compare the number of blocks in the two geo files so see if volmesh has reduced the number of blocks.

Volmesh does not change the names of patchs assigned in build. So the name of your 3d patch will be the same for both a volmesh case and a non volmesh case.

Since your SG case did not accept the 3d patch name but did accept the name BLOCK-NUMBER-1, you sould verify name for the 3d patch by reviewing the name geo file. Make sure that the 3d patch name entered in the command file under grid motion parameters is identical to the name in the geo file. My experience is that both SG and MFR will work with the a USER3D patch name.

Regards,

John

Nyatoto May 13, 2003 07:18

Re: slidding grid
 
Hi John,

Thanks for your contribution.

I do not name the blocks, I create a 3D-patch with a group of blocks( say BLOCK-NUMBER-1,2,3, 4 etc).What i saw VOLMSH does with REBLOCK ON is that the group that comprised the USER3D-XX is given a name like BLOCK-NUMBER-1, and now SG was not accepting my USER3D-XX name.

What i did is that I took the 'fc' and 'geo' files from the my steady state simulation and modified the fc, putting in the transient parametrs. It refused to accept my 3D-patch!! If you said that it should ACCEPT, then I will take your word and relook at the whole thing.

regards nyatoto

Nyatoto May 13, 2003 07:19

Re: sliding grid
 
Hi John,

Thanks for your contribution.

I do not name the blocks, I create a 3D-patch with a group of blocks( say BLOCK-NUMBER-1,2,3, 4 etc).What i saw VOLMSH does with REBLOCK ON is that the group that comprised the USER3D-XX is given a name like BLOCK-NUMBER-1, and now SG was not accepting my USER3D-XX name.

What i did is that I took the 'fc' and 'geo' files from the my steady state simulation and modified the fc, putting in the transient parametrs. It refused to accept my 3D-patch!! If you said that it should ACCEPT, then I will take your word and relook at the whole thing.

regards nyatoto

John May 13, 2003 17:03

Re: sliding grid
 
Hi Nyatoto,

Yes, I use the USER3d patch name in the gid motion parameters. PATCH NAME 'USER3D_XX' I guess you can continue to use the block name as long as the rotating core is composed from a single block.

Regards,

John


Nyatoto May 14, 2003 05:05

Re: sliding grid
 
Thanks John, I will try it again. nyatoto.


All times are GMT -4. The time now is 13:18.