|
[Sponsors] |
June 19, 2007, 17:46 |
interpolate with batch file
|
#1 |
Guest
Posts: n/a
|
Hi,
I would like to interpolate the values from a file name test.res on the mesh from the file named solve.def. I would like to use the command line. In the manual there is something written like -interpolate-iv But when I write: cfx5solve -def solve.def -interpolate-iv test.res ...and so on, it doesn't work. So, what's wrong? Thanks Claudia |
|
June 19, 2007, 19:08 |
Re: interpolate with batch file
|
#2 |
Guest
Posts: n/a
|
Hi,
Try cfx5interp. Glenn Horrocks |
|
June 20, 2007, 04:30 |
Re: interpolate with batch file
|
#3 |
Guest
Posts: n/a
|
Hi Glenn,
this is what i don't want, because I would like to interpolate and start the job in one step, because I am a user of a cluster and I wait a long time to start one job. So I had wo wait for the interpolation and again for the solver run. So can you tell me how the -interpolate-iv command works? thanks |
|
June 20, 2007, 06:25 |
Re: interpolate with batch file
|
#4 |
Guest
Posts: n/a
|
Hi Claudia,
The following works for me: cfx5solve -batch -def solve.def -ini test.res And edit your .def file to include: EXECUTION CONTROL: RUN DEFINITION: Interpolate Initial Values = yes END END You can add this manually to your .def file, or use the -ccl argument. Regards, Michel |
|
June 20, 2007, 06:38 |
Re: interpolate with batch file
|
#5 |
Guest
Posts: n/a
|
Hi Michael,
thanks for the answer, but the -ini switch only works if you have an equal mesh. I need the -interpolate command, because I have different meshs. Or is this command from you: ... Interpolate Initial Values = yes END END ... for interpolating on a different mesh? Thanks Claudia |
|
June 20, 2007, 07:52 |
Re: interpolate with batch file
|
#6 |
Guest
Posts: n/a
|
Hi,
when you say 'it doesn't work' do you mean that it fails, or that the job proceeds but the interpolation doesn't happen? I think the syntax of the command you need should be cfx5solve -def <def> -ini <res> -interp-iv ... Is this what you are specifying? Regards, Johnson |
|
June 20, 2007, 07:54 |
Re: interpolate with batch file
|
#7 |
Guest
Posts: n/a
|
sorry that should be
cfx5solve -def file.def -ini file.res -interp-iv ... |
|
June 20, 2007, 08:24 |
Re: interpolate with batch file
|
#8 |
Guest
Posts: n/a
|
Sorry, the formatting got a bit screwed up before.
Indeed, "-ini test.res" only specifies that the flow solution from test.res should be used as initial guess. And for some reason CFX then also takes the mesh from test.res, not the one specified in the .def file. By adding the following CCL commands: EXECUTION CONTROL: RUN DEFINITION: Interpolate Initial Values = yes END END the flow solution from test.res will be interpolated onto the mesh in solve.def. Perhaps the argument -interp-iv should work, but I don't find it in the solver documentation. Michel |
|
June 20, 2007, 08:34 |
Re: interpolate with batch file
|
#9 |
Guest
Posts: n/a
|
Another go at decent formatting:
EXECUTION CONTROL: RUN DEFINITION: Interpolate Initial Values = yes END END |
|
June 20, 2007, 10:05 |
Re: interpolate with batch file
|
#10 |
Guest
Posts: n/a
|
Dear Johnson and Michel,
thanks for the help. The command: cfx5solve -def file.def -ini file.res -interp-iv ... works! My problem was, that the -ini file.res was missing. I only used -interp-iv file.res that was wrong. I will test the modification in .ccl at another time, but I think that it is also a good way. Thanks for all! Claudia |
|
June 21, 2007, 00:32 |
Different meshes?
|
#11 |
Guest
Posts: n/a
|
Just a clarification, you mentioned different meshes. Are they substantially different, or is one simply a scaled version of the other. The reason i am asking is that I thought you could not change the mesh that much, eg the number of boundary conditions, the connection of the nodes to form elements etc.
Is it possible to do the following. I have a small rectangular tank, and I run a multiphase sim that fills the tank with water to 50% full. I then increase the height of the tank by 20% and remesh it. Can I interpolate the previous transient results onto the new mesh as the inital values, ie the new height part of the tank would contain only air, not water. ps I dont not care about what the air is doing, only the flowing of the water. Thanks Stu |
|
June 21, 2007, 05:57 |
Re: Different meshes?
|
#12 |
Guest
Posts: n/a
|
Stu,
the interpolation should be successful - the overlap will contain the correctly interpolated solution from the previous results file, and the 'new' part should have an approximate solution based on near/surrounding nodes. However, you should probably check the interpolated solution before embarking on the new transient analysis. Johnson |
|
June 21, 2007, 07:17 |
Re: Different meshes?
|
#13 |
Guest
Posts: n/a
|
Thanks for your response.
Where in the documentation does it mention cfx5solve -def file.def -ini file.res -interp-iv ... , specifically the -interp-iv part. I have gone through the solver manual and connot see it in there. Question how do I check the interpolated results, do I just run the sim for 1 more time step and then examine the results? Thanks Stu |
|
June 21, 2007, 09:30 |
Re: Different meshes?
|
#14 |
Guest
Posts: n/a
|
You'll find this option in the command line documentation, i.e. try
cfx5solve -h |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
[swak4Foam] GroovyBC the dynamic cousin of funkySetFields that lives on the suburb of the mesh | gschaider | OpenFOAM Community Contributions | 300 | October 29, 2014 19:00 |
2.0.x on Mac OSX | niklas | OpenFOAM Installation | 74 | March 28, 2012 17:46 |
channelFoam for a 3D pipe | AlmostSurelyRob | OpenFOAM | 3 | June 24, 2011 14:06 |
Version 15 on Mac OS X | gschaider | OpenFOAM Installation | 113 | December 2, 2009 11:23 |
gcc and executable file from Mac to Linux | simone Marras | Main CFD Forum | 0 | April 8, 2007 16:49 |