Compressible heat transfert steady state
Submitted by pcaste12 on April 5, 2012 - 02:58
Hi,
I am currently moving to a test using compressible steady state heat transfer model (trying to simulate a vortex tube) and I got the following error:
Warning:
volume_13 [www.1483.5] RANS Flow: Failed
Unexpected process exit.
--> FOAM FATAL IO ERROR:
Expected a ')' while reading binaryBlock, found on line 15 the word 'ÉZ?'
file: /tmp/of-7e8FAB42/constant/polyMesh/points at line 15.
From function Istream::readEnd(const char*)
in file db/IOstreams/IOstreams/Istream.C at line 111.
FOAM exiting
I probably make a mistake :-(
Could kindly help me?
Many thanks,
Pierre.
Precision has to be single
Have you changed the preference: Physics->OpenFOAM->Precision to Double? It has to be set to Single. This preference is only for writing the case, Caedium does not ship with double precision solvers.
Precision is single
Hi,
I played with it and in the second case it was single. Do I have to restart the whole project from scratch?
That leave me with another question, as far as I know OpenFoam is a double precision solver, so why Caedium Rans Flow is in single precision?
Again, many thanks for your answers.
Pierre.
Single vs double precision
You don't need to restart from scratch, just save the project (.sym) and then reload it using File->Revert.
OpenFOAM solvers can be either single or double precision. Single precision uses significantly less memory so it's preferred for the solvers Caedium supports. Double precision is necessary for very high-aspect ratio cells targeted for Low-Re Turbulence models, LES, and multiphase-flow.
Running... but not too far
Thanks for your tips, it is working now, I have a run which is starting.
Unfortunately it crash after less than 20 iterations.
I ve reviewed the Re number (aprox of course) it is about 3000 for the inlet so the single precision should be ok. Also I define some BC according some test that I have done in OpenFoam (using blockMesh), but some of the definitions you are building are different. Some example in OF use "pressureInletVelocity" but then they have define totalPresure. Again I fixed k to 0.01 (expecting 1% of energy) but then I get 1.5e-6 in the files. I can't also define the outlet temperature according to the fact that I define outlet as "massFlowRate".
Finally I increase the number of cells (on OF I am around 10⁶) to 588k, even thought I get the following error:
Warning:
VT [www.1295.5] RANS Flow: Failed
Solver singularity: GAMG: Solving for p, Initial residual = 0.847719, Final residual = nan, No Iterations 100
Sorry for those simple and stupid questions but I will be keen to have this example working.
Pierre.
Switch BCs around
I can't comment on "some example in OF" - do you think it is correct?
If you want to specify total pressure then try the Inlet-Outlet condition and set Physics:Inlet-Outlet->Type = Atmosphere. However, I haven't found this type of condition to work with a mass flow outlet.
For your case why not try mass flow inlets and then use standard (default, zero gradient) outlets?
To remove the turbulence as a variable, you can try setting Substance->Solver->Turbulence Model = Laminar, given your Reynolds number is quite low.
I find the compressible OpenFOAM solvers to be tricky to setup and even then non-robust. This could just be my lack of detailed understanding.
Note previously when I said Low-Re number turbulence models - it relates to not using wall functions and instead integrating to the wall. It is not a requirement that you have a low Reynolds number flow to use such models.
Test in progress
Thanks for your fast answers.
For the time being, I am not enough good to be able to say if OF example are correct, especially when they dealing with test case that I didn't studied, don't worry..
I was setting up my case with mass flow, somewhere else, I will try with caedium.
I should have turbulence model to get the whole case working in theories, I will also try without to see if it is better but then the results should not be the one expected.
Sorry for that I am not yet at a good level, I am learning...
Thanks again.
Pierre.
Restarting the case from the geometry
For information I have restart the case from the geometry and I get a similar error:
Warning:
VT [www.1273.5] RANS Flow: Failed
Unexpected process exit.
--> FOAM FATAL IO ERROR:
Expected a ')' while reading binaryBlock, found on line 15 an error
file: /tmp/of-b2242415/constant/polyMesh/points at line 15.
From function Istream::readEnd(const char*)
in file db/IOstreams/IOstreams/Istream.C at line 111.
FOAM exiting
Many thanks,
Pierre.