I was running the dnsFoam tutorial case on a 128^3 mesh and I am running into some issues.
I get the following error msg:
Time = 0.175
Number of forced K = 12
k(0.175) = 1.12316e+48
epsilon(0.175) = 2.81871e+49
U.f(0.175) = -1.60366e+22
#0 Foam::error::printStack(Foam::Ostream&) in "/usr/local/OpenFOAM/OpenFOAM-1.5/lib/linux64GccDPOpt/lib
#1 Foam::sigFpe::sigFpeHandler(int) in "/usr/local/OpenFOAM/OpenFOAM-1.5/lib/linux64GccDPOpt/libOpenFOA
#2 ?? in "/lib64/libc.so.6"
#3 Foam::PBiCG::solve(Foam::Field<double>&, Foam::Field<double> const&, unsigned char) const in "/usr/l
#4 Foam::fvMatrix<Foam::Vector<double> >::solve(Foam::Istream&) in "/usr/local/OpenFOAM/OpenFOAM-1.5/ap
#5 main in "/usr/local/OpenFOAM/OpenFOAM-1.5/applications/bin/linux64GccDPOpt/dnsFoam"
#6 __libc_start_main in "/lib64/libc.so.6"
#7 Foam::regIOobject::readIfModified() in "/usr/local/OpenFOAM/OpenFOAM-1.5/applications/bin/linux64Gcc
k and eps seem to go completely off the chart, do you have any idea what is going on?
I also tested dnsFOAM a while and I had to reduce the time step for finer meshes!
I always thought that dnsFoam does not work in parallel. Do you run your 128^3 on a single processor?
I am very interested in your results, since I was never able to successfully use dnsFOAM. Are you getting a nice energy spectrum and do you know who to set the parameters for a certain Reynolds number?
Yeah that makes sense, I'll reduce the time step.
I also found out that dnsFoam doesn't work in parallel. I submitted the 128^3 case to one processor. It is not ideal, but it seems to be working.
I just started looking at dnsFoam so I don't know anything about setting the Re #, but I'll share my results for E(k) when the case is finished.
how are you getting on with dnsFoam, have you managed to get reasonable results?
I am still struggeling to get anything that looks like a energy spectrum !!
I am struggling to get anything remotely close to a reasonable energy spectrum as well.
The solver takes way too long to get a decent calculation done. Although it uses a spectral method, the fact that it cannot be used in parallel is not good. I am trying to figure out a way for it to work in parallel, but it's taking some time to understand the solver completely.
I found a bug in dnsFOAM see here:
If you fix the bug you will get a nice energy spectra, providing the forcing parameters are set correctly.
Thanks for letting me know. I appreciate it.
I'll try to some cases with this modification.
|All times are GMT -4. The time now is 16:46.|