Running buoyantSimpleFoam with oodles data as initialisation
 

Hi, I am fairly new to using OpenFOAM and I am experiencing difficulties running a case in buoyantSimpleFoam in OF v1.6, here are the details:-

• a case of ventilation in a room with inlet and outlets
• initially run in OF v1.4 with oodles by a colleague
• currently trying to run same case with buoyantSimpleFoam then with buoyantSimpleRadiationFoam in order to study the effect of accounting for radiation in solution.

what I have done:-

• changed dimension for p, since initialisation data was run as incompressible and buoyantSimpleFoam is a compressible solver

I get the following errors when I run buoyantSimpleFoam
Starting time loop
Time = 0.002
DILUPBiCG: Solving for Ux, Initial residual = 1, Final residual = 9.96992404881e-06,
DILUPBiCG: Solving for Uy, Initial residual = 1, Final residual = 9.09756548062e-06,
DILUPBiCG: Solving for Uz, Initial residual = 1, Final residual = 9.08723417676e-06,
DILUPBiCG: Solving for h, Initial residual = 1, Final residual = 9.13020536236e-06, N
DICPCG: Solving for p, Initial residual = 0.899530285296, Final residual = 1.52436599
time step continuity errors : sum local = 4.14841430109e-05, global = -2.97540166135erho
max/min : 4.34385853381 -0.172041623519
DILUPBiCG: Solving for nuTilda, Initial residual = 0.524688238581, Final residual = 3
bounding nuTilda, min: -0.865777461241 max: 3.42116944364 average: 2.88491347876e-05
ExecutionTime = 447.08 s ClockTime = 447 s
Time = 0.004
DILUPBiCG: Solving for Ux, Initial residual = 0.92584355044, Final residual = 5.48969
DILUPBiCG: Solving for Uy, Initial residual = 0.714844540147, Final residual = 5.2607
DILUPBiCG: Solving for Uz, Initial residual = 0.986980735918, Final residual = 7.8297
DILUPBiCG: Solving for h, Initial residual = 0.96042233089, Final residual = 3.014970
#0 Foam::error::printStack(Foam::Ostream&) in "/usr/local/OpenFOAM/OpenFOAM-1.6/lib/l
#1 Foam::sigFpe::sigFpeHandler(int) in "/usr/local/OpenFOAM/OpenFOAM-1.6/lib/linux64G
#2 ?? in "/lib64/libc.so.6"
#3 Foam::hPsiThermo<Foam::pureMixture<Foam::sutherlan dTransport<Foam::specieThermo<Fo
in "/usr/local/OpenFOAM/OpenFOAM-1.6/lib/linux64GccDPOpt/libbasicThermophysicalModels.
#4 Foam::hPsiThermo<Foam::pureMixture<Foam::sutherlan dTransport<Foam::specieThermo<Fo
"/usr/local/OpenFOAM/OpenFOAM-1.6/lib/linux64GccDPOpt/libbasicThermophysicalModels.so
#5 main in "/usr/local/OpenFOAM/OpenFOAM-1.6/applications/bin/linux64GccDPOpt/buoyant
#6 __libc_start_main in "/lib64/libc.so.6"
#7 _start at /usr/src/packages/BUILD/glibc-2.9/csu/../sysdeps/x86_64/elf/start.S:116
Floating point exception
aeroflo@gws18:~/OpenFOAM/aeroflo-1.6/run/aircraftCabinHVAC_cp>

any help pointing me in the right direction as to what I am doing wrong will be greatly appreciated. Thanks.

The problem is that the output you posted is badly clipped (incomplete final residual, no number of iterations). Therefor I can only guess: it seems that at the first time-step the nuTilda didn't converge (the bounding is a strong indication) and from then on everything goes. Try switching off everything that can be switched off (turbulence etc) and try again. Try smaller time-steps. Write out every time-step and have a look where the solution "looks funny". BTW: you also changed the value of p, didn't you?

Ups. Just saw: in the first time-step you have negative densities. Man, you have a serious problem

Hey Bernhard,

thanks for the comments. besides changing the dimension for the p file, how can I scale/change the p values? this is a file from a previous run and it is filled with 'nonuniform <Listscalar>' of p fields

Throw all of that away. Honestly. It is from an incompressible run (you said so) and has probably values above and below 0. What you need is the "real" pressure. The one that fits the perfect gas equation (and therefor can only be above zero and somewhere in the range of 1e5 for "room"-conditions). It is probably the source of your negative densities.

Or you write a utility to "transpose" the pressure. Or use funkySetFields. Or maybe there is a util that already does that in the distro

Bernhard

Thanks, I will look into that.