[lcbuijs]

Hi dear Foamers,

Lately I've been consistently running into an error with a specific case I'm running. I'm using a slightly adapted version of buoyantBoussinesqPimpleFoam to simulate RB-convection in a cubical cavity. The adaptations I've made are:

• Add a viscosity model for variable viscosity depending on temperature
• Add variation with temperature of thermal diffusivity and thermal expansion, which I've set to a constant for the time being

For clarity, I'm not using any turbulence models and for now I'm only taking into account variations of viscosity as a T-dependent polynomial, even though the other functionality changes seem to be working properly.

Now, the problem I'm facing is that I'm consistently getting the following error for one specific case:

[0] #0  Foam::error::printStack(Foam::Ostream&) in "/opt/apps/openfoam-2.3.0/OpenFOAM-2.3.0/platforms/linux64GccDPOpt/lib/libOpenFOAM.so"
[0] #1  Foam::sigFpe::sigHandler(int) in "/opt/apps/openfoam-2.3.0/OpenFOAM-2.3.0/platforms/linux64GccDPOpt/lib/libOpenFOAM.so"
[0] #2
[0]  at sigaction.c:0
[0] #3  Foam::incompressible::alphatJayatillekeWallFunctionFvPatchScalarField::updateCoeffs() in "/opt/apps/openfoam-2.3.0/OpenFOAM-2.3.0/platforms/linux64GccDPOpt/lib/libincompressibleRASModels.so"
[0] #4  Foam::fvPatchField<double>::evaluate(Foam::UPstream::commsTypes) in "/home00/lbuys/OpenFOAM/lbuys-2.3.0/platforms/linux64GccDPOpt/bin/myBBPF"
[0] #5  Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh>::GeometricBoundaryField::evaluate() in "/home00/lbuys/OpenFOAM/lbuys-2.3.0/platforms/linux64GccDPOpt/bin/myBBPF"
[0] #6
[0]  in "/home00/lbuys/OpenFOAM/lbuys-2.3.0/platforms/linux64GccDPOpt/bin/myBBPF"
[0] #7  __libc_start_main in "/lib64/libc.so.6"
[0] #8
[0]  in "/home00/lbuys/OpenFOAM/lbuys-2.3.0/platforms/linux64GccDPOpt/bin/myBBPF"
[dutw12229:19548] *** Process received signal ***
[dutw12229:19548] Signal: Floating point exception (8)
[dutw12229:19548] Signal code:  (-6)
[dutw12229:19548] Failing at address: 0x50c00004c5c
[dutw12229:19548] [ 0] /lib64/libc.so.6() [0x3e2fc326a0]
[dutw12229:19548] [ 1] /lib64/libc.so.6(gsignal+0x35) [0x3e2fc32625]
[dutw12229:19548] [ 2] /lib64/libc.so.6() [0x3e2fc326a0]
[dutw12229:19548] [ 3] /opt/apps/openfoam-2.3.0/OpenFOAM-2.3.0/platforms/linux64GccDPOpt/lib/libincompressibleRASModels.so(_ZN4Foam14incompressible47alphatJayatillekeWallFunctionFvPatchScalarField12updateCoeffsEv+0x233) [0x2b1005e94753]
[dutw12229:19548] [ 4] myBBPF(_ZN4Foam12fvPatchFieldIdE8evaluateENS_8UPstream10commsTypesE+0x10) [0x42f7a0]
[dutw12229:19548] [ 5] myBBPF(_ZN4Foam14GeometricFieldIdNS_12fvPatchFieldENS_7volMeshEE22GeometricBoundaryField8evaluateEv+0x1ab) [0x44485b]
[dutw12229:19548] [ 6] myBBPF() [0x42a3b5]
[dutw12229:19548] [ 7] /lib64/libc.so.6(__libc_start_main+0xfd) [0x3e2fc1ed5d]
[dutw12229:19548] [ 8] myBBPF() [0x4273e9]
[dutw12229:19548] *** End of error message ***

The confusing part being that any of my cases with the exact same functionality and slightly different boundary conditions, and also the cases with the same boundary conditions but variation of for example expasion coefficient instead of viscosity, all seem to be working properly. I've even gotten the exact same case at a slightly higher resolution to work, only this very specific case keeps on getting stuck after ~100 iterations.

Any ideas/suggestions for where to look?

Many thanks!

[lcbuijs]

Oh well, I've simply set the alphat boundary condition to a fixed value. That seems to solve the problem so far.

[celez1010]

The error comes from the member function "alphatJayatillekeWallFunctionFvPatchScalarField:: updateCoeffs()"
used in the boundary condition "alphatJayatillekeWallFunction" that you assigned for the turbulent thermal diffusivity alphat.

This boundary condition can be surely used when using wall functions but I am not sure if it's ok also for low-Re simulations. What type of simulation are you doing?

In "updateCoeffs()" the code uses the viscosity and the Prandlt number to compute the thermal diffusivity therefore I think your problem is related to how/where you compute the temperature dependent viscosity (and/or how you changed the code to include a temperature dependent thermal diffusivity).

Also setting a constant value might not be a good solution (it could be an error) since your viscosity (and thermal diffusivity) change with temperature.

Hope this can help.

[lcbuijs]

Originally Posted by celez1010

The error comes from the member function "alphatJayatillekeWallFunctionFvPatchScalarField:: updateCoeffs()"
used in the boundary condition "alphatJayatillekeWallFunction" that you assigned for the turbulent thermal diffusivity alphat.

This boundary condition can be surely used when using wall functions but I am not sure if it's ok also for low-Re simulations. What type of simulation are you doing?

In "updateCoeffs()" the code uses the viscosity and the Prandlt number to compute the thermal diffusivity therefore I think your problem is related to how/where you compute the temperature dependent viscosity (and/or how you changed the code to include a temperature dependent thermal diffusivity).

Also setting a constant value might not be a good solution (it could be an error) since your viscosity (and thermal diffusivity) change with temperature.

Hope this can help.

Thank you for your reply!

I found the problem to be that I manually implemented a value for alpha, while still calculating these wall functions for alphat. I switched that off by using fixedValue (since I'm not using turbulence modelling anyway) and that fixed the problem.