[AlexC]

All,

I am trying to implement the use of rough wall functions in a nozzle case. Roughness seems to make the cases crash for no reason. Any experience determining Ks and Cs values?

Some info:

Diameter: 200 microns
Inlet Pressure: 15 MPa
Outlet Pressure: 220 kPa
Axisymmetric, 5 degree wedge
Solver: custom multiphase flow solver
Ks: 1e-6
Cs: 0.5

-There are no signs of anything bad from the residuals (pyFoamPlotWatcher image attached).

-Here's the last iteration from my outfile where it abruptly stops with a floating point exception.

Minimum pressure is 7.26454 [MPa] Maximum pressure is 15.0082 [MPa] Minimum density is 36.5875 [kg/m3] Maximum density is 718.337 [kg/m3] Maximum velocity is 133.745 [m/s]

Mass flux at centerline = 0
Momentum flux at centerline = 0
Mass flux at frontAndBackPlanes = 0
Momentum flux at frontAndBackPlanes = 0
Mass flux at frontAndBackPlanes_neg = -3.33872e-10
Momentum flux at frontAndBackPlanes_neg = -8.53365e-08
Mass flux at frontAndBackPlanes_pos = -3.33872e-10
Momentum flux at frontAndBackPlanes_pos = -8.53365e-08
Mass flux at inlet = -2.0674e-05
Momentum flux at inlet = -9.16769e-05
Mass flux at outlet = 1.70816e-05
Momentum flux at outlet = 6.56206e-05
Mass flux at wall = 0
Momentum flux at wall = 0
Time = 2.14247e-05 Net mass flux = -3.59313e-06

Mass flux through nozzle Tracer1 at boundary centerline = 0
total tracer value 0
Mass flux through nozzle Tracer1 at boundary frontAndBackPlanes = 0
total tracer value 0
Mass flux through nozzle Tracer1 at boundary frontAndBackPlanes_neg = 1.38041e-09
total tracer value 6530.31
Mass flux through nozzle Tracer1 at boundary frontAndBackPlanes_pos = 1.38041e-09
total tracer value 6530.31
Mass flux through nozzle Tracer1 at boundary inlet = -2.0674e-05
total tracer value 29
Mass flux through nozzle Tracer1 at boundary outlet = 8.60793e-73
total tracer value 1.64044e-66
Mass flux through nozzle Tracer1 at boundary wall = 0
total tracer value 109.401
Courant Number mean: 0.00340225 max: 0.500047 velocity magnitude: 130.963
deltaT = 2.65879e-09
Time = 2.14274e-05

DILUPBiCG: Solving for rho, Initial residual = 1.25259e-05, Final residual = 2.67179e-10, No Iterations 1
volume continuity errors : volume = 1.73046e-11, max error = 0, sum local = 0, global = 0
DILUPBiCG: Solving for x: solution singularity
DILUPBiCG: Solving for y, Initial residual = 1.88362e-05, Final residual = 5.72639e-10, No Iterations 1
DILUPBiCG: Solving for Tracer1, Initial residual = 1.88365e-05, Final residual = 4.22274e-10, No Iterations 1
DILUPBiCG: Solving for epsilon, Initial residual = 1.75258e-05, Final residual = 7.51016e-09, No Iterations 2
DILUPBiCG: Solving for k, Initial residual = 3.66563e-05, Final residual = 2.34053e-08, No Iterations 2

-After solving for k, it solves for Ux.

Any and all advice appreciated!

[AlexC]

Further progress: Here's the crash after compiling foam-extend-3.0 on Debug mode. This is the output of the gdb backtrace:

(gdb) bt
#0 0x00007ffff162ead4 in log () from /lib64/libm.so.6
#1 0x00007ffff5bd67ad in Foam::log (s=-0.0072670797110974744)
at /home/achan/foam/foam-extend-3.0/src/foam/lnInclude/Scalar.H:242
#2 0x00007ffff3af7d9e in Foam::compressible::RASModel::yPlusLam (this=0xd3bb10, kappa=0.40999999999999998,
E=0.95255364484361649) at RASModel/RASModel.C:162
#3 0x00007ffff3bca35a in Foam::compressible::RASModels::mutRoughWallFunctio nFvPatchScalarField::calcMut (
this=0xd41d10)
at derivedFvPatchFields/wallFunctions/mutWallFunctions/mutRoughWallFunction/mutRoughWallFunctionFvPatchScalarField.C:98
#4 0x00007ffff3bc9321 in Foam::compressible::RASModels::mutWallFunctionFvPa tchScalarField::updateCoeffs (
this=0xd41d10)
at derivedFvPatchFields/wallFunctions/mutWallFunctions/mutWallFunction/mutWallFunctionFvPatchScalarField.C:192
#5 0x00000000004c0ff5 in Foam::fvPatchField<double>::evaluate (this=0xd41d10)
at /home/achan/foam/foam-extend-3.0/src/finiteVolume/lnInclude/fvPatchField.C:237
#6 0x00000000004b12d9 in Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh>::GeometricBoundaryField::evaluate (this=0xd3c350) at /home/achan/foam/foam-extend-3.0/src/foam/lnInclude/GeometricBoundaryField.C:308
#7 0x000000000049ba1e in Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh>::correctBoundaryConditions (this=0xd3c278) at /home/achan/foam/foam-extend-3.0/src/foam/lnInclude/GeometricField.C:880
#8 0x00007ffff3b37e4c in Foam::compressible::RASModels::kEpsilon::correct (this=0xd3bb10)
at kEpsilon/kEpsilon.C:335

In the calculation of yPlusLam(), the line with the log is bolded below:

scalar RASModel::yPlusLam(const scalar kappa, const scalar E) const
{
scalar ypl = 11.0;

for (int i=0; i<10; i++)
{
ypl = log(E*ypl)/kappa;
}

return ypl;
}

So somehow E*ypl is returning negative?