Boundary conditions for LaunderGibsonRSTM discussion
 

Hi everyone,

I've set up an incompressible, steady-state simulation of a static body in a freestream using the LaunderGibsonRSTM model for turbulence solving with simpleFoam and would like some discussion regarding appropriate boundary conditions. I am not using wall functions and have what I think is a good mesh of hexas in the boundary layer and tets outside. It is a half mesh,
with symmetry on all walls except the input and output. At the moment I'm using the following boundaryField entries:

0/R: noslipwall{type fixedValue; value uniform (0 0 0 0 0 0);}
inlet{type freestream; freestreamValue uniform (0 0 0 0 0 0};}
symmetry{type symmetryPlane;}
outlet{type freestream; freestreamValue uniform (0 0 0 0 0 0};}

0/U: noslipwall{type fixedValue; value uniform (0 0 0);}
inlet{type freestream; freestreamValue uniform (10 0 0);}
symmetry{type symmetryPlane;}
outlet{type freestream; freestreamValue uniform (10 0 0);}

0/p: noslipwall{type zeroGradient;}
inlet{type freestreamPressure;}
symmetry{type symmetryPlane;}
outlet{type freestreamPressure;}

0/epsilon: noslipwall{type fixedValue; value uniform 1E-10;}
inlet{type freestream; freestreamValue uniform 0.13;}
symmetry{type symmetryPlane;}
outlet{type freestream; freestreamValue uniform 0.13;}

0/k: noslipwall{type fixedValue; value uniform 1E-10;}
inlet{type freestream; freestreamValue uniform 0.1;}
symmetry{type symmetryPlane;}
outlet{type freestream; freestreamValue uniform 0.1;}

0/nut: noslipwall{type fixedValue; value uniform 1E-10;}
inlet{type calculated; value uniform 0.22;}
symmetry{type symmetryPlane;}
outlet{type calculated; value uniform 0.22;}

Do these boundary conditions seem reasonable? I'm using low relaxation values and the solution ticks for some time and then blows up with epsilon being bounded: this makes me doubt the boundary conditions; hence the request for discussion! Please,
any comments about past working cases using this turbulence model would be greatly appreciated!

As a footnote, I'm using the following solvers:

U: smoothSolver. p: GAMG. epsilon: PBiCG. R: PBiCG.
Best regards,

Mark.

Equally, if somebody could give me a source of examples (other than $FOAM_TUTORIALS) I'd be happy to look further myself.
Best regards,
Mark.

Also, I've refined my mesh through the boundary layer and I'd like to turn off the wall functions used with this turbulence model. I've changed my constant/polyMesh/boundary file from using `wall' at the walls of my body, to `patch' in an attempt to do this and set U, R,
epsilon, k, nut to zero; and p to zeroGradient.

Is this the approach other people use to turn off the wall functions?

Best regards,

Mark.

Mark, did you ever figure this out? I'm having the same problem!