[cfdmarkus]

Hi

I believe there is something wrong with nuSgsWallFunction in OF1.6 when used in parallel.

First, in the nuSgs file the boundary condition is set as:
cylinder{ type nuSgsWallFunction; value uniform 0; }

This setup works fine for a serial pisoFoam run!
If I run the same in parallel something is going wrong:

Courant Number mean: 0.01103945 max: 0.38140966
DILUPBiCG: Solving for Ux, Initial residual = 1, Final residual = 69.935898, No Iterations 1001
DILUPBiCG: Solving for Uy, Initial residual = 1, Final residual = 1616.3981, No Iterations 1001
DILUPBiCG: Solving for Uz, Initial residual = 1, Final residual = 15602.459, No Iterations 1001
DICPCG: Solving for p, Initial residual = 1, Final residual = 0.0098202015, No Iterations 354
time step continuity errors : sum local = 0.0001208599, global = -1.8938945e-18, cumulative = -1.8938945e-18
DICPCG: Solving for p, Initial residual = 0.98571483, Final residual = 0.074974329, No Iterations 1001
time step continuity errors : sum local = 0.061158378, global = 7.7675562e-13, cumulative = 7.7675373e-13
DILUPBiCG: Solving for k, Initial residual = 1, Final residual = 78.216523, No Iterations 1001
bounding k, min: -6.8958632e+13 max: 4.9482911e+14 average: 2.6886666e+09
ExecutionTime = 520.46 s ClockTime = 520 s


Note, when nuSgsWallFunction is replace by fixedValue, the problem in parallel mode disappears and all is well. I therefore believe that there is an issue with nuSgsWallFunction in parallel mode. I just don't know what is going wrong.

Best regards
Markus

[cfdmarkus]

I just found out that it does work under OF-1.6.x.