[allenfieldin]

Hi, Nielsen,

I am currently trying to introduce the turbulence model into the waveFoam solver, and I am suffering from the k-epsilon bounding problem. and the error seems to arise suddenly in a certain time step(for my case, it is after 4000 time steps):

time step continuity errors : sum local = 6.57563e-07, global = 2.12232e-09, cumulative = -8.22289e-05
GAMGPCG:  Solving for pdcorr, Initial residual = 1, Final residual = 3.41321e-08, No Iterations 11
GAMGPCG:  Solving for pdcorr, Initial residual = 1.27396e-08, Final residual = 1.27396e-08, No Iterations 0
GAMGPCG:  Solving for pdcorr, Initial residual = 1.27396e-08, Final residual = 1.27396e-08, No Iterations 0
GAMGPCG:  Solving for pdcorr, Initial residual = 1.27396e-08, Final residual = 1.27396e-08, No Iterations 0
Second continuityErrs.H :
time step continuity errors : sum local = 8.20024e-13, global = 9.95863e-15, cumulative = -8.22289e-05
Check Courant No :
Courant Number mean: 0.00308485 max: 0.045008 velocity magnitude: 0.264631
volume continuity errors : volume = 0.394011, max error = 1.95057e-11, sum local = 5.4868e-15, global = 5.10928e-18
No of Alpha subcycle is: 1
Solving Alpha Equation. 
MULES: Solving for alpha1
Liquid phase volume fraction = 0.441953  Min(alpha1) = -3.37222e-29  Max(alpha1) = 1.00107
Relaxing time: 0.03 s
DILUPBiCG:  Solving for Ux, Initial residual = 0.00526774, Final residual = 1.86658e-05, No Iterations 1
DILUPBiCG:  Solving for Uz, Initial residual = 0.00900043, Final residual = 3.5278e-05, No Iterations 1
GAMG:  Solving for pd, Initial residual = 0.00214831, Final residual = 1.85688e-05, No Iterations 1
GAMG:  Solving for pd, Initial residual = 1.85843e-05, Final residual = 9.88188e-08, No Iterations 4
GAMG:  Solving for pd, Initial residual = 9.86914e-08, Final residual = 9.86914e-08, No Iterations 0
GAMGPCG:  Solving for pd, Initial residual = 9.86914e-08, Final residual = 5.34721e-09, No Iterations 2
time step continuity errors : sum local = 1.71494e-09, global = -1.30375e-11
DILUPBiCG:  Solving for epsilon, Initial residual = 0.00087466, Final residual = 1.17416e-09, No Iterations 3
DILUPBiCG:  Solving for k, Initial residual = 0.000304489, Final residual = 6.88509e+29, No Iterations 500
bounding k, min: -1.01028e+31 max: 7.81151e+30 average: 5.85741e+26

I am wondering do you have some better suggestions for turbulence model selection (such as, different turbulence model have different bahavior at the surface) and waht kind of fvSchemes are suitable for wave propagation analysis. Have you already done such kind of analysis? Before k-epsilon, I tried with RNG k-epsilon, but it seems also have the bounding problem.

It is greatly apprecaited if you can give some hint and suggestion on that, it would be helpful for my current work.

I have attached the test case file mesh-fine-0.63-U3-kEpsilon.tar.gz, it is a 2D case(I know it is unsuitable for 2D, but it is just a test case). Hope you can have a minute looking into my case.

Thanks and best regards,

from Allen

[ngj]

Good morning Allen,

A good place to begin with respect to turbulence models for breaking waves is the following reference:

https://journals.tdl.org/icce/index....icle/view/7485

Furthermore, if you are having problems with stability of your turbulence model, then I would suggest to use upwinding on k and epsilon to see, whether this choice in schemes solves the problems.

Kind regards,

Niels