[Lukas7117]

Hi, I’m trying to conduct 3d transient analysis of natural convection. However using k-w SST model I receive messeage that turb. Viscosity ratio is limited. Such thing happens only when I use water as a fluid, for air it’s ok.

After first time step 0.1 (during internal 30-40 iteration) the problem dissapear and residuals are converging. I change the limit to 10^7 for tueb. viscosity ratio.



My computational domain dimension are 80 mm x 100 mm x 150 mm (height). Side walls are adiabatic, heated wall is at the bottom and the top wall is cooled.



I tried diffrent types of mesh, finer, coarse, polyhedra, hexcore. It doesn't help at all.



I'm wondering if I made some mistake during specification of boundary conditions or if my mesh is really inappropriate.

[LuckyTran]

Well, definitely make sure all your boundary conditions are the correct ones because any result would be meaningless if they are not.


Is this a time-accurate simulation or just a transient calculation of an otherwise stationary flow? If time-accurate, then you have to be 100% sure your initial condition is correct. If you're just running transient because you like running transient models, then it just depends on whether your initial guess is good enough or not. Turbulent viscosity ratio being limited is common if your initial guess is poor. It's the turbulence equivalent of your temperature, pressure, or velocity, being a little off. I don't recommend changing the value of the limiter because it is already non-physically large, changing it really doesn't impact anything because the converged solution should never have this level of turbulent viscosity ratio. Again, it's the turbulence equivalent of having a maximum temperature of 5000 K. You can change it, but it's never the problem.