[Daggemann]

Hi,

I am simulating seawater flow in a subsea particle filtration unit, with weak to moderate rotating flow induced by stationary curved fins.

I get convergence using all k-epsilon models using SIMPLE pressure.velocity coupling, but when switching to k-omega (standard and SST) the solution only converge for 1st Order Upwind. When switching to higher order schemes the residual oscilliates around a specific point quite far from the convergence criteria of 0.001. It is mainly the continuity residual that struggles.

I have tried under-relaxation and PISO, but they dont do the job.

As inlet condition I have velocity inlet, and the outlet is an outflow. Most of the flow is laminar, but a certain region is turbulent so I figured I had to use turbulent modelling.

Is there any basic steps I can apply in order to figure out the problem, and is it possible that the k-omega turbulence modelling law just doesn't apply in this case?

ty

[Ralf Schmidt]

Hi!

I have no experience with water, but it could be, that there are problems with your grid - see also: http://www.cfd-online.com/Forums/flu...cillation.html

If just some region is turbulent, you might use a "laminar zone" in this regions to suppresses all turbulence (in the fluid BC).

Another option would be, to change the gradient option (-> define -> models -> solver) . I have no idea, what is really behind this settings, but they helped me sometimes...

Hope, that helps a little.
Ralf