CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   B.C's for k and omega (http://www.cfd-online.com/Forums/main/4281-b-cs-k-omega.html)

 Suppe January 11, 2002 17:34

B.C's for k and omega

Hi all. I'm currently working on implementing Wilcox's k-w model in my flow solver (PNS based). I'm having a lot of problem with the wall boundary condition - I am working on supersoinc turbulent flow past a flat plate. I have used k=0 at the wall, but my problem seems to be with the omega b.c for the wall. Could anyone suggest some good references/ideas please? Thanks!

 Jonas Larsson January 11, 2002 18:06

Re: B.C's for k and omega

I don't have any references with me here but I struggeld lot with this myself several years ago... If I remember correctly Wilcox has a wall-roughness model which gives an omega value based on a surface roughness. Using this model, with a surface roughness set to something which is essentially "smooth" was the most stable and best boundary condition that I found. I think that this is also what Wilcox recommends - the wall-roughness model is surely described in his book (Turbulence Modeling for CFD).

 Suppe January 11, 2002 19:21

Re: B.C's for k and omega

I did look at Wilcox's wall-roughness condition, but I was wondering if there was any other b.c for a smooth wall. I think I'll try out the rough-wall b.c anyway. Thanks for your input!

 Jonas Larsson January 12, 2002 05:41

Re: B.C's for k and omega

If you set the surface roughness paramater (if I remember it correctly Wilcox calls it kr+ or something) to a value well inside the viscous sublayer, say 1, the wall will be smooth - I did a sensitivity analysis on this a long time ago and found that for flows over turbine blades 1 was certainly enough to get a "smooth" blade. It is quite easy for you also to half the kr+ paramater and run againg just to verify that you get the same result.

 Suppe January 12, 2002 12:16

Re: B.C's for k and omega