[bigfootedrockmidget]

A new tutorial has been placed online, explaining the setup of a simulation for the turbulent incompressible flow in a 90 degree pipe bend using wall functions:


https://su2code.github.io/tutorials/Inc_Turbulent_Bend/


More information about turbulence modeling with wall functions in general and with SU2 can be found in the presentation given at the SU2 conference:


https://www.youtube.com/watch?v=x6PMfacKuK8&t=743s


Let us know about your experience!

[savedjuli]

Thanks for your hard work on including the possibility of using wall functions. I do have a doubt, that maybe you can shed some light on it.

I'm running some cfd analysis for a comercial transport aircraft; the complicated cases are for the high Mach numbers, i.e. transonic, and high angles of attack, I need to develop the lift curve up to stall.

I have a very fine mesh, that computationally is very heavy, so takes a lot of time to run; I'm not obtaining the expected results, or even for the really high Mach numbers, I'm not even obtaining convergence; so I was wondering, if I use the wall functions capability, so I can use thicker layers for the boundary layer mesh, so y+ won't be ~1 (like te current one) but bigger, would that help on the convergence issue, and in obtaining accurate results in less time?

Thanks!

[bigfootedrockmidget]

Yes with wall functions you could use for instance y+=30. But please check the paper of nichols and nelson (2005) for applicability. If you have flow separation the wall model results will not perform very well. But I guess it is better than having no results. Still, it might be good to see if you can make the wall resolved case converge by improving the mesh quality or modifying the numerical settings.