[Nek5000-users] Nek5000-users Digest, Vol 81, Issue 23

[nek5000-users at lists.mcs.anl.gov]

Leopold,

Hi...  Interesting LES model; I am also working on a flow where I am 
adapting the Turb Channel work ...

Without analyzing it too deeply, your .usr code looks OK...  Not 
withstanding any problems with the application of the model, have you 
considered using a shorter time step?  The CFL seems kind of high; my 
runs won't work anywhere near that number ;-); although my code is fully 
transient and is adding body force for a rotating turbine blade based on 
time step.  When I was working up my code, I did use time steps that 
would allow runs of several hundred cycles before it blew up.  The fact 
that the program is reducing your DT *could* be an indication that you 
need a shorter DT...

Thanks,
Murph

--------------------------------------------------------------------------------------------------------------------
Dear Neks,

I try to implement the Vreman LES model (see An eddy-viscosity 
subgrid-scale model for turbulent shear flow: Algebraic theory and 
applications, A. W. Vreman, Phys. Fl., 16, 10, 2004) to simulate a 
turbulent jet. To do so, I modified the turbChannel example provided in 
NEK. However, as soon as I impose in the uservp subroutine a 
space-varying viscosity, after some time steps, the simulation blows up 
even if this viscosity remains close from the standard 1/Re viscosity. 
Has anyone ever experienced this kind of issue? Does anyone have an idea 
on how to fix this problem? Below are the .usr file and the log of the 
simulation.

Thanks,
Best regards,

Léopold

-------------------------------------------
.usr file

.........

------------------------------------------------------------------------------------------------------------------------

-- 
Murphy Leo O'Dea
PhD Candidate
Oakland University
mlodea at oakland.edu