[Nek5000-users] Temperature Issue, non physical values for certain Pr num.

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

Hi Michael,

This is a problem arising in the convective terms due to insufficient
resolution.  As a result, the convective term can create over/under
shoots of a convected quantity by numerically creating anti-diffusion
and steepening gradients instead of simply convecting them.  This will
occur where gradients of the convected quantity are high.  As to the
Prandtl # effect, what is happening is that the numerical problem
arising in the convective term is suppressed by the increased physical
diffusion in the temperature equation at lower Prandtl #.  Obviously
this is not good, but the only solution at present in NEK is to increase
resolution.   Another approach would be to use a non-linear convective
scheme such as ENO, TVD, Flux Corrected Transport etc. which add
diffusion only in regions of high (and under-resolved gradients), but
NEK does not (at present) have such schemes, and I believe that
implementing such schemes in the context of Finite element methods is
not all simple.

My two-cents worth would be that adding such schemes would be a great
addition to NEK.  I have run into the same problem you are facing.

Cheers,
Frank 

On Wed, 2010-11-17 at 23:21 -0600, nek5000-users at lists.mcs.anl.gov
wrote:
> Hi,
> 
>  
> 
> I am having an issue with temperature values in my domain. The domain
> consists of two inlets, one at a temperature of 300 and the other 100.
> It appears
> 
> that the interface between the two is giving non-physical
> temperatures, but what is interesting is the magnitude of these
> over/under shoots is dependent on 
> 
> the Prandtl number.  To see the extend of this I saved the max/min
> temperature in the domain for several time steps.
> 
>  
> 
> See the attached image.  The low Pr case had Pr = 0.005, where as the
> other case had Pr = 0.71.  Both cases show non - physical results, but
> the low Pr
> 
> case seems to recover and approach the limits for the domain given by
> the boundary conditions.  In the other Pr = 0.71 case, the
> fluctuations and magnitudes are
> 
> much larger and don't appear to stabilize and recover. 
> 
>  
> 
> Again, I have looked at fld files to confirm that these large
> differences occur at the interface between the hot/cold fluid.  Is
> there a solution to this problem?
> 
>  
> 
> Thanks,
> 
> Michael 
> 
> 
> 
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
-- 
Frank Herbert Muldoon, Ph.D. Mechanical Engineering
Technische Universität Wien (Vienna University of Technology)
Inst. f. Strömungsmechanik und Wärmeübertragung (Institute of Fluid
Mechanics and Heat Transfer)
Resselgasse 3
1040 Wien
Tel: +4315880132232
Fax: +4315880132299 
Cell:+436765203470
fmuldoo (skype)
http://tetra.fluid.tuwien.ac.at/fmuldoo/public_html/webpage/frank-muldoon.html