[Nek5000-users] Problem when running with variable fluid properties

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

Hi again, NEK users.

I think I have found the problem. Following the table on page 17 of
http://nek5000.github.io/NekDoc/Nek_users.pdf, I was mistakenly setting
density in udiff and viscosity in utrans (field=1). Please correct this
typo in the manual. NEK examples are however correct.

Regards
Antonio

2016-03-15 20:14 GMT+01:00 ANTONIO ANTORANZ PERALES <anantora at ing.uc3m.es>:

> Hi,
>
> After successfully running some DNS cases of a turbulent pipe with heat
> transfer in NEK (
> http://www.sciencedirect.com/science/article/pii/S0142727X15000533), I am
> now struggling to run a similar case but with variable fluid properties (Re
> and Pe being functions of temperature).
>
> I am currently using Pn-Pn2 (lx2=lx1-2 in SIZE) method and stress
> formulation (IFSTRS=true), together with IFUSERVP=true (p030=1.0 in REA
> file) to be able to define the viscosity and conductivity functions.
> Initial flow conditions are taken from a previous constant fluid properties
> turbulent pipe case.
>
> My problem is that, when setting IFUSERVP=true, the flow velocities go to
> zero in very few time steps, even though I was setting constant fluid
> properties in USERVP subroutine for trial purposes.
>
> I have checked that the forcing in the momentum equation (subroutine
> USERF) is correctly set.
>
> Could you please give me some indications as to what I can be doing wrong?
> Thanks in advance.
>
> Regards
> Antonio
>
>
>
> --
> ANTONIO ANTORANZ PERALES
> Universidad Carlos III de Madrid
>



-- 
ANTONIO ANTORANZ PERALES
Universidad Carlos III de Madrid
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160330/b802f635/attachment.html>