[Nek5000-users] Outflow divergence

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

Hi Markus,

This question came up earlier for the splitting code...
I apologize that I don't recall the precise resolution
but it's clear that there is a problem for the splitting
case.

I'll try to recall precisely how we fixed it for that
particular case.

Paul


On Sun, 25 Jul 2010, nek5000-users at lists.mcs.anl.gov wrote:

> Howdy neks,
>
> I updated my divergence-for-outflow code from the latest repo of the jet 
> example case (since I have a suspicion the old one is messing with my 
> simulation upstream of the outflow) and found the following issue:
> "
>      if (ifsplit) then
>         do i=1,ntot2
>            dd = dist(i,1,1,1)*dist(i,1,1,1)
>            dd = min(dmax2,dd)
>            if (dist(i,1,1,1).ge.0) usrdiv(i,1,1,1) = cdiv*(1.-dd/dmax2)
>            if (dist(i,1,1,1).ge.0) nd = nd+1
>         enddo
> "
> This seems to add user divergence linearly in the outflow elements (from zero 
> at the farthest distance from the outfow face to some value at it), which is 
> intended, but also adds divergence everywhere else (dd&dist are zero in all 
> non-outflow elements). This of course causes undesired acceleration in domain 
> parts where I'd like to keep a constant flow rate, say.
>
> Was the intention: "if (dist(i,1,1,1).gt.0) usrdiv(i,1,1,1) = 
> cdiv*(1.-dd/dmax2)"?
> How is "ifsplit true" treated differently from "false" in this regard?
>
> Thanks,
> Markus
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>