[Nek5000-users] energy dissipation over a subdomain

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

Hi, Alex!
Thank you very much, sorry for delay, I've been travelling.
Best, Ilias

On Fri, Oct 30, 2015 at 6:21 PM, <nek5000-users at lists.mcs.anl.gov> wrote:

> Hi Ilias,
>
> As I mentioned to you earlier, for the case when you subdomain integration
> consists of whole elements, I would do the simplest
>
>       dimension ybm1(lx1*ly1*lz1), rmask(lx1*ly1*lz1,lelt)
>
>       nxyz = nx1*ny1*nz1
>       ntot  = nxzy*nelt
>       do e=1,nelt
>           call rzero(rmask,ntot)
>           call col3(ybm1,bm1,nxyz)
>           yavg = vlsum(ybm1,nxyz)
>           yavg = yavg/vlsum(ym1,nxyz)
>           if (yavg.gt.y_something) call rone(rmask(1,1,1,e),nxyz)
>       enddo
>
> and then use glsc3 with arguments of bm1, your quantity on GLL mesh, and
> created rmask.
>
> Note that the above approach with modification will work even for the case
> of the integration subdomain boundary cutting through the elements but you
> need to use every GLL's ym1 instead of yavg -- be aware that the spectral
> convergence of this integration with increase of polynomial order will be
> affected since it is equivalent to integration of a step/discontinuous
> function inside an element...
>
> Note also that if the integration subdomain is much smaller than the whole
> domain, you could compute the collocation with your quantity and bm1 inside
> the above loop and do summation yourself
>
> Aleks
>
>
> ------------------------------
> *From:* nek5000-users-bounces at lists.mcs.anl.gov [
> nek5000-users-bounces at lists.mcs.anl.gov] on behalf of
> nek5000-users at lists.mcs.anl.gov [nek5000-users at lists.mcs.anl.gov]
> *Sent:* Thursday, October 29, 2015 1:16 PM
> *To:* nek5000-users at lists.mcs.anl.gov
> *Subject:* [Nek5000-users] energy dissipation over a subdomain
>
> Dear neks!
>
> I would like to compute kinetic energy dissipation over different
> subdomains.
>
> It involves the calculation of the dissipation itself in GLL points,
> and integration over some domain.
>
> I understand now that it is more accurate to involve whole number of
> elements
> for integration,
> and  that for computation of  the integral over
> a subdomain one can use a mask (of 1s and 0es) over the whole domain,
> and after that to use global scalar product like glsc3.
>
> Are there the implementations of these ideas?
> first of all, the creation of a mask, based on geometrical constructions
> like
>       IF      (Y.gt.something) THEN
> ...
>
> and also if somebody has already a procedure of kinetic energy dissipation,
> though here it is more strait: calculation of of velocity gradient and
> contraction of
> its multiplication with itself,
>
> Thank you, Ilias
>
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20151102/a11c1a06/attachment.html>