Re: Digest for amrita-ebook@googlegroups.com - 2 Messages in 1 Topic
28 views
Skip to first unread message
Hans Hornung
Jul 17, 2012, 2:16:58 PM7/17/12
to amrita...@googlegroups.com
I really want to start a new question, but don't know how to do that.
OK, here goes:
I am inserting a body with ASX::InternalBoundary in a Cartesian grid
Along part of this boundary I want to set particular conditions. In the boundary conditions, I tried
Sbdy along J=31 from I=339 to 395: prescribe W'inject
No problem arose when running the script, but the boundary conditions along that line are unchanged by this.
Am I doing something really stupid? Or do I have to use a fitted grid, so that the body is indeed the Sbdy?
Hans.
OK, here goes:
I am inserting a body with ASX::InternalBoundary in a Cartesian grid
Along part of this boundary I want to set particular conditions. In the boundary conditions, I tried
Sbdy along J=31 from I=339 to 395: prescribe W'inject
No problem arose when running the script, but the boundary conditions along that line are unchanged by this.
Am I doing something really stupid? Or do I have to use a fitted grid, so that the body is indeed the Sbdy?
Hans.
On Thu, Mar 1, 2012 at 12:14 AM, <amrita...@googlegroups.com> wrote:
Group: http://groups.google.com/group/amrita-ebook/topics
Hans Hornung <hgho...@gmail.com> Feb 29 03:19PM -0800
Thanks James.
Mattei, did you forget to attach the script or am I too dumb to find it?
Hans.
James Quirk <j...@galcit.caltech.edu> Feb 29 03:25PM -0800
Hans,
On Wed, 29 Feb 2012, Hans Hornung wrote:
> Thanks James.
> Mattei, did you forget to attach the script or am I too dumb to find it?
> Hans.
It's possible that Matei's attachment was stipped off in the
digest you received. If so, you can grab a copy from:
http://groups.google.com/group/amrita-ebook/browse_thread/thread/5b3519cf23eaf16f
James
You received this message because you are subscribed to the Google Group amrita-ebook.
--
You can post via email.
To unsubscribe from this group, send an empty message.
For more options, visit this group.
You received this message because you are subscribed to the Google Groups "amrita-ebook" group.
To post to this group, send email to amrita...@googlegroups.com.
To unsubscribe from this group, send email to amrita-ebook...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/amrita-ebook?hl=en.
James Quirk
Jul 17, 2012, 3:17:30 PM7/17/12
to Hans Hornung, amrita...@googlegroups.com
Hans,
On Tue, 17 Jul 2012, Hans Hornung wrote:
> I really want to start a new question, but don't know how to do that.
>
> OK, here goes:
>
> I am inserting a body with ASX::InternalBoundary in a Cartesian grid
> Along part of this boundary I want to set particular conditions. In the
> boundary conditions, I tried
> Sbdy along J=31 from I=339 to 395: prescribe W'inject
>
> No problem arose when running the script, but the boundary conditions along
> that line are unchanged by this.
>
> Am I doing something really stupid? Or do I have to use a fitted grid, so
> that the body is indeed the Sbdy?
No, you're not doing anything silly. The problem you're coming up
against is one of ordering. Amr_sol's BoundaryConditions are applied
in two passes. The first pass will apply the W'inject by painting
the appropriate state into the ghostcells of any patch that
abutts the specified line. Then in the second pass, the
AXS::InternalBoundary is applied by reflecting the local flow normal
to the L=0 curve in the defined distance function. So if your distance
function sets a wall along J=31 (I=339 to 395) then the AXS
pass will overwrite your W'inject, resulting in no injection.
Hence your observation that W'inject had no affect.
Now the workaround is to define a distance function that
sets the shape of the body, but does not include the
nominal wall where the flow injection is to take place.
Give it a stab then if you get stuck send me a mailit
and I'll look at the problem in more detail.
James
>
> Hans.
>
>
>
> On Thu, Mar 1, 2012 at 12:14 AM, <amrita...@googlegroups.com> wrote:
>
> > Today's Topic Summary
> >
> > Group: http://groups.google.com/group/amrita-ebook/topics
> >
> > - Digest for amrita...@googlegroups.com - 4 Messages in 2 Topics<#135cd53152f015f0_group_thread_0>[2 Updates]
> >
> > Digest for amrita...@googlegroups.com - 4 Messages in 2 Topics<http://groups.google.com/group/amrita-ebook/t/9dab2301b4ab17f9>
> > To unsubscribe from this group, send<amrita-ebook...@googlegroups.com>an empty message.
> > For more options, visit<http://groups.google.com/group/amrita-ebook/topics>this group.
On Tue, 17 Jul 2012, Hans Hornung wrote:
> I really want to start a new question, but don't know how to do that.
>
> OK, here goes:
>
> I am inserting a body with ASX::InternalBoundary in a Cartesian grid
> Along part of this boundary I want to set particular conditions. In the
> boundary conditions, I tried
> Sbdy along J=31 from I=339 to 395: prescribe W'inject
>
> No problem arose when running the script, but the boundary conditions along
> that line are unchanged by this.
>
> Am I doing something really stupid? Or do I have to use a fitted grid, so
> that the body is indeed the Sbdy?
against is one of ordering. Amr_sol's BoundaryConditions are applied
in two passes. The first pass will apply the W'inject by painting
the appropriate state into the ghostcells of any patch that
abutts the specified line. Then in the second pass, the
AXS::InternalBoundary is applied by reflecting the local flow normal
to the L=0 curve in the defined distance function. So if your distance
function sets a wall along J=31 (I=339 to 395) then the AXS
pass will overwrite your W'inject, resulting in no injection.
Hence your observation that W'inject had no affect.
Now the workaround is to define a distance function that
sets the shape of the body, but does not include the
nominal wall where the flow injection is to take place.
Give it a stab then if you get stuck send me a mailit
and I'll look at the problem in more detail.
James
>
> Hans.
>
>
>
> On Thu, Mar 1, 2012 at 12:14 AM, <amrita...@googlegroups.com> wrote:
>
> > Today's Topic Summary
> >
> > Group: http://groups.google.com/group/amrita-ebook/topics
> >
> >
> > Digest for amrita...@googlegroups.com - 4 Messages in 2 Topics<http://groups.google.com/group/amrita-ebook/t/9dab2301b4ab17f9>
> >
> > Hans Hornung <hgho...@gmail.com> Feb 29 03:19PM -0800
> >
> > Thanks James.
> >
> > Mattei, did you forget to attach the script or am I too dumb to find
> > it?
> >
> > Hans.
> >
> >
> >
> >
> >
> > James Quirk <j...@galcit.caltech.edu> Feb 29 03:25PM -0800
> >
> > Hans,
> >
> > On Wed, 29 Feb 2012, Hans Hornung wrote:
> >
> > > Thanks James.
> >
> > > Mattei, did you forget to attach the script or am I too dumb to find
> > it?
> >
> > > Hans.
> >
> > It's possible that Matei's attachment was stipped off in the
> > digest you received. If so, you can grab a copy from:
> >
> >
> > http://groups.google.com/group/amrita-ebook/browse_thread/thread/5b3519cf23eaf16f
> >
> > James
> >
> >
> >
> > You received this message because you are subscribed to the Google Group
> > amrita-ebook.
> > You can post via email <amrita...@googlegroups.com>.
> > Hans Hornung <hgho...@gmail.com> Feb 29 03:19PM -0800
> >
> > Thanks James.
> >
> > Mattei, did you forget to attach the script or am I too dumb to find
> > it?
> >
> > Hans.
> >
> >
> >
> >
> >
> > James Quirk <j...@galcit.caltech.edu> Feb 29 03:25PM -0800
> >
> > Hans,
> >
> > On Wed, 29 Feb 2012, Hans Hornung wrote:
> >
> > > Thanks James.
> >
> > > Mattei, did you forget to attach the script or am I too dumb to find
> > it?
> >
> > > Hans.
> >
> > It's possible that Matei's attachment was stipped off in the
> > digest you received. If so, you can grab a copy from:
> >
> >
> > http://groups.google.com/group/amrita-ebook/browse_thread/thread/5b3519cf23eaf16f
> >
> > James
> >
> >
> >
> > You received this message because you are subscribed to the Google Group
> > amrita-ebook.
> > To unsubscribe from this group, send<amrita-ebook...@googlegroups.com>an empty message.
> > For more options, visit<http://groups.google.com/group/amrita-ebook/topics>this group.
James Quirk
Jul 18, 2012, 11:32:06 AM7/18/12
to Hans Hornung, amrita...@googlegroups.com
Hans,
After studying your mailit, all you need do is to change:
along J=31
to:
along J=1
The BoundaryConditions block applies to the boundaries of patches
which form the boundaries of the computational domain, and the
bottom of your domain is the line J=1. Just to confirm that
this is indeed the fox to your problem, I'll send you some
of the time snapshots I generate at my end.
James
along J=31
to:
along J=1
The BoundaryConditions block applies to the boundaries of patches
which form the boundaries of the computational domain, and the
bottom of your domain is the line J=1. Just to confirm that
this is indeed the fox to your problem, I'll send you some
of the time snapshots I generate at my end.
James
Reply all
Reply to author
Forward
0 new messages