Mesh boundaries

[helios]

Dear all,

I have some questions about mesh boundaries in the case of having more
than one meshes. At the paragraph 6.3.4 of user manual, at the first
bullet, it says:

"Avoid putting mesh boundaries where critical action is expected,
especially fire. Sometimes fire spread
from mesh to mesh cannot be avoided, but if at all possible try to
keep mesh interfaces relatively free of
complicated phenomena since the exchange of information across mesh
boundaries is not yet as accurate
as cell to cell exchanges within one mesh."

I have a case where mesh boundaries crosses the source. It is a little
bit difficult to avoid that. The meshes have a common boundary plane
parallel to the x axis and have the same number of cells in x and z
direction at the boundary (one cell of one domain is totally align
with one cell of another domain - the ideal case 1 at figure 6.2).

Do you expect any loss in accuracy even in the case where the two
grids are totally aligned (case 1 at figure 6.2)?

[Dave McGill]

Hi,

If a mesh boundary is passing through your fire, then you will most likely encounter a numerical instability. (I assume you haven't run this yet.)

Think of it as the meshes having to exchange information with each other at the boundaries. You want to place the boundaries in such a way that this takes place in a relatively quiet area of the model. That usually translales into being at least a couple of meters from the fire.

Dave

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To post to this group, send email to fds...@googlegroups.com.
To unsubscribe from this group, send email to fds-smv+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

[helios]

Thank you for the answer.

No, I have run this and no numerical instability appeared. My source
is 2*10=20 m^2 and
it is a source of heat (20 MW), CO2 and CO (no combustion).

I just wanted to know of how much the results may be affected because
of the mesh boundary crossing the source.
I can't understand why the exchange of information between cells of
different meshes is less accurate than the exchange of information
between two cells in the same mesh.

[Kevin]

Try a simple test case with and without a mesh boundary crossing the fire. The issue has to do with the pressure solver. The pressure equation is solved very accurately within a mesh, but there is a less accurate coupling of the pressure field mesh to mesh. This is why it's better to keep the mesh boundary away from areas where rapid changes in velocity will occur.

> > fds-smv+unsubscribe@googlegroups.com.

[Randy McDermott]

I just want to comment that one of the key improvements in v6 will be better accuracy and stability at mesh interfaces.  There are even efforts to implement a global pressure solution (Susan Kilian's work, Kevin's pressure iteration scheme).

Dave, I appreciate your comments which I am sure are targeted at older versions of the code, I just don't want it to become lore that we cannot have mesh interfaces within a fire.  I do this routinely.  See the Sandia plume cases in the Validation Guide.

Cheers,

R

[Dave McGill]

Hi Randy,

Thanks for the update. Old habits die hard. I will try and live on the wild-side and put a mesh interface within a fire, but it won't be easy.    :  )

Dave

To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/GaveDkdOyD4J.

To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+u...@googlegroups.com.

[William (Ruddy) Mell]

Just FYI, for my work with wildland fire spread through large domains (100s of meters) I cannot avoid mesh interfaces that pass through the fire and, therefore, regions of higher air speeds.

 

Ruddy

> > fds-smv+u...@googlegroups.com.
> > For more options, visit this group at
> >http://groups.google.com/group/fds-smv?hl=en.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/GaveDkdOyD4J.
To post to this group, send email to fds...@googlegroups.com.
To unsubscribe from this group, send email to fds-smv+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

 

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To post to this group, send email to fds...@googlegroups.com.

[Dave McGill]

I still see numerical instabilities in FDS where high velocities cross a mesh boundary.

I recently worked on one where the boundary was in the plane of a door. There was a gap beneath the door with high velocities passing through the gap. The simulation crashed; the solution was to re-configure the meshes so the highest velocities were centered within a mesh rather than at the boundary.

Dave

[Randy McDermott]

Dave,

I don't know what version of the code you are using.  I assume 5.5.3.  As I was saying, there are (theoretically) improvements with the latest SVN.  If you are seeing an instability with the latest repository version of the code, please submit this as an Issue.  Much appreciated.

Best,

Randy

> > fds-smv+unsubscribe@googlegroups.com.

> > For more options, visit this group at
> >http://groups.google.com/group/fds-smv?hl=en.

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/GaveDkdOyD4J.
To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+unsubscribe@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+unsubscribe@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+unsubscribe@googlegroups.com.

> > fds-smv+unsubscribe@googlegroups.com.

> > For more options, visit this group at
> >http://groups.google.com/group/fds-smv?hl=en.

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/GaveDkdOyD4J.
To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+unsubscribe@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+unsubscribe@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To post to this group, send email to fds...@googlegroups.com.

To unsubscribe from this group, send email to fds-smv+unsubscribe@googlegroups.com.

[Dave McGill]

Hi Randy,

Yes, I'm using 5.5.3 on 64-bit LINUX. I have created stripped-down versions of both cases and started them running. If I can duplicate the results, then I will post it to the issue tracker.

I am unable to compile the latest version. The NIST LINUX version usually runs on our cluster. If you want to mail it to me, then I will get it installed and try both cases with it. (I realize that there may be changes to the input parameters that preclude that.)

Dave

[Chris]

We see some differences in the results (for example Temp) if we open
the vents at the Mesh Boundaries and if we don't open the vents. Has
somebody else encountered some differences? Fire and other critical
parameters are not placed at boundaries.

Cheers
Chris

On 8 Mrz., 16:13, Dave McGill <dwt.mcg...@gmail.com> wrote:
> Hi Randy,
>
> Yes, I'm using 5.5.3 on 64-bit LINUX. I have created stripped-down versions
> of both cases and started them running. If I can duplicate the results,
> then I will post it to the issue tracker.
>
> I am unable to compile the latest version. The NIST LINUX version usually
> runs on our cluster. If you want to mail it to me, then I will get it
> installed and try both cases with it. (I realize that there may be changes
> to the input parameters that preclude that.)
>
> Dave
>
> On Thu, Mar 8, 2012 at 8:20 AM, Randy McDermott

> <randy.mcderm...@gmail.com>wrote:

>
>
>
> > Dave,
>
> > I don't know what version of the code you are using.  I assume 5.5.3.  As
> > I was saying, there are (theoretically) improvements with the latest SVN.
> >  If you are seeing an instability with the latest repository version of the
> > code, please submit this as an Issue.  Much appreciated.
>
> > Best,
> > Randy
>
> > On Wednesday, March 7, 2012 12:59:30 PM UTC-5, Dave McGill wrote:
>
> >> I still see numerical instabilities in FDS where high velocities cross a
> >> mesh boundary.
>
> >> I recently worked on one where the boundary was in the plane of a door.
> >> There was a gap beneath the door with high velocities passing through the
> >> gap. The simulation crashed; the solution was to re-configure the meshes so
> >> the highest velocities were centered within a mesh rather than at the
> >> boundary.
>

> >> Dave- Zitierten Text ausblenden -
>
> - Zitierten Text anzeigen -

[Dave McGill]

Hi Randy,

This has been posted as issue 1599. It would be useful to see if the problem occurs with the latest iteration of FDS 6.

Thanks

Dave