Sawtooth=.FALSE. replacement in FDS6?

[Jeremy]

Upon getting my simulation to run, I noticed that the "stair-stepped" geometry was not being smoothed out as I had expected it to be. In FDS5, this functionality was provided by placing SAWTOOTH=.FALSE. on the &OBST lines in order to minimize the impact of "stair-stepped" boundaries. In FDS6 this has been changed. In the user guide, the footnote to section 9.1.7 states that "This functionality [SAWTOOTH=.FALSE.] is replicated by setting FREE_SLIP=.TRUE. on the SURF line that defines the properties of the obstructions." 

It appears that this does not work the same way as it did in FDS5. The resulting simulation provides significantly different results. 

Using the sawtooth verification input file supplied with the FDS5 distribution (sawtooth.fds), we changed the input file to comply with FDS6 syntax (saw.fds). We also had to disable the SLIP_FACTOR variable, as it no longer seems to be supported by FDS6. Both files are attached. We ran the original input using FDS5 and the modified input using FDS6. The results are significantly different. Both SMV outputs were visualized using Smokeview 6.1.5, screencaptures of which are also attached.

Here are the results of the version that was run in FDS5:

These are the results from the FDS6 version:

As can be seen, Smokeview no longer recognizes the sawtooth smoothing boundary condition on the bottom half of the geometry. In addition, not only are the results between the two versions distinctly different, the simulation's upper and lower halves of the FDS6 version show symmetry.  This means that the FREE_SLIP=.TRUE. setting is not smoothing out the stair stepping, and that the turbulence caused by the stair steps is all but unchanged by this parameter, unlike with SAWTOOTH=.FALSE. in FDS5.

Does anyone know of a way to replicate the FDS5 smoothing functionality of SAWTOOTH=.FALSE., in FDS6?

Thank you,
Jeremy

[Randy McDermott]

The comment in the FDS 6.0.1 User Guide is wrong and will be corrected in future releases.

In the current SVN we state the SAWTOOTH has been ELIMINATED.  Further, regarding FREE_SLIP=T, this parameter sets the wall stress to zero (removes viscous friction). It is not equivalent to the SAWTOOTH=.FALSE. functionality from versions of FDS prior to Version 6.

Simply put, DON'T use SAWTOOTH=F, whether using v5 or v6.

 

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/1573780b-6515-473a-9d20-f20fd60024d1%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Jeremy]

Thank you.

Is there some way to replicate the smoothing functionality in v6, though?  If not, is the functionality planned to be added back in?


Also, why shouldn't I use SAWTOOTH=F with v5?

Thanks again,
Jeremy

[Randy McDermott]

The "smooth" functionality never "worked", per se.  It gave different results, but they were never verified or validated.  All SAWTOOTH=F did was turn off one of the vorticity terms.  This is not the same thing as setting the wall normal component of the viscous stress to zero.  There is not way to do this with v6.  The correct way to handle this is with high order immersed boundaries.  This is under development.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/e80c5e0e-888a-4489-a7ad-69130d1b0c41%40googlegroups.com.

[Jeremy]

 Thank you very much.  Is there any chance there is a beta version that has an implementation of the high order immersed boundaries?  We currently have FDS installed in an HPC environment, and would be happy to help test it, if possible.

Thank you again,
Jeremy

[Randy McDermott]

Sorry, but no.

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/cf3d38bd-7177-4a0b-901f-401447e1b825%40googlegroups.com.

[Cian Davis]

First time doing this with FDS6 and it's prompted 2 questions for me.

I would use the some materials (i.e. SURFs) for both curved and
non-curved geometry. Previously, I could just apply SAWTOOTH=.FALSE. to
the OBSTs that make up the curved geometry and use the same SURF_ID for
both. Is there a better way to apply the FREE_SLIP=.TRUE. to the curved
OBSTs than duplicating the SURF and applying the FREE_SLIP=.TRUE. to the
new SURF?

Another advantage previously of the SAWTOOTH=.FALSE. is that smokeview
would smooth them out and it would look nice. Is there a way to get
FDS6 to smooth out the geometry visually only?

Regards,
Cian

> <https://lh3.googleusercontent.com/-r2B7uviTH5w/UxYT7K_UOdI/AAAAAAAAAGQ/FnCfbyLzxOI/s1600/sawtooth_s0000.png>
>
>
>
> <https://lh5.googleusercontent.com/-xFJHOH3xQyA/UxYUmBMWm2I/AAAAAAAAAGY/c5X_sCr3jUQ/s1600/sawtooth_0258.png>

>
>
>
>
>
> These are the results from the FDS6 version:
>

> <https://lh4.googleusercontent.com/-G3dohZIjyDg/UxYVDXTkNBI/AAAAAAAAAGg/DBnxtTplqgk/s1600/saw_s0000.png>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> <https://lh3.googleusercontent.com/-70V-1SRon8U/UxYVUMSYvvI/AAAAAAAAAGo/sxjT_fcUT5g/s1600/saw_0423.png>

>
>
> As can be seen, Smokeview no longer recognizes the sawtooth
> smoothing boundary condition on the bottom half of the geometry. In
> addition, not only are the results between the two versions
> distinctly different, the simulation's upper and lower halves of the
> FDS6 version show symmetry. This means that the FREE_SLIP=.TRUE.
> setting is not smoothing out the stair stepping, and that the
> turbulence caused by the stair steps is all but unchanged by this
> parameter, unlike with SAWTOOTH=.FALSE. in FDS5.
>
> Does anyone know of a way to replicate the FDS5 smoothing
> functionality of SAWTOOTH=.FALSE., in FDS6?
>
> Thank you,
> Jeremy
>
> --
> You received this message because you are subscribed to the Google
> Groups "FDS and Smokeview Discussions" group.
> To unsubscribe from this group and stop receiving emails from it,
> send an email to fds-smv+u...@googlegroups.com

> <mailto:fds-smv%2Bunsu...@googlegroups.com>.

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

> <mailto:fds...@googlegroups.com>.

> To view this discussion on the web visit
> https://groups.google.com/d/msgid/fds-smv/1573780b-6515-473a-9d20-f20fd60024d1%40googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "FDS and Smokeview Discussions" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to fds-smv+u...@googlegroups.com.
> To post to this group, send email to fds...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/fds-smv/CAC6wax0T0RJzwin-nE7yc7ap%3D3WCydgk7_3rWNqQf%3DmydvsVoA%40mail.gmail.com.

[Randy McDermott]

The answer to both your questions is no.

We took SAWTOOTH out because it did not work.  It was not based on sound physics.  We do not want to give false visual impressions of the accuracy with which boundaries are treated.

By the way, these statements apply equally as well to older versions of FDS.  If you are using SAWTOOTH, stop.  If you see anyone else using it, call them out.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/537F3888.6020903%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Marietta_13]

Hi Randy,

I used to use SAWTOOTH in FDS 5.5.3 when modelling fires in curved tunnels. But after seeing this discussion, I took it out and ran one of my old models again. I expected the runtime to be longer but even after 200 hours the job still hadn't completed. The calculation had stopped sometime before the 70 hour mark (perhaps even earlier). 

When I had run the job using SAWTOOTH=.FALSE. , the job would complete in just over 24 hours.

Will you please be able to help me with this?

Thank you

Marietta

>     <mailto:fds-smv%2B...@googlegroups.com>.

[Randy McDermott]

There is a list of possible ways to speed up the calculations in Sec. 1.4 of the FDS User Guide.  But it is very difficult for me to help based on what I know from this thread.  As always, if you believe there to be a bug in the code, we need a simple test case that illustrates the problem.  Whatever the issue is SAWTOOTH is a red herring.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/31e1f7f0-5e6e-47b4-8fc1-5f0feda5241f%40googlegroups.com.

[brookl...@gmail.com]

Hi ,Jeremy!

I am confused with the same problem of you ,SAWTOOTH=.FALSE. and FREE_SLIP=.TRUE. ,could you tell me how to do How do you deal with this problem at  last ,

BEST ragards .

在 2014年3月5日星期三 UTC+8上午2:05:11，Jeremy写道：