RE: Validation Case for HT3D

116 views
Skip to first unread message

Arulnambi Palani

unread,
Jan 25, 2023, 3:58:08 AM1/25/23
to FDS and Smokeview Discussions
Dear Everyone,

I am trying to do a validation case for HT3D model. Currently, I am facing some problems during the set-up of the model. I am referring to a paper for the geometry and its set-up.

The problem I am facing is that, I am unable to apply boundary condition (hot wall) for the solid material. The simulation is running but the temperature of the hot wall is not taken into account. I have tried different definition forms of &OBST and &VENT. But, I couldn't make it work. If the solid is floating in space without attaching to any surface, then the simulation works perfectly. I figured out that the problem is occurring when I place the boundary condition (here, backside surface of the solid) at the exterior mesh domain.

Is there any way to do this set-up? or am I doing anything wrong?

I have attached the fds file and also a figure of my geometrical set-up. In the input file, I have commented the various other definitions that I have considered for OBST/VENT. Please share your feedback.

Thank you.
input.fds
Geometry_setup.png

Randy McDermott

unread,
Jan 25, 2023, 6:30:06 AM1/25/23
to fds...@googlegroups.com
Assuming you are using v6.7.9, you cannot directly apply a TMP_FRONT bc to a thermally thick solid (this will be possible in v7).  For now, you need to follow the example in the verification folder (of v6.7.9) for fds/Verification/Heat_Transfer/ht3d_ibeam.fds.    Basically, what you will do is extend your domain slightly to provide an air gap, set the air to 37 C, and then apply a very high heat transfer coefficient to the surface with H_FIXED... convoluted, but it's the only way with that version of the code.

Alternatively --- and since you are doing research I would suggest this --- update your code to the latest source (or download the latest test bundle) and you can apply TMP_FRONT directly.  Note, however, that the HT3D algorithm has changed in the new code (see Tech Guide) but it is for the better, the case should run faster.

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/a88a0a5d-cb0f-49d0-9f03-a0308c7b021bn%40googlegroups.com.

Kevin McGrattan

unread,
Jan 25, 2023, 10:02:40 AM1/25/23
to fds...@googlegroups.com
The HT3D methodology will change in FDS 7. If you are still using the official 6.7.9 release, you can look at the ht3d_ibeam case in the Examples folder. 

Arulnambi Palani

unread,
Jan 26, 2023, 7:49:28 AM1/26/23
to FDS and Smokeview Discussions
Thank you Randy & Kevin for your feedbacks.

Hi Randy,

I have tried the method you suggested with the current version. I made an air gap and filled it with hot gas (37 C). I was looking for H_FIXED in the manual, but I could not find it. So, I used HEAT_TRANSFER_COEFFICIENT under &SURF. No matter how big a value I increase, the heat transfer within the solid is not happening. Any other suggestions?

Meanwhile, I would also like to observe the simulation with FDS 7. Where can I find the source files for FDS 7?

Thank you.

Kevin McGrattan

unread,
Jan 26, 2023, 9:01:25 AM1/26/23
to fds...@googlegroups.com
We keep the latest test bundles here


If you install this, you need to read the manual because the HT3D is different than it was before.

Arulnambi Palani

unread,
Feb 2, 2023, 10:00:01 AM2/2/23
to FDS and Smokeview Discussions
Hi Kevin,

My question is regarding the manuals for FDS7. I tried making it using the make_file.sh and using texstudio. But I was not successful.

I used the fds source file from GITHUB and the compilation works.

Thank you.

Kevin McGrattan

unread,
Feb 2, 2023, 12:03:34 PM2/2/23
to fds...@googlegroups.com
Unsuccessful -- can you give us more specifics.

Arulnambi Palani

unread,
Feb 6, 2023, 7:25:22 AM2/6/23
to FDS and Smokeview Discussions
Hi Kevin,
It said that the figures were not found. I tried the scripts folder, in which a file tries to copy files from NIST server. But it didn't work.
Anyhow, I was told the link to download the pdf directly.

I am now using FDS7 for the same case discussed in this thread. I am using the new parameter  'NEIGHBOR_SEPARATION_DISTANCE'. With FDS6, the simulation was throwing me segmentation fault. With FDS7, the simulation runs but after 4 seconds, it says numerical instability (Room.out file attached).

Simulation setting: I made sure that 3D solid is not intersecting the computational domain. But as I go for finer mesh settings, the simulation is crashing.

Any suggestion on this front?

Thank you.
input.fds
slurm-547963.out

Kevin McGrattan

unread,
Feb 6, 2023, 9:31:34 AM2/6/23
to fds...@googlegroups.com
There are many things wrong with this case:

1. Your gas phase grid cells are stretched much too much. 
2. Your CELL_SIZE should be more or less the same as your gas phase grid size, and the gas phase grid should be close to an aspect ratio of 1.
3. Why do you start with such a big case? Try something much smaller, and use verification cases to assure that your assumptions are OK.

All this being said, the 3D capability in FDS is limited. If you want a fully 3D solid phase calculator, you need to link to another code. Proceed with caution. Go slow and work your way up. Do not start with a 48 mesh case.

Arulnambi Palani

unread,
Feb 6, 2023, 10:36:34 AM2/6/23
to FDS and Smokeview Discussions
Hi Kevin,
Thank you for your feedback. Why this big case, because I was trying to study the increase of temperature of the solid body through hot gas entering through vent. But, even with 800°C hot air, I could only obtain 100°C on the surfaces of the solid. So, we wanted to check the convection for both 1D and 3D heat transfer model. Therefore, my professor wanted to check Conjugate Heat Transfer with a validation case. For this current case, we have results that can be compared to.

Initially, we could not apply boundary condition (like no slip) on the interface that connects solid and fluid domain (since the split for parallel processors are aligned with the interface,. i.e., boundary condition on the connecting face of the mesh domain). This was not possible in FDS6. In FDS7, it works as far as i tested.

I understand that HT3D model cannot be placed on the exterior domain. Is it also not possible to place exactly at the abutting face of two mesh domains?

Thank you.

Kevin McGrattan

unread,
Feb 6, 2023, 10:41:11 AM2/6/23
to fds...@googlegroups.com
Even if I say it should work, you should still test with a simple case. This HT3D routine is new. Even the old HT3D routine was not used very often, so it is always a good idea to test on simple cases first.

Arulnambi Palani

unread,
Feb 7, 2023, 4:58:40 PM2/7/23
to fds...@googlegroups.com
Hi Kevin,

As you suggested, I have formulated a simple case to study the HT3D model. 

But, unfortunately I couldn't able to view the temperature distribution only within the solid in smokeview (I am still using the smokeview from Version 6.7.9).  I am able see the change in temperature for the air that is on the other side of the solid.

Is this because, my smokeview is not compatible with the new version (FDS 7)?

P.S: I also tried to the output quantities in tecplot using PLOT3D. Even in that, I could see only ambient temperature for the whole solid.

Thank you.


On Mon, Feb 6, 2023, 4:41 PM Kevin McGrattan <mcgr...@gmail.com> wrote:
Even if I say it should work, you should still test with a simple case. This HT3D routine is new. Even the old HT3D routine was not used very often, so it is always a good idea to test on simple cases first.

--
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.

Kevin McGrattan

unread,
Feb 7, 2023, 5:15:55 PM2/7/23
to fds...@googlegroups.com
This feature is not ready yet.
Reply all
Reply to author
Forward
0 new messages