HT3D error 423

[R-Godakandage]

Dear Sir,

I'm using HT3D to model heat transfer in a steel hollow column. I encounter the following error.

ERROR(423): HT3D solid must have at least one face exposed in direction 1: Mesh=14, IOR=-2, IIG=3, JJG=8, KKG=1, I=10
ERROR(423): HT3D solid must have at least one face exposed in direction 1: Mesh=24, IOR=-2, IIG=3, JJG=8, KKG=1, I=10
ERROR(423): HT3D solid must have at least one face exposed in direction 1: Mesh=28, IOR=-2, IIG=3, JJG=8, KKG=1, I=10
ERROR(423): HT3D solid must have at least one face exposed in direction 1: Mesh=15, IOR=-2, IIG=3, JJG=8, KKG=1, I=10
ERROR(423): HT3D solid must have at least one face exposed in direction 1: Mesh=13, IOR=-2, IIG=3, JJG=8, KKG=4, I=10

However, in the model, my obstruction is exposed in the negative Y direction, and it is not at the boundary of a mesh which confuse me with the error. Could you please advise me on solving this error? For your reference, I'm attaching my FDS file and a snapshot of the geometry. In this model, for the thickness of the column faces (obstructions), I have 5 mm, and when I use a thickness of 2 mm or 3 mm, it runs without errors (mesh size is 6.125 mm). I highly appreciate your comment and thanks in advance for your time.

test.txt

[Vojta-Salek]

Hello, I can try to take a look at that. Would it be possible to simplify the case while still maintaining the error? There are 30 meshes and many more VENTs.

[R-Godakandage]

Hello @Vojta-Salek, thanks in advance for your assistance; I simplified it to 5 meshes. please find the fds file and the error here. Thank you.

ERROR(423): HT3D solid must have at least one face exposed in direction 1: Mesh=5, IOR=-2, IIG=3, JJG=8, KKG=4, I=10

testR1.txt

[R-Godakandage]

@mcgratta Dear Sir, I highly appreciate your assistance in resolving this issue I encountered.

[mcgratta]

First, install the latest nightly version of FDS. When you run the case, you will get this error

ERROR(375): OBST Obstruction is VARIABLE_THICKNESS or HT3D and needs a MATL_ID. (MPI Process: 17, CHID: test)

which means that you should add a MATL_ID to your OBST lines.

[R-Godakandage]

Dear Sir, @mcgratta, I added MAT_ID s to the OBST lines and tried with the latest nightly version, but now again, I get ERROR(423). [testR2.txt](https://github.com/user-attachments/files/18821193/testR2.txt)

[mcgratta]

The THICKNESS on the SURF line makes no sense

&SURF ID='steeldummy',
      COLOR='GRAY 60',
      HEAT_TRANSFER_MODEL='LOGLAW',
      MATL_ID(1,1)='STEEL',
      MATL_MASS_FRACTION(1,1)=1.0,
      THICKNESS(1)=1.0E-5,
      HT3D=.TRUE./

When I removed it, the case started to run. When you use the HT3D model, you typically do not assign a THICKNESS unless it is meant to be some thin lining material.

[R-Godakandage]

Thanks a lot Sir @mcgratta . Yes now it is running.

However, this steeldummy SURF is used as a thin lining material to workaround to apply plasterboard insulation on only one face of the steel column. In that case is it suitable to remove that thickness?

[mcgratta]

If you want to use the HT3D model, the plasterboard should be an OBST. You should have an OBST of steel touching an OBST of plasterboard. I suggest you try this with a small domain.

[R-Godakandage]

@mcgratta Thanks a lot, Dr. Kevin. It works with the nightly version; previously, with 6.9.1, it didn't work.

However, when I checked the SMV files, the thickness of the OBSTs had snapped to the cell size(refer to the attached figure). Is this just a representation, or does FDS take this snapping into consideration for the heat transfer? (I hope not when using HT3D). Despite the modelled protection layer thickness (pink coloured) of 10 mm or 5 mm, I get similar temperature variations at locations B1 and B2, which does not make sense (refer to the attached figure).

case1 5mm plasterboard.txt
case1 10mm plasterboard.txt

[R-Godakandage]

[mcgratta]

Your input files are too large to easily diagnose a problem. I suggest that you first try a more simple case where you have only a single mesh and a short column and a small heat source, like a hot object. No fire needed. Just do simple tests. If you see problems, I can more easily help you. Also, submit the simple test case as an issue. Otherwise, this Discussion thread will simply disappear from my radar.

[Prometeo-ianic]

Hello, sorry for jumping into this thread, but I’ve had a similar situation for a while that I wanted to test with HT3D.

I added the MATL_ID line and cell size in the object. I’m also trying to run the test with the night bundle version, but I keep getting the error:

"ERROR(101): Problem with OBST number 1, line number 192 (MPI Process: 0, CHID: wheat)".

I’m attaching my FDS file—what am I doing wrong? I also had another question:

Which device should I use in this case to get
wheat.txt
the internal temperature within the 3D element?

[Prometeo-ianic]

Ok. Thank you. But now I get
ERROR(423): HT3D solid must have at least one face exposed in direction 2: Mesh=1, IOR=-1, IIG=95, JJG=1, KKG=19, I=1
wheat.txt
Which was the same error I got with the previous version of FDS.

[Vojta-Salek]

Look at HT3D Limitations in User's Guide. Point number 1.

[Prometeo-ianic]

Ok Thank you, now the case run. For the internal temperature I have some questions. If I have two layer of obstruction as in my case, when I put the inside wall temperature between these two obstructions, do I get the temperature of the object above or below?

[mcgratta]

The temperature is continuous across the interface.

[mcgratta]

I suggest you add a profile via &PROF to see the temperature profile through the solid.