Error : Numerical Instability with Minimum Density Warning (20 dwelling fire spread experiment)

[Rafia Rizwana]

I am working on a 20 Dwellings Burning simulation with wind velocity 6 m/s and 180 degree direction using PyroSim 2022.3.1208. There were 5 rows with 4 shelter in each row and 4 ignition source were placed in the 1st 4 shelters in the front row. The simulation would run for some time and start giving a minimum density warning. Then the simulation ran about 75s of simulation time and stopped showing a numerical instability problem. I tried to find the cause and added 3 pressure slices and 1 velocity slice. And I observe that a vortex was created in a shelter in the front row at the time of 76s (picture added). I don't know for sure if the vortex is the reason for the numerical instability but after that the simulation gave an error message and stopped. It would be great if anyone can help me find the solution of this problem.

WARNING: Minimum density,    0.111 kg/m3, clipped in Mesh 4

 Time Step:  12900, Simulation Time:     76.04 s

 Time Step:  13000, Simulation Time:     76.39 s

WARNING: Minimum density,    0.111 kg/m3, clipped in Mesh 4

 Time Step:  13100, Simulation Time:     76.64 s

 Time Step:  13200, Simulation Time:     76.66 s

 Time Step:  13300, Simulation Time:     76.71 s

 Time Step:  13400, Simulation Time:     76.87 s

 Time Step:  13434, Simulation Time:     76.95 s

 

ERROR: Numerical Instability - FDS stopped (CHID: No_Cladding(extended_bou)17_12_22_20_Shelters_Fire)

[runfds.exe] subprocess exit: 0

I have attached the fds file and a video of the results. This link contains the simulation files and the results-

https://drive.google.com/drive/folders/1h7efg0wZd58GhnssTc3CyKUZ2ozSpG9q?usp=share_link

[Rafia Rizwana]

I have already tried &CLIP CLIP_DT_RESTRICTIONS_MAX=0 but the simulation still crashes due to numerical instability.

Then I have removed the wind effect and the simulation runs without error, but the fire propagation is not sufficient to burn the 2nd row though it caught fire once.

[Kevin McGrattan]

Why have you put OPEN vents at the ground?

And then you cut little holes in the walls and apply a 6 m/s wind right on these walls. You can see in the Plot3D file that all hell is breaking loose at the floor:

[Randy McDermott]

Please DO NOT post copyrighted material on the forum.

[Rafia Rizwana]

Extremely sorry. I have deleted the message.

I had initially set the floor surface at inert, but back then it was still giving me a numerical instability error. 

This simulation has been inspired by an experimental work. The shelters are arranged and designed accordingly.

Do you suggest me to delete the open vents of XMIN and run again?

[Kevin McGrattan]

Put the ground or floor back and delete the small holes.

[Rafia Rizwana]

Thank you!

Do I need to remove all the holes representing the doors and windows or should I only remove the smaller holes representing windows?

On Mon, 19 Dec 2022, 23:13 Kevin McGrattan, <mcgr...@gmail.com> wrote:

Put the ground or floor back and delete the small holes.

--
You received this message because you are subscribed to a topic in the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/fds-smv/0j_WE1X9K9M/unsubscribe.
To unsubscribe from this group and all its topics, 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/CAAJimDGF8iNfdj5HwGg92BFPJAEtachgNYNm7UkXY4Tp5qhVgg%40mail.gmail.com.

[Kevin]

"... and delete the small holes."

[Rafia Rizwana]

Dear Mr. Kevin, I reviewed the geometry and found there are indeed small holes outside the walls. The primary design of the shelter's was made while validating another paper by R. Walls, where they used cladding material outside with holes. I have deleted the cladding in this case but forgot to delete the holes. Thanks again for your advice. I have attached the updated fds file here. The simulation is still running with minimum density warning (I have attached the log file). Can you please give me any suggestions?

On Mon, Dec 19, 2022 at 11:36 PM Kevin <mcgr...@gmail.com> wrote:

"... and delete the small holes."

--

You received this message because you are subscribed to a topic in the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/fds-smv/0j_WE1X9K9M/unsubscribe.
To unsubscribe from this group and all its topics, 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/e85d479a-b40e-47e0-8472-49c78b8cfb98n%40googlegroups.com.

[Randy McDermott]

I would first try

&CLIP MINIMUM_DENSITY=0.01/

then if that does not work

&MISC SIMULATION_MODE='LES'/

which will tighten a couple of the time step constraints.

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/CAKOmVm1pVDL_8r6ETHG6j_SLuwNLHXS0iWYjAAQvHfNvBpY3Rw%40mail.gmail.com.

[Rafia Rizwana]

Thanks a lot for your suggestion, Mr. Randy McDermott!

If I manually define minimum density, won't it violate mass conservation? I have read something like this in page 53 of FDS User's Manual. Pardon me if I have misinterpreted.

I have another question. If I increase the HRR of the burner, are there any constraints regarding how much HRR should be placed? It appeared to me that in my current simulation, the fire would not be sufficient to burn down the 2nd row of the shelter and propagate accordingly. It would be quite helpful if you could give me any suggestions. 

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/CAC6wax1CHA891_NW1Mx67_Em%2BrcTD3hYTg3TohLfJTn1qsdsCA%40mail.gmail.com.

[Randy McDermott]

Clipping does not conserve mass.  By lowering the clip limit, you are avoiding that (hopefully).

If you had a burner in your experiment, I would match the flow rate of that burner.  Whether there are other issues with the case that are preventing you from matching the experiment, it is very hard to tell without a detailed study.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/CAKOmVm1gMxiVKs_ZTvnExygTBJ5EcKyfpqvfUvDZBu7kmjP3OA%40mail.gmail.com.