Localized leakage function over large obstructions

204 views
Skip to first unread message

Ronnie T

unread,
Aug 30, 2018, 5:25:08 PM8/30/18
to FDS and Smokeview Discussions
Hi,

I'm trying to compare both the ZONE and LEAK functions to see what is appropriate for a larger model. I've consolidated the model to a 3-floor example where the fire is on the ground level, and expect to see smoke moving upwards through the model. However, the smoke appears to be contained within the compartment of origin. Perhaps I'm missing something simple in this iteration?

I have to say after a few iterations, I found that a smaller grid size (0.25 vs. 0.5) was helpful for numerical instability, as was relaxing the velocity tolerance. I could not run this before my 10th iteration. I have also noticed sometimes I have to up the pressure iterations to 20 or higher. 

Also, I ran a single-mesh model but that fails (numerical instability) after 40s, or about 10s after the fire gets going. Further, in previous iterations, I did have to wait until about 20s to run the fire, as the model would not complete the first timestep otherwise. 

Thank you,

Ronnie
Leakage_14.fds

dr_jfloyd

unread,
Aug 30, 2018, 6:37:15 PM8/30/18
to FDS and Smokeview Discussions
You should define a ZONE for each room.  The localized leakage model isn't really intended for use over such large areas. It is meant to represent something like a crack in a door not diffuse leakage over the entire surface of a room. You can try adding HVAC_LOCAL_PRESSURE=.FALSE. to the MISC line.   

Kevin

unread,
Aug 31, 2018, 1:49:26 PM8/31/18
to FDS and Smokeview Discussions
I also suggest you create a very simple test case that runs quickly. Two compartments and a few HVAC lines. Experiment until you get things the way you want, or submit the simplified case.

Ronnie T

unread,
Sep 7, 2018, 7:30:25 PM9/7/18
to FDS and Smokeview Discussions
Hi,

I've taken both of your considerations and put them in the model. The ZONES were defined tightly to allow the vents to be the predominate leakage path. I've defined those loosely as 1m2. This is about as simplified as I could make the model. I've noticed that adding HVAC_LOCAL_PRESSURE=.FALSE. to the MISC line greatly increases the computational expense of the model, at least initially. 

What I don't understand is why the VENTs do not appear to allow airflow in either direction. The fire self-extinguishes around 700-800s, and no transport of smoke is seen from the zone of origin. Perhaps I'm still missing a key component to the input file?

Thank you both!
Leakage_18.fds

Ronnie Thomas

unread,
Sep 7, 2018, 8:27:44 PM9/7/18
to fds...@googlegroups.com
I'd like to update this. As soon as I removed the IOR assigned to the vents, it worked as expected for the localized leakage model. 

--
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/ceYZxcp1dOw/unsubscribe.
To unsubscribe from this group and all its topics, 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/73df51f2-f6fc-4e40-ac72-7ed2c76b0560%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

fde

unread,
Jan 12, 2023, 10:28:39 AM1/12/23
to FDS and Smokeview Discussions
Recently I have encountered an issue (FDS v6.7.9). I am not going to report the case because it is a very large model and I could not simplify enough the model to report, nevertheless I am interested in what else I could try to solve it. I would appreciate if you could give a few ideas.

The geometry is: a few adjacent lobbies. There are doors in between and each door has local leakage VENTs.  The lobby on one end has a OPEN vent with -15 Pa dynamic pressure. The other end lobby has a VENT which is OPEN to 0 Pa (default) and there is a fire here. The issue happens in the lobbies in between. Their pressure drops below -200 Pa. 

I have tried running with these settings:

&PRES MAX_PRESSURE_ITERATIONS =50, PRESSURE_TOLERANCE=1, VELOCITY_TOLERANCE=0.0025/

&MISC  HVAC_PRES_RELAX=0.9/

Neither helped. 

Then  I set HVAC_LOCAL_PRESSURE=.FALSE. and it was solved. I will run the large model with this setting tonight once again to make sure it actually helped and the results are realistic. 

I will report back. 

Ronnie Thomas

unread,
Jan 12, 2023, 10:46:35 AM1/12/23
to fds...@googlegroups.com
I'll check my model later today and let you know if that setting worked. I will report that my models did work out well and the flows were reasonable vs. comparison to CONTAM.

--
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/ceYZxcp1dOw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to fds-smv+u...@googlegroups.com.

dr_jfloyd

unread,
Jan 12, 2023, 11:08:10 AM1/12/23
to FDS and Smokeview Discussions
Is this a model where you have taken a CAD file into Pyrosim so you have a lot of pressure zones due to void spaces? If that is the case, Kevin has been working on that issue and it might  be worth trying with the latest build.

fde

unread,
Jan 12, 2023, 1:06:40 PM1/12/23
to FDS and Smokeview Discussions
Yes, I used a CAD model and converted into obstructions. I will check once again if there is any voids. 

I can't try the latest build since I can't compile it. I will wait for FDS 7 beta to try the same case. 

Thank you for your response. 

Kevin

unread,
Jan 12, 2023, 1:17:26 PM1/12/23
to FDS and Smokeview Discussions
The latest bundles (as of last night) are available here: https://drive.google.com/drive/folders/1X-gRYGPGtcewgnNiNBuho3U8zDFVqFsC

fde

unread,
Jan 12, 2023, 1:22:18 PM1/12/23
to FDS and Smokeview Discussions
Thank you, I will try this one tomorrow. 
Message has been deleted

fde

unread,
Jan 12, 2023, 1:46:50 PM1/12/23
to FDS and Smokeview Discussions
Update on the running case (FS 6.7.9 latest official version) : The run with HVAC_LOCAL_PRESSURE=.FALSE. gives odd results (the leak vents (e.g. below) do not work) probably because I did not assign ZONEs. I will read the User Guide later. 

&VENT ID='node1', SURF_ID='HVAC', XB=57.3,57.3,25.7,25.8,0.1,1.2,/
&VENT ID='node2', SURF_ID='HVAC', XB=57.4,57.4,25.7,25.8,0.1,1.2,/
&HVAC ID='Door', TYPE_ID='LEAK', VENT_ID='node1', VENT2_ID='node2', AREA=1.4717E-4, LEAK_ENTHALPY=.TRUE./


I better should try the latest bundle (link above) first. 

dr_jfloyd

unread,
Jan 12, 2023, 2:14:45 PM1/12/23
to FDS and Smokeview Discussions
With LOCAL_PRESSURE=.FALSE. you only use the ZONE pressure for driving flow. So if both VENTs wind up in the same pressure ZONE (automatically assigned or otherwise), there will be no flow. That may be what you are seeing.

If the new build with keeping HVAC_LOCAL_PRESSURE=T doesn't solve the problem, I suggest opening an issue. 


fde

unread,
Jan 13, 2023, 5:09:17 AM1/13/23
to FDS and Smokeview Discussions
Update:

Initialy I had problems with FDS v6.7.9 official build. Issue was present with both default pressure settings and the tight tolerances (&PRES MAX_PRESSURE_ITERATIONS =50, PRESSURE_TOLERANCE=1, VELOCITY_TOLERANCE=0.0025/)
Then I run two cases with FDS v6.7.9 nightly build (link given above). The run with the same tight tolerances (abovementioned) crashed due to number of iterations being exceeded 50 (official FDS required around 9 with those settings). When I run it with with default pressure settings (no &PRES line) the pressures were okay. 

So the nightly build FDS works fine with the leakages without any pressure settings. 

fde

unread,
Jan 13, 2023, 5:16:21 AM1/13/23
to FDS and Smokeview Discussions
Although I am happy to see a working solution, I also look forward to receive the official release of this version. In order to make use of this void-fixing feature, will we wait for FDS 7 or will there be an official new version of FDS 6?

dr_jfloyd

unread,
Jan 13, 2023, 7:44:47 AM1/13/23
to FDS and Smokeview Discussions
I expect that there will be a release in the first quarter of this year, and hopefully that is a Beta testing version of FDS 7. 

Ronnie T

unread,
Jan 24, 2023, 4:32:34 PM1/24/23
to FDS and Smokeview Discussions
Hi,

Glad to hear there was a working solution found and luckily it was as simple as getting rid of pesky settings. I'm unsure why the models were not first tried with the default settings. I believe my original post indicated things were working fine without messing with the aforementioned settings. 

As to the model size, mine still works (recently revised last month) for an 8-story high rise building with stair pressurization. Not sure if that qualifies a big enough, but certainly proves to me that FDS is going the right direction with regards to how to work with the leakage function. And of course using it for large buildings, although CONTAM is still my program of choice for many reasons. Mostly time and expense.  

If the models are working without messing with specific functions, why would one state there is an issue? I have read and re-read the user guide about 100X in my career. When I first built these models, I had not read carefully enough to catch the need to define zones. Then, if you read the thread, everything worked fine after that. Unsure why the new post used specific settings, but like most software, you need to relax, not constrain, the models and simplify, simplify, simplify. 

The best part of figuring this out a few years ago is that it set precedence in my area for how to set up FDS models regarding passive smoke protection. The results were quite comparable to CONTAM results. 

Glad to hear there is a new Beta version coming out. Can't wait to try the old model and see how it works out. 

Thanks for all the help from the NIST/JH folks, you know we all appreciate your efforts and admire your handling of these posts. 

Ronnie

fde

unread,
Jan 26, 2023, 4:39:13 PM1/26/23
to FDS and Smokeview Discussions
Hi Ronnie,

The model with issue was tried with the default settings in the current version at first. Then it was run with tight tolerances since this is a practice which was recommended by developers in many threads in this forum. Secondly, I have observed in certain smoke spread studies due to high pressure fluctuations in FDS the smoke moves rapidly in and out of compartments through the doors. Although the average pressure level is correct, the visual representation is not realistic and results to wrong temporary outputs. Therefore I am in favor of avoiding these fluctuations by tightening the pressure settings. 

With this practice in mind, I initially run the nightly build FDS version with the tight pressure tolerances since they will result more accurate (lower velocity error) results. It did not run however it allowed me to have an observation; nightly build required more iterations compared to the current version FDS. Then I tried it with default settings. Luckily that was sufficient to run it successfully. 

I hope that explains my setting motives.

Ronnie Thomas

unread,
Jan 26, 2023, 8:09:03 PM1/26/23
to fds...@googlegroups.com
Ah, ok that does make sense. I did not have much information regarding your flows, were they strictly induced by the fire or did you have any HVAC elements?

--
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/ceYZxcp1dOw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to fds-smv+u...@googlegroups.com.

Yuval Hirsch

unread,
Jan 27, 2023, 12:49:23 PM1/27/23
to fds...@googlegroups.com
I also had some issues with the same aversion 
I took a decision to go back to version FDS v6.7.5. Ever since all is working perfectly

Yuval

--
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/0f4a3cfe-e425-4c97-9361-72fd53e3c3f0n%40googlegroups.com.

dr_jfloyd

unread,
Jan 27, 2023, 1:39:32 PM1/27/23
to FDS and Smokeview Discussions
Did you report your issues? If issues aren't reported they cannot be fixed.

fde

unread,
Jan 27, 2023, 4:09:34 PM1/27/23
to FDS and Smokeview Discussions
Yes, it was without HVAC and was mainly influenced by a fire in vicinity. 

Ronnie Thomas

unread,
Jan 27, 2023, 4:47:00 PM1/27/23
to fds...@googlegroups.com
We are using 6.7.9 without issues. I have all sorts of HVAC, leakage and other things going on in the model. Again this is an 8 story high rise but with the caveat that I have high rise floor sizes, probably in the order of 18,000sf (1,670sq m). So there is plenty of room and leakage for the gases to expand. 

Also, I'm curious if you have seen any extinguishment of the fire? What does your output look like?

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/ceYZxcp1dOw/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/CAPDcb2iWzTLRsNpsGw0uTedukuA6-M%2B%2B1Atzfzw8Sm_bQiXptg%40mail.gmail.com.

fde

unread,
Jan 29, 2023, 4:08:27 PM1/29/23
to FDS and Smokeview Discussions
HRR reaches the prescribed values. So there is not a fully extinguishment. 
After seeing the nightly build works while 6.7.9 does not, I did not debug the model. It is likely the source of the issue is these small voids. I will leave it at that for the time being without investing more time on it. If I catch some odd occurance I will report it as always. 

Reply all
Reply to author
Forward
0 new messages