Re: [fds-smv] FDS in a fire investigation

499 views
Skip to first unread message
Message has been deleted

Randy McDermott

unread,
Apr 6, 2017, 7:21:00 AM4/6/17
to FDS and Smokeview Discussions
I would question whether superposition of HRR would work for an under-ventilated fire.  Another approach could be the IGNITION_TEMPERATURE method like was used in the World Trade Center fire investigation---obviously a more expensive approach.  But you might want to see if there are any drastically different qualitative trends using that method.

On Thu, Apr 6, 2017 at 4:54 AM, Bojan <bojan...@gmail.com> wrote:

lo,

I am trying to use FDS to run a simulation of a fire incident which happened in one night club. It is not an official investigation, just my attempt to find out more and see if FDS could help me doing that. The idea was to recreate the fire incident and discover how much time people could live under its conditions. Some of you already helped me with the FDS and Evac problem I had, and I am really grateful for it.

The geometry is more or less like atrium (area 23 m x 13 m and height 7 m, the geometry in the attachment). Six people lost their lives, 5 of them in the upper toilet, 1 person died in front of the lower toilet. All of them had no burning injuries. The victims tried to hide in the toilets, while waiting for the firefighters, who arrived too late (45 minutes after the fire broke out). There are two levels (ground level and upper level) connected with two narrow stairs. Main entrance has area of 2 x 2 m, and the roof opening is 0,7 x 7 m large. At the moment of the incident, there were around 350 people in the night club (200 at the ground level, additional 150 at the upper level). The only way to get out was the main entrance I described before. In order to reach the main entrance, people have to go through the hall, which is 10 meters long. Evacuation time that I get with Evac is around 5 minutes.

I got the photos after the incident which helped me to  estimate the fuel that was in the night club. Taking the worst case, I ended up with 13 MW peak value of HRR. I also developed one more reasonable HRR curve with peak value of 8 MW. The principle I used was the superposition method - it includes combining HRR curves from different furniture burned in a fire lab, and combining them in order to form the main HRR curve. The time between start of burning of different fuel parts I got using radiation formula. I chose CO yield from SFPE Handbook to be 0,08. Then I compared the values I got from FDS with the Purser's experiments in SFPE Hanbook, and got the information that people got unconscious after 10 minutes, and died after additional 20 minutes due to high concentration of CO in the smoke. The protection of the toilet's door I neglected, since the victims opened the door trying to escape, which allowed smoke to come in. The CO concentration is lower than I expected, since the roof opening and the main entrance were immediately opened, allowing relatively good ventilation conditions.

Do you find my approach reasonable?

Thanks.







--
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+unsubscribe@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/d448fcc2-65e5-4a54-b8e7-bd881c6e8a24%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

fde

unread,
Apr 6, 2017, 7:22:44 AM4/6/17
to FDS and Smokeview Discussions
An additional comment on toxicity. HCN is a common product of combustion of sound insulation material which can be found in high amounts in especially nightclubs.  The reason I point out is HCN is 35 times more toxic than CO. You might want to reconsider the ASET calculation taking into account of HCN. 

An example article:


Goodluck on your investigation!
Message has been deleted

Randy McDermott

unread,
Apr 6, 2017, 10:50:06 AM4/6/17
to FDS and Smokeview Discussions
Is the CO yield appropriate for a well-ventilated fire?

On Thu, Apr 6, 2017 at 9:57 AM, blackarrow90 <bojan...@gmail.com> wrote:
The first assumption I had was that the fire was well-ventilated. I ran few simulations, checked the HRR values and oxygen concentrations in FDS, and concluded that I do not have problem with the oxygen (probably due to the dimensions of atrium and openings which were present, as well as fuel spread). That is why I used the experimental HRR curves for different furnitures, which were developed in the fire lab under well-ventilated conditions.

I am aware of the potential of HCN, however I was not able to find value of HCN yield I could use in the simulation. There are not so many experiments in which HCN was measured.

Correct me if I am wrong with both answers I gave above.


On Thursday, April 6, 2017 at 10:54:48 AM UTC+2, blackarrow90 wrote:

lo,

I am trying to use FDS to run a simulation of a fire incident which happened in one night club. It is not an official investigation, just my attempt to find out more and see if FDS could help me doing that. The idea was to recreate the fire incident and discover how much time people could live under its conditions. Some of you already helped me with the FDS and Evac problem I had, and I am really grateful for it.

The geometry is more or less like atrium (area 23 m x 13 m and height 7 m, the geometry in the attachment). Six people lost their lives, 5 of them in the upper toilet, 1 person died in front of the lower toilet. All of them had no burning injuries. The victims tried to hide in the toilets, while waiting for the firefighters, who arrived too late (45 minutes after the fire broke out). There are two levels (ground level and upper level) connected with two narrow stairs. Main entrance has area of 2 x 2 m, and the roof opening is 0,7 x 7 m large. At the moment of the incident, there were around 350 people in the night club (200 at the ground level, additional 150 at the upper level). The only way to get out was the main entrance I described before. In order to reach the main entrance, people have to go through the hall, which is 10 meters long. Evacuation time that I get with Evac is around 5 minutes.

I got the photos after the incident which helped me to  estimate the fuel that was in the night club. Taking the worst case, I ended up with 13 MW peak value of HRR. I also developed one more reasonable HRR curve with peak value of 8 MW. The principle I used was the superposition method - it includes combining HRR curves from different furniture burned in a fire lab, and combining them in order to form the main HRR curve. The time between start of burning of different fuel parts I got using radiation formula. I chose CO yield from SFPE Handbook to be 0,08. Then I compared the values I got from FDS with the Purser's experiments in SFPE Hanbook, and got the information that people got unconscious after 10 minutes, and died after additional 20 minutes due to high concentration of CO in the smoke. The protection of the toilet's door I neglected, since the victims opened the door trying to escape, which allowed smoke to come in. The CO concentration is lower than I expected, since the roof opening and the main entrance were immediately opened, allowing relatively good ventilation conditions.

Do you find my approach reasonable?

Thanks.







--
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+unsubscribe@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
Message has been deleted
Message has been deleted

Randy McDermott

unread,
Apr 27, 2017, 6:12:09 AM4/27/17
to FDS and Smokeview Discussions
I don't have a better reference for CO.  I just wanted to make sure you had considered the question.  The yields will be dramatically different for over- vs. under-ventilated fires.  So if the fire every becomes under-ventilated, those are the conservative numbers I would use.

It is possible to predict HCN, but at present not with simple chemistry.  At present you would have to do your own stoichiometry.  Perhaps that is something we should add to the code.

@Jason, do you see any downside to having a Y_HCN for simple chemistry?


On Thu, Apr 27, 2017 at 5:04 AM, blackarrow90 <bojan...@gmail.com> wrote:
What would you take as reasonable value of CO yield for this type of fire? I found some values in Babrauskas' work here: http://fire.nist.gov/bfrlpubs/fire92/PDF/f92024.pdf The values taken from real-scale test are presented on the page 165, in the last row.

Another question: Since it is not possible to simply put in HCN yield in the code, how I can implement it in my simulation? I looked up FDS User's Guide but I did not manage to measure the HCN concentration. Thanks.
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.

--
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+unsubscribe@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.

dr_jfloyd

unread,
Apr 27, 2017, 7:12:45 AM4/27/17
to FDS and Smokeview Discussions
So if we add Y_HCN, why not Y_HCL, Y_N2O, Y_NO2, ...?  One of the reasons for doing lumped species was rather than adding a multitude of Y_X inputs for all the various toxic products people might be interested in to just give the ability for users to define what ever products they wanted while still being able to use the single-step, fast chemistry, Fuel + Air -> Products.    

Randy McDermott

unread,
Apr 27, 2017, 7:34:24 AM4/27/17
to FDS and Smokeview Discussions
I guess the question is one of how often we users need one particular Y_X over the others.  Clearly, Y_CO and Y_SOOT are important.  My question for the users is, given FED or other practical output considerations, has Y_HCN been elevated to the point of needing special attention?

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

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

dr_jfloyd

unread,
Apr 27, 2017, 9:07:51 AM4/27/17
to FDS and Smokeview Discussions
The simple chemistry input is targeted to well ventilated fires. There you do not make significant quantities of HCN. The N tends to go to N2.

I would really rather not keep adding yields to simple chemistry. This is why we provided the lumped species ability. There is an example in the guide using HCl on how to specify other species.

Kevin

unread,
Apr 27, 2017, 9:34:59 AM4/27/17
to FDS and Smokeview Discussions
Here's another idea for trace species. If you know the yield of, say HCN, just track CO and calculate the concentration of HCN as a fixed fraction of CO. In the simple chemistry model, the yields of all products are fixed, and these products are transported together. The slight change in the stoichiometry of the single combustion reaction due to HCN is negligible. In short, just use CO as the surrogate for all your trace species. If you are interested in FED, just calculate it after the simulation is done.
Reply all
Reply to author
Forward
0 new messages