soot_yield of smoke

[xd]

Hello,

What is a typical soot_yield for a burning car. I suppose the default
value of 0.01 is to low, but I cannot find any information about that
topic. It is quite important to be able to predict the visibility. I
am simulating 2 burning cars (6 MW) in an underground car park.

thanks for the help,
Xavier

[Tom]

Even for a single material, such as a polymer, there is not one well
defined soot yield. Obviously, for a car, there is no "typical" soot
yield seeing as a car consists of thousands of parts.

As with so many other aspects of fire engineering, you have to assume
a credible worst case scenario and present an argument as to why you
are choosing the particular parameters you use in your model. If you
are using visibility as your ultimate acceptance criteria, then it is
even more important to consider a valid worst case to provide these
aspects of your modelling with a sort of safety factor.

Look at what in a car will burn -- the metal?, the seats?, the plastic
interior?, the contents of the tank? etc. What soot yields can you
expect from these materials you think will burn? Consier a reasonable
worst case and use that in your model.

Good luck.

[Kevin]

The default value for soot yield in FDS is based on the default fuel
(propane). And even then, 0.01 is just an order of magnitude estimate
because the soot yield changes with fire size and ventilation
conditions. Work done by Doug Walton, Dave Evans, and many others, on
large pool fires of gasoline, crude oil, diesel fuel, etc, suggest
soot yields on the order of 0.1 (you can search their pubs at
fire.nist.gov --> BFRL Publications On-line). I've also seen under-
ventilated fire soot yields as high as 0.2 and 0.3, but I can't recall
exactly where I've seen those estimates. In any case, 0.1 is a better
starting point. 0.01 is not appropriate.

[clauten]

Very good suggestions. Fire protection engineers applying FDS for
design purposes will recognize that visibility is almost always the
limiting tenability criterion in an ASET/RSET analysis for large
volume spaces so HRR, soot yield, and the 3 or 8 exit sign visibility
factor can have a huge impact on design. So one item to bring up for
discussion: what is meant by a "credible worst case scenario"? Is this
"worst case", or "worst case that may actually occur"? If one
considers the worst case scenario, then overly conservative designs
could result, especially since there are some indications that FDS can
under-predict visibility due to smoke plating. On balance, one usually
tries to embed a certain level of conservatism in design calculations.
I'm just curious how others balance these two concerns.

> > Xavier- Hide quoted text -
>
> - Show quoted text -

[Tom]

A soot yield of 0.1 (perhaps in the area of what you would expect from
ventilation controlled polyurethane-type fires) is indeed a better
starting point than 0.01.

A worst case scenario is not the same as a credible worst case
scenario. You could think up an infinite number of scenarios that no
fire safety measures in the world could adress, but is it reasonable
to account for all of these? These extreme scenarios are in my opinion
better adressed with a risk analysis (if this is deemed necessary),
showing if any of them are probable enough to warrant further
consideration as a design fire/explosion/other scenario.

To adress the concerns you talk about, I would typically employ some
sort of sensitivity study on the HRR (and growth rate) and soot yield
(if necessary). If a reasonable increase in either of these leads to
considerably worse results in the modelling, they might need to be
looked at closer in some way.

Even for less extreme scenarios, a probability study could be
employed. What is the probability that the sprinkler system fails
completely at the same time as the smoke exhaust system for example..
and so on in a probability tree.

[Chris]

&REAC ID='KEROSENE'
SOOT_YIELD=0.042 /

&REAC ID='WOOD'
SOOT_YIELD = 0.01/

&REAC ID='POLYURETHANE'
SOOT_YIELD = 0.10/

In my opinion a soot yield of 0.1 for a car is really high. And a car
has maybe not a HRR of 3 MW this was discussed something like 2 month
ago. The paper is somewhere on the net.

Maybe I'm also wrong. Just an input.
Chris

> > > - Show quoted text -- Zitierten Text ausblenden -
>
> - Zitierten Text anzeigen -

[Chris]

http://www.vtt.fi/inf/pdf/publications/2004/P521.pdf

[Kevin]

I wouldn't put any faith in those soot yields. They were probably
pulled from handbooks.

A burning car is a flashed over fire. The interior is mostly plastic
of one kind or another. I don't think any of us can definitely state
what the soot yield is going to be unless we measure it, and that's
not easy.

> > - Zitierten Text anzeigen -- Hide quoted text -

[Chris]

Yes, that's definitely true.

[Hostikka Simo]

Just talked with Johan about that today.
That publication does not really contain soot yields.
Some soot quantity was measured but the yields have never been
calculated out of it.

Simo

> -----Original Message-----
> From: fds...@googlegroups.com
> [mailto:fds...@googlegroups.com] On Behalf Of Chris
> Sent: Friday, February 15, 2008 4:55 PM
> To: FDS and Smokeview Discussions
> Subject: [fds-smv post:2680] Re: soot_yield of smoke
>
>
> http://www.vtt.fi/inf/pdf/publications/2004/P521.pdf
>
>
> >
>

[Forest]

Yes , I feel that it is a difficult thing to set a right soot_yield
for my fuel too. But the soot_yield has something to do with the
visibility. I know that the defautl visibility value is 30 m in FDS.
But how much visibility is safe or how much is dangerous? In my a
case, the visibility value was 2. I don't konw whether it is a
dangerous visibility. Thanks .

Forest

[clauten]

I think 30 m is just the maximum value on the Smokeview color bar --
if there is no smoke then the visibility would be much greater than 30
m, but Smokeview truncates the scale at 30 m.

There is no one answer as to what is a "dangerous" visibility level.
10 m is visibility level that gets used a lot as a tenability
criterion, at least by US consultancies, but this probably varies by
country and the AHJ may have some input on this. But if you have 2 m
visibility, that's probably "dangerous".

Take a look in the SFPE handbook for a discussion of visibility and
tenability. There's also some journal papers that discuss this. One
widely cited paper is Hadjisophocleous, G.V., Benichou, N., and Tamin,
A.S., "Literature review of performance-based fire codes and design
environment," Journal of Fire Protection Engineering 9: 12-40 1998.

[dr_jfloyd]

Table 13.19

MAXIMUM_VISIBILTY Default value 30 m.
Since visibility is proportional to 1/rho_soot, to prevent divide by
zero we set a minimum value for use in the visiblity calculation. For
most interior spaces 30 m is reasonable ceiling. Since smokeview sets
its contour limits based on the maximum and minumum values, using a
large default for VISIBILITY would result in color contours that were
not meaningful at the scales people are typically interested in (e.g.
like the 2 m you refer to).

> > Forest- Hide quoted text -

[Glenn Forney]

fyi: smokeview would have gotten the 30 m from somewhere, ie it does
not have any internal special case min or max bounds. So 30 m is either
1) a data value in an .ini file, 2) the global max of the data it read
in or 3) a 99'th percentail data value.
glenn

--
Glenn Forney
National Institute of Standards and Technology
100 Bureau Drive, Stop 8663
Gaithersburg MD 20899-8663

Telephone: (301) 975 2313
FAX: (301) 975 4052

Pre-decisional and sensitive information. Not for attribution, distribution, or reproduction.

[Kevin]

No -- 30 m comes from the fact that we cannot have infinite visibility
to start. The visibility calculation requires taking the reciprocal of
a number which can't be zero. Thus, there is some small number
assigned which leads to a visibility of 30 m. You can change this by
setting MAXIMUM_VISIBILITY on the REAC line.

On Mar 5, 10:42 am, clauten <chris.lautenber...@gmail.com> wrote:

[clauten]

Got it -- FDS, not Smokeview, sets the 30 m upper limit (or whatever
the user species for MAXIMUM_VISIBILITY).

read.f90:
EC_LL = VISIBILITY_FACTOR/MAXIMUM_VISIBILITY

(EC_LL must be extinction coefficient lower limit?)

dump.f90:
GAS_PHASE_OUTPUT = VISIBILITY_FACTOR/MAX(EC_LL,EXT_COEF)

where EXT_COEF is the extinction coefficient, so EC_LL prevents divide
by zero and ensures the upper limit of visibility is equal to
MAXIMUM_VISIBILITY.

> > - Show quoted text -- Hide quoted text -

[Kevin]

Good sleuthing -- you've figured out my cryptic variable naming
conventions. You can tell that some of these were conceived of before
Jason converted the code to 132 character line length Fortran 90.

[Forest]

In my case, there is a room of 84mx18m. In fact , one may see 84m
longth distance in day. May I set MAXIMUM_VISIBILTIY=84?
If the visibilty of somepoint by FDS is more samll than 10m , may I
think it is dangerous? I don't use color coutours ,I need konw some
important points visibilty by &DEVC only.

[daserra]

In a real life car fire in an underground car park there will be a lot
of smoke due to the lack of oxygene. The parts of the car that burn
are the fabric and the rubber part. There are tenths of kilograms of
it but only a small amount will burn. So the firefighter are trained
to go in and extinguish the fire first by foam and then ventilate the
smoke out. A large amount of the material is converted into soot.
The whole thing is completely different from a car burning in the
open.

[Forest]

Yes , different fires have different soot denisity, that means
different visibiltiy. So I think that users should set different
Maximum_visibility according different fire character. If one uses
the default visibility(30m), then one will see many smaller
visibility(<10m), it seems like not to accord with the fact. If one
set a right maximu_visibility, one will see a reasonable visibility.
Do you agree my opinion?

> > Xavier- Hide quoted text -

[Kevin]

I do not think you understand what MAXIMUM_VISIBILITY means. It is an
arbitrary distance because FDS and Smokeview cannot work with, or
display, a visibility of infinity. It is assumed in FDS that air is
21% O2, 79% N2, and nothing else. It also assumes that these two gases
are completely transparent to light and thermal radiation. This means
that FDS assume infinite visibility by default at the start of the
calculation, but we cannot plot this in Smokeview. So we just chose 30
m arbitrarily as a cut-off. Please read about all of this more
carefully. Everything that we assume about visibility comes right out
of the SFPE Handbook. FDS-specific implementation is described in the
User's Guide.

> > - Show quoted text -- Hide quoted text -

[Forest]

I think that I don't understand how to set MAXIMUM_VISIBILITY indeed.
Because there are no information in user's guide except Table
13.19.According the guide(12.3.2), there are 3 paramters to control
it. These paramters are SOOT_YIELD, MASS_EXTINCTION_COEFFICIENT and
VISIBILITY_FACTOR in &REAC line. But if I change MAXIMUM_VISIBILITY
only(others are default) , I get different visibilty too. Could you
tell me how to set MAXIMUM_VISIBILITY or to use the default only?
Thanks.

[dr_jfloyd]

In the user's guide we define:

VISIBILITY = VISIBILITY_FACTOR / (MASS_EXTINCTION_COEFFICIENT *
RHO_SOOT)

If RHO_SOOT is zero, then VISIBILITY is infinite; however, doing this
would get a division by zero error. So the actual line of code for
visibility is (seen in clauten post earlier in this thread):

VISIBILITY = VISIBILITY_FACTOR / MAX(EC_LL,MASS_EXTINCTION_COEFFICIENT
* RHO_SOOT)
where EC_LL = MAXIMUM_VISIBILTY/VISIBILITY_FACTOR

MAXIMUM_VISIBILITY only changes VISIBILITY if the MAX statement winds
up using the EC_LL rather than the MASS_EXTINCTION_COEFFICIENT *
RHO_SOOT

Smokeview; however, will set the color scale (as Glenn indicated
earlier in this thread) based on the maximum and minumum values for
the the slice file. Changing MAXIMUM_VISIBILITY, changes the maximum
value in the slice file, and will result in a different color scale.
Which will effect the way the slice file looks when displayed.

[Fredrik (Wuz)]

You must stop focus on "maximum visibility" and "color bars" in
Smokeview. Note that Smokeview only illustrates FDS data based on your
settings. In order to assess safe egress time you must first define a
credible fire scenario and choose appriopriate inputs. Then you
perform your simulation and finally you evaluate the outcome.

E.g. if your building code requires a miniumum visibility of 10 m
during the time needed for escape, the you could set the min and max
bounds in Smokeview for visibility in the intervall of 9.9-10.1.
Smokeview will then show areas with visibility greater than 10.1 in
one color and visibility less than 9.9 in another color. Doing so
helps to evaluate areas where the conditions are untenable.