Implementing PU foam 2-step reaction model?
Marisa Groenewald
properties of polyurethane foam in the library are fictional. Everybody
seems to struggle with modelling PU foam.
I found the article from Kuldeep Prasad et al. "Numerical simulation of
fire spread on polyurethane foam slabs" on the NIST site w.r.t. the 2
step reaction model, where the 1st layer is considered to be PU
releasing TDI, and the second layer polyol.
Material pyrolysis, thermophysical properties and thicknesses are
recommended. I've tried a verification simulation of the experimental in
the article, but can't get the combustion going, so obviously my
interpretation of the pyrolysis model is incorrect.
Attached is the input file I've used. I'll appreciate if someone familiar
with the model can point out my mistake in interpreting the pyrolysis
model.
By the way, I also found a slide show from session 3 of the annual fire
conference in 2009 from the same authors summarizing the result of the
article. There is however a discrepancy with those recommendations and
the one in the article. The heat of reaction for PUF and polyol are
different. The values below are form the article. The ones in the slide
show have values of about 360 kJ/kg (slight difference between the PUF
and the polyol). Which ones are correct?
Regards
Kevin
FDS and other models, and yet there is still no consistent treatment
of the physics or the means of obtaining the properties. Just because
your simulation did not produce sustained pyrolysis does not mean that
you are misinterpreting the pyrolysis model. It is more than likely
that only some of the properties of a very specific kind of foam were
measured, and the rest were estimated from the literature. Your foam
in probably different than the foams used in the papers, and there are
still numerical issues related to the model algortihms which may
explain some of the results as well.
There are no "correct" values for any of these properties. That would
imply that there is a well-characterized type of foam that everyone
works with. My organization does maintain what are known as "standard
reference materials" and apparently there is PU foam that was
originally intended to be studied for fire applications:
https://www-s.nist.gov/srmors/certificates/archive/GM21andGM22.pdf
but I am not sure whether these materials were ever used in this
regard. Even if it were, the listed properties of the foam have
nothing to do with fire, except for the density.
> Foam_lineburner.fds
> 2KViewDownload
philio
pretty well. There are alot limitations not only with the physical
modeling, but ... also a result of the complexities of the foam
itself. For instance, the actual materials can be nearly indentical
but the way the foam is blown, can result in different properties, and
these may yield different results.
However, thats where the creativity comes into play, and Kuldeep's
paper mentions some of these interesting ideas.
I also observed those values for heat of reaction in kuldeeps paper
and in the slides. Im not sure why there is a descrepancy. I find
360 kj/kj hard to believe. I also note the other values he used in
the paper were around ~1960kj/kg.
Im not sure what the real number used was.
Also, a quick note, it can be confusing stating the values for heat of
reaction, combustion. And proper context must be given when doing
so. For instance, the heat of combustion (HOC) of polyurethane foam
is 9600 kj/kg listed in his report.
This is saying the HOC is 9600 kj/kg for the foam, based on the mass
of the entire volume. However, your line in fds, may/should read
(9600/.3333) or 28800 kj/kg. This is because the 9600 was refering to
just 1/3rd of the foam, the pre polyol combustion.
The same applies for the polyol, listed as 17500 kj/kg of the foam.
Polyol constitutes 2/3rd of the mass. Therefore 17500 kj/kg should be
adjusted for your input line in fds, and made (17500/.6666) or 26276
kj/kg.
Perhaps you realized that and I havent helped at all, but I think that
may be what is preventing you from getting "combustion".
Let me know how your work is going. Im still working on solutions to
this PU two step model.
> Foam_lineburner.fds
> 2KViewDownload
Kuldeep
that provides standard reference properties of polyurethane foam.
This value was used in the simulation, but we found that FDS
simulations were quite sensitive to this value.
Subsequently we were able to get DSC measurements on the PUF that was
used in our full scale experiments. The DSC simulations (not performed
at NIST) indicated a value of 360 kJ/kg. Since this value was measured
for the foam used in the experiments, I would trust this value more
than the value obtained for a standard foam.
It should be noted that analysing DSC test data is an art as well.
While analysing these tests, a base line has to be established (not
necessarily horizontal), and the area under the curve has to be
integrated. Different researcher could draw the baseline at different
locations and obtain different numbers for heat of reaction. The
analysis becomes even more difficult when there are multiple peaks.
Other foam properties such as emissivity, absorptivity, radiation
fraction, gas phase reaction play an important role as well (these
numbers have not been measured so far).
Finally we should make sure that the results are not senstivity to
grid spacing (this can be difficult to check in a 3d simulation).
> > 2KViewDownload- Hide quoted text -
>
> - Show quoted text -