material properties database
Matti
I have a short question to you.
Is there a database where I can find very fast material properties?
I am thinking about a database where I can tip my material in and I
will get a list of properties, so that I not spend so mutch time to
search the properties in different lists.
greetings
Matti
Kevin
that contained miscellaneous properties of materials. However, it
seems that the list has disappeared from the "Pages" of this
Discussion Group. I'll check on it.
Bryan Klein
http://groups.google.com/group/fds-smv/browse_thread/thread/d03483acf36b675d/9c55191617ea245b
I started to create a page in May, but decided that our discussion
group is probably not the best place to put it.
-Bryan
--
Bryan Klein
bryan...@nist.gov
National Institute of Standards and Technology
100 Bureau Drive, Mail Stop 8663
Gaithersburg MD 20899
Phone: 301 975 5171
Fax: 301 975 4052
Matti
Should I create a database?
Do you have more properties then in the Material.txt file?
Matti
On 10 Dez., 14:42, "Bryan Klein" <bryanwkl...@gmail.com> wrote:
> It was not a page, only a discussion thread.http://groups.google.com/group/fds-smv/browse_thread/thread/d03483acf...
> I started to create a page in May, but decided that our discussion
> group is probably not the best place to put it.
>
> -Bryan
>
>
>
>
>
>
> > There is no such database. However, there was a list of references
> > that contained miscellaneous properties of materials. However, it
> > seems that the list has disappeared from the "Pages" of this
> > Discussion Group. I'll check on it.
>
> > On Dec 10, 6:57 am, Matti <matthias.zaehrin...@googlemail.com> wrote:
> > > Hello,
>
> > > I have a short question to you.
>
> > > Is there a database where I can find very fast material properties?
>
> > > I am thinking about a database where I can tip my material in and I
> > > will get a list of properties, so that I not spend so mutch time to
> > > search the properties in different lists.
>
> > > greetings
> > > Matti
>
> --
> Bryan Klein
jaynam
for example, I need to know the material properties for Cardboard box
at the moment.
Is there any site i can go to and find the properties?
DavidShep
document the test method that was used to develop the information. At
this time NIST has not provided much guidance about which test methods
should be used to develop the properties. Nor have they provided
guidance about the degree of accuracy that can be expected from the
model depending on which test method was used. I am sure that this
type of research will be done in the future, and since there are
multiple test methods available to measure many of the properties, it
will be essential for your database to contain the test method
information.
Kevin
this nice software that you use for free. How about if you convince
your government or domestic research community to take on the task of
developing a materials database. This would be an excellent
demonstration of international cooperation. What do you say?
I have said in numerous posts and papers that the materials database
in FDS 1 through 4 was a BIG mistake. Why? Because it killed research
in this area. It is amazing how easily people will use ridiculous
values of "cardboard", "upholstery", and "wood", as if these are pure
materials with very well known properties. The original idea of the
database was just to provide a sample of how to input materials
properties. It was never meant to be a comprehensive, authoritative
list.
Kevin
to focus more effort on the development of standard procedures for
testing materials and gathering input data for fire models,
specifically FDS. This effort has failed, for various reasons. This is
why I am trying to develop collaborative relationships with other
institutions.
Glenn Forney
have tried to build and run a 64 bit version of FDS5 on a PC running
Windows x64 .
thanks
--
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.
Hostikka Simo
Fortran.
The executable has been used on 64bit Windows XP.
In some individual case, it was a bit faster and used a little bit less
memory
than running the same case with 32bit code under 64bit windows.
Simo
Glenn Forney
interested in finding out whether cases can be run using more memory
than could have been run under 32bit windows. I'm not as concerned
about the run times as long as it not significantly slower.
Bak, Daniel
database for FDS use.
I am glad at the same time that you are aware that some people think
that all woods are the same and use them inappropriately also.
I fully agree that more cooperation from other countries and research
institutions would help in the objective of developing a database.
However, in the meantime, I would love to see a lean database of key
materials. Those materials would be materials that we be best
documented and least questionable. This database would be also a
"guide" database from which users would, at their discretion and risk,
extrapolate or interpolate to "mimic" behaviors of materials they would
be interested in. Naturally, an even more glaring disclaimer would need
to preface the document, as well as locating a reminder disclaimer on
every page as a header or a footer.
I know my suggestion is not the ideal database that everyone wants to
see, but it is a compromise.
Thanks for your consideration.
Daniel
-----Original Message-----
From: fds...@googlegroups.com [mailto:fds...@googlegroups.com] On
Kevin
wood is about 15,000 kJ/kg, and then say, but this is just a guess, or
this is not validated, or whatever, do you really think that this
number will not be used, and I will not be used as a reference, and by
inference my Institute?
At best an institution like NIST is going to develop test methods. It
is less likely to do the standard tests. What I am trying to convince
my colleagues in the Fire Research Division at NIST to do is develop
standard test methods, and the appropriate mathematical interpretation
of these methods.
A database is pointless without a methodology to obtain the data. We
need the methodology, but everyone keeps screaming for a database, as
if there is some clear way to take a chunk of something, send it to a
lab, and get data for the database.
On Dec 12, 11:37 am, "Bak, Daniel" <Daniel....@dmjmhn.aecom.com>
wrote:
>
> - Show quoted text -
Bryan Klein
I have the Dual Quad-Core Mac Pro with 16GB of RAM, this is a 64bit platform.
After I compile a working 64 bit version of FDS (including the MPI
version), I will run the scaling test example setting the number of
meshes and grid resolutions to exceed the 4GB threshold of 32bit
systems. So far I have not seen a significant speedup in the
processing speed. 3GHz is 3GHz no matter the number of bits, in fact I
have a feeling that there is a little bit of cache overhead in the
chip with 64 bit FDS that makes it slightly slower than the 32 bit
version I am running. I am also building the 64bit version with
multiple code paths depending on the chipset (SSE, SSE2, etc.) This
could also create some slight overhead than a version compiled for my
machine only.
All this to say, I am working on it here too and will report back my
findings. Now to get MPICH2 to build in 64 bit in Leopard. Ugh...
no joy so far. It is not all smug and quirky Mac vs. PC comments in
OSXland today... Come on kitty-cat, don't fail me now!
-Bryan
--
Bryan Klein
bryan...@nist.gov
National Institute of Standards and Technology
dr_jfloyd
steel alloys and they have heat capacities and conductivities that can
vary by 10 % or more. Depending on your application perhaps any old
steel will be reasonable if say it is just serving as a large area
heat sink, but if you plan use the steel temperatute for a structural
calc then you really should know something about the steel you are
building with. If we were to provide a generic steel definition, the
fact is that few user's would make any attempt to determine if that
definition was appropriate for their use. That is exactly what
happened with the prior database. But it was also worse than that.
Many times I have seen inputs where if a key material for a simulation
wasn't in the database, then the user just picked (seeming at random)
any old material rather than attempting to locate properties
themselves.
Dragonit
ย
Glenn Forney
simulated a case that requires more than 3GB of memory (as reported by
the task manager).
Dragonit wrote:
> We are currently experimenting with a 64bit environment -
> FDS5(compiled by Poretti Martino ) Parallel Version.
> plus the MPICH 64 bit and a Win 2003 Server. on a Dual Quad
> processors with 32GB of RAM
>
> will be happy to test benchmark cases and will report tests result as
> soon we have some
>
>
> 2007/12/12, Glenn Forney <glenn....@nist.gov
> <mailto:glenn....@nist.gov>>:
Dragonit
ย
Bak, Daniel
deliberately or by ignorance inappropriate properties of materials. I
content that there will be always users who will not use appropriate
materials for their simulations.
However, I believe that if users choose to ignore disclaimers and
limitations of models or databases, it is their responsibility to accept
consequences. Historically, they were many abuses by users of what fire
models were designed to do. I further contend that if engineering or
scientific communities stopped using models because of the relatively
few abuses, then we would probably loose faith, interest, and support in
performance fire safety, hence, that would spell out the slow death of
fire engineering progress.
I believe that it is reasonable to, using your example of steel, share
at least 3 types of steel materials properties which would reflect 3
general categories, while stressing, as a disclaimer, that any variation
of the specified 3 materials would automatically imply that the user is
on his/her own as to the professional risk they are taking.
Otherwise, if we entertain creating a comprehensive database, we will be
always behind the "8 ball" as new alloys are created practically
everyday.
Please do not take offense by my point of view, I am simply and honestly
sharing with the group a compromising approach.
Thanks, Daniel
dr_jfloyd
opposed to being responsible for developing it and are opposed to it
being considered anything like a NIST Standard Reference which is how
anything released by NIST will be viewed. Any user can upload a file
to this forum or modify an exisiting file. There is only so much we
can (or are willing) to do. At some point the international user
community needs to make contributions. Post material defintions with
references and debate them, discuss test methods for measuring
properties, share the results of testing and validation done for
projects, make it known to those who FUND research in your respective
countries that this is important.
Kevin
and validation. We are trying to increase our efforts in this area.
This means that we accept some responsibility for verifying the
accuracy of the numerics and validating the physical assumptions of
the model, in cooperation with the end user who bears partial
responsibility for V&V'ing the tool that they use for engineering
analysis. We are not in a position to V&V the "database". If we follow
the guidelines for V&V (like ASTM E 1355), any "database" that we
distribute would fall under the same set of guidelines. We've neither
the time nor the personnel to do this. We have to focus on FDS/SMV,
and we look for collaborators who are willing to help with the input
properties.
clauten
for the discussion group that there are no widely-accepted, proven,
validated, standardized, whatever you want to call it test methods or
procedures to obtain all of the "material properties" needed for solid
phase modeling. For example, do we try to run a bunch of separate
small scale tests for thermal properties, reaction kinetics, radiative
characteristics, etc. and then put it all together in a MATL line and
hope that all of the separately measured pieces are reliable enough to
reproduce the bulk behavior of the material in say a Cone Calorimeter
experiment? Or, as someone suggested in a thread last week, do we set
up a "virtual" Cone Calorimeter and adjust the material properties
until the calculations match the experiments? There are advocates for
both approaches, the second one is probably more accessible to
practitioners, but the first approach seems more "scientific".
Also, some of the recent discussions in this forum (not necessarily in
this thread) sound like people are trying to get something for
nothing, or find someone else to do their work for them. Related to
pyrolysis modeling, the answer to the question "where can I find
material properties for yyyy" is usually "nowhere" unless you're
asking about PMMA or wood (and even those aren't as straightforward as
people think) or a material that someone studied for their PhD
dissertation. Unless the community becomes proactive and starts to
contribute experimental data, funding, computer time, manpower, etc.
to help develop test procedures and ultimately a material properties
database, unfortunately such a database is probably wishful thinking.
Matti
a simulation. Often the institution do not know the methods of
simulation.
They can not say anything about the method or programm you made the
simulation. But they can say the parameters a not good/wrong.
And so is often the question "why do you use these parameters?". And
now I have a problem. I often have only a list with some parameters.
But not the method how I get this value. If I have a "database" this
person can go to the internet and look the parameters and why I use
this.
So the thing I need is a mixture between a database and a linklist. I
can search materials and watch a scientific work.
I will do this for me to have a list of documents to search.
Do you know some institution where I perhaps will get these scientific
works.
Matti
Kevin
looking for. Chris L's point is that even if some institution (like
NIST or VTT) were to assemble a list of references or Internet links,
there still remains the problem of interpreting this data. Have you
ever seen a standard test report of a cone calorimeter measurement? Or
in the US a Steiner Tunnel test? Or an ISO 9705 corner test? Or any
standard test? The data contained in these reports cannot be easily
translated into HEAT_OF_REACTION, CONDUCTIVITY, SPECIFIC_HEAT, E, A,
REFERENCE_TEMPERATURE, etc, that FDS uses. These inputs are parameters
in a mathematical equation that is a model of solid phase pyrolysis. A
model; an idealization of reality. Chris's point is that we need to
develop a standard way of interpreting actual bench scale measurements
for use in FDS or any fire model.
Matti
no data, no database
How did you test fds without knowing the parameters?
If you have not a interpretation method, is every simulation
different?
>
> Erfahren Sie mehr >>
Hostikka Simo
The verification test suite contains tests against analytical and
numerical
solutions for simple combustion, radiation and heat transfer problems.
Tests were
not made against experimental data. The pyrolysis verifications
demonstrate that the
pyrolysis model can produce qualitatively realistic burning rates for
charring and
thermoplastic fuels in cone calorimeter. These verifications were not
made against
any particular test, because that would require the determination and
publication
of the specific pyrolysis parameters - something we want to avoid for
the reasons
mentioned in the previous posts.
The validation report, which is coming out soon, contains a large number
of full
scale fire test simulations, demonstrating the accuracy of the code in
such applications.
In these simulations, the material properties have usually been
specified by the
experimentalists.
I don't understand your last comment on every simulation being
different.
Interpretation of test results was the question.
Simo
Matti
my point is, that Kevin had say, that there is no standard way of
interpreting actual bench scale measurements.
My interpretation of that statement is, that you (or the
experimentalist) had interpreted the measurements, otherwise you can
not make a simulation (and could not test fds).
With "every simulation is different" I ment that if someone
interpreted the parameter not like you (because no standart), you will
have different properties and so their is (perhaps) no comparable
between different simulations.
I not want to say, that fds is not valid. You showed this in (real)
tests (and will these show in the validation report).
I am sorry if I had articulate me wrong
Hostikka Simo
Will those differences matter for the final result, is
another qestion.
> With "every simulation is different" I ment that if someone
> interpreted the parameter not like you (because no standart),
> you will have different properties and so their is (perhaps)
> no comparable between different simulations.
>
> I not want to say, that fds is not valid. You showed this in
> (real) tests (and will these show in the validation report).
I didn't think you said that. I just explained how the current
V&V efforcts have been made.
Regards,
Simo
Kevin
measurements, which does not mean that there is NO way to do this.
There are many talented fire protection engineers who use FDS and who
have developed methods to interprete test data. These methods are
justified by a variety of V&V exercises, but there still is no
consensus in the profession about how this is all to be done. Our goal
is to bring all the current knowledge together into a coherent
methodology. Only after this is done will it make sense to assemble a
database.
szilagyics
I'm happy your validation test this will be easier our work. The
Material properties is very interesting subject, and very usefull. But
in my work I experience the example particle board in USA is not same
in Hungary and another countries. This is came from different
manufacturing. Sorry I think the different materials means different
experimentals in different countries. If we have got crutch to this
experimentals the results will be better. I know it will be too hard
work, and maybe impossible. Just it will be good.
Csaba
>
> - Idรฉzett szรถveg megjelenรญtรฉse -
clauten
what was out there in terms of material properties for different
fuels. There's actually quite a bit of info, but it's usually not in
one place. One paper has kinetics, another has thermal properties,
another has heats of reaction, etc. One thing that I've been meaning
to do but just haven't had the time to do is, for a few common
materials, take literature values for material properties from a
multitude of different sources, plug them into a pyrolysis model, and
see how well it "works"--meaning does the calculated behavior come
close to a cone calorimeter (or similar) experiment? Instead, I've
always worked backwards from the cone calorimeter to estimate the
properties. But if anyone wants to try this type of exercise, there's
some tables in Chapter 2 of my dissertation that list a bunch of
literature values for material properties of a few different fuels.
http://me.berkeley.edu/~clauten/lautenberger_dissertation.pdf
However, I'm skeptical that "cherry picking" properties from different
sources would give reasonable predictions of bulk pyrolysis
experiments so V&V is extremeley important for this type of approach.
Just pointing to a paper where someone says the specific heat of wood
is 2000 J/kg-K does not mean you're "good", you still have to do some
work to make sure the model calculations are realistic!
Chris
>
> - Show quoted text -...
>
> read more >>
Kevin
One goal of ours should be to make the cone calorimeter a validation
exercise. That is, you get the fundamental properties of the "stuff"
and use the cone (preferably in non-burning mode) as a validation of
your "model" of the "stuff." Even though the FDS 5 solid phase
description is much better than before, it is still just a model, and
the properties all have some "effective" nature to them.
clauten
should be standard practice for V&V. The flame heat flux confounds the
boundary condition at the front face. If you really want to make
things complicated then there's also the issue of air vs. nitrogen
atmosphere. Even in the absence of a flame, some materials (especially
wood) behave much differently when pyrolyzed in air than in nitrogen
due to exothermic oxidative reactions. So the question then becomes,
for a fire spread simulation where a material initially pyrolyzes in
air contaminated with a little bit of combustion products but later
burns in an oxygen-depleted environment when it's covered by a
diffusion flame, do you try to optimize your material properties for
nitrogen or air?
Kevin
the pyrolysis process. But first we have to generalize the solid phase
routine a bit more, like making it possible to off-gas any listed
SPECies, not just water and "fuel." This has been on the to-do list
since you first posted it way back when. Things have settled down some
so we ought to be able to do some more development work, as opposed to
just bug fixing.
I wonder if it would be possible to validate a candidate set of
material props against a 0% oxygen cone test, and a 10% (or whatever)
oxygen test. Just so that you don't get a flame.
Matti
I created a group to find a standard way for the properties.
http://groups.google.com/group/fds-material-properties
Matti
>
> Erfahren Sie mehr ยป
Bryan Klein
Fracturing the effort, and information, regarding this issue will not
accelerate its resolution.
If you read the discussions carefully you will see that another google
group or containers for material property data is not what's missing,
but rather methods for acquiring data and validating it is needed.
Unless, you feel that the same questions that you posted on the new
group cannot be asked here.
If that is the case, then we have problems that need to be addressed,
since this would not be good.
You are free and welcome to maintain these lines of discussion and
even create pages within the FDS group to consolidate information
related to the topic of turning raw materials into validated FDS
material property data. Coordinated efforts are necessary to move
forward on this topic.
Creating another group is an obvious sign that you think the issue is
important, but in my opinion, it does not help with the solution.
Maybe begin the work to answer the questions on your new group and
post the results within the existing information resources (fds-smv
group) and you will begin to more effectively contribute to solving
this problem.
What is missing from the fds-smv group that made you feel that
creating another group is the solution?
What can we do to help in your opinion?
Feedback is always welcome.
Best Regards,
Bryan
Matti
my major aspect in the creation of a new group is the complexity of
the material properties problem. I think the fds group is too
complex.
You have posts in Smokeview and posts in FDS (problems with different
architecures, problems with materials, ...).
I think if we will think about the material properties problem, we
perhaps will have many posts and so it is very difficulty to search
something.
I am a student and so I not know every method to get material
properties. And the result is that I will have, in you eyes, very easy
questions and this will blow up the group.
What I will is a kind of subgroup where you can find summerized the
material properties problem.
A new group was the easiest way for me.
Perhaps you will have an other option to realize this.
I wish you a happy new year.
Best Regards,
Matti
dr_jfloyd
1) The more separate locations people have to go to for information,
the less likely it is they will all be visited by all people. By
having two FDS forums, you will split the user base and that is to no
one's benefit.
2) Threads will be duplicated on each forum. This hampers discussion.
3) We the developers have more than enough to do managing the issue
tracker and this forum. We have no desire to deal with a third online
task.
4) There are signs that Google may add the ability to create subforums
within a forum. If this happens, itwould do what you want without
fracturing the user base.
BKlein
Click the Pages link to the right and check them out.
As a group member, you are able to create new pages, or edit the
existing pages.
These might help focus the discussion, and members can comment/discuss
the page through a link at the bottom of the page.
What do you think?
-Bryan
Matti
this is a good idea.
I will delete the other group.
Matti