FDS and Smokeview 2006

[Kevin McGrattan <kevin.mcgrattan@nist.gov>]

All

Happy New Year! I'd like to update everyone on our plans for
FDS/Smokeview in 2006. In no particular order, here are the
activities currently underway at NIST and elsewhere:


Development -- we're planning on releasing a major overhaul
of FDS sometime this year, hopefully around summer time.
Simo Hostikka and co-workers at VTT Finland have rewritten
many of the solid phase routines. Their focus is increased
flexibility -- multiple layers of materials, non-planar
geometries, and other improvements to the various pyrolysis
routines. We have in recent years been adding alot of
features in response to user input, but we're at a point now
where the routines are getting too difficult to maintain as
is. A systematic overhaul is necessary.

Prof Arnaud Trouve from the University of Maryland has
accepted a joint appointment at NIST. He and his student, Hu
Zhixin, are working here part-time on various projects, most
notably the development of a better treatment of
under-ventilated fire by FDS. We plan to run a series of
half-scale compartment experiments here in the spring. These
experiments will be similar to those performed by Bill
Pitts, Rik Johnsson and Nelson Bryner about 10 years ago at
NIST, and those performed by Prof Yuri Vandsberger and
students at Virginia Tech. We are trying to develop a simple
two or three step mechanism to predict CO production, soot
production, extinction, etc.

Jason Floyd and Javier Trelles of Hughes Associates are
assisting in this effort under a grant from NIST. Jason
worked with me as a Post-Doc and helped to develop the
original combustion routine in FDS. Javier was also a
Post-Doc here and worked on large outdoor fires.

Combustion Science and Engineering is finishing up its grant
work on smoke detector activation. If any of you have any
comments or issues regarding the smoke detection algorithm
that is currently implemented in FDS (4.05 or 4.06), please
contact Doug Carpenter (dcarpenter@csefire.com).

Glenn Forney continues to make improvements to Smokeview. In
particular, he is investigating this year the idea of using
the GPU (Graphics Processing Unit) instead of the CPU to do
much of the heavy computation in Smokeview. This seems to be
the current trend in scientific visualization because of the
tremendous advance in graphics, mainly due to the gaming
industry. 


Validation -- Over the past few years, we have been working
closely with the US Nuclear Regulatory Commission on a
project to better document the V&V process (Verification and
Validation). The NRC has sponsored large scale experiments
here at NIST, plus the rewriting of both the FDS and CFAST
documentation to better conform to ASTM E 1355, Guide to
Evaluating Fire Models. The NRC now allows its licensees
(Nuclear Power Plants in the US) to use NFPA 805,
Performance-Based Standard for Fire Protection in Light
Water Reactor Electric Generating Plants, as an alternative
to existing requirements. 805 calls for fire models to
undergo a V&V process, and 1355 has been chosen by the NRC
as a guide to follow. This month, the NRC and EPRI (Electric
Power Research Institute) will release a joint publication
for public comment, in which 5 fire models are evaluated for
use in nuclear power plants. The models consist of two sets
of spread sheet calculations (FIVE and Fire Dynamics Tools),
two zone models (CFAST and MAGIC), and one CFD model (FDS).
I will send out more details when the documents are
released. I urge everyone who is interested to have a look
at the reports and send in comments if you have them. The
validation experiments and the corresponding FDS
calculations will form the core of a new FDS Manual. We will
continue to add to this "Validation Guide" in the years
ahead as we make improvements to the various routines.


Maintenance -- I've received numerous inquiries about new
computer hardware, software, etc., especially 64 bit
architectures and enhanced processors. Since we do not have
these machines here, I cannot be of much help if FDS does
not compile or run successfully on these machines. The best
I can do is try to maintain FDS using standard Fortran and C
conventions. I would suggest that you try to compile and run
FDS on any machine that you plan to purchase. I cannot
guarantee that the executable that I compile will run on the
latest hardware.

This year, the compiler that we have been using for Windows
PCs (Compaq Digital Fortran) is being discontinued. We plan
on switching over to the Intel Visual Fortran Compiler. This
should not affect most users, but be aware of it when FDS 5
comes out. There might be unexpected glitches, as is common
with a new compiler. Hopefully, we'll work out these
problems in the months ahead before releasing anything.


That's all for now.

Kevin


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

Telephone: (301) 975 2712
FAX:       (301) 975 4052