SESSION H6 — FIRE
H31 FIRE INDUCED THERMAL AND STRUCTURAL RESPONSE OF THE WORLD TRADE CENTER TOWERS.
Kuldeed Prasad, National Institute of Standards and Technology
Over the past several years, there has been a resurgence of interest in studying the response of building structures to
fires. Typically, the thermal loading for structural analysis of a building on fire is obtained from a standard time
temperature curve or by assuming a spatially uniform enclosure temperature. This decouples the structural analysis
from the fire simulations and, as a result, the structural response to spatially and temporally evolving fires cannot be
predicted. Simulations of the effects of severe fires on the structural integrity of buildings requires a close coupling
between the gas phase energy release and transport phenomena and the stress analysis in the load bearing materials.
A methodology has been developed for coupling CFD simulations of fire growth with finite element models for
thermal and structural analysis. A simple radiative transport model that assumes the compartment is divided locally
into a hot, soot laden upper layer and a cool, relatively clear lower layer is employed to predict radiative fluxes
incident on sub-grid scale structural members. Thermal responses of various structural components on focus floors
of World Trade Center Tower 1, coupled with realistic fire simulations are presented. One of the most striking
observations that emerge from these results is the wide variation of time-temperature curves that hold at different
points in the structure and the lack of resemblance of these curves to the standard time-temperature curve used in
furnace tests of structural elements. Finally the thermal response is used to predict the reduction in load carrying
capacity of the structure as a function of time. Sensitivity of the load carrying capacity to imperfections in
fireproofing, structural damage and fire growth is discussed.
H32 QUANTIFICATION OF FIRE SIGNATURES FOR PRACTICAL SPACECRAFT MATERIALS.
Randy Vander Wal, Jane Novak, NCSER, NASA Glenn Research Center
In microgravity there exists a lack of natural buoyant convective flow due to the nature of the symmetrical
environment. More important than the different physical characteristics of flames in microgravity however are the
differences in combustion products formed by fires in a low-gravity environment compared to those formed in
normal gravity. Common pre-fire events such as smoldering and pyrolysis generate substantial yields of volatiles
(partial decomposition products) that could be used as pre-fire signatures. The trick is knowing what compound or
compounds to look for and in what concentrations and their dependence upon buoyant or forced convective flow.
The purpose of this research is first to develop a facility that allows for ground-based combustion testing that
brackets flow conditions as may be encountered in a low-gravity environment. Secondly, the research seeks to
expand the knowledge base of combustion signatures and products evolved from practical spacecraft materials in
such environments. Finally, the data from the experiments may be used to guide development of improved fire
sensors for space vehicles and habitats. The chosen signatures for analysis include light gases, condensables, and
particulates.
H33 DETERMINATION OF PYROLYSIS TEMPERATURE FOR INFINITE RATE KINETICS MODELS OF CHARRING
MATERIALS.
Won Chan Park 1, Arvind Atreya1, Howard R Baum2
1
University of Michigan 2Building and Fire Research Laboratory, National Institute of Standards and Technology
An energy and mass balanced method to determine the pyrolysis temperature for infinite rate kinetics models is
proposed. The concept of the proposed method is to find the pyrolysis temperature that consumes the same amount
of energy to produce the same amount of mass when using the pyrolysis temperature (infinite rate kinetics) model as
when using the finite rate kinetics model for the entire charring process. The resulting pyrolysis temperature has the
form of pyrolysis rate weighted average temperature. Comparison between finite rate and infinite rate models for
various boundary conditions, geometries, heats of decomposition, kinetic parameters and assumptions used in the
literature were studied to assess the proposed method. It was found that the pyrolysis temperature (infinite rate
kinetics) model showed good agreement on pyrolysis time but predicted a larger mass loss rate for the beginning and
the final stages of the charring process and a smaller mass loss rate during the middle stage when compared with the
finite rate kinetics model. Extensive numerical studies on various factors influencing charring material pyrolysis
showed that heat flux, sample thickness, heat of decomposition and kinetic parameters are the most important
factors to determine an appropriate pyrolysis temperature. Char thermal conductivity, virgin solid specific heat and
virgin solid density have a lesser effect on the pyrolysis temperature. For practical application, the model
predictions are compared with experimental measurements of wood cylinder pyrolysis and other methods of
estimating the pyrolysis temperature.
H34 ULTRA FINE WATER MIST EXTINCTION OF A DIFFUSION FLAME.
Ramagopal Ananth, Richard C. Mowrey, Navy Technology Center for Safety & Survivability
Ultra fine water mist (< 10 µm drop size) is being investigated as a haoln1301 alternative for fire suppression due to
its high evaporation rates and minimal wetting of electronic components. In this work, we perform two-
dimensional, Lagrangian computations for the interactions of mono-disperse droplets with an axisymmetric
diffusion-flame formed in a cup burner. The computations show that the droplets entrained toward the flame base
play a critical role in flame extinction. As the droplet concentration is increased to the extinction concentration, the
flame lifts from the burner and blows-off mainly due to the combined effects of oxygen dilution and heat absorption.
Chain-branching reactions involving OH and H occur at a maximum rate at the flame base. The maximum rate is
found to be reduced by a factor of 5 to a critical value at extinction. The droplet extinction concentration decreases
from 15 to 10.5 mass % as the droplet size is increased from 4 to 32 µm. The large droplets are found to penetrate
the flame base better than the small drops, and are more effective despite the decreased evaporation rate. As the
droplet size is increased further however it is expected that the reduced evaporation rate will outweigh the increased
penetration and lead to decreased effectiveness. The predicted extinction concentration for droplets is in
quantitative agreement with the recent measurements of Fisher and co-workers at the Naval Research Laboratory.
H35 INTERACTION OF SMALL WATER DROPLETS WITH A METHANE NON-PREMIXED JET FLAME.
G.J. Liao1, A.U. Modak2 R.J. Kee2, J.-P. Delplanque1
1
Mechanical & Aeronautical Engineering Dept., University of California, Davis 2Colorado School of Mines
The focus of this research is the numerical investigation of the effects of small (10–50 microns) water droplets on a
methane co-flow diffusion jet flame. The targeted application is water-mist fire suppression. The numerical
simulation of mist/flame interactions is a computationally intensive endeavor for a multiphase reacting flow model
that constitutes an elliptic system of partial differential equations. This work aims at providing a computationally
efficient approach. Thus, a hybrid Eulerian-Lagrangian boundary layer formulation has been developed to describe
the interaction between diffusion flame and water mist. It is based on a previously developed numerical solution of
the diffusion flame obtained using a boundary-layer approach. The implementation of a two-way coupling strategy
between the Eulerian boundary-layer diffusion flame model and the Lagrangian water mist model is described. The
effect of droplet diameter and water injection mass flow rate on flame characteristics is evaluated based on the
predicted droplet trajectories, evaporated water distribution, heat extraction rate and oxidizer displacement.
H36 EVACUATION STUDY USING INTEGRATED FIRE/EVACUATION ENVIRONMENT.
Sergei A. Filatyev1, Angela K. Mellema2, Alok R. Chaturvedi2, Jay P. Gore1
1
School of Mechanical Engineering, Purdue University 2Purdue Homeland Security Institute
As complexity of buildings increase, it becomes more and more challenging to provide a satisfactory level of fire
safety. In the recent years it has been a tendency to move away from prescriptive fire codes to a performance-based
approach. Evacuation software is usually used to study egress time with little attention to possible fire development.
This is normally attributed to existence of separate fire and evacuation software developed without knowledge of
each other and thought of possible integration. In the present study an integrated fire/evacuation system was
developed. To simulate fire propagation NIST large-eddy simulation code FDS was used. The evacuation software
was designed to simulate human behavior during evacuation by tracking action of every individual in a building,
taking into account environmental effects (e.g. temperature, carbon monoxide, smoke, etc.) on the behavior and
health of each evacuee. The algorithm is based on a generalized force model and designed to simulate crowd
dynamics in densely-populated areas. The created shared environment was designed to provide the bridge between
multiple simulations for data transfer and model interaction. The integrated system was applied to study the Rhode
Island Nightclub fire that occurred on February 20, 2003 in West Warwick, RI. The accident data was used to
validate the model. The integrated system was also used for fire evacuation simulation experimentation to study
possible changes to the internal layout and the exit number on people health and death toll. The results showed that
minor design changes could have prevented the majority of the casualties. It was found that a created simulation
environment can provide suggestions for fire safety enhancement and can be used as a tool in fire safety design.
H37 AN INVESTIGATION OF CANOPY BULK DENSITY EFFECTS ON THE DYNAMICS OF CROWN FIRE INITIATION
Watcharapong Tachajapong, Jesse Loranzo, Shankar Mahalingam,
Department of Mechanical Engineering, University of California
David Weise, Forest Fire Laboratory, Pacific Southwest Research Station, USDA Forest Service
Crown fire initiation is studied by using detailed physical modeling based on a Large Eddy Simulation (LES). This
model is used to gain a better understanding of transition from surface to crown fire. In the studies conducted thus
far, we have investigated the effects of variation in height of crown fuel on crown fire initiation. Three distinct
regions can be identified within the structure of the surface fire: (1) a continuous flame region, (2) an intermittent
flame region, and (3) a hot plume gas region. Crown fuel ignition occurs when crown base height is in the
continuous flame region, while it will not occur when the crown base height is in the hot plume gas region.
Therefore, we will focus on crown fuel ignition when the crown base height is in the intermittent flame region. In
the intermittent flame region, the flame shape and height changes with time over the course of pulsation. This
causes the flame to impinge on the crown fuel base and the hot gas is forced to flow pass through the crown fuel.
Under certain conditions, it is observed that the crown fuel bulk density affects this impingement of flame and the
ignition of crown fire. The crown fuel properties used were estimated from live chamise (Adenostoma
fasciculatum) with fuel moisture content of 80% (dry basis). As the crown fuel bulk density was increased from
0.1kg/m3 to 0.26kg/m3, it was observed that the average hot gas velocity inside the crown matrix decreases from
0.27 m/s to 0.21 m/s, resulting in less air entrained into the crown fuel and more heat accumulated inside crown fuel
matrix. The higher bulk density also shows effects on the surface fire. As the hot gas flow into crown fuel matrix
was retarded, the average hot gas temperature at crown fuel base increases from 520 K to 561 K due to the mixing
rate of air and combustible gas around the base of crown fuel increases. Although higher fuel bulk density means
more fuel needs to be heated, the increase in accumulated heat within crown fuel matrix was higher than the
additional heat needed by fuel. Thus, the average crown fuel temperature increases from 398 K to 420 K. In this
study, three dimensional LES is used to model the surface to crown fire transition and to highlight and quantify the
physical processes involved in the unsteady interaction between the surface fire and crown fuel with different crown
fuel bulk densities.
H38 PILOTED IGNITION TO FLAMING IN SMOLDERING FIRE-RETARDED POLYURETHANE FOAM.
Olivier Putzeys1, A. Carlos Fernandez-Pello1, David Urban2
1
University of California, Berkeley 2NASA Glenn Research Center
Experimental results are presented on the piloted transition from smoldering to flaming in the fire-retarded
polyurethane foam Pyrell® (35.3 and 64.0 kg/m3). The experiments are conducted with small parallelepiped
samples vertically placed in a wind tunnel. Three of the vertical sample sides are insulated and the fourth side is
exposed to an upward oxidizer flow of variable oxygen concentration and to a variable radiant heat flux. The gases
emitted from the smoldering reaction pass upward through a pilot, which consists of a coil of resistance heating
wire. The results show that Pyrell® undergoes a weak smoldering reaction that requires significant assistance in the
form of external heating, due to the fire retardant additives. In addition, the fire retardants are shown to directly
affect the piloted ignition by inhibiting gas-phase reactions. The piloted transition to flaming is observed in oxygen
concentrations of 23% and above in both low-density and high-density Pyrell®. Comparisons with previous
experiments show that the piloted transition from smoldering to flaming is possible under a wider range of external
conditions (i.e. lower oxygen concentration) than the spontaneous transition from smoldering to flaming.