fire

Article
Evaluation of Available Safety Egress Time (ASET) in
Performance-Based Design (PBD) Using CFAST
Hyo-Yeon Jang 1 and Cheol-Hong Hwang 2, *

1 Department of Disaster Prevention, Graduate School, Daejeon University, 62 Daehak-ro, Dong-Gu,

Daejeon 34520, Republic of Korea; [email protected]
2 Department of Fire and Disaster Prevention, Daejeon University, 62 Daehak-ro, Dong-Gu,
Daejeon 34520, Republic of Korea
* Correspondence: [email protected]; Tel.: +82-42-280-2595

Abstract: In South Korea, the need to link fire and evacuation simulations to compare the available
safety egress time (ASET) and required safety egress time (RSET) in real time when implementing
performance-based design in buildings is increasing. Accordingly, the Consolidated Model of Fire
Growth and Smoke Transport (CFAST) has been discussed as an alternative to the fire dynamics
simulator, which requires high computational costs, sufficient experience in fire dynamics numerical
calculations, and various input parameters and faces limitations in integration with evacuation
simulations. A method for establishing a reasonable computational domain to predict the activation
times of smoke and heat detectors has been proposed. This study examined the validity of using
CFAST to predict factors relevant to the ASET evaluation. The results showed that CFAST, which
solved empirical correlations based on heat release rates, predicted high gas temperatures similarly.
Moreover, the applicability of the visibility distance calculation method using smoke concentration
outputs from CFAST was examined. The results suggest that despite the limitations of the zone
model, CFAST can produce reasonable ASET results. These results are expected to enhance the
usability of CFAST in terms of understanding general fire engineering technology and simple fire
dynamics trends.

Keywords: performance-based design (PBD); available safety egress time (ASET); fire dynamics
simulator (FDS); consolidated model of fire growth and smoke transport (CFAST); fire simulation
Citation: Jang, H.-Y.; Hwang, C.-H.
Evaluation of Available Safety Egress
Time (ASET) in Performance-Based
Design (PBD) Using CFAST. Fire 2024, 1. Introduction
7, 108. https://doi.org/10.3390/
In South Korea, predicting outcomes for various fire scenarios based on the pur-
fire7040108
pose and use of different buildings is essential for implementing performance-based de-
Academic Editor: Maged A. Youssef sign (PBD). Generally, PBD is adopted for the fire safety design of buildings, employing
engineering-based simulations to address fire hazards. PBD evaluates the fire safety of
Received: 4 March 2024
buildings by comparing the available safety egress time (ASET) and required safety egress
Revised: 21 March 2024
time (RSET) obtained through fire and evacuation simulations, respectively. This involves
Accepted: 22 March 2024
Published: 25 March 2024
assessing time-related factors for the safety of building occupants in scenarios with a high
fire risk. The ASET > RSET is considered safer than the ASET < RSET [1]. For fire sim-
ulation, the fire dynamics simulator (FDS) and Consolidated Model of Fire Growth and
Smoke Transport (CFAST) are generally used [2,3]. They were developed and validated
Copyright: © 2024 by the authors. by the National Institute of Standard and Technology (NIST). For evacuation simulations,
Licensee MDPI, Basel, Switzerland. Pathfinder [4] developed by Thunderhead Engineering, Simulex [5], and buildingEXO-
This article is an open access article DUS [6] are utilized. In conducting these simulations independently, the temperature and
distributed under the terms and concentrations of chemical species from fire simulations are not considered in the location
conditions of the Creative Commons predictions of occupants in evacuation simulations [7].
Attribution (CC BY) license (https://
ASET is calculated from the primary results of fire simulations for selected fire scenar-
creativecommons.org/licenses/by/
ios based on the purpose in the PBD implementation. Figure 1a shows the PBD process
4.0/).

Fire 2024, 7, 108. https://doi.org/10.3390/fire7040108 https://www.mdpi.com/journal/fire

 Fire 2024,
Fire 2024, 7, x FOR PEER REVIEW
7, 108 2 of 13
2 of 13

andevaluation
and evaluationfactors
factors for
for fire
fire and evacuation
evacuationsimulations
simulations[8]. ASET
[8]. ASET evaluates
evaluatesthe the
impact
impact
atatthe ◦ C), toxic gas concentrations (O
theheight
heightofofthe
thebreathing
breathing limit line line (1.8
(1.8m)
m)ofofheat
heat(60(60°C), toxic gas concentrations (O2 2
below15%,
below 15%,COCO22 above
above 5%,5%, and
and CO COabove
above1400
1400ppm),
ppm),and and visibility
visibilityrelated to the
related height
to the height
ofofthe
thesmoke
smokelayer
layerfollowing
following life
life safety
safetystandards
standards[9].[9].The
Thevisibility
visibilityis assessed
is assessedwhenwhenit isit is
below5 m
below 5m inin general
general facilities
facilities andand below
below 1010
mm in in relatively
relatively high-ceiling
high-ceiling assemblyand
assembly andsales
sales facilities.
facilities. DuringDuring this process,
this process, engineers
engineers mustmust simultaneously
simultaneously consider
consider thethe funda-
fundamental
mental prediction
prediction uncertainties
uncertainties of the offirethe fire modeling
modeling and systematization
and the the systematization of of interpre-
interpretation
tation methods and numerical result verification for accurate
methods and numerical result verification for accurate ASET evaluation. The RSET ASET evaluation. The RSET is the
is the sum of evacuation delay times from evacuation simulations,
sum of evacuation delay times from evacuation simulations, starting from the activation starting from the acti-
vation
point of point of the detectors
the detectors in the
in the fire fire simulations.
simulations. A building
A building is considered
is considered safe when
safe when the final
the final ASET is larger than the RSET, thereby providing
ASET is larger than the RSET, thereby providing a margin of safety. Figure 1b a margin of safety. Figure
shows 1b the
shows the major parameters that determine the evaluation results
major parameters that determine the evaluation results among the various parameters among the various pa-
rameters calculated for the ASET in the PBD evaluations conducted over the past five
calculated for the ASET in the PBD evaluations conducted over the past five years. In
years. In 94% of the cases, ASET was determined by visibility [10], thereby indicating the
94% of the cases, ASET was determined by visibility [10], thereby indicating the dominant
dominant influence and significance of visibility in the ASET evaluation [11].
influence and significance of visibility in the ASET evaluation [11].

Figure 1. (a) Fire and evacuation simulation timeline for performance-based design (PBD) in build-
Figure 1. (a) Fire and evacuation simulation timeline for performance-based design (PBD) in building
ing fires [8]; (b) Statistical proportions of the evaluation factors that determine the available safety
fires [8]; time
egress (b) Statistical proportions of the evaluation factors that determine the available safety egress
(ASET) [10].
time (ASET) [10].
FDS, commonly used in PBD implementation, continuously performs validation and
FDS, commonly
verification used in PBD
against numerous implementation,
full-scale fire experimentscontinuously performs
and is primarily validation
applied due toand
verification against numerous full-scale fire experiments and is primarily
its ability to efficiently interpret complex fire dynamics using a large eddy simulation applied due to its
(LES) technique [12,13]. However, obtaining reliable predictions with FDS requires accu-
ability to efficiently interpret complex fire dynamics using a large eddy simulation (LES)
rate physical and numerical input parameters. Specifically, the heat release rate per unit
technique [12,13]. However, obtaining reliable predictions with FDS requires accurate
area (HRRPUA)
physical has the input
and numerical highest impact on numerical
parameters. Specifically,solutions
the heat as release
a modeling
ratefire
persource
unit area
[14]. For accurate
(HRRPUA) has thepredictions of visibility
highest impact and CO solutions
on numerical and CO2 concentrations
as a modelingconsidered
fire sourcein[14].
fireaccurate
For safety standards, numerous
predictions user-required
of visibility and COinputand COparameters, such as the
2 concentrations CO and soot
considered in fire
yield standards,
safety generated during a fire, are
numerous required. Therefore,
user-required studies on establishing
input parameters, such as the databases
CO andofsoot
the experimentally
yield generated during measured
a fire, heat release rate
are required. (HRR), CO
Therefore, concentration,
studies and sootdatabases
on establishing yield
offor
thevarious combustibles
experimentally under fire
measured heatconditions
release ratehave been reported
(HRR), [15–17]. Toand
CO concentration, accurately
soot yield
predict
for variousdetector activationunder
combustibles times fire
in the RSET of the
conditions PBD
have implementation
been in buildings,
reported [15–17]. ex-
To accurately
perimentally
predict detectormeasured physical
activation timesinput
in the parameters
RSET of arethepresented [18,19]. As such,
PBD implementation in the use
buildings,
of various input
experimentally parameters
measured can lead
physical to different
input parameters riskare
assessments
presentedfor similar
[18,19]. Asfire scenar-
such, the use
ios, a critical drawback depending on the user’s discretion in PBD implementation.
of various input parameters can lead to different risk assessments for similar fire scenarios, Fur- a
thermore,
critical the grid
drawback size significantly
depending affectsdiscretion
on the user’s the accuracy
in PBDof numerical solutions,
implementation. thereby
Furthermore,
rendering the interpretation results less reliable and dependent on extensive
the grid size significantly affects the accuracy of numerical solutions, thereby rendering the numerical
experiences. results less reliable and dependent on extensive numerical experiences.
interpretation
CFAST, a typical fire modeling tool applying the zone model, assumes uniform prop-
erties by dividing the computation domain into upper hot and lower cool layers and
considering the interrelation of heat and mass flux across the boundary. Each zone is
assumed to be a control volume, and changes over time are considered through empirical
Fire 2024, 7, 108 3 of 13

correlations using the law of conservation of mass and energy. Thus, understanding fire dy-
namics is necessary to conceptually simplify fire behavior, with an existing barrier to entry.
Although small computational domains provide a relatively good prediction performance,
large domains or aspect ratios can reduce the reliability of the results. Based on the nature of
the model, the separation of heat and the concentration between the upper and lower layers
must be considered, thereby necessitating similar dimensions for the width and height of
the computational domain [20]. However, CFAST has advantages including easy operation
after short-term training. It can be used on lower-performance computers and requires a
significantly low computational cost. It has faster computation times than other analysis
tools and a good prediction performance in smaller computational domains. Owing to
these advantages, fire engineers and researchers are continuously verifying the accuracy of
CFAST based on experimental results [21,22] and have obtained reasonable results from
the perspective of the computational cost [23]. Therefore, by providing a setting approach
to understand fire dynamics trends through the simple assumptions of CFAST, its practical
application for achieving simple and targeted results has been examined.
This study was conducted as part of a project aimed at developing a program for
establishing evacuation plans by fire safety managers in a building fire scenario. The
applicability of CFAST, a relatively simple analysis tool, was examined when evaluating
the ASET and RSET within the same domain through the real-time integration of fire
and evacuation simulations, which is necessary for enhancing the reliability of life safety
assessment [24,25]. Based on the assumption of the uniform properties of the zone model
within each zone, fire modeling in areas (corridors) with large horizontal aspect ratios
requires careful consideration. Previous studies have compared the activation times of
smoke and heat detectors predicted by FDS and CFAST, thereby reporting measures for
setting computational domains [26]. This study presents a method for utilizing CFAST
in evaluating ASET, which has a significantly lower computational cost than FDS. In the
future, the prediction performance of CFAST on real buildings will be investigated from
the perspective of the practical application of the methodology described in this paper.

2. Conditions and Methods for Fire Simulation

2.1. Fire Simulation Conditions
To review the physical quantities evaluated in life safety standards, as predicted by
smoke and heat spread in fire models, FDS version 6.7.9 and CFAST version 7.7.0 were applied.
Figure 2a shows the computational domain of the FDS considered for the ASET evaluation,
which is set in accordance with the standard compartment. In PBD implementation, a
single floor is examined; hence, a corridor shape extended by 12.0 m from the opening was
considered. In this setup, fire is started inside a standard compartment, and heat and smoke
are configured to move into the corridor, while the stairwell geometry limits the flow from the
doorway. The corridor section is set to 16.0 m (x) × 2.4 m (y) × 2.4 m (z), and the staircase is
set to 4.8 m (x) × 4.8 m (y) × 7.2 m (z). As CFAST cannot consider structures, the opening of
the standard compartment was excluded for a direct comparison of the prediction results. The
physical quantities considered for the ASET were set at a height of 1.8 m from the floor and
spaced 0.2 m apart from the exit to minimize the influence of boundary conditions. Predicted
physical quantities included the gas temperature and volume fractions of O2 , CO2 , and CO,
including visibility. CFAST does not output visibility over time; however, we post-processed
visibility using smoke concentration data based on the importance and need for visibility, as
shown in Figure 1b.
 Fire 2024,
Fire 2024, 7, x FOR PEER REVIEW
7, 108 4 of 13
4 of 13

Figure
Figure 2.2.Schematic
Schematicof
ofthe
the computational
computational domain
domainfor
forthe
theprediction
predictionof of
physical quantities
physical related
quantities to to
related
the ASET. (a) FDS; (b) CFAST.
the ASET. (a) FDS; (b) CFAST.
Figure2b
Figure 2bshows
shows thatthat the
the computational
computationaldomain domaininin the CFAST
the CFAST is divided
is divided andandset up
set up
with regular intervals. The reason for this approach is that our previous studies showed
with regular intervals. The reason for this approach is that our previous studies showed
how to utilize the CFAST, which has the advantage of a low computational cost. Owing
how to utilize the CFAST, which has the advantage of a low computational cost. Owing
to the limitations relative to applying the zone model in areas with large aspect ratios, it
to the limitations relative to applying the zone model in areas with large aspect ratios, it
was reported that the computational domain needed to be divided at regular intervals to
was reported that the computational domain needed to be divided at regular intervals to
accurately predict the physical quantities. Studies on application measures for utilizing
accurately predict the physical quantities. Studies on application measures for utilizing
heat and smoke detectors in the CFAST to accurately evaluate RSET have been reported.
heat andon
Based smoke detectorscorrelations
the prediction in the CFAST to accurately
of the CFAST forevaluate RSET have
the temperature, flow been
rate,reported.
and
Based on the prediction correlations of the CFAST for the temperature,
smoke concentration, careful consideration is required to divide the computational flow rate, anddo-smoke
concentration, careful consideration
main for heat and smoke detectors [26]. is required to divide the computational domain for
heat and smoke detectors [26].
In predicting temperature using heat detectors, the influence of velocity is dominant.
EvenIn ifpredicting
the area oftemperature
interest is close to the
using firedetectors,
heat source, if the
it is influence
divided from the fire room,
of velocity the
is dominant.
ceiling jet velocity is fixed at 0.1 m/s [20]. Although undivided computational domains
Even if the area of interest is close to the fire source, if it is divided from the fire room,are the
appropriate for accurately predicting high gas temperatures, in PBD, fire simulations are
ceiling jet velocity is fixed at 0.1 m/s [20]. Although undivided computational domains
conducted
are appropriate for the
forentire building
accurately or a floor,high
predicting as needed. Hence, wheninsePBD,
gas temperatures, ing the
firecomputa-
simulations
tional
are domain for
conducted based theonentire
the structure
buildingofor thea building,
floor, as the probability
needed. Hence, thatwhen
the fire room the
setting
and exit are in domain
computational the same based
computational domain is of
on the structure very
thelow. To predict
building, thethe smoke concen-
probability that the
tration,
fire roomthe andcomputational
exit are in the domain
sameiscomputational
appropriate for domain
dividing is thevery
fire room,
low. To while mass the
predict
transfer
smoke should be ensured
concentration, between adjacent
the computational domain compartments.
is appropriate In for
thisdividing
study, the thedivided
fire room,
computational
while mass transfer domain required
should for the between
be ensured prediction of various
adjacent physical quantities
compartments. In thiswas set, the
study,
including
divided the FDS anddomain
computational CFAST with a square
required for thefireprediction
source of of 1.0various
m in the center of
physical each
quantities
compartment.
was set, including the FDS and CFAST with a square fire source of 1.0 m in the center of
For fire simulation input data, the prediction results of the physical quantities for
each compartment.
evaluating
For firethe ASET under
simulation inputvarious
data,firethegrowth curves
prediction were applied,
results as shownquantities
of the physical in Figure for
evaluating the ASET under various fire growth curves were applied, as shown inchem-
3. The temperature of the gases generated from the flame, the concentration of the Figure 3.
ical species, and the visibility can differ significantly from the maximum HRR (HRRmax)
The temperature of the gases generated from the flame, the concentration of the chemical
species, and the visibility can differ significantly from the maximum HRR (HRRmax ) and
fire growth rate. In this study, the prediction results of the FDS and CFAST were compared
based on changes in the HRRmax and fire growth rate. Figure 3a shows the fire conditions
considered for review based on changes in the fire growth rate. Four conditions—slow,
 Fire 2024, 7, x FOR PEER REVIEW 5 of 13

Fire 2024, 7, 108

and fire growth rate. In this study, the prediction results of the FDS and CFAST were com- 5 of 13

pared based on changes in the HRRmax and fire growth rate. Figure 3a shows the fire con-
ditions considered for review based on changes in the fire growth rate. Four conditions—
slow, medium,
medium, fast, andfast, and ultrafast—outlined
ultrafast—outlined in NFPA in NFPA
92 [27]92were
[27] considered
were considered
basedbased
on theonsame
the same
HRR max ofHRR
1.0 max
MW.of Additionally,
1.0 MW. Additionally, all the conditions
all the conditions were set were set to HRR
to reach reachmax
HRR 1.0atMW
atmax
1.0 were
and MW and were maintained
maintained for 1000 s. forFigure
1000 s.3b Figure
shows 3bthe
shows the fire conditions
fire conditions considered
considered for review
for review
based based in
on changes on the
changes
HRRmax in the
. ToHRR max. To
further furtherthe
examine examine
physicalthequantities
physical quantities
evaluated in
evaluated
the in theFDS
ASET using ASET using
and FDS and
CFAST, the CFAST,
slowestthe fireslowest
growthfire growth
rate rate (α =was
(α = 0.0029) 0.0029)
set.was
All the
set. All themaintained
conditions conditions the
maintained
slow firethe slow rate,
growth fire growth
although rate,
HRRalthough
max wasHRR
assumed
max was toas-
reach
sumed to reach 0.5, 1.0, 1.5, and 2.0 MW and was maintained for 1000
0.5, 1.0, 1.5, and 2.0 MW and was maintained for 1000 s. Normal heptane was used as the s. Normal heptane
was
fire used as
source forthe fire source for various
implementing implementing variouscurves
fire growth fire growth
[28]. curves
From a[28]. From a com- cost
computational
putational cost perspective, the FDS required approximately 25,920
perspective, the FDS required approximately 25,920 min, whereas the computation min, whereas the com-
time of
putation time of the CFAST
the CFAST was 155,500 times higher. was 155,500 times higher.

Figure3.3. Fire
Figure Fire conditions
conditions considered
consideredfor
forthe
thepredictive
predictiveperformance
performance comparison of the
comparison of FDS and and
the FDS
CFAST:(a)
CFAST: (a)Fire
Fire growth
growth rate
rate (slow,
(slow, medium,
medium,fast,
fast,and
andultrafast); (b)(b)
ultrafast); Maximum
Maximumheatheat
release rate rate
release
(HRR max
(HRRmax ).).

2.2.Fire
2.2. FireSimulation
Simulation Method
Method
Thecriteria
The criteria for
for evaluating
evaluating the theASET
ASETininPBD PBD forfor
specific target
specific buildings
target buildings are presented
are presented
in Table 1. The fire simulations assess the time required to reach
in Table 1. The fire simulations assess the time required to reach a temperature a temperature of 60 °C
of or
60 ◦ C
ormore
more at at
a height
a height of 1.8 m (breathing
of 1.8 m (breathinglimit limit
line) from
line) the
from floor
theafter
floora fire
afterstarts.
a fireThe time The
starts.
when
time the concentrations
when the concentrations of various chemical
of various species exceed
chemical speciestheir
exceedcritical values
their is evalu-
critical values is
ated, such such
evaluated, as when it is below
as when 15% for
it is below 15% O2,for
above
O2 , 5% for CO
above 5%2,for
and CO 1400 ppm for CO. This
2 , and 1400 ppm for CO.
evaluation
This includes
evaluation the time
includes the when the visibility
time when for occupants
the visibility drops to
for occupants 5 m or
drops to less
5 m for
or less
light-reflecting
for light-reflectingobjects. UsingUsing
objects. a conservative approach,
a conservative the ASET
approach, theof ASET
the building
of theisbuilding
evalu- is
ated basedbased
evaluated on theonsmallest time value
the smallest time among these physical
value among quantities
these physical for reaching
quantities their
for reaching
respective critical values. The major input parameters of the fire
their respective critical values. The major input parameters of the fire simulation simulation that can sig- that
nificantly alter the physical quantities that influence ASET, such as the
can significantly alter the physical quantities that influence ASET, such as the heat of heat of combustion,
CO yield, and soot yield, were set identically to those in previous studies [26].
combustion, CO yield, and soot yield, were set identically to those in previous studies [26].
Table 1. Performance Criteria of Life Safety in the Domestic PBD Standard at 1.8 m.
Table 1. Performance Criteria of Life Safety in the Domestic PBD Standard at 1.8 m.
Parameter Value
GasParameter
temperature Over Value
60 °C
OGas
2 volume fraction
temperature Under
Over15%
60 ◦ C
CO volumefraction
O22 volume fraction Over 5%15%
Under
CO2 volume fraction Over 5%
CO volume fraction Over 1400 ppm
Visibility (for a light-reflecting object) Under 5 m (other facility)

The FDS calculates the physical conditions at the center of each cell in a three-
dimensional grid as a function of time, which is based on fundamental equations for
Fire 2024, 7, 108 6 of 13

mass, momentum, and energy. In this case, a simple chemistry method is applied, thereby
determining the amount of combustion products simply through yield values. Specifically,
the concentrations of the chemical species are implemented in proportion to the fuel supply
using user-inputted CO and soot yields [12]. The generated CO and soot move based on
the dynamics equations of the gas phase implemented by heat release rates and buoyancy
from the fire source. However, CFAST applies Bernoulli’s equation to predict the flow rate
and pressure using the mass and energy conservation laws [20]. Therefore, by dividing the
area where the fire source is located from areas without it, the concentration of the chemical
species in the upper and lower layers of each section is output owing to the pressure differ-
ence at the boundary conditions. Similar to the FDS, the CO and soot produced per unit
mass of consumed fuel are specified, thereby enabling the implementation of the chemical
species and smoke concentration.
Particularly, the smoke concentration predicted based on the user-specified soot yield
is closely related to visibility, a major physical quantity in determining ASET. Specifically,
in both FDS and CFAST, the smoke concentration is predicted using Equation (1). During
a fire, the smoke concentration is calculated along with other major products, which are
determined by the intensity of the monochromatic light passing through a distance L [12,20].
 
1 I
D = − log10 = Klog10 e, K = Km ρYs (1)
L I0

where D represents the optical density, I0 is the light intensity without smoke, and I is the
light intensity after passing through smoke. The calculated optical density is then redefined
by the light extinction coefficient K. Additionally, K is a major variable for evaluating
visibility within the computational domain, which is expressed as the product of Km (mass
specific extinction coefficient, m2 /kg) and ρYs (density of smoke particulate, mg/m3 ) [12].
The default value of K, an average of the various fuels and conditions, is 8700 m2 /kg [29].
ρYs is determined in real time by the smoke generated in a fire and is directly influenced by
the user-inputted soot yield.
The FDS predicts visibility caused by the smoke concentration using Equation (2).
S represents visibility, and C is a dimensionless specific constant based on the type of object
viewed through smoke, with C = 8 for light-emitting signs and C = 3 for reflective signs.
The default value of C used in FDS visibility calculations is 3, set for reflective signs [30].

S = C/K (2)

As CFAST does not directly output visibility, it applies Equation (2), which is used in
FDS, to evaluate the visibility caused by smoke from a fire. The coefficient K, as predicted in
CFAST, and the visibility constant set to 3 for reflective markers such as in FDS are utilized.
Since FDS assumes a maximum visibility distance of 30 m, any value of S exceeding 30 m
is set to 30 m in post-processing [12]. Previous studies have reported on the predictive
performance by comparing the activation times of heat and smoke detectors in the FDS
with those in the CFAST to confirm the feasibility of calculating RSET using CFAST. This
approach can contribute to the completeness of programs aimed at developing evacuation
plans by fire safety managers.

3. Results and Analysis

3.1. Prediction of ASET Evaluation Factors Based on Fire Models
Figure 4 compares the gas temperatures predicted at the 1.8 m breathing limit line
using FDS and CFAST. The X-axis shows the gas temperatures predicted using FDS, and
the Y-axis shows those predicted using CFAST, thereby enabling a direct comparison of
the physical quantities predicted by both tools. The critical value for the PBD is 60 ◦ C.
This value is reached in all fire scenarios regardless of the fire growth rate or HRRmax .
Figure 4a shows that as the fire growth rate increases, CFAST tends to under-predict the
gas temperatures. This under-prediction by CFAST seems to arise because the model does
 3. Results and Analysis
3.1. Prediction of ASET Evaluation Factors Based on Fire Models
Figure 4 compares the gas temperatures predicted at the 1.8 m breathing limit line
using FDS and CFAST. The X-axis shows the gas temperatures predicted using FDS, and
the Y-axis shows those predicted using CFAST, thereby enabling a direct comparison of
Fire 2024, 7, 108 the physical quantities predicted by both tools. The critical value for the PBD is 60 °C. This 7 of 13
value is reached in all fire scenarios regardless of the fire growth rate or HRRmax. Figure
4a shows that as the fire growth rate increases, CFAST tends to under-predict the gas tem-
peratures.the
not consider Thisexact
under-prediction by CFAST
ceiling jet velocity seems
due to arise
to the because
division the model
of the does not domain
computational
consider the exact ceiling jet velocity due to the division of the computational domain
where the fire source is located and the area predicting the ASET. Specifically, CFAST
where the fire source is located and the area predicting the ASET. Specifically, CFAST as-
assumes
sumesa aceiling
ceiling jet
jet velocity
velocity ofof0.1
0.1 m/s
m/s in computational
in computational domainsdomains separated
separated from the from
fire the fire
room,room, as reported in previous studies [20]. Figure 4b shows that although the range of the range of
as reported in previous studies [20]. Figure 4b shows that although the
predicted gas
the predicted gastemperatures
temperatures varies with the
varies HRR
with the , bothmax
maxHRR the ,FDS and
both CFAST
the FDS predict sim- predict
and CFAST
ilar gas
similar gas temperatures.
temperatures.

Figure
Figure 4. Comparisonofofgas
4. Comparison gas temperatures
temperatures forfor
different fire models
different according
fire models to the (a)to
according fire growth
the (a) fire growth
rate and (b) maximum heat release rate.
rate and (b) maximum heat release rate.
Figure 5 compares the concentrations of the chemical species predicted at positions
Figure
0.2 m from5 compares the
the exit using concentrations
the of the chemical
fire model. By observing species of
the concentrations predicted at positions
O2, CO2, and
0.2 mCOfrom the exit using the fire model. By observing the concentrations
at varying fire growth rates, we confirmed that the concentration is predicted within of O 2 , CO2 , and
CO atthe same range when only the fire growth speed differs, while the HRR max remains thewithin the
varying fire growth rates, we confirmed that the concentration is predicted
samesame.
range Furthermore,
when onlythe thefaster
fire the fire growth
growth speedspeed, thewhile
differs, more similar
the HRRthemax
results of the the same.
remains
predicted O 2, CO2, and CO concentrations using FDS in the initial stages of the fire. Con-
Furthermore, the faster the fire growth speed, the more similar the results of the predicted
versely, slower speeds tend to result in over-prediction. Looking at the predicted concen-
O2 , CO 2 , and CO concentrations using FDS in the initial stages of the fire. Conversely,
trations of the chemical species to evaluate ASET using the fire models, the critical values
slower speeds tend to result in over-prediction. Looking at the predicted concentrations
are not reached under all conditions. The concentrations of the chemical species predicted
of the
using CFAST species
chemical to evaluate
are over-predicted ASET using
compared the
to those fireFDS,
using models, thesuggesting
thereby critical values
that are not
ASET values would be smaller than those of FDS. From a conservative perspective, CFAST
reached under all conditions. The concentrations of the chemical species predicted using
CFASTcan be
areutilized to grasp simple
over-predicted trends into
compared firethose
engineering.
using FDS, thereby suggesting that ASET
Fire 2024, 7, x FOR PEER REVIEW
values would be smaller than those of FDS. From a conservative perspective,8 CFAST
of 13
can be
utilized to grasp simple trends in fire engineering.

Figure 5. ComparisonofofOO2,, CO
5. Comparison CO2, ,and CO concentrations for different fire models against the fire
Figure 2 2 and CO concentrations for different fire models against the fire
growth rate. (a) O2 concentration; (b) CO2 concentration; (c) CO concentration.
growth rate. (a) O2 concentration; (b) CO2 concentration; (c) CO concentration.
Figure 6 compares the concentrations of the chemical species predicted at the 1.8 m
Figure 6 compares the concentrations of the chemical species predicted at the 1.8 m
breathing limit line using FDS and CFAST based on the changes in HRRmax. The X-axis
breathing limit
shows the line using predicted
concentrations FDS andusingCFASTFDS,based onY-axis
and the the changes in predicted
shows those HRRmax .us-
The X-axis
ingthe
shows CFAST, thereby enabling
concentrations a direct
predicted comparison
using between
FDS, and both tools.
the Y-axis showsFigure 6apredicted
those shows using
that thereby
CFAST, as the HRR max increases,
enabling the comparison
a direct oxygen concentration
between predicted usingFigure
both tools. the fire6a
models
shows that as
decreases. As the HRRmax increases, compared with FDS, CFAST tends to over-predict the
oxygen concentration. As previously mentioned, the critical value for oxygen concentra-
tion is 15%, and this critical value is reached in the 2000 kW condition predicted using
CFAST. Examining Figure 6b,c, which compare and illustrate the volume fractions of CO2
and CO predicted using the fire models, the concentrations of CO2 and CO increase as the
HRRmax increases. The critical value for CO2 in PBD is 5%. However, the CO2 concentra-
 Figure 5. Comparison of O2, CO2, and CO concentrations for different fire models against the fire
growth rate. (a) O2 concentration; (b) CO2 concentration; (c) CO concentration.

Fire 2024, 7, 108 Figure 6 compares the concentrations of the chemical species predicted at the 1.8 m 8 of 13
breathing limit line using FDS and CFAST based on the changes in HRRmax. The X-axis
shows the concentrations predicted using FDS, and the Y-axis shows those predicted us-
ing CFAST, thereby enabling a direct comparison between both tools. Figure 6a shows
the HRRmax increases, the oxygen concentration predicted using the fire models decreases.
that as the HRRmax increases, the oxygen concentration predicted using the fire models
As the HRRmax
decreases. increases,
As the compared
HRRmax increases, with FDS,
compared CFAST
with FDS, CFASTtendstendstotoover-predict the oxygen
over-predict the
concentration. As previously
oxygen concentration. mentioned,
As previously mentioned,thethe
critical
criticalvalue
value forforoxygen
oxygen concentration is
concentra-
15%,tion
andis this
15%,critical
and thisvalue
criticalisvalue is reached
reached in thein 2000
the 2000
kWkW conditionpredicted
condition predicted using
using CFAST.
CFAST. Examining
Examining Figure
Figure 6b,c, which6b,c,compare
which compare and illustrate
and illustrate the the volume
volume fractions of
fractions of CO
CO2 2 and CO
and COusing
predicted predicted using
the fire the fire models,
models, the concentrations
the concentrations of COof CO 2 and CO increase as the
and CO increase as the HRRmax
2
HRRmax increases. The critical value for CO2 in PBD is 5%. However, the CO2 concentra-
increases. The critical value for CO2 in PBD is 5%. However, the CO2 concentrations
tions predicted by the CFAST and FDS do not reach this critical value. The critical value
predicted by the CFAST and FDS do not reach this critical value. The critical value for CO
for CO is 1400 ppm, and studies have shown that neither of the predicted CO concentra-
is 1400 ppm,
tions of theand studies
fire models have
have shown
reached this that neither
critical of the
value. The predictedofCO
concentrations theconcentrations
chemical of
species
the fire predicted
models haveusing CFAST
reached arecritical
this over-predicted compared
value. The with those predicted
concentrations using species
of the chemical
FDS. using CFAST are over-predicted compared with those predicted using FDS.
predicted

Figure 6. Comparison of
6. Comparison ofOO2,2CO 2, and
, CO CO concentrations based on the maximum HRR. (a) O2 con-
Figure 2 , and CO concentrations based on the maximum HRR. (a) O2
centration; (b) CO2 concentration; (c) CO concentration.
concentration; (b) CO2 concentration; (c) CO concentration.
As CFAST does not directly output visibility, as described in Section 2.2, it was post-
As CFAST does not directly output visibility, as described in Section 2.2, it was post-
processed using the method employed by FDS. Thus, visibility predicted at the 1.8 m
processed
breathingusing
limit the
line method employed
using the FDS and CFAST by is
FDS.
shownThus, visibility
in Figure predictedmen-
7. As previously at the 1.8 m
tioned, limit
breathing when line
post-processing
using thethe FDS optical
anddensity
CFASTpredicted
is shownby the CFAST into
in Figure visibility,
7. As a
previously men-
limitwhen
tioned, of 30 m is applied when the
post-processing thevalue is exceeded.
optical Consequently,
density predicted by compared
the CFAST with FDS,
into visibility, a
limitCFAST
of 30 mmaintains a visibility
is applied whenofthe 30 value
m for an extended duration
is exceeded. after smokecompared
Consequently, generation.with FDS,
The critical value for visibility in PBD is 5 m. All fire scenarios reach
CFAST maintains a visibility of 30 m for an extended duration after smoke generation. this critical value The
Fire 2024, 7, x FOR PEER REVIEW
critical value for visibility in PBD is 5 m. All fire scenarios reach this critical value regardless 9 of 13

of the fire growth rate or HRRmax . Figure 7a shows that as the fire growth speed increases,
the visibility predicted by CFAST is over-estimated compared with that predicted by FDS.
regardless of the fire growth rate or HRRmax. Figure 7a shows that as the fire growth speed
This discrepancy exists because
increases, CFASTpredicted
the visibility predictsbysmoke
CFAST is concentration in the relatively
over-estimated compared av-
with that pre-
eraged wide range ofdicted by FDS. layer,
the upper This discrepancy exists because
unlike FDS, which CFAST predicts
calculates smoke concentration
visibility based onina
the relatively averaged wide range of the upper layer, unlike FDS, which calculates visi-
localized grid for smoke concentration at a height of 1.8 m. Furthermore, Figure 7b shows
bility based on a localized grid for smoke concentration at a height of 1.8 m. Furthermore,
that the predicted visibility
Figure 7bbased
showson thatchanges in the
the predicted HRRbased
visibility max does not exhibit
on changes considerably
in the HRR max does not

exhibit considerably quantitative differences.

quantitative differences.

Figure
Figure 7. Comparison of 7. Comparison
visibility for theof visibility
different for fire
the different
models firebased
modelson
based
theonfire
the growth
fire growthrate
rate and
and
maximum HRR. (a) Fire growth rate; (b) HRRmax.
maximum HRR. (a) Fire growth rate; (b) HRRmax .
3.2. Comparison of Predicted Results of Major ASET Evaluation Factors Using Fire Models
This study compared the raw data of physical quantities reaching critical values
among the ASET evaluation factors, as stipulated by the life safety standards in Table 1,
through the fire scenarios considered. As shown in Figure 1, the statistically significant
ranking of the evaluation factors of the ASET when performing PBD was analyzed and
found to be visibility, followed by the gas temperature. Similar physical quantities were
Fire 2024, 7, 108 9 of 13

3.2. Comparison of Predicted Results of Major ASET Evaluation Factors Using Fire Models
This study compared the raw data of physical quantities reaching critical values
among the ASET evaluation factors, as stipulated by the life safety standards in Table 1,
through the fire scenarios considered. As shown in Figure 1, the statistically significant
ranking of the evaluation factors of the ASET when performing PBD was analyzed and
found to be visibility, followed by the gas temperature. Similar physical quantities were
selected for the fire scenario applied in this study. Specifically, the prediction results for the
gas temperature at the 1.8 m breathing limit line in FDS and in the upper layer in CFAST,
including visibility, were compared and analyzed.
Figure 8a shows the raw data for the changes in the fire growth rate, thereby indicating
a similar trend to the predictions of FDS, although the onset of the temperature rise differs.
As the fire growth rate increases, the initial temperature rise in the CFAST appears slower.
The critical value for the gas temperature in life safety standards is 60 ◦ C, and it is influenced
by the change in the fire growth rate. Consequently, the times when the gas temperature
reaches the critical value are 279, 143, 77, and 45 s in FDS and 290, 164, 99, and 62 s
in CFAST. Figure 8b shows the gas temperature changes based on the HRRmax , thereby
revealing overall similar trends. When comparing the results predicted by FDS and CFAST,
CFAST tends to under-predict the maximum gas temperature as the HRRmax increases.
Regarding changes in the HRRmax , the critical value of the gas temperature (60 ◦ C) is
similarly predicted. Listing the times when the gas temperature reaches the critical value,
FDS shows 263, 267, 273, and 275 s, whereas CFAST consistently reaches the same time of
Fire 2024, 7, x FOR PEER REVIEW 10 of 13
290 s for all the conditions. Hence, regardless of the changes in the fire growth rate and
HRRmax , the gas temperatures predicted by both fire models correspond.

Figure8.8.Comparison
Figure Comparison of
of the
the gas
gastemperature
temperatureusing
usingFDS and
FDS CFAST
and in terms
CFAST of changes
in terms in theinfire
of changes the fire
growth rate and maximum HRR. (a) Fire growth rate; (b) HRR
growth rate and maximum HRR. (a) Fire growth rate; (b) HRRmax . max.

Figure99compares
Figure compares the the visibility
visibilitypredicted
predictedbybyFDS FDSand and CFAST
CFAST post-processed
post-processed intointo
visibility.Observing
visibility. Observingthe the predicted
predicted results
results forfor changes
changes in the
in the fire growth
fire growth rate,rate,
CFAST CFAST
shows a
showsvisibility
slower a slower reduction
visibility reduction
than FDS,than FDS, although
although it reachesit the
reaches thevalue
critical critical valueThe
faster. faster.
reason
The reason for the difference in the fire models is that while FDS calculates
for the difference in the fire models is that while FDS calculates numerous grid points, numerous grid
points,divides
CFAST CFASTthemdivides them
into twointo two
zones andzones and expresses
expresses each zone eachaszone as an average
an average value.
value. Therefore,
Therefore, visibility does not decrease at the initial stages of a relatively
visibility does not decrease at the initial stages of a relatively low smoke concentration low smoke con- but
centration but does as the fire grows and a sufficient smoke concentration is generated.
does as the fire grows and a sufficient smoke concentration is generated. Although the
Although the initial reduction in visibility differs, the trend is very similar, with FDS
initial reduction in visibility differs, the trend is very similar, with FDS reaching the critical
reaching the critical value in 50, 84,151, and 281 s and CFAST reaching it in 58, 88, 135, and
value in 50, 84,151, and 281 s and CFAST reaching it in 58, 88, 135, and 212 s. The delayed
212 s. The delayed reduction in visibility under the slow condition of FDS is a ributed to
reduction in visibility under the slow condition of FDS is attributed to fluctuations caused
fluctuations caused by fresh air entering at the boundary condition of the exit. Owing to
by fresh air entering at the boundary condition of the exit. Owing to an increase in the
an increase in the HRR, the visibility predicted by FDS and CFAST exhibits similar trends,
HRR,
and athe visibilitydifference
significant predictedinby FDS
the time and CFASTthe
reaching exhibits
criticalsimilar
value istrends,
observed andunder
a significant
the
slow condition. This indicates the need to examine the applicability in various computa-
tional domains, rather than a lack of predictive accuracy for a specific fire growth speed.
 Therefore, visibility does not decrease at the initial stages of a relatively low smoke con-
centration but does as the fire grows and a sufficient smoke concentration is generated.
Although the initial reduction in visibility differs, the trend is very similar, with FDS
reaching the critical value in 50, 84,151, and 281 s and CFAST reaching it in 58, 88, 135, and
Fire 2024, 7, 108 212 s. The delayed reduction in visibility under the slow condition of FDS is a ributed10 toof 13
fluctuations caused by fresh air entering at the boundary condition of the exit. Owing to
an increase in the HRR, the visibility predicted by FDS and CFAST exhibits similar trends,
and a significant
difference difference
in the time reachingin the
the critical
time reaching
value isthe critical value
observed underisthe
observed under theThis
slow condition.
slow condition.
indicates the need This indicates the
to examine the applicability
need to examine the applicability
in various in various
computational computa-
domains, rather
tional
than domains,
a lack rather than
of predictive a lack for
accuracy of predictive
a specificaccuracy for aspeed.
fire growth specific fire growth speed.

Figure9.9.Comparison
Figure Comparison of
of visibility
visibility using
usingFDS
FDSandandCFAST
CFAST regarding changes
regarding in the
changes fire fire
in the growth rate rate
growth
and
and maximumHRR.
maximum HRR.(a)
(a)Fire
Fire growth
growth rate;
rate; (b)
(b) HRR
HRRmax.
max .

Figure 10 quantitatively compares the times when the gas temperature and visibility
reach their critical values in FDS and CFAST, thereby examining the potential applicability
of CFAST. According to the life safety standards in Table 1, ASET is assessed using the time
required for the gas temperature to exceed 60 ◦ C and for the visibility to drop below 5 m at
a height of 1.8 m from the floor. Accordingly, this study applied a similar concept, with
the X-axis corresponding to how long it takes to reach the critical values in FDS and the
Y-axis corresponding to that for CFAST. Figure 10a shows that the ASET evaluated using
the gas temperatures predicted by CFAST is extremely similar to that of FDS. The time
difference for reaching the critical values predicted by FDS and CFAST is 20 s slower, with
a difference of 7% to 41% relative to the time predicted by the FDS. The time required to
reach 60 ◦ C under all conditions, regardless of the fire growth rate, can result in a difference
of 20 s compared to FDS. This phenomenon is attributed to the model limitations of the
CFAST, wherein the computational domain is divided between the fire source and exit
area, thereby fixing the ceiling jet flow at 0.1 m/s [20]. Despite these limitations, CFAST is
suitable for predicting gas temperatures considering its substantial advantage in terms of
its computational cost.
FDS predicts the smoke concentration at a localized location at a point, while CFAST
predicts the average smoke concentration of the upper layer in the exit zone. Figure 10b
shows that CFAST predicts a smaller ASET for visibility than FDS, which calculates the
smoke concentration at localized positions. In the slow, CFAST under-predicts by 70 s, and
in the ultrafast condition, it over-predicts by a maximum of 6 s. This indicates a –25% to
12% range in the time difference for reaching the critical value for visibility compared with
FDS. Such results do not indicate the limitations of accuracy for specific fire growth rates.
Rather, further examination in terms of applicability in different computational domains
will be conducted. Additionally, for CFAST, differences in the time needed to reach the
critical value of predicted visibility with changes in the fire growth rate can be observed.
This is attributed to the basic assumptions, such as the mass and energy conservation
laws, of the zone model in CFAST, thereby resulting in pressure differences between the
upper and lower layers in divided spaces [20]. Hence, faster fire growth rates lead to a
rapid transfer of smoke concentration in interested computational domains owing to the
higher pressure in each divided domain. The possibility of utilization was examined when
determining the ASET for the buildings using CFAST, which has a very low calculation
 the critical value of predicted visibility with changes in the fire growth rate can be ob-
served. This is a ributed to the basic assumptions, such as the mass and energy conserva-
tion laws, of the zone model in CFAST, thereby resulting in pressure differences between
the upper and lower layers in divided spaces [20]. Hence, faster fire growth rates lead to
Fire 2024, 7, 108 a rapid transfer of smoke concentration in interested computational domains owing to the 11 of 13
higher pressure in each divided domain. The possibility of utilization was examined when
determining the ASET for the buildings using CFAST, which has a very low calculation
cost.The
cost. Themethod
methodforforand
and ability
ability of
of evaluating
evaluatingthe
thetime
timeneeded
needed to to
reach thethe
reach critical values
critical values
for the gas temperature, chemical species concentration, and visibility through the
for the gas temperature, chemical species concentration, and visibility through the smoke smoke
concentrationininCFAST
concentration CFASTare
are expected
expected to
to be
be highly
highlybeneficial
beneficialfor
forengineers.
engineers.

Figure10.
Figure 10.Comparison
Comparison of of the
the times
timesneeded
neededtoto
reach thethe
reach critical value
critical by FDS
value and CFAST.
by FDS (a) Gas
and CFAST. (a) Gas
temperature;(b)
temperature; (b)Visibility.
Visibility.

4. Conclusions
This study examined the use of CFAST as part of a project to develop a program to
help fire safety managers establish evacuation plans in a building fire. We focused on the
two most important variables for assessing ASET in building fire risk assessment, namely,
visibility and gas temperature. The following are the main conclusions.
Generally, as fire modeling for a single floor of a building is performed, using CFAST
in setting the computational domain may involve dividing the area of interest and fire
source. Therefore, even if the area of interest is very close to the fire source, by dividing
it from the fire room, a fixed flow rate of 0.1 m/s for the gas temperature is evaluated for
potential use. Despite these limitations, CFAST is very useful, as it shows an excellent
predictive performance for the gas temperature compared with FDS.
The predicted results for the concentration of the chemical species and visibility were
over-predicted in CFAST compared to in the FDS. This is because of the zone model
limitation of CFAST, which divided the space into hotter upper and cooler lower layers,
contrary to FDS, which could predict the concentration at localized positions.
When comparing the times needed to reach the critical value for visibility, as de-
termined by the smoke concentration outputs from CFAST, differences were observed
depending on the fire growth rate. Overall, CFAST tended to under-predict visibility. This
is believed to be caused by the pressure difference between the upper and lower layers of
the smoke generated from the flame in each compartment, which is released by the zone
model, which is based on the basic assumption of the conservation of mass and energy.
Therefore, CFAST was sufficiently usable from a conservative perspective.
Based on these results, the potential use of the output values of CFAST, which requires
a significantly lower computational cost than FDS, for ASET evaluation factors was assessed.
The research results are expected to enhance the usability in the understanding of general
fire engineering technology and simple fire dynamics trends.

Author Contributions: Conceptualization, H.-Y.J.; methodology, H.-Y.J. and C.-H.H.; validation, H.-
Y.J. and C.-H.H.; formal analysis, H.-Y.J.; investigation, H.-Y.J. and C.-H.H.; resources, C.-H.H.; data
curation, H.-Y.J.; writing—original draft preparation, H.-Y.J.; writing—review and editing, C.-H.H.;
visualization, H.-Y.J.; supervision, C.-H.H.; project administration, C.-H.H.; funding acquisition,
C.-H.H. All authors have read and agreed to the published version of the manuscript.
Fire 2024, 7, 108 12 of 13

Funding: This paper was supported by the “National Fire Agency” R&D program (20016433) and
the Korea Agency for Infrastructure Technology Advancement (KAIA) grant funded by the Ministry
of Land, Infrastructure, and Transport (Grant RS-2022-00156237).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data used to support the findings of this study are available from
the corresponding author upon request.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. McGuire, J.H.; Tamura, G.T. Simple analysis of smoke-flow problems in high buildings. Fire Technol. 1975, 11, 15–22. [CrossRef]
2. McGrattan, K.; Hostikka, S.; McDermott, R.; Floyd, J.; Weinschenk, C.; Overholt, K. Fire Dynamics Simulator User’s Guide, NIST SP,
6th ed.; National Institute of Standards and Technology: Gaithersburg, MD, USA, 2015; Volume 1019.
3. Peacock, R.D.; Reneke, P.A.; Forney, G.P. CFAST. Consolidated Fire and Smoke Transport: User’s Guide, NIST TN 1889v2, 7th ed.;
National Institute of Standards and Technology: Gaithersburg, MD, USA, 2021; Volume 2.
4. Thunderhead Engineering. Pathfinder User Manual; Thunderhead Engineering: Manhattan, KS, USA, 2020.
5. IES Virtual Environment. Egress: Simulex User Guide; IES Virtual Environment: Glasgow, UK, 2015.
6. Galea, E.R.; Lawrence, P.J.; Gwynne, S.; Filippidis, L.; Blackshields, D.; Cooney, D. buildingEXODUS v6.3 User Guide; GUEL
Subsidiary of the University of Greenwich: London, UK, 2017.
7. Korhonen, T.; Hostikka, S.; Heliovaara, S.; Ehtamo, H. ‘FDS+ Evac: An Agent Based Fire Evacuation Model.’ Pedestrian and Evacuation
Dynamics 2008; Springer: Berlin/Heidelberg, Germany, 2010.
8. Jang, H.Y.; Hwang, C.H. Revision of the input parameters for the prediction models of smoke detectors based on the FDS. Fire Sci.
Eng. 2017, 31, 44–51. [CrossRef]
9. National Fire Agency of Republic of Korea. Performance Based Design Methods and Standards on Fire-Fighting System Installation Act;
National Fire Agency of Republic of Korea: Sejong, Republic of Korea, 2016; Volume 3-a, Annex 1.
10. Ahn, S.H. Study on the Improvement of Performance-Based Design (PBD) through the Analysis of the Whole PBDs. Ph.D.
Dissertation, The Graduate School of Daejeon University, Daejeon, Republic of Korea, 2018.
11. Hsieh, H.R.; Chung, H.C.; Kawabata, N.; Seike, M.; Hasegawa, M.; Chien, S.W.; Shen, T.S. Assessment Method Integrating
Visibility and Toxic Gas for Road Tunnel Fires Using 2D Maps for Identifying Risks in the Smoke Environment. Fire 2023, 6, 173.
[CrossRef]
12. McGrattan, K.; Hostikka, S.; McDermott, R.; Floyd, J.; Weinschenk, C.; Overholt, K. Fire Dynamic Simulator; Technical Reference
Guide, NIST SP, 6th ed.; National Institute of Standards and Technology: Gaithersburg, MD, USA, 2015; Volume 1018-1.
13. Brohez, S.; Caravita, I. Fire induced pressure in airthigh houses: Experiments and FDS validation. Fire Saf. J. 2020, 114, 103008.
[CrossRef]
14. NRC; EPRI. Verification and Validation of Selected Fire Models for Nuclear Power Plant Applications; NUREG-1824; EPRI 1011999 [Final
Report]; United States Nuclear Regulatory Commission: Washington, DC, USA, 2007.
15. Mun, S.Y.; Cho, J.H.; Hwang, C.H. Effects of external heat flux and exhaust flow rate on CO and soot yields of acrylic in a cone
calorimeter. Appl. Sci. 2021, 11, 5942. [CrossRef]
16. Acherar, L.; Wang, H.Y.; Coudour, B.; Garo, J.P. Assessment of Semi-Empirical Soot Modelling in Turbulent Buoyant Pool Fires
from Various Fuels. Thermo 2023, 3, 424–442. [CrossRef]
17. Kaczorek-Chrobak, K.; Fangrat, J. PVC-based copper electric wires under various fire conditions: Toxicity of fire effluents.
Materials 2020, 13, 1111. [CrossRef] [PubMed]
18. Jang, H.-Y.; Hwang, C.-H. Obscuration threshold database construction of smoke detectors for various combustibles. Sensors 2020,
20, 6272. [CrossRef] [PubMed]
19. Yoon, G.Y.; Han, H.S.; Mun, S.Y.; Park, C.H.; Hwang, C.H. DB construction of activation temperature and response time index for
domestic fixed-temperature heat detectors in ceiling jet flow. Fire Sci. Eng. 2020, 34, 35–42. [CrossRef]
20. Peacock, R.D.; McGrattan, K.; Forney, G.P.; Reneke, P.A. CFAST Consolidated Fire and Smoke Transport: Technical Reference Guide,
NIST TN 1889v1, 7th ed.; National Institute of Standards and Technology: Gaithersburg, MD, USA, 2021; Volume 1.
21. Williamson, J.; Beyler, C.; Floyd, J. Validation of numerical simulations of compartment fires with forced or natural ventilation
using the fire and smoke simulator (FSSIM), CFAST and FDS. Fire Saf. Sci. 2011, 10, 1277–1288. [CrossRef]
22. Floyd, J. Comparison of CFAST and FDS for Fire Simulation with the HDR T51 and T52 Tests; United States Department of Commerce,
Technology Administration, National Institute of Standards and Technology: Gaithersburg, MD, USA, 2002.
23. Patterson, N.M. Assessing the Feasibility of Reducing the Grid Resolution in FDS Field Modelling; University of Canterbury:
Christchurch, New Zealand, 2002.
24. Oliva, D.; Somerkoski, B.; Tarkkanen, K.; Lehto, A.; Luimula, M. Virtual reality as a communication tool for fire safety—
Experiences from the VirPa project. In Proceedings of the GamiFIN Conference 2019, Levi, Finland, 8–10 April 2019;
pp. 241–252.
Fire 2024, 7, 108 13 of 13

25. Yun, H.S.; Hwang, C.H.; Choi, J.H. Evaluation of Predictive Performance of CFAST for Fire Characteristics in Single and
Multi-storey Buildings. Trans. Korean Soc. Mech. Eng. B 2022, 46, 441–451. [CrossRef]
26. Jang, H.-Y.; Hwang, C.-H. Configuration approaches of CFAST for prediction of smoke and heat detector activation times in
corridor fires. Appl. Sci. 2023, 13, 13161. [CrossRef]
27. National Fire Protection Association. NFPA 92 Standard for Smoke Control System; National Fire Protection Association: Quincy,
MA, USA, 2012.
28. Tewarson, A. Generation of heat and chemical compounds in fires. In SFPE Handbook of Fire Protection Engineering, 3rd ed.;
DiNenno, P.J., Ed.; National Fire Protection Association: Quincy, MA, USA, 2002.
29. Mulholland, G.W.; Croarkin, C. Specific extinction coefficient of flame generated smoke. Fire Mater. 2000, 24, 227–230. [CrossRef]
30. Jin, T.; Yamada, T. Irritating effects of fire smoke on visibility. Fire Sci. Technol. 1985, 5, 79–90. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.