Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com
Available online at www.sciencedirect.com
Available online at www.sciencedirect.com
Availableonline
Available online at www.sciencedirect.com
www.sciencedirect.com
Available online at
at www.sciencedirect.com
Available online at www.sciencedirect.com
Procedia Engineering 00 (2017) 000–000
ScienceDirect
Procedia Engineering 00 (2017) 000–000
Procedia Engineering 00 (2017) 000–000
Procedia Engineering 00 (2017) 000–000
Procedia Engineering
Engineering00211 (2018)
(2017) 343–352
000–000
Procedia Engineering 00 (2017) 000–000
Procedia Engineering 00 (2017) 000–000
Effect of Platform Screen Doors on Mechanical Smoke
Effect
2017 of Platform
8th International Screen
Conference Doors
on Fire onandMechanical
Science Smoke
Fire Protection Engineering
Effect
Effect of Platform Screen Doors on Mechanical Smoke
Effect of Platform
of(on Exhaust
the Screen
in Subway
Development
Platform
Effect of Platform Screen Doors on
on Mechanical
DoorsStation
of Performance-based Fire Code) Smoke
Fire
Mechanical Smoke
Effect of Platform ExhaustScreen in Subway
Screen
DoorsStation
Doors
on Mechanical
on Fire
Mechanical
Smoke
Smoke
Exhaust
Exhaust Da-yan
in
in Subway
Subway
LI, Guo-qing
Station
Station
ZHU* Fire
Fire
Exhaust
Exhaust Da-yanin
in Subway
Subway
LI, Guo-qing Station
ZHU* Fire
Station Fire
Exhaust in
School of Safety Engineering,Da-yan
Subway
LI,
China University
School of Safety Engineering,Da-yan
Guo-qing
LI, Guo-qing Station Fire
ZHU* Xuzhou,
of Mining and Technology,
ZHU* Xuzhou,
221116, PR China
Da-yan
Da-yan LI,
LI, Guo-qing
China University
Guo-qing ZHU*
of Mining and Technology,
ZHU*
221116, PR China
School of Safety Engineering, China University of Mining and Technology, Xuzhou, 221116, PR China
Abstract Da-yan LI, Guo-qing ZHU*
School of Safety Engineering, China University of Mining and Technology, Xuzhou, 221116, PR China
School of Safety Engineering, China University of Mining and Technology, Xuzhou, 221116, PR China
School of Safety Engineering, China University of Mining and Technology, Xuzhou, 221116, PR China
Abstract
In subway station fires,School
Abstract of Safety
the most Engineering,
immediate threatChina University of
to passengers' lifeMining
is notand
theTechnology, Xuzhou,
direct exposure to221116,
fire, butPRthe
China
smoke inhalation with hot
Abstract
air and toxic gases.fires,
Thus,
In subway
Abstract station thethemost
study on efficient
immediate smoke
threat control is of
to passengers' lifetheoretical
is not the and practical
direct significance.
exposure to fire, butInthethis paper,inhalation
smoke the fire dynamics
with hot
Abstract
simulator
In
air subway (FDS)
and toxicstationisfires,
gases. usedthe
Thus, tothemost
study the
oneffects
immediate
study ofsmoke
open
threat
efficient mode
controlofisplatform
to passengers' screen
lifetheoretical
of is not doors
the and
direct (PSDs)
exposure
practical on mechanical
to fire,
significance. butInthe smoke
this paper,exhaust
the fireindynamics
inhalation subway
with hot
In
air subway
Abstract
station
and
simulatorfire.
toxicstation
The
gases.
(FDS) isfires,
fires is
Thus,
used the
to most
located
the onimmediate
study
study the
onmiddle threat
of
efficient
effects of thetoplatform.
smoke
open passengers'
control
mode And
ofis life
ofthe isresults
not the
theoretical
platform screen direct
indicate
and exposure
that open
practical
doors to fire,
platform
significance.
(PSDs) on butIn
mechanical the
screen smoke
thisdoors inhalation
on
paper,
smoke both
the
exhaustfirein with
sides of hot
the
dynamics
subway
In
air subway
and toxicstation
gases. fires,
Thus, thethemost immediate
study on threat
efficient to passengers'
smoke control is lifetheoretical
of is not the and
direct exposure
practical to fire, butInthe
significance. smoke
this inhalation
paper, the fire with hot
dynamics
In subway
platform,
simulator
station station
can not
(FDS)
fire. The isfires,
only
used
firesThus, the
reduce
to
is locatedmost
the
study immediate
gas
the effects
on onmiddle threat
temperature,
of to
but
open
of smoke passengers'
also
mode
the platform.improve
of
And life
platform
the is
the not the
screen direct
efficiency of
doors
results indicate exposure
mechanical
(PSDs)
that open to
on fire,
smoke but
mechanical
platform screen the
exhaust,smoke
smoke
doors inhalation
which are
exhaust
onthe
both with
of hot
beneficial
in subway
sides to
the
air and
simulatortoxic gases.
(FDS) isonly
used the study
tothemost
study the efficient
effects ofsmoke
open control
mode ofis of theoretical and practical significance. In this paper, fire dynamics
air
In andevacuation.
safety
station
platform,toxic
subway
fire.
can gases.
station
The
notfiresThus,
fires,
is the
located
reduce study
on
the ontemperature,
the
gas efficient
immediate
middle threat
of theto control
platform.
but also isplatform
passengers'
And
improveof
life
the is screen
theoretical
not
results
the the doors
and
direct
indicate
efficiency of (PSDs)
practical
exposure
that open
mechanicalon
to mechanical
significance.
fire,
platform
smoke butInthe
screen smoke
this
exhaust, doors exhaust
paper,
smoke the fire
inhalation
on
which both
are indynamics
subway
with
sides of
beneficial hot
the
to
simulator
station (FDS)
fire. The is used to study the middle
effects of
of the
open mode of platform screen doorsthat
(PSDs) on mechanical smoke exhaust in subway
simulator
air
platform,
safety
station can
fire. notfires
(FDS)
andevacuation.
toxic The
isonly
gases.
fires
isreduce
used
Thus,
is
located
located
on gas
tothestudy
study
the
on
the
the
effects
middle
ofsmoke
ontemperature,
efficient
of
open
the
platform.
but mode
controlAnd
of
also improve
platform. And ofthe
isplatform
the
results
the
results
indicate
screen
theoretical doors
and
efficiency open
(PSDs)
practical
of
indicate that
platform
on
significance.
mechanical
open smoke
platform
screen
mechanical
In this
exhaust,
screen
doors
smoke on
paper,
which
doors on
both
exhaust
the
are sides
fire
both
of the
indynamics
subway
beneficial
sides of to
the
platform,
station
simulator
safety can
fire. notfires
The
(FDS)
evacuation. only
is isreduce
used located
to the
on gas
study the temperature,
middle
effects of
of thebut
open also improve
platform.
mode And
of thethe
platform efficiency
results of that
indicate
screen doors mechanical
open
(PSDs) smoke
platform
on exhaust,
screen
mechanical which
doors
smoke on are
bothbeneficial
exhaust sides
in to
of the
subway
© 2017
platform, The Authors. Published by Elsevier Ltd. Selection and peer-review under responsibility of the
can not only reduce the gas temperature, but also improve the efficiency of mechanical smoke exhaust, which are beneficial to Academic Committee
safety
station evacuation.
platform, can
fire. notfires
The onlyisreduce
located the
on gas
the temperature,
middle of thebut also improve
platform. And the the efficiency
results of that
indicate mechanical
open smoke screen
platform exhaust, which
doors on are
bothbeneficial
sides of to
the
of ICFSFPE
© 2017
safety 2017. Published by Elsevier Ltd. Selection and peer-review under responsibility of the Academic Committee
The Authors.
evacuation.
© 2018
safety
platform,
© 2017 The
canAuthors.
The not only Published
evacuation.
Authors. reduce the by
Published gas
byElsevier
ElsevierLtd.
temperature, butSelection
Ltd. also improveand the efficiency of
peer-review mechanical
under smoke exhaust,
responsibility of the which are beneficial
Academic Committee to
of ICFSFPEunder
Peer-review 2017.responsibility of the organizing committee of ICFSFPE 2017.
© 2017
safety
of The Authors.
2017. Published by Elsevier Ltd. Selection and peer-review under responsibility of the Academic Committee
evacuation.
ICFSFPE
© 2017 The
Keywords: Authors.
subway Published
station fire, platform by Elsevier
screen Ltd. Selection
doors, mechanical smoke and peer-review
exhaust under responsibility of the Academic Committee
of ICFSFPE
© 2017 2017. Published by Elsevier Ltd. Selection and peer-review
The Authors. under responsibility of the Academic Committee
of ICFSFPE
Keywords: 2017.
subway station fire, platform screen doors, mechanical smoke exhaust
of ICFSFPE
© 2017 The
Keywords: 2017.
Authors.
subway Published
station fire, platform by Elsevier
screen Ltd. Selection
doors, mechanical smoke and peer-review
exhaust under responsibility of the Academic Committee
of ICFSFPE
Keywords: 2017.
subway
Nomenclature station fire, platform screen doors, mechanical smoke exhaust
Keywords: subway station fire, platform screen doors, mechanical smoke exhaust
Keywords: subway station fire, platform screen doors, mechanical smoke exhaust
Nomenclature
bKeywords: plume radiusfire,
subway station
Nomenclature (m)platform screen doors, mechanical smoke exhaust
Nomenclature
bcp plume radius
specific heat at(m)constant pressure ( kJ / (kg K) )
bNomenclature
Nomenclature
c plume radius (m)
gb p specific heat
acceleration
plume radiusdue at
(m)constant
to gravity (m/s2) ( kJ / (kg K) )
pressure
cp
bNomenclature specific
plume heat at
radius (m)constant pressure2 ( kJ / (kg K) )
g
bcmp acceleration
plume
the plume
specific radius due
mass
heat (m)
at to gravity (m/s ) ( kJ / (kg K) )
flow ratepressure
constant (kg/s)
gbccmp specific
plume
the heat
acceleration
radius
plume massat
due
(m)constant
to
flowgravity
rate pressure
(m/s
(kg/s)
2 ( kJ / (kg K) )
)
gt p specific heat due
acceleration
required safe
at constant
to gravity
egress time (m/s22) ( kJ / (kg K) )
pressure
(s)
gcmsp the plume
acceleration
specific mass
heat due
at flow ratepressure
to gravity
constant (kg/s)
(m/s2) ( kJ / (kg K) )
gtm acceleration
the plumesafe
required due
mass to gravity
flow
egress rate (s)
time (m/s )
(kg/s)
gttm the plumetime mass of flow rate (s)
(kg/s)
s required the smoke layer2)fall to smoke screens (s)
mcs  acceleration
required
the plumesafe due
mass to gravity
egress
flow time (m/s
rate (kg/s)
tzmcs required safe
time egress
of the time
smoke (s)layer
(m)fall to smoke screens (s)
tt s  height
the plume
required
required
abovemass
safe
time
safe
the
egress
egress
point
flow source
rate
time
of the smoke
time
(kg/s)
(s)
(s)layer fall to smoke screens (s)
zt0c c
s virtual
required origin
height above (m)
timethe point
of the source
smoke (m)fall to smoke screens (s)
layer
zzD t0cs required
height safe
time
above egress
offuel
the the time
smoke
point (s)(m)
source layer
(m)fall to smoke screens (s)
tz c diameter
virtual
required
height
of the
origin
time
above
(m)
of
the the source
smoke
point source layer
(m) fall to smoke screens (s)
zzD
H t00c virtual
diameter
height
required origin
of smoke
of
above
time (m)
the screens
offuel
the the source
point (m)(m)
source
smoke (m)fall to smoke screens (s)
layer
zD virtual
height
diameter origin
above
of (m)
the
the point
fuel source
source (m)
zH Q 0 height
virtual
total of
energysmoke
origin (m) screens
release rate (m)(m)
(kW)
zD
H 0 diameter
virtual
height of
origin
above
height of of the
smoke(m)
thefuel source
point
screens source (m)
(m)(m)(m)
D Q0 diameter
total the fuel source
H
zD Q heightenergy
diameter
virtual of
origin
convective
release
smoke
of the
(m)
enegryfuel rate
screens
source
release
(kW)
(m) (m)(kW)
rate
H QC heightenergy
total of smoke screens
release rate (m)
(kW)
H
D Q heightenergy
diameter
total of of
convective smoke
the screens
fuel
release
enegry source
rate
release (m)
(kW)(m)(kW)
rate
TQ0C centerline
total energy temperature
release rate(K)(kW)
H Q height of
convective
total energy smoke
enegry screens
release
release rate(K) (m)
(kW) (kW)
rate
T
TQ0CC centerline
convective
ambient airtemperature
enegry release
temperature rate (kW)
(K)
TQ0C total energy
convective
centerline releaserelease
enegry
temperature rate(K)(kW)
rate (kW)
TQ
V0C convective
ambient
centerline air enegry release
temperature
the exhausttemperature
capacity of(K) rate (kW)
(K)
exhaust fan ((m3/s))
TQ centerline
ambient
convective airtemperature
temperature
enegry release(K)
(K)
rate (kW)
T0
VT 0
 C centerline
the
ambient airtemperature
exhaust capacity
temperature of(K)
exhaust
(K) fan ((m3/s))
VT
Greek

ambient
the
centerline
symbols air
exhaust temperature
capacity
temperature of (K)
exhaust
(K) fan ((m33/s))
VT0 ambient air capacity
temperature (K)

the exhaust of exhaust fan ((m3/s))
Greek
VT symbols
the exhaust
ambient air capacity
temperature of exhaust
(K) fan ((m3/s))
V
Greek the exhaust
symbols capacity of exhaust fan ((m /s))
Greek
V symbols
the exhaust capacity of exhaust fan ((m3/s))
Greek symbolsauthor. Tel.: +86-158-5249-6101; fax: +86-516-8359-0598.
* Corresponding
Greek symbols
E-mail address: grace_ldy@163.com
Greek symbolsauthor. Tel.: +86-158-5249-6101; fax: +86-516-8359-0598.
* Corresponding
* Corresponding
E-mail author. Tel.: +86-158-5249-6101; fax: +86-516-8359-0598.
address: grace_ldy@163.com
*1877-7058 © 2018
Corresponding
E-mail
The
author. Authors.
Tel.: Published by
+86-158-5249-6101;
address: grace_ldy@163.com fax:Elsevier Ltd.
+86-516-8359-0598.
* Corresponding author. Tel.: +86-158-5249-6101; fax: +86-516-8359-0598.
*Peer-review
E-mail address:under
Corresponding responsibility
grace_ldy@163.com
author. of the organizing
Tel.: +86-158-5249-6101; committee of ICFSFPE 2017
fax: +86-516-8359-0598.
E-mail address: grace_ldy@163.com
10.1016/j.proeng.2017.12.021
E-mail address: grace_ldy@163.com
* Corresponding author. Tel.: +86-158-5249-6101; fax: +86-516-8359-0598.
E-mail address: grace_ldy@163.com
 344 Da-yan
Da-yan LIProcedia
LI et al/ et al. / Procedia Engineering
Engineering 00 (2017)211 (2018) 343–352
000–000

 flue gas density (kg/m3)

0 plume velocity (m/s)

1. Introduction

Fire threat is a major concern in subway stations d ue to platform confinement and stack effect. With the
continuous improvement of operational safety requirements of city subway station, the platform screen doors (PSDs) almost
became the subway construction standards. The original intention of PSD is to protect passengers’ safety, they also have a
certain effect on subway station’s fire conditions. The most immediate threat to passengers' life is the smoke inhalation with
hot air and toxic gases in case of subway station fire. Thus, the study on effect of PSDs on smoke exhaust in subway station
fire has become one of the most essential research directions of fire safety science.
According to past research, many experiments were carried out to investigate smoke movement under various smoke-
control operating modes in platform of a subway station and innovative achievements were accomplished. The smoke
generator and heater are used for simulating the smoke movement at subway station fires. Many researches have been
carried out to predict smoke behavior or movement and related topics in underground space such as tunnel [1-4] and subway
system [5-7]. A Computational Fluid Dynamics (CFD) analysis was performed to demonstrate the capability of the
emergency ventilation system to control the smoke and heat generated by fire in a station equipped with PSDs [8]. Jae
Seong Roh et al. [9], performed fire simulation and evacuation simulation to estimate the effect of PSDs on passenger’s life
safety in a subway train fire. The results showed that the passengers in platform with PSDs have more available time than
passengers in case without PSDs in modeled subway station.
In this study, the fire dynamics simulator (FDS) is used to study the effect of open mode of PSDs on mechanical smoke
exhaust in subway station fire. The subway station is island-platform station with three stories below the ground. This paper
assumes that the fire is located on the middle of the platform and three calculation scenarios are designed.

2. Fire modeling

In order to estimate the effect of open mode of PSDs on mechanical smoke exhaust in subway station fire, simulations of
smoke movement in case of subway station fire are carried out by using the fire field model FDS, which was developed by
National Institute of Standards and Technology (NIST). The model has been verified by many domestic and foreign full-
scale fire experiments. The FDS describes fire-driven flows based on field model. The spatial distribution and changes over
time of status parameters were obtained by using partial differential equation. When the simulation runs, study area is
divided into a number of control volume, and they are solved by iteration of differential equations. The governing equations
for FDS code can be found in McGrattan [10].

2.1. Subway station geometry

The subway station chosen by the paper is island-platform station with three stories below the ground, which is 335 m
long, 94 m wide and 15 m high. Platform floor is in basement 3, station hall is in basement 2, and station exit and
commercial streets are in basement 1. Transfer hall and transfer channel are on the right side of station, the left side is
equipment space. There are four stairs and four escalators between platform floor and station hall, effective platform floor is
140 m long, 14 m wide and 4.5 m high. The plane models of station are shown in Figure 1.

(a)
 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352 345
Da-yan LI et al./ Procedia Engineering 00 (2017) 000–000

(b)

(c)

Figure 1 Plane view of (a) platform floor, (b) station hall and (c) basement one

2.2. Determination of smoke emission

In the smoke and smoke engineering design calculation, the fire smoke generation is an important design parameter to
determine the amount of smoke, which depends mainly on the size of plume mass flow rate. There are many models for
calculating the plume mass flow rate, such as the ideal plume model, the Zukoski plume model, the McCaffrey plume model
and the NFPA model. The Heskestad plume model is widely recognized internationally, and is adopted by the NFPA92B to
calculate the plume mass flow rate. Therefore, this paper chooses the Heskestad plume model to calculate the platform
plume mass flow.
Furthermore, Heskestad optimizes the ideal plume model by introducing the concept of virtual origin:
(1) The point source is relaxed by introducing a “virtual origin” at height ：

zo 0.083Q2 5  1.02D c (1)

Where D is the diameter of the fuel source in [m] and Q is the total energy release rate in [kW].
(2) Account wil be taken of the fact that spme plume properties depend on the convective enegry release rate, QC

QC  0.7Q (2)

(3) The “top hat” profiles across the plume for velocity and temperature will be replaced by Gaussian profiles。
(4) The Boussinesq approximation will be removed so that large density differences can be taken into account.
Plume radius b：

b 0.12 T0 T  ( z  z0 )
12
 (3)

Plume velocity 0 ：

g
u0 3.4( )1 3  Q1 3  ( z  z0 )1 3 (4)
 C T

Centerline temperature T0 ：

g   2 C 2
T0 9.1( ) 1 3  Qc2 3  ( z  z0 ) 1 3 (5)
T
 346 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352
Da-yan LI et al/ Procedia Engineering 00 (2017) 000–000

Where the convective energy release rate is give in [kW], the height above the point source (z) in [m], the temperatures in
[K], the heat capacity (of ambient air) in [ kJ / (kg k) ], and other properties of ambient air are given in SI units. The
ambient air properties are assumed to be T  293 K,   1.1 kg/m3 , cp  1.0 kJ / (kg k) and g = 9.81 m/s2. Also, observe
that the above equations are valid only above the mean flame hight (L)：

L 0.235Q2 5  1.02D (6)

Heskestad gave the following equations for the plume mass flow rates above and below the mean flame height：


 13 53 3
m 0.071Qc  ( z  z0 )  1.92 10 Qc (z>L)
 (7)
m  0.0056Qc z / L
 (z<L)

where the plume mass flow rate is give in [kg/s]

Figure 2 Schematic diagram of Heskestad plume model

To ensure that the smoke generated by the fire all excluded, the exhaust capacity of exhaust fan (V) should be greater
than the amount of flue gas production. If not consider the effects of smoke vertical wall, then:

V  mp /  (8)

where  is the flue gas density in [kg/m3] and it can be estimated by the formula as follow:

353
 (9)
T

According to the BSI DD240 recommended formula, the flue gas temperature can be determined by the following
formula:

980 z / Q 2/5  0.08


T 78.4( z / Q 2/5 ) 0.08  z / Q 2/5  0.20 (10)
 2/5 5/3 2/5
23.9  ( z / Q ) 0.20  z / Q

In order to improve the efficiency of mechanical smoke exhaust, smoke screens are usually added to increase the storage
space. Suppose the height of smoke screens is H, the required safe egress time is ts, the required time of the flue gas layer
fall to smoke screens is tc. To ensure the safety of human evacuation, the following requirements must meet.
 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352 347
Da-yan LI et al./ Procedia Engineering 00 (2017) 000–000

tc  t s (11)

HA
tc  ts (12)
mp
V


The effective length of the selected subway station is 140 m, the effective width of the platform is 14 m, and the effective
area of the public area is about 2000 m2 in this paper. According to the regulations of Subway Design Specifications: "Apart
from meeting the needs of upgoing and downgoing passengers, the amount of pedestrian staircase and escalator should be
verified by the principle that the accident evacuation time of platform level is not more than 6 minutes "[13] .This means
that the specification specifies the minimum required safe egress time at the platform level to be 6 minutes. The formula (8)
can be expressed as:

V  mp /   25H (13)

According to the analysis above, suppose the height of smoke screens is 1.0 m, the escape time is 6 min, the fire is station
luggage fire and the heat release rate is 721 kW based on the subway station geometry selected in this paper. So the smoke
extraction volume under the Heskestad plume model can be calculated and compared with the specification requirements.
The comparison results are shown in Table 1.

Table 1 Comparison of the smoke extraction volume getting from different methods

Calculation model Heskestad plume model Specification

smoke extraction volume (m3/h) 0.8 105 1.2 105

According to the comparison results in Table 1, it can be concluded that the smoke extraction volume calculated by
Heskestad model is less than the specified amount of smoke in specification. Therefore, consider the safety and air leakage,
this paper selected the smoke extraction volume calculated according to the specification.

2.3. Fire scenarios

This paper assumes that the fire occurs at the platform level and the platform luggage fire is selected as the fire form. It is
assumed that two suitcases catch fire at the same time for security reasons, so the maximum fire power is 721 kW. The FDS
geometry of the subway station is shown in Figure 3. The grid is made of a structured mesh with cubic cells of 1 cm.The fire
source is located between the two escalator openings in the platform layer, as shown in Figure 4. To study the smoke
propagation in subway station fires, three calculation scenarios are designed as shown in Table 2, the smoke volume is
determined as 1.2 105 m3 /h according to the calculation of the previous section. 18 smoke vent is built at Platform floor,
and nature ventilation is chosen as the ventilation ways at station hall.

Table 2 Scenario and parameter setting of fire in the middle of the platform

Fire scenario Fire form Maximum fire power /kW The open mode of PSDs
Fire scenario 1 Platform luggage fire 721 Close PSDs
Fire scenario 2 Platform luggage fire 721 Open PSDs on one left side of the platform
Fire scenario 3 Platform luggage fire 721 Open PSDs on both sides of the platform

Figure 3 The FDS geometry of the subway station

 Da-yan LI et al/ Procedia Engineering 00 (2017) 000–000
348 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352

Figure 4 Fire location in the middle of the platform

3. Results and discussion

3.1. Smoke propagation

Figure 5 shows the smoke propagation process of each fire scenario. It can be seen from Figure 5 (a) that when the
baggage in the middle of the platform is on fire, the smoke rises in the form of axisymmetric plume and reaches the ceiling
at 6 s. After reaching the ceiling, the smoke flows radially. At 34 s, the smoke plume change from two-dimensional
movement to one-dimensional movement under the action of the side wall. At 86 s, some smoke spread from the openings
of the escalator 2 and escalator 3 to the station floor layer. At 224 s, the smoke layer is stable. The smoke spread process in
fire scenario 2 and fire scenario 3 also follow this law, as shown in Figure 5 (b) and Figure 5 (c), but with different stages
time. The more PSDs are open, the longer time the smoke spread to the station floor cost, and the more convenient for
personnel evacuation.

(a) (b)

(c)

Figure 5 Smoke spreading process of (a) fire scenario 1, (b) fire scenario 2 and (c) fire scenario 3

3.2. Smoke layer height

Figure 6 shows smoke layer height history in each scenario. Smoke layer heights can be obtained by reasonable setting
relative optical parameters. It can be seen from Figure 6 that the smoke layer is rapidly declining at the beginning of the fire,
and the descending rate of the smoke layer is almost the same regardless of whether the PSDs are opened or not. With the
continuous combustion of the fire source, the smoke layer height decreased to about 3 m from the ground in fire scenario 1,
 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352 349
Da-yan LI et al./ Procedia Engineering 00 (2017) 000–000

about 3.3 m from the ground in fire scenario 2, and about 3 m from the ground in fire scenario 3. Therefore, the opening of
the PSDs can effectively reduce the thickness of smoke layer.

Figure 6 Smoke layer height history in each scenario

3.3. Ceiling temperature

When the smoke layer is stable, the ceiling temperatures of three fire scenarios are compared and the results are shown in
Figure 7. The maximum smoke temperatures in the fire scenario 1 to 3 are 165°C, 113°C and 93°C respectively. It can be
observed that the platform ceiling temperature in fire scenario 1 and fire scenario 3 are obvious symmetrical, and the
maximum smoke temperature of the ceiling is not right above the fire source, but on the left side of the fire source. In fire
scenario 2, the ceiling temperature on the right side of fire source is slightly higher than the left side. The highest smoke
temperature of ceiling is located in the right side of the fire source. In addition, the right side of smoke temperature curve in
fire scenario 2 is above the right side of smoke temperature curve in the fire scene 1, indicating that the smoke will
accumulate to the right side of ceiling when PSDs on the left side of the platform are open. The temperature distribution of
ceiling smoke is shown in Figure 8. It shows that the more PSDs are opened, the more area where the smoke temperature
exceeds 60 °C. When open PSDs only on the left side of the platform, the smoke exceeding 60 °C is concentrated on the
close side. This is because when open PSDs only on the left side of the platform, the air flow is unstable, and the smoke
generated a certain tilt, resulting in the ceiling smoke temperature is higher than temperature in case the PSDs are not
opened. Overall, when the platform catches fire, open the PSDs can effectively reduce the ceiling smoke temperature, but in
order to prevent the smoke and heat generated by fire accumulating to one side of ceiling, should also open PSDs on both
sides of platform.

Figure 7 The ceiling smoke temperature at different distance from Figure 8 The temperature distribution of ceiling smoke
fire source in each scenario
 350 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352
Da-yan LI et al/ Procedia Engineering 00 (2017) 000–000

3.4. Vertical smoke temperature

Figure 9 shows the temperature distribution of vertical smoke. The change of smoke temperature above the fire source is
basically the same and the smoke temperature decreases gradually with the increase of the height from the ground. The
vertical smoke temperature 0.5 m above the fire source in scenario 2 is highest, and lowest in scenario 3. The fire plume
temperature would increase when the PSDs on the one side of the platform are open, indicating that the fire is fully burning.
The fire plume temperature would decrease when the PSDs on both sides of the platform are open, which is beneficial to
safety evacuation.
The distribution of vertical smoke temperature about 2 m from the right side of fire source in fire scenario 2 is distinct
from fire scenario 1 and fire scenario 3, as shown in Figure 9 (b). The vertical smoke temperature in fire scene scenario 2
increases as the distance from the ground increases, and it is greater than that in the other two scenarios. However, the
vertical smoke temperature, about 2 m from the right side of fire source in fire scenario 2, is basically lower than that in the
other two scenarios, as illustrated in Figure 9 (d). This is further proof that only open the PSDs on the one side of platform
will destroy the smoke stratification law. When the distance from the ground increases, the vertical smoke temperature in
fire scenario 1 and fire scenario 3 decrease first, then increase as shown in Figure 9 (b, d). That occurs since the influence of
flame on measuring point, which is near the fire source, is more intense compared to smoke. With the increasing of height,
the measuring points will be mainly affected by the thermal radiation of smoke, while the temperature of smoke is lower
than the temperature of the flame. Figure 9 (c, e) show the distribution of vertical smoke temperature about 4 m from the
fire source in each scenario. The temperature is mainly affected by the smoke layer and increases as the distance from the
ground increases.

(a) The vertical smoke temperature above the fire source

(b) 2 m on the right side of fire source (c) 4 m on the right side of fire source
 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352 351
Da-yan LI et al./ Procedia Engineering 00 (2017) 000–000

(d) 2 m on the left side of fire source (e) 4 m on the left side of fire source

Figure 9 The temperature distribution of vertical smoke

3.5. Wind velocity at the escalator

The wind speed curves at each escalator are shown in Figure 10. The wind velocity at the escalator 1 is negative,
illustrating the fact that there is a downward wind to prevent the spread of smoke from the platform to the station hall. And
it is observed that the more PSDs are opened, the smaller the wind velocity is when reaching a steady state. The wind
velocity at the escalator 2, which is close to the fires, is negative in the early stages of the fire. However, a change of wind
occurred due to the rise of smoke temperature. The change time of winds in fire scenario 1 to fire scenario 3 are 45 s, 96 s
and 152 s respectively. The smoke generated by the station fire may spread to the station hall after the wind changed, and
affect the safe egress of the occupants. The escalator 3 is also close to the fire source and there is a change of the wind
direction in fire scenario 1 and scenario 2. Nevertheless, when the PSDs on both sides of platform are open, the wind
velocity at the escalator 3 is negative. The wind velocity at the escalator 4 is always negative and smaller when the PSDs on
both sides of platform are open.
Therefore, when the platform is on fire, the wind velocity at the escalator near the fire source will change its direction
and may carry smoke to the station hall. However, it will increase the time of change after opening the PSDs. The more the
number of PSDs open, the longer the time required for change. The wind velocity at the escalator away from the fire source
is always negative, and open the PSDs will reduce the wind velocity. However, it can be seen from the analysis of the
smoke diffusion process that although the wind speed at the escalator becomes smaller, the time of the smoke spreading to
the station floor becomes longer, so that the critical wind velocity of smoke spread from the platform to station hall can be
reduced.

(a) (b)
 352 Da-yan LI et al. / Procedia Engineering 211 (2018) 343–352
Da-yan LI et al/ Procedia Engineering 00 (2017) 000–000

(d)
(c)

Figure 10 Wind velocity at the (a) escalator 1, (b) escalator 2, (c) escalator 3 and escalator 4

4. Conclusions

This paper studied the effect of open mode of PSDs on mechanical smoke exhaust in subway station fire. The simulation
results show that when the fire source is located in the middle of the platform, the air flow rate at the opening of the
escalator is reduced when the platform screen door on the both side of platform is opened, but still can extend the time of
the smoke spread to the station hall, and effectively reduce the height of the smoke layer. When the PSDs on the one side of
the platform is open, the fire smoke will gathered at the platform side, leading to smoke temperature on the side of the roof
are higher than temperature in case the PSDs are not opened. In brief, open both sides of PSDs, can not only reduce the
ceiling smoke temperature and the vertical smoke temperature, but improve the efficiency of mechanical smoke exhaust,
which are beneficial to safety evacuation. The results give the best open mode of PSDs on mechanical smoke exhaust which
will provide reference for designer.

Acknowledgements

This work was supported by Natural Science Foundation of Jiangsu Province (BK20160270), National Key Research,
Development Plan (2016YFC0802907), the Priority Academic Program Development of Jiangsu Higher Education
Institutions (PAPD) and Postgraduate Research and Innovation Plan Project in Jiangsu Province (KYLX16_0575).

References
[1] Oka Y., Atkinson GT., 1995. Control of smoke flow in tunnel fires, Fire Safety Journal 25, p. 305.
[2] Y. Wu, M. Z. A. Bakar, 2000. Control of smoke flow in tunnel fires using longitudinal ventilation systems - a study of the critical velocity, Fire Safety
Journal 35, p. 363.
[3] Roh J S, Ryou H S, Kim D H, 2007. Critical velocity and burning rate in pool fire during longitudinal ventilation, Tunnelling & Underground Space
Technology Incorporating Trenchless Technology Research 22, p. 262.
[4] Roh J S, Yang S S, Hong S R, 2007. Tunnel Fires: Experiments on Critical Velocity and Burning Rate in Pool Fire During Longitudinal Ventilation,
Journal of Fire Sciences 25, p. 161.
[5] C. H. Cha, J. K. Kim, 1999. Smoke control in subway tunnels, Korean Journal of Air-Conditioning and Refrigeration Engineering 28, p. 425.
[6] Daniel Gabay, 2004. Compared fire safety features for metro tunnels, Safe & Reliable Tunnels. Innovative European Achievements First International
Symposium, 4-6 February, Prague, Czech Republic.
[7] S. R. Lee and H. S. Ryou, 2004. An experimental study of the effect of the aspect ratio on the critical velocity in longitudinal ventilation tunnel fires,
Journal of Fire Sciences, 23, p. 119.
[8] Miclea P C, McKinney D M, 2000. CFD simulation of a vehicle fire in a station equipped with platform screen doors[C]//BHR GROUP
CONFERENCE SERIES PUBLICATION. Bury St. Edmunds; Professional Engineering Publishing, 43, p. 305
[9] Roh J S, Ryou H S, Yoon S W, 2010. The effect of PSD on life safety in subway station fire, Journal of Mechanical Science and Technology, 24, p. 937.
[10] McGrattan K, Klein B, Hostikka S, 2009. Fire dynamics simulator (Version 5): user’s guide (NIST special publication 1019-5). Washington: NIST.