Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 71 (2014) 145 152

Study on Smoke Control Strategy in a High-rise Building Fire

Yuan Yua,b, Yan-yan Chua,b,*, Dong Lianga,b
a

Safety Engineering Research Center, Department of Engineering, Sun Yat-sen University, Guangzhou 510006, China
b
Guangdong Provincial Key Laboratory of Fire Science and Technology , Guangzhou 510006, China

Abstract

During the high-rise building fire, it is crucial to win the time for occupant evacuation. One of the possible methods is to
utilize HVAC operations to control/slow down smoke propagation on the fire floor. It is also possible to supply fresh air to
the path of the evacuation so that occupants will not breathe into the poison gases. HVAC operations can control heat
change and smoke conditions and change the fluid flow directionally so that it is possible to optimize the HVAC operations
to make the building safer during fire. In this paper, the effects of HVAC operations, air supply system and mechanical
smoke exhaust system will be studied. A smoke propagation model using the Large Eddy Simulation will be developed to
study the smoke propagation under different HVAC operations. Simulation results show the temperatures at the fire room
exit for different supply air quantities. Results also show that smoke propagation method is affected by the building
construction, air supply and smoke exhaust system. The smoke control strategy is investigated.

by Elsevier
Ltd.
access
under
CC BY-NC-ND
license. under responsibility of the Academic Committee
2014
2014 Published
The Authors.
Published
by Open
Elsevier
Ltd.
Selection
and peer-review
Peer-review
of ICPFFPE under
2013. responsibility of School of Engineering of Sun Yat-Sun University
Keywords: fire science, fire safety, smoke propagation, HVAC, large eddy simulation

Nomenclature
u, v, w

velocity in the direction of x, y, and z (m/s)

glutinosity coefficient of turbulence

D diffuse coefficient of component

Y
concentration of the component
M component mol mass;
Su , S v , and S w momentum equation source terms in x, y, and z directions
S Y component transport equation source
S q energy equation source.

* Corresponding author. Tel.: +86-020-39332927; fax: +86-020-39332927.

E-mail address: [email protected]

1877-7058 2014 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Peer-review under responsibility of School of Engineering of Sun Yat-Sun University
doi:10.1016/j.proeng.2014.04.021

146

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

1. Introduction
In the building fire, smoke is often the leading cause of fatalities. Smoke often travels a long distance from the fire floor
[1]. Some research on smoke propagation has been developed. Air curtain could be utilized for confinement of fire-induced
smoke and carbon monoxide transportation along channels [2]. Comparison results between louver system and chimney
system show that the chimney system provides a more stable smoke and temperature stratification inside the building and a
higher hot smoke layer above the floor [3]. Fire simulation results on the underground shopping mall show that the location
of the mechanical exhaust vents and mechanical exhaust rate can greatly impact the smoke extraction [4]. It is important to
impede smoke propagation in the building fire utilizing all kinds of systems such as HVAC system.
In order to create comfortable and safety work condition, modern intelligent buildings are always equipped with
advanced HVAC system with fully control. HVAC system makes air in the building flow directionally. HVAC operations
have been extensively studied in energy saving. Energy can be saved effectively if HVAC operations can be optimized with
the information of occupants. On the other hand, the effect of HVAC operations on the building safety has been less studied
during fire, fire smoke plume shape will be varied and some smoke moves to return air fan. It is clearly that HVAC
operations can influence the smoke propagation in the building and can effect on fire alarm system, smoke exhaust system
and evacuation system. It is potentially feasible to utilize HVAC to delay smoke propagation and help evacuation. A smoke
propagation model is developed to study the smoke propagation under different HVAC operations using the Large Eddy
Simulation. Simulation results are performed to study the temperatures at the fire room exit for different supply air
quantities. Results also show that smoke propagation method is affected by building construction, air supply and smoke
exhaust system and smoke control strategy is studied for building safety.

2. Problem Setup and Model Development

In the high-rise building, HVAC operations are usually fully controlled and operations are optimized for energy saving.
On the other hand, the building fire in the high-rise building is much more dangerous comparing to the fire in a small
building. Fire alarm and sprinkler systems are many times the last defence for the building fire. Since the evacuation of
occupants is extremely difficult, it is important to utilize all means to slow down the smoke propagation and supply fresh air
to the place with occupants.

Fig.1. Smoke Simulation Model.

A numerical model will be developed in this paper to study the feasibility of utilizing HVAC operations to enhance the
safety of the building. A typical high-rise building is made up by an office zone and a core. Elevators, equipment rooms and
evacuation staircase lie in the core. The office zone is usually divided into several compartments by the walls based on
tenements requirement. Every office zone has at least an exit to corridor. A typical layout is shown in Fig. 1. Building area
of the fire zone is set to be 2700 m2. Area of the office zone is about 2000 m2 and area of the central core is 700 m2.

147

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

Building height of the fire zone is about 4 m and clear height is 2.7 m. Locations of the walls and exits are also presented in
Fig. 1. Smoke extraction vents lies in office zone homogeneously as black areas in Fig. 1. VAV ventilators of HVAC are
made as air makeup vents as grey areas in Fig. 1. A fire is set at the upper right corner; we will study the effect of HVAC
operations on the smoke propagation and discuss the control strategy.

3. Simulation Mathematics Model

Smoke propagation during the fire belongs to buoyancy flow caused by heat in a complicated geometry. Smoke density
fluctuating could be ignored for simplified mathematics model of fire field. Smoke propagation in the fire floor can be
simulated by the Large Eddy Simulation. The fundamental movement control equations of fire smoke movements are
shown as follows:
Mass Equation:

r
+
t

(r u ) (r v)
+
+
x
y

(r w)
= 0
z

(1)

Momentum Equation:

( r u )
( r uu ) ( r uv)
( r uw)
+
+
+
= div( mgrad (V )) x
z
t
y

p
+ Su
x

(2)

( r u )
( r vu ) ( r vv)
( r vw)
+
+
+
= div( mgrad (V )) x
z
t
y

p
+ Sv
y

(3)

( r u )
( r wu ) ( r vw)
( r ww)
+
+
+
= div( mgrad (V )) x
z
t
y

p
+ Sz
z

(4)

Component Equation:

( r Yi )
( r Yu
( r Yi v)
( r Yi w)
i )
+
+
+
= div( r Dgrad (Yi )) + SY
x
z
t
y

(5)

( r T )
( r Tu) ( r Tv)
( r Tw)
+
+
+
= div( r Ggrad (Ti )) + Sq
x
z
t
y

(6)

Energy Equation:

State Equation of Gas:

p = Rr T (Y / M ) = Rr T (

Yco Yco2 Yc
+
+ )
28 44 12

(7)

4. Design of Fire Scenarios

To study the HVAC operations on fire propagation, we will setup different fire scenarios. In this paper, we will study
one case. Fire source is supposed as propane gas combustor and lie in the right top of the story. Fire source is in the floor
and the area is about 4 m2. Fire location is shown in Fig. 1. Office fire is supposed developing with the sprinklers action.
2

On the assumption that fire heat release develops according to t , heat release can be obtained based on Equation (8).

Q = 0.0469 t 2
The maximum fire power is set to be 10MW and fire develops to the maximum fire power at 180s.

(8)

148

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

5. Simulation Results
5.1. Discuss of Orthogonal Experiments
FDS from NIST is selected to study the smoke control parameters/strategies. Meshes are divided using the nonuniform partitioning method. Total number of mesh is 108h108 h15. Simulation time is 1800s. Time step is set based on
CFL value, which is sufficient for the stability of simulations.
Smoke propagation will be simulated when air supply and smoke exhaust system are at different conditions.
Parameters of HVAC system such as supply air quantity, supply vent height and distance from fire to supply vent effect can
cause different fire control results. We will develop a control strategy by changing air supply box parameters.
Location of fire room exit and fire source is shown in Fig. 1. Smoke propagation is simulated when door, HVAC and
smoke exhaust systems are at different operating states. All doors of evacuation staircase, air supply fans and smoke exhaust
fans are opened at the time of burning. Figs. 2(a) to 2(d) show the contour with temperature lower than 40 on the fire
floor at 100s, 200s, 800s and 1200s when the ratio of supply air quantity and smoke exhaust is set to be 0.8 m3 s-1.

(a)

(b)

(c)

(d)

Fig. 2. Contour of temperature is 40 in Fire Room (a) at 100 s, (b) at 200 s, (c) at 800s,(d) at 1200 s.

To determine the control strategy, it is important to understand the effect of various HVAC equipments on the smoke
propagation and sensitivity of HVAC operations. We will study three parameters in this paper, e.g., supply air quantity,
supply vent height and distance from fire to supply vent, and their effects. The orthogonal experiments are designed and
three factors and three grades are shown in Table 1. Orthogonal experiment simulation results are shown in Table 2.
Discussion on factor effectiveness is based on the average temperature of fire room exit at 400s.
Orthogonal experiment results show that supply air quantity, supply vent height and distance from fire to supply vent
can all effect temperature of fire room exit. But influence is complicated. Based on range analysis, the effect of supply air
quantity is more important than supply vent height, that is more important than distance from fire to supply vent.
Experimental results show that supply air quantity is the major factor and effect of supply vent height is not clear and need
to be further investigated.

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

149

Table 1. Orthogonal Experiment Factors and Grades

Factor Level

Supply Air Quantity

Q/ m3 s-1

Supply Vent Height

H/m

1
2
3

2
3
4

0
1.3
2.7

Distance from Fire to

Supply Vent
L/ m
35
25
15

Table 2. Orthogonal Experiments Simulation Results

Factors
Item

Supply Air Quantity

Q/ m3 s-1

Temperature of Exits
/

1
2
3
4
5
6
7
8
9
K1
K2
K3
Range
Factor force
Sequence

1(2)
1(2)
1(2)
2(3)
2(3)
2(3)
3(4)
3(4)
3(4)
36.1
34.2
29.8
6.3

1(0)
2(1.3)
3(2.7)
1(0)
2(1.3)
3(2.7)
1(0)
2(1.3)
3(2.7)
32.4
31.2
36.4
4

Experiment Results
Distance from Fire to Average Temperature
of Exits within 400 s
Supply Vent
L/ m
/
1(35)
37.1
2(25)
33.0
3(15)
38.1
2(25)
32.4
3(35)
33.5
1(15)
36.6
3(35)
27.1
1(15)
27.1
2(25)
34.6
33.6
33.3
33.1
0.5
3

5.2. Discussion of Moderate Supply Air Quantity

Temperatures of fire room exit supply at different supply air quantities are studied when the ratios of supply air
quantity and smoke exhaust are 0.4, 0.8, and 1.6. Temperature distributions at the exit at 100s, 400s, and 800s are shown in
Fig. 4. FDS simulation results show that the heat field in the fire room will change with the supply air quantity. From Fig.
4(a), it is shown that temperature at the exit is similar for different supply air quantities when fire has not developed entirely.
From Fig. 4(b) and 4(c), it is shown that temperature at the exit (See Fig. 1) is lower when the ratio of supply air quantity
and smoke exhaust is 0.8 than 0.4 because the supply air to fire room can slow down smoke downward motion. So supply
air can increase the smoke filling time and help people evacuation. It also shows that temperature at the exit is higher when
the ratio of supply air quantity and smoke exhaust is 1.6 than 0.8 because supply air quantity is too large to push smoke to
the exit area.
Simulation results show that temperatures at the fire room exit can be reduced by moderate supply air quantity and too
large supply air quantity may change the smoke flow to the exit and disturb people evacuation.

150

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

(a)

(b)

(c)
Fig. 3. Temperature Isotherm of Fire Room Exit aat 100 s, (b) at 400 s, (c) at 800 s.

5.3. Discuss of Fire Control Strategy on the Fire Floor

At the current practice, during the fire, fire source location can be detected by fire alarm system and then smoke
exhaust system of fire floor and neighbor floor will be turned on. Fresh air vents in fire floor will be turned on and Fresh air
vent in neighbor floor will be turned off. To enhance safety of the building, HVAC operations can be utilized to slow down
the smoke propagation into the place with occupants and to direct the fresh air to there. For rooms without fire in the fire
floor, HVAC system will be active to prevent smoke propagation.
Six experiments are designed to study how to operate HVAC system in the rooms without the fire on the fire floor.
Design of fire scenarios include only number 1 room, as shown in Fig. 1, start air supply and smoke exhaust system, only
number 2 room start air supply and smoke exhaust system, only number 3 room start air supply and smoke exhaust system,
only number 1 room start air supply without smoke exhaust system, only number 2 room start air supply system without
smoke exhaust and only number 3 room start air supply without smoke exhaust system. All doors in corridor are open.
Smoke visibility simulation results with different control strategy are shown in Fig. 4 and Fig. 5 at 400 s and 800 s.
Smoke propagate in corridor through door 1 (see Fig. 1). Air supply system and smoke exhaust system change smoke flow
direction in corridor.
In Fig. 4 and Fig. 5, (a) shows the visibility when only number 1 room start air supply and smoke exhaust system.
Smoke does not flow in core because of positive pressure draft in room 1. (b) shows the visibility when only number 2 room
start air supply and smoke exhaust system. Smoke propagation in core. (c) shows the visibility when only number 3 room

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

start air supply and smoke exhaust system. Smoke propagation in room 3 through door 3 Simulation results show that
turning on air supply and smoke exhaust system of some room can delay smoke propagation in corridor behind the room. So

turning on air supply and smoke exhaust system of room without occupants in fire floor can delay smoke propagation in
evacuation path.
(d) shows the visibility when only number 1 room start air supply without smoke exhausts system. (e) shows the
visibility when only number 2 room start air supply system without smoke exhaust and (f) shows the visibility when only
number 3 room start air supply without smoke exhaust system. Smoke does not propagation in room 3 through door 3.
Comparison between figure of turn on air supply system and smoke exhaust system (See Fig. 5 (c))with turn on air supply
system and turn off smoke exhaust system (See Fig. 5 (f))show that turning on HVAC system and smoke exhaust system at
the same time cannot delay smoke propagation in room without fire.
Smoke visibilities of Fig. 4 and Fig. 5 show that smoke propagation can be slowed down when HVAC systems in the
rooms without fire in the fire floor are turned on and smoke exhaust systems are turned off. Smoke propagation in fire floor
will be changed when air supply and smoke exhaust systems are in action.

(a)

(b)

(d)

(e)

(c)

(f)

Fig. 4. Smoke Visibilities with Different Room Control (a), (b), (c), (d), (e), (f) at 400 s.

(a)

(b)

(d)

(e)

(c)

(f)

Fig. 5. Smoke Visibilities with Different Room Control (a), (b), (c), (d), (e), (f) at 800 s.

6. Conclusion
Simulation cases results by FDS show what the major factor for smoke movement is and how supply air quantity and
smoke control strategy effect smoke propagation and help people evacuate.
(1) Orthogonal experiment results show that supply air quantity is the major factor.

151

152

Yuan Yu et al. / Procedia Engineering 71 (2014) 145 152

(2) Air supply quantity simulation results show that temperatures at the fire room exit can be reduced by moderate
supply air quantity and too large supply air quantity may change the smoke flow to the exit and disturb people evacuation.
(3) Smoke control strategy simulation results show that turning on HVAC system and smoke exhaust system at the
same time cannot delay smoke propagation in room without fire and turning on air supply and smoke exhaust system of
some room can delay smoke propagation in corridor behind the room.
Those research show that door controllers, air supply system and smoke exhaust system can be used to isolate the fire
area after occupants leave the area, contain fire and smoke, and make para-safety area safer and we can develop control
strategy using sensor readings to control door, air supply, smoke exhaust system. But smoke propagation in building
network is affected by building construction, natural supply air, supply air and smoke exhaust system, etc., which we will
study in future.

Acknowledgements
This work was supported by Guangdong Natural Science Foundation ( S2011040001483) and Guangdong Provincial
Key Laboratory of Fire Science and Technology (2010A06080101.)
References
[1]W.Z. Black, George W and Woodruff, Smoke movement in elevator shafts during a high-rise structural fire, Fire Safety Journal, 2008, Vol. 3, pp.
1~15.
[2]L.H. H, J.W. Zhou, R. Huo, W. Peng and H.B. Wang, Confinement of fire-induced smoke and carbon monoxide transportation by air curtain in
channels, 2007, Vol. 12, pp. 327~334.
[3]Z. Dong Chen and David Yung, Numerical Study of Two Air Intake Strategies for a New Fire Laboratory, Fire Protection Engineering, 2005, Vol. 17,
pp. 27~41.
[4] Yuan Shusheng and Zhang Jian, 2007, Large eddy simulations of fire smoke flow and control in an underground shopping mall, Journal of University
of Science and Technology of China, Vol. 37, pp. 61-69.
[5] Jong-Yoon Kim and Sung-Wook Yoon, 2006, A study on the smoke control characteristic of the longitudinally ventilated tunnel fire using PIV ,
Tunnelling and Underground Space Technology, Vol. 21, pp. 302.
[6] Dong-Ho Rie, Myung-Whan Hwang and Seong-Jung Kim, 2006, A study of optimal vent mode for the smoke control of subway station fire,
Tunnelling and Underground Space Technology, Vol. 21, pp. 300-301.
[7] Su M D, 1990, Algebraic modeling of large eddy simulation and its application, Science in China (Series A) Vol. 33, pp. 185195.
[8] McGrattan K B, 2001,Computational fluid dynamics and fire modeling,National Institute of Standards and Technology Vol. 21, pp. l50.
[9] Woodburn PJ and Britter RE, 1996, CFD simulation of a tunnel Part 1. Fire Safety, Vol. 26, pp. 35-62.
[10] Woodburn PJ and Britter RE, 1996, CFD simulation of a tunnel Part 2. Fire Safety Vol. 26, pp. 63-90.
[11] XU Liang, ZHAN G Heping , YANG Yun , et al, 2004, Preparatory study on fire design in performance-based design . Fire Science and
Technology, Vol. 23, pp. 129~132.
[12]J.Liu and X.Y.Lu, 2002, Application of Large Eddy Simulation to Smoke Movement, Fire Safety ScienceVol. 2, pp. 515.
[13] McGrattan K and Forney G, 2004, Fire Dynamics Simulator (Version 4), Technical Reference Guide [R]. NIST Special Publication, Vol. 2, pp.
10~18.