Chapter 15

Ventilation and fire safety for high-rise buildings

Dahai Qi1

This chapter presents the most recent progress in high-rise building design and
control methods for achieving energy efficiency and fire safety. First, basic
knowledge of high-rise ventilation is introduced. Then, the challenges of designing
and controlling high-rise ventilation are pointed out. State-of-the-art energy effi-
ciency and safety research on the modeling, control, and design of high-rise ven-
tilation is also presented. Lastly, two case studies on high-rise fire smoke control
and atrium fire smoke control are introduced and discussed. Readers will come
away from this chapter with an understanding of the theory of high-rise ventilation
and the design challenges related to fire safety concerns, as well as a knowledge
about designing safe and energy efficient high-rise ventilation systems.

15.1 Background
The National Fire Protection Association (NFPA) (NFPA, 2018) defines high-rise build-
ings as buildings higher than 23 m or seven stories. Due to the limited land supply and
high population density of cities, high-rise buildings have become increasingly popular,
particularly, during periods of economic growth. More than 100 high-rise buildings have
been erected in 142 cities and in 2019; among those the city with the most high-rise
buildings was New York with 6,034 (SkyscraperPage.com, 2019). Although many cities
already have a number of high-rise buildings, the number continues to increase, especially
in cities with fast economic growth. For example, at its peak, there were over 150 high-
rise buildings under construction in recent years in Toronto (DH Toronto Staff, 2018).
However, the rising demand for space cooling in high-rise buildings is putting
pressure on the electrical systems. Air conditioners and fans for space cooling consume
around 20% of the total electricity used in buildings around the world (OECD/IEA,
2018). High-rise buildings usually consume more energy due to their higher cooling
load than mid- and low-rise buildings. The high cooling load of high-rise buildings is
caused by the high internal heat gain (e.g. lights, equipment) and wide use of huge
glazed façades (Yuan et al., 2018). Therefore, to mitigate the strain of cooling energy
demand, it is essential to reduce the cooling load of high-rise buildings.

1
Department of Civil and Building Engineering, Université de Sherbrooke, Sherbrooke, Canada
448 Handbook of ventilation technology for the built environment

Building ventilation is an effective solution for maintaining indoor air quality,

cooling indoor spaces, and reducing building cooling loads, i.e. ventilative cooling
(IEA, 2018). Ventilation systems can be categorized as natural ventilation (NV),
mechanical ventilation (MV), and hybrid ventilation (HV) (also referred to as mixed-
mode ventilation). Previous studies have found that depending on the local climate and
building ventilation design, ventilation can reduce cooling-related energy consumption
by 56%–86% (Malkawi et al., 2016; Hu and Karava, 2014). There are more than
50 large cities with a significant ventilative cooling potential of more than 2,000 h/year
(Chen et al., 2017). In cold climates, such as Canada and Northern Europe, high-rise
ventilative cooling can be used for a long time throughout the year, not only during
shoulder seasons but also during the summer (Artmann et al., 2007). Furthermore, cold
climates have large diurnal temperature variations and relatively low nighttime outdoor
temperatures even in the summer. The characteristics of cold climates are more ben-
eficial to high-rise ventilative cooling than other climates. A high-rise building struc-
ture can be cooled during the night and becomes a huge heat sink during the daytime to
reduce cooling loads and thus reduces peak electricity demands.
However, fire safety concerns, associated with the stack effect in large vertical
spaces, are one of the big issues of the use of ventilation in high-rise buildings. In high-
rise buildings, cool fresh outdoor air can pass through floors, move upward through the
vertical spaces due to the stack effect, arrive at different floors to remove indoor heat, and
exit the buildings during NV and HV. The vertical spaces can be atria, stairwells, double-
skin façades, and elevator shafts, which are the critical structures of high-rise buildings.
During regular operations, many existing features and functions of these large spaces can
contribute to the stack effect and high-rise ventilation potentially providing a maximum
level of energy savings. However, during a fire outbreak, fire-generated smoke laden
with toxic gases can spread far from the origin of the fire deep throughout the building
through these large vertical spaces, endangering occupants, damaging property, and
creating challenges for firefighters. For example, it was reported that around
145,000 structure fires in high-rise buildings occurred per year from 2009 to 2013 in the
United States, causing an average of 40 deaths, 520 injuries, and $154 million in property
damage per year (Ahrens, 2016). The problem of smoke control in high-rise buildings
can be further complicated by interactions with dynamic weather conditions, including
variable winds, temperatures, and building ventilation system operations.
Knowing that energy efficient and safe high-rise buildings are essential for the
sustainable development of cities and society, this chapter provides state-of-the-art
high-rise ventilation design and fire smoke control methods, which take into con-
sideration the energy efficiency and building safety issues in high-rise ventilation,
including NV, MV, and HV.

15.2 Ventilation types

15.2.1 Natural ventilation
NV refers to the intentional introduction of air naturally passing through open
windows, doors, grilles, etc. (ASHRAE, 2017). Based on the driving forces, NV
 Ventilation and fire safety for high-rise buildings 449

can be categorized as buoyancy-driven NV, wind-driven NV, or wind- and

buoyancy-driven NV, which are driven by the pressure difference across the
building envelope caused by wind, stack effect, or both.
The indoor–outdoor air density difference caused by the indoor–outdoor tem-
perature difference generates the airflow movement of buoyancy (stack effect)
ventilation in high-rise buildings (Wood and Salib, 2013). Equation (15.1) can
calculate the pressure difference of the stack effect at any vertical location, while
neglecting the vertical density gradient.
 
Ti  T0
ps ¼ r0 gðHNPL  H Þ (15.1)
Ti
where ps is the stack pressure difference (Pa), r0 is the outdoor air density (kg/m3),
Ti and T0 are the absolute indoor and outdoor temperature (K), g is the gravitational
acceleration (m/s2), HNPL and H are the height of the neutral pressure level (NPL)
and the height above the reference plane respectively (m).
For wind-driven NV, the wind-driven force is dominant, which can be further
classified as single-side ventilation or cross ventilation. Single-side ventilation
refers to fresh air entering the room and exhausting through the same side. Cross
ventilation is the airflow between the windward side and leeward side of a build-
ing’s envelope. The relationship between wind pressure and wind speed is
demonstrated in (15.2) (ASHRAE, 2017):

U2
pw ¼ Cp r (15.2)
2
where pw is the windward pressure relative to the outdoor static pressure (Pa), r is
the outdoor air density (kg/m3), U is the wind speed (m/s), Cp is the wind surface
pressure coefficient, dimensionless. Cp can be decided by the geometry and loca-
tion of the building, as well as wind direction. The equation to calculate Cp can be
found in the 2001 ASHRAE Handbook—Fundamentals (Chapter 16). For high-rise
buildings, it should be noted that the wind speed increases parabolically as the
building height increases (Günel and Ilgin, 2014) and thus creates large wind
pressure differences along the building façade according to (15.2). Hence, when
designing wind-driven NV in high-rise buildings, the wind pressure variation along
the façade must be considered, because the high wind speed at the upper floors may
result in uncomfortable indoor air velocity and unacceptable wind-induced noise.
During the winter, the indoor temperature is higher than the outside temperature and
the airflow inside the building will rise driven by buoyancy. Due to the air density
differences of indoor and outdoor air, there exists an NPL, above which the indoor
pressure is higher than the outdoor pressure, so the air tends to exfiltrate to the outside.
The NPL can be located in the middle of the building (Figure 15.1(a)), or lower or higher
(Figure 15.1(b)), depending on the indoor and outdoor air temperatures, as well as the
building structures. During the transient seasons, when indoor and outdoor temperatures
are similar, the indoor and outdoor pressure profiles can be parallel, and the outdoor
windward pressure can be higher than the indoor pressure as shown in Figure 15.1(c).
450 Handbook of ventilation technology for the built environment

Inside
Height
Inside Neutral plane level

Height
Height
Neutral plane level

Outside
Outside Inside Windward

(a) Pressure (b) Pressure (c) Pressure

Figure 15.1 Pressure profile of a high-rise building. (a) Neutral plane is in the
middle of the building; (b) neutral plane is above the middle of the
building; (c) there is no neutral plane

Applying (15.1), it can be found that the pressure difference generated by

buoyancy depends on the height above the NPL. Hence, when using buoyancy-
driven ventilation for ventilative cooling, its cooling performance is usually highly
dependent on the building structure, i.e. the high vertical spaces, such as the ven-
tilation shaft, atrium, solar chimney, and double-skin façade (Sha and Qi, 2020a).
To improve buoyancy-driven ventilation, the design of the parameters of these
vertical architectural elements, e.g. the cross-sectional size, height, and the location
and size of the openings, should be optimized. For example, Moosavi et al. (2015)
recommended designing a large inlet to outlet opening ratio for achieving high
cooling performance in an atrium. Furthermore, to make full use of the vertical
space, these architectural elements can be combined together. For example, the
atrium, double-skin façade, and solar chimney can be integrated into one building
to generate a higher airflow rate (Ding et al., 2005).
When designing wind-induced ventilation, indoor airflow and airflow patterns
greatly impact its indoor heat removal performance. The characteristics of the
building, including the building positioning, floor planning, and building façade,
must be considered to optimize the indoor airflow rate and airflow patterns (Sha
and Qi, 2020a). The building positioning refers to the arrangement of buildings on
the ground surface according to their locations and dimensions. For example, the
location of the building should avoid the shelter effect caused by the surrounding
buildings or structures, which can influence the wind pressure distributions on a
building and reduce the airflow rate of NV (Chen et al., 2008). Floor planning is the
planning of floor partitions and door locations, etc. The reduction of partitions and
the creation of wind-path on the floor plan can enhance the airflow rate of NV
(Zhou et al., 2014). The elements on the building façade, e.g. balconies and
openings, also need to be optimized to achieve effective NV on cooling (Ai et al.,
2011; Wang et al., 2007).

15.2.2 Mechanical ventilation

MV refers to the air movement into or out of a building by using mechanical
equipment, such as fans, ductwork, and grilles (ASHRAE, 2017). The operation of
mechanical equipment affects pressure differences across the building envelope
 Ventilation and fire safety for high-rise buildings 451

and thus air change rates, which are essential for building energy efficiency and
safety (ASHRAE, 2017).
One of the approaches for designing the MV system to reduce the building
cooling load is mechanical night ventilation. Medved et al. (2014) found that the
energy savings of mechanical night ventilation can reach around 50% compared
with only using a mechanical cooling (MC) system. Zhang et al. (2018) investi-
gated the control of the air exchange rate (ACH) for night MV based on the
building energy simulation (BES). The results showed that the optimal ACH of
nighttime MV is higher than the ACH for maintaining indoor air quality, which
results in a 47% reduction in energy consumption over the entire summer.
However, an appropriate control strategy of the MV system must be used to
minimize cooling-energy consumption based on the chiller cooling and mechanical
ventilative cooling energy performances. Otherwise, mechanical ventilative cool-
ing may consume more energy than chiller cooling, because the fans in MV sys-
tems can consume a considerable amount of energy. This point can be seen in a
study conducted by Kolokotroni and Aronis (1999). The MV system is usually
designed for maintaining indoor air quality, especially in high-rise buildings. High
resistance due to the long ductwork spanning many floors and/or low fan efficiency
can generate high-energy consumption of the MV system.
With an appropriate control strategy, the energy efficiency and maximum flow
rate of fans are the key factors in determining the cooling performance of
mechanical ventilative cooling (Sha and Qi, 2020b). The fans should be energy
efficient to provide ventilative cooling, i.e. their specific fan power (SFP) must be
low enough. SFP is a parameter for quantifying the energy efficiency of fans, which
is defined as the electric power that is needed to transport one unit of air. A low
SFP requires low resistance of the ductwork and small energy loss on the belt or
motor in the fans. A higher maximum fan flow rate can provide more ventilative
cooling, but it is meaningless to design a very high fan flow rate because the
cooling demand is not always very high during the day. Therefore, an optimal
maximum fan flow rate should exist. In practice, a decentralized system that spans
fewer floors than the centralized system in high-rise buildings is recommended in
the design, because a decentralized system can more easily satisfy the fan flow rate
and energy efficiency requirements than the centralized system.

15.2.3 Hybrid ventilation

An HV system is defined as a system that provides a comfortable internal environ-
ment by making use of both NV and MV at different times (Heiselberg, 2002). HV
avoids the disadvantages of NV, such as uncertain performance caused by weather
conditions, and improves the degree of individual control of the indoor climate.
According to Heiselberg (2002) and Li and Heiselberg (2003), the types of HV
systems can be classified as an alternate use of NV and MV, fan-assisted NV, and
stack and wind-assisted MV. The alternate use of NV and MV means that the NV
and MV systems are independent, and a control strategy can switch between the NV
and MV systems. The fan-assisted NV system is an NV system combined with fans
452 Handbook of ventilation technology for the built environment

that are used during periods when the NV is weak. The stack and wind-assisted MV
refers to an MV system that uses stack and wind forces to reduce the need for fans.
The main challenge of HV systems is the control strategy and coordination of NV
and MV systems. The advanced HV system can save as much energy as possible while
maintaining the high indoor environmental performance requirements. Different control
methods of HV systems have been previously proposed. For example, Hu and Karava
(2014) reported that a model-predictive control strategy to control the opening of the NV
system in the HV system can effectively reduce the total cooling energy demand by
85%. Chen et al. (2018) proposed a control method based on the reinforcement learning
technique, which is a model free. The reinforcement learning control strategy can treat
the environment as an unknown black box, and the reinforcement learning algorithm can
learn from the interaction with the environment and find the optimal control decision.

15.3 Smoke control for high-rise fires

15.3.1 Pressurization system for stairwells
The pressurization system includes the vertical space and fans. By supplying suf-
ficient air into the vertical space, such as stairwells and elevator shafts, the inside
pressure can be high enough to prevent smoke spreading into the protected spaces.
If there is fire, the pressurized stairwells could provide a smoke-free safety route
for escape and firefighters. According to the number of air supply locations, the

Pressurization fan Pressurization fan

Duct

Duct shaft
Stairwell Stairwell

(a) (b)

Figure 15.2 (a) Stairwell pressurization with roof fan; (b) stairwell pressurization
by multiple injection with roof fan
 Ventilation and fire safety for high-rise buildings 453

pressurization can be categorized as a single-injection pressurization system where

the air is supplied at one floor (top or bottom of the stairwell, see Figure 15.2(a)), or a
multiple injection pressurization system where the air is supplied at multiple floors
(see Figure 15.2(b) and (c)). The weakness of the single-injection pressurization
system is that it may fail for spaces far from the air supply location, such as in tall
buildings. The single-injection pressurization system may also not work when the
doors near the air supply location are open, which reduces the pressure. Therefore, in
high-rise buildings, to maintain positive pressure in the tall stairwell, the multiple
injection pressurization system is often applied, which includes a duct with multiple
air supply openings and a fan. The fan can be mounted on the roof of the building or
at the ground level.

15.3.1.1 Minimum design pressure difference

To prevent smoke spread to the stairwell, the pressurization system should create
an adequate pressure difference across the stairwell door. The required minimum
pressure across the closed stairwell door, Dpmin , can be calculated by using
 
1 1
Dpmin ¼ DpSF þ 3;460h  (15.3)
To TF
where Dpmin is the minimum design pressure difference, Pa; DpSF is the pressure
difference safety factor, Pa; h is the distance above the neutral plane, m; To and TF
are the absolute surrounding temperature and hot gas temperature, K.
NFPA requires that stairwells should be pressurized to maintain a 0.10-in. water gauge
(25 Pa) across a closed stairwell door for a non-sprinklered building (see Table 15.1), based
on the assumption of a hot gas temperature of 927 C next to the stairwell door and a 0.03-
in. water gauge (7.5 Pa) pressure difference safety factor. Table 15.1 presents the required
minimum design pressure difference across smoke barriers, which depends on the ceiling
height and the building type (with or without sprinklers).

15.3.1.2 Maximum design pressure difference

Besides the minimum design pressure difference, there is also a maximum design
pressure difference, which is used to avoid pressurization systems creating forces
that are too great to open the door. The force required to open a side-hinged

Table 15.1 Minimum design pressure across smoke barriers

Building type Ceiling height in m (ft) Design pressure difference

in Pa (in.w.g.)
Sprinklered Any 12.5 (0.05)
Non-sprinklered 2.75 (9) 25 (0.10)
Non-sprinklered 4.56 (15) 35 (0.14)
Non-sprinklered 6.41 (21) 45 (0.18)
Note: The table presents minimum design pressure differences developed for a gas temperature of
1,700 F (927 C) next to the smoke barrier.
Source: NFPA (2012).
454 Handbook of ventilation technology for the built environment

swinging door needs to be great enough to overcome the door closer and the
pressure difference across the closed door, which can be expressed by (Klote and
Milke, 2002)
kd WADP
F ¼ Fdc þ (15.4)
2ðW  d Þ
where F is the total door opening force (N), Fdc is the force to overcome the door
closer (N), W is the door width (m), A is the door area (m2), DP is the pressure
difference across the door (Pa), d is the distance from the doorknob to the edge of
the knob side of the door (m), Kd is the coefficient (dimensionless).
The relationship between the pressure difference across the closed door, DP,
and the resultant force can be expressed as
2ðW  d ÞðF  Fdc Þ
DP ¼ (15.5)
kd WA
Figure 15.3 shows the relationship between the pressure difference and door
opening force to overcome the pressure difference for a door 2.13 m in height (H) and
a distance from the doorknob to the edge of the knob side of the door, d of 0.06 m. The
pressure difference across the closed door, caused by the pressurization system, cannot
be too high; otherwise, a person cannot open the stairwell door to escape through the
stairwell if there is a fire. NFPA 101 (life safety code) states that the force required to
open any door as a means of egress shall not exceed 133 N (30 lb). The force to
overcome a door closer is normally greater than 13 N. According to Figure 15.3, for a

180
2(W – d )FP
P= W = 0.81 m
KdWA
W = 0.91 m
150
W = 1.02 m
W = 1.12 m
Pressure difference (Pa)

120

90

60

30

0
0 30 60 90 120 150
FP (N)

Figure 15.3 Relationship between pressure difference across the door and the
force to overcome the pressure difference (H ¼ 2.13 m, d ¼ 0.06 m)
 Ventilation and fire safety for high-rise buildings 455

door 2.13 m in height and d ¼ 0.06 m, if the force to overcome a door closer is 43 N, to
satisfy the NFPA 101 requirement (total door opening force should be less than 133 N),
the pressure difference cannot be higher than 88 Pa.

15.3.2 Smoke ventilation

One of the primary concerns for the use of pressurization systems is the possible
excessive pressure difference across the closed door, leading to high forces required
to open the door and preventing occupants from being able to open them. This issue
should be carefully considered especially for tall buildings higher than 30 m (Lay,
2014). One of the alternative solutions is to use a dedicated smoke exhaust system
to exhaust smoke by stack effect and/or exhausting fans. This method would reduce
the risk of smoke infiltrating into the stairwell so that the pressurization system for
the stairwell could create a lower pressure difference across the closed door or not
even need to be in operation depending on the fire smoke control performance of
the whole system.
One or two dedicated shafts can be installed in the high-rise building to
exhaust smoke as indicated in Figure 15.4. Dampers are installed on each floor. If a
fire occurs on one of the floors, the damper on the fire floor opens and the smoke is
exhausted through the smoke shaft, while the dampers on the non-fire floors remain
closed so that the smoke will not infiltrate into the non-fire floors. The shaft can be
designed without fans, where the smoke is exhausted due to the stack effect, or with
fans to remove the smoke mechanically.
Estimating the smoke temperature inside the smoke shaft is helpful for
designing the smoke shaft and smoke exhaust fan at the top of the shaft for safe
operation.

Pressurization fan With or without fan Pressurization fan With or without fan
Dedicated shaft Dedicated shaft Damper

Fire floor Fire floor

Stairwell Stairwell

(a) (b)

Figure 15.4 (a) Smoke exhaust system with one dedicated shaft; (b) smoke
exhaust system with two dedicated shafts
456 Handbook of ventilation technology for the built environment

The smoke temperature inside the shaft can be expressed as

Tshx ¼ Tb þ Tf  Tb expðjaÞ (15.6)
where Tshx is the temperature profile along the vertical direction of the dedicated
shaft (x) where the smoke is exhausted ( C), Tf is the fire room temperature ( C),
Tb is the non-fire room temperature ( C).
j is relative height:
x
j¼ (15.7)
H
where H is the vertical distance between the fire floor and the top of the shaft (m).
a is the temperature attenuation coefficient:
PH
a¼ (15.8)
_ P Rt
mC
where P is the perimeter of the shaft (m), m_ is the mass flow rate (kg/s), CP is the
specific heat capacity of the smoke (J/kg K), Rt is the thermal resistance between
two sides of the shaft (m2 K/W).


TshH ¼ Tb þ Tf  Tb expðaÞ (15.9)
The EN 12101-3 Standard of the European Committee for Standardization
classifies smoke exhaust fans by temperature (EN 12101-3 Standard, 2002), e.g. the
F200 fan must resist 200 C for at least 2 h. The shaft has a perimeter of P ¼ 5 m
and a height of H ¼ 50 m, and a thermal resistance of Rt ¼ 0.224 (m2 K)/W. If the
EN 12101-3 F200 fan is used in the shaft, for an exhaust fan with a given flow rate
of 1.6 kg/s, the maximum fire would be Tf ¼ 374 C. This means that the fan would
withstand a fire smoke of 374 C. A solution could be to change for another fan with
a higher temperature rating or to add another F200 fan to share the smoke flow rate.
Therefore, (15.9) could be used in practice to evaluate the thermal performance of
mechanical exhaust fans installed at the top of the shaft.

15.4 Case studies

15.4.1 Evaluation of high-rise fire smoke control*
High-rise buildings comprise many complex structures, like atriums, stairwells, ele-
vator shafts, corridors, and compartments, which also leads to difficulties in designing
smoke control systems in these structures. The smoke control systems could interact
when they operate together. For example, the required fan speed of the stairwell and
elevator shaft-pressurization systems is strongly coupled (Miller and Beasley, 2009).

*
Section 15.4.1 is modified from the conference paper published by ASHRAE: Qi D, Soubra M,
Mashayekh S, Wang L. CFD Modeling of Full-size High-rise Fire Smoke Spread and Smoke Control.
ASHRAE 2017 Annual Conference, Long Beach, CA. 2017.
 Ventilation and fire safety for high-rise buildings 457

To design a reliable smoke control system, simulation techniques to predict smoke

movement and the performance of the smoke control system are often required. To
investigate smoke spread and control performance in high-rise buildings, the compu-
tational fluid dynamic (CFD) modeling is preferable. In this case study, the fire
dynamics simulator (FDS) is used, which is a widely used CFD tool in the prediction of
smoke spread inside high-rise buildings (McGrattan et al., 2013). FDS is a CFD model
based on large eddy simulation (LES) of turbulent flows, which was developed by the
US National Institute of Standards and Technology. The calculation of LES is much
more time-consuming than other CFD models based on the Reynolds-averaged
Navier–Stokes modeling of turbulences. In recent years, high-performance computing
(HPC), using clustered supercomputers, has been more frequently applied in many
CFD simulations. HPC can be used in fire dynamics simulation of whole-building fire
smoke spreads and controls in full-size high-rise buildings.
In this case study, a full-size 30-story tower with fire smoke spread was studied.
Different smoke control strategies were designed and modeled, including sprinklers only,
a smoke exhaust system, pressurization system and dedicated smoke exhaust shaft system,
and a new approach for high-rise smoke removal (Lay, 2014). To evaluate the smoke
control performance, the results of smoke/air temperature distributions, mass flow rates,
and pressure distributions in the stairwells were compared among different strategies.

15.4.1.1 Full-size high-rise fire smoke modeling

A 30-story high-rise building, 93 m in height, as shown in Figure 15.5(a), was
modeled. The bottom three floors are car parking, which is ignored in the

Window

Stairwell 1
Door

Fire
room
Window
Corridor

Elevator
shaft
Stairwell 2

Fifth f loor Smoke

shaft
(Case 5)

(a) (b)

Figure 15.5 Schematic of the building: (a) vertical schematic of the whole
building and (b) plan schematic of fifth floor
458 Handbook of ventilation technology for the built environment

Table 15.2 Designs of different fire smoke control cases

Cases Smoke control method Windows

(W (m)H (m))
Case 1 Sprinkler only, no fan on roof of stairwell 20.8
Case 2 Exhaust fan on the roof of Stairwell 2 20.8
Case 3 Pressurization fan on roof of Stairwell 1 21.6
Case 4 Pressurization fan on roof of two stairwells 21.6
Case 5 Dedicated smoke exhaust shaft with roof extraction fan 20.8

Note: In Cases 2–5, the volume flow rate of the fan is 20 m3/s.

simulation. The staircases in each tower do not span to the bottom four floors. The
building includes two stairwells, one elevator shaft, corridors, and residential rooms
(see Figure 15.5(b)). The door size is 1 m2 m (H).
In this case study, a fire is assumed in a fifth-floor room (see Figure 15.5(b)).
The area of the fire room is 70 m2 and the heat release rate (HRR) is 8,000 kW,
which is approximately the total maximum HRR of four chairs (Babrauskas, 1983).
The sprinklers were assumed to be installed in the fire room, which are spaced
4 m4 m. According to the NFPA Standard 13 (NFPA, 2002), the water volumetric
flow rate of the sprinklers is set at 60 L/min, and the triggering temperature is 74 C.
The ambient temperature is 20 C.
The study focuses on the smoke spread inside the two stairwells that are the
evacuation routes for people. To evaluate the performance of different smoke
control systems under the worst situation, the window of the fire room and the
doors between the room and the stairwells were set as open (assuming the resident
opened them due to panic). Five cases with different smoke control systems were
designed, and the relevant information is listed in Table 15.2. Case 1 is the base
case without any protections in the stairwells. Case 2 includes a smoke exhaust
system in Stairwell 2. Cases 3 and 4 use a stairwell pressurization system in the
stairwell with different settings. Case 5 has a dedicated smoke exhaust shaft with a
cross-sectional area of 2 m2 m, which is shown in Figue 15.5(b).
The simulation models were created using PyroSim, a graphical user interface
for FDS (Thunderbird Engineering, 2015). The total number of cells is around
500,000. The size of each cell is 0.5 m which ensures that there are at least four
cells in each opening. Temperature and pressure distributions inside the stairwells
were recorded at each floor of the stairwells. Mass flow rates at the stairwell
openings were measured by flow measuring devices.

15.4.1.2 Results
Figure 15.6 demonstrates the transient smoke/air temperature in the two stairwells at
the 5th (fire floor), 17th, and 29th floors for different cases. In Case 1, the maximum
smoke temperature reaches 450 C at around 220 s on the fire floor of Stairwell 2,
which is higher than 330 C of the maximum smoke temperature in Stairwell 1 at the
 Ventilation and fire safety for high-rise buildings 459
500 500
Temperature (°C)

Temperature (°C)
5th floor 5th floor
400 400 17th floor
17th floor
300 29th floor 300 29th floor
200 200
100 100
0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Time (s) Time (s)

Case 1– Stairwell 1 Case 1– Stairwell 2

5th floor 100 5th floor

Temperature (°C)
100
Temperature (°C)

17th floor 17th floor

80 29th floor 80 29th floor
60 60
40 40
20 20
0 0
0 100 200 300 400 500 0 100 200 300 400 500
Time (s) Time (s)

Case 2– Stairwell 1 Case 2– Stairwell 2

30 120
Temperature (°C)

Temperature (°C)

5th floor
90 17th floor
20 29th floor
5th floor 60
10 17th floor
29th floor 30
0 0
0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800
Time (s) Time (s)

Case 3– Stairwell 1 Case 3– Stairwell 2

30 30
Temperature (°C)
Temperature (°C)

20 20
5th floor 5th floor
10 17th floor 10 17th floor
29th floor 29th floor
0 0
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
Time (s) Time (s)

Case 4– Stairwell 1 Case 4– Stairwell 2

60 5th floor 60
Temperature (°C)
Temperature (°C)

5th floor
50 17th floor 50 17th floor
40 29th floor 40 29th floor
30 30
20 20
10 10
0 0
0 200 400 600 800 1,000 0 200 400 600 800 1,000
Time (s) Time (s)

Case 5– Stairwell 1 Case 5– Stairwell 2

Figure 15.6 Comparison of temperature profiles of smoke/air inside the stairwells

for different cases
460 Handbook of ventilation technology for the built environment

same time. After reaching the maximum temperature, the smoke temperature
decreases to the ambient temperature, which is around 22 C. This is due to the
operation of sprinklers in the fire room and the lack of fresh air for combustion since
there is no other opening for fresh air flowing into the building except for the open
window in the fire room. In Case 2, the stairwell smoke temperature jumps to 90 C in
a very short time (around 30 s) and then remains at around 70 C for most of the time
since the MV fan on the roof of Stairwell 2 operates and fresh air continuously flows
into the fire room through the open window.
For Case 3, the fire floor smoke temperature of Stairwell 1 increases to a
maximum of 23 C around 100 s. The temperature then drops to 21 C, because the
pressurization fan works well. Since the smoke spreads to Stairwell 2 (no pres-
surization in Stairwell 2), the smoke temperature increases to around 90 C by 400 s
then becomes stable. Both stairwells in Case 4 are smoke free, due to the two
pressurization fans on the roof. The air temperatures remain at the ambient tem-
perature, 20 C. For Case 5, most of the smoke is exhausted by the smoke shaft, and
only a little smoke infiltrates into the stairwells. The smoke temperature on the fire
floor of the stairwells is only a little bit higher than the ambient temperature, which
is around 25 C on average.
Figure 15.6 shows that Case 1 (sprinkler system only) is not enough to extin-
guish the fire, and the smoke can spread throughout the tower. It can also be seen
that the smoke spreads through stairwells faster in Cases 1 and 2 than in Cases 3–5.
Therefore, for Cases 1 and 2, both stairwells are not safe for evacuation. However,
the cases with a pressurization system or smoke shaft (Cases 3–5) could have at
least one safe stairwell that smoke cannot flow into. To further investigate the
performance of the smoke control systems in Cases 3–5, the mass flow rate and
pressure distribution inside the stairwell are discussed later.
According to Figure 15.6, smoke temperatures stabilize after 300 s indicating a
stable (or “steady state”) smoke flow. The average mass flow rate and pressure
distribution inside the stairwells during the time range of 300–400 s are presented
in Figures 15.7 and 15.8.
Figure 15.7 compares smoke/air mass flow rates flowing through the open
doors of the two stairwells in Cases 3–5. The negative mass flow rate indicates that
the smoke/air mass flow exfiltrated from stairwells to the floors whereas positive
values mean smoke infiltrated into the stairwells.
For Stairwell 1 in Case 3, pushed by the roof pressurization fan, the smoke
leaves the fifth floor, while little airflow infiltrates into the stairwell for the upper
floors of the stairwell, so the upper floors are smoke free and the temperature
increases a little (see Figure 15.6). The range of stairwell gauge pressure is 9–18 Pa
(see Figure 15.8), which is lower than the NFPA 92 requirement of pressure dif-
ference across any closed stairwell doors in the pressurized stairwells, which is
25–88 Pa. Considering the stairwell’s doors are set as open and smoke cannot be
found inside the pressurized stairwell, the design of Stairwell 1 with the roof
pressurization fan with a volume flow rate of 20 m3/s can be acceptable.
For Case 4, most of the air pressurized by the roof fan leaves from the bottom
opened doors of the stairwells (exfiltration), because of the large pressure difference
 Ventilation and fire safety for high-rise buildings 461

Floor number 29 29

Floor number
25 25
21 21
17 17
13 13
9 9
5 5
–3 –2 –1 0 1 2 3
–3 –2 –1 0 1 2 3
Mass f low rate (kg/s)
Mass flow rate (kg/s)
Case 3– Stairwell 2
Case 3– Stairwell 1

29 29
26
Floor number

Floor number
25 23
21 20
17 17
14
13
11
9 8
5 5
–20 –10 0 10 –40 –30 –20 –10 0 10 20
Mass flow rate (kg/s) Mass f low rate (kg/s)
Case 4– Stairwell 1 Case 4– Stairwell 2

29 29
Floor number

Floor number

26 26
23 23
20 20
17 17
14 14
11 11
8 8
5 5
–2 –1 0 1 2 –2 –1 0 1 2
Mass flow rate (kg/s) Mass f low rate (kg/s)
Case 5– Stairwell 1 Case 5– Stairwell 2

Figure 15.7 Comparison of mass flow rate through stairwell doors for different
cases (negative mass flow rate: exfiltration from stairwell; positive
mass flow rate: infiltration into stairwell)

across the stairwells’ doors. As is shown in Figure 15.8, the largest gauge pressure in
both stairwells exceeds 500 Pa, which far exceeds the required range of 25–88 Pa
defined by NFPA 92 for stairwell pressurization systems. This high gauge pressure
means that the required force to open the door can reach 1,000 N, indicating exces-
sive pressure in the stairwell, which may result in potential problems for people to
safely evacuate (e.g. people cannot open the stairwell door easily).
Figure 15.7 shows that Case 5 has smoke infiltrating into Stairwell 1 from the
fire floor. However, Stairwell 2 is free of smoke because the smoke is exhausted
through the smoke shaft, which is near shaft 2. The gauge pressure of Stairwell 2 at
the fire floor is 92 Pa (see Figure 15.8), which is a little bit higher than the NFPA 92
462 Handbook of ventilation technology for the built environment

100

80

Height (m)
60

40

20

0
–100 0 100 200 300 400 500 600
Stairwell pressure (Pa)

Figure 15.8 Comparison of stairwell pressure distribution for different cases

requirement. In summary, the best smoke control strategy of all five cases is
probably Case 3 (pressurization fan on the roof of Stairwell 1) because of its simple
implementation and lower working pressures in the evacuation stairwell. It appears
that Case 5 (smoke shaft with exhaust fan) may be an alternative approach, but
more in-depth analysis needs to be conducted to investigate its performance.

15.4.1.3 Conclusion
This case study uses CFD simulation of fire smoke spread and control in a real
high-rise building with two stairwells. Five smoke control methods, including
sprinklers only, mechanical exhaust system, two different pressurization systems,
and a smoke shaft system, were simulated and compared in the FDS simulations.
The comparison of five smoke control methods was shown by using the tempera-
tures, pressures, and mass flow rates of smoke inside the two stairwells. The
simulation results show that pressurizing only one stairwell with a fan volumetric
flow rate of 20 m3/s is enough to keep the evacuation stairwell safe for people.
However, if pressurization fans are installed in both stairwells, excessive pressure
differences across the stairwells could occur leading to a potential failure of the
pressurized system (e.g. people cannot open the stairwell door easily). Therefore,
the installation of pressurization systems in both stairwells should be carefully
analyzed and evaluated. In addition to the pressurization system, an added dedi-
cated smoke exhaust shaft can help to achieve smoke-free conditions in the nearest
stairwell. The dedicated smoke exhaust shaft can be an alternative method, espe-
cially when the stairwell pressurization system cannot be applied.

15.4.2 Atrium fire smoke control†

An atrium,† as a popular architecture element of high-rise buildings, can take
advantage of its large vertical space to generate the stack effect for NV (Wood and

†
Section 15.4.2 is modified from the conference paper published by ASHRAE: Sha H, Qi D. A Novel
Ventilation Approach of Large Vertical Space for Achieving Fire Safety and Energy Efficiency.
2020 ASHRAE Virtual Conference, 2020.
 Ventilation and fire safety for high-rise buildings 463

Salib, 2013). However, an atrium may also become the main route of smoke spread
when fire occurs. To avoid smoke spread along the atrium, the height of the atrium
is often limited by fire safety regulations. Segmentations are often applied in the
atrium, which limits the height of the space and reduces the size of the smoke
control zone. For example, the large vertical atrium in the Concordia EV building is
divided into five sub-atria by segmentations (Hu and Karava, 2014), and thus each
sub-atrium is only one-fifth of the height of the building. However, this solution
decreases the NV potential of the atrium, because the added segmentations increase
the airflow resistance and reduce the flow rates of NV (Sha and Qi, 2019).
A state-of-the-art ventilation design is proposed to balance the fire safety
concerns and potential of NV on the reduction of cooling energy consumption. This
ventilation approach consists of two elements: the segmentation slab and the
independent ventilation shaft. The segmentation slab is designed to limit the
smoke. The ventilation shaft can be used to exhaust smoke when fire occurs and
provide an airflow path for NV during daily use. This design was evaluated by
using FDS simulations and BESs in a case study of a 30 (L)12 (W)30.6 (H)-m3
atrium in Montreal, Canada. With a comparison of the traditional ventilation
method (i.e. the atrium without segmentation and ventilation shaft), the smoke
layer height predicted by FDS is used to evaluate the fire protection performance.
The cooling load in the atrium predicted by BES is used to evaluate the NV energy
performance.

15.4.2.1 Fire smoke and energy modeling

FDS was applied for the CFD smoke simulations, and the relevant information is
summarized in Table 15.3. The scenarios of the atrium with or without the new
ventilation design were considered. The atrium without the new design is called the
traditional design, which does not have the segmentation slab and the shaft. In the
new ventilation design, a segmentation is added at 16.5 m of the atrium, and a
ventilation shaft with a cross-sectional area of 1 (L)12 (W) m2 is added on a
sidewall of the atrium (please see Figure 15.9(a)). Apart from the segmentation and
shaft, the other settings are the same, i.e. the bottom openings and HRR of the fire
source are the same for both scenarios. Two HRRs were considered here: 2 and 5
MW, representing a relatively small fire source and a large fire source. Since the
FDS simulation accuracy is highly dependent on the mesh size, the grid indepen-
dence study was also conducted on grid sizes of 0.125, 0.25, and 0.5 m. The results

Table 15.3 Cases for fire smoke simulations

Cases Atrium Heat release Exhaust rate Bottom opening

design rate (MW) (m3/s) size (m2)
SC 1 No new design 2 212 211
SC 2 New design 2 80 78
SC 3 No new design 5 295 298
SC 4 New design 5 115 115
464 Handbook of ventilation technology for the built environment

Smoke exhaust fan

Top opening

Upper opening
Shaft
Segmentation slab Upper inlet

P1 Lower inlet Lower opening

Bottom opening

Fire

(a) (b)

Figure 15.9 Schematic of the atrium: (a) CFD simulation model; (b) BES model

Table 15.4 Energy simulation case design

Cases Cooling mode NV design Weather conditions

for NV
Traditional design-MC MC All openings closed /
New design-MC MC All openings closed /
Traditional design- MCþNV Inlet opening, 12 m2 Ta<Tin and
MNV Top outlet opening, 15  Ta  24 C
4 m2 Td  13.5 C;
vw  7.5 m/s
New design-MNV MCþNV Ventilation shaft Same as above
and inlet
openings, 12 m2
Top outlet openings,
4 m2

Note: Ta, ambient temperature; Tin, indoor temperature; Td, dew temperature; vw, wind speed.

of the 0.125 and 0.25 m grids demonstrate little difference in the smoke layer
height. Hence, a grid size of 0.25 m was chosen for all the FDS simulations.
EnergyPlus was used to conduct the BES for evaluating the NV energy perfor-
mance of two scenarios of the atrium. The summer period (from June to September of
typical meteorological weather data) was simulated. Four cases were designed and
illustrated in Table 15.4; the combination of the atrium with or without the new ven-
tilation design and with or without NV. The design of the atrium with the new venti-
lation design can be seen in Figure 15.9(b). To calculate the cooling load inside the
atrium, the setpoint of MC is 24 C. The atrium is occupied between 8:00 and 22:00.
The construction materials and internal heat gain (i.e. people, equipment, and lighting
load) are defined in the same way as the prototype building developed by the US
 Ventilation and fire safety for high-rise buildings 465

Department of Energy (Goel et al., 2014). The glazing ratios of the west and north
facades are set at 0.9 and 0.7, respectively. The NV is calculated by using the airflow
network model, which is integrated into EnergyPlus.

15.4.2.2 Results
Figure 15.10 presents the smoke layer height variation at location P1 (P1 is between
the entrance door and fire source, as Figure 15.10(a) shows) in all four cases. It can
be seen that the cases with the new ventilation design (Cases SC 2 and SC 4) have
larger smoke-free spaces than the other two cases (Cases SC 1 and SC 3). For Cases
SC 1 and SC 3, the smoke layer height is around 15–20 m, but there is almost no
smoke in Case SC 2 and only around a 5-m smoke layer depth in Case SC 4. After
adding the segmentation, the new design can keep the entire upper atrium smoke-free
and effectively remove the smoke in the lower atrium. Therefore, the new ventilation
design can contribute to better smoke control performance. In addition, the cases with
the new ventilation design require lower exhaust airflow rates and smaller opening
areas than the cases without the new design (see Table 15.3).

30 30
25 Smoke 25
Segmentation
20
Height (m)

Height (m)

20
15 15
10 10 Smoke

5 5
0 0
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)

Case SC1 Case SC2

30 30

25 Smoke 25
Segmentation
20 20
Height (m)

Height (m)

15 15
Smoke
10 10
5 5
0 0
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)

Case SC3 Case SC4

Figure 15.10 Smoke layer height

466 Handbook of ventilation technology for the built environment

In BES, the monthly cooling load from June to September was simulated.
These two ventilation design methods (the new ventilation design and traditional
design) have similar energy performance in terms of reducing the cooling load.
Compared with the cases with only MC, both traditional and new design methods
with NV (traditional design-MNV and new design-MNV) can reduce about 18% of
the total cooling load. The total NV airflow rates of the two design methods are
almost the same. However, it should be noted that the total cooling load of the new
design-MC case is around 22% higher than that of the traditional design-MC case,
which is caused by the increased thermal mass of the segmentation slab and double
internal heat gain. Although the new design has a higher total cooling load, the
atrium with the new design can accommodate twice as many occupants as the
traditional design, because of the segmentation slab.

15.4.2.3 Conclusion
This state-of-the-art novel ventilation design can contribute to better fire safety
performance and maintain the NV potential. The results of CFD simulations and
BESs proved that the atrium with the novel ventilation design has a larger smoke-
free space and almost the same energy savings as NV.

References

Ahrens M. High-rise building fires. 2016. https://doi.org/10.20965/jdr.2007.p0236.

Ai ZT, Mak CM, Niu JL, and Li ZR. The assessment of the performance of bal-
conies using computational fluid dynamics. Build Serv Eng Res Technol.
2011;32:229–43. https://doi.org/10.1177/0143624411404646.
Artmann N, Manz H, and Heiselberg P. Climatic potential for passive cooling of
buildings by night-time ventilation in Europe. Appl Energy. 2007;84:187–201.
ASHRAE. ASHRAE Handbook of Fundamentals. Atlanta, GA: American Society
of Heating, Refrigerating and Air-Conditioning Engineers; 2017. https://doi.
org/10.1017/CBO9781107415324.004.
Babrauskas V. Upholstered furniture heat release rates: Measurements and esti-
mation. J Fire Sci. 1983;1:9–32.
Chen H, Liu N, Zhang L, Deng Z, and Huang H. Experimental study on cross-
ventilation compartment fire in the wind environment. Fire Saf Sci.
2008;9:907–18. https://doi.org/10.3801/iafss.fss.9-907.
Chen Y, Norford LK, Samuelson HW, and Malkawi A. Optimal control of HVAC and
window systems for natural ventilation through reinforcement learning. Energy
Build. 2018;169:195–205. https://doi.org/10.1016/j.enbuild.2018.03.051.
Chen Y, Tong Z, and Malkawi A. Investigating natural ventilation potentials across
the globe: Regional and climatic variations. Build Environ. 2017;122:386–96.
Ding W, Hasemi Y, and Yamada T. Natural ventilation performance of a double-
skin façade with a solar chimney. Energy Build. 2005;37:411–8. https://doi.
org/10.1016/j.enbuild.2004.08.002.
 Ventilation and fire safety for high-rise buildings 467

EN 12101-3 Standard. Smoke and heat control systems—Part 3: Specification for

powered smoke and heat exhaust ventilators. Brussels, Belgium. 2002.
Goel S, Rosenberg M, Athalye R, and Xie Y. Enhancements to ASHRAE Standard
90.1 prototype building models. 2014.
Günel MH and Ilgin HE. Tall buildings: Structural systems and aerodynamic form.
2014. https://doi.org/10.4324/9781315776521.
Heiselberg P. Principles of Hybrid Ventilation. Aalborg: Aalborg University; 2002.
Hu J and Karava P. Model predictive control strategies for buildings with mixed-
mode cooling. Build Environ. 2014;71:233–44.
IEA. Annex 62 ventilative cooling design guide. 2018.
Klote JH and Milke JA. Principles of smoke management. 2002.
Kolokotroni M and Aronis A. Cooling-energy reduction in air-conditioned offices
by using night ventilation. Appl Energy. 1999;63:241–53. https://doi.org/
10.1016/S0306-2619(99)00031-8.
Lay S. Pressurization systems do not work & present a risk to life safety. Case Stud
Fire Saf. 2014;1:13–7. https://doi.org/10.1016/j.csfs.2013.12.001.
Li Y and Heiselberg P. Analysis methods for natural and hybrid ventilation-a cri-
tical literature review and recent developments. Int J Vent. 2003;1:3–20.
https://doi.org/10.1080/14733315.2003.11683640.
Malkawi A, Yan B, Chen Y, and Tong Z. Predicting thermal and energy perfor-
mance of mixed-mode ventilation using an integrated simulation approach.
Build Simul. 2016;9:335–46.
McGrattan K, McDermott R, Weinschenk C, and Overholt K. Fire dynamics
simulator user’s guide. Gaithersburg, MD. 2013.
Medved S, Babnik M, Vidrih B, and Arkar C. Parametric study on the advantages
of weather-predicted control algorithm of free cooling ventilation system.
Energy. 2014;73:80–7. https://doi.org/10.1016/j.energy.2014.05.080.
Miller RS and Beasley D. On stairwell and elevator shaft pressurization for smoke
control in tall buildings. Build Environ. 2009;44:1306–17. https://doi.org/
http://dx.doi.org/10.1016/j.buildenv.2008.09.015.
Moosavi L, Mahyuddin N, and Ghafar N. Atrium cooling performance in a low
energy office building in the Tropics, a field study. Build Environ.
2015;94:384–94. https://doi.org/10.1016/j.buildenv.2015.06.020.
NFPA. NFPA 101 life safety code. Quincy, MA: NFPA. 2018.
NFPA. NFPA 13: Standard for the installation of sprinkler systems. Quincy, MA:
National Fire Protection Association. 2002.
NFPA. NFPA 92: Standard for control smoke systems. Quincy, MA: NFPA. 2012.
OECD/IEA. The future of cooling opportunities for energy-efficient air con-
ditioning. 2018.
Sha H and Qi D. A review of high-rise ventilation for energy efficiency and safety.
Sustain Cities Soc. 2020a;54:101971. https://doi.org/10.1016/j.scs.2019.101971.
Sha H and Qi D. Energy and fire safety performance of atrium ventilation in high-
rise buildings. IBPSA 2019, Rome. 2019.
Sha H and Qi D. Investigation of mechanical ventilation for cooling in high-rise buildings.
Energy Build. 2020b;228:110440. https://doi.org/10.1016/j.enbuild.2020.110440.
468 Handbook of ventilation technology for the built environment

SkyscraperPage.com. Global cities & buildings database. 2019. [cited 2019 Sep
18]. Available from: http://skyscraperpage.com/.
DH Toronto Staff. The 21 tallest buildings under construction in Toronto right now.
Daily Hive. 2018.
Thunderbird Engineering. PyroSim user manual. 2015.
Wang L, Wong NH, and Li S. Facade design optimization for naturally ventilated
residential buildings in Singapore. Energy Build. 2007;39:954–61. https://
doi.org/10.1016/j.enbuild.2006.10.011.
Wood A and Salib R. Natural ventilation in high-rise office buildings. Chicago.
2013.
Yuan S, Vallianos C, Athienitis A, and Rao J. A study of hybrid ventilation in an
institutional building for predictive control. Build Environ. 2018;128:1–11.
Zhang Y, Wang X, and Hu E. Optimization of night mechanical ventilation strategy in
summer for cooling energy saving based on inverse problem method. J Power
Energy. 2018;232:1093–102. https://doi.org/10.1177/0957650918766691.
Zhou C, Wang Z, Chen Q, Jiang Y, and Pei J. Design optimization and field
demonstration of natural ventilation for high-rise residential buildings. Energy
Build. 2014;82:457–65. https://doi.org/10.1016/j.enbuild.2014.06.036.