Journal of Building Engineering 54 (2022) 104561

Contents lists available at ScienceDirect

Journal of Building Engineering

journal homepage: www.elsevier.com/locate/jobe

Numerical analysis on the hazards of open stairwell doors in

high-rise residential buildings
Philip McKeen a, *, Zaiyi Liao a
a
Department of Architectural Science, Ryerson University, Toronto, M5B 2K3, Canada

A R T I C L E I N F O A B S T R A C T

Keywords: High-rise buildings have unique risks associated with fire as egress is often limited to stairwells.
Smoke spread Protecting these vertical paths from the infiltration of smoke is essential for safe evacuation.
Stairwell Passive fire safety strategies have addressed this by using fire-rated compartmentation. Fire-rated
Air leakage stairwell doors are commonly compromised in residential buildings, as observations have found
Airflow them propped open or failing to close correctly. Many recent fire tragedies have demonstrated the
MURB
spread of smoke into the stairwell can render egress paths untenable and carry smoke throughout
the building. The consequence of stairwell door openings on smoke spread in high-rise multi-unit
residential buildings is evaluated in this study. The smoke spread is predicted within a 10-storey
multi-unit residential building using computational fluid dynamics. Tenability predictions in the
stairwells determine available safe egress time for an array of leakage and door parameters. This
illustrates the relative hazard to egress where stairwell doors are not properly closed.

1. Introduction
Smoke movement within a building is a function of pressure differentials. These exist due to environmental conditions, such as
stack effect, wind, mechanical systems, and those created by fire, such as buoyancy and thermal expansion. For uncontained high-rise
fires, smoke can quickly spread and create untenable environments far from the fire origin. Within multi-unit residential buildings
(MURBs), vertical shafts can facilitate the spread of smoke. The size of the leakage area between compartments can affect the quantity
of smoke entering the stairwells and influence the tenability of egress pathways.
Fires on lower floors can be particularly hazardous as the smoke rises in vertical shafts. The MGM Grand Hotel fire in Las Vegas in
1980 involved a ground floor fire that killed 85 people, with the majority located on upper floors [1]. Recently, the 2022 Bronx, New
York, apartment fire resulted in the spread of smoke into the stairwell. While the fire began in a lower residential suite, many fatalities
occurred in the stairwells [2]. Evidence suggests that the apartment door and several stairwell doors remained open, creating a “flue”
or “stack-driven flow” that reduced tenability rapidly.
As buildings become taller and more densely populated, the importance of protecting the egress path grows. Building codes have
evolved to include various passive and active fire safety strategies for tall buildings. This evolution creates a disparity in the life safety
of older buildings as upgrades are perceived to be cost-prohibitive. Nonetheless, open stairwell doors may significantly compromise
the stairwells of both new and old buildings.

* Corresponding author.
E-mail address: pmckeen@ryerson.ca (P. McKeen).

https://doi.org/10.1016/j.jobe.2022.104561
Received 6 February 2022; Received in revised form 8 April 2022; Accepted 22 April 2022
Available online 2 May 2022
2352-7102/Crown Copyright © 2022 Published by Elsevier Ltd. All rights reserved.
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

1.1. Stairwells as smokeproof enclosures

Historically, passive fire strategies of compartmentalization have been relied upon to physically obstruct the smoke movement with
walls, floors, and doors. Several active fire strategies evolved in the 1980s, with NFPA 92(A) addressing smoke control in buildings
using pressurization to protect evacuations paths [3]. This expanded the methods for achieving “smokeproof enclosures” of the egress
stairwell. In general, this can be achieved in three ways: 1) naturally ventilated space, 2) mechanical ventilation of lobbies, or 3)
stairwell pressurization.
While many modern MURBs have both a stairwell pressurization and sprinkler system, plenty of existing building stock does not.
This can lead to larger fires, with greater temperatures and smoke production. An NFPA survey of building fires between 2009–13
revealed that only 42% of high-rise MURBs contained sprinkler systems [4]. Even where such systems exist, their performance can be
less than desirable. Pressurization systems do not always prevent smoke from spreading into shafts, especially for high-rise applica­
tions, due to stack effect and variation in airflow resistance amongst floors [5]. It was estimated that 35% of pressurization systems
might fail to function as intended [6].
The 2017 Grenfell Tower fire demonstrated that smoke spread remains a hazard even with compartmentation and smoke control
systems [7]. The Grenfell Tower inquiry noted the bespoke mechanically ventilated corridors likely did not perform as intended or as
required. Other problems that may reduce performance include human interference with the operation and improper maintenance.
Inadequate smoke control performance may be found in many older buildings with ineffective systems or even newer ones that lack
proper maintenance [2].

1.2. Open doors in stairwells

Observations of stairwell doors permanently propped open are common in MURBs (Fig. 1). In this example, stairwell doors on
multiple floors were permanently open. A 1970’s study on fire doors revealed that an average of 17% of fire doors were propped open
in several apartment buildings inspected in the UK [8]. Despite prevalent requirements for door closers, building surveys have found
the devices not installed or failing to latch properly [9].
Open doors have led to several fatalities in recent fires [2]. Occupants fleeing the fire origin have often left doors open, allowing
smoke to spread to the corridors. Additionally, most smoke control systems are intended to perform with properly closed doors and
may be rendered ineffective when several openings exist.
Analyzing the potential smoke spread in a building can determine the hazard to life safety. This paper demonstrates the conse­
quence of smoke spread due to openings in stairwell compartments. Using Fire Dynamics Simulator (FDS) software, the influence of
leakage area is parametrically examined for a 10-storey case-study building. Simulations determine the flow of smoke and concen­
trations in the egress paths. The impact on visibility is predicted to determine tenability and the available safe egress time (ASET). The
relative performance of different scenarios demonstrates the significant influence leakage parameters have on smoke spread. These
findings can apply to other buildings and provide insight into the inherent risk to egress.

2. Theory and methodology

Smoke spread throughout buildings can be analyzed using computational fluid dynamic (CFD) simulations [10]. For this purpose,
an idealized model is developed utilizing relevant properties derived from a range of existing MURBs. Environmental conditions and
fire characteristics (size, location, growth rate) can be specified based on the likelihood of occurrence. The range and appropriateness
of these parameters are explored in this section (Fig. 2). Predictions of smoke spread can be compared against tenability criteria to
determine the performance of each scenario.

Fig. 1. Observations of stairwell doors propped open in a high-rise MURB in China.

2
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 2. Process of developing an idealized model to evaluate smoke spread in a typical MURB.

2.1. The idealized model

The layout of MURBs typically consists of a perimeter of residential units served by a corridor and vertical shafts. Construction cost,
zoning, and design regulations contribute to a common range of building footprints and aspect ratios. A study of high-rise MURBs
geometries in China found that approximately 60% had a plan aspect ratio between 1:1 and 1:2 [11].
Airflow through the numerous compartments is analogous to the flow through a complex system of pipes and orifices [12]. The
primary leakage paths can be represented by the building envelope, A1, the corridor/suite doors, A2, doorways to vertical shafts, A3.
Changes to the pressure regime can change the airflow in magnitude and direction, as illustrated in Fig. 3. During building fires, these
paths allow the spread of smoke.
Airtightness experiments of existing buildings [13,14] demonstrate the correlation of airflow at specific pressure differentials
according to the power-law Equation (1). The pressure exponent for the flow regime ranges from 0.5 for turbulent to 1.0 for laminar
[15,16].

V̇ = C(ΔP)n (1)

V̇ = volumetric ​ flow ​ rate, ​ m3 /s

C = flow ​ coefficient, m3 /(s · Pan )
n = flow ​ exponent
ΔP = pressure ​ differential, ​ Pa
The leakage through gaps in walls, floors, and doors can also be described by the flow equation appropriate for sharp-edged orifices
(Equation (2)). The orifice can be defined as a cross-sectional area perpendicular to the flow of air or the effective flow area, Ae . The
discharge coefficient, CD, accounts for the flow regime, often ranging between 0.5 for large openings and 0.7 for narrow openings [13].
√̅̅̅̅̅̅̅̅̅
2ΔP
V̇ = CD Ae (2)
ρ

CD = discharge ​ coefficient
Ae = effective ​ flow ​ area, m2

Fig. 3. A variety of airflow patterns may exist within a MURB floor. Note direction of flow may change depending on pressure differentials.

3
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

ρ = density ​ of ​ gas, ​ kg/m3

ΔP = pressure ​ differential, Pa
When tailored with the proper coefficients, these equations demonstrate good agreement with experiments and are deemed
acceptable for the study of smoke spread in buildings [13]. While Equations (1) and (2) differ, the flow can be closely correlated for
most building pressure differentials.
For the exterior envelope, airtightness is often described by a relative range encompassing most buildings. Tamura and Shaw
defined this range as tight, average, and leaky to measure approximately 0.5, 1.5, and 3.0 L/s/m2 at 75 Pa, respectively [17]. Fig. 4
illustrates the airflow when using a flow exponent of 0.65. For comparison, airtightness measurements of a MURB constructed in 1986
determined the pre-retrofit and post-retrofit airtightness to be 3.9 L/s/m2 and 1.9 L/s/m2 at 75 Pa, respectively [18]. The total leakage
area may change based on the surface area dictated by the building geometry. The definition of tight, average, and leaky envelopes is
loosely defined but indicates an approximate range that can be used in idealized models.
For interior compartments, leakage occurs along walls, floors, and doors. The majority of the leakage area is often found at doors,
even when closed. Leakage area is limited to the size of the gap between door and frame and regulated by standards such as NFPA 80
[19], which requires a maximum gap of 3 mm (1/8 in.) along the jambs and head and 10 mm (3/8 in.) at the threshold. For a standard
size door, this perimeter gap could present a cross-sectional area of approximately 250 cm2, although the effective area will be affected
by the door frame and gaskets. For doors required to be smoke-leakage rated, building codes commonly require the air leakage to meet
specific standards. The International Building Code (IBC) [20] and National Building Code of Canada (NBC) [21] reference standard
UL1784 [22] for a maximum leakage of 15.24 L/s per m2 (3 CFM per ft2) at 25 Pa. The leakage rate of this standard is depicted in Fig. 5
for a typical 2 m2 door. More stringent requirements are found in the British Standard Institution (BSI) BS476–31.1, with a maximum
leakage rate of 3 m3/h/m (measured along the frame) at 25 Pa [23]. The standard criteria do not consider the leakage at the door
threshold or the wall and frame junction. Depending on construction, the stairwell wall surface may contribute to the leakage area, as
much as 3.5 cm2/m2 as per NFPA 92 [3,24].
While the tightness of doors could be minimal when built to these strict tolerances, evidence suggests that older buildings and
workmanship can result in far more leakage. Field test of a MURB constructed in 1986 determined that the undercut depth of suite
doors ranged from 0 mm to 22 mm, with an average of 5.9 mm [18,25]. The testing of several MURB doors revealed a range in airflow
between 30 and 210 L/s at 75 Pa. Leakage testing performed in the 1980s found stairwells door assembly leakage rate between 113 and
271 L/s at 75 Pa [26]. Fig. 5 depicts the measured leakage area of several doors and shows the resulting airflow. Considering a
spectrum of leakage measurements of suite and stairwell doors, an approximate range found within buildings may be described as
“tight” (100 cm2), “average” (200 cm2), and leaky (300 cm2).

2.1.1. Leakage paths and airflow through buildings

Variation in the leakage area of envelopes, corridors, and stairwells can affect the airflow through the system [27]. Klote
demonstrated this occurrence in stairwells, for which airflow increased significantly when additional doors were opened [28]. The
effective flow area is defined as the cross-sectional area capable of producing equal airflow for a given pressure differential, defined by

Fig. 4. Typical range of airtightness for building envelopes as determined by several blower door test and an orifice approximation defined as “tight”, “average”
and “leaky".

4
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 5. Typical range of airtightness for closed doors from experiments, reference standard UL1784, and the orifice approximation defined as “tight”, “average”
and “leaky".

Equation (4). The most restrictive path greatly influences the effective area in a series of leaks.
( )− 1
∑n
1
2

Ae = (4)
2
i=1
A i

Airtightness can change significantly when occupant behavior is considered. For instance, a tight building envelope can provide a
large leakage area when operable windows are opened. A tight corridor can dampen the fluctuation in the effective area resulting from
exterior openings.

Fig. 6. Comparison of the volumetric flow through a leakage path using flow equations and physically modeled orifices in CFD.

5
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

2.1.2. Validation and analysis of CFD modeling

In CFD simulations, a leakage path can be defined by a flow equation or a physical model of the orifice. In FDS, the correlation
between these is illustrated in Fig. 6. The power-law equation describes a field measurement of airtightness. An orifice equation can
also represent this with an area of 200 cm2 using a reference pressure of 10 Pa. However, the physically modeled orifice in FDS tends to
overpredict. This can be compensated for by reducing the orifice opening area to 150 cm2.

2.1.3. Airflow through closed doors

Smoke spread will occur along the leakage path between closed doors and frames. Physically modeling such a gap is impractical in
large-scale CFD simulations due to the computational cost. In FDS, this leakage path can be modeled with “vents” which utilize the
orifice equation to determine the mass flow based on the average pressure differential along its boundaries. This makes it necessary to
describe the door with multiple vents, as the pressure gradient will vary with height above the floor.
Due to buoyancy, the flow of smoke through a closed door will typically occur only when the ceiling smoke layer descends beneath
the head of the door. For this reason, the leakage paths should follow the geometry of a door frame. Buoyant forces also result in a
localized neutral pressure plane (NPP) between the floor and ceilings of compartmented space. The surface vent method in FDS does
not allow for bi-directional flow along its surface. However, using multiple vents permits the bi-directional flow of gas between the two
compartments. Fig. 7 illustrates the effect of the modeling approach on the smoke movement through a closed door, represented by a
single vent, two vents, and a door frame composed of six vents.

2.1.4. Validation of stairwell experiments in FDS stairwell simulations

The flow of hot gas through a shaft can be predicted in FDS, and many scale experiments have produced measurements that can
establish some validation. Ji et al. experiment of buoyant plumes, conducted at 1/3rd scale for a conventional stairwell [29], measured
the temperature along the center of the stair shaft for near steady-state conditions. A full description of the experimental conditions can
be found in Ji et al. study [29].
The stairwell experiments were replicated in FDS in the 1/3rd and full scale using the Froude modeling concept to demonstrate the
correlation of gas flow and temperature distribution. To properly resolve the fire, a grid consisting of 4 cm cells in the 1/3rd scale and
12 cm cells in the full-scale simulation plume [30]. Scissor stairs were also modeled to compare the temperature distribution with
conventional stairs (Fig. 8).
FDS simulations of the conventional stairwell produced good agreement amongst temperature rise and airflow measurements for
the fire sizes tested. A comparison of steady-state temperature measurements from the 1/3rd scale experiment and the full-scale FDS
simulation is depicted in Fig. 9. Temperature predictions nearest to the fire closely match experiment results, often within 10%, and
deviated with distance from the fire source.
Fig. 10 shows the temperature distribution of the 76 kW fire experiment of the conventional stairwell compared to the simulation
prediction for both the conventional and scissor stairwells. The 1/3rd and full-scale simulations strongly correlate with the prediction
but are most accurate nearest the fire origin. Scissor stairwell simulations produced significantly higher temperature predictions.
However, this can be expected due to a reduced hydraulic area and less airflow through the shaft. These simulations demonstrate that
reasonable agreement can be achieved in the prediction of the buoyant rise of hot gas within vertical shafts.

2.2. Environmental conditions

Pressure differentials created by mechanical ventilation, stack effect, and wind can drive gas flow throughout a building. The
environmental conditions are numerous and can be nullified to demonstrate performance in their absence. However, it is worth noting
that they can significantly influence smoke spread, and some examples are briefly described.
Stack effect becomes an increasingly dominant environmental force as building height increases and where the climate is colder [1,
5,31]. The stack pressure is a function of the building height and indoor/outdoor temperature. The magnitude of the height and
temperature difference can lead to significant pressures (Equation (1)).
( )
gP 1 1
ΔP = − h (5)
R To Ti

where.

Fig. 7. Smoke spread through one vent, two vents, and six vents and geometry used in the FDS model.

6
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 8. Simulation geometry for full scale conventional and scissor stair shafts (left) and FDS simulations of these shafts (right).

Fig. 9. Comparison of temperature for thermocouple array in the experiment [29] and FDS predictions for various fires sizes.

h = building height, m
g = gravitational acceleration, m/s2
P = absolute atmospheric pressure, Pa
R = gas constant of air, J/(kg•K)
To = absolute temperature of outside air, K
Ti = absolute temperature of inside air, K
Pressure from wind often results in horizontal driven flow, compared to the vertical flow-through vertical shafts [32]. However,
wind gusts affect a compartment fire’s fire dynamics by providing ventilation and increasing combustion and heat release. A
compartment fire exposed on the windward side would result in smoke being pushed into the building and adversely affect the
conditions.

2.3. Design fire

A design fire is used to define the fire growth, heat release rate, and soot production deemed appropriate for the purpose intended.
Many guides and articles have broadly addressed the appropriateness of a design fire and the process of justification [13]. Fire safety
engineers usually attempt to assess the design fire scenario with the probability of occurrence and magnitude of hazard associated with

7
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 10. Temperature rise from 76 kW fire experiment and comparison to FDS simulation, scale and stair type.

it.
For MURBs, fire hazard and probability vary depending on fire location. Fires are most frequent in residential suites, and fires in the
kitchen are most common [4]. However, the fire hazard and growth potential are significantly higher in bedrooms or living rooms due
to more combustible surfaces, including furniture and finishes. The size of a fire can vary significantly due to fuel load, oxygen
availability, and compartment geometry. While fires can grow beyond the compartment of origin, the focus of the design fire is to
capture the production of heat and soot for the necessary evacuation duration.
Open calorimeter experiments of residential suite mock-ups have revealed a range of HRR profiles that a design fire can emulate.
Once past the initial incubation stages, a fire often grows exponentially and can be described by the t-squared fire equation (Equation
(6)). The growth rate is dependant on the fuel and environment, with fires in residential compartments often between fast, α = 0.04689
kW/s2, and ultra-fast, α = 0.1878 kW/s2 [33].

Q̇ = αt2 (6)

Q̇ = heat release rate of fire, kWα = fire growth coefficient, kW/s2 t = time, ​ s
The soot concentration can vary between fuel and combustion conditions [34]. Studies on the flaming combustion of fuels revealed
that smoke produced from soot yields beyond 0.12 g/g produces similar obscuration of visibility, and was used in simulations [35]. The
mass optical density depends upon the material burned and combustion conditions [36]. Uncertainty exists as to what the actual fuel
components of fire will be. Generally, the residential environments can expect a mix of plastics and wood products as the majority of
the fuel source. The default value used in FDS is 8700 m2/kg, suggested for flaming combustion of wood and plastics. Equation (7)
demonstrates the relationship between soot produced and fire size [35,37,38].

Q̇
ṁs = Ys (7)
Hc,eff

ṁs = mass ​ flux ​ of ​ smoke, ​ kg/s

Ys = soot ​ yield, ​ g/g
Q̇ = heat release rate, kW
Hc,eff = effective ​ heat ​ of ​ combustion, ​ MJ/kg

Justifying the appropriate design fire ascertains a safety level is a challenge in performance-based design. The fire duration and
peak HRR can vary greatly depending on the fuel load and distribution. Experiments of residential compartment fires revealed peak
HRR ranged between 1.8 MW and 8.0 MW, highest when fast-burning fuels were used [33].
While variation in design fire size and growth rate will affect the duration of tenability, the relative performance between different
building scenarios remains relevant. The design fire used in this study is prescribed to grow fast (α = 0.04689), reaching 1.4 MW at 147
s, and maintained for the remainder of the simulation. The fire size was selected to ensure the fire was well ventilated under all
scenarios.

8
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

The design fire is located within a residential suite on the 2nd floor. Fires on the lower levels are often the most dangerous due to the
combustion gases buoyancy and stack effect. Windows are often broken during fire growth or at flashover, ensuring ventilated con­
ditions. To represent broken windows, an opening of 3.0 m2 is provided to the fire compartment, and for simplicity, the opening exists
from the ignition. The door from the fire compartment to the corridor is left open [39].

2.4. Tenability criteria

Tenability along the egress path determines the duration of safe egress. Tenability criteria consider a wide range of hazards to those
evacuating, including visibility, heat, and exposure to toxic gases. The ASET can be estimated by determining when such criteria are
compromised.
Visibility is often considered a leading indicator of tenability [5,40–42]. The lack of visibility can impede egress, but is contingent
on the complexity the egress route and occupant familiarity. A decrease in visibility has been demonstrated to reduce the walking
speed of evacuees [43]. Reducing visibility under 5.0 m may prevent occupants from attempting to evacuate entirely [44]. For
stairwell egress, the visibility criterion should allow evacuees to see the landing directly below [5]. While the lack of visibility does not
harm the individual, studies have established a correlation between the concentration of smoke and harmful gases like carbon
monoxide [31].

3. Smoke spread simulation

Smoke spread simulations assess the hazards of open stairwell doors in a 10-storey MURB. An idealized model, illustrated in Fig. 11,
was developed based on typical layouts and aspect ratios. The stairwell consists of two separate stairwells (SW1 and SW2), opening
into a corridor on each floor. The parameters appropriate for MURBs were evaluated and implemented into the idealized model. The
determination of fire size, tenability criteria, environmental conditions, and simulation parameters appropriate for a MURB fire are
discussed below. Several simulation scenarios were then examined to evaluate the influence on the smoke spread. Simulations are
conducted in the absence of smoke control systems. This represents a proportion of buildings that lack a smoke control system or have a
system that is not reliable or ineffective.

3.1. Building idealization and parameters

Through building idealization, relevant airflow paths were identified and simplified into parameters for numerical simulation. The
model does not consider leakage paths through floors as airflow through these paths is proportionally low compared to vertical shafts
[13]. While elevator doors also constitute a leakage path, their leakage area can vary greatly. Elevators are not included in this
simulation study, as the objective is to compare the relative performance of stairwell shafts exclusively. However, they can change the
flow dynamics and pose a conduit for smoke spread to other floors.
Leakage parameters were determined using experimental data reflecting industry standards and are defined in Table 1. The leakage

Fig. 11. Plan of 10-storey MURB used in smoke spread simulation showing idealized paths of leakage, sections, and sensor locations used to measure visibility in
simulations.

9
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Table 1
Idealized component parameters.

Flow Path ID Flow Path Description Orifice Area Discharge Coefficient (Cd)
2 2 2
A1 Exterior walls (tight, 2.9 cm /m ) 700 cm per floor 0.5
A1 Exterior walls (average, 5.8 cm2/m2) 1430 cm2 per floor 0.5
A1 Exterior walls (leaky, 11.6 cm2/m2) 2850 cm2 per floor 0.5
A2 Corridor to Suites (doors closed) 440 cm2 0.5
A3 Stairwell to building (“average” doors closed) 220 cm2 0.5
A3 Stairwell to building (“leaky” doors closed) 440 cm2 0.5
A3 Stairwell to building (doors opened) 22,000 cm2 a

a
CFD calculates fluid flow through open doors.

area of stairwell doors represents the approximate average airtightness testing of door assemblies, as illustrated in Fig. 5. Accounting
for the door and wall assembly, the leakage area of stairwell doors was selected as “average” (220 cm2) and leaky (440 cm2). The range
of leakage parameters was utilized to gauge the influence on smoke spread through open and closed stairwell doors.
Fig. 11 depicts the location of the design fire used in the FDS simulation. The location of windows which are assumed to be broken
and open during the simulation is noted. Sensor locations in the stairwell are used to measure tenability and record visibility and
temperature. Building cross-sections used to illustrate the spread of smoke and simulation visuals are referenced as S1, S2, and S3.

3.1.1. FDS simulation parameters and boundary conditions

Simulations were conducted in FDS version 6.6.0 using default properties. The simulation domain was modeled to contain the
illustrated building design, with a width of 27.0 m, a depth of 14.0 m, and a height of 30.0 m. Simulations were run with a cell spacing
of 0.1 m requiring 11,340,000 cells. The simulations were run on an Intel Xeon E5-2650 CPU with 24 cores. Parallel processing with
multiple-mesh arrangements required approximately five days per simulation.
Boundary and environmental were modeled based on parameter investigation and building idealization. Interior surfaces utilized
the material properties of concrete. Closed-door leakage and envelope leakage were modeled using FDS’s ‘VENT’ surfaces, applying the
orifice equation. Open doors were modeled as physical openings and solved by CFD directly. The exterior temperature and initial
interior temperature were specified to be 20 ◦ C. The exterior temperature difference can result in stack pressures that change the
airflow characteristics of the building significantly; despite this being of interest, the simulations conducted in this study aim first to
investigate the effect of door opening size on the smoke spread.

3.1.2. Tenability criteria

A visibility criterion of 5.0 m was used for FDS simulations. For evaluation of tenability in simulations, visibility is measured in the
stairwell compartments, 2.0 m above the center of each landing, as indicated in Fig. 11. When visibility is reduced below 5.0 m, it is
assumed that the path is no longer tenable. The measurements are time-averaged over 10 s to discount the effect of puffs of smoke into
the stairwell when visibility is otherwise maintained.

3.2. Simulation scenarios

Simulations were conducted to evaluate the effect of stairwell door openings on the smoke spread and tenability during a suite fire.
Parametric analysis was conducted by simulating the effect of increasing stairwell door leakage characteristics from tight to open.
Three series of simulations were conducted to consider the influence of stairwell doors.
• Simulation Scenario 1: Closed vs Open Stairwell Doors
This set of simulations investigates the worst-case scenario of all stairwell doors being open. Open doors were modeled with
physical openings of approximately 2.2 m2. This is compared to all stairwell doors being closed, with a leakage area of 220 cm2,
representing tight doors or 440 cm2, representing leaky doors. Simulations were conducted with tight, average, leaky envelopes for
each door condition.
• Simulation Scenario 2: Partial Stairwell Door Openings
The effect of only specific stairwell door openings existing is evaluated. Stairwell doors on the fire floor are specified as closed, with
all remaining stairwell doors open (Closed 2nd Floor). Stairwell doors on the fire floor are specified as open, with all other stairwell
doors closed (Open 2nd floor). Stairwell doors on the fire floor and the 10th floor are specified as open, with all other stairwell doors
closed (Open 2nd, 10th floor).
• Simulation Scenario 3: Dynamic Stairwell Door Openings
It is expected that stairwell doors will be opened and closed during egress. The effect of occupant behavior on stairwell doors is
considered in this simulation. Stairwell doors are set to open and close throughout this simulation.
• Sensitivity Study: Influence of open vents
Stairwells may contain a variety of openings, which may be part of passive smoke control systems. The influence of several common
openings within the shaft and corridor are evaluated on tenability and smoke spread.

10
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

4. Results and discussion

Smoke spread simulations were conducted in FDS for the 10-storey building illustrated in Fig. 11. The time required for smoke
infiltration into corridors was recorded amongst different scenarios (Table 2). However, this prediction does not indicate the tenability
of the egress path, as the dilution of smoke can allow for tenable conditions to persist.
Therefore, tenability along the egress path was assessed based on the measured visibility in both stairwell compartments (SW1 and
SW2), as recorded in Table 2. The stairwell with the lesser ASET duration was used to represent the scenario’s overall performance.
Simulation of closed stairwell doors with a leakage area of 220 cm2 formed a baseline allowing comparison of overall smoke spread and
tenability in stairwells to all other simulations. The relative ASET represents the change in ASET time for each simulation compared to
the baseline simulations.

4.1. Simulation scenario 1: open vs closed doors

Closed stairwell doors (220 cm2) provided an ASET of 435 s. When the leakage area was increased, closed stairwell doors (440 cm2)
provided an ASET of 305 s, or 71% of the baseline. An ASET of just 150 s resulted, or 36% of the baseline with open stairwell doors.
Visibility measurements of door conditions under tight, average, and leaky envelopes are depicted in Fig. 12.
The smoke spread was most affected by the size of the leakage area of stairwell doors. Open stairwell doors resulted in the rapid
spread of smoke to corridors. Closed doors significantly limited the smoke traveling between the stairwell and the corridors (Fig. 13).
Envelope tightness influenced the spread of smoke, most notable when stairwell doors were open (Table 2).

4.1.1. Closed stairwell door scenarios

Closed stairwell doors prevented the direct entry of smoke into the stairwell, significantly improving the ASET. As buoyant smoke
and hot gas accumulate within the corridor, the smoke layer descends. The buoyant hot gas layer creates a vertical pressure differential
along the corridor. This results in bi-directional flow at stairwell door locations. Smoke infiltrates the stairwell and occurs faster for the
leakier door (Fig. 14). Stairwell doors with leakage of 220 cm2 maintained stairwell tenability for 465 s (SW1) and 410 s (SW2).
Stairwell doors with leakage of 440 cm2 maintained stairwell tenability for 340 s (SW1) and 310 s (SW2).

4.1.2. Open stairwell door scenarios

Open stairwell doors allow for the direct infiltration of smoke and hot gas. Tenability rapidly declines with an ASET of just 150 s.
The buoyant forces spread smoke vertically in the stairwell and into corridors. As the buoyant smoke rises in the stairwell, it expands
into corridors due to the open door conditions. Predictions show the smoke reaches the 10th floor in 325 s.
Envelope conditions were noted to influence the volumetric smoke entering the stairwells. Flow-through open stairwell doors
reached a steady-state average of 0.7 m3/s for tight envelopes, 0.9 m3/s for average envelopes, and 1.2 m3/s for leaky envelopes. Tight
envelopes reduced the vertical rate of smoke spread, and leaky envelopes accelerated it (Fig. 15). However, the tenability of stairwells
was similar in all scenarios, as visibility was reduced below 5.0 m at approximately the same time.
It is not necessarily a strategy or suggestion that tight envelopes benefit smoke spread. However, the building tightness can in­
fluence the flow and stratification of smoke. For instance, tight envelopes reduced the infiltration of smoke, and shaft pressure
increased (Fig. 16), even if only marginally. For leaky envelopes, smoke entering stairwells can exfiltrate through the building

Table 2
Smoke spread simulation Results.

11
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 12. Visibility in the stairwell for closed and open doors for tight, average and leaky envelopes (along cross-section S1).

Fig. 13. Smoke spread into corridors for closed and open doors with average envelope tightness (along cross-section S2).

Fig. 14. Smoke flow from the fire floor corridor into the stairwell for tight, leaky, and open stairwell doors (along cross-section S3).

envelope, minimizing pressure build-up. While this does not have a tremendous impact on tenability, it does affect the quantity of
smoke within the stairwell. Tight envelopes reduce the rise of smoke but may promote its expansion into lower floor corridors faster
(Table 2).

12
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 15. Flow rate through open stairwell doors for Tight, Average and Leaky envelopes.

Fig. 16. Pressure in stairwell with doors open for Tight, Average and Leaky envelopes measured at 300 s.

4.2. Simulation scenario 2: select stairwell openings

4.2.1. Closed stairwell doors on the 2nd floor
With stairwell doors on the fire floor closed, open stairwell doors on other floors yield only a slight influence on the spread of
smoke. What is worth noting is that the spread of smoke into the corridors on upper floors can be significant when open doors exist, as
illustrated in Table 2. Smoke enters the 3rd-floor corridor at 390 s. The ASET of 440 s represents an 8% increase compared to stairwell
doors closed. The increased tenability is a consequence of dilution resulting from unimpeded smoke spread into corridors. While
visibility is maintained for a slightly longer duration, the spread of smoke to corridors is undesirable.

13
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

4.2.2. Open stairwell doors on the 2nd floor

With stairwell doors on the fire floor open, the smoke spread rate is influenced by the opening of stairwell doors on other floors
(Fig. 17). Simulation Scenario 1 demonstrated the rapid spread of smoke into the stairwell when all stairwell doors were open. With all
stairwell doors open, airflow from the fire floor into the stairwells reaches 0.7 m3/s and 1.2 m3/s for SW1 and SW2, respectively
(Fig. 18). When only the 2nd-floor stairwell doors are open, this is reduced to 0.6 m3/s for SW1 and SW2. The combination of other
open stairwell doors can accelerate the spread of smoke in the stairwell and into corridors.
With only stairwell doors open on the 2nd floor, the flow of smoke is restricted due to increased shaft pressure, as illustrated in
Fig. 16. This results in increased smoke logging into the lower floor corridors. This is noted in Table 2, where the upper floor corridors
remain free of smoke for a longer duration. Open stairwell doors on the 2nd and 10th-floor allow smoke to rise faster within the stair
shaft and enter corridors on higher floors (Fig. 17).

4.2.3. Open stairwell doors on the 2nd and 10th floor

The opening of the stairwell doors on the 2nd and 10th floor creates a unique condition where SW1 and SW2 compartments are
joined in a closed circuit. The hot gas entry into SW1 and SW2 gives rise to pressure vertically in the shaft. Due to the geometry of the
fire floor, the flow of hot gas into the stairwell is most dominant in SW2. The pressure build-up eventually forces the rising hot gas in
SW1 to reverse (Fig. 17). The flow into the stairwell reaches 2.8 m3/s for SW2 and is reduced to 0.1 m3/s for SW1 (Fig. 18). It can be
understood that the opening of other stairwell doors increases the flow of gas from the fire floor and can create unique flow conditions
that change the expected smoke movement.

4.3. Simulation scenario 3: dynamic door opening

Simulations with doors closed do not consider the operation of stairwell doors, which is necessary for a building evacuation. Its
effect should be considered when comparing the smoke spread performance. While occupant behavior is difficult to predict, simple
assumptions can be made to account for the egress of building occupants. In this instance, an egress model accounts for the evacuation
of 180 occupants, who begin to exit 60 s after the fire begins. The operation of opening and closing the stairwell doors is specified at 5 s
to approximate the time required for door closers to operate. This operation occurs at each door at 60 s intervals.
Within the smoke spread simulation, the intermittent opening of stairwell doors permits the influx of smoke in pulses. Thus
tenability predictions are sensitive to when the doors are opened. The operation of the door on the fire floor would likely not occur
once the corridor has become filled with smoke. However, accounting for some operations of the doors reveals a conservative scenario
for comparison.
The predicted ASET is approximately 260 s (Table 2), or a 35% reduction from the baseline. The system’s performance with the
specified openings is significantly better than open door simulations, which yielded an ASET of 150 s. Fig. 19 shows the smoke spread
and stairwell visibility resulting from the opening and closing of doors.
The time required for smoke to enter the corridors on floors throughout MURB is illustrated in Fig. 20. Open stairwells result in the
rapid spread of smoke to corridors. Closed doors delay the presence of smoke for a significant duration. The operation of doors makes it

Fig. 17. Visibility in the stairwell for open doors and select door openings (along cross-section S1).

14
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Fig. 18. Rate of flow into stairwell compartments, SW1 and SW2, for various door openings.

Fig. 19. Smoke spread and visibility in stairwell results for the Dynamic Door Openings and applied egress model.

Fig. 20. Predicted time for the smoke to enter corridors for door leakage areas with average building envelope.

15
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

likely that smoke will enter the stairwell and other corridors. Fig. 20 does not indicate the smoke concentration, which can be expected
to be significantly higher for the open door scenario. However, smoke in the corridor and egress path may result in increased panic and
trouble in evacuation.
While the presence of all stairwell doors open represents a worst-case scenario, the opening of just some stairwell doors may still
yield significant consequences on the smoke spread. With only stairwell doors on the fire floor open, smoke rises and enters the
corridors slower. The presence of other open doors may accelerate the rise of smoke in the stairwell, as demonstrated when the 10th-
floor doors are open. This scenario results in the “flue” or “stack driven” flow and increases smoke spread to corridors on upper floors.

4.4. Sensitivity Study: influence of open vents

Stairwells may incorporate a variety of openings as part of the design or passive smoke control strategy implemented at the time of
construction. Additionally, smoke control systems may introduce permanent openings, even when not operating. The influence of
several common stairwell openings on the smoke spread was examined to highlight the sensitivity of predictions to variations in vented
scenarios. The simulations of the closed door (220 cm2 and 440 cm2) and the open 2nd floor stairwell doors were used to demonstrate
the influence of various openings. This included an opening of the top of the shaft, the bottom of the shaft, and the corridor vent,
illustrated in Fig. 21.

4.4.1. Stairwell with bottom vent (BV)

The opening at the bottom of the stairwell is required in some codes [21]. When ambient temperatures are less than interior
temperatures, the pressure of the shaft rises and can reduce infiltration from the corridors. In simulations, an opening consisted of a 1.8
m2 vent at the bottom of each stairwell. However, a difference in ambient temperature is necessary to induce such stack pressures. This
resulted in a negligible effect on the smoke spread and tenability and is not shown in Fig. 21.

4.4.2. Stairwell with combined top and bottom vent (TBV)

For some buildings, ventilation of the internal stairwell has relied on openings at the top and bottom of the shaft. The rise of smoke
entering the shaft is intended to be diluted by air entering the bottom [7]. This scenario is simulated with 1.5 m2 vents at the top and
1.8 m2 vents near the bottom. Where a top vent exists, the spread of smoke into the stairwell can occur at an increased rate. However,
some benefits include the dilution of smoke and the exhaust of smoke at the top of the shaft. Fig. 21 shows the spread of smoke
corridors on lower floors was reduced. Amongst the upper floors, smoke infiltration occurred moderately faster. A consequence of the
top vent is that smoke may rise faster and infiltrate corridors faster.

4.4.3. Stairwell with corridor vent (CV)

Opening vents or windows within the corridor or stair vestibule are utilized in some buildings and may satisfy current building
codes as a method of natural ventilation smoke control [20]. The openings can limit the accumulation of hot gas, reduce the pressure of
the corridor relative to the stairwell and help sustain a neutral pressure plane. Its performance may be compromised during cold
ambient temperatures, as it represents a direct leakage path to the exterior, and stack effect may cause infiltration. While this may

Fig. 21. Influence of natural vents or openings at the corridor, bottom and top of the stairwell.

16
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

dilute smoke, it may also promote the flow into stairwells.

In simulations, an external opening reduced the accumulation of smoke and hot gas within the corridor. A significant reduction in
smoke spread occurred where stair doors were closed, extending ASET to 930 s. However, open stairwell doors resulted in a nearly
negligible change in ASET, at just 150 s. This emphasizes the necessity of stairwell doors remain closed for natural ventilation stra­
tegies to be effective.

5. Summary
Smoke spread simulations illustrate the importance of the compartment integrity of the egress path. The leakage and opening
parameters examined show a significant variation in the ASET. While fire size, location, and environmental conditions may influence
the duration of ASET, the simulations reveal the relative performance that may be expected.
• Open Stairwell Doors: In all scenarios involving open stairwell doors on the fire floor, the ASET was between 145 and 150 s.
• Closed Stairwell Doors: In scenarios with closed stairwell doors (220 cm2), ASET was 405 s. While closed stairwell doors are ideal,
the ASET
• Dynamic Openings: It should be recognized that the operation of doors may result in the introduction of smoke into the stairwell.
When considering door openings due to egress, ASET was 260 s.
Open stairwell doors were also noted to exacerbate the sensitivity of smoke spread to environmental conditions. For instance, the
exterior envelope tightness had minimal effect on the smoke spread, except when stairwell doors were open. In which case, a leaky
envelope resulted in an increased rate of smoke spread. While simulations were conducted in the absence of ambient temperature or
pressure differential, results imply that open stairwell doors will increase the smoke spread.
Lastly, open stairwell doors may significantly hinder any natural ventilation that could otherwise limit smoke spread. Simulations
demonstrated that open vents the corridor or vestibule reduced smoke spread into the stairwell. However, the benefit of such openings
was nullified when the stairwell door remained open.

5.1. Limitations and further study

Simulations were conducted in the absence of active fire protection systems such as sprinklers or pressurization systems. Envi­
ronmental conditions did not include wind or stack effect, using the same interior and ambient temperature. Such conditions can
produce significant pressure differentials within the building and affect flow dynamics. Stack effect in a cold climate can increase
airflow into vertical shafts and upper floor corridors. With door openings to stairwells, hot gas flow into vertical shafts may further
increase the stack pressure differential between the ambient environment. Further study on the influence of stack effect and wind is
warranted as it will likely aggravate the spread of smoke throughout.

6. Conclusion
Numerical simulations evaluated the effect of stairwell door openings on smoke spread in a 10-storey MURB. The study utilizes a
residential suite fire scenario with an open corridor door. The smoke spread simulations provide quantification of the hazard when
stairwell door openings exist. A range of parameters was evaluated to reveal the influence of door openings and leakage areas,
combined with occupant evacuation.
Tenability within the stairwells was determined from visibility predictions based on smoke obscuration. Simulations with closed
stairwell doors predicted an ASET of 405 s and formed the baseline for comparison with other scenarios. The ASET was reduced to 36%
of the baseline with open stairwell doors. As evacuations require door operation, some infiltration of smoke from the fire floor may be
expected. The predicted ASET was reduced to 64% of the baseline scenario when door operation was considered using a basic egress
model.
Leakage of the envelope, corridor and vertical shafts can affect the smoke spread significantly. Open stairwell doors on the fire floor
are most problematic; however, additional open doors can further exacerbate smoke spread, turning the vertical shaft into a chimney.
Envelope airtightness had a noticeable impact on smoke spread only when stairwell door leakage was significant. The smoke spread
potential posed by open stairwell doors is higher for leakier building envelopes. Tight envelopes reduced the movement of gas from
corridors into stairwells. However, the effect of the building envelope was negligible when stairwell doors were closed on the fire floor.
The hazard posed by open stairwell doors is apparent from smoke spread simulations. While the smoke spread is subject to many
environmental conditions, the consequence of open stairwell doors is a severe reduction in ASET. The existence of stairwell doors that
are propped open is a hazard that exists in occupied buildings and requires the attention and education of building operators and
residents.

Funding
This work was partially funded by the Natural Sciences and Engineering Research Council of Canada [NSERC DG 2016–04176] and
Ryerson University, Canada.

Author statement
Philip McKeen: Conceptualization, Methodology, Analysis.

17
P. McKeen and Z. Liao Journal of Building Engineering 54 (2022) 104561

Zaiyi Liao: Conceptualization, Methodology, Analysis.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing in­
terests: The authors declare no conflict of interest.

References
[1] Clark County Fire Department, MGM Fire Investigation Report, Las Vegas, 1981.
[2] N. Fandos, Two open doors created ‘flue effect’ of deadly smoke at Bronx high-rise," the New York times [Online]. Available: https://www.nytimes.com/2022/
01/10/nyregion/bronx-apartment-fire-smoke.html, 10 1 2022. (Accessed 22 January 2022).
[3] National Fire Protection Association, NFPA 92 Standard for Smoke Control Systems, National Fire Protection Association, Quincy, MA, USA, 2021.
[4] M. Ahrens, High-Rise Building Fires," NFPA, National Fire Protection Association), Quincy, MA, USA, 2016.
[5] J.H. Klote, Stairwell Smoke Control by Ventilation, ASHRAE, Virginia, 2011.
[6] S. Lay, Pressurization Systems Do Not Work & Present a Risk," Case Studies in Fire Safety, 2014, pp. 13–17.
[7] B. Lane, Grenfell Tower — Fire Safety Investigation," London, 2018.
[8] G.J. Langdon Thomas, G. Ramachandran, Improving the effectiveness of the firecheck door, Fire Int. 3 (27) (1970) 73–80.
[9] "Survey Reveals Scale of Fire Door Abuse, News: Construction Manager Magazine, Construction Manager - News, 2016.
[10] L. Nafiseh, B. Behnam and F. Peyman, "A BIM-based framework for evacuation assessment of high-rise buildings under post-earthquake fires," J. Build. Eng., vol.
43, 2021.
[11] L. Xinyi, Y. Runming, L. Meng, C. Vincenzo, Y. Wei, W. Wenbo, S. Alan, L. Baizhan, Developing urban residential reference buildings using clustering, Energy
Build. 169 (2018) 417–429.
[12] T. Tanaka, Vent flows, in: The SFPE Handbook of Fire Protection Engineering, fifth ed., National Fire Protection Association, Quincy, MA, 2016, pp. 455–485.
[13] J.H. Klote, J.A. Milke, Principles of Smoke Management, American Society of Heating, Refrigerating and Air-conditioning Engineers, Inc., Atlanta, 2002.
[14] G.T. Tamura, Computer Analysis of Smoke Movement in Tall Buildings, National Research Council Canada, Ottawa, 1969.
[15] N.B. Hutcheon, G.O. Handegord, Building Science for a Cold Climate, Institute for Research in Construction (Canada), Ottawa, 1995.
[16] J.F. Straube, E.F. Burnett, Building Science for Building Enclosures, Building Science Press, 2005.
[17] G.T. Tamura, C.Y. Shaw, Studies on Exterior Wall Air Tightness and Air Infiltration of Tall Buildings, ASHRAE Transactions, 1976.
[18] L. Ricketts. A Field Study of Airflow in a High-Rise Multi-Unit Residential Building, University of Waterloo, Waterloo, ON, Canada, 2014.
[19] Nfpa, Standard for Fire Doors and Other Opening Protectives, NFPA, 2019.
[20] IBC, International Building Code, International Code Council, Washington DC, 2015.
[21] National Research Council, National Building Code of Canada, Code of Canada, Ottawa, 2015.
[22] Underwriters’ Labratories, Standard for Air Leakage Test of Door Assemblies," UL, Northbrook, 1784, p. 2015.
[23] British Standards Institute. British Standards BS476, Part 31: Methods for Measuring Smoke Penetration through Doorsets and Shutter Assemblies, AMD8366,
London, 1983.
[24] G.T. Tamura, C.Y. Shaw, Air leakage data for the design of elevator and stair shaft, Build. Eng. (1976) 179–190.
[25] RDH Building Engineering Ltd, Air Leakage Control in Multi-Unit Residential Buildings," Vancouver, 2013.
[26] D. Gross, W. Haberman, Estimating Air leakage through doors for smoke control, Fire Saf. J. (1991) 171–177.
[27] B.W. Gulay, C. Stewart, G.J. Foley, Field Investigation Survey of Airtightness, Air Movement and Indoor Air Quality in High Rise Apartment Buildings Summary
Report, Canada Mortgage and Housing Corporation, Winnipeg, 1993.
[28] J.H. Klote, Considerations of Stack Effect in Building Fires, National Institute of Standards and Technology, Gaithersburg,, 1989.
[29] L. Li, J. Ji, C. Fan, J. Sun, X. Yuan, W. Yuan, Experimental investigation on the characteristics of buoyant plume movement in a stairwell with multiple openings,
Energy Build. 68 (2014) 108–120.
[30] G. Zhao, T. Beji, B. Merci, Study of FDS simulations of buoyant fire-induced smoke movement in a high-rise building stairwell, in: IAFSS 12th Symposium,
Elsevier Ltd., 2017.
[31] A.H. Gager III, G. Dominguez, Tenability criteria in unique situations and atypical buildings, in: The 3rd Fire and Evacuation Modeling Technical Conference,
Málaga, 2016.
[32] D. Gross, W.L. Haberman, Analysis and prediction of air leakage through door assemblies, in: Proc. 2nd Int. Syrup. On Fire Safety Science, Tokyo, 1988.
[33] C.L. Chow, W.K. Chow, Heat Release Rate of Accidental Fire in a Supertall Building Residential Flat, Building and Environment, 2010, pp. 1632–1640.
[34] D.A. Purser, Combustion toxicity, in: The SFPE Handbook of Fire Protection Engineering, Society of Fire Protection Engineers Fifth ed., 2016, pp. 2207–2307.
[35] W. Węgrzyńskia, G. Vigneb, Experimental and numerical evaluation of the influence of the soot yield on the visibility in smoke in CFD analysis, Fire Saf. J.
(2017) 389–398.
[36] J.H. Klote, Tenability and Open Doors in Pressurized Stairwells, ASHRAE Transactions: Symposia, Atlanta?, 2004.
[37] C. Croarkin, G. Mulholland, Specifc extinction coefficient of flame generated, Fire Mater. (2000) 227–230.
[38] F. Ouf, A. Coppalle, J. Yon, J. Vendel, Measurement of the mass specific extinction coefficient of acetylene, toluene and polymethyl methacrylate soot particles,
in: Fire Safety Science, 9, 2008.
[39] K. Richardson, Fire Safety in High-Rise Apartment Buildings, CMHC, 2016.
[40] R.S. Hicks, K. Miller, How to Use NFPA 92 to Design Smoke Control Systems, CFE Media LLC, Downers Grove, 2017.
[41] V. Sanjay, A.K. Das, Numerical assessment of hazard in compartmental fire having steady heat release rate from the source, Build. Simulat. 11 (2018) 613–624.
[42] M.M. Barsim, M.A. Bassily, H.M. El-Batsh, Y.A. Rihan, M.M. Sherif, Performance of impulse ventilation system in an underground car park fire: case study,
J. Build. Eng. 29 (2020) 1–13.
[43] T. Jin, Visibility through Fire Smoke (No. 42), Report of Fire Research Institute of Japan, 1976.
[44] X. Zhou, Numerical Analysis of Fire Performance Index of Teaching Building, Journal of Xi’an University of Science and Technology, 2013, pp. 404–410.

18