Case Studies on Evacuation Elevator Systems in Supertall Buildings

Abstract

In 2009, the International Building Code (IBC) introduced a new requirement for an additional exit stair or evacuation elevators in buildings (except residential buildings) over 420 ft. (128 m) tall. This new requirement has emerged as a critical innovation in occupant evacuation in supertall buildings, offering the potential for faster egress during emergencies. In the 2018 edition of the IBC, the analysis of full building evacuation with elevators is further required to demonstrate an evacuation time of less than 1 h. This paper examines the implementation of evacuation elevators in accordance with recent building code developments, such as the International Building Code (IBC), Chinese “General Code for Fire Protection of Buildings and Constructions” (GB 55037-2022), and the Korean Building Code (KBC). This study provides a comparative analysis of these regulations, highlighting the evolving acceptance of elevator-based evacuation methods. Since the requirements on means of egress in the codes in China and Korea follow similar concepts in the IBC, case studies of supertall buildings in China and Korea, where evacuation elevators were integrated into the overall egress strategy, are carried out to demonstrate the capability of evacuation elevators in achieving the IBC’s requirement for full evacuation within one hour. Using computer-based egress modeling, the study evaluates practical solutions for reducing evacuation times, exploring factors such as elevator capacity, speed, and coordination with traditional stairwell egress. The results suggest that, while evacuation elevators can significantly improve evacuation efficiency, achieving the one-hour target remains a challenge in complex, high-occupancy environments. The study indicates the importance of optimizing the balance between stair and elevator usage and explores the future role of artificial intelligence (AI) in enhancing evacuation systems.

1. Introduction

In traditional building codes, the smokeproof exit stairs are always considered the primary evacuation paths in supertall buildings. The use of elevators is prohibited for the self-evacuation of occupants (i.e., evacuation without being required to be accompanied by the fire service or other emergency personnel) in the event of a fire.

The full evacuation of the World Trade Center towers after the terrorist bombing in 1993, which took approximately 5 h for tens of thousands of building occupants exiting via exit of stairs [1], introduced discussions and studies on the safety of supertall buildings in the fire protection industry. The discussions and studies raised great concerns among public given the fact that it usually takes several hours to evacuate from high-rise buildings via exit stairs, due to the large number of floors and long evacuation distances, especially with occupant loads of tens of thousands in a building.

In the first decade of 2000, the Fire Research Institute of the National Institute of Standards and Test (NIST) of the United States intensified and accelerated the research of safe evacuation through the use evacuation elevators in high-rise buildings, especially the potential use of evacuation elevators in supertall buildings. [2,3]. Based on the research, the 2009 International Building Code (IBC) [4] introduced a new requirement of an additional exit stairway or evacuation elevators in supertall buildings over 420 ft (128 m). It is well understood in the fire protection industry in the United States that the reason behind certain changes in the 2009 IBC of requiring an additional exit stairway in supertall buildings was intended to promote the use of evacuation elevators in supertall buildings [5].

Since then, the developments in elevator technology and design strategies [6,7], such as electrical power and elevator system monitoring, as well as fire alarm system interface, have promoted elevator systems to be utilized for occupant self-evacuation during fire or other emergency situations. The methodology of elevator evacuation has been gradually accepted by the modern building codes, including, but certainly not limited to, IBC, the National Fire Protection Association (NFPA) 5000, Building Construction and Safety Code [8], American Society of Mechanical Engineers (ASME) Elevator Safety Code (ASME A17.1-2010) [9], the Chinese General code for fire protection of buildings and constructions (hereafter referred to as GB 55037) [10], the Korean Building Code [11], etc. The elevator evacuation methodology has been evaluated and incorporated into building codes around the world [12,13,14].

Some research, combined with the analysis of built projects, has identified the elements that should be implemented in tall buildings to reduce risks in the use of evacuation elevators [15,16,17], and to measure their effectiveness against other systems [18]. In the past 20 years, there are more and more supertall buildings around the world that have incorporated evacuation elevators into the designs. Some of the most notable built projects that have benefited from this new and evolving approach include: Burj Khalifa [19] (Dubai, UAE, 829 m, 163 floors), Lotte World Tower [20] (Seoul, South Korea, 555 m, 123 floors), Shanghai International Finance Center [21] (495 m, 101 floors). The further questions that have been raised for the evacuation elevators in supertall buildings are focused on how many elevators should be provided for the tens of thousands of occupants in the building, and what the criteria should be.

A new requirement was added in 3008.1.1 of the 2018 edition of the IBC, which requires that the number of elevators available for occupant evacuation shall be determined based on an egress analysis on one of the two scenarios:

1. Full-building evacuation, where the analysis demonstrates that the number of elevators provided for evacuation results in an evacuation time less than 1 h.

2. Evacuation of the five consecutive floors with the highest cumulative occupant load where the analysis demonstrates that the number of elevators provided for evacuation results in an evacuation time less than 15 min.

The full building evacuation of a supertall building is considered in this study. In the fire protection industry, few studies have been performed to evaluate the full evacuation time in supertall buildings with evacuation elevators [22,23]. The studies that compare the results of full evacuation with the newly proposed 1 h criteria in IBC, especially, are very limited. Since the requirements regarding the means of egress in the codes in China and Korea follow similar concepts in the IBC, although some requirements are more conservative in the Chinese and Korean codes, the egress modeling simulations based on the real projects in this study can also reflect the evacuation systems in supertall buildings with the IBC requirements. In Chinese GB 55037, refuge floors/Areas of Refuge (AOR) are required in supertall buildings, and the exit stairs are required to transfer at the refuge floors. Similarly, KBC requires a refuge floor every 30 floors or fewer in a high-rise building, and exit stairs connected to the refuge floor should be installed in a way that allows for access to the upper and lower floors though refuge areas. Also, at least one of elevators to be provided in a building should be utilized as an occupant evacuation elevator with additional safety measures. In this study, two real supertall projects with evacuation elevators designed with the requirements of the Chinese GB 55037 and KBC, respectively, are reviewed with Pathfinder egress modeling in this study. The case studies are performed to demonstrate the practical solutions to reducing the time of full building evacuation to the 1 h limitation using evacuation elevators in supertall buildings.

2. Comparison of the Major Requirements on Means of Egress in IBC, Chinese GB 55037 and KBC

A comparison on the requirements of the means of egress in IBC, Chinese GB 55037 and KBC found that the Chinese GB 55037 and KBC generally have more stringent requirements for the evacuation systems. The comparison is summarized as follows.

2.1. Vestibules at Exit Stairs and Stair Pressurization Systems

IBC requires a vestibule to be provided at the exit stair for smokeproof exit stairs in high-rise buildings or a pressurization system to be provided in stairwells instead of vestibules. The configuration and method of operation of the pressurization systems sometimes may be modified by amendments to the IBC in some major cities in US. The amendments are made to adapt local climate conditions, also accommodating the practices and protocols of the local fire services. For instance, some major cities [24] in Texas require stair vestibules and stair pressurization systems in high-rise buildings.

The GB 55037 requires both stair vestibules and stair pressurization systems in high-rise buildings. The requirement for pressure difference and supply injection points in the Chinese GB 55037 are not identical to those in the IBC. However, the Chinese GB 55037 is considered more stringent than IBC on this requirement.

KBC, regarding the requirements of the pressurization system for stairwells and vestibule in high-rise buildings, is similar to the Chinese GB 55037. KBC requires vestibules for the exit stairs in high-rise buildings. In addition, KBC requires the pressurization systems in buildings greater than 11 stories and 3 stories below grade.

2.2. Refuge Floors and the Segmentation of Each Exit Stair

The requirement of refuge floors, and the segmentation of an exit stair on the refuge floors, are perhaps the most significant differences in the life safety techniques utilized between IBC, GB 55037 and KBC.

The IBC does not require the refuge floors in tall buildings.

GB 55037 requires a refuge floor every 50 m in height in high-rise buildings, and the exit stairs are required to break and transfer at the refuge floors. This requirement can potentially help to reduce the large stack effect in stairwells, and prevent smoke blocking the entire exit stair from the bottom to the top. However, the gathering of large numbers of occupants on refuge floors may create other risks related to large crowd control [25]. These risks can be mitigated through the careful design and configuration of the refuge floor and related elements, as well as the alert management of occupant evacuation during an emergency event [26,27]. These risks need to be addressed on an individual basis for each project. For assembly areas with large numbers of occupants (such as those occupants that can be found on refuge floors) the IBC requires an increase in exit-stair capacity to accommodate the increased occupant load. Although the GB 55037 is more stringent on the refuge floor, it is less restrictive in terms of the quantity of exit stairs and the related elements required for the large number of occupants on refuge floors [28,29].

In Korea, refuge floors are required in high-rise buildings. Refuge floors should be located every 30 floors and protected by a mechanical pressurization system or a natural ventilation system. According to the KBC, the size of refuge floors is required to be able to hold 50% of the total combined occupant load from all the floors located in between the two refuge floors (if a fire occurs on a refuge floor, the occupants on that level will travel down to the next refuge floor or keep traveling down the stairs). Based on this occupant load, the refuge area must be sized by providing a minimum of 0.28 square meters of clear floor area space per person.

2.3. Fire Standpipes and Indoor Fire Hydrants

The IBC (section 905) requires fire standpipes to be installed at the landings in the exit stairwells. This configuration provides certain protections for firefighters.

The Chinese standard GB 50974-2014 “Technical Specifications for Fire Water Supply and Fire Hydrant Systems” [30] 7.4.7 requires that the indoor fire hoses in high-rise buildings be in the stairwell, on the stair landings, or in the stair vestibules. When the indoor fire hoses in high-rise buildings are provided in stair vestibules, the fire hoses are easy to access for firefighting and the required protection for firefighters is provided. In addition, disruptions between the occupant evacuation flows and the firefighters operating the fire hoses in the stairwell can be reduced.

Korean Fire Code (KFC) also required indoor fire hydrants in every floor of a building. Fire hoses from the hydrants are required to cover all the area on each floor. Two risers for hydrants should be installed in high-rise buildings. Per KFC, the indoor hydrants are not allowed inside the stairwell. Therefore, it is not likely that the occupant evacuation flows will be interrupted by the operating of a fire hose.

2.4. Fire Service Elevators

The IBC (sections 403 and 3007) requires all buildings with an occupied floor over 36.6 m above the lowest level of fire-service vehicle access to be provided with a designated fire-service access elevator that is protected with an elevator lobby no less than 14 square meters in size and be provided with protected access to an exit stair.

GB 55037 (Section 2.2) requires that supertall buildings must be equipped with fire service elevators and elevator lobbies. Firefighters can usually approach the fire floor in an emergency via fire service elevators without blocking the exit stairs.

Per the KBC, fire service elevators shall be provided in buildings over 31 m in height. The firefighter’s elevator should have a dedicated lobby. It should not be shared with a vestibule of a special egress staircase and shall be separated by a 2 h fire separation. The KBC requires the fire-service elevator lobby and the vestibule of the exit stair at different locations.

2.5. Evacuation Elevators

In buildings (other than residential occupancies such as apartments, condominiums, or hotels) more than 128 m in height, the IBC (section 403) requires an additional interior exit stair, in addition to the minimum number of exit stairs typically required. This additional stair can be omitted if a compliant occupant evacuation elevator system is provided. The requirement of an additional exit stairway in supertall buildings in the 2009 IBC was intended to promote the use of evacuation elevators in supertall buildings.

No. 57—“Enhanced Technical Requirements for Fire Protection Design of Buildings with Height Greater Than 250 Meters (2018)” [31]. No. 57 Ordinance requires that supertall buildings over 250 m shall be equipped with evacuation elevator(s) and an additional exit stair as well.

Per the KBC and the City of Seoul Ordinances, high-rise buildings should be designed with the required elevator that can be used in an emergency to implement an emergency evacuation strategy that includes the use of the building’s elevator systems to assist in evacuation. No additional exit stair is required.

With all the other requirements discussed in Section 2.1, Section 2.2, Section 2.3, Section 2.4, Section 2.5 above, the requirements for an additional exit stair and evacuation elevators sometimes might be unnecessarily stringent. This analysis utilizes computer egress modeling to evaluate the efficiency of the evacuation elevator in a supertall building in different scenarios and conditions.

3. Egress Modeling Simulations on Supertall Buildings

For large and complex buildings, egress studies can be carried out by evacuation drills, but full evacuation drills require considerable time and cost for planning and execution. Although partial evacuation drills are sometimes conducted to familiarize occupants with the egress protocols for their building and to evaluate their response time, drills of any type may interrupt normal operations and daily business, and are extremely costly. An egress modeling analysis [32] that simulates the egress of pedestrians from buildings has been developed and refined over decades. Currently, such software includes, but is not necessarily limited to EXODUS [33], PEDFLOW [34], and Pathfinder [35].

The Pathfinder model has been used in this analysis. It is an agent-based egress simulator that uses steering behaviors to model occupant motion. It consists of the three following modules: a graphical user interface, the simulator, and a 3D results viewer. Pathfinder uses a combination of steering mechanisms and collision handling to control how the occupant follows their seek curve. These mechanisms allow the occupant to deviate from the path while still heading in the correct direction toward their goal. The movement environment is a 3D triangulated mesh designed to match the real dimensions of the building being modeled. Additional technical information for Pathfinder can be found in the Pathfinder user’s guide and in the Pathfinder Technical Reference [36]. In the Pathfinder model, the simulated movement of the agents is calculated over a triangulated mesh on each floor. Floors are connected with exit stairs and elevators. The validation studies of the Pathfinder model have been performed to verify the fundamental speed–density and specific flow–density simulations with the SFPE correlations. The comparisons show that Pathfinder correctly simulates the speed–density curve in the SFPE Handbook [37], and the capability of evacuation models to simulate evacuation using elevators was verified, which were reported in the Verification and Validation document of Thunderhead Engineering in 2015 [38]. The predictions of the Pathfinder model were reported as satisfactory.

The premovement delay was not considered in these Pathfinder modeling studies. It was assumed that all occupants start moving without any delay at time 0 (time of alarm and notification for full building evacuation). All occupants were familiar with the locations of the exit stairs and the ground exits. No panic or chaos was included in the model and no occupant needs assistance. In reality, the delay, unfamiliarity, panic, and assistance may all slow down the egress process significantly. This study just considers the very basic egress scenario and human behavior, so as to focus on the impacts from the evacuation elevators.

This egress modeling analysis on the supertall building in China is a continued study of a previous one documented in the “Study on the Reliability of the Exit Stair System in the Super High-rise Buildings Based on Egress Modeling Analysis” [39] in 2020. Evacuation elevators are then added in the protected refuge floors in the project building to evaluate the efficiency of the evacuation elevator. In the modeling analysis, occupants reach the refuge floors via exit stairs; some occupants may choose to wait for the evacuation elevators, the others continue down via the exit stairs.

A supertall building designed in Korea is also studied to compare the evacuation strategies per KBC. Multiple lifeboats from several sky-lobby levels are studied with the egress modeling analysis. In the egress models, occupants are assumed able to walk without aid. The occupant walk speed is defined as a function of density and terrain, based on the SFPE Handbook. The maximum walk speed on horizontal surfaces is assumed as approximately 1 m/s, which is similar to brisk walking. The maximum walking speed on stair is approximately 0.75 m/s. The fire is assumed as limited to the fire floor and not impacting any elevators.

3.1. The Simulated Building in China

In this study, a supertall building project in China is evaluated, which is designed by Skidmore, Owings and Merrill (SOM) [40].

1.

Design of the Building in China

The building is an 87-story tower, 407 m in height, which consists of a commercial podium (floors 1–8), office floors (floors 11–20, 22–31, 33–41 and 43–51), service apartments (floors 53–62), a hotel (floors 64–76), and observation and club house levels (floors 77–78). Refuge floors are dispersed between these sections on floors 9–10, 21, 32, 42, 52, 63 and 74. The section and floor plans of the subject building are illustrated in Figure 1.

The occupancy densities are calculated as 10 m²/person in the offices, 30 m²/person in mechanical plant rooms, and two persons per room in the hotel. The total occupant load of the building is estimated at approximately 16,151 persons. Chinese code requires the exit stair to be a minimum of one meter in width for every 100 occupants. The office floors are approximately 3000 m²; after deducting the building core, the office area is approximately 2000 m², which results in approximately 200 occupants on the office floor. The egress was simulated with scenarios of two (2) exit stairs, three (3) exit stairs, and one scenario of the three (3) exit stairs blocked in the section between two refuge floors; refer to the stair in the red circle in Figure 1. A fire is assumed occurring on Floor 25.

2.

Evacuation Elevator in the Building (China)

It is intended that one of the fire service elevators will be operated by the security staff of the building or by the fire department during an emergency. The other one will serve as the evacuation elevator, and will pick up evacuees on refuge floors; the evacuees are then transported to and discharged on the ground floor. The occupants on each floor reach the refuge floors via exit stairs. Some occupants will wait there for the evacuation elevator, while other occupants will continue to evacuate down the stairs.

According to Bukowski’s studies on evacuation elevators, the basic operation of elevators in evacuations in this egress model can be summarized as follows:

Each elevator has one discharge floor, where the elevator starts at the beginning of the simulation and where it will take occupants it has picked up. In this study, the evacuation elevator has seven (7) pickup floors, which are the refuge floors’ areas.

The elevator is called to a pickup floor by an occupant when they come within one-half meter of the elevator door.

The elevator uses a priority system to serve called floors. By default, floors are served from top to bottom; however, other floors can be given higher priority to simulate fire floors. Once the elevator has picked up occupants (nearly always completely filling the elevator car), it will only travel to the discharge floor, where it will discharge all the occupants. It will not stop at any other floors to pick up more occupants.

One of the two fire service elevators (lifts) is assumed to be used as the evacuation elevator, and it is highlighted in Figure 2.

As a reference, the super high-speed elevators installed in Shanghai Tower in 2017 are Mitsubishi Electric-designed elevators capable of traveling at an incredible peak speed of 20.5 m per second (67 ft/s). The evacuation elevator simulated in this study utilized the speed of 10 m/s. The other characteristics of the evacuation elevator are listed in

Table 1. Evacuation elevator in the supertall building (China).

Evacuation Elevator	High Speed Elevator
Nominal Load	25 persons
Maximum Velocity	10 m/s
Acceleration	2.5 m/s2
Door Open + Close Time	7.0 s
Elevator Door Width	1.25 m

:

The time for the elevator traveling between the discharge floor and the pickup floor is calculated as:

$$t=t_{up}+t_{load}+t_{down}+t_{unload}$$

(1)

where t_up is the time from the ground floor to the pickup floor.

t_down is the time from the pickup floor (refuge floor) to the ground floor.

t_load is the pickup time, including door open and close time.

t_unload is the unload time, including door open and close time.

The highest pickup floor is the refuge floor on Floor 74 (340 m). Assuming occupants enter the elevator by 2 persons/second, it will take approximately 12.5 s to load/unload 25 persons (elevator capacity), and then the time for the super high-speed elevator running a cycle is slightly less than 2 min, determined in seconds as follows:

t_up = Acceleration time + travel time at maximum speed + deceleration time = 4 + 30 + 4 = 38 s.

t_load = Door open and close time + time for occupant entering the elevators = 7 + 12.5 = 19.5 s.

t_down = Acceleration time + travel time at maximum speed + deceleration time = 4 + 30 + 4 = 38 s.

t_unload = Door open and close time + time for occupant exiting the elevators = 7 + 12.5 = 19.5 s.

3.

The simulated scenarios (China)

The scenarios simulated for the supertall building in China are labeled as C1 to C4.

C1

Two (2) exit stairs are available, no evacuation elevator.

C2

One additional exit stair is provided per the No. 57 Ordinance, totaling three exit stairs.

C3

One of the three exit stairs is blocked between two refuge floors (F22–F31), with no evacuation elevator.

C4

One exit stair is blocked between two refuge floors, and one evacuation elevator is provided.

In scenarios C3 and C4, Floor 25 is assumed as the fire floor. The exit stair highlighted in the red circle in Figure 1 is assumed to be blocked between floors 22 to 31. The modeled building is illustrated in Figure 3.

3.2. The Simulated Building in Korea

The building simulated is one of the supertall buildings in Korea, which is approximately 600 m high.

1.

Design of the Building in Korea

The 100+ story building consists of offices, observation decks, lobbies, kitchens, and a restaurant. It also includes basement levels for parking, mechanical rooms, auditorium, gym facilities, cafeteria, kitchen, and lobbies.

There are approximately 40 elevators, including shuttle elevators, passenger elevators and service/fire-service elevators in this project. Also, three (3) exit stairs and five (5) refuge floors, that are dispersed within 30 floors in this building according to the KBC requirement. Figure 4 shows the floor plan of Floor 30 in the project.

2.

Evacuation elevator in the building (Korea)

A set of designated elevators are designed to be used as lifeboats/evacuation elevators during an emergency, which allows them to be used as an additional means to evacuate occupants in the building. During emergency evacuation, trained staff/operators will control the elevators to shuttle occupants from the designated refuge floors to the primary exit discharge floor. In this analysis, approximately 30 out of the 40 elevators in the building were used as lifeboat elevators. Firefighter’s elevators are not counted as evacuation elevators. Pathfinder egress modeling is used to analyze the efficiency of the evacuation elevators in this supertall building, by estimating the evacuation time with (1) the use of stairs only, and (2) the use of exit stairs and associated evacuation elevators from refuge floors.

According to the 5th Edition of the SFPE Handbook, the average walking speed of 1.1 m/s was used as the maximum walking speed for all occupants. The travel speed associated with occupants is specified as the input for the egress models. The design parameters of the evacuation elevators provided by the elevator manufacturer are as listed in

Table 2. Evacuation elevator in the supertall building (Korea).

Refuge Floor	Elev. Capacity (Persons)	Acceleration (m/s2)	Max. Velocity (m/s)	Open/Close Time (s)
102	27	13	10	2.8/3.9
83	30	11.7	9	2.8/3.9
58	30	11.7	9	2.8/3.9
33	30	7.9	6	2.8/3.9
07	27	6.9	5	2.8/3.9

.

Figure 5 shows a diagram of the evacuation system of the building, including evacuation elevators, exit stairs and refuge floors. Each dedicated evacuation elevator group consists of 4 to 6 elevators serving between a pickup floor to the ground floor. These elevators are used as passenger elevators in normal operation and converted to evacuation elevators during entire building evacuation.

3.

The simulated scenarios (Korea)

A total building evacuation simulation is conducted to estimate the evacuating time using the exit stairs in the core, as well as to assess the efficiency of lift-assisted evacuation. For this study, Pathfinder is used in the modeling analysis. This evacuation analysis includes all the occupied floors of the tower above ground, while the occupants on the basement levels are excluded from this study. The scenarios simulated for the supertall building in Korea are summarized as K1 to K3.

K1

Entire building evacuation; all three exit stairs are available, no elevators.

K2

A total of 50% of the occupant load used the exit stairs and the remaining 50% of the occupants used lifeboat elevators.

K3

A total of 35% of the occupant load used the exit stairs and the remaining 65% of the occupants used lifeboat elevators.

Figure 6 shows the Pathfinder model and the occupants evacuating using the lifeboats/evacuation elevator at the refuge floor.

4. Simulation Results

The Pathfinder modeling results of the supertall buildings in China and Korea are summarized in the following sections.

4.1. The Building Simulated (China)

Four (4) evacuation scenarios were considered, with an occupant load of approximately 16,000 in the supertall building. Scenario C1 is considered as an ordinary evacuation, with two exit stairs available, and occupants can use any exit stairs for evacuation. In Scenario C2, one additional exit stair is added per the No. 57 Ordinance of China Fire Protection Bureau, and all three exit stairs are available. In Scenario C3, one of the three exit stairs is blocked between two refuge floors. In Scenario C4, one evacuation elevator is utilized. The highlighted yellow vertical band in Figure 7 represents the hoistway of the evacuation elevator.

Figure 8 shows the operation of the evacuation elevator in the Pathfinder model. The evacuation elevator runs between the first floor to the refuge floors to pick up the evacuees, and the model assumes optimal evacuation decisions, in which the occupants estimate the time of waiting for the elevator versus the time of exiting through the exit stairs, and then selects the more optimal decision.

Scenario C1 required approximately 238 min to evacuate the entire building. When an additional exit stair is added in C2, the total evacuation time was 170 min. One of the exit stairs is blocked between the fire section in C3, which resulted in approximately 186.5 min being needed to evacuate the entire building. Scenario C4 demonstrates the congestion by occupants waiting for the elevator in the elevator lobby, which obstructs the egress movement of occupants via the exit stair in the shared elevator lobby and the stair vestibule (See Figure 8). This kind of congestion reduces the efficiency of the exit stair. However, this congestion can be significantly reduced by (1) providing a larger elevator lobby, or (2) arranging an elevator lobby and transfer stairs not to interrupt evacuation flows. The simulation results are summarized in

Table 3. Comparison of evacuation time in the supertall building in China.

Scenarios	Note	Evacuation Time
C1	Ordinary Evacuation: Two stairs available	14,268 [s] = 238 min.
C2	Ordinary Evacuation: Three stairs available	10,188 [s] = 170 min.
C3	One of three stairs is blocked between Floors 22–32, no elevator	11,187 [s] = 186.5 min
C4	One of three stairs is blocked between Floors 22–32, one evacuation elevator used	9880 [s] = 165 min

. One elevator is added in Scenario C4 based on Scenario C3, and the evacuation time is reduced by approximately 20 min compared to C3. It is expected that the optimized evacuation scenario involves all three (3) exit stairs and one elevator, which can be derived from Scenario C2 with 20 min reduction in evacuation time, resulting in approximately 150 min (2.5 h) of full evacuation time.

Figure 9 illustrates the evacuation time in the above four scenarios. The evacuation time of Scenario C1 is considered as a reference for the comparison of the evacuation time with other evacuation scenarios. In Scenario C3, due to the blocking of one exit stair in a section of the fire floor, the evacuation speed decreased, resulting in an increase of about 10% in the overall evacuation time from C3 to C2. In Scenario C4, even with the obstruction of a stair in the section where the fire occurs, the use of a high-speed elevator for evacuation shows that the entire evacuation process is essentially the same time required for the evacuation time with three stairs in the Scenario C2. It is found that the evacuation time is longer than 2.75 h in all the four scenarios.

It is found that an additional stair can reduce the full evacuation time from approximately 4 h to 3 h; when fire blocks one of the exit stairs in the section where the fire occurs, the evacuation time slightly increases 15 to 20 min. With an evacuation elevator and an additional stair that is blocked in the fire section (Scenario C4, which is the scenario described in the Chinese No. 57 Ordinance) the 15-to-20 min loss in evacuation time in Scenario C3 can be fully compensated by the section of the exit stair, and the total evacuation time can be reduced to less than 3 h. The simulation indicates that with one additional stair and an evacuation elevator, the full evacuation time can be reduced significantly, but it is still far away to the 1 h goal. Adding more exit stairs is not favorable, due to the construction cost and the loss of floor area. Using more of the elevators which are used in normal operation for evacuation seems to be another path. The next case study, in which as many as 30 elevators are used as evacuation elevators, based on a supertall project in Korea, is discussed as follows.

4.2. The Simulated Building (Korea)

Three (3) evacuation scenarios were considered with an occupant load of approximately 28,000, which is estimated based on both International Building Code and Korean Building Code.

Table 4. Evacuation by elevators in the supertall building in Korea.

Refuge Floor (Evacuation Elevators)	Occupant Loads (Total Loads Served by the Elevators)	Evacuation Time of the Floor
102	903	1635 (s) = 27.0 min.
83	3221	4016 (s) = 67.0 min.
58	4293	4701 (s) = 78.5 min
33	4271	3537 (s) = 59 min.
07	4491	3028 (s) = 50.5 min.
Total	17179	98 min

shows the summary of occupant loads using evacuation elevators from the refuge floors and the fastest evacuation time of 98 min is reached when approximately 60–70% of the total occupant loads use lifeboats/evacuation elevators. The occupants on the floors below Floor 7, where the first refuge floor and evacuation elevators are located, used exit stairs to evacuate. According to the occupant load factors in KBC and IBC, there are approximately 1500 to 2400 persons estimated on floors 1 to 6 in total. Each floor has 250 to 400 persons, and each of the three exit stairs quickly reaches at its full capacity.

Figure 10 illustrates the evacuation time in the three scenarios.

Table 5. Evacuation scenario summary and results (Korea).

Scenarios	Description	Total Evacuation Time
K1	100% of occupants using three exit stairs.	163 min
K2	50%/50% of occupants using stairs and lifeboat elevators, respectively.	109 min
K3	30%/70% of occupants using stairs and lifeboat elevators, respectively.	98 min

shows the scenarios’ summary and the total evacuation results for each scenario. The total evacuation time is 163 min, which is a result of 100% of the building population only using the exit stairs for evacuation. Using elevators in addition to the exit stairs helps reduce the evacuation time.

However, the elevator usage comprises only 4701 s out of the total evacuation time of 5933 s (98 min), which occurs for about (4701/5933) = 79% of the total evacuation time. In other terms, the usage time of the stairs is about 20% longer than the usage time of the elevators. The occupants on the upper floors spending 21 more minutes descending via stairs could have been making use of the capacity in the elevators. Therefore, increasing the percentage of occupants using elevators, especially on the upper levels, could significantly reduce the total evacuation time. Nonetheless, the total evacuation time of the supertall building in Korea is found to be reduced from almost three (3) hours to approximately 1.5 h by using dozens of evacuation elevators.

Apparently, how to direct evacuees to use appropriate evacuation routes and improve the efficiency of evacuation elevators and exit stairs is a big concern and challenge in real emergency procedures concerning full evacuation in supertall buildings. The simulation of the building in China indicates that one additional exit stair and one emergency elevator can reduce the full evacuation time a little bit, but the evacuation time is still much longer than 1 h. One elevator is far from enough for evacuation. The simulation of the Korean building indicates that as many as 30 elevators in a supertall building as the scale in this study should be considered for evacuation. The simulation of the Korean building shows that the efficiency of elevators and exit stairs can be largely impacted by the numbers of occupants who select elevators or stairs. When many elevators are provided to tens of thousands of evacuees, the complexity of the emergency operation is far beyond the capability of the traditional method of emergency operation. It is expected that the efficiency of the evacuation elevators and occupant evacuation can be significantly improved by applying artificial intelligence (AI) technology to the evacuation systems.

5. Monitoring and Control of the Evacuation Elevators

Due to the large number of floors and long evacuation distances, especially with occupant loads of tens of thousands in the building, it usually takes more than 2 h to evacuate from supertall buildings by using exit stairs. In the evacuation of the World Trade Center towers following the terrorist bombing in 1993, the evacuation time took approximately 5 h for tens of thousands of building occupants to successfully and safely traverse some five million flights of stairs.

The two supertall projects in Asia in this study indicate that the evacuation elevators may be a solution to significantly reducing the total evacuation time in supertall buildings. The evacuation elevators should be carefully designed and installed according to the fire protection requirements and equipped with appropriate control systems. Artificial intelligence (AI) may be able to greatly improve the efficiency of using evacuation elevators, by directing occupants waiting for the next evacuation elevator or walking down the exit stairs.

The following functions can be achieved in the elevator evacuation module:

(1) Information can be collected via the closed-circuit television cameras (CCTV) that are installed in elevator cars, stairs, and refuge floors. The crowd movements during an evacuation will be monitored by CCTV. The digital image processing methods can be applied to track pedestrian movements and analyze the pedestrian tracking algorithms to estimate the crowd movement behavior [41].

(2) Information transmission: Through uplink and downlink data communication, the timely and accurate transmission of front-end data is achieved in the event of unstable communication signals. During emergency evacuation, reliable audio and video calls can be made, and elevator dispatch commands can be issued correctly.

(3) Information management and real-time monitoring: A software platform will be developed based on B/S architecture to achieve multimedia information management and the real-time monitoring of elevator operation status, as well as the internal and external environment of elevator cars [42].

(4) The collected on-site information and the crowd movement behavior will be used to estimate the evacuation times by elevators and by stairs. The analysis results will be used to control the evacuation signals and speakers by the central dispatch system, directing the occupants to use appropriate exit paths.

The central dispatch system will control the elevator controller through the elevator evacuation control terminal in emergency. The evacuation controller will send control commands to the evacuation elevators automatically, based the AI system [43].

The efficient usage of evacuation elevators and exit stairs will be achieved by evacuating occupants ideally using the elevators and exit stairs at the same time. AI technology can be used to greatly increase the efficiency of the evacuation elevators [44,45].

6. Conclusions

The evacuation strategies of supertall buildings were studied by scenarios involving (1) two exit stairs; (2) two + one additional exit stair; (3) two stairs + one additional exit stair + one evacuation elevator; and (4) multiple lifeboats from several refuge levels with egress modeling analysis based on two supertall projects in Asia. The study indicates that adding exit stair(s), which is very costly in terms of construction and ordinary operation, could only reduce the evacuation time by a small amount (from 4 h to 3 h in the model); using elevators, which can be used as passenger elevators under ordinary operation, could significantly reduce the evacuation time to close to 1 h, although the 1 h goal has not been achieved yet in the study.

The evacuation system in supertall buildings in China was established in accordance with the requirements of No. 57 Ordinance of China Fire Protection Bureau, i.e., both an additional stair and an evacuation elevator are provided. This is more restrictive than the requirements in IBC, which provide the choice of either an additional stair or the use of occupant evacuation elevators, and those in KBC as well. The No. 57 Ordinance requirements appear to be unnecessarily conservative in supertall buildings with the additional exit stair in addition to evacuation elevator. The use of more additional evacuation elevators would have great potential to further improve the total evacuation time.

The study with the supertall building in Korea indicates that evacuation elevators expedite the evacuation process, significantly reducing the total evacuation time from the building. The simulation results show bottlenecks being created near the elevator lobby entrances in the refuge floors, as occupants queue in the lobby area to use the lifts. Stair transfers also occur at each refuge floor, which could also exacerbate the situation; hence, determining the optimal balance between occupants using the stairs and lifts could be a critical factor in affecting the overall evacuation time.

To avoid congestion in front of the evacuation elevators, some consideration should be taken in the design, as follows: The lobbies of the evacuation elevator on the pickup floors should not overlap with the flow path of the exit stairs. The lobbies should be able to accommodate 100 to 200 persons with a density of approximately 1.5–2.5 persons/m² (4–7 ft²/person), etc.

As required in the IBC, Chinese GB 55037 and KBC, all evacuation elevators and their lobbies in supertall buildings should be enclosed with fire-resistant-rated shafts with a fire resistive rating of no less than 2 h. In addition, the shaft should be waterproofed to protect the emergency equipment from the water discharged from sprinklers, and a smoke-proof enclosure should be provided to protect the occupants. All the redundant and reliable emergency power should be provided to the emergency equipment.

This study indicates that the emergency/evacuation elevators should be the best solution to reducing full evacuation time in supertall buildings. It would be expected that the efficiency of evacuation elevators in supertall buildings can be greatly boosted by AI technology, by directing the evacuees’ movements to appropriate paths towards evacuation elevators or exit stairs; when dozens of elevators are provided to tens of thousands of evacuees, the complexity of the emergency operation is far beyond the capability of the traditional method of emergency operation. Further studies should be performed to analyze how the AI technology could be incorporated in the evacuation systems of supertall buildings.