Next Article in Journal
Environmentally Sustainable Lithium Exploration: A Multi-Source Remote Sensing and Comprehensive Analysis Approach for Clay-Type Deposits in Central Yunnan, China
Previous Article in Journal
Optimization of Freshwater–Saline Water Resource Mixing Irrigation Under Multiple Constraints
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Case Report

Optimizing Key Evacuation Features for Safer Egress in Complex Buildings with Underground Connections: A Simulation-Based Approach to Resilient and Sustainable Design

1
Urban Infrastructure Research Division Disaster Management Research Center, Seoul 04516, Republic of Korea
2
Department of Architectural Engineering, Inha University, Incheon 22212, Republic of Korea
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(8), 3731; https://doi.org/10.3390/su17083731
Submission received: 17 February 2025 / Revised: 10 April 2025 / Accepted: 17 April 2025 / Published: 21 April 2025
(This article belongs to the Section Hazards and Sustainability)

Abstract

:
This study explores the impact of key evacuation features on occupant safety in complex buildings with underground connections in Seoul, the city with the highest concentration of such buildings in the country. By analyzing factors like exit spacing, exit width, stairwell distances, and stairway configurations, the study assesses evacuation safety using fire and evacuation simulations, comparing available safe egress time (ASET) with required safe egress time (RSET). Reducing interior exit facility spacing from the legal standard of 100 m to 50 m improved evacuation time by 77.5% (from 36 min to 8 min and 7 s), with a further reduction to 40 m improving performance by an additional 23.3% (to 6 min and 13 s). In downward evacuations, reducing the walking distance to exterior exits from over 50 m to 30 m cut evacuation time by at least 59.9% (from 23 min and 55 s to 9 min and 35 s), ensuring successful evacuations. These findings demonstrate that optimizing evacuation routes, addressing bottlenecks, and improving evacuation feature standards can significantly enhance safety and minimize casualties. By adjusting building design and fire safety regulations, these optimizations promote resilient urban infrastructure, reduce disaster-related socio-economic impacts, and inform evidence-based policies, offering valuable insights for policymakers and guiding future improvements in fire safety and evacuation protocols.

1. Introduction

As of 2024, Seoul has 211 complex buildings with underground connections, the highest among all Korean cities [1]. According to Article 2 of the Special Act on Management of Disasters in Super High-Rise Buildings and Complex Buildings with Underground Connections, a complex building with underground connections refers to a structure with at least 11 floors or an accommodation capacity of 5000 persons, with its underground section connected to a subway station or underground shopping arcade. These buildings often house facilities for cultural activities, commerce, transportation, businesses, lodging, and amusements [2]. Hence, the failure of the initial response after a disaster can lead to significant human and property damage in such buildings [3,4,5,6,7]. Specifically, through underground passages in these buildings, fires can spread to nearby structures such as large-scale retail facilities and multiplex cinemas, increasing evacuation times, and exacerbating confusion among occupants [8,9].
In 2021, a fire originating from an underground restaurant kitchen in a composite building in Namyangju City, Gyeonggi Province, spread rapidly because of inadequate fire compartmentalization and malfunctioning fire shutters. Further, the smoke continued to spread to nearby stations for over seven hours, resulting in 41 injuries and approximately KRW 9.4 billion in property damage [10]. Additionally, the enclosed nature of underground spaces often creates a sense of confinement, psychological distress, and fear of structural collapse, which can make escape more challenging for occupants [8,9].
To prevent potential large-scale casualties, the Korean government has established evacuation standards as part of building regulations and fire-related laws. As part of the performance-based fire safety design, egress safety evaluations are conducted, and their outcomes are integrated into building designs. Engineering-based approaches such as computer simulations are often employed to assess the effectiveness of alternative floor plans and fire safety equipment, allowing for adjustments during the design [11]. These considerations are reflected in building management, safety education, and training programs for occupants.
Using fire and evacuation simulations, this study aims to analyze the impact of key evacuation features on evacuation safety in complex buildings with underground connections. Specifically, it focuses on features that directly affect evacuation outcomes, including exit spacing, exit width, stairwell distances, direct stair access to ground floors, stairway and landing widths, and effective widths of stairway entrances. This study examines the impact of modifications to evacuation features on safety. Three evacuation simulation scenarios are analyzed: the first represents the building’s original design, while the second and third incorporate enhancements to key evacuation features, specifically addressing the critical bottleneck areas identified in the initial analysis. Egress safety is assessed by comparing the available safe egress time (ASET)—the time required by occupants to safely evacuate before hazards such as smoke occur—with the required safe egress time (RSET)—the time required for complete evacuation. To determine ASET, the Fire Dynamics Simulator (FDS) was used to calculate visibility (m), temperature (°C), and carbon monoxide (CO) concentrations (ppm) based on the initial fire conditions. Successful evacuation requires RSET to be shorter than ASET; otherwise, evacuation safety is compromised. Evacuation times were analyzed using Simulex software (VE2015), and strategies to improve bottlenecked evacuation features were developed based on the simulation results. This study proposes managerial implications and strategies to enhance egress safety in complex buildings with underground connections. By bridging architectural design, regulatory standards, and fire safety simulations, this study provides a comprehensive, evidence-based approach to enhancing egress efficiency in underground connected buildings, setting a precedent for future evacuation planning in high-density urban environments.

2. Research Methods

Computer modeling simulations aid in assessing the critical conditions and overall safety of underground spaces, allowing for comparisons between different evacuation feature options [12,13]. Evacuation simulations help analyze fire scenarios by evaluating the ASET and RSET. In this context, evacuation is considered successful only if all occupants evacuate within the ASET [14]. In this study, the ASET was derived using the FDS, whereas the RSET was based on evacuation simulations of occupants conducted using Simulex software.
To evaluate the effect of spatial design features on evacuation performance in underground buildings, three representative scenarios were developed based on a baseline architectural layout. Each scenario isolated a critical design parameter known to influence evacuation efficiency: exit placement, stair width, and the presence of vertical connections between underground levels.
The simulation framework integrates fire and evacuation modeling tools to assess the interaction between fire hazards and occupant movement. Fire Dynamics Simulator (FDS) was first used to simulate fire growth, smoke propagation, temperature distribution, and visibility over time. These outputs were used to determine the available safe egress time (ASET), which defines the period during which evacuation is tenable.
Following this, Simulex was employed to model human evacuation behavior based on the same architectural layout. The tool calculated the required safe egress time (RSET), accounting for occupant travel distances, density-driven slowdowns, and stairwell flow. A scenario was deemed successful if RSET remained within ASET thresholds for all occupants.
This two-step simulation allowed for evaluation of evacuation feasibility under architectural constraints, independent of fire suppression systems or behavioral interventions.

2.1. Building Outline

The target building used for the simulations is a representative complex building with underground connections in Seoul, South Korea. The building consists of four basement and four above-ground levels, forming an interconnected underground complex. The building houses various facilities, including shopping malls, multiplex cinemas, aquariums, exhibition halls, conference rooms, restaurants, and offices (Table 1).
Figure 1 displays exterior and interior views of the target building while Figure 2 illustrates floor plans of the simulated building from the first basement level (B1) to the fourth ground level (G4), with marked locations of staircases and escalators relevant to evacuation analysis. The first basement floor primarily consists of a large shopping mall, a multiplex cinema, and an aquarium. It includes exits connected to underground subway stations, as well as staircases and escalators providing vertical access to the first floor. The four ground floors above contain spaces such as conference rooms, exhibition halls, restaurants, offices, and an auditorium. The first floor has five exits leading to the outside, and both staircases and escalators provide vertical circulation throughout the building. The first basement floor houses a large shopping mall, which attracts approximately 100,000 visitors on weekdays and 150,000 on weekends. The underground shopping mall contains more than 200 stores and is connected to a nearby subway station and hotel.

2.2. ASET and RSET Calculation

In this study, the Fire Dynamics Simulator (FDS), developed by the National Institute of Standards and Technology (NIST), was used to evaluate occupant survivability under various fire scenarios. FDS enables precise modeling of fire behavior, including fire spread, temperature changes, and smoke movement, and is widely used in fire safety research and building design [15,16]. Specifically, we used FDS to analyze visibility, temperature, and CO concentration, providing key insights into occupant survivability. These thresholds, summarized in Table 1, are critical for ensuring safe evacuation. Data were collected at main entrances, exits, underground routes, and stairwells on each floor. To determine the maximum ASET, a visibility threshold of 5 m was used as the limiting factor, as it was reached earlier than the thresholds for temperature (60 °C) and CO concentration (500 ppm) [17]. As shown in Table 2 and Table 3, this threshold aligns with safety guidelines recommending 5–10 m visibility in familiar spaces and over 15–20 m in unfamiliar environments to ensure effective evacuation [17].
The input and output data for the FDS simulations are listed in Table 4. The input data for simulation initialization include the initial temperature, ignition source, internal airflow velocity, fire size, simulation duration, building size, and grid resolution. The ignition is assumed to occur on a polyurethane sofa, with the fire size set at 3.5 MW, based on fire experiment data from the NIST in the United States. The mesh within the FDS was configured to the optimal size using a rectangular prism (2 m × 2 m × 2 m), as recommended by the developer.
The output data obtained from the simulation included the visibility (m), temperature (°C), and CO distribution. Over a simulation runtime of 3600 s (60 min), the propagation of smoke (visibility) as well as the temperature and CO distributions were measured at designated points on each floor of the building. These measurement points, totaling 8–9 per floor, were specified at locations such as 1.5 m above the ground, roughly equivalent to the breathing level height. Visibility, temperature, and CO concentrations were measured at the ignition point, entrances/exits (including underground evacuation features, main exits, and evacuation stairs), and corridors.
For the RSET analysis, Simulex software was used to assess the egress performance as occupants evacuated the building during a fire scenario. The key input data for Simulex include the number of evacuees, walking speeds, and evacuation initiation times. The output data consists of the number of evacuees per exit at each time interval and the final evacuation completion time.
The evacuation capacity was calculated using the maximum occupancy per unit area, as specified in the National Fire Protection Association (NFPA) 101 Code for commercial and public assembly buildings in the United States. A detailed breakdown of floors is presented in Table 5. To determine the evacuation capacity of each floor, the usable floor area (m2) was multiplied by the occupancy load per unit area. The usable floor area refers to the effective space available for occupancy, such as auditoriums, offices, conference rooms, restaurants, and exhibition halls, excluding non-usable areas, such as mechanical rooms, storage rooms, and staircases. The occupancy load per unit area was derived from the NFPA 101 code standards [18], as 0.71 persons/m2 for exhibition and conference rooms, 0.36 persons/m2 for restaurants, and 0.1 persons/m2 for offices (Table 5). Based on these calculations, the total evacuation capacity was determined for each floor. Subsequently, occupants were randomly assigned to positions within the designated areas on each floor to simulate realistic evacuation scenarios.
Moreover, to account for the significant impact of occupant characteristics on evacuation outcomes, the “Japan: Hall/Hotel+” option was selected in Simulex because of the similarity between the physical characteristics and residential behavior of Koreans and Japanese. The detailed specifications are provided in Table 6. These dimensions determine the spatial representation of evacuees, ensuring realistic movement interactions within the simulation. The total body radius (0.25 m) accounts for both the torso and shoulder area, while the torso (0.15 m) and shoulder circles (0.10 m) define individual body structure. These settings are crucial in modeling crowd movement in environments such as shopping malls and halls, where spatial constraints significantly impact evacuation dynamics. In addition, the walking speed settings in the simulation are designed to replicate realistic movement behaviors across different terrains. These parameters ensure an accurate representation of evacuee movement, particularly the reduced speeds when navigating stairs, which aligns with real-world evacuation behavior.

2.3. Simulation Scenario

In this study, three evacuation simulation scenarios were analyzed to assess egress safety, as summarized in Table 7. The first scenario was based on the original design of the target building, whereas the second and third scenarios incorporated improvements to several evacuation features in the areas identified as the most severe bottlenecks in the first scenario.
Specifically, in accordance with Article 8 of the “Regulations on the Determination, Structure, and Installation Standards of Underground Public Pedestrian Facilities”, the distance between underground exit stairs connected to main roads was set to 100 m in the first scenario. Additionally, as per Article 11(1) of the “Regulations on Standards for Evacuation and Fire Protection Structures of Buildings”, the distance from the evacuation stairs on the ground floor (designated as the evacuation floor) to the building exit was set to 50 m for buildings with fire-resistant structures or non-combustible materials. The entrance width of the underground evacuation features was set to 8 m, and the widths of the stairs and handrails, along with the effective width of the stair entrances, were set to 1.6 m.
Specific parameters were adjusted in Scenarios 2 and 3 to optimize the evacuation performance during fire emergencies. These adjustments included reducing the distance between the underground exit stairs connected to the main road, shortening the distance from the evacuation stairs to the main exits, and increasing the width of both stairs and stair entrances, as well as the handrails. In summary, in Scenarios 2 and 3, the aim was to enhance the overall evacuation efficiency by minimizing travel distances and improving the capacity of stairways and exits.
Figure 3 and Figure 4 illustrate the differences among the three scenarios in terms of staircases, escalators, and outside exits for the 1st basement and above-ground (ground to 4F) floors, respectively. Notably, the number of staircases and exits increases with each successive scenario. In the diagrams, newly added exits compared to the previous scenario are highlighted with red circles, while existing exits from the previous scenario are marked with yellow squares.
In all simulation scenarios, it was assumed that the fire originated from the polyurethane sofa in the dining area of the basement-floor shopping mall. The evacuation routes were determined using the Simulex evacuation simulation program. The primary protocol involved occupants on the basement floors (B1–B4) evacuating upward and accessing the main roads outside via underground exit stairs. Potential delays due to congestion were considered, although delays resulting from accidents were not considered. Conversely, the occupants on the ground floors (1F–4F) evacuated downward. These occupants used evacuation stairs and escalators, avoiding elevators, to reach the primary exits leading outside the ground floor via the most expedient route available. Our simulation assumes full building occupancy with all occupants initiating evacuation simultaneously upon fire detection, which represents a worst-case scenario for egress demand. No modeling of fire alarms, sprinklers, or delayed responses is incorporated, in order to isolate the effects of spatial configurations on evacuation outcomes. All evacuees are treated as mobile adults with standardized travel speeds, and behavioral variability is not modeled. These assumptions aim to ensure that the influence of architectural design parameters can be clearly evaluated within the defined scope.

3. Results

3.1. Analysis of ASET Using the FDS

In the Fire Dynamics Simulator (FDS) simulation, key input parameters were defined to model fire behavior and its impact on the built environment. The ignition source was a polyurethane sofa, a highly flammable material commonly used in construction and furnishings. The fire scenario was characterized by a peak heat release rate (HRR) of 3.5 MW, representing a significant fire event. To analyze fire growth and smoke propagation, the simulation was conducted for 3600 s (60 min). The computational domain covered a 752 m × 332 m × 32 m building, discretized into 998,656 grid cells (376 × 166 × 16) to ensure sufficient resolution for accurate fire and smoke dynamics analysis.
The maximum ASET for each floor was calculated using the FDS as shown in Table 8. On the first basement floor, the time taken to reach the fire hazard threshold for visibility (5 m) was 8 min and 10 s, while the temperature (60 °C) and CO (500 ppm) levels could not be measured, as the maximum allowable safe egress times for both measures exceeded the 60 min simulation duration, making measurement impossible. Therefore, the shortest ASET based on the visibility was set as the maximum allowable safe egress time on the floor.
On the first ground floor, the time taken to reach the fire hazard threshold for visibility (5 m) was 9 min and 51 s, with temperature (60 °C) and CO (500 ppm) levels again unmeasurable. The maximum ASET for the second floor was 13 min and 45 s, respectively. On the third and fourth floors, the fire hazard threshold for visibility (5 m) was not reached within the 1 h simulation duration, making it impossible to measure the ASET for these floors.
Regarding the time-dependent smoke spread, 6 min after the onset of the fire, a significant amount of smoke began accumulating in parts of the first basement floor (shopping mall) and quickly spread to the upper floors, including the first ground floor, via both direct and emergency staircases (Figure 5). At 8 min and 10 s (the maximum allowable safe egress time), smoke filled the entire building, including the first basement floor. After 36 min, the smoke density increased further and completely engulfed the building (Figure 5).
In terms of temperature and CO distribution over time, the temperature around the origin of the fire increased steadily until 10 min, reaching a maximum of 30 °C throughout most areas by 25 min. However, the temperature did not exceed the fire hazard threshold of 60 °C. High concentrations of CO were observed around the origin of the fire within the first 5 min, spreading rapidly across the affected floor within 20 min before gradually decreasing.

3.2. Analysis of RSET Using the Simulex

For occupants on the first basement floor and below, the analysis of the RSET using the Simulex software yielded the following results across the three scenarios as summarized in Table 9. In the first scenario, in which evacuation ultimately failed, the maximum ASET of 8 min and 10 s was shorter than the final evacuation completion time of 36 min. In the second scenario, the evacuation was successful because the final evacuation time of 8 min and 7 s was within the safety limit of 8 min and 10 s, as calculated using the Fire Dynamics Simulator (FDS). Within the first 5 min of evacuation, 28,898 individuals (92.5% of the maximum occupancy of 31,236 on the basement floors) could evacuate safely by directly reaching the main road via underground exits to ground level. In the third scenario, the evacuation was successful, with a final evacuation time of 6 min and 13 s, well within the maximum allowable safe egress time of 8 min and 10 s. This scenario met the international evacuation standard time limit of 6 min, as defined by the U.S. NFPA 130 Code, ensuring successful evacuation. Additionally, within the first 5 min of evacuation initiation, 98.9% of the maximum occupancy (30,896 individuals) completed the evacuation.
Three scenarios were analyzed for occupants on the first ground floor and above and the results are summarized in Table 10. In the first scenario, the evacuation was unsuccessful because the final evacuation completion time of 41 min and 19 s exceeded the maximum allowable safe egress time of 9 min and 51 s. Similarly, in the second scenario, the evacuation failed because the final evacuation completion time of 22 min and 56 s surpassed the allowable time limit.
However, in the third scenario, the evacuation was successful. The final evacuation completion time of 9 min and 35 s was within the maximum allowable safe egress time of 9 min and 51 s. In this scenario, within the first 5 min of evacuation initiation, 33,869 individuals (90.5% of the maximum occupancy of 37,384 across the first through fourth floors) successfully evacuated through the exit on the first ground floor, which was designated as the evacuation floor.
In Summary, in Scenario 1, the default design led to prolonged and inefficient evacuation, as reflected by the extended completion times of over 36 min for the basement level and more than 41 min for the upper floors. This scenario experienced considerable congestion along evacuation paths due to excessive distance between exits (100 m) and limited vertical circulation capacity, contributing to a bottleneck effect. Scenario 2 introduced moderate improvements in exit spacing and vertical access, reducing exit distance to 50 m. As a result, the evacuation was completed within the ASET for the basement floor (8 min 7 s vs. 8 min 10 s), while the ground floor still exceeded the ASET slightly. Nevertheless, the evacuation rate was significantly improved, with more than 92% of occupants successfully evacuating within 5 min from the basement, indicating a substantial improvement in early-stage egress effectiveness. Scenario 3 demonstrated the most efficient evacuation performance due to more aggressive design optimizations. With exit spacing reduced to 40 m and improved stair and landing dimensions (2 m width), this scenario achieved rapid occupant flow, with nearly 99% of basement occupants and 90.5% of upper-floor occupants evacuating within 5 min. The total evacuation time remained well below the critical thresholds (6 min 13 s for B1; 9 min 35 s for 1F), satisfying both the calculated ASETs and international standards such as NFPA 130.
Figure 6 displays the cumulative number of evacuees across three scenarios for the first basement floor and the ground floor. For both floors, the first scenario shows the slowest evacuation rate at the beginning, with a gradual increase in the number of evacuees over time. The cumulative number increases steadily, indicating that evacuations are happening more slowly or in a less efficient manner compared to the other scenarios. In the second scenario, the evacuation starts off more quickly than in the first scenario and continues at a relatively consistent rate. The curve shows a smoother and more efficient evacuation process compared to the first scenario. The third scenario shows the most rapid early evacuation, with a steep increase in the number of evacuees at the beginning. This suggests that the architectural settings implemented in this scenario lead to quicker movement of people out of the building, especially in the initial stages.
In addition, Figure 7 illustrates the evacuation process for occupants on the first basement floor, captured 2 min after the fire incident, as simulated under the first, second, and third scenarios. In the first scenario, the internal spacing of the underground exit facility was set at approximately 100 m, which resulted in high congestion along the evacuation route. In contrast, the second and third scenarios featured adjustments to the internal spacing, set to 50 and 40 m, respectively, significantly reducing the density of occupants and alleviating congestion along the evacuation path.
As shown in Figure 8, the ground floor levels experienced significant reductions in congestion along the evacuation routes in the second and third scenarios, similar to the improvements observed for the basement floors. This reduction in congestion facilitated smoother evacuation and enhanced overall safety during the fire scenario.

4. Discussions

This study examined the impact of key evacuation features on occupant safety in complex buildings with underground connections. Given that Seoul has the highest concentration of such buildings in the country, understanding and enhancing evacuation strategies are crucial for minimizing casualties and property damage during emergencies. Through detailed fire and evacuation simulations, this study provides valuable insights into the effectiveness of existing evacuation protocols and identifies key modifications to improve egress safety.
The evacuation performance varied based on occupant location and evacuation routes. For occupants below the 1st basement floor (upward evacuation), when the interior spacing of underground exit facilities was 50 m, the total evacuation time was 8 min and 7 s, which was within the maximum survivability threshold of 8 min and 10 s, ensuring a successful evacuation. Reducing the spacing to 40 m further decreased the evacuation time to 6 min and 13 s, meeting the international standard threshold of 6 min set by the NFPA 130 Code (U.S.). However, when the legal standard spacing of 100 m was applied, severe bottlenecks at the 1st basement floor evacuation stair entrance resulted in evacuation failure, with a total evacuation time of 36 min. For occupants above the ground floor (downward evacuation), when the walking distance from the staircase on the ground floor (evacuation floor) to the exterior exit was 30 m, the evacuation was successful, with a total evacuation time of 9 min and 35 s, staying within the maximum survivability threshold of 9 min and 51 s. However, when the walking distance exceeded 50 m, the total evacuation time ranged from 23 min and 55 s to 41 min and 19 s, surpassing the survivability threshold and resulting in evacuation failure.
The findings emphasize the importance of optimizing evacuation routes by addressing bottleneck scenarios and enhancing installation standards for evacuation features. The analysis revealed significant differences in evacuation performance across the three simulated scenarios, demonstrating that successful evacuations were achieved through improved configurations, including reduced distances to exits and increased stairway and landing widths. These findings contribute to the existing body of knowledge on disaster management in high-density urban environments and are expected to inform policy decisions regarding building regulations and fire safety protocols.
To enhance the effectiveness of evacuation routes, a dual-path strategy was developed to account for the varying locations of occupants. This strategy involved utilizing both underground exit facilities and ground-floor (evacuation floor) exits. During a fire incident, occupants in the basement are likely to encounter severe bottlenecks at the entrance to the evacuation stairs on the first basement floor. To mitigate this issue, providing direct evacuation routes to the streets via underground exit facilities is critical. Additionally, occupants on and above the ground floor can safely evacuate through exterior exits at ground level.
To implement this dual-path evacuation strategy effectively, specific improvements in the installation standards of key evacuation features are necessary. The interior spacing of underground exit facilities should be maintained within 50 m to ensure sufficient space for safe navigation during emergencies. Furthermore, the walking distance from the stairs on the ground floor (evacuation floor) to the exterior exits should not exceed 30 m. These adjustments significantly reduce the time required for occupants to reach safety and enhance overall evacuation efficiency.
Installing at least one underground plaza adjacent to an underground walkway is essential for emergency evacuation. According to Article 6 of the “Regulations on the Decision, Structure, and Installation Standards for Underground Public Walkways”, these underground plazas must be designed to facilitate quick and efficient escape routes. Additionally, sunken areas—spaces located below ground level and open to outside air—should be constructed up to the third basement floor. These areas serve as walking and resting spaces and provide direct stair access to ground level. Such enhancements align with the guidelines set forth by the General Building Review Standards of the Seoul Architectural Commission and ensure that occupants can evacuate safely to ground level during emergencies.
However, while the study focused on three representative scenarios involving variations in stair width, exit location, and vertical connections, we acknowledge that additional parameters such as phased evacuation, variable occupant loads, and delayed response times can further impact egress performance. Future research should explore these dimensions to generate more comprehensive design and policy guidance for complex urban buildings. In addition, although this study focuses primarily on architectural configurations affecting evacuation, it is essential to recognize that fire compartmentalization, emergency signage, and suppression systems are integral to a comprehensive fire safety strategy. In the present simulations, compartmental boundaries were not explicitly modeled, which may affect smoke propagation and tenability assumptions. Similarly, wayfinding signage and dynamic evacuation cues were not included in behavioral modeling. Despite these limitations, the design variables examined remain critical in determining baseline egress capacity. Future studies should integrate these elements in alignment with regulatory frameworks such as NFPA 101 and national fire codes to enhance model realism and applicability.

5. Conclusions

This study analyzed evacuation safety in underground connected buildings using fire and evacuation simulations. Successful evacuation was achieved when underground exit spacing was ≤50 m for upward evacuation and walking distance to exterior exits was ≤30 m for downward evacuation. Exceeding these limits resulted in severe bottlenecks and evacuation failure, with times surpassing survivability thresholds. Results showed that optimizing key features—wider stairways, shorter exit distances, and a dual-path strategy—significantly improved egress efficiency. These findings provide critical insights for disaster management and urban fire safety policies.
Addressing bottlenecks and enhancing evacuation standards proved essential for optimizing evacuation routes. Simulations revealed that reducing exit distances and increasing stair and landing widths improved evacuation efficiency. These findings contribute to disaster management in high-density urban areas and inform future building regulations and fire safety protocols.
Beyond fire safety, this study supports sustainable urban resilience by integrating safer egress strategies into urban planning. The results inform evidence-based policies and regulations while minimizing casualties, and economic disruption, and fostering resilient communities.
Limitations include reliance on specific simulation parameters and a focus on one building type, which may limit generalizability. Future research should examine diverse buildings, occupancy levels, and behavioral factors, with longitudinal studies assessing long-term evacuation effectiveness. Continuous improvements in evacuation planning are essential for safer urban environments. In addition, this study focused specifically on analyzing key evacuation design features and, therefore, minimized the influence of other fire safety components, such as fire protection systems, compartmentation, and fire alarms. While this approach helped isolate the effects of evacuation design, it represents a limitation of the study. Future research could incorporate these additional factors to provide a more holistic assessment of evacuation performance under fire conditions.

Author Contributions

Conceptualization, Y.-S.B. and M.C.; methodology, Y.-S.B. and M.C.; software, M.C.; formal analysis, Y.-S.B. and M.C.; writing—original draft preparation, M.C.; writing—review and editing, Y.-S.B. and M.C.; visualization, M.C.; supervision, Y.-S.B.; funding acquisition, M.C. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grants funded by the Korean government (MSIT) (RS-2024-00337975). This work was supported by the Seoul Institute (2024-PR-12).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. National Fire Agency. Available online: https://www.nfa.go.kr/nfa/news/pressrelease/press/;jsessionid=ZkojwvnyqkINJxrm+2r7ENIS.nfa12?boardId=bbs_0000000000000010&mode=view&cntId=2212&category=&pageIdx=&searchCondition=&searchKeyword (accessed on 7 November 2024).
  2. Special Act on Management of Disasters in Super High-Rise Buildings and Complex Buildings with Underground Connections, Korea, Article 2. 2024. Available online: https://elaw.klri.re.kr/eng_mobile/viewer.do?hseq=65275&type=part&key=12 (accessed on 16 February 2025).
  3. Li, X.; Chen, W.; Wang, C.; Kassem, M.A. Study on evacuation behavior of urban underground complex in fire emergency based on system dynamics. Sustainability 2022, 14, 1343. [Google Scholar] [CrossRef]
  4. Wang, D.; Yang, Y.; Zhou, T.; Yang, F. An investigation of fire evacuation performance in irregular underground commercial building affected by multiple parameters. J. Build. Eng. 2021, 1, 102146. [Google Scholar] [CrossRef]
  5. Shan, S.; Guo, X.; Wei, Z.; Sun, W.; Zheng, H.; Pan, H.; Lin, J. Simulation analysis of evacuation processes in a subway station based on multi-disaster coupling scenarios. Int. J. Disaster Risk Reduct. 2023, 1, 103998. [Google Scholar] [CrossRef]
  6. Akizuki, Y. Evacuation route design based on visibility for reducing evacuation delays. Fire Saf. J. 2024, 1, 104099. [Google Scholar] [CrossRef]
  7. Bae, S.; Cha, H. A Method on Developing 3D/BIM-Based Real Time Fire Disaster Information Management. Korean J. Constr. Eng. Manag. 2023, 24, 3–12. [Google Scholar]
  8. Park, J.G. A study on the risks of underground spaces. J. Disaster Res. 2004, 6, 16–26. [Google Scholar]
  9. Hong, W.H.; Kim, T.H.; Jeon, K.Y. A study on the initial evacuation and rescue activities in underground spaces during emergencies: Focusing on the Daegu subway fire incident. Korean J. Archit. Inst. Korea 2006, 22, 263–270. [Google Scholar]
  10. Fire Prevention News. Available online: https://www.fpn119.co.kr/156233 (accessed on 8 November 2024).
  11. Santos, G.; Aguirre, B.E. A Critical Review of Emergency Evacuation Simulation Models; Disaster Research Center: Newark, DE, USA, 2004. [Google Scholar]
  12. Kallianiotis, A.; Papakonstantinou, D.; Arvelaki, V.; Benardos, A. Evaluation of evacuation methods in underground metro stations. Int. J. Disaster Risk Reduct. 2018, 31, 526–534. [Google Scholar] [CrossRef]
  13. Cui, C.; Shao, Q.; Liu, Y.; Han, G.; Liu, F.; Han, X. A review of the evolution and trends in research on the emergency evacuation of urban underground spaces. Buildings 2023, 13, 1325. [Google Scholar] [CrossRef]
  14. Purser, D. Developments in Tenability and Escape Time Assessment for Evacuation Modelling Simulations. In Evacuation Modeling Trends; Springer: New York, NY, USA, 2016; pp. 25–53. [Google Scholar] [CrossRef]
  15. McGrattan, K.B.; McDermott, R.; Vanella, M.; Mueller, E.; Hostikka, S.; Floyd, J. Fire Dynamics Simulator User’s Guide; NIST special publication: Gaithersburg, MD, USA, 2024; p. 1019. [Google Scholar]
  16. McGrattan, K.; Hostikka, S.; Floyd, J.; Baum, H.; Rehm, R.; Mell, W.; McDermott, R. Fire Dynamics Simulator (Version 5) Technical Reference Guide; NIST Special Publication: Gaithersburg, MD, USA, 2010; p. 1018. [Google Scholar]
  17. Hurley, M.J.; Gottuk, D.T.; Hall, J.R., Jr.; Harada, K.; Kuligowski, E.D.; Puchovsky, M.; Watts, J.M., Jr.; Wieczorek, C.J. (Eds.) SFPE Handbook of Fire Protection Engineering; Springer: New York, NY, USA, 2015. [Google Scholar]
  18. NFPA 130; Standard for Fixed Guideway Transit and Passenger Rail Systems. National Fire Protection Association: Quincy, MA, USA, 2023.
Figure 1. Exterior and Interior Views of the Target Building.
Figure 1. Exterior and Interior Views of the Target Building.
Sustainability 17 03731 g001
Figure 2. Floor Plans (First Basement to Fourth Ground floor) of the Simulated Building.
Figure 2. Floor Plans (First Basement to Fourth Ground floor) of the Simulated Building.
Sustainability 17 03731 g002aSustainability 17 03731 g002b
Figure 3. Changes in Evacuation Features Across Three Scenarios for the B1 Floor. M: Multiplex, Aq: Aquarium.
Figure 3. Changes in Evacuation Features Across Three Scenarios for the B1 Floor. M: Multiplex, Aq: Aquarium.
Sustainability 17 03731 g003aSustainability 17 03731 g003b
Figure 4. Changes in Evacuation Features Across Three Scenarios for the Above-Ground Floors. E: Exhibition Hall, C: Conference Room, R: Restaurant.
Figure 4. Changes in Evacuation Features Across Three Scenarios for the Above-Ground Floors. E: Exhibition Hall, C: Conference Room, R: Restaurant.
Sustainability 17 03731 g004aSustainability 17 03731 g004b
Figure 5. Smoke Spread Analysis during Fire Simulation with FDS.
Figure 5. Smoke Spread Analysis during Fire Simulation with FDS.
Sustainability 17 03731 g005aSustainability 17 03731 g005b
Figure 6. Cumulative Number of Evacuees across Three Scenarios.
Figure 6. Cumulative Number of Evacuees across Three Scenarios.
Sustainability 17 03731 g006
Figure 7. Screenshots of the Evacuation Simulation Taken 2 min After Fire Initiation on the B1 Floor.
Figure 7. Screenshots of the Evacuation Simulation Taken 2 min After Fire Initiation on the B1 Floor.
Sustainability 17 03731 g007
Figure 8. Screenshots of the Evacuation Simulation Taken 2 min After Fire Initiation on the 1F Floor.
Figure 8. Screenshots of the Evacuation Simulation Taken 2 min After Fire Initiation on the 1F Floor.
Sustainability 17 03731 g008
Table 1. Key Facilities and Their Effective Area on Each Floor of the Simulated Building.
Table 1. Key Facilities and Their Effective Area on Each Floor of the Simulated Building.
FloorTotal Floor AreaKey Facilities and Their Effective Area
B2 to 4-Parking lot
B1101,199 m2Shopping mall (69,300 m2), Multiplex (4218 seats), Aquarium (5750 m2)
1F (Ground floor)72,103 m2Exhibition hall A (10,368 m2), Exhibition hall B (8810 m2), Meeting rooms (1817 m2), Restaurants (1713 m2)
2F72,103 m2Auditorium (186 seats), Meeting rooms (1262 m2), Offices (6722 m2), Restaurants (1713 m2)
3F72,103 m2Auditorium (1058 seats), Exhibition hall A (10,348 m2), Exhibition hall B (7281 m2), Meeting rooms (6579 m2),
4F72,103 m2Auditorium (450 seats), Meeting rooms (468 m2), Offices (4621 m2)
Table 2. Fire Hazard Thresholds for Visibility, Temperature, and Carbon Monoxide during a Fire [16].
Table 2. Fire Hazard Thresholds for Visibility, Temperature, and Carbon Monoxide during a Fire [16].
FactorAllowable Threshold
Visibility≥5 m
Temperature≤60 °C
Carbon monoxide (CO)≤500 ppm
Table 3. Visibility Thresholds and Smoke Density Levels Impacting Evacuation [16].
Table 3. Visibility Thresholds and Smoke Density Levels Impacting Evacuation [16].
VisibilitySmoke Density Levels
20–30 m· Very faint presence of smoke
-
    Smoke detectors activate at this concentration level
-
    Individuals unfamiliar with the building may begin to experience difficulties evacuating when the concentration exceeds this level
5 m· Familiar occupants may start to experience difficulty during evacuation
3 m· A sense of dimness is felt
-
    Evacuation may require occupants to feel their way forward gradually
1–2 m· Visibility is almost completely obstructed
≤1 m· Visibility is nearly zero, creating a state of darkness. Induction lights are also not visible
Table 4. Input and Output Data for FDS Simulations.
Table 4. Input and Output Data for FDS Simulations.
DataVariableValue
InputInitial temperature25 °C (room temperature)
Ignition sourcePolyurethane
Internal airflow velocity0 m/s
Fire size (HRR)3.5 Mw
Simulation duration3600 s (60 min)
Building size and number of gridsBuilding size: 752 m (X), 332 m (Y), 32 m (Z)
Number of grids: 998,656 (376 × 166 × 16)
OutputVisibility (m)Smoke density at measurement points during simulation time
Temperature (°C)Temperature changes at measurement points during simulation time
Carbon monoxide distributions (ppm)Changes in level of carbon monoxide at measurement points during simulation time
Table 5. Evacuation Capacity per Floor.
Table 5. Evacuation Capacity per Floor.
AreaSize (m2)Unit CapacityEvacuation Capacity (Persons)
B1
Shopping mall
(200 stores)
69,3000.36 person/m224,948
Multiplex21,2711 seat/person4218
Aquarium57500.36 person/m22070
1F (Ground Floor)
Exhibition hall19,1780.71 person/m213,616
Meeting room18170.71 person/m21290
Restaurant17130.36 person/m2617
2F
Performance hall62,4061 seat/person186
Office67220.1672
Meeting room12620.71896
Restaurant17130.36617
3F
Performance hall47,8951 seat/person1058
Exhibition hall17,6290.71 person/m212,516
Meeting room65790.71 person/m24672
4F
Performance hall1 location1 seat/person450
Office46210.1462
Meeting room4860.71 person/m2332
Total68,620
Table 6. Evacuee Specifications in Simulex Simulation.
Table 6. Evacuee Specifications in Simulex Simulation.
Body Type IndexBody Type NameTotal Radius (m) of Body CircleRadius (m) of Main Torso CircleRadius (m) of Shoulder Circles
27Japan: Hall/Hotel+0.250.150.10
Unimpeded walking velocity on flat terrainDistributed variation of unimpeded walking velocity on flat terrainMultiplication factor for walking speed down stairsMultiplication factor for walking speed up stairsBody color (0–19)
1.00.00.60.4513
Table 7. Settings for Key Evacuation Features for 1st, 2nd, and 3rd Simulation Scenario.
Table 7. Settings for Key Evacuation Features for 1st, 2nd, and 3rd Simulation Scenario.
Evacuation FeaturesSimulation Scenario
1st2nd3rd
Exit stairs connected to the main roadDistance between exit stairs~100 m50 m40 m
Width of exit stairs8 m10 m10 m
Distance from the evacuation stairs 1 on ground floor to the outside exit≥50 m50 m30 m
Distance from main service areas to the direct stairs 2 in each floor≥50 m50 m30 m
Width of stairs and landings1.6 m2 m2 m
Effective width of stair entrances1.6 m2 m2 m
1: a staircase that is constructed with fire-resistant materials, equipped with backup lighting, maintain specified distances from other openings, has restricted window sizes, provides doors that open in the evacuation direction and automatically close during fires, and connect directly to a shelter floor or ground level. 2: a staircase that connects directly from every floor of a building (except for the shelter floor) to either the ground floor or outside exit.
Table 8. Analysis of Time Taken to Reach Fire Hazard Thresholds with FDS.
Table 8. Analysis of Time Taken to Reach Fire Hazard Thresholds with FDS.
FloorVisibility (5 m)Temperature (60 °C)Carbon Monoxide (CO, 500 ppm)
B18 min 10 s (490 s)Unable to measure *Unable to measure
1F9 min 51 s (591 s)Unable to measureUnable to measure
2F13 min 45 s (825 s)Unable to measureUnable to measure
3FUnable to measureUnable to measureUnable to measure
4FUnable to measureUnable to measureUnable to measure
* The value could not be measured because it exceeded the simulation time limit of 60 min.
Table 9. Results of Simulex Simulation Analysis for the B1 Floor.
Table 9. Results of Simulex Simulation Analysis for the B1 Floor.
Time (min)1st Scenario2nd Scenario3rd Scenario
Evacuated PersonsCumulative TotalEvacuated PersonsCumulative TotalEvacuated PersonsCumulative Total
1132613264892489275097509
225163842884213,73412,93020,439
325336375740221,136695527,394
423608735518726,323260329,997
5227811,013257528,89889930,896
6202513,038141130,30931231,208
7194014,97872431,0332831,236
8193216,91019231,225--
9182518,7351131,236--
10165820,393----
12279020,393----
14175323,183----
16114924,936----
18118326,085----
2084827,268----
2265428,116----
2455328,770----
2645129,323----
2845330,227----
3038530,612----
3223730,849----
3421931,068----
3616831,236----
Total31,23631,23631,23631,23631,23631,236
Mean867 persons/min3904 persons/min5206 persons/min
Table 10. Results of Simulex Simulation Analysis for the 1F Floor.
Table 10. Results of Simulex Simulation Analysis for the 1F Floor.
Time (min)1st Scenario2nd Scenario3rd Scenario
Evacuated PersonsCumulative TotalEvacuated PersonsCumulative TotalEvacuated PersonsCumulative Total
17417412995299516,33816,338
218922633728510,280588922,227
320344667477315,053488927,116
421396806348618,539400631,122
520638869278721,326274733,869
6211810,987240423,730185935,728
7192712,914208425,81494236,670
8181214,726194027,75440537,075
9170416,430185929,61325737,332
10169218,122173031,3435237,384
12321421,336262033,963--
14288024,216152135,484--
16237226,588102036,504--
18183228,42058337,087--
20176530,18510937,196--
22153531,72010037,296--
2495932,6798837,384--
2682433,503----
2877234,275----
3071634,991----
3268035,671----
3449136,162----
3639836,560----
3839136,951----
4036837,319----
426537,384----
Total37,38437,38437,38437,38437,38437,384
Mean911 persons/min1625 persons/min4153 persons/min
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Bae, Y.-S.; Choi, M. Optimizing Key Evacuation Features for Safer Egress in Complex Buildings with Underground Connections: A Simulation-Based Approach to Resilient and Sustainable Design. Sustainability 2025, 17, 3731. https://doi.org/10.3390/su17083731

AMA Style

Bae Y-S, Choi M. Optimizing Key Evacuation Features for Safer Egress in Complex Buildings with Underground Connections: A Simulation-Based Approach to Resilient and Sustainable Design. Sustainability. 2025; 17(8):3731. https://doi.org/10.3390/su17083731

Chicago/Turabian Style

Bae, Yoon-Shin, and Minji Choi. 2025. "Optimizing Key Evacuation Features for Safer Egress in Complex Buildings with Underground Connections: A Simulation-Based Approach to Resilient and Sustainable Design" Sustainability 17, no. 8: 3731. https://doi.org/10.3390/su17083731

APA Style

Bae, Y.-S., & Choi, M. (2025). Optimizing Key Evacuation Features for Safer Egress in Complex Buildings with Underground Connections: A Simulation-Based Approach to Resilient and Sustainable Design. Sustainability, 17(8), 3731. https://doi.org/10.3390/su17083731

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop