Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Der Artikel stellt eine Studie zu den Brandrisiken von Batterie-Elektrofahrzeugen (BEVs) in Tiefgaragen vor, wobei der Schwerpunkt auf den einzigartigen Herausforderungen liegt, die von der geschlossenen Umgebung ausgehen. Er vergleicht das Verhalten von BEV-Bränden in offenen und geschlossenen Räumen und verdeutlicht die Intensität und Schnelligkeit von Bränden in Tiefgaragen. Die Studie umfasst umfassende Brandtests, einschließlich des Einsatzes eines Prüfstandes, der einen Eckabschnitt einer Tiefgarage simulieren soll. Die Ergebnisse zeigen, dass BEV-Brände auf engstem Raum aufgrund der Ansammlung heißer Gase und des Verpuffungspotenzials zu größeren Schäden führen können. Die Erkenntnisse sind entscheidend für die Neubewertung der Brandschutzstandards und die Information der Ersthelfer bei der Bewältigung von Brandeinsätzen.
KI-Generiert
Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.
Abstract
Battery-only electric vehicle (BEV) fires have recently emerged as a significant safety concern in modern society, particularly when they occur in car parking spaces, where the consequences can be severe. While relatively few fatalities and injuries are reported, the associated economic losses are substantial. Car park structures present several challenges from a fire safety perspective, including high energy content, confined geometry, and difficulties in detecting and accessing to the fire’s origin due to smoke accumulation in spaces with limited ventilation and exits. This study investigates this fire hazard by conducting a full-scale fire test on a modern BEV in an instrumented test rig that simulates a segment of an underground car park. The data obtained were compared with standard fire curves to assess the hazard’s characteristics and with the data from a companion study to identify differences between BEV fires in underground and surface car parks. The primary difference observed was deflagration venting, which occurred shortly after the initial ignition of combustible gases accumulated beneath the ceiling, lasting until 13 min and 5 s. This phenomenon led to a rapid initial growth of the BEV fire, which burned more intensely for a shorter duration in the semi-enclosed configuration comparted to the open configuration. The enclosed fire recorded an average ceiling-jet temperature of approximately 1100°C and a peak incident heat flux exceeding 225 kW/m2. Additionally, thermal runaway within the lithium-ion battery cells was analysed to understand the adverse effects of the enclosed environment on thermal runaway propagation. This fire testing provides critical insights for developing firefighting strategies including the proper preparation of equipment and the design of fire protection systems. The principal findings could inform tactics for mitigating unforeseen fire risks and contribute to revisions of fire safety regulations.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
In 2006, a small startup in Silicon Valley announced its intention to manufacture a fully electric powered vehicle as an alternative to conventional internal combustion engine vehicles (ICEVs), aiming to promote clean transportation. Since then, Battery-only Electric Vehicle (BEV) technology, alongside lithium-ion battery (LIB) advancements, has rapidly progressed, driving growth in both the market and supporting industries over the past two decades. By 2023, global sales of electric vehicles (EVs), including plug-in hybrid electric vehicles (PHEVs), surpassed 14 million units, representing approximately 18% of the global automobile market—an impressive six-fold increase from 2018 [1]. Of these, over 40 million EVs were in operation worldwide, with 70% being BEVs. Despite the rise in BEV adoption, there have been sporadic fire incidents linked to LIB packs in these vehicles. While the frequency of BEV fires is lower than that of ICEV-related fires, the public has grown increasingly concerned about the risks associated with thermal runaway in LIBs and the potential for fire propagation.
Despite their growing popularity, BEVs still represent a minority in the global automotive market, and the field of BEV and LIB technologies remains complex and ever-evolving. Current knowledge about fires involving BEVs and LIBs is limited, prompting early studies to focus primarily on fundamental research. However, practical concerns about the risk of BEV fires, particularly in urban environments like underground car parks and multi-storey parking structures, have increased recently. These locations pose unique challenges for fire safety due to the following reasons: (1) High-energy combustibles—Car parks contain significant numbers of automobiles, each a potential source of fuel; (2) Enclosed construction—Structural elements such as columns, slabs, beams, and walls create partially enclosed spaces, facilitating the accumulation of hot combustion gases; (3) Limited ventilation—Small ventilation openings allow smoke to accumulate beneath the ceiling, obscuring the origin of the fire and complicating firefighting efforts; (4) Restricted access—Limited entrances and exits impede firefighters’ ability to reach the fire source quickly. The behaviour of flames in semi-enclosed spaces, such as car parks, is markedly different from that in open environments. Flames in these confined spaces tend to exhibit more intense heat and higher severity levels. Additionally, the introduction of BEV fires addresses a new layer of uncertainty to these risks. This is further complicated by the installation of BEV charging stations in these areas, promoted by government initiatives to transition to cleaner transportation modes. This development persists despite sporadic fire incidents associated with BEV charging, adding complexity to the management of fire safety in these structures.
Anzeige
Although the frequency of automobiles fire accidents in parking areas is relatively low compared to other fire categories across various countries [2, 3], these incidents, when they occur, often result in significant economic losses, potentially impacting the growing BEV market. For instance, as shown in Fig. 1, notable fire incidents in parking structures include: the Liverpool’s Echo Arena in the UK in 2017, which resulted in damage to approximately 1400 vehicles [4]; Stavanger Airport in Norway in 2020, affecting around 300 vehicles [5]; and a high-rise residential building complex in Korea in 2021, impacting 600 vehicles [6]. A particularly severe incident occurred recently at a 1581-household residential complex in Korea [7], where a fire initiated in a BEV spread rapidly, consuming nearly 90 vehicles and causing extensive damage to around 800 more. This event also damaged infrastructure such as ceiling fixtures for plumbing pipes and electric cables, leading to a 9-day outage of power and water that affected approximately 480 households. The scale of this disaster triggered nationwide concern about BEV safety, prompting a surge in ad hoc parking restrictions for BEVs and adversely affecting the national battery industry.
Previous research has focused on understanding the failure conditions of LIBs, including the associated electrochemical and thermochemical reactions through bench- or mid-scale experiments [8]. Following these foundational studies, researchers have explored the risks associated with BEV fires and their differences from ICEV fires by conducting full-scale car fire experiments in open settings [9‐14]. These tests have primarily assessed the intensity and duration of BEV fires by measuring the heat release rate, fire growth rate, and total heat released over time. The next phase of research is shifting towards more realistic conditions, examining full-scale BEV fires under scenarios that closely mimic actual environments [15]. In the early 2000s, similar effects focused on ICEV fires, where experiments were conducted in a confined setup using a 30-m-long, 15-m-wide, and 2.5-m-tall test rig designed to emulate an enclosed car park [16]. Additionally, ceilings were installed above the test specimens to simulate the thermal feedback from hot smoke layers [17]. Despite these advancements, there remains a significant gap in our understanding of how modern BEV fires behave in confined spaces that replicate full-scale underground car parks.
To protect life, property, and the fire service, fire protection standards set essential minimum requirements for parking structures. These include provisions for both active and passive fire protection systems, ventilation systems, and the fire-resistant qualities of structural elements [2, 18‐21]. The criteria for these standards consider several factors, such as the potential fire’s intensity and duration, the combustibility of materials used, and the area of openings for natural ventilation. Typically, underground car parks necessitate more rigorous fire protection measures, including mechanical ventilation and exhaust systems, fire alarms, and automated sprinkler systems. The shift from ICEV to BEV as a fire source prompts a re-evaluation of the sufficiency of current fire detection, extinguishing, and ventilation standards. BEV fires, for instance, differ markedly from ICEV fires in several respects: they may involve toxic and combustible gases released from LIBs prior to ignition, spontaneous ignition characterised by jet flames, and potential for LIB reignition after initial fire suppression efforts [12, 14]. Moreover, the enclosed design of underground car parks could modify the intensity and duration of BEV fires relative to those occurring in open environments. Such differences raise another questions about the adequacy of current standards for the fire-resistance of structural members. Given these consideration, it is imperative to reassess fire protection codes and firefighting strategies for modern BEV fires in underground car parks. This reassessment should be informed by data from comprehensive full-scale fire testing to ensure that the evolving fire risks associated with BEVs are adequately managed.
In the companion paper [14], the characteristics of modern BEV fires were experimentally examined in an open configuration and compared to those of ICEV and older EV fires. Building upon this, the current study investigates the effects of a confined environment, typical of parking structures, on the combustion characteristics of BEV fires. This includes an analysis of burning behaviour, intensity, and duration. To this end, a test rig measuring 7.8 m in length and width, and 2.3 m in height, was constructed to replicate a corner section of an underground car park, encompassing three adjacent parking bays. Thermocouples and heat flux gauges were strategically placed within the rig to observe temperature distributions across the confined area and to estimate the heat fluxes that might affect nearby cars or structural elements. A five-door sports utility BEV with a 72.6 kWh electric capacity was positioned at a designated location within the instrumented rig. To closely monitor the fire’s spread, thermocouples were installed in the motor, passenger, and trunk compartments of the vehicle. To simulate an internal failure of a LIB cell within the battery pack, a 575-Watt electric heating sheet was used to initiate heating in a single cell. Additional thermocouples and voltmeters were placed within the LIB modules to track thermal runaway propagation. CCTV cameras encircling the rig provided continuous observation of the BEV’s burning behaviour. The entire setup was placed under the hood of a 10-MW-scale oxygen consumption calorimetry at Korea Conformity Laboratories (KCL) to measure the BEV fire’s growth rate and duration by quantifying its Heat Release Rate (HRR) and Total Heat Released (THR).
Anzeige
2 Experimental
2.1 Test Rig Design
Figure 2a and b illustrate the test rig constructed to emulate a corner area of an underground car park. As mandated by the Korean Enforcement Rules of the Parking Lot Act [22], the minimum dimensions for parking bays are defined as 2300 mm in width and 5000 mm in length for a standard type, and 2500 mm by 5100 mm for an extended type. In typical underground car parks, it is common to find three parking bays positioned between the structural columns of buildings. For this experiment, as shown in Fig. 2c, three parking bays each measuring 2500 mm wide and 5000 mm long were constructed. This design accommodates the conventional width of automobiles (1900 mm) and provides a 600 mm distance between parked vehicles to allow for door openings. According to the Korean Rules on Housing Construction Standards [23], the minimum ceilings heights for buildings constructed before and after 2019 are 2300 mm and 2700 mm, respectively. The test rig was designed with an internal height of 2300 mm to ensure representativeness of typical underground parking conditions. The rig was partially enclosed, with fireproof gypsum boards installed on only two of the four sides, reflecting the assumption that the parking structure would extend in the direction of the two open sides.
In enclosed environments, fire plumes are known to accumulate beneath the ceiling, a condition not present in open configurations where airflow remains unrestricted. This accumulation of hot gases can lead to simultaneous radiant and convective heating of combustibles within the space, creating thermal conditions that are generally more severe than those encountered in outdoor fires. Such conditions are typical in partially enclosed spaces, such as underground and open-sided multistorey car parks. To replicate and optimise these conditions in the full-scale fire testing, 1000-mm-high skirts were suspended from the ceiling of the rig, as depicted in Fig. 2a. This design ensures the formation of a hot gas layer in the upper section of the space. Additionally, an opening of approximately 1200-mm2 was incorporated in the upper corner of the rig. This opening serves a purpose: it allows the escape of combustion gases into the large-scale calorimeter hood and therefore facilitates accurate measurements of HRR and THR using the oxygen consumption method.
To accurately assess the thermal conditions of underground car parks during a BEV fire, the rig was equipped with K-type thermocouples and Schmidt-Boelter-type water-cooled heat flux metres. As detailed in Fig. 2b, six thermocouples (TC_GB, GW, GR, GCRW, GFW, and GRW) were strategically positioned to measure the temperatures of the upper gas layer directly above the test vehicle, capturing the intense thermal activity typical of such fires. Additionally, three heat flux metres (HF_FD, RD, and R) were installed to estimate the incident heat flux impacting nearby structural elements such as columns and walls, BEV charging stations, and vehicles parked adjacent to the test BEV. Moreover, as illustrated in Fig. 3a, five thermocouple trees were constructed, each equipped with four vertically arranged thermocouples (TR_x_L, M, U1, and U2 where x = 1 through 5). These devices are designed to measure the vertical temperature distributions within the partially enclosed space. Positioned at the base of these trees, heat flux metres (HF_1, 2, 3, 4, and 5) with upward-facing measuring surfaces further help monitor the thermal dynamics. The spatial arrangement within the rig was methodically planned to facilitate comprehensive data collection. As shown in Fig. 3b, the thermocouple trees were uniformly distributed across the space, dividing it into six designated areas (Areas 0–5). The test BEV was situated in Area 0, located at the intersection of two walls, maintaining a clearance of approximately 300 mm from each wall to ensure unimpeded measurement by nearby sensors. This placement strategy, combined with the sensor array depicted in Fig. 4a, allows for an extensive evaluation of the horizontal temperature distribution and heat fluxes throughout the rig, offering critical insights into the fire behaviour within this controlled setting.
Figure 3
Schematics of the measurement equipment and layout
The test vehicle employed in this study is a five-door electric sports utility vehicle (SUV), produced on a dedicated electric vehicle platform in 2021. It measures 4635 mm in length, 1890 mm in width, 1647 mm in height, and has a curb weight of approximately 2030 kg. Analysis by an automotive benchmarking company (A2Mac1) indicates that non-metallic materials constitute about 33%–36% of the total mass of similar electric SUVs. In the full-scale fire tests [14], it was observed that one BEV lost between 17.6 and 18.4% of its mass as a result of complete combustion. Specifically, the mass of the LIB pack decreased by 6.2%, and the mass of the BEV body (excluding the pack) reduced by 21.8%. From these figures, the total fire load of the vehicle used in this experiment is estimated to be between 357 kg and 374 kg, considering the BEV’s total curb weight of 2030 kg.
During the testing, the vehicle’s front door windows were kept open to lead to internal fire dynamics. Thermocouples were strategically placed inside the vehicle to monitor the progression of the fire, as depicted in Fig. 4b. Additionally, the temperatures of the tyres (TC_WF and WR) were measured. This is crucial because the composite rubber materials in tyres can rapidly propagate flames, particularly in cases where thermal runaway occurs in the LIB pack situated beneath the passenger compartment. To simulate a natural progression of the fire, no external ignition sources were introduced in the test setup. The aim was to observe the inherent fire spread behaviour under realistic conditions where the vehicle’s own materials and design characteristics could influence the development and intensity of the fire.
The test involved inducing thermal runway in a single LIB cell (NMC 622 pouch-type) enclosed within the metal housing of the LIB pack. A 575-Watt electric heating sheet measuring 90 × 65 mm was used to initiate the process. To facilitate this, the LIB pack was physically removed from the BEV, and its top cover was opened for access. Figure 5a shows the arrangement of modules within the LIB pack and the placement of the heating elements, thermocouples, and voltmeters. The LIB pack, with a capacity of 72.6 kWh, consists of 30 modules connected in series, where each module is formed by connecting six assemblies in series, with each assembly comprising two cells connected in parallel. The specifications of this configuration are presented in Table 1. During the test, three heating sheets were attached to the outer surfaces of the end-cells in Modules 8, 12, and 16. Of these, only one sheet (H-M) was activated to trigger thermal runaway, with the remaining two sheets (H-11-4 and 12-0) reserved as backups in case of a malfunction in the heating process. Prior to disassembly, the BEV was fully charged under normal conditions. Post-disassembly, the voltage across the modules was measured, as shown in Fig. 5b, with each module registering approximately 24.82 V, or about 4.17 V per cell, indicating a charge level exceeding 95%. Thermocouples were affixed to both the top and bottom surfaces of the metal casing of the LIB pack, as depicted in Fig. 5c. Following instrumentation, the LIB pack was reinstalled into the BEV to commence the testing phase.
Figure 6a through e show the time-dependent variations in several thermophysical quantities observed during the BEV fire. Key moments in the test are marked on the graphs with vertical arrows, annotated with numbers or small letters, corresponding to major observed events. These events are catalogued chronologically in Table 2. For analytical clarity, the development of the BEV fire was segmented into six district stages. These stages are labelled with capital letters on the graph and are defined based on principles commonly applied to compartment fires [24], as outlined below:
Figure 6
Time dependent measurements: (a) Heat release rate of the BEV fire; (b) Gas–temperature variation with time beneath the test-rig ceiling; (c) temperature variations with time on the surfaces of LIB pack housing; (d) heat flux variation with time at the floor level in Areas 1–5; and (e) temperature variations with time inside the BEV
The Main Events Occurred During Testing and the Definition of Stages
No
Main events
Time
No
Stages
Period
1
The onset of heating a single LIB cell
00′00″
A
LIB-cell heating
00′00″–08′24″
2
The first vent-gas observation around the LIB pack
08′24″
B
Gas venting
08′24″–21′28″
3
A deflagration with the first ignition
21′28″
C
Fire growth
21′28″–22′28″
4
Achievement of an indicative heat-flux level in Area 1
22′28″
D
Indicative heat-flux
22′28″–25′15″
5
Achievement of indicative heat-flux levels in Areas 4 & 5
22′58″
E
Fully developed fire
25′15″–34′00″
6
Achievement of an indicative heat-flux level in Area 2
23′16″
F
Fire decay
34′00″–
7
Achievement of an indicative heat-flux level in Area 3
25′15″
–
–
–
8
The end of thermal runaways
29′58″
9
The onset of fire decay stage
34′00″
(1)
Stage A: LIB-Cell Heating Stage
The experiment commenced with the initiation of an internal failure in a single LIB cell, marked as Event 1. This process began at time zero with the activation of the heating sheet attached to the cell’s exposed surface. The temperature of the heating element progressively increased to around 200°C then sharply rose to above 400°C at Event 2, as depicted in Fig. 6b. This spike in temperature indicated the onset of thermal runaway in the targeted LIB cell within Module 8.
(2)
Stage B: Gas Venting Stage
Subsequent to Stage A, light grey gases were initially observed emerging beneath the vehicle, as illustrated in Fig. 7a. These gases then ascended and accumulated beneath the rig’s ceiling, as shown in Fig. 7b. The culmination of this stage occurred at 21 min and 28 s into the test (Event 3), marking the transition to ignition. This stage, which lasted 13 min and 5 s, is characterised by the gradual venting and accumulation of combustible gases.
Figure 7
The observation of the first ignition and a deflagration: (a) at 8 min 24 s (Event 2); and (b) at 21 min 27 s
Event 3 marked a critical junction in the experiment, as an initial spark was observed around the LIB pack beneath the BEV, captured within the lower yellow dotted oval in Fig. 8a. The upper oval in the same figure illustrates the spontaneous ignition of hot flammable gases surrounding the vehicle. With no external ignition sources in the test setup, it is likely that the spark and subsequent ignition originated from within the LIB pack during internal thermal runaways, and then became visible externally. As depicted in Fig. 8b, the majority of the vapour cloud that had accumulated beneath the ceiling during Stage B combusted almost instantaneously. This rapid combustion, demonstrated in Fig. 8b through 8e, propelled a substantial volume of combustion gases outward at subsonic speeds, indicative of a deflagration vent [25]. This marked the onset of measurable HRR. Following this event, a fire quickly escalated, progressing to spontaneous burning within the partially enclosed space, as shown in Fig. 8f through h. This period of rapid fire growth, defined from 21 min and 28 s to 22 min and 28 s, saw gas temperatures beneath the ceiling (TC_GB, GW, GR, GCRW, GFW, and GRW) spike to above 1000°C within a remarkably short duration, despite maintaining a level around 80°C just before ignition. This dramatic temperature rise highlights the intense and swift nature of the fire development during this stage.
Figure 8
The observation of deflagration at Event 3: (a) at 21 min 28 s (Event 3–1); (b) at 21 min 28 s (Event 3–2); (c) at 21 min 28 s (Event 3–3); (d) at 21 min 28 s (Event 3–4); (e) at 21 min 29 s (Event 3–5); (f) at 21 min 30 s (Event 3–6); (g) at 21 min 32 s (Event 3–7); (h) at 21 min 41 s (Event 3–8); (i) at 21 min 54 s (Event 3–9); and (j) at 22 min 1 s (Event 3–10)
Figure 6d presents the incident heat fluxes measurements at floor level across Areas 1–5 (HF_1, 2, 3, 4, and 5, as indicated in Fig. 3a. On the graph, sky-coloured arrows annotated with small letters (Events 4–7) mark the instances when heat flux levels reached 20 kW/m2 in each area. This heat flux threshold is akin to the flashover point observed in compartment fires, which represents the critical level of heat flux necessary to cause the spontaneous autoignition of most flammable materials within a compartment [26]. However, given that this experiment was conducted in a simulated section of an underground car park rather than a fully enclosed space, this heat flux condition is described as a critical level indicative of untenable conditions. It signals the potential for spontaneous ignition of combustible materials in adjacent vehicles and other items within the area. This critical condition first occurred in Area 1, outside the immediate fire zone, at 22 min and 28 s (Event 4). The remaining areas reached this critical heat flux level within the subsequent 2 min and 47 s. This timeframe is defined as the critical heat-flux stage, during which the thermal environments of the semi-enclosed area reached a peak, as illustrated in Fig. 9.
Figure 9
The achievement of critical heat flux levels in Areas 1–5: (a) at 22 min 28 s (Event 4); (b) at 22 min 58 s (Event 5); (c) at 23 min 16 s (Event 6); and (d) at 25 min 15 s (Event 7)
Following Stage D, the fire transitioned into the fully developed stage, during which the majority of the BEV’s combustibles burned completely. This phase lasted for 8 min and 35 s. Throughout Stage E, the gas temperature soared to the upper measurement limit of the K-type thermocouples, approximately 1260°C, as shown in Fig. 6b. It is noteworthy that sequential thermal runaways of the BEV’s LIBs also occurred during this stage, lasting a total of 21 min and 35 s, and concluded at 29 min and 58 s (Event 8). The endpoint of these thermal runaways was determined using voltage data from the LIBs, details of which will be discussed in the following section.
(6)
Stage F: Fire decay stage
The intensity of the BEV fire began to diminish 34 min after the test commenced, marking the beginning of the fire decay stage (i.e., Event 9). This stage characterises the gradual weakening of the fire as the available combustible materials—specifically, the fire load of approximately 357 kg–374 kg within the BEV—were consumed, and the thermal outputs decreased.
It should be noted that the combustion gases generated by the BEV fire were intentionally directed through a 1200-mm2 opening positioned on the centre axis of the calorimeter. This setup was designed to ensure that all gases were fully captured by the hood, allowing for accurate measurements of HRR and THR. However, as depicted in Fig. 9d, some gases overflowed from the rig—intended to simulate underground car parks—beyond Stage D. This overflow resulted in an underestimation of the recorded HRR and THR values.
3.2 Fire Spread in the Vehicle
Vehicle fires can be categorised as compartment fires within the realm of fire safety engineering. This classification is due to the car’s metal body acting as a spatial separator, which can influence flame behaviour differently compared to an unconfined fire [27]. Generally, the growth of a fire within a vehicle is influenced by the type and amount of combustible materials present and the ventilation conditions of the compartment containing the fire source. Once the fire becomes sufficiently intense, it can gradually spread to adjacent compartments. In vehicles, smaller compartments such as the motor and passenger areas often experience fires that are more dependent on ventilation than on the fuel available, distinguishing them from typical building enclosure fires.
During this test, the initial ignition of light grey flammable gases was observed around the LIB pack, and almost instantaneously in the BEV’s passenger compartment, as depicted in Fig. 8a. Supporting this observation, temperature readings from thermocouples mounted in the passenger cabin, shown in Fig. 6e, confirmed the ignition location. At the moment of first ignition (Event 3), temperatures surged above 400°C at TC_WR, CRF, and CRS, while remaining below 200°C at TC_CFS, T, and M; similarly, lower temperatures were noted at TC_WF, CD, and CFF. In Stage C, radiation from the hot gas jet accumulating at the ceiling significantly heated the upper sections of the BEV, including the top, windows and window seals, a phenomenon highlighted by the yellow dotted oval in Fig. 8i. This effect is specific to the semi-enclosed conditions of the test rig and would not typically occur in fires in surface car parks. As shown in Fig. 8j, the intense heat expedited the melting of the window glass, allowing air to flow into the passenger compartment and intensifying the growth of the ventilation-controlled compartment fire. By Stage D, temperatures at TC_CRS, CD, WF, WR, and T rapidly climbed to above 500°C, indicating that the fire had spread to the front motor compartment after consuming most of the flammable materials in the passenger area. With the onset of Stage E, significant temperature increases to nearly 900°C were recorded at TC_M, CFF, CFS, and CRF, marking the combustion of the majority of the interior flammables. These temperatures reached their peak within approximately 11 min and 30 s and persisted through Stage F, as the thermal conditions inhibited cooling in the affected areas.
During this test, the interior temperatures of the BEV began to rise at Event 3, as depicted in Fig. 6e. However, the temperature increase of the LIB pack commenced at Event 2 due to a sequence of thermal runaways, as evidenced by Fig. 6c which shows the temperature–time profiles of the LIB pack’s exterior metal housing. Typically, the temperature of the case’s upper cover (TC_PTF, PTC, and PBF) was higher than that of the lower container (TC_PBF, PBC, and PBR). This difference is attributed to the buoyancy-driven movement of hot gases and the top cover’s lower heat capacity compared to the more robustly designed lower container, which is thicker to protect the internal LIB cells from external impacts during driving. In the early gas venting stage (Stage B), the temperatures at the middle, back, and front of the casing rose rapidly in the order of TC_PTC, PTR, PTF, PBC, PBR, and PBF. This progression mirrors the pattern of thermal runaway propagation within the pack, a topic that will be explored further in the discussion section. During Stage B, the top-middle and -rear regions of the housing reached temperatures above 400°C, and it is important to note that the metal cover heated to over 500°C merely from the convective flow of the unlit gases. At Event 3, only minor temperature spikes were observed in the pack (TC_PBC and PBR), coinciding with the first ignition in the passenger compartment. Throughout Stages E and F, the temperatures of the upper cover, which had already elevated in earlier stages, either remained steady or increased gradually, while the temperatures of the lower container continued to rise slowly. The sustained high temperatures until the end of the test were facilitated by the metal casing’s high heat capacity and the ongoing thermal conditions.
4 Discussion
4.1 The Severity of BEV Fires in Underground Car Parks
To understand the severity of the BEV fire within the partially enclosed test-rig, several critical aspects were analysed: (1) HRR—Comparing the difference in HRR between semi-enclosed and open fires; (2) Initial Fire Growth—Measuring the upper fire-plume temperature during the initial stages of fire development; (3) Average Fire-Plume Temperature—Assessing the temperature during the fully developed fire stage; (4) Duration of Buring—Timing the overall duration of the fire; (5) Fire-plume Temperature Distributions—Examining temperature distributions within the semi-enclosed area; (6) Pre-Combustion Vapour Cloud Temperature—Determining the temperature necessary to activate fire suppression systems; and (7) Incident Heat Flux—Measuring the incident heat flux in the vicinity of the fire origin. This test addresses a worst-case scenario for BEV fires in underground car parks, which manifests only under specific conditions such as reaching the lower flammable (or explosive) limit. It is crucial to note that such tests are not indicative of frequent occurrences but represent severe, albeit rare, hazards. In light of the paucity of data on such accidental hazards, the comparative analysis of gas temperatures at identical locations (TC_GB, GW, GR, GCRW, GFW and GRW) between the semi-enclosed and open BEV fire setups offers essential insights into the risks associated with BEV fires in underground car parks. While the semi-enclosed BEV fire delineates a worst-case scenario, the data for the open BEV fire are drawn from three repeated full-scale tests (Tests 3–5 as noted in [14]), providing a robust basis for understanding the variations and potential severity of such fires.
Figure 10a presents the HRRs from two BEV fires generated in partially enclosed (red) and open (black) spaces, corresponding to simulated underground and surface car park environments, respectively, as port of the current and companion tests [14]. These tests aim to capture the differing fire behaviours typical of underground versus surface car park settings. The data reveal that the BEV fire in the underground setting burns more intensely and consumes fuel more rapidly compared to the surface BEV fire. It should be noted, however, that the HRR recorded during the present test may be underestimated due to the overflow of combustion gases during measurement, as previously discussed in Sect. 3.1. Consequently, underground BEV fires are expected to cause more significant damage, a factor largely attributed to the rapid rate at which the fire spreads to adjacent combustibles.
Figure 10
Comparisons of data obtained in the enclosed and open configurations [14]: (a) HRR variations with time; and (b) Experimental temperature–time profiles with fire curves
Figure 10b illustrates the time-dependent temperature variations of the upper fire-plume in partially enclosed (red) and open (black) BEV fires. In the experimental data, bars and circles represent the maximum (or minimum) and average temperatures recorded by the thermocouples (i.e., TC_GB, GW, GR, GCRW, GFW, and GRW as shown in Fig. 4a), respectively. To assess the initial growth phases of these fires, conventional fire curves from fire safety engineering were superimposed on the experimental data. For comparative analysis, the onset of these curves was aligned with Event 3, marking the moment of initial ignition and deflagration in the test.
Standardised fire curves are commonly employed to assess the fire resistance of structural elements such as columns, slabs, beams, and walls across various industries including building construction, tunnel, offshore platforms, and petrochemical facilities. These elements are subjected to a thermal load equivalent to a fire curve in a controlled furnace setting. The cellulosic fire curve, designed based on typical burning rates of materials used in general building constructions, contrasts with the chemical fire curves, which are formulated to represent the maximum thermal loads from the most severe fire scenarios, in response to dramatic tunnel fires in the last century [28]. For instance, the RWS fire curve is derived from a full-scale fire test involving 1500 l of petrol over a 4-m2 burning area with a 0.45-m2 opening, simulating a catastrophic crash involving a 50-m3 petrol tanker in a tunnel, representing the epitome of a worst-case scenario. It is important to note that the thermal load generated by a BEV fire in a confined space does not equate to those modelled by these standardised fires. However, during the initial growth stage of the fire (i.e., Stage C), the gas temperature in the BEV fire escalated as rapidly as, or even faster than, those depicted by the cellulosic and hydrocarbon fire curves, which represent severe and maximal fire conditions, respectively. Remarkably, the temperature in the BEV fire reached approximately 1000°C within just one minute, aligning with the rapid escalation observed in the hydrocarbon fire curve, which is noted for having the fastest growth rate among the curves.
The dramatic rise in temperature was triggered by the near-simultaneous ignition of an extensive flammable vapour cloud, composed of off-gas and electrolyte fumes from the fifteen LIB modules, which had accumulated beneath the ceiling over a period of 13 min and 5 s. This accumulation set the stage for a deflagration event, driven by several factors: (1) Pre-combustion thermal energy of the flammable gases, which increased the likelihood of ignition; (2) Elevated gas density within the confined space, reaching the lower flammable limit necessary for combustion; and (3) Presence of a thermal or electrical trigger within the LIB pack that catalysed the ignition. It is important to note that this explosive burst of energy, resulting in rapid combustion, is rarely observed in conventional ICEV fires in similar semi-enclosed configurations. In typical ICEV fire scenarios, the thermal energy from combustibles is gradually released through a mild burning process once smoke formation is evident. Concerning Stage E, the single BEV fire in the semi-enclosed space lasted approximately 8 min and 15 s, which is notably shorter than the durations predicted by standard fire curves. However, it is anticipated that if multiple vehicles were parked in close proximity to the fire source, the duration of Stage E would be extended. This extension would be due to deflagration vents and fire plumes igniting nearby vehicles, as observed in this test, thereby intensifying and prolonging the fire scenario.
The temperature variations of the upper fire-plume in the semi-enclosed BEV fire (red) were compared with those of the open BEV fire (black), as depicted in Fig. 10b. These measurements were taken at identical spatial coordinates during the current and companion tests [14]. Throughout Stages D, E, and F, the average temperature of the hot combustion gases in the semi-enclosed setting was more than double that observed in the open BEV fire. This significant increase indicates that the confined space produced at least sixteen times more concentrated radiant heat from the upper fire-plume. According to the fundamental principles of radiant heat transfer, this intensification accelerated the spontaneous ignition of the vehicle’s remaining combustible components far more than in the open space. Consequently, as evidenced in Fig. 10a, the vehicle burned more intensely and quickly in the confined configuration than it would have in the open area.
Although the installation of automated fire suppression systems—such as sprinkler and water mist systems—in car parking areas is not yet mandated in most countries, their effectiveness in preventing the spread of fire within these environments is widely recognised. The thermal bulbs in these systems are typically designed to activate at temperatures ranging from approximately 57°C to 79°C, varying by environmental conditions. During the test, the temperature of the uncombusted gas reached 57°C, a critical threshold for the activation of fire suppression systems, 12 min and 16 s after the initial gas venting (Event 2) as depicted in Fig. 11a. This measurement was taken at the ceiling level (TC_G), where fire suppression systems are commonly installed. Subsequently, the temperature at this location rose to approximately 80°C. Concurrently, temperatures at lower levels, specifically at the cabin floor (TC_CRF) and the pack lid (TC_PTR), escalated to around 140°C and 560°C, respectively, as shown in Fig. 11b and c.
Figure 11
Time-dependent temperature variations in the early stages A and B: (a) Gas-temperature variation with time beneath the test-rig ceiling; (b) temperature variations with time inside the BEV; and (c) temperature variations with time on the surfaces of LIB pack housing
Figure 12a and b illustrate the horizontal temperature distributions of the hot fire-plume at the ceiling height (z = 2270 mm) following the deflagration vent, capturing the major events (Events 3–9, as listed in Table 2) along the x- and y-axes as outlined in Fig. 3. These axes include temperature readings from thermocouples TC_GB, GW, GR, GCRW, GFW, GRW, and TR_i_U2 (where i = 1, 2, and 5). At each event, the area around the rig’s corner (Area 0) consistently recorded the highest ceiling temperatures, ranging from 542.8°C to 1263.7°C, with temperatures generally decreasing outward in both the x- and y-directions. At the time of the initial ignition point (Event 3), ceiling temperatures near the fire source in Areas 1 and 5 were approximately 520°C and 400°C, respectively. These temperatures rose significantly, reaching over 815°C and 650°C, respectively, following the establishment of critical heat-flux levels during Events 4 and 5. Such thermal conditions could allow significant radiant heat to impinge on the combustible exterior components of vehicles parked in Areas 1 and 5, potentially causing spontaneous ignition. Throughout Events 3–9, the temperature of the upper gas layer in Area 2 varied from 216.5°C to 772.4°C, paralleling the upper gas temperatures observed in the surface car park BEV fire shown in Fig. 10b.
Figure 12
Temperature distributions within the test rig: (a) Ceiling-jet temperature distribution in x-direction at z = 2270 and y = 2800 mm; (b) ceiling-jet temperature distribution in y-direction at z = 2270 and x = 1550 mm; and c Gas temperature distribution in z-direction at y = 2800 mm
Figure 12c illustrates the vertical temperature distributions in Areas 0, 1, and 2 at the major events. The x-axis represents temperature, and the y-axis corresponds to the z-coordinate (height). The values at z = 1200 mm and 400 mm in Area 0 reflect the average temperatures in the passenger compartment and around the LIB modules within the pack, respectively. In general, the temperatures increased along the z-direction. The most significant temperature variations along the z-axis were recorded in Areas 0, 1, and 2, reaching maximum temperatures of 819.1°C, 601.4°C, and 518.6°C, respectively. At the height of the tyres and LIB pack (z = 400 mm), the ambient temperatures exceeded 500°C in Area 1 and 300°C in Area 2. These elevated temperatures in the lower regions could contribute to the ignition of any remaining combustible materials.
In BEV fires occurring in underground car parks, radiant heat from the fire source and ceiling jet can pose significant risks to flammable exterior materials on surrounding vehicles, including coating layers, window seals, tyres, and other parts. This heat can also affect materials inside the passenger compartment by penetrating through transparent window glass. A fire in this scenario may spread more rapidly than it would in a surface car park. Figure 13a presents the heat fluxes at predetermined locations, as indicated in Fig. 4a, alongside the corresponding regional temperatures at the height of the heat flux gauges, shown in Fig. 3, over the same time period. During this test, the incident heat flux on exposed surfaces of nearby objects—such as structural elements, BEV charging stations, or vehicles parked near the fire source in Area 1—rose sharply, exceeding 225 kW/m2. At the same time, ambient temperatures surpassed 580°C in Stage E. It is important to note that heat flux levels above 225 kW/m2 represent extremely hazardous conditions. For comparison, the maximum radiation levels considered survivable for tunnel users and firefighters equipped with protective gear are generally around 2.5 kW/m2 and 5 kW/m2, respectively. This heat flux is comparable to, or more severe than, those produced in building compartment fires, as illustrated in Fig. 13b. The coloured bars in the figure represent the peak heat fluxes measured in the current test, while the black-patterned bars denote values from previous research [29‐35]. Based on the heat flux data obtained from this test, any vehicle parked and charging equipment near the fire source in an underground car park faces a high risk of spontaneous ignition once the fire begins.
In this test, the electric heating sheet triggered an internal short circuit in a single LIB cell within the metal housing, leading to the initiation of the BEV fire. Subsequently, the initial thermal runaway (TR) appeared to propagate to adjacent cells, though this process remained invisible from the outside. To better understand the rate and direction of TR propagation within the pack, time-dependent variations in the temperature and voltage of each of the thirty LIB modules were analysed, as shown in Fig. 14. In this figure, solid lines represent temperature data, while dotted lines denote voltage measurements. Vertical arrows numbered on the graph symbolise the instants of TR and key events, which are listed in chronological order in Table 3. To improve visual clarity, the entire time-dependent data set has been divided into three graphs, shown in Fig. 14a, b, and c, with less significant data lines omitted for simplicity.
Figure 14
Variations in temperature and voltage of LIB modules: (a) From 8 to 18 min; (b) from 18 to 24 min; and (c) from 24 to 30 min
Achievement of critical heat-flux levels in Areas 4 & 5
22′58″
2
The first vent-gas observation around the LIB pack
08′24″
B17
The onset of thermal runaway at Module 29
23′00″
B1–2
The first failure of the LIB assembly at Module 8
08′55″
6
Achievement of a critical heat-flux level in Area 2
23′16″
B2
The onset of thermal runaway
at Module 26
10′37″
B18
The onset of thermal runaway at Module 13
25′20″
B3
at Module 7
11′05″
7
Achievement of a critical heat-flux level in Area 3
25′15″
B4
at Module 5
13′12″
B19
The onset of thermal runaway
at Module 14
25′20″
B5
at Module 4
13′37″
B20
at Module 15
25′40″
B6
at Module 27
13′58″
B21
at Module 22
25′53″
B7
at Module 1
14′02″
B22
at Module 19
25′58″
B8
at Module 6
14′26″
B23
at Module 11
25′59″
B9
at Module 12
15′08″
B24
at Module 9
26′09″
B10
at Module 25
18′42″
B25
at Module 21
27′19″
B11
at Module 23
19′35″
B26
at Module 10
27′22″
B12
at Module 30
20′00″
B27
at Module 16
27′33″
B13
at Module 3
20′25″
B28
at Module 20
28′00″
B14
at Module 24
20′32″
B29
at Module 18
28′11″
B15
at Module 2
20′45″
B30
at Module 17
28′22″
3
A deflagration with the first ignition
21′28″
8
The end of thermal runaways
29′58″
4
Achievement of a critical heat-flux level in Area 1
22′28″
9
The onset of fire decay stage
34′00″
The first failure of the LIB assembly, consisting of two cells, was identified at 8 min and 55 s (Event B1–2). However, the initial TR was observed 32 s earlier (Event 2), indicating a 32-s delay in TR propagation from the initially triggered cell to the adjacent cell within Module 8’s LIB assembly. It is noteworthy that this delay occurs because the two cells are electrically connected in parallel to form one assembly, forming a single assembly. The assembly’s voltage only begins to drop once both cells have fully failed, marking the onset of complete assembly failure.
The first TR in Module 8 triggered adjacent Modules 26 and 7, initiating their TRs within approximately 1 min and 35 s. TRs continued to propagate steadily through the LIB modules until a cascade of TR initiations occurred around 13 min and 12 s (Event B4), at which point the LIB pack’s interior temperature exceeded 500°C. This elevated thermal condition likely accelerated the TR propagation. As the sequence progressed, the interior temperature rose globally to around 600°C and locally to approximately 900°C. Another series of TRs occurred between Event B10 (18 min and 42 s) and Event 3 (the deflagration with the first ignition).
The initial TR in Module 8 propagated sequentially to Modules 26, 7, 5, 4, 27, 1, 6, 12, 25, 23, 30, 3, 24, 2, 28, and 29, all of which are located in the rear and rear-mid areas of the LIB pack, as shown in Figs. 5a, 14a and b. TRs in these modules occurred earlier than in the front and front-mid modules (Modules 9–22). This can be attributed to the presence of two pressure relief holes at the rear-top of the LIB pack, near Modules 2, 3, 28 and 29. Since heat was primarily transferred by convection within the housing, it is likely that the hot gases produced by the initial group of modules (including Modules 8, 26, and 7) surged towards the housing’s edge, heating the nearby modules (Module 1, 5, 25, and 27). As the hot gases were expelled through the pressure relief holes, they transported heat to Modules 2, 3, 28, and 29 at the rear of the pack.
The deflagration at 21 min and 28 s was triggered by the near-simultaneous ignition and combustion of all flammable gases generated by fifteen modules, which had accumulated for approximately 13 min and 5 s in the underground car park. After ignition, the BEV fire progressed rapidly, reaching Stage D in about one minute. However, the internal temperature of the LIB pack did not rise significantly during this time. In Stage E, the remaining thirteen modules—Modules 13, 14, 15, 22, 19, 11, 9, 21, 10, 16, 20, 18, and 17—failed in rapid succession within 3 min and 20 s, as depicted in Fig. 14c. All TRs concluded at 34 min (Event 9), lasting a total of 25 min and 37 s in the partially enclosed BEV fire test. This duration is nearly half the 51 min and 15 s that TRs persisted during the BEV fire in the surface car park, as shown in Fig. 15 [14].
Figure 15
Temperature–time profiles of the LIB pack’s interior and exterior [14]
The growing market share of BEVs has raised significant concern about the risks of BEV fires in underground car parks, presenting a serious societal threat. The primary risk of a fire involving a single modern BEV in an open configuration was experimentally investigated in the companion research [14]. That study analysed key aspects including fire intensity, duration, growth rate, heat release rate contributions, and the impact of jet flames discharged from the LIB pack. This study, however, focused on (1) examining the secondary risk associated with a partially enclosed configuration, specifically simulating a corner segment of an underground car park, and (2) assessing how this secondary threat could exacerbate the primary hazard.
A key distinction between BEV fires in open and semi-enclosed configurations is the occurrence of a deflagration vent. This violent flow was caused by the near-simultaneous ignition of the grey smoke that had accumulated beneath the ceiling for approximately 13 min, emitted by 15 LIB modules. This phenomenon led to a rapid initial fire growth, faster than or comparable to conventional fire curves used in fire safety engineering. Such an explosive burst of energy is rarely observed in open BEV fires, or in ICEV fires under conditions similar to the current semi-enclosed setup.
The secondary key difference was the intensity and duration of burning in the semi-enclosed BEV fire, which were higher and shorter, respectively, compared to the open BEV fire. Following the deflagration vent, the BEV experienced rapid combustion, with radiation from the ceiling plume accelerating fire development during the growth stage. In the fully developed fire stage, the average combustion gas temperatures above the vehicles were approximately 1100°C in the confined fire and 500°C in the open fire. This suggests that underground car parks are subject to more severe burning than surface car parks, due to at least sixteen times more intense radiant heat feedback from the ceiling jet to the combustible materials. The incident heat fluxes on the gauges at a height of 1270 mm exceeded 225 kW/m2, a highly dangerous condition comparable to, or even more hazardous than, typical building compartment fires. The durations of the burning stages, from ignition to full fire development, were approximately 12 min and 30 s for the semi-enclosed BEV fire and 20 min and 30 s for the open BEV fire. The thermal runaway durations in the LIB packs were measured at approximately 25 min and 37 s for the semi-enclosed fire and 51 min and 15 s for the open fire.
Consequently, the thermal characteristics of the BEV fire largely dictated its burning behaviour in open space, while in the semi-enclosed configuration, the fire behaviour was primarily influenced by the restricted environment. This confinement resulted in heat being reflected back onto the fire source through both radiation and convection. A semi-enclosed vehicle fire, which burns more intensively and rapidly, could pose a significant threat to nearby objects, including adjacent cars, structural members (such as columns, slabs, beams and walls), and BEV charging stations.
The findings from this study could primarily aid first responders in managing BEV fire incidents and, secondarily, inform the reassessment of current fire protection standards and codes related parking structure design. The grey gases released by LIB packs during the early stages of BEV accidents pose a significant deflagration risk and should be forcefully vented from confined spaces by first responders using appropriate equipment. The experimental data presented here can also contribute to a broader public understanding of the risks associated with BEV fires in underground car parks, supporting the reconsideration of fire protection requirements. However, further studies are needed to provide a more comprehensive basis for revising existing standards.
Declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Boehmer H, Klassen M, Olenick S (2020) Modern vehicle hazards in parking structures and vehicle carriers. Report FPRF-2020-07, Fire Protection Research Foundation
3.
Meraner C (2023) Car park fires: a review of fire incidents, progress in research and future challenges. In: 7th International Conference on Fires in Vehicles—FIVE 2023, April 24–25, Stavanger, Norway, pp 7–23
Storesund K, Sesseng C, Mikalsen RF, Holmvaag OA, Steen-Hansen A (2020) Evaluation of fire in Stavanger airport car park 7 January 2020, RISE-report 2020:91, RISE Fire Research
Watanabe N, Sugawa O, Suwa T, Ogawa Y, Hiramatsu M, Tomonori H, Miyanoto H, Okamoto K, Honma M (2012) Comparison of fire behaviors of an electric-battery-powered vehicle and gasoline-powered vehicle in a real-scale fire test. In: 2nd International Conference on Fires in Vehicles—FIVE 2012, September 27–28, Chicago, USA, pp 195–205
10.
Lecocq A, Bertana M, Truchot B, Marlair G (2014) Comparison of the fire consequences of an electric vehicle and an internal combustion engine vehicle. In: 2nd International Conference on Fires in Vehicles—FIVE 2012, September 27–28, Chicago, USA, pp 183–194
11.
Lam C, MacNeil D, Kroeker R, Lougheed G, Lalime G (2016) Full-scale fire testing of electric and internal combustion engine vehicles. In: 4th International Conference on Fire in Vehicles, October 5–6, Baltimore, USA, pp 95–106
12.
Willstrand O, Bisschop R, Blomqvist P, Temple A, Anderson J (2020) Toxic gases from fire in electric vehicles, RISE Report 2020:90, RISE Fire Research
13.
Cui Y, Liu J, Cong B, Han X, Yin S (2022) Characterization and assessment of fire evolution process of electric vehicles placed in parallel. Process Saf Environ Prot 116:524–534. https://doi.org/10.1016/j.psep.2022.08.055CrossRef
Sturm P, Fobleitner P, Fruhwirt D, Galler R, Wenighofer R, Heindl SF, Krausbar S, Heger O (2022) Fire tests with lithium-ion battery electric vehicles in road tunnels. Fire Saf J 134:103695. https://doi.org/10.1016/j.firesaf.2022.103695CrossRef
16.
Joyeux D, Kruppa J, Cajot L-G, Schleich J-B, van de Leur P, Twilt L (2002) Demonstration of real fire tests in car parks and high buildings, European Commission, Technical Steel Research, Final Report 2002, EUR-29466 EN
17.
Shintani Y, Kakae N, Harada K, Masuda H, Takahashi W (2004) Experimental investigation of burning behavior of automobiles. In: 6th Asia-Oceania Symposium on Fire Science and Technology, 17–20 March 2004, Daegu, Korea, pp #6B-5
18.
Special detailed requirements based on occupancy and use. In: International Building Code (IBC), 2021 Edn. Chapter 4 International Code Council
19.
Fire protection and life safety systems. In: International Fire Code (IFC), 2018 Edn. Chapter 9 International Code Council
20.
NFPA 88A—Standard for parking structures, 2019 Edn. National Fire Protection Association
21.
Elsagan N, Mozaffari MH (2023) Literature review on parking of electric vehicles, Report A1–02202.1, NRC CNRC
Bengtsson G (2001) Enclosure fires. Swedish Rescue Services Agency NRS Tryckeri, Sweden
25.
NFPA 68—Standard on explosion protection by deflagration venting, 2018 Edn. National Fire Protection Association
26.
Peacock RD, Reneke PA, Bukowski RW, Babrauskas V (1999) Defining flashover for fire hazard calculations. Fire Saf J 32:331–345CrossRef
27.
NFPA 921—Guide for Fire and Explosion Investigations, 2021 Edn. National Fire Protection Association
28.
Taillefer N, Carlotti P, Lemerle C, Avenel R (2013) Ten years of increased hydrocarbon temperature curves in French tunnels. Fire Technol 49:531–549CrossRef
29.
Behnke WP (1984) Predicting flash fire protection of clothing from laboratory tests using second-degree burn to rate performance. Fire Mater 8(2):57–63CrossRef
30.
Fang JB (1980) Static pressures produced by room fires. NBSIR 80-1984, National Bureau of Standards
31.
Krasny J, Rockett JA, Huang D (1988) Protecting fire fighters exposed in room fires: comparison of results of bench scale test for thermal protection and conditions during room flashover. Fire Technol 24:5–19CrossRef
32.
Holcombe BV, Hoschke BN (1986) Do test methods yield meaningful performance specifications? Performance of Protective Clothing (ASTM STP 900), 1st edn. American Society for Testing and Materials, West Conshohocken, PA, pp 327–339CrossRef
33.
Dale JD, Weckman B, Ackerman M (1995) HFOG (House Full of Gas) III Part B: The Explosion and Aftermath. In: Proceedings in spring technical meeting, The Combustion Institute, Canadian Section, pp #43m
34.
Hwang CH, Lock A, Bundy M, Johnson E, Ko GH (2011) Effects of fuel location and distribution on full-scale underventilated compartment fires. J Fire Sci 29:21–52CrossRef
35.
Bwalya A, Gibbs E, Lougheed G, Kashef A (2015) Heat release rates of modern residential furnishings during combustion in a room calorimeter. Fire Mater 39:685–716CrossRef
