Characterization of Thermal Runaway of Lithium Ternary Power Battery in Semi-Confined Space

Abstract

In some new energy aircraft powered by lithium-ion batteries (LIBs), the LIBs operate in semi-confined spaces. Therefore, studying the thermal runaway (TR) characteristics of LIBs in such spaces is significant to safety research of new energy aircraft. This paper investigated TR of LIBs in semi-confined space by using external heating, and compared it with the TR characteristics in open space in terms of behavior characteristics and temperature changes of lithium ternary power batteries in semi-confined spaces. The results show that the TR process of LIBs can be subdivided into seven different stages according to the TR characteristics of LIBs. Compared with the TR process of the LIB in open space, TR of the LIB in semi-confined space has an additional explosion stage. In terms of temperature, the maximum TR temperature of the LIB in open space is 708 °C, and the maximum heating rate is 72.3 °C/s, while the maximum temperature in semi-confined space is 552 °C, and the maximum heating rate is 32.1 °C/s. This study is beneficial for the subsequent provision of certain theoretical guidance for LIBs use in semi-confined environments.

1. Introduction

With continuous advancements in the new energy industry, an increasing number of vehicles, including motorcycles, automobiles, and aircraft, are transitioning from fossil fuels to electric power. This shift not only enhances daily convenience but also contributes to global sustainable development. Currently, the primary obstacle to further progress in new energy technologies lies in battery development [1], with aerospace electrification posing particularly significant challenges. Among various battery types available in the market, ternary lithium-ion batteries (LIBs) have demonstrated notable advantages. Their lightweight design, extended lifespan, high energy density, and stable charge and discharge performance have led to successful implementation in the new energy electric aircraft [2,3].

However, studies indicate that increased energy density in batteries correlates with more severe consequences of thermal runaway (TR) incidents, including fires and explosions [4,5]. Compared to other battery types, ternary lithium batteries are prone to heat accumulation under extreme conditions such as high temperatures, rapid charge and discharge cycles, mechanical impacts, and other stressful operational environments. If this accumulated heat cannot be effectively dissipated, the battery temperature will escalate uncontrollably, triggering TR in LIBs [6,7]. Research confirms that TR constitutes the primary cause of LIB-related accidents [8,9,10,11]. To mitigate such risks, scholars have extensively investigated the critical factors influencing TR in LIB systems.

Zhang X. et al. [12] conducted TR propagation experiments at the electric vehicle battery system level and found that individual cells in a module experience sequential and simultaneous propagation, while modules exhibit simultaneous propagation. The harm caused in multiple modules is greater compared to simultaneous TR in a single cell. Liu J. et al. [13] investigated the effects of charging rate, ambient temperature, and aging on TR induced by overcharging, and it was shown that high charging rates and ambient temperatures increase the risk of TR. Lee S K. et al. [14] investigated the electro-thermal behavior of LIBs under different conditions, including internal short circuits of various sizes, shapes, and locations, and found that the size of the internal short circuit is a key factor in determining the onset of TR. Zhang Q. et al. [15] investigated the lower explosion limit of TR gas of LIBs under different charging and discharging environments and proposed a corresponding guideline for the safe use of LIBs. Jin C. et al. [16] investigated the effect of the heating power and the heating energy on LIB components by experimentation and simulation, and it was found that the accumulation of thermal energy was the cause of the TR phenomenon triggered by external heating. Jie D. et al. [17] systematically investigated the TR and combustion characteristics of the combined conditions “heating + overcharging” and “heating + short-circuit”. Compared with the individual conditions, the heating + overcharging condition shortens the safety valve opening time, reduces the TR temperature, and increases the HRR peak, which is the most dangerous scenario.

In addition to the above studies, when TR occurs, the difference of the space environment in which the LIB is located also affects the severity of the consequences of TR accidents. Liu T. et al. [18] carried out overcharging tests using large-size LIBs of different capacities in homemade airtight test containers under different conditions. The results show that changes in battery squeeze pressure are the most sensitive parameter in the early stage of TR and can be used as a reliable early warning signal. Yanhui Liu et al. [19] investigated the criticality of TR of a ternary LIB stack under a thermal boundary in a low-pressure laboratory and measured the temperature characteristics of the battery for safe deflation and TR, and they found that a low-pressure environment can promote the safe deflation of LIBs earlier and more strongly. Wang Z. et al. [20] investigated the effect of spacing on the TR propagation of battery packs in open and closed space environments, and the study showed that spacing and cell alignment play an important role in the TR propagation of batteries. Chen X. et al. [21] conducted experiments on 18650-type LIBs in dynamic pressure chambers and investigated the effects of ambient pressure and heating power on the TR characteristics. The results show that violent behaviors such as jet fire, deflagration, and explosion occur only at high ambient pressure and high heating power; and the hazards of fire risk and toxic/combustible gas emission are closely related to TR behaviors. Liu J. et al. [22] explored the factors affecting the TR of 18650-type LIBs in confined space, and the study showed that a lower state of health accelerates the side reactions before TR. Mao B. et al. [23] investigated the deflagration behavior of LIB modules containing different cathode materials through TR tests under varying ventilation conditions within a confined space. The study found that battery combustion comprised jet flames and deflagration, with NCM811 batteries exhibiting higher hazard potential. Additionally, the deflagration pressure initially increased and subsequently decreased as the vent opening angle was enlarged.

To summarize, there are many kinds of reasons affecting the consequences of TR accidents of LIBs, and currently, most scholars mainly focus on experiments of LIBs in open or closed environments, and most of the experimental materials are 18650-type LIBs. However, for some electric aircraft battery installation locations, the battery is installed semi-confined behind the pilot, which will have restricted gas exchange with the outside world, creating a semi-confined space. Currently, due to the fact that new energy aircraft are still in the development stage, there is limited research on the TR characteristics of LIBs used in aviation power systems under specific environmental conditions. Therefore, this paper investigated the TR characteristics of power LIBs in semi-confined environments by combining experimental parameters. The study aims to explore how semi-confined spaces affect battery TR behavior, providing a theoretical foundation for the application of power LIBs in electric aircraft operating under such environmental constraints.

2. Experimental Setting

2.1. Battery Specimen

The battery sample selected in the electric aircraft series soft-pack, large-size, high-energy-density ternary LIB (as shown in Figure 1), which made in China, with relevant specific parameters provided in

Table 1. Experimental LIB sample parameters.

Name	Parameters
Size	220 × 160 × 10 mm
Capacity	50 Ah
Rated Voltage	3.6 V
Operating Voltage Range	2.5 V–4.3 V
Type	Ternary lithium-ion battery
Cathode	Ternary cathode material
Anode	Graphite

.

2.2. Experimental Apparatus

In LIB TR experiments, external heating is considered the best repeatable triggering method [16]. In this article, the LIB TR experimental platform was built in semi-confined space and open space. The main equipment used in the experiment is the stand, inorganic flame-retardant board, electric heating plate (220 × 160 mm), K-type thermocouple, temperature data collector, video camera, and voltage detector.

The electric heater is placed flat just below the battery to ensure that the lower surface of the battery is covered by the entire heating surface of the plate. The temperature of the LIB will continue to rise due to the heating of the heater. When TR is triggered in the battery, the power of the heater is switched off. To obtain the temperature changes of LIBs during TR activation, a total of six K-type thermocouples were arranged at the positive and negative lugs, the positive and negative sides, the center, and the bottom of the battery, as shown in Figure 2.

2.2.1. Semi-Confined Space LIB TR Experiment Platform

Figure 3 shows the schematic diagram of the semi-confined space LIB TR experiment platform. Semi-confined space experiments rely on the smoke toxicity test box to complete. The experimental box has a combustion chamber size of 90 mm long, 78 mm wide, and 100 mm high. The front side is equipped with an explosion-proof glass for observing the experimental process, and a video camera is set up at a distance of 1 m from the front side for recording the TR process. The front door of the unit has a good seal when the limit screws are tightened; there is a vent at the top (unpainted), and there are twelve (3 × 4) 5 mm unsealed diameter holes in the side of the box. As the combustion chamber is not sealed, there are some pores that can exchange a certain amount of gas with the outside world, thus constituting a semi-confined space environment.

2.2.2. Open-Space LIB TR Experiment Platform

The experimental platforms for studying LIB TR in open and semi-enclosed spaces utilize identical core equipment. The key difference in open space is that the experimental equipment is set up to allow free gas exchange with the external environment, enabling LIB TR gas to interact freely with the outside atmosphere. The specific experimental device is shown in Figure 4.

2.3. Experimental Methods

It has been shown that the higher the LIB state of charge (SOC), the lower the battery stability and the greater the fire hazard [24]. In this study, 100% SOC LIBs with the highest fire risk were selected for TR experiments. Before the start of the experiments, the batteries were set aside for a period of time, and after the battery voltage was stabilized, the 100% SOC experimental samples were obtained using a Neware battery cycler (Made in China) with a constant current charging with 10 A (0.2 C) to a cut-off voltage of 4.2 V. All cells were subjected to three charge and discharge cycles for stabilizing the internal chemistry of the cells before preparation and the start of the experiments.

The procedure is performed as follows: First, place the test battery on the heating plate and sequentially activate the camera, temperature data collector, and voltage data collector, followed by initiating the electric heating plate. When TR is detected through observable LIB shell expansion and smoke emission, immediately deactivate the heating plate while maintaining continuous monitoring of TR progression. The experiment ends when the voltage reaches 0 V and temperatures exhibit a sustained downward trend.

3. Results and Discussion

3.1. Analysis of TR Process

Figure 5 and Figure 6 show the TR process of an LIB under semi-confined space and open space, respectively. As can be seen from the figures, 100% SOC LIBs in the early stage of TR produce smoke, without flame generation, and in the later stage of the performance of the burning state. Based on this, this paper analyzed the TR of the LIB into two main stages, namely the flame-free TR development stage and the flame TR stage.

According to the characteristics of the TR phenomenon of LIBs in two kinds of space, the stage of TR development without flame can be further divided into the following: heating stage (stage I), bulging stage (stage II), smoke injection stage (stage III), and the stage of TR with flame can be further divided into flame injection stage (stage IV), complete combustion stage (stage VI), explosion stage (stage VI), and the end of TR reaction stage (stage VII). The TR characteristics of each stage are described below.

(1)

Flame-free TR development stage

At this time, the LIB is affected by external heating, and TR is in the birthing process. From the experimental results, it can be seen that regardless of the open space or semi-confined space, the LIB TR process is similar in their basic characteristics, there are stages I–III.

Stage I is the initial stage of TR; the battery is continuously heated by the heating plate, and no obvious changes are observed in the LIB under the two kinds of space. When the battery is continuously heated for a certain period of time, gas is generated inside the battery due to the TR reaction, which makes the battery shell start to deform and expand, and the TR of the battery proceeds to Stage II. In this stage, the battery will first expand slightly, and it will rapidly expand and bulge, eventually forming a pike-shaped structure with a wide center and two narrow sides. Stage III is the stage in which smoke is ejected from the inside of the battery to the outside. Some of the responses at this stage are shown below [25,26,27,28]:

$${{(\mathrm{CH}}_2{\mathrm{OCO}}_2\mathrm{Li})}_2\to {\mathrm{Li}}_2{\mathrm{CO}}_3{+\mathrm{C}}_2{\mathrm{H}}_4{+\mathrm{CO}}_2+{\displaystyle \frac{1}{2}}{\mathrm{O}}_2$$

(1)

$${2\mathrm{Li}+(\mathrm{CH}}_2{\mathrm{OCO}}_2{\mathrm{Li})}_2\to {2\mathrm{Li}}_2{\mathrm{CO}}_3{+\mathrm{C}}_2{\mathrm{H}}_4$$

(2)

$${2\mathrm{Li}}^{+}+2e^{-}{+\mathrm{C}}_3{\mathrm{H}}_4{\mathrm{O}}_3\left(\mathrm{EC}\right)\to {\mathrm{Li}}_2{\mathrm{CO}}_3{+\mathrm{C}}_2{\mathrm{H}}_4$$

(3)

$${2\mathrm{Li}}^{+}+2e^{-}{+\mathrm{C}}_5{\mathrm{H}}_{10}{\mathrm{O}}_3\left(\mathrm{DEC}\right)\to {\mathrm{Li}}_2{\mathrm{CO}}_3{+\mathrm{C}}_2{\mathrm{H}}_4{+\mathrm{C}}_2{\mathrm{H}}_6$$

(4)

$$2{\mathrm{Li}}^{+}+2e^{-}+{\mathrm{C}}_3{\mathrm{H}}_4{\mathrm{O}}_3(\mathrm{DMC})\to {\mathrm{Li}}_2{\mathrm{CO}}_3+{\mathrm{C}}_2{\mathrm{H}}_6$$

(5)

$${2\mathrm{Li}}^{+}+2e^{-}{+\mathrm{C}}_4{\mathrm{H}}_6{\mathrm{O}}_3\left(\mathrm{PC}\right)\to {\mathrm{Li}}_2{\mathrm{CO}}_3{+\mathrm{C}}_3{\mathrm{H}}_6$$

(6)

When the internal pressure exceeds the pressure limit of the shell, the battery shell will rupture, resulting in a large amount of gas due to the internal high pressure and the formation of high-speed gas flow to the outside world.

The process can be explained by Equations (7) and (8) [29]:

$$PV=nRT$$

(7)

where P is the pressure (Pa); V is the volume of the gas (m³); T is the temperature (K); n is the amount of substance of the gas (mol); and R is the molar gas constant (J/(mol·K)).

$$P_{\mathrm{in}}=P_{\mathrm{out}}+P_{\mathrm{valve}}$$

(8)

where P_in refers to the internal pressure of the battery, which is derived from the production of gases inside the battery and the accumulation of reaction heat; P_out refers to the external pressure of the battery; and P_valve refers to the rupture pressure of the polymer shell of the battery.

When self-heating occurs in the battery due to TR, the internal temperature T of the battery continues to increase, the internal pressure P_in and the battery volume V increase with it, and in spite of when the volume increases to the tolerance limit, P_in continues to increase. Since the external pressure of the battery (the environmental pressure) does not change significantly, the P_valve also continues to increase with P_in. From when the internal pressure exceeds the shell pressure limit, the shell ruptures.

(2)

TR stage with flame

Stage IV is the stage of flame jet formation. This phase occurs when the self-exothermic reaction triggering LIB TR leads to sustained temperature accumulation. Under these high-temperature conditions, the high-pressure gases produced within the battery (as described by Equations (1)–(6)) carry substantial thermal energy outward. Upon contact with oxygen and reaching the ignition threshold of the ejected vapors, these combustible mixtures culminate in jet flame formation.

Stage V is the stage of complete battery combustion, where the TR of LIBs in the two spaces begins to show significant differences. As the reaction progresses, the battery temperature rises sharply. Owing to the structural characteristics of soft-package LIBs, the initial gas-jet flame emerging from rupture points gradually propagates across the battery surface, transitioning into fierce combustion. In open spaces, unrestricted gas exchange with the air sustains a vigorous and persistent flame. Conversely, in semi-confined spaces, progressive oxygen depletion occurs during combustion. Impeded gas exchange prevents timely oxygen replenishment, leading to gradual flame diminishment until eventual extinction.

Stage VI marks the explosion phase, a phenomenon unique to LIB TR in semi-confined spaces. Although flames are extinguished by the end of Stage V in semi-confined spaces, the combustion chamber retains high temperatures, and partial TR reactions continue. Under such conditions, combustible fumes from LIB TR accumulate due to restricted gas circulation. During combustion, limited air exchange in semi-confined spaces allows oxygen to gradually seep into the chamber. When this oxygen interacts with accumulated TR-generated combustible gases, reignition occurs upon reaching ignition temperatures, ultimately triggering an explosion. In contrast, open spaces facilitate unobstructed gas exchange and maintain sufficient oxygen levels. Combustible gases either burn instantly or disperse into the atmosphere, preventing accumulation. Consequently, this explosion phase is absent in open environments.

Stage VII is the end stage of battery TR. At this stage, the TR reaction of the LIB has ended, the flame goes out, and the temperature gradually decreases. The surface of the battery is black and burnt, and the internal diaphragm is white powder.

For the flame-free TR development stage, there is no significant difference in the performance of the LIB’s TR process in semi-confined space and open space. At this time, TR of the LIB is mainly affected by the high-temperature heat source and the space environment where the battery is located, and the gas exchange has no direct influence on TR.

In the flame TR stage, semi-confined spaces pose higher risks due to restricted gas exchange. This limitation leads to TR-generated combustible gas accumulation and inefficient heat dissipation. When these gases contact outside air, the likelihood of combustion increases significantly, potentially triggering explosions. In contrast, open spaces allow unobstructed gas dispersion and rapid heat dissipation, preventing gas accumulation. Consequently, semi-confined environments exhibit greater TR-related hazards compared to open settings.

3.2. Characterization of TR Temperature Change

(1)

Flame-free TR development stage

Figure 7a,b show the temperature change images of TR Stages I–III in semi-confined space and open space. In order to provide a clear comparison of the temperature change curves of Stages I–III under the two spaces, the curves were localized and enlarged to obtain Figure 7c,d.

As shown in Figure 7c, under the semi-confined space, the temperature of each part of the battery shows a gradual and slightly slow increase due to the influence of heating by the heating plate and the poor air circulation in the semi-confined space, so that the heat cannot be dissipated in time. In the open space, smooth gas exchange results in the temperature of all parts of the battery still being maintained at a certain level, without a significant increase in the trend. In general, during certain durations of heating, such as 0–165 s under semi-confined space and 0–155 s under open space, the overall battery temperature can be maintained at about 40 °C without any obvious rapid warming phenomenon. This is due to the fact that the temperature at this stage is still within the maximum operating temperature range of the LIB, making the battery tolerant to external temperature changes.

A comparative analysis of Figure 7c,d reveals that the Stage I–III temperatures during the TR process do not show significant differences in the two different spatial scenarios. It can be seen that Stages I–III are not the main stages that have a critical impact on the TR results.

(2)

TR stage with flame

Figure 8a,b show TR Stage IV–VII temperature changes of LIBs in semi-confined and open spaces, respectively. Figure 8c,d then correspondingly show the localized magnified images of Stages IV–VII of Figure 8a,b. This process is the most dangerous TR stage during the entire TR period, with the largest fire and the fastest rate of warming.

In open spaces, flame temperature benefits from sufficient oxygen supply, enabling complete combustion with continuous and intense flames. The highest temperature (709 °C) occurs on the negative side at 184 s. In semi-confined spaces, however, oxygen levels progressively decline as combustion proceeds. Rising internal pressure from increasing temperatures further restricts oxygen replenishment to the combustion chamber. This leads to incomplete combustion and eventual flame extinction, with the peak temperature (482.6 °C) observed at the cell center in 352 s.

A single derivative of the temperature rise curve yields the corresponding rate of warming of each part of the LIB during TR. Under semi-confined space, the maximum warming rate of 22.7 °C/s appears at the positive side at 311 s, and the sharp increase in temperature lasts for a total of 92 s (taking the warming rate of any part >5 °C/s as the start time and ending when the warming rate of all parts is <5 °C/s); in open space, the maximum warming rate of 72.3 °C/s is observed at the negative side at 158.5 s, and the sharp temperature increase lasts for a total of 78 s. This indicates that in terms of peak temperature and heating rate, TR of LIBs is much more intense in open spaces than in semi-confined spaces.

For Stage VII, the temperature drop in the semi-confined space is relatively smooth, while the temperature in the open-space LIB after the end of TR experiences a rapid drop, and then tends to slow down; this is due to the free exchange of gas between the LIB and the outside world under open space. The heat can be emitted to the outside world in a timely manner, while in the semi-confined space, due to the restriction of exchange of gas with the outside world, the heat cannot be emitted in a timely manner to the outside environment.

3.3. Exploration of TR Characteristics of LIB Under Semi-Confined Space

Compared with TR of an LIB in open space, semi-confined space has more obvious differences in some TR stages due to certain limitations in gas circulation.

The TR of LIBs triggered by heat involves the evaporation of organic solvents, the decomposition of the SEI layer and separator, the chemical transformation of positive and negative electrodes, and their interactions [28]. These processes generate substantial quantities of combustible gases, releasing significant heat to increase the temperature, which further accelerates the TR reaction.

In the flame-free TR development stage, batteries in semi-confined spaces subjected to heating exhibit an initial warming trend across all components. However, during this process, TR progression is mainly governed by the battery’s internal reactions, while environmental constraints and gas flow exert minimal influence.

During TR combustion in LIBs, environmental ventilation critically determines hazard pathways. In open spaces, unhindered gas exchange maintains sufficient oxygen supply, allowing complete combustion with sustained intense flames. This enables a higher T_max, faster heating rates, and rapid post-reaction cooling, with TR risks primarily dependent on the magnitude of T_max.

Semi-confined spaces exhibit different behavior. Oxygen depletion during combustion, combined with restricted replenishment due to rising internal pressure from heat accumulation, leads to premature flame extinction. Consequently, these environments experience lower T_max values alongside reduced heating rates, while slow heat dissipation prolongs thermal persistence. Crucially, ongoing TR reactions continuously generate combustible gases that accumulate within the combustion chamber due to limited diffusion. Simultaneously, structural constraints allow gradual external air ingress while trapping internal heat, creating delayed explosion risks when accumulated gases mix with infiltrated oxygen. This mixture may reach concentrations exceeding the lower explosive limit, triggering explosion.

Compared to open space, restricted air circulation in semi-confined spaces amplifies three interconnected hazards: impeded heat dissipation, accumulation of flammable gases, and unstable oxygen supply. Current understanding remains incomplete in quantifying relationships between ventilation levels and TR severity, necessitating systematic investigations to clarify these mechanisms.

In future designs of new energy aircraft, if the semi-confined space exists within the LIB operating environment, the TR risk assessment must consider not only the T_max of TR events but also the potential for TR-induced toxic fume generation and diffusion, which may trigger explosions and other hazardous scenarios. Addressing these factors holistically could enhance aircraft safety.

3.4. Future Perspectives

This study investigated the TR behavior of aviation power LIBs in semi-confined environments under simulated practical conditions. Despite achieving insights into TR, certain limitations necessitate further investigation. Experimental constraints prevented precise quantification of temperature thresholds corresponding to sequential TR phases, such as electrolyte decomposition, gas venting, and ignition. Additionally, comprehensive analysis of post-failure battery debris remains essential for assessing secondary hazards. Furthermore, while the current research focused on single-cell behavior, the propagation dynamics of TR within multi-cell battery modules under realistic aerospace operating scenarios require systematic exploration [30,31,32]. Addressing these gaps through advanced instrumentation and multi-scale modeling will significantly enhance the safety prediction capabilities for aviation energy storage systems.

4. Conclusions

In this paper, the behavioral and temperature characteristics of TR triggered by thermal abuse of LIBs in semi-confined space were studied. The process of TR and the temperature change of the battery were analyzed through experiments, in order to understand and analyze differences in TR for LIBs in semi-confined space compared with open space. The main conclusions are as follows:

1.

The TR behavior of LIBs in semi-confined space and open space is different. The TR process triggered by thermal abuse can be divided into two main stages: the flame-free TR development stage and the flame TR stage. Among them, the flame-free TR development stage includes heating, bulging, and jetting. The flame TR stage includes flame injection, complete combustion, explosion, and the end of the TR reaction. In the TR process of LIBs, no significant difference was observed between the TR performance in semi-confined space and open space during the flame-free TR development stage. However, during TR with flame, the two cases exhibit significantly different characteristics. It is particularly noteworthy that the explosion stage only occurs in the TR process of LIBs in semi-confined space, which is a unique stage differentiating it from open space.

2.

The TR temperatures of LIBs in semi-confined space and open space are different. In the flame-free TR development stage, the temperature of both semi-confined space and open-space LIBs can be stabilized at about 40 °C for a certain length of time. In the stage of TR with flame, the TR flame under the open space environment is violent and persistent, with a maximum temperature of 709 °C and a maximum warming rate of 72.3 °C/s. For the semi-confined space, due to the low level of gas circulation with the outside world, combustion ends early, the heat is not fully released, the maximum temperature is only 482.6 °C, and the maximum temperature rises at a rate of 22.7 °C/s.

3.

The key factor influencing differences in TR outcomes between semi-confined and open spaces lies in air circulation. In open spaces with unrestricted airflow, TR reactions of LIBs are markedly intensified. This manifests through a sharp temperature increase in flames and an accelerated heating rate. Conversely, semi-confined spaces with restricted ventilation suppress TR reactions, resulting in premature flame extinguishment and diminished heating rates. However, such restricted airflow may concurrently allow combustible gas accumulation in localized zones, substantially elevating explosion risks and overall hazard potential. For future electric aircraft designs implementing LIBs in semi-confined compartments, meticulous evaluation of spatial impacts on TR propagation is imperative to optimize safety standards.