In recent years, the widespread adoption of lithium ion batteries in various fields such as off-grid energy storage, electric tools, and new energy vehicles has raised significant concerns regarding safety and reliability. As a power source for electric vehicles, the lithium ion battery market has expanded rapidly, but issues like range anxiety and safety risks have become hot topics. Abuse scenarios such as high heat, mechanical shock, and compression can trigger thermal runaway reactions in lithium ion batteries, potentially leading to severe thermal propagation. Thermal propagation refers to a chain reaction where a single cell undergoes thermal runaway and spreads throughout the battery system. During this process, battery cells and combustible materials within the battery pack can combust violently, releasing large amounts of heat, toxic and flammable gases, posing serious threats to personal safety, property, and the environment. With numerous electric vehicle fire incidents reported annually, standards like GB 38031-2020 have mandated the inclusion of thermal propagation testing for power batteries as a key evaluation item. This standard, along with international guidelines like UL-9540A, EVS-UN GTR, and ISO-26262, aims to ensure that passengers have sufficient time to escape in the event of a fire. In this study, we focus on nail penetration as a trigger method for thermal runaway in lithium ion battery packs, as it is one of the most probable causes of severe thermal events in electric vehicles. We conducted experiments on both battery modules and packs to verify their safety performance against thermal propagation requirements, analyzed the duration and characteristics of heat diffusion from cells to the pack, and evaluated the effectiveness of fire-fighting water in extinguishing thermal events. Throughout this article, we will emphasize the importance of understanding lithium ion battery behavior under stress conditions.
To begin, we delve into the experimental setup and parameters. The lithium ion batteries used in this study are nickel-cobalt-manganese ternary types, with a positive electrode active material of nickel-manganese-cobalt oxide (in a ratio of 5:2:3) and a negative electrode active material of graphite. We tested two samples: Sample 1 is a module consisting of two prismatic cells connected in series, and Sample 2 is a battery pack comprising a battery management system, cooling system, and 16 modules, each with six cells in series. The battery pack has a rated voltage of 364.5 V and a rated energy of 61.3 kWh. Prior to testing, both the module and pack were charged to 100% state of charge (SOC). The nail penetration tests were conducted according to GB 38031, using a stainless steel needle with a diameter of 6 mm and a cone angle of 45°, penetrated at a speed of 8 mm/s. Temperature monitoring points were strategically placed on the module and pack surfaces to capture thermal behavior. For instance, on the module, nine points (T1 to T9) were set, with T1 at the penetration site and T2 on the cell surface near the insulation plate. On the battery pack, eight points (T-1 to T-8) were used, with T-7 near the pressure relief valve and penetration point. The test rig included a hydraulic press to drive the needle, simulating real-world abuse conditions. This setup allows us to systematically analyze the thermal propagation characteristics of lithium ion battery systems.
The results from the module test revealed critical insights into thermal runaway initiation. Upon nail penetration into cell 1, the module voltage dropped rapidly, indicating internal short circuits due to contact between active materials, current collectors, and the needle. Within 1.5 seconds, an explosion and fire occurred, lasting 8.5 seconds, with temperatures at T1 and T11 (above the pressure relief valve of cell 2) spiking to 528.73°C and showing significant convective heat transfer. The heat release during this phase primarily stemmed from Joule heating from short circuits, chemical reactions, and combustion, with combustion contributing the most. Notably, the needle penetrated only about 16% of the cell’s thickness before triggering thermal runaway, highlighting the sensitivity of lithium ion batteries to mechanical abuse. After the fire subsided, smoke emission persisted for 540 seconds, with cell 1’s surface temperature remaining high at 277.5°C and the inter-cell temperature at 171.9°C, posing a thermal risk to adjacent cells. However, due to directional flame ejection from the pressure relief valve and the presence of insulation plates, cell 2 did not undergo thermal runaway, and the module voltage stabilized at 4.19 V, indicating no thermal propagation. The heat transfer modes observed included convection (rapid temperature rise at T1 and T11), radiation (gradual rise at T6 and T7), and conduction (sustained heating at multiple points), as summarized in the temperature-rate curves. For example, the maximum heating rates were 98.37°C/0.1s for dT1/dt and 78.85°C/0.1s for dT11/dt, while dT6/dt and dT7/dt were much lower at 3.28°C/0.1s and 3.37°C/0.1s, respectively. Post-test inspection showed that cell 1’s internal materials were ejected, the pressure relief valve was torn, and cell 2’s casing was blackened by smoke, underscoring the violent nature of thermal runaway in lithium ion batteries.
For the battery pack test, we observed a more complex thermal propagation process, which we divided into three stages based on smoke or fire events. Stage 1 involved the initial nail penetration into the target module, triggering thermal runaway with spark and flame ejection from the penetration hole. Temperature at T-7 surged to 649.3°C within seconds, reflecting rapid energy release from short circuits. Other points like T-6 also showed quick rises due to heat accumulation against the pack casing. This stage lasted approximately 1000 seconds, with no significant thermal propagation to other cells, meeting the GB 38031 requirement of no propagation within 5 minutes. Stage 2 began at 1150 seconds, when thermal runaway spread to an adjacent cell, causing T-7 to spike to 762°C and other points like T-4 and T-6 to increase. The duration of this stage was about 550 seconds, half that of Stage 1, indicating an acceleration in thermal propagation. Stage 3 commenced at 1500 seconds, with further spread to a third cell, pushing T-7 to 767°C and sustaining high temperatures across multiple points. Throughout these stages, we noted that heat was directed and concentrated within the pack, posing risks to wiring and flammable components. To simulate firefighting in electric vehicles, we applied standard fire-fighting water at 0.45 MPa pressure during Stage 3, spraying continuously for 10 minutes. This reduced surface temperatures to ambient levels, but upon reducing the water flow by half, the pack reignited after 14 minutes, leading to additional fire events and eventual complete destruction after 2 hours and 46 minutes. This demonstrates the challenges in extinguishing lithium ion battery fires, as water may only cool the surface without addressing internal heat sources.
To better summarize the thermal behavior, we can use formulas to model heat generation and transfer. For instance, the total heat release $Q_{\text{total}}$ during thermal runaway in a lithium ion battery can be expressed as:
$$ Q_{\text{total}} = Q_{\text{Joule}} + Q_{\text{chem}} + Q_{\text{comb}} $$
where $Q_{\text{Joule}}$ is the Joule heat from short circuits, $Q_{\text{chem}}$ is the heat from chemical reactions, and $Q_{\text{comb}}$ is the heat from combustion. Based on prior studies, $Q_{\text{comb}}$ often dominates, accounting for over 70% of energy release within 60 seconds. The temperature rise $\Delta T$ at a given point can be related to the heat flux $q$ and thermal properties:
$$ \Delta T = \frac{q \cdot A \cdot t}{m \cdot c_p} $$
where $A$ is the area, $t$ is time, $m$ is mass, and $c_p$ is specific heat capacity. For convective heat transfer, we have:
$$ q_{\text{conv}} = h \cdot (T_{\text{surface}} – T_{\text{ambient}}) $$
with $h$ as the convective heat transfer coefficient. These equations help quantify the rapid temperature changes observed in our tests, particularly at points like T1 and T-7.
Additionally, we present tables to encapsulate key data from the experiments. Table 1 summarizes the battery parameters used in this study:
| Sample | Type | Configuration | Rated Voltage (V) | Rated Energy (kWh) | State of Charge |
|---|---|---|---|---|---|
| Sample 1 | Module | 2 cells in series | 7.4 | 0.5 | 100% SOC |
| Sample 2 | Battery Pack | 16 modules (6 cells each in series) | 364.5 | 61.3 | 100% SOC |
Table 2 outlines the temperature monitoring points and their locations for the battery pack test:
| Temperature Point | Location Description | Maximum Temperature Recorded (°C) |
|---|---|---|
| T-1 | Near pack casing, upper region | 150 |
| T-2 | Mid-section of pack side | 180 |
| T-3 | Lower region near cooling system | 120 |
| T-4 | Adjacent to target module | 300 |
| T-5 | Above wiring harness | 200 |
| T-6 | Near penetration hole on casing | 500 |
| T-7 | At pressure relief valve, penetration site | 767 |
| T-8 | Opposite side of pack | 100 |
Table 3 details the stages of thermal propagation in the battery pack, including durations and key events:
| Stage | Start Time (s) | End Time (s) | Duration (s) | Key Events | Peak Temperature at T-7 (°C) |
|---|---|---|---|---|---|
| 1 | 0 | 1000 | 1000 | Nail penetration, initial fire, smoke | 649.3 |
| 2 | 1150 | 1700 | 550 | Thermal runaway spreads to second cell, fire reignites | 762 |
| 3 | 1500 | 2000+ | >500 | Spread to third cell, continuous burning, firefighting applied | 767 |
From these results, we can derive insights into the acceleration of thermal propagation. The reduction in stage durations suggests that once thermal runaway initiates in a lithium ion battery pack, subsequent events may occur faster due to pre-heating and cumulative heat buildup. This aligns with models of thermal runaway propagation, where the heat transfer rate $ \dot{Q} $ between cells can be expressed as:
$$ \dot{Q} = k \cdot \Delta T \cdot A / d $$
where $k$ is thermal conductivity, $\Delta T$ is temperature difference, $A$ is contact area, and $d$ is distance. As $\Delta T$ increases with each event, $\dot{Q}$ rises, accelerating propagation. Furthermore, the energy release during nail penetration can be estimated using the short-circuit current $I_{sc}$ and internal resistance $R_{int}$:
$$ Q_{\text{Joule}} = I_{sc}^2 \cdot R_{int} \cdot t $$
In our tests, the rapid voltage drop indicates high $I_{sc}$, contributing to immediate heat generation. This underscores the vulnerability of lithium ion batteries to internal shorts, a critical safety concern for electric vehicles.
Regarding firefighting effectiveness, our experiment with water spray highlights limitations. The heat removal rate $ \dot{Q}_{\text{cool}} $ from water can be calculated as:
$$ \dot{Q}_{\text{cool}} = \dot{m} \cdot c_{p,\text{water}} \cdot (T_{\text{out}} – T_{\text{in}}) + \dot{m} \cdot L_v $$
where $\dot{m}$ is water mass flow rate, $c_{p,\text{water}}$ is specific heat, $T_{\text{out}}$ and $T_{\text{in}}$ are outlet and inlet temperatures, and $L_v$ is latent heat of vaporization. At full flow, $ \dot{Q}_{\text{cool}} $ was sufficient to cool the surface, but reduced flow lowered it, allowing internal heat to reignite cells. This implies that standard firefighting may not penetrate the sealed battery pack of a lithium ion battery system, necessitating additional measures like integrated extinguishers or thermal barriers.
In conclusion, our study on lithium ion battery packs triggered by nail penetration reveals that thermal runaway can initiate rapidly with minimal penetration depth, releasing immense energy through combined heat sources. While modules may contain propagation with insulation, packs show staged propagation with accelerating events, even when meeting regulatory delays. The lithium ion battery’s design must account for directed heat fluxes to protect adjacent components, and firefighting strategies require enhancements beyond surface cooling. We recommend further research into materials that mitigate short-circuit risks and advanced thermal management systems to suppress propagation. Ultimately, ensuring the safety of lithium ion batteries in electric vehicles demands a multifaceted approach, combining robust testing, intelligent design, and effective emergency response protocols. Throughout this work, we have emphasized the critical role of understanding lithium ion battery behavior under abuse conditions, as it directly impacts the widespread adoption of clean energy technologies.
To expand on the discussion, we can consider the implications for battery management systems (BMS). A BMS in a lithium ion battery pack typically monitors voltage, temperature, and current to prevent abuse conditions. However, during thermal runaway, traditional BMS may be overwhelmed by rapid changes. We propose incorporating predictive models based on real-time data, such as using the Arrhenius equation for reaction rates:
$$ k = A \cdot e^{-E_a / (R T)} $$
where $k$ is the rate constant, $A$ is the pre-exponential factor, $E_a$ is activation energy, $R$ is gas constant, and $T$ is temperature. By tracking temperature rises, a BMS could anticipate thermal runaway and trigger countermeasures like isolating cells or activating coolant. Additionally, the use of phase-change materials (PCMs) in lithium ion battery packs could absorb heat during thermal events, as described by:
$$ Q_{\text{PCM}} = m_{\text{PCM}} \cdot \Delta H_f $$
where $\Delta H_f$ is latent heat of fusion. Integrating such materials might delay propagation, providing more time for evacuation or firefighting.
Another aspect is the environmental impact of lithium ion battery fires. The release of toxic gases, such as hydrogen fluoride (HF) from electrolyte decomposition, poses health risks. The concentration $C$ of HF over time $t$ can be modeled as:
$$ C(t) = C_0 \cdot e^{-kt} + \frac{G}{V} \cdot t $$
where $C_0$ is initial concentration, $k$ is decay constant, $G$ is generation rate, and $V$ is volume. This highlights the need for ventilation and gas detection in battery enclosures. Furthermore, the recycling of damaged lithium ion batteries becomes complex after thermal events, as materials degrade and become hazardous. Developing safer electrolytes, such as solid-state alternatives, could reduce these risks in future lithium ion battery designs.
In terms of standardization, our findings align with ongoing updates to safety codes. For instance, the GB 38031 requirement for no thermal propagation within 5 minutes is a step forward, but our data shows that even after this period, propagation can accelerate. We suggest extending test durations and including scenarios like partial water exposure to better simulate real-world accidents. International collaboration on lithium ion battery safety standards, referencing UL-9540A and ISO-26262, is crucial for global consistency. Additionally, public education on handling electric vehicle fires, such as avoiding water immersion due to electrical risks, can save lives.
Finally, we reflect on the economic aspects. The cost of lithium ion battery packs is decreasing with technology advances, but safety incidents can lead to recalls and reputational damage. Investing in robust testing, like the nail penetration method we employed, can mitigate long-term losses. Insurance models for electric vehicles might also incorporate thermal propagation risks, incentivizing manufacturers to enhance safety. As the lithium ion battery market grows, continuous innovation in materials science, such as using silicon anodes or high-nickel cathodes, must balance energy density with safety, ensuring that lithium ion batteries remain a viable and secure power source for the future.
Through this comprehensive analysis, we have explored multiple dimensions of lithium ion battery safety, from experimental data to theoretical models. The repeated emphasis on lithium ion battery throughout this article underscores its centrality in modern energy systems, and our hope is that this research contributes to safer and more reliable applications. As we move towards a sustainable energy future, the lessons learned from thermal propagation studies will be invaluable in shaping the next generation of lithium ion battery technologies.