As the global shift toward electric vehicles accelerates, the rapid growth in the adoption of new energy vehicles, particularly in China, has brought the safety of EV power batteries to the forefront. Thermal runaway incidents in these batteries pose severe threats to vehicle operation and passenger safety, driven by factors such as overcharging, short circuits, mechanical impacts, and environmental temperature fluctuations. These conditions can lead to heat accumulation, triggering a chain reaction that rapidly propagates through the battery system. Developing effective early warning and protection mechanisms is critical for advancing the safety standards of China EV battery technologies. This study focuses on analyzing the causes and evolution of thermal runaway, reviewing current monitoring and protective strategies, and proposing optimized solutions to enhance the overall thermal safety of EV power battery systems. By integrating multi-source data fusion and high-efficiency thermal management, we aim to provide a robust framework for mitigating risks associated with thermal runaway in China’s burgeoning EV market.
Thermal runaway in EV power batteries is primarily initiated by internal short circuits, over-discharging, uneven aging, or external mechanical damage. These factors can cause electrolyte decomposition, separator melting, or intense reactions between cathode and anode materials, releasing substantial heat and initiating a self-sustaining chain reaction. During the thermal diffusion phase, high-temperature zones quickly transfer to adjacent cells, propagating the失控 state uncontrollably. This process is often accompanied by vigorous gas emission and rapid pressure increases, which further compromise the structural integrity of the battery. Early signs of thermal runaway include abnormal temperature rises, voltage fluctuations, and localized heat accumulation. Without timely intervention, these can escalate into thermal eruptions or even combustion and explosion. Understanding these mechanisms is essential for designing effective预警 systems for China EV battery applications.
Current early warning technologies for EV power batteries rely heavily on sensor deployment, thermal feature monitoring, and advanced algorithms. In sensor布设, thermocouples, thermistors, and infrared temperature sensors are commonly used to monitor cell surfaces and key nodes within battery modules. Some systems integrate voltage, current, and pressure sensors to obtain multi-dimensional data. Sensor placement typically targets high-risk areas, such as thermal conduction interfaces between cells, central regions of the battery pack, and ends of heat dissipation paths. While high-density sensor arrays improve temperature resolution, they can introduce data transmission delays and increase wiring complexity, especially in space-constrained EV power battery systems. Recent advancements include flexible distributed sensor arrays and fiber Bragg grating technology, which enable high-precision distributed temperature monitoring and offer resistance to electromagnetic interference, enhancing stability in high-voltage environments common in China EV battery setups.
Thermal feature monitoring focuses on extracting abnormal signals from temperature variations, employing methods like multi-point temperature difference analysis and temperature rise rate evaluation to identify local thermal anomalies. Disturbances in temperature distribution can indicate散热 failures or internal heating irregularities. Infrared thermal imaging provides high identification accuracy for tracking hotspot expansion paths but faces limitations in vehicle applications due to size and interference issues. Some studies utilize heat flux change modeling to improve the recognition accuracy of thermal source activation behaviors. Multi-variable协同 judgment performs better in high-rate or aging scenarios, enhancing the stability and accuracy of预警 systems for EV power batteries.
In预警算法研究, data-driven approaches have seen significant progress with the support of large-scale monitoring data. Typical methods include support vector machines, convolutional neural networks, and recurrent neural networks, which enable deep feature extraction and trend prediction from multi-dimensional sensor information. Modeling methods that incorporate thermal physical laws attempt to introduce cell heat conduction, thermal radiation, and chemical reaction models to dynamically simulate the evolution of thermal runaway, offering advantages in theoretical rigor and prediction breadth. For instance, a dynamic risk scoring model for a动力电池模组 can be represented as:
$$ R(t) = \int_{0}^{t} \left( \alpha \cdot \frac{dT}{dt} + \beta \cdot \Delta T + \gamma \cdot \frac{d^{2} T}{dt^{2}} \right) e^{-\lambda (t-\tau)} d\tau $$
Here, \( \frac{dT}{dt} \) is the temperature change rate, \( \Delta T \) is the difference between the current cell temperature and the average, \( \frac{d^{2} T}{dt^{2}} \) is the temperature acceleration, \( \alpha \), \( \beta \), and \( \gamma \) are weight coefficients, \( \lambda \) is a time decay factor, \( \tau \) is the historical time variable, and \( t \) is the current time. This model dynamically assesses the risk trends of thermal runaway evolution in different regions, incorporating time decay to weight historical data and increasing sensitivity to abnormal development speeds. Another model based on Bayesian updates for thermal anomaly prediction is constructed as follows:
$$ P(H_i | D_t) = \frac{P(D_t | H_i) P(H_i)}{\sum_{j=1}^{n} P(D_t | H_j) P(H_j)} $$
In this equation, \( H_i \) represents the hypothesis of the i-th thermal runaway state, \( D_t \) is the multi-dimensional sensor observation data at time \( t \), \( P(H_i) \) is the prior thermal risk distribution, \( P(H_i | D_t) \) is the posterior probability of \( H_i \) occurring given the observation data \( D_t \), \( P(D_t | H_i) \) is the observation probability density under a specific state, \( n \) is the total number of thermal runaway states, and \( j \) is the state index variable. This model is suitable for real-time thermal runaway probability updates in uncertain environments, automatically adjusting warning response levels based on data streams to improve the identification capability of early complex thermal behaviors in China EV battery systems.
Protection technologies for EV power batteries involve cell structure optimization, thermal management systems, and safety isolation design. Mainstream cell forms include cylindrical, prismatic, and pouch structures, each with distinct differences in heat diffusion paths, shell strength, and gas release methods. Cylindrical cells, with their metal shells and good thermal symmetry, exhibit superior unit volume heat conduction efficiency but are less effective in横向热阻隔 under dense packing. Prismatic cells offer compact structure and high integration but suffer from concentrated heat sources and uneven internal pressure, which can accelerate local过热扩散. Pouch cells, using aluminum-plastic film shells, have better散热性能 but lack mechanical strength, making them prone to tearing and eruption under thermal runaway conditions. Porous散热通道 designs have been applied in some high-end cells, incorporating micro-scale ventilation layers to enhance longitudinal heat conduction rates and slow横向热扩展 rates. The following figure illustrates the heat diffusion paths of different cell structures under thermal runaway conditions, highlighting the importance of structural design in China EV battery safety.
Thermal management systems play a core role in the thermal safety of EV power batteries, not only maintaining optimal operating temperatures but also responding quickly to early thermal runaway risks. Current mainstream technologies include air cooling, liquid cooling, and phase change cooling. Air cooling is cost-effective but has limited散热能力; liquid cooling offers stronger thermal regulation performance and is widely used in mid-to-high-end vehicles; phase change material cooling absorbs latent heat to delay temperature rises, making it suitable for buffering thermal shocks and often used in combination with active cooling systems. These systems are vital for ensuring the reliability of China EV battery packs under varying operational conditions.
Safety isolation design is a critical component in thermal runaway protection for EV power batteries, focusing on isolating heat and reactions effectively. Common methods involve adding low thermal conductivity, high heat capacity materials between cells or at module boundaries, such as ceramic fiber pads, silicone layers, or inorganic mineral composite layers, to slow heat spread and prevent thermal propagation. Structural designs must balance thermal conductivity control and gas pressure relief, incorporating diversion cavities and exhaust channels to direct high-temperature喷气 away from critical structural areas. Shell isolation using layered structures can improve thermal insulation while maintaining mechanical strength. Fuse bridges or thermal断路 elements can interrupt current in the early stages of失控, curbing the escalation of thermal reactions. Module-level isolation techniques include honeycomb-type insulation layers and removable thermal resistance chambers, effectively achieving independent control of local faults. Combined with thermal management system联动 design, this helps establish a协同机制 between temperature control and insulation, enhancing the overall anti-runaway capability of China EV battery systems.
In engineering applications and case analyses, we examine a mainstream model from a domestic new energy vehicle enterprise in China, which launched a high-performance pure electric SUV in 2022. This vehicle uses a ternary lithium battery module with a nominal voltage of 355 V and a capacity of 85 kW·h, consisting of 24 modules with 12 prismatic cells each. The battery pack features a liquid cooling thermal management system, equipped with a battery management system for temperature monitoring,均热 control, and over-temperature protection. Targeted at urban family users and short-distance business applications, the vehicle offers high power output and fast-charging capabilities, with annual sales exceeding 50,000 units initially. The design incorporated liquid cooling散热 plates and multi-point thermal sensor systems to meet requirements for use in environments above 40 °C.
A fault case回顾 reveals that prior to the incident, the vehicle operated continuously for over 8 hours in high-temperature conditions, with two instances of high-power fast charging, keeping the battery pack temperature above 50 °C for extended periods. Data from the battery management system showed abnormal temperature rises in module 17, with a heating rate exceeding 1.5 °C/min, but temperature sensors were only placed at the module ends, failing to capture the ongoing high-temperature development in the internal central cells. The thermal runaway trigger point was cell 6 in the module center, where high internal resistance led to a rapid temperature surge beyond 120 °C during high-rate discharge, causing intense side reactions, electrolyte decomposition, and gas generation that increased pressure, compressed the separator, and induced an internal short circuit. Heat accumulation eventually breached the module encapsulation, resulting in a high-temperature eruption that ignited adjacent cells, with the thermal reaction spreading quickly to the left-side modules and ultimately causing the battery pack to combust.
In response to the issues identified in this case, the enterprise collaborated with research institutions to systematically optimize thermal safety technologies, upgrading aspects such as sensor system deployment, thermal management path optimization, structural isolation enhancement, and intelligent algorithm improvement. The upgraded system added embedded flexible sensors in the module centers to the original single-point temperature monitoring, enabling real-time capture of high-temperature areas in the central regions. The number of sensor points increased from 48 to 96, doubling spatial thermal perception accuracy. After system upgrades, the enterprise validated 200 vehicles of this model under high-temperature operating conditions and conducted cell-level thermal runaway simulation experiments. The table below compares key thermal safety indicators before and after optimization in typical operating environments.
| Indicator Category | Pre-optimization Performance | Post-optimization Performance | Improvement Rate |
|---|---|---|---|
| Maximum temperature difference within module | 18.6 °C | 6.3 °C | Reduced by 66.1% |
| Peak temperature rise rate | 2.3 °C/min | 0.8 °C/min | Reduced by 65.2% |
| Thermal runaway propagation time | 42 s | >180 s | Extended by 328.6% |
| Local alarm response time | 75 s | 35 s | Advanced by 53.3% |
| Adjacent cell temperature rise during runaway | 24.5 °C | 4.7 °C | Reduced by 80.8% |
Analysis of the table shows that post-optimization, the heat diffusion paths between cells were effectively cut off, with the maximum temperature difference within modules significantly reduced from 18.6 °C to 6.3 °C, indicating more balanced internal temperatures. Local temperature rise rates slowed substantially, dropping to one-third of the original system in simulated single-cell abnormal heating tests. The thermal runaway propagation time increased from 42 seconds to over 180 seconds, providing more than 2 minutes of additional response time for the control system, effectively reducing system-level accident risks. Alarm response times shortened to within 35 seconds, allowing the system to perform active regulation or isolation operations before reaching critical thermal runaway states. These improvements underscore the effectiveness of integrated warning and protection strategies for China EV battery systems.
In conclusion, thermal runaway in EV power batteries remains a significant challenge for enhancing the safety performance of new energy vehicles, particularly in the context of China’s rapidly expanding EV market. Establishing efficient early warning mechanisms and multi-layered protection systems is of paramount importance. This study, starting from mechanism analysis, combines sensing technologies, thermal management strategies, and structural isolation solutions to propose a systematic technical pathway. Case analyses validate the practical value of key measures in engineering applications. Future efforts should continue to advance sensor accuracy, algorithm intelligence, and integrated thermal protection design to support the safe and sustainable development of the China EV battery industry. By focusing on these areas, we can further mitigate the risks associated with thermal runaway and ensure the long-term reliability of EV power battery systems worldwide.