Overcharge-induced thermal runaway represents the most prevalent and hazardous failure mode for lithium-ion batteries. This article establishes a simulation model for the thermal runaway temperature of lithium batteries, systematically analyzing the mechanisms by which key parameters influence battery temperature evolution throughout the entire overcharge thermal runaway process.
Lithium-ion batteries serve as the core energy storage component in fields such as new energy storage and electric vehicles. Their thermal characteristics directly determine service safety, cycle life, and energy output efficiency, with overcharge thermal runaway risk being a key bottleneck limiting the large-scale application of high-capacity batteries. The accuracy of temperature estimation and thermal behavior prediction heavily relies on precise thermal model construction. Current mainstream battery thermal models can be categorized into electrochemical-thermal coupled models and electro-thermal coupled models. The electro-thermal coupled model, due to its balance between computational feasibility and engineering applicability, has become a core technical path for temperature estimation research.
Although significant progress has been made in electro-thermal coupled models, existing research still exhibits limitations. Some studies have implemented battery pack temperature field simulations through electro-thermal coupled models combined with STAR-CCM+ software, but model validation lacks targeted experimental support. The parameters for heat generation and dissipation modules were not calibrated with customized experimental data, making it difficult to match the complex electro-thermal coupling effects in real-world scenarios. Other research has constructed multi-level “cell-module-pack” simulation systems on platforms like MATLAB/Simulink to achieve comprehensive electrical/thermal performance prediction. However, the superposition of multi-scale coupling and real-time state estimation leads to a significant increase in model complexity and computation time, failing to meet the requirements for real-time engineering control.
To address these issues, this study takes a 280Ah lithium iron phosphate battery as the research subject. Overcharge thermal runaway experiments at 0.3C and 1C rates are conducted to obtain dynamic voltage-temperature response data throughout the lifecycle, providing direct experimental support for model construction. Building on this, an electro-thermal coupled simulation model is proposed that balances accuracy, efficiency, and adaptability to extreme conditions. The electrical part employs a first-order RC equivalent circuit model, which significantly reduces computational complexity and time while ensuring accuracy in capturing dynamic voltage behavior. The thermal part innovatively integrates multiple heat generation mechanisms such as joule heat, polarization heat, and entropic heat, combined with convective heat dissipation paths, to accurately simulate the internal microscopic electro-thermal coupling process of the battery. Heat generation parameters are dynamically adjusted according to the State of Charge (SOC) stage, enhancing the authenticity of heat generation prediction across all overcharge stages. All model parameters in this study are calibrated and optimized using experimental data, ensuring high predictive accuracy under both normal charging and extreme conditions like overcharge thermal runaway. This research not only fills the gap in the “experiment-simulation” closed-loop for high-capacity lithium iron phosphate batteries under overcharge conditions but also provides a technical solution with both engineering practicality and scientific rigor for real-time temperature estimation and thermal safety early warning in Battery Management Systems (BMS).
Thermal Runaway Principles and Experimental Setup
Principles of Thermal Runaway
Thermal runaway in a lithium-ion battery is an irreversible process characterized by “heat accumulation-chain reaction-energy爆发.” Its core involves a series of violent reactions of the positive electrode, negative electrode, electrolyte, and separator once the temperature exceeds a safety threshold. The entire process is continuous without obvious external intervals. In the preheating trigger stage (100-200°C), factors like external short circuits, overcharge, mechanical damage, or battery aging causing lithium plating and SEI film instability initiate temperature rise. The SEI film decomposes first, generating compounds like Li2CO3, LiF, hydrocarbons, and gases such as CO2 and C2H4, releasing a small amount of heat. Subsequently, the exposed metallic lithium reacts slowly with the electrolyte, and the separator begins to soften and shrink, further accumulating heat. At this stage, there are no obvious external abnormalities in the battery, but the voltage shows slight fluctuations. In the intense exothermic gas generation stage (200-400°C), the temperature突破 the decomposition threshold of the positive electrode material. The deeply delithiated positive electrode material decomposes, releasing a large amount of heat causing a sharp temperature rise. The electrolyte undergoes chain decomposition simultaneously, producing flammable and explosive small molecules like methane, ethane, ethylene, as well as CO and CO2. The amount of gas increases exponentially, leading to a sharp rise in internal battery pressure. During the runaway propagation stage (400-600°C), high temperatures trigger violent redox reactions between the graphite negative electrode and the electrolyte, as well as decomposition products from the positive electrode. The separator completely melts, causing an internal short circuit. The internal pressure突破 the shell’s limit, leading to rupture. High-temperature gases and electrolyte喷射 out. Oxidizing substances produced from positive electrode decomposition react violently with the negative electrode,加剧 heat release. Some products react with residual moisture to生成 highly toxic gases. In the final爆发 stage (above 600°C), the喷射 high-temperature electrolyte ignites upon contact with air. Oxygen released from positive electrode decomposition and reaction heat sustain intense combustion, accompanied by flames, thick smoke, and the emission of toxic gases, completing the entire thermal runaway process.
Experimental Design
To investigate the electro-thermal coupling characteristics and the evolution of thermal runaway in lithium-ion batteries under overcharge conditions, constant current overcharge experiments were designed and conducted. The experimental platform was built around a 280Ah lithium iron phosphate cell. The battery was subjected to a 1C constant current overcharge until thermal runaway occurred. The simulation model was constructed based on this experiment.
The experimental battery was固定 in a battery fixture to prevent expansion or deformation during the experiment. The experimental site was set up, and all measurement equipment was connected. This included K-type thermocouples placed at various positions: on the large surface of the battery (1), on the battery positive terminal (2), to measure air flow temperature (3), near the QR code area (4), and under the large surface of the battery (5). A multi-channel data logger, charge/discharge tester power cables, insulating pads, etc., were also connected. All measurement equipment was debugged and synchronized in time. The multi-channel data logger was then started to collect and record temperature data from various positions on the battery surface. The experimental plan for lithium-ion battery overcharge thermal runaway is summarized in the table below:
| Experiment | Experiment Type | Charging Current | Experimental Purpose |
|---|---|---|---|
| Experiment 1 | Constant Current Overcharge | 0.3C | Comparative analysis of different current magnitudes |
| Experiment 2 | Constant Current Overcharge | 1C | – |
After all preparatory work was completed, the charge/discharge tester was activated to perform constant current overcharge on the battery. Two separate constant current overcharge experiments were conducted on lithium-ion battery samples using different charging currents. Key data such as battery voltage in the test program were monitored simultaneously until the battery completely underwent thermal runaway, after which heating was stopped. The setup was left to stand for over an hour to ensure no safety risks, then the recorded data was collected and saved.
The voltage and temperature data curves from the two experiments are shown in the provided figures. Taking the 1C overcharge thermal runaway temperature data as an example, the results show a clear thermal gradient across different positions of the battery. The lowest temperature, recorded underneath the large surface, reached a peak of 592.48°C during thermal runaway, with fluctuations occurring during the venting process. The temperature on the upper large surface peaked at 532.31°C, while the temperature recorded in the QR code area peaked at approximately 292.62°C. The air flow temperature increased rapidly during venting and thermal runaway, reaching a peak of 448.72°C. The 0.3C overcharge thermal runaway temperature data showed: the upper large surface temperature peaked at 541.80°C, and the temperature underneath the large surface peaked at 525.14°C during thermal runaway. Notably, as thermal runaway approached, the temperature rise rate underneath the large surface significantly accelerated and further intensified during the runaway stage, exhibiting typical thermal runaway acceleration characteristics.
Comparing the temperature data from lithium-ion battery overcharge thermal runaway at two different charging rates, it was found that their temperature change trends were consistent. However, a lower rate overcharge took longer to reach thermal runaway, and the maximum battery temperature was slightly lower.
Model Construction and Parameter Setting
Simulation Construction Process
Based on the two overcharge thermal runaway experiments, a simulation model was constructed. The model built in Simulink is a circuit model. The flowchart for building the simulation model is illustrated in the provided figure.
The overall electro-thermal coupled model for lithium-ion battery overcharge thermal runaway built in Simulink is shown in the provided figure. The electrical part of the simulation model employs the Thevenin equivalent circuit model, known for its simple structure, high computational efficiency, and ease of integration with experimental data. Based on the mechanisms of electrochemical polarization and concentration polarization, it uses an equivalent circuit composed of an ideal voltage source, an ohmic resistor, and a parallel RC network to accurately capture the dynamic changes of voltage with time and current during charge/discharge cycles, forming the basis for the electro-thermal coupled model. Based on Kirchhoff’s laws, the state equations for the first-order RC equivalent circuit model are as follows:
$$ U_{R0} = I \times R_0 $$
Where \( R_0 \) is the battery’s internal resistance.
$$ U_1 = I(t) – \frac{1}{C_1} \int_0^t I \, dt $$ (Note: The original text seems to have a formatting issue here. The correct formula for voltage across the RC branch is typically derived from \( U_1 = I R_1 (1 – e^{-t/(R_1 C_1)}) \) or its differential form. The integral form might be representing charge, but the equation as written is dimensionally inconsistent. A more standard form for the state of the polarization voltage is: \( \dot{U}_1 = -\frac{1}{R_1 C_1} U_1 + \frac{1}{C_1} I \) )
Let’s represent the polarization voltage dynamics correctly. The voltage across the parallel RC branch (U1) is governed by:
$$ \frac{dU_1}{dt} = -\frac{U_1}{R_1 C_1} + \frac{I}{C_1} $$
Where \( R_1 \) is the battery’s polarization resistance and \( C_1 \) is the battery’s polarization capacitance.
The terminal voltage \( U_t \) is given by:
$$ U_t = U_{oc} – I R_0 – U_1 $$
Where \( U_{oc} \) is the battery’s open-circuit voltage, \( I \) represents the battery’s operating current, and \( U_t \) represents the measurable terminal voltage.
Battery Parameter Setting
Based on cycle charge/discharge experiment results and electrochemical impedance test results, parameters such as ohmic internal resistance, polarization internal resistance, and polarization capacitance were analyzed and calculated, laying the foundation for subsequent parameter calculations. The battery capacity selected for the simulation model is 280Ah. The charging current magnitude is determined in conjunction with the battery’s rated capacity. Referring to the experimental conditions, a 1C charging current magnitude is chosen for the simulation, meaning the charging current is 280A. Referring to the experimental temperature, the battery’s initial temperature is set to 15°C.
The setting of the battery’s heat generation module is based on experimental data, comprehensively considering the heat generation principles throughout the entire charging process and the changes in battery heat generation at different stages. The heat generation data obtained from experiments is imported into the simulation model as a reference to simulate the heat generation situation of the battery throughout the entire overcharge process.
Comprehensively considering the structural parameters of the battery itself, the parameters are as shown in the table below:
| Parameter | Unit | Positive Electrode | Separator | Negative Electrode |
|---|---|---|---|---|
| Thermal Conductivity | W/(m·K) | 1.48 | 0.334 | 1.04 |
| Density | kg/m³ | 1500 | 492 | 2660 |
| Particle Radius | μm | 1.15 | – | 10 |
| Convective Heat Transfer Coefficient | W/(m²·K) | 7.17 | ||
| In-plane Thermal Conductivity | W/(m·K) | 38.746 | ||
| Through-plane Thermal Conductivity | W/(m·K) | 1.0958 | ||
And the environmental conditions during the overcharge experiment are used to set the relevant heat dissipation data for the battery in the simulation model, as shown in the table below:
| Parameter | Value | Unit |
|---|---|---|
| Battery Temperature | 15 | °C |
| Ambient Temperature | 15 | °C |
| Mass | 0.5 | kg |
| Heat Dissipation Area | 4 | m² |
| Specific Heat Capacity | 1.4 | J/(kg·K) |
Calculation of Battery External Characteristic Parameters
Based on the input conditions for the lithium-ion battery model: data such as current and temperature, the voltage magnitude of the lithium-ion battery and its State of Charge (SOC) value during charging are calculated. The battery voltage is calculated from the input current magnitude and the RC module inherent to the battery model itself. The battery’s SOC is calculated using the Ampere-hour integral method.
Calculation of SOC for the lithium-ion battery: Since the charging current magnitude is known in the simulation model, the battery’s SOC during charging is calculated using the Ampere-hour integral method:
$$ SOC_t = SOC_0 – \frac{1}{3600 \times C_{batt}} \int_0^t I_t \, dt $$
The overcharge process in the simulation model is implemented by modifying the integration limits in the SOC calculation formula, allowing the SOC to exceed 1 during the overcharge stage and introducing an additional heat source from side reactions when the SOC exceeds a certain threshold.
Battery Heat Generation and Dissipation Modules
Battery Heat Generation Model
Under normal operating conditions, the main sources of heat generation in a battery are joule heat caused by ohmic internal resistance, polarization heat, and entropic heat.
Joule Heat (\(Q_j\)): Primarily occurs during the charge and discharge of a lithium-ion battery. Ohmic heat is generated when current flows through the battery’s internal materials (such as the separator and electrodes). Joule heat is calculated as:
$$ Q_j = I_t^2 R_0 $$
Where \( I_t \) is the current and \( R_0 \) is the ohmic resistance.
Polarization Heat (\(Q_1\)): Arises from the deviation between the electrode potential and its equilibrium potential during electrochemical reactions. This manifests as the consumption of part of the chemical energy into released heat during the thermal diffusion movement of ions. The calculation formula for polarization heat is:
$$ Q_1 = I_t^2 R_1 $$
Where \( R_1 \) represents the equivalent polarization resistance.
Entropic Heat (\(Q_a\)): Primarily the heat generated by entropy changes in electrochemical reactions. Heat released from side reactions due to battery aging is slow and can be neglected. The formula for entropic heat is:
$$ Q_a = I T \frac{\partial U_{oc}}{\partial T} $$
Where \( \frac{\partial U_{oc}}{\partial T} \) is the entropy coefficient, which depends entirely on the State of Charge (SOC) and is independent of the battery’s operating temperature at that SOC.
In summary, combining the heat generation formulas gives the heat generated per unit time by a single lithium-ion battery cell:
$$ Q = Q_j + Q_1 + Q_a = I_t^2 (R_0 + R_1) + I T \frac{\partial U_{oc}}{\partial T} $$
Under fault conditions, additional side reactions occur inside the battery, altering its heat generation mechanism. Corresponding to different SOC stages, there are segmented fault heat generation modules.
When the battery SOC ≤ 0.8 and remains within the normal range, the battery is in the normal charging stage with normal heat generation. When the battery SOC > 0.8, the difference between terminal voltage and open-circuit voltage increases, causing part of the input electrical energy to be unable to fully convert into chemical energy and be released as heat. When the SOC exceeds 1.1, the battery is in a severe overcharge state. The energy charged into the battery cannot be stored in the electrode materials and is rapidly released as heat in a short time through side reactions. This process is often accompanied by irreversible reactions such as electrolyte decomposition, electrode lithium plating, and SEI layer damage.
Battery Heat Dissipation Module
The heat generated during the operation of a lithium-ion battery needs to be released to the external environment through three methods: heat conduction, heat convection, and heat radiation. Among these, heat convection is the process of heat transfer facilitated by the relative displacement of fluid particles, which relies on both medium flow and involves conductive phenomena. Specifically, the heat generated by the battery is first transferred through the casing to the surrounding cooling medium and then dissipated via fluid motion. This process is based on Newton’s law of cooling. Relevant parameters are fitted using experimental data, and a temperature-triggered mechanism is introduced to dynamically adjust parameters, ultimately enhancing the model’s adaptability to different operating and fault conditions.
Based on experimental temperatures, the ambient temperature in the simulation is set to 15°C, the battery heat dissipation area to 4 m², the battery mass to 0.5 kg, and the battery’s specific heat capacity to 1.4 J/(kg·K).
$$ q_b = h (T_i – T_j) $$
Where \( q_b \) represents the convective heat transfer rate per unit area per unit time between the battery surface and the fluid; \( T_i \) represents the instantaneous temperature of the battery, \( T_j \) represents the ambient temperature, and \( h \) is the convective heat transfer coefficient.
$$ h = \frac{\lambda}{l} Nu = \frac{\lambda}{l} \times \begin{cases} 0.446 \, Re^{0.5} Pr^{0.4} \\ 0.023 \, Re^{0.8} Pr^{0.4} \end{cases} $$
Where \( Nu \) represents the Nusselt number; \( Re \) represents the Reynolds number; \( Pr \) represents the Prandtl number; \( \lambda \) is the thermal conductivity of the coolant in W/(m·K); \( l \) represents the characteristic length (e.g., diameter of cooling pipe) in m. The choice between the two equations typically depends on the flow regime (laminar or turbulent) and geometry.
According to fluid mechanics principles, the magnitude of the Reynolds number is generally used to determine whether the coolant flow is laminar or turbulent. In engineering applications, a Reynolds number of 2300 is usually regarded as the critical threshold: exceeding 2300 indicates turbulent coolant flow; otherwise, it is laminar.
$$ Re = \frac{\rho v l}{\mu} $$
Where \( \rho \) represents the density of the coolant in kg/m³; \( v \) represents the fluid velocity in m/s; \( \mu \) represents the dynamic viscosity of the fluid in Pa·s.
$$ Pr = \frac{\mu C_p}{\lambda} = \frac{\nu}{\alpha} $$
Where \( \mu \) represents viscosity; \( C_p \) represents specific heat capacity in J/(kg·K); \( \lambda \) represents thermal conductivity; \( \alpha \) is thermal diffusivity in m²/s; \( \nu \) represents kinematic viscosity.
There is a significant difference in heat transfer intensity between natural convection and forced convection. When the battery surface temperature is in a low range, heat dissipation is dominated by natural convection, and the heat transfer coefficient remains at a low level. Once the battery temperature exceeds a specific threshold, the temperature difference with the environment increases, air convection effects随之 enhance, thereby enabling dynamic correction of heat dissipation intensity.
In summary, based on experimental data and practical considerations, the electro-thermal coupled model for a single battery cell is derived.
$$ C M \Delta T = Q_g – Q_s = I_t^2 (R_0 + R_1) + I T \frac{\partial U_{oc}}{\partial T} – h (T_i – T_j) $$
Where \( Q_g \) represents heat generation, \( Q_s \) represents heat dissipation. Temperature changes can trigger dynamic parameter adjustments within the battery. By coupling the equivalent circuit model with the heat generation and dissipation models, the electro-thermal coupling characteristics of the battery under different operating conditions can be more accurately reflected.
Experimental Verification and Result Analysis
Simulation Result Comparative Analysis
The provided figures showcase the comparison between experimentally measured voltage and temperature data and the simulation model’s voltage and temperature data for a battery under 1C overcharge conditions, along with corresponding error analyses between simulation and experiment.
Analyzing the 1C overcharge thermal runaway voltage results: In the initial charging phase (t < 1500 s), both experimental and simulated voltages rise slowly, with the simulation results highly consistent with the experiment, indicating the model accurately describes the electrochemical behavior in this stage. As charging continues (1500–2500 s), the experimental voltage gradually enters a plateau, and the simulation effectively describes this voltage plateau feature, maintaining alignment with the experimental data. At the end of charging (t ≈ 2500 s), the experimental voltage surges to the cutoff voltage. The simulation model’s cutoff is 26V, while the experimental value is 30.27V, a difference less than 5 V. This proves the simulation model accurately describes the voltage response in the final overcharge stage. In the error analysis plot, the voltage error curve shows a distinct peak. This peak exists because the voltage predicted by the simulation lags slightly behind the experimental value in time, but the predicted voltage peak is relatively accurate.
Analyzing the 1C overcharge thermal runaway temperature results: In the initial charging stage (t < 1500 s), both experimental and simulated temperatures show an upward trend, with the simulation data slightly lower than the experimental data. As charging continues (1500–2500 s), the experimental temperature rise amplitude is lower than the simulation data, but the simulation model accurately captures this拐 point characteristic. Upon entering the thermal runaway stage (t ≈ 2500 s), the experimental temperature surges to 532.31°C, while the simulation’s maximum temperature is 520.83°C. The difference is less than 12°C, with an error of 2.1%, which is within a reasonable error range. This indicates the simulation model can accurately describe the intense heat generation phenomenon during the process of overcharge-induced thermal runaway. In the temperature decay stage, the experimental temperature decreases slowly, while the simulated temperature drops rapidly, resulting in lower temperatures later on. This discrepancy is primarily because the simplified heat dissipation model does not fully account for residual reactions and structural heat retention effects after thermal runaway. In the temperature error analysis plot, a peak is also visible. However, this peak actually highlights an advantage of this temperature anomaly simulation method. It can not only accurately simulate the temperature peak at the moment of overcharge thermal runaway anomaly but also do so slightly earlier in time than the experimental value. This simulation method offers a temporal advantage for detecting and preventing battery thermal runaway.
Comparing the maximum temperature values from different experimental locations with the maximum temperature from the simulation shows that the simulation temperature is closest to the temperature on the upper large surface of the lithium-ion battery, where the prediction effect is best. This aligns with the comparison result shown earlier for 1C overcharge thermal runaway temperature. The comparison between temperatures at different locations and the simulation temperature is illustrated in the provided bar chart.
To validate the effectiveness of the electro-thermal coupled model, the simulation results for voltage and temperature were compared with experimental values. Not only was the error between experimental and simulated values calculated, but error evaluation metrics were also computed. The coefficient of determination R² measures the degree to which the model explains the variability in the actual data through statistical methods, with a range of [0,1]. A value of R² closer to 1 indicates a higher degree of fit between the model and the experimental data. The comparative data analysis is shown in the table below:
| Current | Parameter | Maximum Value | Error | R² |
|---|---|---|---|---|
| 1C | Voltage / V | Experiment: 30.27 Simulation: 26.07 |
4.20 | 0.758 |
| Temperature / °C | Experiment: 532.31 Simulation: 520.83 |
11.48 | 0.880 |
Regarding voltage, the simulation model effectively reproduces the overall trend of the experimental voltage curve, including the平稳 rise at the start of charging, the gradual change in the middle period, and the steep growth at the end. The coefficient of determination R² is 0.758, indicating the model explains about 76% of the variation trend in the voltage changes, with fitting accuracy at an acceptable level.
Regarding temperature, the model effectively captures the overall battery temperature rise trend and the thermal runaway拐点. The simulation curve closely matches the experimental situation at the position of the temperature peak. The coefficient of determination R² reaches 0.880, indicating the model explains about 88% of the temperature change trend. The peak temperature difference is less than 12°C, meeting the precision requirements for engineering analysis and early warning.
Extended Simulation Result Analysis
Based on the accuracy of the 1C simulation results, extended simulations were conducted. The provided figure presents the comparative results of the simulation model’s temperature and voltage data for battery overcharge at 0.5C and 2C rates.
Comparative analysis results: Under low-rate 0.5C overcharge, the irreversible energy loss caused by current is minimal, heat diffusion is relatively sufficient, temperature rise is relatively gentle, and voltage growth remains stable. Initial charging stage: The voltage curve remains generally stable, while the temperature remains基本 stable around 25°C, indicating weak thermal effects at low rates. Middle charging stage: The voltage increases slowly in an approximately linear manner, temperature rises slightly, but the rate of temperature increase remains low. Late charging stage: The voltage rise accelerates significantly, and the temperature also increases substantially, forming a curve拐点. After the battery overcharge leads to thermal runaway, both temperature and voltage show a rapid衰减 trend.
Under high-rate 2C overcharge, internal resistance losses and polarization effects are significantly amplified, leading to明显 accelerated temperature rise and a more pronounced tendency for thermal runaway. Initial stage: Voltage increases rapidly, with a noticeably faster rate of change compared to the 0.5C condition, while temperature shows no significant fluctuation. As charging continues: The voltage curve rises陡峭, and concurrently, the temperature also exhibits rapid攀升 characteristics. Finally, upon entering the decay stage: Both voltage and temperature迅速回落.
Compared with 1C rate overcharge thermal runaway, the change patterns of temperature and voltage under these three charging rates are highly similar. Temperature rises slowly in the initial stage, accelerates from the middle to late stages as SOC increases, and exhibits a sharp jump followed by decay during the overcharge thermal runaway stage. The differences in temperature changes among the three rates are mainly manifested in time and rate. A larger charging rate causes thermal runaway to occur earlier. The consistent pattern of curve changes under the three charging rates indicates that the model captures the essential physical mechanisms applicable to different charging rates.
Conclusion
Based on overcharge experiments conducted on a 280Ah lithium iron phosphate battery at 0.3C and 1C rates, this paper proposed and established a simulation model using a first-order RC equivalent circuit coupled with multi-condition heat generation to characterize the temperature behavior of a lithium-ion battery throughout the entire process of constant current overcharge leading to thermal runaway.
(1) An electro-thermal coupled model for normal charging and overcharge was constructed. Under 1C overcharge conditions, the comparison between the simulation model (peak temperature 520.83°C) and experimental results (peak temperature 532.31°C) showed an error of approximately 12 degrees Celsius, with an R² of 0.88. This demonstrates its ability to accurately simulate battery temperature changes under both normal and fault conditions.
(2) By conducting extended simulations under 0.5C and 2C overcharge, it was found that the patterns of result changes remained highly consistent. This can provide model support for the development of early thermal runaway fault warning strategies in battery management systems and the optimization of thermal management structures.