As lithium ion batteries (LIBs) increasingly power electric vehicles and energy storage systems, their thermal safety remains a critical concern. Thermal runaway (TR), a chain reaction of exothermic processes triggered by local overheating, poses significant risks of catastrophic failure. This study investigates the TR characteristics of LIBs under localized overheating at battery terminals, compares the impacts of different heating positions, and proposes an optimized liquid cooling strategy to mitigate thermal hazards.
1. Thermal Runaway Mechanism in Lithium Ion Batteries
LIBs generate heat during operation due to internal resistance and electrochemical reactions. Under normal conditions, this heat dissipates efficiently. However, localized overheating—caused by loose connectors, high-rate charging/discharging, or mechanical damage—can disrupt thermal equilibrium, initiating TR. Key exothermic reactions driving TR include:
- SEI decomposition:Rsei=Aseiexp(−Ea,seiRT)Rsei=Aseiexp(−RTEa,sei)
- Anode-electrolyte reaction:Rne=Aneexp(−Ea,neRT)Rne=Aneexp(−RTEa,ne)
- Cathode-electrolyte reaction:Rpe=Apeexp(−Ea,peRT)Rpe=Apeexp(−RTEa,pe)
- Electrolyte decomposition:Re=Aeexp(−Ea,eRT)Re=Aeexp(−RTEa,e)
The total heat generation (QtotQtot) is the sum of contributions from these reactions:Qtot=∑HiWiRi(i=sei, ne, pe, e)Qtot=∑HiWiRi(i=sei, ne, pe, e)
where HiHi is the enthalpy, WiWi is the mass fraction, and RiRi is the reaction rate.
2. Modeling and Validation
A 3D lumped thermal model was developed to simulate TR propagation in a 50 Ah LiFePO44 battery. Key parameters are listed in Table 1.
Table 1: Material properties of lithium ion battery components
| Component | Density (kg/m³) | Specific Heat (J/kg·K) | Thermal Conductivity (W/m·K) |
|---|---|---|---|
| Electrode | 3846 | 1100 | (1, 18, 18) |
| Positive Tab | 2700 | 900 | 238 |
| Negative Tab | 8960 | 385 | 400 |
| Liquid Cooling Plate | 2700 | 900 | 238 |
| Coolant (Water) | 998.2 | 4182 | 0.6 |
The energy conservation equation governs heat transfer:ρCp∂T∂t=∇⋅(k∇T)+QtotρCp∂t∂T=∇⋅(k∇T)+Qtot
where ρρ, CpCp, and kk denote density, specific heat, and thermal conductivity, respectively.
Model validation against experimental data (Table 2) confirmed accuracy, with TR trigger time errors below 5%.
Table 2: Validation of simulated TR trigger times
| Heating Area (cm²) | Experimental TR Time (s) | Simulated TR Time (s) | Error (%) |
|---|---|---|---|
| 54 | 1378.3 | 1419.5 | 5.3 |
| 81 | 1339.2 | 1404.4 | 7.4 |
| 108 | 1662.0 | 1396.0 | 19.3 |
3. Thermal Runaway Characteristics Under Terminal Overheating
Localized heating (225 W) was applied to the terminal, bottom, and front surfaces to compare TR behaviors.
3.1 Impact of Heating Power
Increasing heating power accelerated TR initiation and elevated peak temperatures (Table 3). At 225 W, TR triggered 3891.5 s earlier than at 100 W, with a 27.9°C higher peak temperature.
Table 3: TR characteristics under varying heating powers
| Parameter | 100 W | 156 W | 225 W |
|---|---|---|---|
| Self-heating Temp (°C) | 113.9 | 99.8 | 104.1 |
| TR Trigger Temp (°C) | 239.3 | 229.0 | 218.3 |
| Peak Temp (°C) | 696.9 | 721.5 | 724.8 |
| TR Trigger Time (s) | 5120.5 | 2111.0 | 1299.0 |
3.2 Impact of Heating Position
Terminal overheating caused severe thermal accumulation near tabs due to limited heat dissipation paths. Key findings include:
- Higher peak temperatures: Terminal heating resulted in 705.5°C vs. 695.1°C (bottom) and 698.3°C (front).
- Faster temperature rise: TR duration reduced by 17.2% (vs. bottom) and 10.9% (vs. front).
- Poor temperature uniformity: Standard deviation (TσTσ) was 31.8 for terminal heating, compared to 38.4 (bottom) and 35.7 (front).
4. Thermal Management Using Oblique-Channel Liquid Cooling
To address localized overheating, an oblique-channel liquid cooling plate (LCP) was designed (Figure 1). Compared to straight-channel LCPs, oblique channels improved heat dissipation by aligning with temperature gradients.
Key performance metrics:
- Temperature reduction: Oblique LCP lowered average battery temperature by 9.0% vs. straight LCP.
- Uniformity improvement: Temperature standard deviation decreased by 18.9%.
- Optimal tilt angle: A 27.5° channel angle minimized TσTσ and peak temperature.
The cooling efficiency was modeled using fluid dynamics equations:∂∂t(ρwv⃗)+∇⋅(ρwv⃗v⃗)=−∇P+μ∇2v⃗∂t∂(ρwv)+∇⋅(ρwvv)=−∇P+μ∇2v∂ρw∂t+∇⋅(ρwv⃗)=0∂t∂ρw+∇⋅(ρwv)=0∂∂t(ρwCwTw)+∇⋅(ρwCwv⃗Tw)=∇⋅(kw∇Tw)∂t∂(ρwCwTw)+∇⋅(ρwCwvTw)=∇⋅(kw∇Tw)
where ρwρw, v⃗v, PP, and μμ represent coolant density, velocity, pressure, and viscosity.
5. Discussion and Implications
- Terminal overheating: Small contact areas and high power density exacerbate thermal accumulation, necessitating proactive thermal management.
- Cooling plate design: Oblique-channel LCPs enhance heat extraction by leveraging directional heat flow, critical for high-rate applications.
- Safety protocols: Real-time monitoring of terminal resistance and temperature gradients can preempt TR.
6. Conclusion
This work demonstrates that terminal overheating in lithium ion batteries significantly accelerates TR severity compared to bottom or front heating. The proposed oblique-channel liquid cooling plate effectively mitigates thermal accumulation, reducing peak temperatures by 9.0% and improving uniformity by 18.9%. These insights advance LIB safety in high-power applications, emphasizing the need for targeted thermal management strategies.