In the past decade, the rapid growth of electric unmanned aerial vehicles (UAVs) has created new requirements for compact, lightweight and high-rate propulsion batteries. The core component of such a propulsion system is often a single high-power pouch cell or a small module that must deliver traction current for only a few minutes before being recharged. In this study, I refer to this complete energy-delivery system as a traction battery pack. The traction battery pack investigated here is based on a LiCoO₂/graphite pouch cell, with a nominal capacity of 2.25 A h, a nominal voltage of 3.85 V and a maximum continuous discharge current of 10 C. This type of traction battery pack is normally used in consumer drones, which require a very high instantaneous discharge rate but have limited cooling area. The consequence is that the traction battery pack often experiences fast aging and a non-negligible risk of thermal runaway. I therefore combined experiments and multi-physics simulation to understand how the traction battery pack heats up and degrades under realistic UAV operating conditions.
1. Battery and Experimental Setup
The commercial pouch cell used in the traction battery pack has external dimensions of 7.8 mm × 61 mm × 34 mm and a mass of 33.7 g. Its positive electrode is LiCoO₂, the negative electrode is graphite, and the electrolyte is LiPF₆ dissolved in ethylene carbonate and dimethyl carbonate. The cell is soft-packed with aluminum laminated film. Table 1 lists the main characteristics declared by the manufacturer and used during the experiments.
| Property | Value |
|---|---|
| Nominal capacity | 2.25 A h at 0.5 C |
| Nominal voltage | 3.85 V |
| Charge cutoff voltage | 4.4 V |
| Discharge cutoff voltage | 3.0 V |
| Maximum charge current | 1 C (2.25 A) |
| Maximum discharge current | 10 C (22.5 A) |
| Mass | 33.7 g |
| Initial DC internal resistance | 4.8 mΩ |
| Discharge ambient temperature range | −20 °C to 60 °C |
I built a test bench composed of a battery cycler, a programmable temperature and humidity chamber, a multi-channel thermometer, T-type thermocouples and a data acquisition system. All tests were performed in a single-cell configuration, but the measured voltage, current and temperature responses are representative of the behavior of the larger traction battery pack under identical current densities.
The experiments included discharge tests at different C-rates, discharge tests at different ambient temperatures, HPPC direct-current internal resistance measurements, entropic heat coefficient measurements, and accelerated aging tests. Before each characterization test, the capacity was measured at 1 C between 4.4 V and 3.0 V. The charging protocol was constant-current constant-voltage (CC-CV): 1 C current to 4.4 V, followed by 4.4 V constant-voltage until the current decreased below 0.05 C.
2. Experimental Thermal and Electrical Characteristics
2.1 Effect of Discharge Rate and Ambient Temperature
Figure 2 of the original test campaign shows the discharge voltage curves at 1 C, 2 C, 3 C and 4 C in a 25 °C environment. As the discharge rate increased, the initial voltage decreased and the delivered capacity decreased. Table 2 summarizes representative temperatures measured on the cell surface. The 1 C discharge ended at about 29.2 °C, whereas the 4 C discharge ended at about 42.3 °C. This temperature rise was caused not only by the larger current, but also by the higher overpotential and ohmic loss at high C-rates.
| Discharge rate | Ambient temperature (°C) | Observed trend |
|---|---|---|
| 1 C | 25 | Small temperature rise; end temperature 29.2 °C |
| 2 C | 25 | Higher initial voltage drop; faster temperature rise |
| 3 C | 25 | Further increase of temperature slope |
| 4 C | 25 | End temperature 42.3 °C; clear “N” shape in the early temperature curve |
When the ambient temperature was varied while the discharge current was fixed at 1 C, the discharge capacity changed significantly. At 40 °C the measured discharge capacity was 2280 mAh, whereas at −10 °C the discharge capacity was only 2179 mAh. At 25 °C the measured 1 C capacity was 2250 mAh. This observation indicates that the capacity of the traction battery pack is more sensitive to ambient temperature than to discharge rate in the range considered. It also demonstrates that lithium-ion transport in the electrolyte and through the solid-electrolyte interface is strongly slowed at low temperature.
2.2 DC Internal Resistance
The direct-current internal resistance was measured using the HPPC method. I applied a 10 s discharge pulse after a rest period. The 10 s resistance was calculated as
$$\mathrm{DCR}_{10s} \;=\; \frac{U_{\text{rest}} – U_{\text{10s pulse}}}{I_{\text{pulse}}},$$
where \(U_{\text{rest}}\) is the cell voltage after the rest period before the pulse, \(U_{\text{10s pulse}}\) is the voltage at the end of the 10 s discharge pulse, and \(I_{\text{pulse}}\) is the pulse current. The measured direct-current resistance increased as the ambient temperature decreased. In the SOC range from about 20% to 80%, the resistance was relatively flat at moderate temperatures, while the low-temperature curve showed a stronger SOC dependence. This behavior is relevant for thermal modeling because the internal resistance directly contributes to the irreversible heat generation rate of the traction battery pack.
2.3 Entropic Heat Coefficient
The reversible heat of a lithium-ion traction battery pack depends on the entropic heat coefficient, which is the derivative of the open-circuit voltage with respect to temperature. I measured the full-cell entropic heat coefficient by placing the cell in a temperature chamber and recording the open-circuit voltage after long rests at −5 °C, 5 °C, 15 °C, 25 °C, 35 °C and 45 °C for each state of charge. The full-cell coefficient was evaluated as the slope of voltage versus temperature.
$$\frac{\mathrm{d}U_{\text{cell}}}{\mathrm{d}T}\;=\;\frac{U_{\text{cell}}\left(T_{1}\right)-U_{\text{cell}}\left(T_{2}\right)}{T_{1}-T_{2}}.$$
The measured coefficient changed with discharge depth and ranged from about −1.1 mV K⁻¹ to +0.29 mV K⁻¹. In the range 0% to 35% depth of discharge (DOD), the coefficient was negative, corresponding to exothermic reversible heat during discharge. Between 35% and 50% DOD, the coefficient became positive and the reversible heat changed sign; this caused the hump or cooling interval in the surface temperature curve. Beyond 50% DOD, the entropic heat coefficient became negative again and decreased quickly, resulting in a rapid temperature increase near the end of discharge. This reversible heat behavior is one of the main reasons for the “N” shape observed in the measured temperature rise curves.
3. Electrochemical-Thermal Coupling Model
To analyze the internal heat-generation distribution and the aging behavior of the traction battery pack, I developed a physics-based model coupling a pseudo-two-dimensional electrochemical model with a three-dimensional thermal model. The electrochemical model describes lithium diffusion in active particles, lithium transport in the electrolyte, charge conservation in both phases, and interfacial charge-transfer kinetics. The thermal model solves the heat conduction equation in the pouch cell and accounts for anisotropic heat conduction.
3.1 Electrochemical Model
The active material particles are assumed to be spherical and uniformly distributed. Solid-phase lithium diffusion inside a particle is governed by Fick’s law:
$$\frac{\partial c_{s,i}}{\partial t} \;=\; \frac{D_{s,i}}{r^{2}}\,\frac{\partial}{\partial r}\left(r^{2}\,\frac{\partial c_{s,i}}{\partial r}\right), \qquad i=\mathrm{pos},\,\mathrm{neg},$$
where \(c_{s,i}\) is the solid-phase lithium concentration, \(D_{s,i}\) is the solid diffusion coefficient, and \(r\) is the radial coordinate. The boundary condition at the particle surface is written as
$$-D_{s,i}\,\frac{\partial c_{s,i}}{\partial r}\Bigg|_{r=R_{p,i}} \;=\; J_{i},$$
where \(J_{i}\) is the pore-wall flux of lithium ions. The solid diffusion coefficient is temperature dependent and follows an Arrhenius law:
$$D_{s,i}\left(T\right) \;=\; D_{s,i,\mathrm{ref}}\,\exp\!\left[\frac{E_{\mathrm{act},i}}{R}\,\left(\frac{1}{T_{\mathrm{ref}}}-\frac{1}{T}\right)\right].$$
Electrolyte diffusion in the porous electrodes and separator is described by
$$\varepsilon_{e}\,\frac{\partial c_{e}}{\partial t} \;=\; \frac{\partial}{\partial x}\left(D_{e}^{\mathrm{eff}}\,\frac{\partial c_{e}}{\partial x}\right) \;+\; \frac{1-t_{+}^{0}}{F}\, j_{\mathrm{Li}},$$
with the effective electrolyte diffusivity \(D_e^{\mathrm{eff}} = D_e \varepsilon_e^{\mathrm{Brug}}\). The interfacial reaction is described by the Butler-Volmer equation:
$$j_{\mathrm{Li}} \;=\; a_{v}\, i_{0}\, \left[\exp\!\left(\frac{\alpha_{a} F \eta}{R T}\right) \;-\; \exp\!\left(-\frac{\alpha_{c} F \eta}{R T}\right)\right],$$
where \(a_v\) is the specific surface area of the porous electrode, \(i_0\) is the exchange current density, \(\alpha_a\) and \(\alpha_c\) are transfer coefficients, and \(\eta\) is the local surface overpotential. The exchange current density is calculated as
$$i_{0} \;=\; F k_{i}\, \sqrt{c_{e}\, c_{s,\mathrm{surf}}\, \left(c_{s,\mathrm{max}}-c_{s,\mathrm{surf}}\right)}.$$
The local overpotential is related to the solid-phase potential \(\phi_s\), the electrolyte-phase potential \(\phi_e\), the equilibrium potential \(U_{\mathrm{eq}}\), and the SEI film resistance \(R_{\mathrm{film}}\) on the negative electrode:
$$\eta \;=\; \phi_{s} – \phi_{e} – U_{\mathrm{eq}} – j_{\mathrm{Li}} R_{\mathrm{film}},$$
with \(j_{\mathrm{Li}} = a_v J\).
3.2 Heat Source and Thermal Model
The total volumetric heat generation in the traction battery pack was divided into reversible reaction heat, polarization heat and ohmic heat. The local polarization heat is
$$q_{\mathrm{pol}} \;=\; j_{\mathrm{Li}}\,\eta.$$
The reversible reaction heat is
$$q_{\mathrm{rev}} \;=\; j_{\mathrm{Li}}\, T\, \frac{\mathrm{d}U_{\mathrm{eq}}}{\mathrm{d}T}.$$
The ohmic heat includes electronic conduction in the solids, ionic conduction in the electrolyte and contact resistance effects. Tab heat generation is also considered through an additional term. The final energy balance equation used in the thermal model is
$$\rho_{b}\, C_{p,b}\,\frac{\partial T}{\partial t} \;=\; \nabla \cdot \left(\mathbf{k}\,\nabla T\right) \;+\; q_{\mathrm{rev}} \;+\; q_{\mathrm{pol}} \;+\; q_{\mathrm{ohm}} \;+\; q_{\mathrm{tab}}.$$
Heat transfer from the outer surfaces of the cell to the environment is modeled by combined convection and radiation:
$$-\mathbf{k}\,\nabla T \cdot \mathbf{n} \;=\; h\,\left(T_{\mathrm{surf}} – T_{\mathrm{amb}}\right) \;+\; \varepsilon_{\mathrm{rad}}\,\sigma_{\mathrm{SB}}\,\left(T_{\mathrm{surf}}^{4} – T_{\mathrm{amb}}^{4}\right).$$
The cell has a layered pouch architecture. Because the layers are much thinner in the through-plane direction than in the in-plane directions, I used orthotropic thermal conductivities. The parameter values used in the model are summarized in Table 3.
| Parameter | Negative electrode | Separator | Positive electrode | Negative current collector | Positive current collector |
|---|---|---|---|---|---|
| Material | Graphite | PE/PP type | LiCoO₂ | Cu | Al |
| Thickness (µm) | 34.5 | 8 | 26 | 6 | 13 |
| Particle radius (µm) | 5 | — | 1 | — | — |
| Solid phase volume fraction | 0.58 | — | 0.48 | — | — |
| Electrolyte volume fraction | 0.332 | 0.45 | 0.30 | — | — |
| Maximum solid concentration (mol m⁻³) | 31 507 | — | 56 250 | — | — |
| Initial solid concentration (mol m⁻³) | 25 330 | — | 19 020 | — | — |
| Solid conductivity (S m⁻¹) | 100 | — | 10 | 6×10⁷ | 3.8×10⁷ |
| Brugeman exponent | 4.1 | 2.3 | 1.5 | — | — |
| Density (kg m⁻³) | 1347 | 900 | 2500 | 8910 | 2700 |
| Specific heat (J kg⁻¹ K⁻¹) | 1437.4 | 1842.4 | 745.8 | 386 | 905 |
| Thermal conductivity (W m⁻¹ K⁻¹) | 1.04 | 0.344 | 1.58 | 401 | 238 |
The effective in-plane and through-plane thermal conductivities of the cell were calculated from the layer resistances in series and parallel. In the validation model, the average in-plane conductivity was 63.76 W m⁻¹ K⁻¹ and the through-plane conductivity was 1.11 W m⁻¹ K⁻¹. Natural convection on the outer surfaces was represented by a heat transfer coefficient of 25 W m⁻² K⁻¹.
3.3 Model Validation
I validated the coupled model by comparing simulated voltage curves and surface temperature curves with the experiments. The validation covered 0.5 C charging and 1 C, 2 C, 3 C and 4 C discharging in a 25 °C environment. Table 4 gives the maximum observed relative errors. The errors were larger in the middle of the 4 C discharge, where the electrochemical polarization changes quickly, but the overall agreement was sufficient for the subsequent thermal and aging analyses.
| Validation target | Condition | Maximum relative error |
|---|---|---|
| Cell voltage | 0.5 C charge and 1 C–4 C discharge at 25 °C | 5.56% at 4 C discharge |
| Surface temperature | 1 C–4 C discharge at 25 °C | 4.25% at late 4 C discharge |
4. Thermal Characteristics Analysis
4.1 Effect of Discharge Rate
Using the validated multi-physics model, I analyzed the internal heat generation of the traction battery pack at 1 C, 2 C and 4 C. The total heat-generation rate was lowest at the beginning of discharge, increased gradually, then showed a transient drop around 35% DOD and finally rose sharply after 80% DOD. This behavior is closely related to the entropic heat coefficient. At 35%–50% DOD, the reversible heat is endothermic and reduces the net heat rate. At the end of discharge, the reversible heat becomes strongly exothermic and the total heat-generation rate increases rapidly.
The relative contributions of the three heat sources are summarized in Table 5. At 1 C, the reversible reaction heat contributed about 58% of the total heat. With increasing discharge rate, the irreversible heat contributed more: at 4 C the reaction heat share decreased to about 31%, ohmic heat increased to roughly 6%, and polarization heat became dominant with roughly 63%. This explains why high-rate discharge produces a much faster surface temperature rise in the traction battery pack.
| Discharge rate | Reversible reaction heat share | Polarization heat share | Ohmic heat share |
|---|---|---|---|
| 1 C | 58% | Broadly below 40% | Small |
| 2 C | Intermediate | Increased | Increased |
| 4 C | 31% | 63% | 6% |
The spatial analysis showed that the positive electrode was the largest heat source among the cell components, because LiCoO₂ has a large entropic heat contribution. When the cell was discharged at higher C-rates, the heat share of the negative electrode and separator increased while the positive-electrode share decreased. I also observed that the irreversible heat of the negative electrode was larger than that of the positive electrode. The separator produced only a small amount of ohmic heat and therefore contributed very little to the total heat of the traction battery pack.
4.2 Effect of Ambient Temperature
At a constant discharge rate of 2 C, the heat-generation behavior depended strongly on ambient temperature. When the ambient temperature was decreased from 25 °C to 5 °C, the end-of-discharge temperature rise increased from about 8.3 °C to 10 °C and the average heat-generation rate increased from 190.4 kW m⁻³ to 235.4 kW m⁻³. In contrast, when the ambient temperature increased from 25 °C to 45 °C, the average heat-generation rate decreased to 173.7 kW m⁻³ and the end-of-discharge temperature rise decreased to about 7.5 °C. The lower ambient temperature increased the electrolyte viscosity and internal resistance, thus increasing polarization heat. At 5 °C, irreversible heat accounted for more than 60% of the total heat, while at 45 °C the reversible reaction heat was largely dominant.
4.3 Effect of Electrode Thickness and Particle Size
Electrode thickness and active-particle radius are important design variables for the traction battery pack. Increasing the electrode thickness increases the lithium-ion diffusion path and the electronic conduction path. In the simulation, I increased the positive and negative electrode thicknesses together while maintaining the designed capacity ratio. The discharge voltage became lower and the average cell temperature became higher at the same C-rate. The total heat-generation rate increased with electrode thickness. The irreversible heat fraction also increased because ohmic resistance became larger.
Particle radius affects the specific surface area and thus the effective exchange current density. I simulated both the positive and negative particle radii. With larger particles, the specific surface area was reduced, the local current density at the particle surface became higher, and the total heat generation increased. Table 6 summarizes the qualitative effect of these structural parameters on the heat-generation composition of the traction battery pack.
| Design parameter | Tendency | Main thermal consequence |
|---|---|---|
| Increase positive/negative thickness | Higher ohmic resistance | Higher total heat rate and higher irreversible heat share |
| Increase particle radius | Lower particle surface area | Higher polarization heat; reversible heat weakly affected |
| Increase thickness | Lower external surface area per volume | Higher core temperature and larger internal temperature gradient |
4.4 Temperature Uniformity
Because the UAV traction battery pack is frequently discharged at very high current, its temperature uniformity is especially critical. I used the difference between the core temperature and the surface temperature of the cell as an indicator of temperature uniformity. The core temperature was always higher than the surface temperature. Heat is generated throughout the volume, but the through-plane thermal conductivity is low, so heat cannot easily escape from the center of the cell.
I evaluated the effect of the external convection coefficient \(h\). For an ambient temperature of 25 °C, increasing \(h\) from 50 W m⁻² K⁻¹ to 200 W m⁻² K⁻¹ significantly reduced the mean cell temperature but increased the core-to-surface temperature difference because only the outer surface is cooled. At a 10 C discharge, the largest core-to-surface temperature difference increased from about 3.5 °C at \(h=50\) W m⁻² K⁻¹ to about 4.8 °C at \(h=200\) W m⁻² K⁻¹. The temperature difference is therefore not only a material property; it also depends on the external thermal management strategy.
Cell thickness is another important factor. Since the cell volume was kept constant in the simulation, increasing thickness reduced the effective cooling area and increased the core temperature. The simulation showed that when the total cell thickness is increased by 1 mm, the peak temperature at an 8 C discharge rises by about 1 °C and the core-to-surface temperature difference by about 0.6 °C. For a thickness of 11.6 mm, the maximum core-to-surface difference exceeded 6 °C at 8 C. This finding indicates that a very thick traction battery pack may suffer from uneven aging and local degradation unless the through-plane thermal conductivity is improved.
I also simulated the effect of increasing the through-plane thermal conductivity from 1.11 W m⁻¹ K⁻¹ to 3 W m⁻¹ K⁻¹. This change represents either thinner layers, better interfacial contact, or the use of a thermally conductive separator. With the improved through-plane conductivity, the core temperature was closer to the surface temperature. At a 10 C discharge and \(h=200\) W m⁻² K⁻¹, the maximum temperature difference decreased to about 2.6 °C instead of 4.8 °C. Therefore, a high-conductivity internal architecture is a very effective way to improve the temperature uniformity of the traction battery pack.
5. Aging Model and Aging Behavior
5.1 SEI Film Growth Mechanism
The accelerated aging experiment was performed at approximately 25 °C and with a 1 C charge / 3 C discharge cycle. The cell completed 450 cycles. Under this moderate-temperature condition, the main aging mechanism in the traction battery pack is the continued growth of the solid electrolyte interphase on the negative electrode. I therefore extended the electrochemical-thermal model with an SEI side-reaction model. The side reaction consumes lithium ions and electrolyte, increases the negative-electrode film resistance, and causes capacity fade. A simplified chemical reaction for SEI formation is
$$\mathrm{S} \;+\; \mathrm{Li}^{+} \;+\; \mathrm{e}^{-} \;\rightarrow\; \mathrm{P}_{\mathrm{SEI}},$$
where \(\mathrm{S}\) represents solvent molecules and \(\mathrm{P}_{\mathrm{SEI}}\) represents the SEI products. The side-reaction current density is modeled as
$$J_{\mathrm{SEI}} \;=\; -\frac{F\, c_{\mathrm{EC}}}{\dfrac{\delta_{\mathrm{film}}}{D_{\mathrm{EC}}} \;+\; \dfrac{1}{k_{\mathrm{SEI}}}\,\exp\!\left(\dfrac{\alpha_{\mathrm{SEI}}\,F\,\eta_{\mathrm{SEI}}}{R\,T}\right)}\left(1+H\,K\right),$$
where \(c_{\mathrm{EC}}\) is the solvent concentration, \(\delta_{\mathrm{film}}\) is the SEI film thickness, \(D_{\mathrm{EC}}\) is the solvent diffusion coefficient in the film, \(k_{\mathrm{SEI}}\) is the SEI reaction rate constant, and \(\eta_{\mathrm{SEI}}\) is the side-reaction overpotential. The factor \(H\) accounts for graphite particle expansion, while \(K\) describes the influence of the local negative-electrode state of charge. SEI film growth changes the effective film resistance according to
$$\frac{\partial \delta_{\mathrm{film}}}{\partial t} \;=\; -\frac{M_{\mathrm{SEI}}}{\rho_{\mathrm{SEI}}\,a_{v,\mathrm{neg}}\,F}\,J_{\mathrm{SEI}}, \qquad R_{\mathrm{film}} \;=\; R_{\mathrm{film},0} + \frac{\delta_{\mathrm{film}}}{\kappa_{\mathrm{SEI}}},$$
where \(M_{\mathrm{SEI}}\) and \(\rho_{\mathrm{SEI}}\) are the SEI molar mass and density, and \(\kappa_{\mathrm{SEI}}\) is the SEI ionic conductivity. The extra irreversible heat caused by the film resistance is
$$q_{\mathrm{SEI}} \;=\; \frac{J_{\mathrm{total}}^{2}\,R_{\mathrm{film}}}{a_{v,\mathrm{neg}}}.$$
By adding this heat source to the energy balance, the aging model becomes an electrochemical-thermal-aging coupled model. The main parameters used in the SEI model are listed in Table 7.
| Parameter | Value |
|---|---|
| EC concentration in electrolyte (mol m⁻³) | 4541 |
| SEI side-reaction transfer coefficient | 0.5 |
| SEI molar mass (kg mol⁻¹) | 0.16 |
| SEI density (kg m⁻³) | 1690 |
| SEI ionic conductivity (S m⁻¹) | 5×10⁻⁶ |
| Graphite expansion factor | 6.7 |
| Initial SEI thickness (nm) | 1 |
| SEI equilibrium potential (V) | 0.4 |
5.2 Comparison between Simulation and Aging Test
The accelerated aging test was conducted in a 25 °C chamber. The charge current was limited to 1 C and the discharge current was fixed at 3 C. The discharge cutoff voltage was 3.0 V. The available capacity was measured every 50 cycles. The model predicted the capacity fade with good accuracy. During the first 200 cycles, the maximum relative error between the simulated capacity and the measured capacity was 1.1%; between cycle 200 and cycle 450, the relative error was reduced to 0.68%. The comparison of simulated and measured discharge voltage curves at different aging states also showed good agreement.
5.3 Heat-Generation Evolution with Aging
After 450 cycles, the measured 1 C discharge capacity decreased to 92.6% of the initial value. The capacity fade appeared almost linear, but the simulation showed that the SEI growth rate was initially rapid and then gradually slowed down. Figure 10 in the original data series shows that the film resistance and film thickness increase quickly at the early stage and approach a lower growth rate as the cycle number increases.
The aging of the traction battery pack has a direct effect on its heat-generation behavior. I compared a fresh cell with a cell that had undergone 450 accelerated cycles, both discharged at 2 C and 25 °C. The average total heat-generation rate increased from 190.4 kW m⁻³ for the fresh cell to 278.4 kW m⁻³ for the aged cell. Table 8 presents the relative contributions of reversible heat and irreversible heat before and after aging.
| State | Reaction heat share | Polarization heat share | Ohmic heat share | Main heat-generating component |
|---|---|---|---|---|
| Fresh cell | 48.4% | 48.2% | Small | Positive electrode (54.3%) |
| After 450 cycles | 33.3% | 64.7% | Small | Negative electrode (60%) |
During aging, the positive-electrode polarization heat remained relatively stable, but the negative-electrode polarization heat increased significantly. This trend is consistent with the SEI growth mechanism on the negative electrode: a thicker SEI increases the negative-electrode film resistance and therefore the local overpotential. The result is that the dominant heat source shifts from the positive electrode to the negative electrode as the traction battery pack ages.
5.4 Influence of Cycle Rate
I simulated cycling at 1 C, 2 C and 3 C discharge currents with the same 1 C charge protocol in a 25 °C environment. The simulations show that higher discharge current accelerates capacity loss. After 450 cycles, the relative capacity difference between the 1 C-cycled cell and the 3 C-cycled cell was about two percentage points. The increase in cycle rate accelerated SEI film growth because high current produces stronger polarization, higher cell temperature and more pronounced particle stress. However, the effect of cycle rate was moderate compared with the effect of ambient temperature.
5.5 Influence of Ambient Temperature
Ambient temperature is an even stronger aging factor for the traction battery pack. I simulated cycling at 25 °C, 32 °C and 40 °C at a 1 C discharge rate. At 25 °C, the simulated relative capacity after 450 cycles was 92.6%. At 32 °C, the final relative capacity was lower. At 40 °C, the relative capacity decreased to 78.9%, which is below the usual end-of-life threshold of 80%. Higher temperature accelerates the SEI side reaction because the reaction rate constant is temperature dependent. At the same time, the higher cell temperature increases the total heat-generation rate, which creates a positive feedback that can accelerate thermal aging.
Table 9 compares the SEI film resistance and SEI film thickness after 450 cycles for two different ambient temperatures.
| Ambient temperature | Relative capacity after 450 cycles | Increase in SEI film resistance | Increase in SEI film thickness |
|---|---|---|---|
| 25 °C | 92.6% | Baseline | Baseline |
| 32 °C | Intermediate value | Larger than at 25 °C | Larger than at 25 °C |
| 40 °C | 78.9% | +0.073 Ω m² compared to 25 °C | +363 nm compared to 25 °C |
In addition to the total capacity loss, the model predicted that the total heat-generation rate increases with aging and that the irreversible heat fraction becomes larger. Therefore, an aged traction battery pack is more prone to overheating than a fresh one under the same discharge current. Thermal management is thus especially important for aging packs. By reducing the number of high-temperature and high-rate cycles, the SEI growth rate of the traction battery pack can be reduced and its useful lifetime prolonged.
6. Conclusion
In this work, I combined experiments and model simulations to study the thermal characteristics and aging behavior of a LiCoO₂/graphite soft-pack cell intended for a high-rate UAV traction battery pack. The experimental results revealed that increasing discharge rate or decreasing ambient temperature lowers the discharge voltage, reduces the usable capacity and raises the surface temperature rise of the traction battery pack. The entropic heat coefficient of the cell explained the “N” shape in the temperature-rise curves during discharge.
I then developed an electrochemical-thermal coupled model of the traction battery pack. The model was validated against measured voltage and surface temperature curves. The simulations showed that at low discharge rates, reversible heat dominates, while at high discharge rates or low ambient temperatures, irreversible heat becomes dominant. The positive electrode was the largest heat source in the fresh cell. Increasing electrode thickness and particle radius both increased irreversible heat generation. Under forced convection cooling, the core-to-surface temperature difference of the traction battery pack increased with the heat transfer coefficient and with cell thickness. A higher through-plane thermal conductivity significantly improved internal temperature uniformity.
Finally, I extended the model with an SEI growth aging sub-model. The aging model was validated against 450-cycle accelerated experiments. The model captured the capacity fade and the gradual increase in negative-electrode polarization heat. As the cell aged, the total heat-generation rate increased and irreversible heat became more important. Both a high ambient temperature and a high discharge rate accelerated SEI growth and capacity fade. Among the two, ambient temperature was the more influential factor in the simulated range. The results indicate that limiting high-temperature and high-discharge-rate operation is essential for extending the life of the traction battery pack.