In this work, I systematically investigate the design and thermal performance of a phase-change-material composite fin structure for high-voltage battery thermal management through numerical simulations and topology optimization. The proposed framework is intended to mitigate the excessive temperature rise and thermal runaway risk that constrain the practical deployment of high-energy-density lithium-ion cell modules. I first develop a validated heat-generation model for both prismatic and cylindrical high-voltage battery cells. Then, I perform topology optimization of the metallic fins embedded in the phase-change material region, using either the average temperature or the thermal compliance as the objective function. The optimized fin topologies are compared with conventional straight fins over a wide range of fin-to-phase-change-material volume ratios. Later, I reconstruct the optimized two-dimensional fin layout into a practical three-dimensional geometry and couple it with forced-air cooling schemes. I examine the effect of fin extension height, air velocity, and three air-cooling arrangements, namely one-side cooling, two-side co-current cooling, and two-side counter-current cooling. Finally, I evaluate the capability of the topology-optimized fin arrangements to suppress thermal-runaway propagation between adjacent cells.
The results demonstrate that topology-optimized fins are markedly superior to straight fins in reducing the temperature of high-voltage battery modules. For the prismatic cell, the topology-optimized design lowers the battery temperature by 8.2% and 10.1% at volume ratios of 0.1 and 0.6, respectively, compared to the straight-fin configuration under 3C discharge. In the cylindrical cell, hexagonal arrangement with optimized fins provides more uniform heat spreading than the square arrangement. The optimized fins also show clear benefits in thermal-runaway scenarios: in the prismatic model, a neighboring healthy cell only reaches 67.2°C when the optimized topology is present, while the corresponding value for a disconnected straight-fin design is 92.6°C. In the cylindrical model, the healthy cell temperature under optimized topology is 83.2°C, compared to 88.9°C for the straight-fin baseline. Moreover, the two-side counter-current air cooling scheme reduces the cell-to-cell temperature difference by nearly 90% compared with one-side cooling. These findings suggest that the topology-optimized phase-change composite fin structure is an attractive candidate for next-generation high-voltage battery thermal management systems.
1. Introduction
High-voltage battery systems are becoming the dominant energy-storage solution in electrified transportation because of their high energy density, long cycle life, and low self-discharge rate. Nevertheless, lithium-ion cells are extremely sensitive to operating temperature. The optimum window for a high-voltage battery pack is generally between 293.15 K and 323.15 K, and the maximum temperature difference among cells in one module should not exceed 5 K. Excessive temperature accelerates degradation, reduces available power and capacity, and may eventually trigger thermal runaway, which has been a major safety concern for electric vehicles. On the other hand, too low a temperature increases internal resistance and restricts charge and discharge capability. Therefore, an effective battery thermal management system is indispensable.
Conventional thermal management technologies for high-voltage battery packs include air cooling, liquid cooling, phase-change-material cooling, heat-pipe cooling, and various hybrid schemes. Air cooling is simple and reliable, but suffers from limited heat-transfer coefficient and spatial temperature non-uniformity. Liquid cooling offers superior heat-removal capability, yet it requires pumps, manifolds, and sealing, and raises concerns about leakage and parasitic power. Phase-change materials can absorb large amounts of latent heat during melting, providing passive temperature control without extra power consumption. However, most phase-change materials have very low thermal conductivity. The accumulated heat in a highly integrated high-voltage battery battery module cannot be dissipated rapidly, which may lead to local overheating and failure of the phase-change cooling strategy. To mitigate this problem, researchers usually embed metallic fins into the phase-change material to increase the effective conductivity and spread the heat more evenly.
The geometry of the fin plays a decisive role in the overall thermal performance. Traditional straight fins are widely used because of their ease of manufacturing, but they often fail to distribute heat optimally inside the phase-change domain. Many studies rely on trial-and-error or empirical rules to choose fin height, thickness, spacing, and number. Such an approach is time consuming and cannot fully explore the design space. In recent years, topology optimization has emerged as a powerful tool to generate high-performance conductive structures. By distributing a limited amount of highly conductive material within a design domain, topology optimization can yield unconventional branching or tree-like pathways that efficiently deliver heat from the cell surface to the entire phase-change region. In this thesis, I apply density-based topology optimization to design metallic fins for both prismatic and cylindrical high-voltage battery modules. The target is to minimize the average temperature of the battery while respecting a given volume fraction of aluminum with respect to the phase-change material.
The contribution of the present work is fourfold. First, a numerical heat-generation model of high-voltage battery cells is constructed and validated against experimental data. Second, topology optimization is performed with different volume ratios, and the resulting fin structures are compared with conventional straight fins. Third, a practical three-dimensional fin geometry is reconstructed and coupled with forced-air cooling to explore the effect of fin height, velocity, and cooling arrangement. Fourth, the optimized fin structures are tested in thermal-runaway-propagation simulations to evaluate their safety-enhancement potential.
2. Methodology
2.1 Numerical battery model and heat generation
Lithium-ion cells can be simplified as homogeneous heat sources for system-level thermal management studies. The governing energy equation is
$$\rho C_{p}\frac{\partial T}{\partial t}=\nabla\cdot\left(k\nabla T\right)+q,$$
where ρ is the effective density, Cp is the effective specific heat, k is the effective thermal conductivity, T is the temperature, t is time, and q is the volumetric heat-generation rate. In the present model, radiation is neglected and the boundary heat loss is described by Newton’s law of cooling,
$$q_{s}=h\left(T_{s}-T_{\infty}\right),$$
with the convective heat-transfer coefficient h, surface temperature Ts, and ambient temperature T∞.
The heat-generation rate is computed from the semi-empirical Bernardi equation:
$$q=\frac{I}{V}\left[ \left(U_{ocv}-U\right)+T\frac{dU_{ocv}}{dT}\right]=\frac{I}{V}\left(IR_{e}+T\frac{dU_{ocv}}{dT}\right),$$
where I is the current, V is the cell volume, Uocv is the open-circuit voltage, U is the terminal voltage, Re is the total internal resistance, and dUocv/dT is the entropic heat coefficient. During a constant-current discharge, the first term represents irreversible ohmic and polarization heat, while the second term represents reversible entropic heat.
For the prismatic cell, I use a validated 3C-discharge heat-generation correlation. For the cylindrical cell, a similar expression is applied based on measured internal resistance and entropic coefficient. The discharge rates considered are 1C, 2C, and 3C. The environment is initially at 25°C. By comparing the predicted cell temperature histories with experimental data, the maximum deviation is found to be less than 5%, which confirms that the model is reliable for the subsequent topology-optimization and thermal-management studies.
2.2 Geometry and thermophysical parameters
In this work, the high-voltage battery modules contain either prismatic cells or cylindrical 18650 cells. The main geometric and thermal properties are listed in the following tables.
The prismatic cell has dimensions 148 mm × 27 mm × 91 mm, a rated capacity of 51 Ah, and a mass of 816.47 g. Its anisotropic thermal conductivity is 14 W/(m·K) in the in-plane directions and 1.31 W/(m·K) through the thickness. The cylindrical cell is an 18650 type with radius 18 mm, height 65 mm, capacity 2.6 Ah, and mass 47.5 g. The cylindrical cell has a much higher axial conductivity, 37.6 W/(m·K), but a very low radial conductivity, 0.2 W/(m·K). Table 1 summarizes these parameters.
| Parameter | Prismatic cell | Cylindrical cell |
|---|---|---|
| Nominal capacity (Ah) | 51 | 2.6 |
| Mass (g) | 816.47 | 47.5 |
| Dimensions (mm) | 148 × 27 × 91 | R = 18, H = 65 |
| Specific heat capacity (J/(kg·K)) | 1050.31 | 1200 |
| Thermal conductivity (W/(m·K)) | kxy = 14, kz = 1.31 | kxy = 0.2, kz = 37.6 |
| Density (kg/m³) | 2245.3 | 2873.5 |
The phase-change material selected for this study is n-eicosane, a paraffin with a phase-change temperature range from about 35°C to 37°C. Its latent heat is 241 kJ/kg. The thermal properties of the solid and liquid phases are listed in Table 2. Because paraffin has a low thermal conductivity, pure conduction in the phase-change material is insufficient for high-rate discharge of a high-voltage battery pack.
| State | Melting temperature (°C) | Density (kg/m³) | Specific heat capacity (kJ/(kg·K)) | Thermal conductivity (W/(m·K)) |
|---|---|---|---|---|
| Solid | 35 | 810 | 1.9 | 0.39 |
| Liquid | 37 | 770 | 2.2 | 0.157 |
| Latent heat | 241 kJ/kg | |||
Aluminum is used as the fin material because of its high thermal conductivity and light weight. Its properties are presented in Table 3.
| Property | Value |
|---|---|
| Specific heat capacity (J/(kg·K)) | 900 |
| Thermal conductivity (W/(m·K)) | 238 |
| Density (kg/m³) | 2700 |
The required volume of phase-change material is determined by assuming that all the heat generated at 3C discharge during 1200 s is stored as latent heat. The calculation is given by
$$V_{pcm}=\frac{\int_{0}^{1200}q\left(t\right)dt}{\rho_{pcm}L},$$
where L is the latent heat. For the prismatic cell, the calculated volume is about 23.144 cm³; for the cylindrical cell, the required volume is about 13.715 cm³. These volumes are kept constant when the fin-to-PCM volume ratio ω varies, so the total area occupied by the fins and PCM is adjusted accordingly.
2.3 Topology optimization methodology
The topology optimization problem is solved in a two-dimensional domain that represents one cross-section of the fin-PCM composite layer adjacent to one battery wall. The design variable γ is a pseudo-density that varies continuously between 0 and 1. A value of γ=1 denotes aluminum fin material, while γ=0 denotes pure phase-change material. By using the density-based method, the local effective properties are interpolated as
$$k\left(\gamma\right)=k_{pcm}+\gamma\left(k_{fin}-k_{pcm}\right),$$
$$\rho\left(\gamma\right)=\rho_{pcm}+\gamma\left(\rho_{fin}-\rho_{pcm}\right),$$
$$C_{p}\left(\gamma\right)=C_{p,pcm}+\gamma\left(C_{p,fin}-C_{p,pcm}\right).$$
To avoid numerical artefacts such as checkerboard patterns and mesh-dependent solutions, a Helmholtz-type density filter is applied:
$$-\nabla\cdot\left(r_{\min}^{2}\nabla\tilde{\gamma}\right)+\tilde{\gamma}=\gamma,$$
where rmin is the filter radius. The filtered density $\tilde{\gamma}$ is then sharpened by a hyperbolic-tangent projection:
$$\bar{\gamma}=\frac{\tanh\left(\beta\left(\tilde{\gamma}-\gamma_{\beta}\right)\right)+\tanh\left(\beta\gamma_{\beta}\right)}
{\tanh\left(\beta\left(1-\gamma_{\beta}\right)\right)+\tanh\left(\beta\gamma_{\beta}\right)},$$
where β=8 and γβ=0.5 are used in the optimization. The purpose of the projection is to reduce grey-scale transition regions and obtain a more crisp fin structure.
Two objective functions are examined. The first is the thermal compliance,
$$J_{1}=\int_{\Omega}qT\,d\Omega,$$
and the second is the volume-average temperature,
$$J_{2}=\frac{1}{|\Omega|}\int_{\Omega}T\,d\Omega.$$
The volume constraint is
$$\int_{\Omega}\bar{\gamma}\,d\Omega \le \omega V_{pcm},$$
where ω denotes the fin-to-PCM volume ratio. In this study, I solve the steady-state heat-conduction equation within the design domain for topology generation, and then insert the obtained optimized fin layouts into the full transient battery model to evaluate the cooling performance.
3. Topology optimization results for the prismatic high-voltage battery
3.1 Effect of objective function
I first compare the optimized fin topologies obtained with the thermal-compliance objective and the volume-average-temperature objective. Figure 3-8 of the original thesis indicates that the average-temperature objective produces lower battery temperatures than the thermal-compliance objective. At low volume ratios, the difference is more evident. For ω=0.1, the thermal-compliance optimized structure yields a final temperature of 41.26°C, while the average-temperature optimized structure yields only 39.58°C. For ω=0.3, the corresponding values are 40.37°C and 37.77°C. Because the average-temperature objective directly relates to cell temperature control, all subsequent topology optimizations in this work use the average-temperature objective.
Table 4 compares the final battery temperatures at the end of 3C discharge under different volume ratios for the prismatic cell. The topology-optimized fins always outperform straight fins at the same volume ratio.
| Fin-to-PCM volume ratio ω | Straight-fin temperature (°C) | Topology-optimized fin temperature (°C) |
|---|---|---|
| 0.1 | 44.3 | 40.3 |
| 0.3 | — | 38.0 (approximately) |
| 0.4 | — | 37.4 |
| 0.5 | — | 37.2 |
| 0.6 | 41.3 | 37.1 |
It can be seen that, for a volume ratio of 0.1, the topology-optimized fin reduces the battery temperature from 44.3°C to 40.3°C, a reduction of 8.2%. At ω=0.6, the optimized fin yields 37.1°C versus 41.3°C for the straight fin, corresponding to a reduction of about 10.1%. With increasing ω from 0.1 to 0.6, the battery temperature decreases by about 7.9%. However, the incremental benefit becomes very small after ω=0.4. Therefore, the topology-optimized fin structure with ω=0.4 is considered to be the most economical choice for the prismatic high-voltage battery module.
The optimized fin layouts exhibit a tree-like branching structure that extends from the battery wall into the phase-change domain. This branching pattern enables the heat to be delivered to a much larger portion of the phase-change material rather than being concentrated only near the heat source. As a result, larger portions of the PCM can melt simultaneously and absorb latent heat, thereby reducing the cell temperature and improving the uniformity of the temperature field.
4. Topology optimization for cylindrical high-voltage battery cells
The cylindrical high-voltage battery can be arranged in either a square lattice or a hexagonal lattice. The shape of the phase-change domain is correspondingly a square or a regular hexagon surrounding each cylindrical cell. I kept the PCM volume constant and changed the side length of the design domain when the volume ratio was varied. Topology optimization was performed separately for each lattice geometry.
Figure 3-18 and Figure 3-19 in the original dissertation show the optimized fin structures for square and hexagonal lattices. In both cases, the fins become thicker and more branched as ω increases. The optimized topology spreads the highly conductive aluminum from the cell surface toward the corners of the phase-change domain, which enhances heat diffusion to regions that would otherwise remain inactive during melting.
Table 5 summarizes the final battery temperatures for the square arrangement with straight fins and topology-optimized fins.
| Volume ratio ω | Straight-fin temperature (°C) | Topology-optimized temperature (°C) |
|---|---|---|
| 0.1 | 45.4 | 42.7 |
| 0.4 | — | 40.2 |
| 0.5 | — | 40.0 |
At ω=0.1, the topology-optimized fin lowers the temperature by about 6% compared with the straight-fin design. Increasing ω from 0.1 to 0.5 reduces the temperature by 6.2%. In the hexagonal arrangement, the optimized topology performs even better. At ω=0.1, the topology-optimized structure gives a final temperature of 41.7°C versus 45.2°C for the straight-fin case, which is a reduction of about 7.8%. The results are shown in Table 6.
| Volume ratio ω | Straight-fin temperature (°C) | Topology-optimized temperature (°C) |
|---|---|---|
| 0.1 | 45.2 | 41.7 |
| 0.5 | — | 39.5 |
Comparing the square and hexagonal topologies, the hexagonal layout provides lower cell temperatures because the six vertices offer more paths for heat conduction and make the optimized fins more evenly distributed around the cell. For ω=0.1, the hexagonal topology reaches 41.7°C, whereas the square topology reaches 42.7°C. At ω=0.5, the difference is reduced to about 0.6°C. Consequently, the hexagonal arrangement appears more favorable for cylindrical high-voltage battery cells from a thermal-management viewpoint.
5. Three-dimensional reconstruction and forced-air-cooling performance
5.1 From two-dimensional topology to manufacturable fin geometry
The raw topology-optimized fins have irregular zig-zag boundaries, which are difficult to mass-produce and may cause non-negligible computational issues in three-dimensional fluid simulations. Therefore, I reconstructed the optimized two-dimensional topology into a simpler geometry with straight edges, while retaining the overall branch distribution. The reconstructed fin is easier to fabricate by conventional extrusion or forging. The battery temperature using the reconstructed fin is shown in Figure 4-6 in the original work. The reconstructed fin causes only a very small temperature increase of about 0.1°C when compared with the exact topological fin. This demonstrates that the reconstructed layout retains most of the thermal benefit of the optimized topology.
5.2 Mesh independence
Because the reconstructed three-dimensional model contains both solid domains and external air passages, an unstructured tetrahedral mesh was used to discretize the complex geometry. I performed a mesh-independence study on a simplified single-side-air-cooled high-voltage battery module with a fin extension of 5 mm. The number of mesh elements ranged from 170,000 to 850,000. The predicted average battery temperatures are shown in Table 7.
| Mesh count | Average battery temperature (°C) |
|---|---|
| 170,000 | 38.7 (approximately) |
| 320,000 | 38.65 |
| 590,000 | 38.66 |
| 850,000 | 38.64 |
The temperature difference between 320,000 and 850,000 elements is less than 0.1%, so all subsequent simulations were performed with approximately 320,000 elements, with slight modifications for the different air-cooling geometries.
5.3 Effect of fin extension height
For the air-cooling study, I fixed the air velocity at 1.5 m/s and varied the fin extension height above the phase-change layer from 0 mm to 20 mm. Fin extension increases the heat-transfer area exposed to air and also provides a direct conductive path from the PCM and the cell to the cooling air. Table 8 lists the predicted battery temperatures at the end of 3C discharge.
| Fin extension height (mm) | Battery temperature (°C) |
|---|---|
| 0 | 40.6 |
| 5 | 38.7 |
| 10 | 38.3 |
| 15 | 38.1 |
| 20 | 37.8 (approximately) |
The temperature reduction is significant when the fin height is increased from 0 to 10 mm, but it becomes only approximately 0.3°C when the height increases beyond 10 mm. Thus, 10 mm is chosen as a good compromise between cooling efficiency, material consumption, and fan-power requirements. Higher fins increase air-flow resistance and packaging volume without producing meaningful temperature reduction.
5.4 Comparisons of air-cooling schemes
I then evaluated three air-cooling arrangements: single-sided cooling (air only on one side), double-sided co-current cooling (the same flow direction on both sides), and double-sided counter-current cooling (opposite flow directions on the two sides). The computations were performed under 3C discharge, with a fixed fin height of 10 mm. Five air velocities were considered: 1.0, 1.5, 2.0, 2.5, and 3.0 m/s.
Figure 4-19 in the original work reveals that both double-sided air-cooling schemes decrease the average battery temperature much more effectively than single-sided cooling. At a given wind speed, the two double-sided schemes provide almost identical average temperatures. For instance, at 1.0 m/s, single-sided cooling gives 38.6°C, while both double-sided schemes give approximately 37.7°C. At 3.0 m/s, the single-sided result is 37.8°C, while the double-sided result is about 36.9°C. Interestingly, further increasing the velocity from 2.5 to 3.0 m/s yields only a 0.1°C improvement; therefore, 2.5 m/s is sufficient for practical high-voltage battery thermal management.
The largest difference between the cooling arrangements appears in the cell-to-cell temperature difference. Under single-sided cooling, the temperature difference between the cell nearest the inlet and that nearest the outlet increases with velocity. Under double-sided co-current cooling, the cell-to-cell temperature difference is also relatively large because cold air enters from one end on both sides and heats up along the flow path. Under double-sided counter-current cooling, however, the high-temperature side on one side corresponds to the low-temperature side on the other side; the two effects compensate, yielding a nearly uniform cell temperature distribution. The temperature differences at the end of discharge are summarized in Table 9.
| Cooling scheme | Battery temperature (°C) | Cell-to-cell temperature difference (°C) |
|---|---|---|
| Single-sided | 38.0 | 0.15-0.18 |
| Double-sided co-current | 37.0 | 0.16-0.20 |
| Double-sided counter-current | 37.0 | ≤ 0.01 |
The double-sided counter-current scheme reduces the cell-to-cell temperature difference by almost 90% compared with single-sided cooling. This scheme does not significantly increase the pressure loss per unit heat removal compared with co-current flow, for the same total flow rate. Therefore, for a battery pack built with a topology-optimized phase-change composite fin structure, the double-sided counter-current air-cooling mode is the recommended choice because it simultaneously ensures low average temperature and excellent temperature uniformity.
6. Thermal runaway suppression
6.1 Model setup for thermal runaway
Thermal runaway is an extreme event in which a high-voltage battery cell releases its stored energy in a very short period because of internal short-circuit, mechanical abuse, or overheating. In the simulations, the onset temperature of thermal runaway is set to 150°C. The heat-generation rate during runaway is estimated from the electrical energy stored in the cell and the total release time. For the prismatic cell, the runaway volumetric heat source is 136.3 MW/m³. For the cylindrical cell, the corresponding value is 129.3 MW/m³. In the first 60 seconds, both cells are discharged at 3C and remain in a safe state. After 60 seconds, one cell switches to the thermal-runaway heat-generation mode for 10 seconds, while the neighboring healthy cell continues with normal 3C discharge. The total simulated time is 500 seconds.
6.2 Square battery thermal runaway
I compared three fin configurations for the prismatic module: connected straight fins, disconnected straight fins, and topology-optimized fins. The volume ratio is fixed at 0.4. The connected straight fins create a continuous high-conductivity path between the two cells, which accelerates heat transfer from the runaway cell to the healthy cell. In that case, the healthy cell temperature exceeds the 150°C thermal-runaway threshold within the 500 s simulation. Disconnecting the straight fins weakens the thermal bridge and protects the healthy cell better. However, the topology-optimized fins provide the best protection. At 500 s, the healthy cell in the topology-optimized model has a maximum temperature of only 67.2°C, whereas the disconnected straight-fin model reaches 92.6°C. Table 10 reports the healthy cell temperatures at the end of the simulation.
| Fin structure | Healthy cell temperature (°C) |
|---|---|
| Connected straight fin | >150 (thermal runaway) |
| Disconnected straight fin | 92.6 |
| Topology-optimized fin | 67.2 |
The topology-optimized fin layout effectively blocks direct heat conduction paths while simultaneously distributing heat from the runaway cell into the phase-change material over a large volume. Since the PCM melts and absorbs latent heat, the temperature of the healthy cell remains far below the critical threshold. This confirms the safety advantage of topology-optimized fins in high-voltage battery modules.
6.3 Cylindrical battery thermal runaway
For the cylindrical cell in the hexagonal arrangement, the same thermal-runaway test is performed. The straight-fin structure is compared with the topology-optimized fin structure at ω=0.4. At 500 s, the healthy cell temperature in the straight-fin model is 88.9°C, while in the topology-optimized model it is 83.2°C. Table 11 lists the results. Although the absolute difference is smaller than in the prismatic case, the optimized fins still reduce healthy-cell heating and slow the propagation of thermal runaway. Overall, the topology-optimized fin structure can act as a thermal fuse that interrupts the propagation of thermal runaway between neighboring cells.
| Fin structure | Healthy cell temperature (°C) |
|---|---|
| Straight fin | 88.9 |
| Topology-optimized fin | 83.2 |
7. Discussion
The present numerical investigation confirms the potential of topology-optimized fins for high-voltage battery thermal management. Several important observations can be drawn from the simulations.
First, the optimized fin distribution strongly depends on the shape of the phase-change domain. For the prismatic cell, the optimized fins resemble narrow branching trees that emanate from the heated wall. For the cylindrical cell in a hexagonal PCM domain, the fins branch toward the vertices of the hexagon, making more efficient use of the available phase-change material. These topologies are difficult to anticipate using intuition alone.
Second, there is an optimum fin volume ratio beyond which cells exhibit diminishing returns. Increasing the aluminum fraction improves the effective thermal conductivity but simultaneously reduces the PCM volume. When the fin volume ratio is higher than 0.4, the cell temperature does not reduce appreciably because the PCM latent-heat capacity has become the limiting factor. A ratio of about 0.4 offers a good design point for both prismatic and cylindrical high-voltage battery cells.
Third, combining a phase-change composite with forced-air cooling is beneficial only if the fin arrangement allows heat to be externally removed. The fin extension height must be selected to balance improved heat transfer with pressure loss and structural requirements. In this work, a height of 10 mm appears to be a reasonable choice. Air cooling can effectively remove heat from the side of the phase-change layer, but the temperature uniformity of a battery module is highly dependent on the flow arrangement. Double-sided counter-current cooling is particularly attractive because it cancels the inlet-to-outlet temperature gradient on a module scale.
Fourth, the thermal-runaway simulations indicate that the topology-optimized fins provide better protection than straight fins because the optimized topology inherently creates narrow, discontinuous, and distributed metallic paths instead of a long continuous straight metallic strip. The topology prevents a single low-resistance heat path from appearing between adjacent cells, while still allowing efficient heat extraction during normal operation. This dual function is important for the passive safety of high-voltage battery packs.
It should be noted that the current simulations are based on the assumption of homogeneous cell heat generation and temperature-independent thermophysical properties except for the phase change. In a real high-voltage battery, current density, state-of-charge, aging, and contact resistance may introduce additional local variations. Future work should incorporate three-dimensional thermal-electrochemical coupling and experimental tests on additively manufactured topology-optimized fins. Nevertheless, the trends and conclusions presented here are expected to remain valid because the model has been quantitatively validated against discharge-temperature data under multiple C-rates.
8. Conclusion
In this thesis, I have presented a systematic numerical study on the topology optimization of phase-change cooling structures for high-voltage battery systems. The main conclusions can be summarized as follows:
(1) A numerical heat-generation model of prismatic and cylindrical lithium-ion cells was established and verified with experimental data within 5% error. The model successfully reproduces the temperature evolution under 1C, 2C, and 3C discharge conditions.
(2) Topology-optimized fins outperform conventional straight fins in lowering the temperature of high-voltage battery modules. With the prismatic cell and ω=0.1, the optimized fin reduces the final 3C-discharge temperature from 44.3°C to 40.3°C, a decrease of 8.2%. With ω=0.6, the reduction is about 10.1%. For the cylindrical cell, the hexagonal topology is superior to the square topology; the hexagonal optimized fin design reduces the temperature by about 7.8% when ω=0.1. In all cases, ω=0.4 provides an economical compromise between cooling enhancement and available latent-heat storage.
(3) Reconstructing the optimized topology into a simpler manufacturable shape has only a minor influence on battery temperature. Air cooling improves heat removal significantly when the fin extends above the PCM surface. A fin height of 10 mm and an air velocity of 2.5 m/s are appropriate for the present configuration.
(4) Among the three air-cooling arrangements, double-sided counter-current cooling achieves the best overall thermal performance. It reduces the average battery temperature as effectively as double-sided co-current cooling, and it suppresses the cell-to-cell temperature difference by nearly 90% compared with the single-sided scheme. Thus, for a high-voltage battery module with phase-change composite fins, double-sided counter-current air cooling is strongly recommended.
(5) The topology-optimized fins demonstrate superior thermal-runaway-propagation suppression. In the prismatic battery, the healthy cell temperature remains at 67.2°C at 500 s, whereas a disconnected straight fin leads to 92.6°C and a connected straight fin even triggers thermal runaway in the adjacent cell. In the cylindrical battery, the topology-optimized structure also yields a lower healthy-cell temperature (83.2°C versus 88.9°C). Optimizing the fin geometric layout is therefore an effective strategy to improve both normal heat dissipation and thermal-runaway safety of high-voltage battery packs.
The present study provides a solid basis for further design optimization and experimentation of phase-change composite fins. Future research should include multi-physics coupling, additive manufacturing of optimized fin structures, and tests with realistic dynamic drive cycles to fully realize the benefits of topology optimization for high-voltage battery thermal management.