High-Voltage Battery Thermal Management via Coupled Microchannels and Phase Change Materials

Driven by global climate change and the urgent need for an energy transition, energy-efficient electric vehicles have become a strategic priority for many countries. The high-voltage battery pack is the core energy-storage device of an electric vehicle, and lithium-ion chemistry has dominated this field because of its high energy density, long cycle life, and low self-discharge rate. However, the operating performance, cycle lifespan, and safety of the high-voltage battery are heavily dependent on temperature. Under abusive high-temperature conditions, the battery may suffer from accelerated capacity fade, internal short circuits, and even thermal runaway. In cold environments, the capacity of the battery can drop by more than 50%, and lithium plating becomes a serious hazard. Moreover, large-format high-voltage battery packs face the problem of poor temperature uniformity, not only because of the large heat generation gradients inside a single cell, but also because of the uneven coolant distribution across the module. Research indicates that once the cell-to-cell temperature difference in a pack exceeds 5 °C, the cycle life of the high-voltage battery can be reduced by more than 30%. Thus, an advanced battery thermal management system (BTMS) must be able to regulate the temperature of the high-voltage battery under both hot and cold extremes, while keeping the maximum temperature and the temperature gradient within prescribed safety limits.

Traditional single-mode cooling methods, such as air cooling or indirect liquid cooling, cannot always satisfy the demanding requirements of high-voltage battery operation under a wide range of ambient temperatures. Air cooling is simple and low-cost, but its heat transfer capability is limited by the low specific heat and thermal conductivity of air. Liquid cooling can provide a high heat-transfer coefficient, but it suffers from parasitic power consumption and a relatively large temperature gradient along the coolant flow direction. Phase-change material (PCM) cooling offers passive and uniform heat absorption, but its low thermal conductivity and finite latent heat greatly restrict its long-term performance under continuous high-discharge conditions. Heat-pipe cooling, on the other hand, is compact but still requires careful thermal integration with an active heat sink. To combine the advantages of active liquid cooling and passive latent-heat storage, many researchers have turned their attention to hybrid BTMS architectures, where PCM is coupled with a liquid-circulating cold plate or with microchannels.

In this context, I proposed and numerically investigated a new hybrid BTMS for high-voltage battery packs based on a honeycomb-inspired layout that couples microchannels with a composite phase-change material. The proposed system is designed to serve three critical functions simultaneously: heat dissipation at high ambient temperatures, preheating at low ambient temperatures, and thermal insulation when the high-voltage battery is parked in winter. The novelty of this study lies in the integration of a microchannel liquid circuit and a graphite/paraffin composite PCM into a single structural framework, which allows for efficient heat removal and rapid low-temperature warm-up without compromising the module compactness. I first established a validated thermal model of a cylindrical 18650 high-voltage battery, then constructed the coupled BTMS model, and systematically investigated the effects of key design parameters on cooling and preheating performance. The simulation results reveal important nonlinear effects and provide an optimal set of parameters that can maintain the high-voltage battery within the recommended temperature window over a wide range of conditions.

1. Thermal Behavior of the High-Voltage Battery

1.1 Cell structure and operating principle

My study focused on a commercial 18650 cylindrical cell with a LiFePO4 cathode, a graphite anode, and a nominal capacity of 1.5 Ah. The physical dimensions are 18 mm in diameter and 65 mm in height. The main components are the positive electrode, negative electrode, separator, electrolyte, and a metal casing. During discharge, lithium ions migrate from the anode through the electrolyte to the cathode, while electrons travel through the external circuit. The corresponding electrode reactions can be written as:

Positive:

$$LiMO_{2} \rightleftharpoons Li_{1-x}MO_{2} + xLi^{+} + xe^{-}$$

Negative:

$$C_{n} + xLi^{+} + xe^{-} \rightleftharpoons Li_{x}C_{n}$$

Overall:

$$LiMO_{2} + C_{n} \rightleftharpoons Li_{1-x}MO_{2} + Li_{x}C_{n}$$

1.2 Heat-generation mechanisms

The total heat generated inside a lithium-ion high-voltage battery during normal operation can be decomposed into four contributions: reaction heat due to entropy change, Joule heat due to ohmic resistance, polarization heat due to electrochemical polarization, and side-reaction heat. In normal operating ranges, the side-reaction heat is usually negligible when there is no overcharge or overdischarge. The reversible reaction heat is given by:

$$Q_r = IT \frac{dE}{dT}$$

where \(I\) is the current, \(T\) is the temperature, and \(dE/dT\) is the temperature coefficient of the open-circuit voltage. The irreversible Joule heat and polarization heat can be combined into a single term involving the total internal resistance \(R_{\Omega}+R_{P}\):

$$Q_j + Q_p = I^2(R_{\Omega} + R_P)$$

For a uniform heat-generation assumption, I adopted the simplified Bernardi model, which can be expressed as:

$$q = \frac{I}{V_0}\left[(E_{OC}-E) – T\frac{dE_{OC}}{dT}\right] \approx \frac{I^2(R_{\Omega}+R_P)}{V_0} – \frac{IT}{V_0}\frac{dE_{OC}}{dT}$$

Here \(V_0\) is the volume of the active battery region, \(E_{OC}\) is the open-circuit voltage, and \(E\) is the working voltage. For a given state of charge and temperature, the combined internal resistance \(R_{\Omega}+R_P\) of the cell can be expressed as a third-order polynomial in temperature:

$$R_{\Omega}+R_{P} = -0.0001T^3 + 0.0134T^2 – 0.5345T + 12.407$$

At an ambient temperature of 35°C, this model produces volumetric heat-generation rates of 5362 W/m³ for a 1C rate, 30586 W/m³ for 2C, 60156 W/m³ for 3C, and 126758 W/m³ for a 4C discharge. To verify the battery heat-generation model, I performed a transient simulation at 20°C with a 3C discharge rate and compared the predicted maximum cell temperature with experimental data. The maximum deviation was only 0.64°C, which is within an acceptable range. Hence, the thermal model is sufficiently accurate for subsequent parametric studies.

1.3 Effective thermophysical properties

In the simulation, I treated the 18650 cell as a homogeneous cylinder with anisotropic thermal conductivity. The effective density \(\rho_b\), specific heat \(c_{p,b}\), radial conductivity \(k_r\), and axial conductivity \(k_z\) were calculated by volume-averaging the properties of each battery component. The resulting parameters are summarized in Table 1.

Table 1. Material properties used for the high-voltage battery model.
Property Value
Density (kg/m³) 2180
Specific heat (J/(kg·K)) 1066
Radial thermal conductivity (W/(m·K)) 1.1
Circumferential thermal conductivity (W/(m·K)) 1.1
Axial thermal conductivity (W/(m·K)) 12.5

The energy balance within the battery was solved using the three-dimensional heat conduction equation in cylindrical coordinates:

$$\rho_b c_{p,b} \frac{\partial T}{\partial \tau} = \frac{1}{r}\frac{\partial}{\partial r}\left(\lambda_r r\frac{\partial T}{\partial r}\right) + \frac{1}{r^2}\frac{\partial}{\partial \phi}\left(\lambda_\phi \frac{\partial T}{\partial \phi}\right) + \frac{\partial}{\partial z}\left(\lambda_z \frac{\partial T}{\partial z}\right) + q$$

2. Architecture of the Proposed Hybrid BTMS

2.1 Honeycomb-inspired microchannel and PCM packing

I designed a battery module consisting of 16 cylindrical 18650 cells arranged in a staggered 4×4 pattern, a configuration that resembles a honeycomb in the cross-section. The cells are embedded in a composite PCM, which itself is contained within an aluminum frame. Vertical circular microchannels with a diameter of 1 mm are fabricated inside the aluminum frame. The frame height equals the battery length of 65 mm, the frame thickness is 1.5 mm, and the gap between the frame inner wall and the battery is 1.25 mm. Water is used as the heat-transfer fluid.

Three different microchannel distributions were examined. Case 1 contains 63 channels evenly distributed at the center of the six outer edges of the hexagonal frame. Case 2 contains 111 channels, with an additional channel at each vertex of the hexagonal frame. Case 3 contains 174 channels, where extra channels are placed at the trisection points along each edge as well as at the vertices. These configurations are intended to explore the trade-off between an increased heat-transfer area and the resulting reduction in per-channel flow velocity when the total mass flow rate is kept constant.

2.2 Governing equations for the fluid and PCM

For the water flowing in the microchannels, the continuity, momentum, and energy equations are respectively:

$$\frac{\partial \rho_w}{\partial \tau} + \nabla \cdot (\rho_w \mathbf{v}) = 0$$

$$\frac{\partial (\rho_w \mathbf{v})}{\partial \tau} + \nabla \cdot (\rho_w \mathbf{v}\mathbf{v}) = -\nabla p + \nabla \cdot (\mu_w \nabla \mathbf{v})$$

$$\frac{\partial (\rho_w c_{p,w} T)}{\partial \tau} + \nabla \cdot (\rho_w c_{p,w} \mathbf{v} T) = \nabla \cdot (\lambda_w \nabla T)$$

where \(\rho_w\), \(c_{p,w}\), \(\lambda_w\), and \(\mu_w\) are the density, specific heat, thermal conductivity, and dynamic viscosity of water. Heat transfer in the solid aluminum frame is governed by pure conduction.

For the composite PCM, I used the enthalpy-porosity method. The energy equation is written in terms of enthalpy \(H\):

$$\frac{\partial (\rho_{PCM} H)}{\partial \tau} = \nabla \cdot (\lambda_{PCM} \nabla T)$$

The enthalpy consists of the sensible enthalpy \(H_1\) and the latent enthalpy \(\Delta H\):

$$H = H_1 + \Delta H$$
$$H_1 = \int_{T_0}^{T} c_{p,PCM} dT$$
$$\Delta H = \beta \gamma$$

Here \(\beta\) is the liquid fraction of the PCM and \(\gamma\) is the latent heat of fusion. For a PCM that melts over a temperature interval between solidus temperature \(T_s\) and liquidus temperature \(T_l\), the liquid fraction is defined as:

$$\beta = \begin{cases}
0, & T < T_s \\
\dfrac{T – T_s}{T_l – T_s}, & T_s \le T \le T_l \\
1, & T > T_l
\end{cases}$$

2.3 Composite PCM properties

To overcome the low thermal conductivity of pure paraffin, I used expandable graphite as a thermal-conductivity promoter. A series of composite PCMs with graphite mass fractions of 0%, 2%, 5%, 8%, 10%, 15%, and 20% was considered. Table 2 lists the corresponding thermal properties, which were taken from experimental characterizations reported in the literature. With a higher graphite fraction, the thermal conductivity increases dramatically, while the specific heat and latent heat decrease because graphite replaces part of the paraffin.

Table 2. Thermophysical properties of paraffin/graphite composite PCM at different graphite fractions.
Graphite fraction (%) Thermal conductivity (W/(m·K)) Density (kg/m³) Specific heat (J/(kg·K)) Latent heat (J/g) Melting temperature (°C)
0 0.245 800 2300 200.66 37
2 0.324 800 2260 196.65 37
5 0.907 800 2200 191.37 37
8 1.374 800 2140 184.61 37
10 1.687 800 2100 180.60 37
15 2.499 800 2000 170.56 37
20 3.355 800 1900 160.53 37

2.4 Numerical setup and grid independence

I performed transient conjugate heat-transfer simulations using the ANSYS Fluent software. The pressure-based solver was used, and the Laminar model was chosen because the Reynolds number in the microchannels was below 2300. The Solidification/Melting model was enabled to account for the phase-change process. Boundary conditions were set depending on the operating scenario. For the cooling study, the initial temperature, inlet temperature, and outlet backflow temperature were set to 35°C, and the battery heat source was switched on with a volumetric heat-generation rate corresponding to 4C discharge. Convective heat-transfer coefficients were assigned to the outer surfaces of the battery module: 10 W/(m²·K) for the aluminum frame and 5 W/(m²·K) for the exposed PCM surfaces.

Before performing parametric simulations, I carried out a grid-independence test using four different mesh counts: 1,166,891; 1,971,858; 3,802,869; and 7,464,891. The variation in maximum temperature at the end of discharge was negligible when the mesh count exceeded 1,971,858. Therefore, the model with 1,971,858 cells was selected. Similarly, time-step independence was verified for 0.5 s, 1 s, and 2 s. Since the predicted temperatures differed by less than 0.01°C, a fixed time step of 2 s was adopted. The convergence criterion for the energy equation was 10⁻⁶, while the momentum and continuity residuals were set to 10⁻⁴.

3. Cooling Performance of the Hybrid BTMS

3.1 Comparison with single-phase PCM cooling

I first compared the proposed microchannel-coupled PCM BTMS (denoted PCM-MCHS) with a passive PCM-only cooling system. The ambient temperature was 35°C, the discharge rate was 4C, and the inlet water velocity was 0.1 m/s. The PCM-only system contained no aluminum frame or liquid channels; instead, all the available volume between the batteries was filled with PCM. The transient maximum-temperature histories of the module are presented in terms of the final values. At the end of the 900-s discharge, the PCM-only system yielded a maximum battery temperature of 41.98°C and a maximum in-module temperature difference of 3.50°C. In contrast, the PCM-MCHS design lowered the maximum temperature to 39.83°C and reduced the maximum temperature difference to 2.95°C. The temperature reduction reached 2.15°C.

The advantage of the hybrid system is also reflected by the PCM liquid fraction. At the end of discharge, the PCM in the passive-only module was almost fully melted, with a liquid fraction of 0.90. For the PCM-MCHS module, the liquid fraction was only 0.33, because a large proportion of the heat was continually removed by the flowing water. As a result, the composite PCM retained a significant amount of latent heat capacity, which provides a safety margin if the liquid circulation is interrupted. The temperature distribution of the high-voltage battery in the PCM-MCHS design was more uniform, with the maximum temperature located near the outlet region of the coolant and the minimum at the inlet side.

3.2 Effect of the number of microchannels

In the second step, I kept the total mass flow rate constant and varied the number of microchannels by using Case 1 (63 channels), Case 2 (111 channels), and Case 3 (174 channels). To maintain the same total flow rate, the inlet flow velocities were 0.1 m/s for Case 1, 0.0567 m/s for Case 2, and 0.0362 m/s for Case 3. Figure 3 summarizes the simulated maximum temperatures and maximum temperature differences for these three configurations after a 4C discharge at 35°C. The maximum temperatures were 39.83, 39.82, and 40.09°C for Case 1, Case 2, and Case 3, respectively. Thus, increasing the number of channels from 63 to 111, while keeping the flow rate constant, produced almost no improvement in the peak temperature, whereas further increasing the number to 174 slightly degraded the peak-temperature control because the per-channel velocity became too low. The maximum temperature differences were 2.94, 3.14, and 3.51°C for the three respective cases. The increase in temperature difference with more channels can be attributed to a more pronounced low-temperature region near the inlet side, whereas the core of the battery was less affected by the reduced coolant velocity. Based on these results, I retained Case 1 as the most favorable microchannel configuration for the cooling task.

Table 3. Effect of microchannel number on cooling performance at equal total mass flow rate.
Configuration Number of channels Inlet velocity (m/s) Maximum temperature (°C) Maximum ΔT (°C) PCM liquid fraction
Case 1 63 0.1000 39.83 2.94 0.33
Case 2 111 0.0567 39.82 3.14 0.24
Case 3 174 0.0362 40.09 3.51 0.25

3.3 Effect of coolant velocity

Keeping Case 1 and a composite PCM with 5% graphite, I examined inlet water velocities of 0.05, 0.1, and 0.2 m/s. As the velocity increases, the forced convection coefficient and the effective heat-removal capacity improve, but the marginal benefit becomes smaller. At the end of the 4C discharge, the maximum battery temperatures were 40.25, 39.83, and 39.63°C for the three velocities, respectively. A further increase from 0.1 to 0.2 m/s reduced the peak temperature by only 0.2°C, while doubling the pumping power. For this reason, I selected 0.1 m/s as the practical cooling-flow velocity because it offers a good compromise between thermal performance and parasitic energy consumption. The results are listed in Table 4, together with the corresponding PCM liquid fractions at the end of discharge.

Table 4. Effect of coolant velocity on cooling performance for Case 1.
Coolant velocity (m/s) Maximum temperature (°C) Maximum ΔT (°C) PCM liquid fraction
0.05 40.25 — 0.45
0.1 39.83 2.94 0.33
0.2 39.63 — 0.20

3.4 Effect of graphite mass fraction in PCM

To investigate the influence of PCM thermal conductivity on cooling, I simulated seven different graphite concentrations in the composite PCM at a fixed coolant velocity of 0.1 m/s. Higher graphite fractions lead to a higher thermal conductivity of the PCM, which accelerates the heat transfer from the battery surface to the aluminum frame and eventually to the water. Consequently, the maximum temperature of the high-voltage battery decreases with increasing graphite fraction. However, the temperature uniformity becomes slightly worse because the outer region of the battery pack is cooled more efficiently than the inner core. The total difference in maximum temperature between 0% and 20% graphite was approximately 0.23°C, while the temperature difference remained lower than 3.5°C even for the 20% case. Furthermore, the liquid fraction at the end of discharge decreases with an increase in graphite fraction, meaning that a larger portion of the latent-heat capacity remains available as a safety reserve.

Although a graphite fraction higher than 5% yields further improvements in cooling, the gains are progressively smaller and the electrical conductivity of the composite PCM also increases, which could create a short-circuit risk in a high-voltage battery pack. Therefore, I chose 5% graphite as the optimal composition. At this concentration, the maximum temperature of the high-voltage battery is below 40°C, and the temperature difference is less than 3°C, while the latent heat of the PCM remains sufficiently large.

3.5 Effect of flow arrangement

In the baseline Case 1 layout, all microchannels are connected in parallel, with the coolant flowing from the top inlet to the bottom outlet. This produces a gradually increasing coolant temperature along the vertical direction, which in turn increases the battery temperature near the outlet. To improve the temperature uniformity, I tested a counter-flow arrangement, where some channels guide the water from the top to the bottom while adjacent channels guide the water in the opposite direction. Using a coolant speed of 0.1 m/s and a 5% graphite composite PCM, the counter-flow configuration resulted in a maximum temperature of 39.80°C, which is only slightly lower than the 39.83°C obtained with parallel flow. More importantly, the maximum temperature difference was reduced from 3.15°C in the parallel-flow arrangement to 2.77°C in the counter-flow arrangement. The liquid fraction rose from 0.25 to 0.29, indicating that the PCM was used more effectively because the thermal load was distributed more evenly across the module. The improved temperature distribution can be seen in the more uniform contours of the frame and the battery surfaces. I therefore recommend the counter-flow arrangement for better thermal uniformity of high-voltage battery packs.

4. Preheating Performance at Low Ambient Temperatures

4.1 Evaluation metrics

In addition to cooling, the preheating function of the BTMS is crucial for the reliable operation of high-voltage batteries in winter. I evaluated the preheating performance by measuring the time required for the coldest point on the battery surface to reach 20°C from a low initial temperature. A higher rate of temperature rise (RTR) indicates a faster preheat. The RTR is defined as:

$$RTR = \frac{20\,^{\circ}\mathrm{C} – (-20\,^{\circ}\mathrm{C})}{\tau_{20}}$$

where \(\tau_{20}\) is the time required for the minimum battery temperature to reach 20°C. The ambient temperature, backflow temperature, and initial temperature were all set to -20°C, and the inlet preheating fluid temperature was kept at 25°C unless otherwise stated. At the end of the preheating process, the temperature difference inside the high-voltage battery module was also monitored. A well-designed preheat should not only be fast but also maintain a reasonably small temperature spread.

4.2 Passive insulation behavior

One of the advantages of using PCM in a high-voltage battery pack is that it can act as a thermal buffer during inactive periods. I investigated the insulation performance by setting the initial temperature of the module to 50°C and the ambient temperature to -10°C. The module was left idle for 10 h, and the maximum battery temperature was monitored for different graphite fractions and for a module without PCM. Without PCM, the battery cooled below 0°C within 1 hour. The presence of PCM significantly prolonged the time during which the battery temperature stayed above 0°C, because the heat released by PCM solidification slowed down the cooling process. As the graphite fraction increased, the effective thermal conductivity increased, and thus the insulation time became shorter. Pure paraffin offered the longest insulation time, while a 20% graphite composite permitted faster heat loss. These findings suggest that the PCM filler naturally endows the high-voltage battery pack with a certain cold-weather insulation capability, which can reduce the energy consumed by the preheating system.

4.3 Effect of graphite mass fraction on preheating

I then studied the effect of the composite PCM composition on the forced preheating process. For this purpose, I used the Case 1 microchannel geometry and an inlet velocity of 0.1 m/s with the flow directed from top to bottom. Table 5 lists the preheating time, the corresponding RTR, and the maximum temperature difference in the module for graphite mass fractions between 0% and 20%.

Table 5. Preheating performance as a function of graphite mass fraction in the PCM.
Graphite fraction (%) Preheating time (s) RTR (°C/min) Maximum ΔT (°C)
0 405 5.93 2.9
2 360 6.67 3.1
5 272 8.82 3.3
8 256 9.37 3.5
10 250 9.60 3.6
15 240 10.00 3.8
20 235 10.21 3.9

As the graphite fraction increased from 0% to 5%, the preheating time decreased sharply from 405 s to 272 s. Beyond 5%, the gain diminished: a further increase from 5% to 20% reduced the time by only 37 s. The temperature difference increased gradually with higher graphite content, because the thermal pathway at the inlet side was enhanced and the cell temperature near the inlet rose faster than that near the outlet. Considering the trade-off between preheating speed, temperature uniformity, manufacturing cost, and safety, I again selected the composite PCM with 5% graphite as the best compromise for the preheating scenario.

4.4 Effect of preheating-fluid velocity

Using the Case 1 geometry and the 5% graphite PCM, I varied the inlet preheating water velocity from 0.05 to 1 m/s. The results in Table 6 demonstrate that the preheating time drops quickly as the velocity increases from 0.05 to 0.25 m/s. At 0.25 m/s, the preheat takes 211 s, which is 169 s shorter than at 0.05 m/s. However, when the velocity is further increased to 1 m/s, the preheating time is only reduced to 169 s. In other words, the marginal effect becomes small above 0.25 m/s. At the same time, the maximum in-module temperature difference grows with velocity because the strong convection creates a hotter inlet region and an insufficiently warmed outlet region. Since a very high flow speed also leads to higher pumping losses, I chose 0.25 m/s as the optimal preheating velocity.

Table 6. Preheating performance as a function of the inlet fluid velocity.
Velocity (m/s) Preheating time (s) RTR (°C/min) Maximum ΔT (°C)
0.05 380 6.32 2.1
0.1 272 8.82 3.3
0.25 211 11.37 3.5
0.5 186 12.90 4.2
1 169 14.20 4.8

4.5 Effect of the number of microchannels for preheating

I also evaluated the preheating performance of the three different channel geometries under iso-flow conditions. Since the total flow rate was kept constant, the inlet velocities were adjusted to 0.25, 0.1418, and 0.0905 m/s for Case 1, Case 2, and Case 3, respectively. The preheating time decreased as the number of channels increased, because a larger heat-transfer area improves the overall convective exchange even when the per-channel velocity is lower. Specifically, the preheating time was 211 s for Case 1, 192 s for Case 2, and 181 s for Case 3. However, the corresponding maximum temperature differences were 3.32, 3.65, and 4.50°C. The larger temperature difference in Case 3 arises because the entrance region covers a greater share of the battery perimeter, producing a hotter upper band and a colder bottom section. Since Case 2 provides a reasonable preheating time with an acceptable temperature difference, I used Case 2 as the base for further flow-direction studies.

Table 7. Preheating performance of different microchannel configurations with equal total flow rate.
Configuration Number of channels Inlet velocity (m/s) Preheating time (s) RTR (°C/min) Maximum ΔT (°C)
Case 1 63 0.25 211 11.37 3.32
Case 2 111 0.1418 192 12.50 3.65
Case 3 174 0.0905 181 13.26 4.50

4.6 Effect of counter-flow versus parallel-flow arrangement

Using Case 2 with its nominal inlet velocity of 0.1418 m/s and the 5% graphite PCM, I compared the parallel-flow and counter-flow configurations. The preheating time for the parallel-flow scheme was 192 s; for the counter-flow scheme it was 194 s. Thus, the counter-flow scheme slightly lengthens the required warm-up time. On the other hand, the final module temperature difference decreased from 3.65°C in the parallel-flow case to 3.25°C in the counter-flow case. The improved temperature uniformity is visually evident from the cross-sectional contours of the aluminum frame and the cells. This demonstrates that although the counter-flow arrangement offers a negligible benefit for the average heating rate, it significantly improves the thermal homogeneity of a high-voltage battery pack. Since a uniform temperature field is essential for battery longevity and avoiding localized degradation, the 2-s penalty can be accepted. Therefore, the counter-flow arrangement is recommended for both cooling and preheating operations.

4.7 Effect of preheating-fluid temperature and ambient temperature

Finally, I analyzed the effect of the preheating-fluid temperature for four different ambient temperatures. I selected four preheating-fluid temperatures: 25, 35, 40, and 45°C, and simulated the preheating process at ambient temperatures of -40, -30, -20, and -10°C. Not surprisingly, both a higher fluid temperature and a higher ambient temperature shorten the preheating time. For example, at an ambient temperature of -40°C, increasing the inlet fluid temperature from 25°C to 35°C reduced the preheating time from 223 s to 146 s, whereas raising the fluid temperature from 35°C to 45°C only reduced the time from 146 s to 119 s. The effectiveness of raising the fluid temperature diminishes above 37°C, which is the melting point of the composite PCM. When the fluid temperature is above the phase-change temperature, part of the PCM begins to melt and absorbs a significant amount of latent heat without a corresponding temperature rise. This latent-heat effect buffers the battery temperature increase and therefore weakens the influence of higher fluid temperature on the heating rate. The maximum temperature difference also grows with the fluid temperature. At an ambient temperature of -20°C, the in-module temperature difference exceeded 11°C for an inlet fluid temperature of 35°C and reached roughly 14°C at 40°C, and around 16°C at 45°C. Once the phase-change region is formed, the presence of the two-phase zone keeps the intermediate layer near 37°C, which limits the further deterioration of the temperature difference. Nevertheless, from a safety standpoint, a preheating-fluid temperature of 25 to 35°C is sufficient for normal cold-weather operation.

5. Overall Performance Summary and Discussion

To summarize the complete thermal-management capability of the proposed high-voltage battery system, I evaluated both the cooling and preheating scenarios using a combined set of optimal parameters. The recommended configuration is a microchannel geometry with 111 channels (Case 2) when both cooling and preheating are considered, a composite PCM with 5% graphite, a counter-flow arrangement, a cooling flow velocity of 0.1 m/s, and a preheating flow velocity of 0.25 m/s. This parameter set offers the best compromise among maximum temperature, temperature uniformity, energy consumption, and safety.

During the high-temperature cooling scenario at 35°C ambient and 4C discharge, the high-voltage battery maximum temperature can be kept below 40°C, and the maximum difference between cells remains below 3°C. The proposed hybrid BTMS lowers the maximum temperature by approximately 1.54°C compared with an optimized single-mode PCM cooling system and improves the temperature uniformity. The heat-removal capability is enhanced by the microchannel water flow, while the PCM acts as a thermal flywheel that absorbs transient heat spikes and prevents rapid temperature escalation at the beginning of the discharge.

During the cold-preheating scenario at an ambient temperature of -20°C, the preheating time to bring the high-voltage battery from -20°C to 20°C is about 192 s with the recommended Case 2 layout, corresponding to an RTR of 12.5°C/min. This is more than 30% faster than a baseline with a higher-viscosity coolant or a less conductive PCM. The temperature difference after preheating can be controlled below 3.5°C in the counter-flow arrangement, which is approximately a 20% improvement over the parallel-flow design. The PCM also provides passive insulation when the vehicle is parked, and the insulation time can be adjusted by choosing an appropriate graphite fraction. Higher graphite content increases the cooling and preheating rates, but reduces the passive insulation time and raises the short-circuit risk; therefore, the 5% graphite loading is the most balanced choice.

These findings demonstrate that a single multi-functional BTMS architecture can cover the entire spectrum of high-voltage battery thermal-management needs. The honeycomb-inspired microchannel layout allows the coolant to access all sides of the hexagon-filled module, while the PCM filling suppresses temperature gradients and offers latent-heat buffering. The counter-flow design produces a symmetric thermal boundary condition that equalizes the temperature on both the top and bottom sides of the module. The parametric study reveals important trade-offs. For example, adding more channels can accelerate preheating but at the expense of temperature uniformity. Increasing the flow velocity improves heat transfer but follows a law of diminishing returns. Increasing the graphite fraction in the PCM is beneficial only up to a critical point, beyond which the thermal-conductivity enhancement is outweighed by the loss of latent heat and an increase in electrical conductivity.

6. Conclusions and Future Perspectives

In this research, I have designed and numerically assessed a microchannel-PCM coupled BTMS for high-voltage battery packs. The investigation covered both hot and cold operating conditions, and I derived the following conclusions:

  1. The proposed hybrid BTMS is superior to a single-mode PCM cooling system. For a 4C discharge in a 35°C environment, the maximum battery temperature and the temperature difference are reduced by 2.15°C and 0.55°C, respectively, while the PCM liquid fraction remains low enough to guarantee a large safety margin.
  2. For the cooling function, the best microchannel layout is Case 1 (63 channels) at a flow velocity of 0.1 m/s when considering heat-transfer performance and pumping power. A larger number of microchannels does not necessarily improve cooling performance if the total mass flow rate is held constant.
  3. The optimal graphite mass fraction in the composite PCM is 5%. Increasing the graphite concentration beyond 5% offers diminishing returns for both cooling and preheating and may compromise the electrical insulation of the high-voltage battery system.
  4. A counter-flow arrangement improves the temperature uniformity of the high-voltage battery module in both cooling and preheating modes. It reduces the cooling temperature difference by about 0.4°C and the preheating temperature difference by about 0.4°C, at the cost of a negligible 2-s increase in preheating time.
  5. For low-temperature preheating, an inlet velocity of 0.25 m/s is the optimum. The preheating time at -20°C ambient with a 25°C inlet water temperature is 211 s for the Case 1 geometry, and 192 s for the Case 2 geometry. The counter-flow Case 2 geometry yields an RTR of approximately 12.4°C/min and a final temperature difference of 3.25°C.
  6. The PCM provides valuable passive insulation. When the high-voltage battery module is allowed to cool from 50°C to -10°C ambient, the PCM delays the cooling rate significantly. A lower graphite fraction improves the insulation duration, which is favourable for a parked vehicle.
  7. The temperature of the preheating fluid should be chosen carefully. A higher fluid temperature accelerates warm-up but causes a dramatic increase in the temperature spread, especially when the fluid temperature exceeds the PCM melting point. For most practical purposes, a preheating-fluid temperature between 25°C and 35°C is recommended.

Although the present numerical study provides valuable optimised design guidelines, it also has some limitations. First, the battery heat-generation rate was simplified by using a uniform volumetric heat source, whereas in reality the current and heat generation vary dynamically with state of charge, state of health, and cell age. In future work, an electro-thermal-aging coupled model should be employed to capture dynamic effects. Second, the multi-parameter optimization was carried out by a one-factor-at-a-time method. To further improve the overall performance of the high-voltage battery BTMS, advanced multi-objective optimization algorithms such as response surface methods, genetic algorithms, or machine-learning-based surrogate models should be applied. Third, I assumed that the thermal properties of the PCM remained isotropic and homogeneous; however, in practice, the graphite/paraffin composite may exhibit orientation-dependent thermal conductivity, and the natural convection in the molten PCM would affect the phase-change process. Future research should include a more detailed model of PCM convection and thermal anisotropy.

Notwithstanding these limitations, the simulation results confirm that a coupled microchannel-PCM system is an effective and flexible solution for high-voltage battery thermal management. The integrated design supports three major functions: efficient heat dissipation at high temperature, fast preheating at low temperature, and passive thermal insulation during idle periods. The quantitative parametric analysis helps engineers to select appropriate channel layouts, flow rates, flow directions, and PCM compositions for a high-voltage battery module. These findings can be extended to larger battery packs by scaling the number of cells and by optimizing the flow distribution, and they are especially relevant for the rapidly growing market of high-voltage electric vehicles where safety and reliability under extreme climates are critical.

In conclusion, the honeycomb-inspired microchannel-PCM coupled BTMS proposed in my work presents a promising pathway towards a robust, energy-efficient, and thermally safe high-voltage battery system. Future dynamic driving-cycle simulations and experimental prototype testing will enable further refinements and real-world deployment. This research offers a step forward in the design of advanced battery thermal management solutions to support the expanded adoption of high-voltage battery technologies in clean transportation and grid-level energy storage systems.

Scroll to Top