Bidirectional Counter-Flow Cooling for Energy Storage Battery

I present a numerical investigation into the thermal management of grid-scale energy storage battery packs under high-rate charge and discharge conditions. The increasing penetration of renewable energy requires energy storage battery systems to handle abrupt power imbalances, often leading to high C-rate operations. In this study, I propose a bidirectional counter-flow liquid cooling plate that is placed on the large side surfaces of each battery cell. The cooling performance of this novel configuration is systematically compared with a conventional bottom-placed single-direction cooling plate and a side-placed single-direction cooling plate. Simulations are carried out at 1C, 3C, and 5C charge/discharge rates. The results demonstrate that the bidirectional counter-flow heat exchange plate simultaneously satisfies the requirements for maximum temperature and temperature uniformity under high C-rate conditions. The findings confirm that the proposed design can effectively mitigate thermal risks and improve the service life of energy storage battery packs.

1. Introduction

Renewable energy sources such as wind, solar, hydro, and biomass are playing an increasingly important role in the transition toward a low-carbon energy system. However, these sources are intermittent and geographically distributed, which causes fluctuations in grid frequency and voltage. Energy storage battery systems have emerged as a key technology to smooth these fluctuations through peak shaving and valley filling. Among various storage technologies, lithium-ion based energy storage battery packs are widely adopted because of their modular design, fast response, and flexible deployment.

During operation, lithium-ion energy storage battery cells generate heat due to internal resistance and electrochemical reactions. If this heat is not removed effectively, the cell temperature rises, which accelerates degradation and may eventually lead to thermal runaway. More importantly, temperature gradients within a battery pack cause uneven aging among cells. It is widely accepted that the optimal operating window for lithium-ion energy storage battery cells is between 293 K and 313 K, with a maximum allowable temperature difference of 5 K across the pack. Exceeding this temperature difference can reduce the cycle life by more than 30%. A temperature above 393 K is considered a high-risk threshold because the lithium intercalated in graphite reacts with the electrolyte and binder, possibly causing fire or explosion.

In grid applications, serious challenges arise during sudden increases in renewable generation or peak load demands. The energy storage battery must charge or discharge at a high rate to stabilize the grid. High-rate operation leads to intense heat generation and rapid temperature rise. Conventional air cooling and phase-change cooling are often insufficient to handle such extreme conditions. Liquid cooling, by contrast, offers a much higher heat transfer coefficient and is therefore recognized as an effective solution for energy storage battery thermal management.

Many previous studies have investigated liquid-cooled plates with serpentine channels, parallel straight channels, and tree-like fractal structures. Those designs are usually optimized for normal charging rates, and their performance under high C-rate conditions remains inadequate. The main issue is that a typical liquid cooling plate attached only to the bottom of the cell cannot extract heat uniformly from the cell because of the anisotropic thermal conductivity of the electrode stack and because of the long heat conduction path from the top to the bottom. Even when the cooling plate is moved to the large side surface, the cooling liquid warms up along its flow path, creating a persistent temperature gradient in the flow direction. Such gradients violate the 5 K requirement during high-rate operation of energy storage battery packs.

To overcome these limitations, I design a bidirectional counter-flow heat exchange plate that contains two sets of channels with opposite flow directions. The two inlet ports are placed at the bottom while the two outlet ports are placed at the top. This configuration produces a reverse thermal compensation effect that flattens the temperature distribution across the cell. In addition, an air gap is introduced between neighboring cooling plates to reduce thermal interference, further improving the temperature uniformity. The proposed scheme is particularly beneficial for grid-scale energy storage battery systems that must operate safely under fluctuating charge and discharge rates.

The remainder of this article is organized as follows. Section 2 describes the geometric configurations and the mathematical model. Section 3 presents the simulation results for three different cooling strategies under various C-rates. Section 4 provides a comparative analysis of the heat transfer mechanisms. Section 5 summarizes the main conclusions and their practical implications for energy storage battery thermal management.

2. Numerical Model

2.1 Geometric Configurations

I constructed a three-dimensional model of a battery pack consisting of ten prismatic lithium iron phosphate cells connected in series. Each cell has a nominal capacity of 135 Ah, a nominal voltage of 3.2 V, and dimensions of 945 mm × 14 mm × 90 mm. The density is 1715 kg/m³, the specific heat capacity is 1100 J/(kg·K), and the anisotropic thermal conductivity is 18.3 W/(m·K) in the tangential directions and 1.1 W/(m·K) in the normal direction. This anisotropy is caused by the layered electrode-separator structure inside the cell.

Three cooling schemes are considered. In Scheme 1, a conventional single-direction straight-channel cooling plate is placed underneath the bottom surface of all cells. The coolant enters from one side and leaves from the other side, so the flow is unidirectional. In Scheme 2, the same type of single-direction cooling plate is installed between the large side surfaces of adjacent cells. In Scheme 3, my proposed bidirectional counter-flow cooling plate is installed on the large side surfaces of each cell. The bidirectional plate contains two independent channel groups. The coolant enters through two bottom ports and leaves through two top ports, forming a cross-counter flow pattern. A thin air gap is included between the cooling plate and the adjacent cell in Scheme 3 to minimize unnecessary heat exchange between neighboring cells. In all schemes, a thermally conductive silicone pad is applied between the cooling plate and the cell to reduce contact resistance. The main geometric parameters of the cells and cooling plates are summarized in Table 1.

Table 1: Physical and geometrical parameters of the energy storage battery cell and cooling plate
Parameter Value
Cell capacity (Ah) 135
Nominal voltage (V) 3.2
Cell dimensions (mm) 945 × 14 × 90
Cell density (kg/m³) 1715
Cell specific heat capacity (J/(kg·K)) 1100
In-plane thermal conductivity (W/(m·K)) 18.3
Through-plane thermal conductivity (W/(m·K)) 1.1
Coolant type 50% ethylene glycol solution
Coolant inlet temperature (K) 293.15
Total coolant flow rate (L/min) 12

2.2 Mathematical Model

The heat generation inside the lithium-ion energy storage battery cell is calculated using the Bernardi equation. The total volumetric heat generation rate is expressed as

$$ q = \frac{I}{V_b} \left[ \left( E_0 – U \right) – T \frac{dE_0}{dT} \right] = \frac{1}{V_b} \left( I^2 R – I T \frac{dE_0}{dT} \right) \tag{1} $$

where \( V_b \) is the cell volume, \( E_0 \) is the open-circuit voltage, \( U \) is the terminal voltage, \( T \) is the cell temperature, \( I \) is the applied current, and \( R \) is the internal resistance. The first term on the right-hand side represents irreversible ohmic heating, while the second term accounts for reversible entropic heating. At high C-rates, the ohmic term dominates because it is proportional to \( I^2 \).

The anisotropic three-dimensional heat conduction inside the cell is governed by

$$ \rho C_p \frac{\partial T}{\partial t} = \lambda_x \frac{\partial^2 T}{\partial x^2} + \lambda_y \frac{\partial^2 T}{\partial y^2} + \lambda_z \frac{\partial^2 T}{\partial z^2} + Q \tag{2} $$

where \( \rho \) is the density, \( C_p \) is the specific heat capacity, \( \lambda_x,\lambda_y,\lambda_z \) are the thermal conductivities in the three orthogonal directions, and \( Q \) is the volumetric heat generation rate. Because the cell is thin, the heat conduction in the thickness direction is much weaker than in the other two directions, which makes the top-to-bottom temperature management particularly challenging.

For the fluid domain, the continuity, momentum, and energy equations are solved. The continuity equation is

$$ \frac{\partial \rho_w}{\partial t} + \nabla \cdot \left( \rho_w \mathbf{v} \right) = 0 \tag{3} $$

where \( \rho_w \) is the coolant density and \( \mathbf{v} \) is the velocity vector. The momentum equation reads

$$ \frac{\partial \mathbf{v}}{\partial t} + \left( \mathbf{v} \cdot \nabla \right)\mathbf{v} = -\frac{\nabla p}{\rho_w} + \frac{\mu}{\rho_w} \nabla^2 \mathbf{v} + \mathbf{g} \tag{4} $$

where \( p \) is pressure, \( \mu \) is dynamic viscosity, and \( \mathbf{g} \) is gravitational acceleration. The energy equation for the coolant is

$$ \frac{\partial}{\partial t}\left( \rho_w C_{p,w} T_w \right) + \nabla \cdot \left( -k_w \nabla T_w \right) + \rho_w C_{p,w} \mathbf{v} \cdot \nabla T_w = 0 \tag{5} $$

where \( C_{p,w} \) is the coolant specific heat capacity, \( T_w \) is the coolant temperature, and \( k_w \) is the coolant thermal conductivity. The energy equation for the solid cooling plate is

$$ \frac{\partial}{\partial t}\left( \rho_n C_{p,n} T_n \right) + \nabla \cdot \left( -k_n \nabla T_n \right) = 0 \tag{6} $$

in which \( \rho_n, C_{p,n}, T_n, k_n \) are the density, specific heat capacity, temperature, and thermal conductivity of the plate material.

The boundary conditions are specified as follows. The initial temperature of the energy storage battery cell and the ambient temperature are both set to 298 K. The coolant inlet is set as a velocity inlet with a constant temperature of 293.15 K and a total flow rate of 12 L/min. The coolant outlet is set as a pressure outlet. The heat generation rates corresponding to 1C, 3C, and 5C charge/discharge conditions are applied as volumetric heat sources in the cell domain. A steady-state solver is used to evaluate the long-term thermal behavior under continuous operation.

For the numerical simulation, the computational domain is discretized using Fluent Meshing. To ensure accuracy near the thermal critical interfaces, local mesh refinement is applied to the silicone pads, the cooling plates, the air gap, and the fluid domains. Six different mesh densities are selected to conduct a grid independence study. The pressure drop between the inlet and outlet is used as the monitoring parameter. Table 2 lists the grid counts and the corresponding pressure drops.

Table 2: Grid independence verification
Number of cells (million) Inlet-outlet pressure drop (kPa)
1.33 13.60
2.36 13.90
2.89 14.20
3.16 14.80
4.23 14.90
5.13 14.93

The variation in pressure drop becomes negligible when the mesh size exceeds 4.23 million cells. Therefore, a mesh with 4.23 million cells is selected for all subsequent simulations to balance numerical accuracy and computational efficiency.

3. Results and Discussion

3.1 Performance at 1C Charge/Discharge Rate

At a conventional 1C rate, the heat generation is relatively moderate. I first examine the temperature distribution of the energy storage battery pack using Scheme 1, where the cooling plate is attached to the bottom of the pack. The simulation shows that the cell temperature ranges from 294.5 K to 301 K. The maximum temperature is 301 K, which is below the upper limit of the ideal operating window. However, the maximum temperature difference across the pack is 6.5 K, which exceeds the recommended threshold of 5 K. This is because the cooling plate only contacts the bottom surface of the cells. Heat generated near the top of the cell must travel through the low through-plane thermal conductivity of the electrode stack before reaching the cooled bottom surface. The long heat conduction path creates a significant top-to-bottom thermal gradient. In contrast, the variation along the coolant flow direction is relatively small, confirming that the thermal bottleneck is in the thickness direction of the cell.

In Scheme 2, the single-direction cooling plate is relocated to the large side surface of the cell. The contact area is significantly increased, and the heat conduction path from the center of the cell to the cooling surface is shortened. As a result, the maximum temperature drops to 294.7 K and the maximum temperature difference is only 1.5 K. The temperature distribution now becomes much more uniform, with the highest temperature located in the middle region of the cell and the lowest temperatures near the side surfaces. This configuration satisfies both the maximum temperature and the temperature uniformity constraints for energy storage battery at 1C operation.

Scheme 3 employs the proposed bidirectional counter-flow plate on the side surfaces. At 1C, the maximum temperature is 294 K and the maximum temperature difference is only 1 K. The bidirectional flow channels provide a more even coolant temperature distribution compared with the unidirectional flow. Moreover, the air gap between the cooling plate and the adjacent cell prevents cross-interference, which further improves the uniformity. Thus, Scheme 3 outperforms the other two schemes even under conventional conditions. Table 3 summarizes the results for all three schemes at 1C.

Table 3: Thermal performance comparison at 1C
Scheme Maximum temperature (K) Maximum temperature difference (K)
1: Bottom cooling 301.0 6.5
2: Side unidirectional cooling 294.7 1.5
3: Side bidirectional counter-flow cooling 294.0 1.0

3.2 Performance at 3C Charge/Discharge Rate

High-rate operation is the most relevant scenario for grid-scale energy storage battery reserves that must respond quickly to power fluctuations. At 3C, the heat generation rate is nine times larger than at 1C, which imposes severe thermal stress on the cooling system.

For Scheme 1, the simulated maximum temperature reaches 357 K. This is dangerously close to the thermal runaway threshold of 393 K, and the maximum temperature difference between the highest and lowest cell temperatures is as large as 52 K. Such a large gradient will inevitably cause accelerated capacity fade and may create local hot spots where electrolyte decomposition can initiate. Clearly, the bottom cooling configuration is inadequate for high C-rate operation of a large-format energy storage battery.

Scheme 2 significantly improves the cooling effectiveness because the side-mounted plate has a much larger heat transfer area. The maximum temperature at 3C is 307 K, which is below the acceptable limit for lithium-ion cells. However, the temperature difference across the battery pack is about 13 K, far exceeding the 5 K requirement. The temperature pattern shows a clear increasing trend in the coolant flow direction. The cells near the outlet are hotter than those near the inlet because the coolant absorbs heat along its path. Even though the total flow rate is the same as in Scheme 1, the lack of counter-flow compensation results in a persistent thermal gradient. This gradient would lead to inhomogeneous aging of the energy storage battery pack and reduce its overall reliability.

Scheme 3, the bidirectional counter-flow plate, yields a maximum temperature of only 299 K and a maximum temperature difference of 4.8 K. Both metrics satisfy the strict criteria for high-rate energy storage battery operation. The unique reverse-flow arrangement allows the cooling liquid in one channel to exchange heat with the liquid in the opposite channel. Regions that would otherwise become hotter because of the coolant warming are balanced by cooler liquid entering from the opposite direction. The resulting temperature distribution is remarkably uniform, with the hottest region located in the central area of the cell and the coolest region near the side surfaces, as expected for a well-designed side cooling approach. The air gap helps isolate adjacent cells and prevents thermal coupling, which is particularly beneficial when a single cell experiences higher heat generation than its neighbors.

To provide a quantitative view, Table 4 lists the simulated maximum temperature and maximum temperature difference for all three schemes at 3C.

Table 4: Thermal performance comparison at 3C
Scheme Maximum temperature (K) Maximum temperature difference (K)
1: Bottom cooling 357.0 52.0
2: Side unidirectional cooling 307.0 13.0
3: Side bidirectional counter-flow cooling 299.0 4.8

3.3 Performance at 5C Charge/Discharge Rate

To assess the robustness of the proposed design under extreme grid events, I also simulate a 5C charge/discharge condition. Such a high rate is generally not sustained for long periods, but it may occur during voltage sag compensation or frequency regulation in a weak grid.

For Scheme 1, the temperature distribution ranges from 363 K to 470 K. The maximum temperature of 470 K is far above the 393 K thermal runaway threshold. In this situation, the energy storage battery will almost certainly undergo thermal runaway, potentially leading to fire or explosion. The bottom cooling plate fails completely to control the heat accumulation because the heat transfer rate from the cell to the cooling plate is limited by the low through-plane thermal conductivity and the small contact area.

Scheme 2 lowers the maximum temperature to 332 K, which is still significantly above the recommended operating range of 293–313 K. The maximum temperature difference is 39 K. Although thermal runaway is not immediately triggered, prolonged operation at 5C would cause severe capacity degradation and possibly internal short circuits due to uneven electrode expansion. The large temperature gradient also induces differential states of charge among parallel-connected cells, leading to current imbalance and further localized heating.

Scheme 3 maintains the maximum temperature at 308 K and the maximum temperature difference at 14 K. Although the temperature difference exceeds the ideal 5 K limit, it is important to note that a 5C rate is an emergency condition that occurs only for a short duration. The maximum temperature remains well below the risky threshold of 393 K, so the energy storage battery pack can safely operate without immediate danger. The bidirectional counter-flow plate effectively prevents hot spots and limits overheating even under such extreme thermal load. If the total coolant flow rate were temporarily increased or the inlet temperature lowered, the temperature difference could be further reduced. Table 5 lists the results at 5C.

Table 5: Thermal performance comparison at 5C
Scheme Maximum temperature (K) Maximum temperature difference (K)
1: Bottom cooling 470.0 107.0
2: Side unidirectional cooling 332.0 39.0
3: Side bidirectional counter-flow cooling 308.0 14.0

3.4 Comparison of Coolant Temperature Distribution

To explain the superior performance of the bidirectional counter-flow concept, I analyze the coolant temperature distribution inside the cooling plates at 3C. In Scheme 1, the coolant temperature rises by about 15 K from the inlet to the outlet. The temperature is highest in the center of the plate and lower near the edges. This adverse distribution reduces the local heat transfer capability near the outlet and creates a non-uniform cooling effect. In Scheme 2, the temperature rise is about 10 K along the flow direction. The side plate has a larger width, which improves the thermal contact, but the monotonic rise in coolant temperature still causes a directional temperature gradient in the cell. In Scheme 3, the temperature variation across the entire bidirectional plate is only about 1 K. The two opposing streams exchange heat with each other so that the coolant temperature remains nearly uniform throughout the plate. This self-compensating mechanism is analogous to a counter-flow heat exchanger, where the temperature difference between the two streams is maintained as small as possible. The near-isothermal surface of the cooling plate is the key factor that enables the energy storage battery pack to achieve a temperature difference below 5 K at 3C.

The pressure drop of the bidirectional plate is slightly higher than that of the single-direction plate because of the double-channel entry and exit arrangement. Nevertheless, the pressure drop remains acceptable for the given flow rate. The total pressure drop for all schemes is approximately 14.9 kPa, as shown in Table 2. A low pressure drop is beneficial because it reduces the parasitic pumping power. The proposed design does not require any additional flow rate or a lower coolant temperature, which makes it easy to retrofit into existing energy storage battery thermal management systems.

4. Heat Transfer Mechanism Analysis

The main heat transfer path inside the energy storage battery cell is from the internal heat generation regions to the side surfaces where the cooling plate is attached. The cell thickness is only 14 mm, but the through-plane thermal conductivity is as low as 1.1 W/(m·K). Therefore, the temperature difference between the center of the cell and the side surface is controlled by the heat flux and the thermal resistance. When the cooling plate is placed on only one side, the heat flows primarily in one direction, and the temperature gradient becomes large. When the cooling plate is placed on both sides or when the cell is clamped between two cooling plates, the heat conduction path is halved and the temperature gradient is reduced.

The bidirectional counter-flow plate further improves the situation by offering an almost uniform surface temperature. According to the one-dimensional heat conduction equation

$$ \Delta T_{\text{cell}} = \frac{q L^2}{2 k_{\perp}} \tag{7} $$

where \( L \) is the half-thickness of the cell (about 7 mm) and \( k_{\perp} \) is the through-plane thermal conductivity. If the surface temperature is constant, the internal maximum temperature rise is proportional to the volumetric heat generation rate and the square of the thickness. With the side-mounted cooling plates, the effective heat conduction length is reduced compared with bottom cooling. Therefore, the maximum temperature rise inside the cell is significantly reduced. The uniform surface temperature of the bidirectional plate further ensures that the temperature field inside the cell is symmetric and uniform, thereby avoiding hot spots.

The reverse-flow channels in Scheme 3 also enhance convective heat transfer inside the cooling plate. The local Nusselt number can be estimated as

$$ \text{Nu} = \frac{h D_h}{k_w} \tag{8} $$

where \( h \) is the convective heat transfer coefficient and \( D_h \) is the hydraulic diameter. Because the coolant flows in narrow channels at a moderate velocity, the flow is turbulent or transitional, giving a relatively high Nusselt number. The counter-flow arrangement maintains a higher temperature difference between the coolant and the cell surface along the entire length, which improves the effective heat transfer rate compared with a unidirectional configuration where the coolant temperature increases and reduces the local driving temperature difference.

In addition, the air gap between adjacent cooling plates in the bidirectional scheme acts as a thermal insulation layer. It prevents the heat from one cell being transferred to another through the cooling plate itself. In a unidirectional side cooling arrangement, the cooling plate serves as a heat spreader that may unintentionally transfer heat from a hotter cell to a colder cell. This coupling can cause a reversal of the desired heat flow and deteriorate the pack-level temperature uniformity. The air gap breaks this thermal bridge and makes each cell thermally independent, which is desirable for a modular energy storage battery system.

5. Discussion

The simulation results clearly demonstrate that the bottom cooling configuration (Scheme 1) is not suitable for high-C-rate energy storage battery applications. Even at 1C, the temperature difference exceeds 5 K, and at 3C the maximum temperature approaches a dangerous level. The side unidirectional cooling configuration (Scheme 2) is a significant improvement but still suffers from a large temperature gradient along the flow direction. This gradient becomes unacceptable at high C-rates because the coolant temperature rise is amplified by the increased heat load. The side bidirectional counter-flow cooling configuration (Scheme 3) solves this problem through the reverse thermal compensation mechanism. The two sets of channels operate in opposite directions, so the coolant outlet from one channel is near the inlet of another channel. The hot coolant near the outlet transfers heat to the cold coolant near the inlet, effectively smoothing the temperature distribution.

It is worth noting that the maximum temperature difference at 5C for Scheme 3 is 14 K, which is larger than the 5 K criterion. This is because the total heat generation rate at 5C is 25 times higher than at 1C. To maintain the temperature difference below 5 K under such an extreme rate, one could increase the coolant flow rate, decrease the coolant inlet temperature, or use a phase-change material in conjunction with the liquid cooling plate. In practical grid operations, however, a 5C rate is rarely sustained for more than a few minutes. The proposed bidirectional counter-flow plate limits the maximum temperature to 308 K, which is safely below the thermal runaway threshold. Therefore, the design offers an adequate safety margin for emergency operations. For continuous high-rate operation at 3C, the bidirectional plate meets all requirements without any additional modification.

The versatility of the proposed concept makes it attractive for various energy storage battery system configurations. It can also be applied to cylindrical cells or pouch cells by adapting the channel layout. The manufacturing of the bidirectional plate is not significantly more complex than that of a conventional plate, since the two channel layers can be machined as mirror images and bonded together. The additional cost is expected to be modest, while the improvement in thermal uniformity provides a substantial benefit for extending battery life and reducing the risk of thermal runaway.

6. Conclusions

I have conducted a detailed numerical simulation to evaluate the thermal management performance of a bidirectional counter-flow heat exchange plate for grid-scale energy storage battery packs. The proposed scheme was compared with two conventional liquid cooling arrangements: bottom-mounted unidirectional cooling and side-mounted unidirectional cooling. The main conclusions are as follows:

  1. At a conventional 1C rate, the bottom cooling configuration exhibits a maximum temperature difference of 6.5 K, which violates the temperature uniformity requirement of energy storage battery packs. The side-mounted unidirectional and bidirectional configurations both perform well, with maximum temperature differences of 1.5 K and 1.0 K, respectively.
  2. At a high 3C rate, the bottom cooling configuration results in a maximum temperature of 357 K and a temperature difference of 52 K, making it completely unsuitable. The side-mounted unidirectional configuration reduces the maximum temperature to 307 K but still has a 13 K temperature difference. The side-mounted bidirectional counter-flow configuration achieves a maximum temperature of 299 K and a temperature difference of 4.8 K, satisfying both safety and uniformity criteria.
  3. At an extreme 5C rate, the bottom cooling configuration leads to a maximum temperature of 470 K, which is a serious fire hazard. The side-mounted unidirectional configuration reaches 332 K with a 39 K temperature difference. The bidirectional counter-flow plate keeps the maximum temperature at 308 K and the temperature difference at 14 K, providing a safe operational envelope for short-term emergency conditions.
  4. The superior performance of the bidirectional counter-flow design arises from the uniform coolant temperature distribution across the entire cooling surface and the air gap that decouples adjacent cells. The design effectively acts as a counter-flow heat exchanger, balancing the thermal load naturally.
  5. The proposed bidirectional counter-flow heat exchange plate represents a promising technical solution for improving the safety and reliability of grid-scale energy storage battery systems, especially in applications subject to high-rate charging and discharging due to renewable energy variability.

Future work will investigate dynamic operating conditions with variable C-rates and will explore the feasibility of integrating phase-change materials with the bidirectional plate to further enhance its performance under sustained extreme loads. The optimization of channel dimensions and the coolant flow rate will also be studied to achieve the best trade-off between thermal performance and pumping power.

Scroll to Top