The pursuit of higher energy density and greater integration efficiency is a central focus in electric vehicle (EV) battery pack development. Traditional Cell-to-Module (CTM) architectures often result in a volumetric utilization rate of only around 40%, imposing significant limitations on vehicle range and lightweight design. The advent of Cell-to-Pack (CTP) technology, which eliminates intermediate module assemblies and integrates cells directly into the pack, offers a transformative solution. This approach can increase volumetric utilization by 10-30% and push the gravimetric energy density beyond 200 Wh/kg. However, this heightened integration within the EV battery pack presents formidable thermal management challenges. The increased energy density and reduced internal air gaps exacerbate heat accumulation, making efficient and uniform cooling paramount for ensuring safety, performance, and longevity. This paper details the design of a module-less EV battery pack, incorporating a symmetric serpentine channel liquid cooling plate, and presents a comprehensive numerical analysis of its thermal performance under various operational and boundary conditions.
1. EV Battery Pack Design and Thermal Management System
1.1 Battery Cell Parameters and Pack Architecture
The core building block of the EV battery pack is a commercial prismatic lithium nickel manganese cobalt oxide (NMC) cell. The key parameters are summarized in Table 1. A total of 96 such cells are configured in a series arrangement to form the complete pack.
| Parameter | Value |
|---|---|
| Cell Dimensions (L×W×H, mm) | 148 × 78 × 103 |
| Cell Mass (g) | ≤ 2750 |
| Nominal Capacity (Ah) | 156 |
| Nominal Voltage (V) | 3.7 |
| Internal Resistance (mΩ) | ≤ 0.6 |
| Total Pack Energy (kWh) | 55.4 |
| Pack Nominal Voltage (V) | 355.2 |
The designed module-less EV battery pack achieves high integration. The 96 cells are organized into six stacks of 16 cells each. These stacks are housed within a structural lower enclosure. Adjacent cell stacks are separated by insulating partitions. Critically, the conventional module housings, bolts, and nuts are eliminated. Cells within a stack are physically secured and thermally coupled using a thermally conductive structural adhesive. To mitigate thermal runaway propagation risk, thin sheets of thermal insulation mica are placed between individual cells. A liquid cooling plate is positioned at the bottom of the pack, below the cell assemblies, with the thermal interface material ensuring efficient heat transfer from the cells to the cooling plate.
1.2 Liquid Cooling Plate Design
The liquid cooling system is pivotal for managing the thermal load of the high-density EV battery pack. A cold plate with a symmetric serpentine flow channel geometry was designed, fabricated from aluminum to balance thermal performance and weight. The serpentine design promotes effective heat absorption across the large surface area of the pack base. The cooling plate’s dimensions are 1580 mm × 1028 mm × 20 mm, ensuring full coverage of the cell array footprint.
2. Thermal Modeling and Simulation Setup
2.1 Cell Heat Generation Model
The heat generation rate of a lithium-ion cell is typically modeled using the framework proposed by Bernardi et al. This model accounts for irreversible Joule heating and reversible entropic heat. The volumetric heat generation rate \( q_{gen} \) (W/m³) is given by:
$$ q_{gen} = \frac{I}{V} \left[ (E_{OCV} – V_t) – T \frac{dE_{OCV}}{dT} \right] $$
where \( I \) is the current (A), \( V \) is the cell volume (m³), \( E_{OCV} \) is the open-circuit voltage (V), \( V_t \) is the terminal voltage (V), and \( T \) is the absolute temperature (K). The term \( \frac{dE_{OCV}}{dT} \) is the entropy coefficient. For simulation efficiency, a lumped model is often used, calculating the total heat generation per cell \( Q_{cell} \) (W) as:
$$ Q_{cell} = I^2 R_{int} + I T \Delta S $$
where \( R_{int} \) is the internal resistance and \( \Delta S \) represents the entropic heat term. For the subject NMC cell, the entropic heat coefficient was taken as 11.16×10⁻³ V. Based on this, the heat generation per cell at a 1.0C discharge rate is approximately 12.5 W, leading to a total EV battery pack heat load of 1.2 kW.
2.2 Thermal Properties and Anisotropy
The layered structure of a prismatic cell results in anisotropic thermal conductivity. The in-plane (x, y) and through-plane (z) effective thermal conductivities are calculated using a composite material approach:
Effective through-plane conductivity:
$$ \lambda_z = \frac{\sum L_i}{\sum (L_i / \lambda_i)} $$
Effective in-plane conductivity:
$$ \lambda_{x,y} = \frac{\sum L_i \lambda_i}{\sum L_i} $$
where \( L_i \) and \( \lambda_i \) are the thickness and thermal conductivity of each layer (current collectors, electrodes, separator). The average specific heat capacity \( C_{p,cell} \) and density \( \rho_{cell} \) are obtained via mass-weighted averaging:
$$ C_{p,cell} = \frac{1}{m_{cell}} \sum (c_{p,i} m_i) $$
$$ \rho_{cell} = \frac{m_{cell}}{V_{cell}} $$
The thermal properties of all materials used in the EV battery pack model are listed in Table 2.
| Material / Component | Density (kg/m³) | Specific Heat (J/(kg·K)) | Thermal Conductivity (W/(m·K)) |
|---|---|---|---|
| Battery Cell (Averaged) | 2345 | 979.6 | λ_x = 1.21, λ_y = λ_z = 17.45 |
| Cooling Plate (Aluminum) | 2702 | 903 | 237 |
| Coolant (50% Glycol-Water) | 1071 | 3300 | 0.384 |
| Thermal Interface Adhesive | 1200 | 1240 | Variable (0.5 – 1.5) |
| Inter-cell Insulation (Mica) | 1230 | 1457 | 0.23 |
2.3 Numerical Simulation Framework
A three-dimensional computational fluid dynamics (CFD) and conjugate heat transfer model was built using commercial software (STAR-CCM+). The model geometry was simplified by excluding minor components like busbars and bolts that have negligible impact on overall thermal behavior. The cells, thermal interface adhesive, cooling plate, and inter-cell insulation were fully resolved. A polyhedral mesh was generated, and a grid independence study confirmed that a mesh size of approximately 4.5 million cells yielded results independent of further refinement.
Boundary conditions included: a constant heat generation source for each cell based on discharge rate; a mass flow inlet for the coolant; and a pressure outlet. The initial temperature for the EV battery pack (cells, adhesive, insulation) was set to 42°C, while the cooling plate and inlet coolant were initialized at 25°C for baseline cases. Simulations were conducted for constant discharge rates (0.5C to 2.0C) and dynamic profiles based on the China Light-duty Vehicle Test Cycle (CLTC).
3. Results and Analysis of Thermal Performance
3.1 Impact of Discharge Rate on EV Battery Pack Cooling
The thermal performance of the EV battery pack was first evaluated under constant discharge rates ranging from 0.5C to 2.0C, with fixed coolant conditions (25°C inlet, 20 L/min flow rate). As expected, the maximum cell temperature (\(T_{max}\)) and the temperature difference within the pack (\(\Delta T_{pack}\)) increased significantly with higher discharge rates due to the quadratic relationship between current and Joule heating. The results, summarized conceptually below, show that while the cooling system manages the 1.5C rate adequately, the 2.0C rate pushes the pack to its thermal limits under these fixed conditions, highlighting the need for adaptive control.
$$ T_{max} \propto I^2 \cdot R_{int} $$
$$ \Delta T_{pack} \propto \frac{Q_{gen}}{\lambda_{eff}} $$
This finding is critical: a one-size-fits-all cooling strategy is insufficient. The thermal management system for an EV battery pack must adapt its parameters—specifically coolant temperature and flow rate—based on the demanded discharge rate (i.e., vehicle power demand) to maintain optimal temperature and uniformity.
3.2 Optimization of Thermal Interface Material
In a module-less EV battery pack, the thermal interface material (TIM), often a structural adhesive, plays a dual role: providing mechanical fixation and facilitating heat transfer from cells to the cooling plate. Its thermal conductivity (\(\lambda_{TIM}\)) is a key design parameter. Simulations were conducted varying \(\lambda_{TIM}\) from 0.5 to 1.5 W/(m·K) across different discharge rates.
The analysis revealed a trade-off: while a higher \(\lambda_{TIM}\) reduces the overall thermal resistance between the cell and the coolant, thereby lowering the average and maximum pack temperature, it can also exacerbate vertical temperature gradients within the cells. Heat is drawn more rapidly from the cell bottom, potentially increasing the top-to-bottom temperature difference (\(\Delta T_{cell}\)). An optimal value must balance overall cooling efficiency with cell-level temperature uniformity. For this specific EV battery pack design, a \(\lambda_{TIM}\) of 0.9 W/(m·K) was identified as providing the best compromise, effectively managing the peak temperature while keeping intra-cell and inter-cell \(\Delta T\) within acceptable limits (typically < 5°C).
3.3 Influence of Coolant Parameters and Operating Strategy
Building on the finding that cooling parameters must be adaptive, a parametric study was performed to map the optimal coolant inlet temperature (\(T_{in}\)) and mass flow rate (\(\dot{m}_{coolant}\)) for different steady-state discharge rates. The primary goals are to maintain \(T_{max} < 45-50°C\) (safety limit) and \(\Delta T_{pack} < 5°C\) (uniformity target).
The results can be generalized as follows: For a given \(T_{in}\), increasing \(\dot{m}_{coolant}\) enhances convective heat transfer, reducing \(T_{max}\). However, the benefit diminishes at higher flow rates due to the logarithmic nature of convective cooling, while pump power consumption increases linearly. Lowering \(T_{in}\) is highly effective at reducing \(T_{max}\) but can increase \(\Delta T_{pack}\) if the cooling is too aggressive, as it creates a steeper temperature gradient from the cooled bottom to the adiabatic top of the pack. The optimal operating points for the studied EV battery pack are summarized in Table 3.
| Discharge Rate (C-rate) | Recommended Inlet Temp. (°C) | Minimum Coolant Flow Rate (L/min) | Key Thermal Outcome |
|---|---|---|---|
| 0.5 | 20 – 25 | 10 | Excellent uniformity, very low ∆T |
| 1.0 | 25 | 10 | Balanced performance, ∆T ~ 3.5°C |
| 1.5 | 25 | 25 | Manages peak load, ∆T ~ 5°C |
| 2.0 | 25 | 30 | Maintains safety limit at high load |
This table provides a foundational look-up table for a rule-based thermal management controller for the EV battery pack.
3.4 Temperature Uniformity Under Dynamic CLTC Profile
Real-world EV driving involves highly dynamic power demands. The pack’s thermal performance was evaluated under three consecutive CLTC cycles, representing urban and suburban driving. A critical scenario was tested: a low-flow internal coolant circulation mode (simulating a thermal soaking condition or low-power pump operation) versus a no-coolant-flow (natural convection) scenario.
With an initial pack temperature of 38°C and an internal coolant loop starting at 25°C with a minimal flow of 2 L/min, the system demonstrated remarkable temperature homogenization. The maximum temperature stabilized around 35-37°C across the three cycles, but more importantly, the maximum temperature difference within the EV battery pack was maintained at approximately 0.25°C. In contrast, the no-flow scenario resulted in a \(\Delta T_{pack}\) of about 1.56°C. This underscores a vital principle for EV battery pack thermal management: even a minimal, actively circulated coolant flow is immensely superior to passive conditions for achieving temperature uniformity, which is crucial for balancing cell aging and maximizing usable capacity.
The temperature evolution can be described by the energy balance for the coolant in an internal loop:
$$ \dot{m}_{coolant} C_{p,coolant} (T_{out} – T_{in}) = \dot{Q}_{pack} – \dot{Q}_{loss} $$
where \( \dot{Q}_{pack} \) is the total heat generated by the pack and \( \dot{Q}_{loss} \) is heat loss to ambient. The small, continuous flow acts as an efficient internal heat exchanger, redistributing thermal energy from warmer to cooler regions of the pack.
4. Discussion: Towards an Intelligent Thermal Management System
The design and analysis presented herein confirm the viability of the module-less architecture for high-performance EV battery packs. The symmetric serpentine cooling plate provides an effective base for heat extraction. However, the studies clearly move the discussion beyond a fixed-design solution towards the necessity of an intelligent, adaptive thermal management system (TMS).
The performance of an EV battery pack is not static; its optimal thermal state is a function of multiple variables: instantaneous discharge/charge rate (I), state of charge (SoC), ambient temperature, and cell aging. A modern TMS should integrate the following insights from this work:
- Variable Flow Control: The pump speed should be modulated based on the power demand. High flow rates are reserved for aggressive driving or fast charging, while low flow rates suffice for city driving (as shown by the CLTC analysis) to save pump energy and improve overall vehicle efficiency.
- Coolant Temperature Control: The chiller or heater setpoint should be adjusted dynamically. A warmer coolant can be used under low loads to improve cell efficiency and reduce heater energy in cold climates, while a cooler coolant is essential for high loads. The trade-off between \(T_{max}\) and \(\Delta T_{pack}\) must be managed by the control algorithm.
- Thermal Interface Optimization: The selection of \(\lambda_{TIM} = 0.9 \, \text{W/(m·K)}\) is design-specific. Future materials development focusing on higher conductivity while maintaining mechanical properties and cost will further benefit module-less EV battery pack designs.
- State Estimation Integration: The thermal model used for simulation can be reduced in order and embedded in the Battery Management System (BMS) for real-time estimation of core temperatures and heat generation, providing the necessary inputs for the adaptive TMS controller.
The ultimate goal is to maintain every cell in the EV battery pack within a narrow, optimal temperature window (e.g., 25°C – 35°C) with minimal spatial variation, thereby maximizing power capability, slowing degradation, and ensuring safety under all operating conditions.
5. Conclusion
This work presented a comprehensive design and thermal analysis of a module-less EV battery pack employing a CTP architecture. A detailed CFD model incorporating anisotropic cell properties was developed and validated. The key conclusions are:
- The designed EV battery pack with a symmetric serpentine liquid cooling plate demonstrates robust cooling capability, effectively handling discharge rates up to 2.0C through appropriate selection of coolant parameters.
- The thermal conductivity of the interface adhesive in a module-less pack presents a design trade-off. An optimal value of 0.9 W/(m·K) was identified for this configuration, balancing peak temperature reduction with pack temperature uniformity.
- Coolant inlet temperature and flow rate must be treated as dynamic control variables, not fixed parameters. Optimal setpoints vary significantly with the discharge rate to simultaneously manage maximum temperature and temperature spread within the EV battery pack.
- Under dynamic driving profiles like the CLTC, even a minimal active coolant circulation (e.g., 2 L/min) is profoundly effective at enhancing temperature uniformity within the EV battery pack compared to a passive no-flow scenario, reducing the maximum temperature difference by over 1.3°C in the studied case.
The evolution of EV battery pack technology towards higher integration demands a parallel evolution in thermal management strategy—from a static, hardware-centric approach to an adaptive, software-defined system that intelligently controls cooling resources to optimize the pack’s thermal state in real-time. The findings and methodologies outlined here provide a foundation for the development of such advanced thermal management systems for next-generation electric vehicles.