As electrified transportation advances toward higher energy density and higher power density, the lithium-ion cell and the surrounding pack hardware are exposed to increasingly severe thermal duty. My work here started from a simple but frequently overlooked observation: the overwhelming majority of thermal-safety countermeasures target the lithium-ion cell itself, while the metallic conductors that carry the full pack current are treated as a purely electrical detail. In this study I treat the aluminium busbar as a thermal component of equal importance to the lithium-ion cell, and I quantify how an undersized busbar converts electrical current into a persistent local heat source that the battery management system (BMS) cannot see and the liquid-cooling loop cannot compensate.
Mainstream thermal-safety strategies for lithium-ion cell systems can be grouped into four families: power limitation enforced by the BMS, active thermal management through optimised cold-plate design, cell-inversion architectures that protect the occupant compartment, and thermal-barrier materials that interrupt propagation between neighbouring cells. All four families aim either at improving the intrinsic safety of the lithium-ion cell or at strengthening the active cooling path. The current-carrying capability of the busbar, by contrast, has rarely been evaluated as a thermal-safety variable in its own right. A systematic investigation of how mechanical and electrical interconnect structures influence thermal safety is therefore a necessary complement to existing protection mechanisms, particularly for the prevention of spontaneous ignition in scenarios without mechanical abuse.
Inside a battery pack, the busbar aggregates and transports the current of many individual lithium-ion cell units. Depending on the raw material, busbars are classified as copper busbars or aluminium busbars. Copper offers superior electrical conductivity and good mechanical strength and is therefore commonly used for module-to-module connections and terminal outputs, whereas aluminium is favoured for cell-to-cell connections because of its low density and low cost. The differing physical properties of copper and aluminium determine different overcurrent coefficients. For any battery system design, ensuring that the busbar overcurrent capability, that is the maximum permissible continuous current, exceeds the operating demand of the system is a core criterion of thermal-safety design. When the actual working current of the pack exceeds the rated overcurrent value of the busbar, the conductor cross-section temperature rises rapidly through the Joule heating effect, which is described by the fundamental relation
$$Q = I^{2} R t$$
where \(Q\) is the generated heat, \(I\) is the instantaneous current, \(R\) is the equivalent resistance of the conduction path and \(t\) is the duration of current flow. This thermal hazard can escalate into pack-level overheating and ultimately into a thermal-safety event that threatens both occupants and the vehicle system. Because the resistance of a conductor is governed by its geometry and resistivity,
$$R = \rho \frac{L}{A}$$
with \(\rho\) the resistivity, \(L\) the conduction length and \(A\) the cross-sectional area, the current density
$$J = \frac{I}{A}$$
becomes the single most informative design index for busbar sizing. A lithium-ion cell pack that migrates from a low-rate charging profile to a high-rate fast-charging profile changes its current density dramatically without any change in the physical busbar, and this mismatch is the subject of the present study.
Test Bench and Instrumentation
I built the experimental platform around a 800 V class charge and discharge system, a walk-in thermal chamber, a dual-channel water chiller, a multi-channel temperature acquisition system and a multi-channel voltage sampling board. The configuration allowed me to record the temperature rise of busbars and of the lithium-ion cell surfaces simultaneously during charging and discharging. The principal equipment and its specifications are summarised in the following table.
| Subsystem | Key specification | Purpose in this study |
|---|---|---|
| Charge and discharge unit | 800 V, 600 A, 300 kW, accuracy ±1‰ | Pack-level fast charging and typical duty-cycle emulation |
| Walk-in thermal chamber | Door width 1.8 m, internal width ≈2.1 m, height 1.8 m, depth 3.5 m; −40 °C to 130 °C; control accuracy ±1 °C | Ambient-temperature control for pack and module charging |
| Water chiller | Dual channel: 5.5 kW × 2 @ 100 °C; 5.5 kW × 2 @ 20 °C; 3.8 kW × 2 @ 5 °C | Emulation of a 3 kW liquid-cooling loop |
| Temperature acquisition | Multi-channel; accuracy 0.5 °C; resolution 0.01 °C | Direct monitoring of busbars, weld spots and terminal posts |
| Voltage acquisition | 6½-digit multimeter for static BMS accuracy; isolated multi-channel dynamic board with 1 mV accuracy | Static and dynamic voltage precision verification |
Pack Selection and Pre-Treatment
Pack-A was selected as the primary test object. It is a 1P96S pack with an MTP architecture, composed of three 1P32S sub-modules. To establish the validity of the conclusions across the whole service life, three Pack-A samples with different accumulated mileage were used as a comparison group and labelled Pack-A, x# with x = 1, 2, 3. Three additional packs of similar construction, designated Pack-B, Pack-C and Pack-D, were introduced as control groups. The basic parameters are listed below.
| Parameter | Unit | Pack-A | Pack-B | Pack-C | Pack-D |
|---|---|---|---|---|---|
| Overall dimensions | mm | 2020 × 1430 × 151 | 2020 × 1430 × 151 | 1320 × 1340 × 145 | 1320 × 1340 × 145 |
| Busbar dimensions | mm | 50 × 1.6 × 117.5 | 58.5 × 2.0 × 115 | 50.3 × 1.5 × 77.6 | 50.3 × 1.5 × 77.6 |
| Busbar cross-section | mm² | 80 | 117 | 75 | 75 |
| Rated energy @ 1/3C | kWh | 88.3 | 88.3 | 67.8 | 67.8 |
| Cooling method | — | Bottom liquid cooling | Bottom liquid cooling | Bottom liquid cooling | Bottom liquid cooling |
For pack-level testing I first performed an appearance inspection and recorded the baseline information of each pack. I then opened the pack and selected sixteen busbars in the module nearest to the main positive output terminal, attaching temperature sensors at those locations so that the hottest conduction path in the pack would be captured. The sampling points used for the module-level and cell-level work are listed below together with their numbering; positions 1# to 4# correspond to instrumentation introduced inside the lithium-ion cell, while positions 5# and 6# correspond to external surface instrumentation.
| Index | Measurement location | Level |
|---|---|---|
| 1# | Positive electrode laser weld | Internal |
| 2# | Negative electrode laser weld | Internal |
| 3# | 5 mm below the root of the positive tab | Internal |
| 4# | 5 mm below the root of the negative tab | Internal |
| 5# | Centre of the large face | External |
| 6# | Bottom of the large face | External |
| 7# | BMS NTC location on the top cover | External |
| 8# | Positive terminal post | External |
| 9# | Negative terminal post | External |
| 10# | Middle of the negative busbar | External |
| 11# | Middle of the positive busbar | External |
Test Procedures
Three procedures were designed, differing in test object and thermal boundary condition. Procedure one was a pack-level constant-current test. Each pack was discharged at 1/2C until the relay opened. After a rest period, the pack was charged at 2C while busbar temperatures and BMS-reported temperatures were logged continuously. Charging was terminated when the NTC temperature exceeded 55 °C or when any other BMS protection condition was triggered. The thermal management logic opened the coolant loop above 34 °C and closed it below 32 °C, with the loop disabled during discharge and rest, coolant inlet at 22 °C and a flow rate of 12 L/min.
Procedure two was a map-based fast-charging test. The pack was placed in the thermal chamber, soaked until its internal temperature matched the setpoint, and adjusted to 3 % state of charge. It was then charged according to the vehicle fast-charging map, in which the current is indexed by state of charge and by lithium-ion cell temperature, while busbar temperature, BMS temperature, current and state of charge were recorded synchronously. Charging stopped at 95 % state of charge or upon triggering a BMS protection. The thermal management logic disabled the coolant loop below a maximum temperature of 34 °C, enabled it with a 22 °C inlet and 12 L/min flow between 34 °C and 39 °C, and enabled it with an 18 °C inlet and 15 L/min flow above 39 °C.
Procedure three was a small-module test built from Pack-A lithium-ion cell units in a 1P3S configuration. The module was soaked to the target ambient temperature, discharged to 3 % state of charge, charged according to the fast-charging map until 95 % state of charge or a protection trigger, and then discharged at 1/2C to 3 % state of charge. Three consecutive cycles were performed. The coolant loop opened above 35 °C and closed below 32 °C and was disabled during discharge and rest, with a 20 °C inlet at either 1 L/min or 3.6 L/min. The four sides and the top of the small module were wrapped in thermal insulation to reproduce the single-sided heat-rejection boundary that a lithium-ion cell experiences inside a real pack. The differences between the three procedures arise from the differences in test object and boundary condition: procedures one and two use the standard Pack-A thermal management condition, with procedure two additionally reproducing the two-stage 34/39 °C cooling algorithm of the on-vehicle BMS, while procedure three lowers the cooling threshold and flow rate to match the thermal characteristics of a single lithium-ion cell and to allow a flow-rate sensitivity comparison.
Constant-Current Fast Charging of Pack-A
Three Pack-A samples with accumulated mileages of 16 000 km, 54 758 km and 176 438 km were subjected to 2C fast charging both with and without liquid cooling. The purpose was to verify the excessive temperature rise caused by insufficient busbar overcurrent capability and to check whether an unsuitable sampling arrangement produces an unreported overtemperature condition. The measured maximum temperatures are collected in the table below.
| Sample | Cooling | 2C duration (s) | Max busbar temperature (°C) | Max BMS temperature (°C) | Difference (°C) |
|---|---|---|---|---|---|
| Pack-A, 1# | No coolant | 192.0 | 62.83 | 36.0 | 26.83 |
| Pack-A, 1# | Coolant on | 182.4 | 56.86 | 32.0 | 24.86 |
| Pack-A, 2# | No coolant | 245.2 | 65.00 | 36.0 | 29.00 |
| Pack-A, 2# | Coolant on | 245.2 | 60.44 | 32.0 | 28.44 |
| Pack-A, 3# | No coolant | 224.2 | 64.52 | 37.0 | 27.52 |
| Pack-A, 3# | Coolant on | 217.1 | 56.81 | 32.0 | 24.81 |
With liquid cooling active, Pack-A, 1# sustained 2C charging for 182.4 s and the aluminium busbar reached 56.86 °C at the end of charging, whereas the maximum temperature reported by the BMS was 32 °C, giving a difference of 24.86 °C. Pack-A, 2# sustained 245.2 s, the busbar reached 60.44 °C and the BMS reported 32.0 °C, a difference of 28.44 °C. Pack-A, 3# sustained 217.1 s, the busbar reached 56.81 °C and the BMS reported 32.0 °C, a difference of 24.81 °C. Without liquid cooling the differences were 26.83 °C, 29.00 °C and 27.52 °C respectively.
Two conclusions follow immediately. First, compared with the commonly specified pack-level temperature-difference requirement of less than 10 °C, Pack-A exceeds the limit substantially during fast charging. Second, and more importantly, the magnitude of the discrepancy is essentially the same with and without liquid cooling. This means that the Joule heating produced in the busbar under high current far exceeds the compensation capability of the cooling system. Even when the cooler is strengthened, the gap between the busbar and the BMS-monitored lithium-ion cell surface remains above 20 °C in every case. The charging strategy actually executed by the BMS is therefore not matched to the optimal charging strategy for the lithium-ion cell at that moment, and operating the lithium-ion cell above its map current for extended periods is a serious challenge to both safety and cycle life. In addition, because the aluminium busbar is mechanically joined to the cell terminal posts, a hot busbar can conduct heat back into the lithium-ion cell, which further increases the overtemperature risk of the busbar under high current.
A first-order thermal model explains why cooling cannot suppress this effect. For a busbar segment of thermal capacitance \(C_{th}\) exchanging heat with a coolant sink through a lumped conductance \(hA_s\) while generating \(I^2R\) internally, the transient temperature is
$$T(t) = T_{\infty} + \frac{I^{2} R}{h A_{s}} \left( 1 – e^{-t/\tau} \right), \qquad \tau = \frac{C_{th}}{h A_{s}}$$
so that the steady-state excess temperature is proportional to \(I^{2}R / (hA_s)\). Increasing the coolant flow raises \(h\) only weakly, whereas raising the current raises the generated heat quadratically. A busbar operating at the design current density will therefore always dominate the heat balance once the current is doubled, exactly as observed in the experiments above.
Assessment of BMS Temperature Sampling Locations
To determine whether the placement of the BMS NTC is responsible for the observed blind spot, I attached sensors to three separate locations on the surface of three lithium-ion cell units in Pack-A, 1# and compared them with the NTC position under the map charging profile. During the initial stage of charging, the NTC location was warmer than the upper point, which was warmer than the midpoint, which was warmer than the lower point. As charging progressed, the three surface points rose faster than the NTC location and ended 4 °C to 5 °C above it. Notably, during the fast-charging phase, defined here as currents above 400 A, the NTC location tracked the relative high point of the cell surface temperature, and the difference from the maximum surface temperature remained within 5 °C, which satisfies normal engineering requirements. Because the NTC value slightly exceeds the maximum surface temperature, a BMS that indexes its charging map on this measurement is conservative and will not command a current larger than intended, so the NTC location itself is a sound design choice. The core deficiency of the thermal management design of this pack therefore does not lie in the NTC position. It lies in the fact that the NTC monitors the lithium-ion cell surface while the dominant heat source is the busbar.
Map-Based Fast Charging of Pack-A
Ambient temperatures of 25 °C, 30 °C and 38 °C were used to represent different usage scenarios for map-based fast charging, and the resulting busbar and NTC temperatures are summarised below.
| Ambient temperature (°C) | Peak busbar temperature (°C) | Peak BMS temperature (°C) | Difference (°C) |
|---|---|---|---|
| 25 | 71.30 | ≈50 | 29.37 |
| 30 | 74.41 | ≈50 | 28.74 |
| 38 | 80.79 | ≈57 | 30.33 |
The results expose a system-level overtemperature risk under high-load operation. The busbar peak temperature reaches 80.79 °C in the real charging regime, while the difference from the BMS-monitored temperature stays close to 30 °C and does not shrink appreciably when the cooling loop intervenes. The existing thermal design of this pack therefore cannot effectively control the local overheating produced by busbar overcurrent. During charging, the busbar behaves as a continuous heat source and conducts heat back into the wound electrode stack, causing abnormal temperature rise in the lithium-ion cell and threatening pack-level thermal safety.
Module-Level Temperature Rise in a 1P3S Configuration
The pack-level results showed that under both constant-current and map-based charging, the busbar temperature rises abnormally in samples of different state of health, and that a large discrepancy exists relative to the BMS reading. This indicates a strong coupling between busbar overcurrent capability and the charging strategy: whenever the charging strategy is revised or the current demand is changed without re-verifying the busbar overcurrent capability, the busbar will overheat and inject heat into the lithium-ion cell through the terminal post. Weld quality, which raises local thermal resistance, may act as an additional aggravating factor. To rule out the possibility that the BMS simply fails to cover the hottest region of the lithium-ion cell, and to verify the temperature back-flow effect from the busbar into the cell, precise internal temperature measurement is required. The 1P3S small module provides exactly this capability. The test matrix and the maximum temperatures recorded at each instrumented position are given below.
| Ambient (°C) | Charging strategy | Module | Flow (L/min) | Coolant (°C) | 1# | 2# | 3# | 4# | 5# | 6# | 7# | 8# | 9# | 10# | 11# | Ambient measured (°C) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 35 | 30 min | 1# | 3.6 | 20 | 63.6 | 62.7 | 49.2 | — | 47.5 | 45.8 | 46.8 | 66.2 | 63.6 | — | — | 35.0 |
| 35 | 30 min | 2# | 1.0 | 20 | 71.6 | 74.7 | 54.6 | 56.0 | 52.2 | 49.5 | 53.5 | 74.3 | 69.7 | 70.5 | 75.8 | 35.0 |
| 40 | 30 min | 1# | 3.6 | 20 | 71.7 | 70.8 | 55.7 | — | 52.4 | 50.2 | 53.4 | 75.4 | 72.3 | — | — | 40.8 |
| 40 | 30 min | 2# | 1.0 | 20 | 79.4 | 78.6 | 58.7 | 60.2 | 55.5 | 52.7 | 59.0 | 81.6 | 76.8 | 78.0 | 83.3 | 40.0 |
| 35 | 40 min | 1# | 3.6 | 20 | 52.7 | 54.9 | 46.4 | — | 45.1 | 43.4 | 43.6 | 54.2 | 53.5 | — | — | 35.0 |
| 35 | 40 min | 2# | 1.0 | 20 | 55.6 | — | 48.5 | 49.8 | 46.9 | — | 45.5 | 60.0 | 55.2 | 56.5 | 60.0 | 35.0 |
| 40 | 40 min | 1# | 3.6 | 20 | 57.1 | 58.4 | 50.0 | — | 48.2 | 46.4 | 47.3 | 59.2 | 58.2 | — | — | 40.0 |
| 40 | 40 min | 2# | 1.0 | 20 | 60.7 | — | 52.5 | 54.3 | 50.8 | — | 49.9 | 65.5 | 60.5 | 61.9 | 65.6 | 40.0 |
Increasing the coolant flow rate does suppress the overall pack temperature to a measurable degree, but as already indicated by the pack-level results, the cooling strategy does not reduce the temperature difference between the terminal posts and the NTC. The NTC value at position 7# is significantly lower than the values at the positive and negative terminal posts, positions 8# and 9#, with a gap of 15 °C to 20 °C. In the 35 °C ambient and 30 min charging case, for example, position 7# reads 46.8 °C while position 8# reads 66.2 °C, a difference of 19.4 °C. At the lower coolant flow rate of 1 L/min the gap widens further to 20 °C to 25 °C; under 40 °C ambient temperature and the 30 min strategy, position 7# reads 53.4 °C while position 8# reads 75.4 °C, a difference of 22 °C. The positive terminal post reaches as high as 81.6 °C at 40 °C ambient, and at that moment the large-face temperature at position 5# and the tab-root temperature at position 3# also reach their peaks, demonstrating a strong correlation between the internal temperature rise of the lithium-ion cell and the heat source at the terminal post. This verifies the conduction and spreading of heat from the terminal region and shows that a hot post causes abnormal heating of the whole lithium-ion cell.
The thermal network that describes this behaviour can be written as
$$T_{weld} – T_{cool} = Q_{weld} \left( R_{weld} + R_{tab} + R_{plate} + R_{conv} \right) + I^{2} R_{bulk} \left( R_{tab} + R_{plate} + R_{conv} \right)$$
where \(Q_{weld}\) is the heat generated in the weld spot, \(R_{weld}\) is the weld thermal resistance, \(R_{tab}\) and \(R_{plate}\) are the conductive resistances of the tab and the cold plate, and \(R_{conv}\) is the convective resistance of the coolant loop. The same network applies in the reverse direction to the lithium-ion cell surface, which is why the NTC located on the top cover always lags the terminal post:
$$\Delta T_{blind} = \max \left( T_{busbar},\ T_{post} \right) – \max \left( T_{NTC} \right)$$
Busbar Cross-Section Versus Temperature Rise
To confirm that insufficient overcurrent capability is a fundamental design defect rather than a transient artefact, I compared two packs of similar architecture under identical charging conditions. The busbar of Pack-A has a cross-section of approximately 80 mm², whereas the busbar of Pack-B has a cross-section of approximately 117 mm². The theoretical continuous overcurrent capability of the Pack-A busbar is about 200 A, while the fast-charging peak current exceeds 500 A, which represents a severe shortfall. The corresponding current densities are highly illuminating:
$$J_{A} = \frac{500}{80} = 6.25\ \text{A/mm}^{2}, \qquad J_{B} = \frac{500}{117} = 4.27\ \text{A/mm}^{2}$$
and the steady-state overcurrent limit predicted by a lumped convective balance
$$I_{max} = \sqrt{\frac{h\,P\,A_{cs}\,\Delta T_{allow}}{\rho}}$$
scales with \(A_{cs}\) at fixed geometry, so a 46 % larger cross-section permits a 46 % larger permissible current at the same allowable temperature rise. The experimental comparison at 40 °C ambient and identical charging profiles shows that the Pack-A busbar peaks at 67.96 °C against 57.95 °C for Pack-B, a penalty of 10.01 °C, with a local temperature rise of roughly 28 °C. The result confirms that an under-sized cross-section is the root cause of the local overheating problem.
| Metric | Pack-A | Pack-B | Relative difference |
|---|---|---|---|
| Busbar cross-section (mm²) | 80 | 117 | +46 % |
| Current density at 500 A (A/mm²) | 6.25 | 4.27 | −32 % |
| Approximate continuous rating (A) | 200 | 292 | +46 % |
| Peak busbar temperature at 40 °C ambient (°C) | 67.96 | 57.95 | −10.01 °C |
| Local temperature rise (°C) | ≈28 | ≈18 | −10 °C |
Heat Path and the Role of the Weld Spot
Multi-point temperature monitoring inside the lithium-ion cell reveals a clear hierarchy of heat transfer under fast charging: heat generated at the weld spot propagates to the U-shaped busbar, then to the tab, then into the wound core and finally to the outer case. The peak temperature measured at the weld spot reaches 84.7 °C while the NTC reports only 54.9 °C, a gap approaching 30 °C. This temperature gradient confirms that insufficient busbar overcurrent capability is the root cause of local overheating and that the weld spot is the core region where the thermal risk is concentrated. A larger busbar cross-section with a well-controlled weld geometry lowers the current density, flattens the gradient and reduces the peak temperature of the whole conduction path.
| Location in the conduction path | Representative peak temperature (°C) | Gradient relative to coolant |
|---|---|---|
| Weld spot | 84.7 | Highest |
| U-shaped busbar | ≈80 | High |
| Terminal post | 75 to 82 | High |
| Tab root | 55 to 60 | Intermediate |
| Wound core surface | 50 to 55 | Intermediate |
| NTC on the top cover | 45 to 59 | Lowest of the monitored points |
Design Implications
The combined evidence supports a set of quantitative design rules that treat the busbar as a thermal-safety component rather than a purely electrical one. The most important of these are collected below.
| Design rule | Expression | Recommended value |
|---|---|---|
| Continuous overcurrent margin | \(I_{rated} \ge k I_{rms}\) | \(k \ge 1.5\) |
| Peak current margin | \(I_{rated} \ge k_{p} I_{peak}\) | \(k_{p} \ge 0.6\) for the fast-charge peak |
| Current density limit | \(J = I/A\) | \(J \le 2.5\ \text{A/mm}^{2}\) continuous |
| Pack temperature uniformity | \(\Delta T_{pack} = \max(T_i) – \min(T_i)\) | \(\Delta T_{pack} \le 10\ ^\circ\text{C}\) |
| Monitoring offset tolerance | \(\Delta T_{blind} = \max(T_{busbar}) – \max(T_{NTC})\) | \(\Delta T_{blind} \le 5\ ^\circ\text{C}\) |
| Cooling and overcurrent co-optimisation | \(Q_{gen} \le h A_{s} \Delta T_{allow}\) | Verify at the fast-charge peak |
Two further observations deserve emphasis. First, the overcurrent requirement must be re-verified whenever the charging strategy changes. A busbar selected for a conventional charging rate may still be inadequate once the market demands a fast-charging capability, because the thermal margin that existed at the original current density disappears quadratically with current. Second, the effectiveness of the liquid-cooling system must be assessed jointly with the busbar design. As the experiments show, strengthening the coolant loop cannot compensate for a structurally under-sized conductor, because the thermal resistance between the hot spot and the coolant remains the bottleneck. Only a combined optimisation of the electrical conduction path and the thermal path can deliver a genuine improvement in thermal management efficiency.
Conclusions
I have examined the reliability of BMS temperature monitoring and the thermal-safety risk arising from busbar overtemperature through systematic testing at the pack, module and single lithium-ion cell levels. The following conclusions can be drawn.
First, a significant temperature difference exists between the busbar and the BMS measurement, and an unreported overtemperature risk is objectively present. In 2C fast-charging tests on packs with different accumulated mileage, the difference between the actual busbar temperature and the BMS-reported temperature remained above 20 °C in every case. The BMS therefore contains a protection blind spot: when the busbar reaches its safe temperature limit, the protection mechanism cannot intervene in time. This behaviour was reproduced under both constant-current and map-based charging, and it persisted with and without liquid cooling.
Second, the BMS sampling location is well chosen but cannot sense an overheated busbar. Comparing multi-point surface temperatures with the NTC position in a lithium-ion cell module confirmed that during fast charging the NTC reflects the high point of the upper cell surface and remains within 5 °C of the maximum surface temperature, which means the location itself is correctly designed. Combined with the busbar measurements, however, the results show that monitoring the lithium-ion cell temperature alone cannot detect an abnormal heat source located in the conductor.
Third, the heat-transfer path is clear and the weld spot is the core region of local overheating. Internal multi-point monitoring reveals a stepwise gradient from the weld spot to the U-shaped busbar, the tab, the wound core and the cell surface. The weld spot reaches 84.7 °C while the NTC reads 54.9 °C, a difference of nearly 30 °C, which identifies insufficient busbar overcurrent capability as the root cause of the local overheating.
Fourth, a direct correspondence exists between busbar design and thermal safety. An appropriate weld geometry and an adequate busbar cross-section effectively suppress the temperature rise of the local heat source and reduce hot-spot risk by lowering the current density and flattening the current distribution. Synergistic optimisation of the cooling system and the busbar overcurrent capability, including coolant flow rate and the layout of the heat-conduction path, further strengthens thermal management efficiency and safeguards the thermal safety and long-term reliability of the pack under complex operating conditions.
Taken together, these results show that the busbar structure is a critical element of pack thermal management. A busbar with insufficient overcurrent capability becomes a persistent heat source during fast charging, produces an abnormal temperature rise, creates a thermal monitoring blind spot in the BMS, and cannot be compensated by the existing liquid-cooling system. The findings provide a quantitative link between electrical interconnection and thermal safety and offer both a theoretical basis and experimental validation for optimising the thermal design of lithium-ion cell battery packs. It must also be noted that the engineering background of this issue stems from a rapid evolution in market demand for fast charging: in the early design phase, busbars are typically selected on the basis of conventional charging rates, and whether a busbar that satisfied conventional duty still possesses sufficient overcurrent capability, and whether its coordination with the cooling system remains effective, must be verified whenever the charging envelope is expanded.