Experimental Thermal Assessment of Novel Dual-Terminal Architecture for Cylindrical Li-Ion Battery Packs Under Variable Discharge Rates

Abstract

A novel architectural design is proposed to optimize the thermal management of lithium-ion batteries (LiBs) through a software-enabled switching mechanism. This approach addresses critical challenges such as hot-spot generation, peak temperature rise, and uneven thermal distribution—issues commonly observed in conventional single-terminal battery modules (STBMs). The proposed dual-terminal configuration integrates an enhanced battery pack structure with a software-enabled switching algorithm that identifies the 50% depth of discharge (DoD) and toggles the current path between two terminals to supply the load. Correspondingly, the module also incorporates the division of four thermal zones and four regions concept in the battery module (BM). Experiments were conducted to evaluate the performance of the proposed model at five different C-rates: 0.5C, 0.75C, 1C, 1.25C, and 1.5C. The results demonstrate that the software-enabled dual-terminal switching (Se-DTS) consistently outperforms the STBM across three key aspects. First, in terms of peak temperature, Se-DTS achieved reductions of 19.33%, 17.83%, and 12.72% at C-rates of 1C, 1.25C, and 1.5C, respectively. Second, in thermal distribution, Se-DTS improved performance, with an 86.1% reduction at 1.25C. Third, regarding hot-spot reduction, improvements of 100% (regional level) and 72.22% (zonal level) were observed at 1.25C, while at 1.5C, an 80% improvement was achieved at the zonal level, without using a cooling system.

1. Introduction

Lithium-ion batteries (LiBs) have solidified their position as the preeminent electro-chemical energy storage technology, primarily due to their superior energy density, extended cycle life, and advantageous power-to-weight ratio [1]. This has led to their widespread adoption in diverse applications ranging from electric vehicles (EVs) and aerospace systems to consumer electronics and grid-scale energy storage. However, the escalating performance demands, particularly for high-rate charging and discharging, have intensified the challenges associated with thermal management [2]. Inadequate control of heat generated during operation can severely compromise LiB safety, accelerate degradation, and reduce overall system efficiency, making advanced thermal management a critical design imperative [3,4]. Studies consistently show that even minor temperature non-uniformities (e.g., >5 °C) across cells within a pack can led to differential aging and diminished performance [5], issues exacerbated in densely packed, large-scale assemblies where, as Rawat S. et al. recently investigated, achieving thermal homogeneity remains a significant challenge [6].

A primary determinant of a battery’s thermal behavior is the discharge rate, or C-rate, which dictates the current magnitude and, consequently, the extent of internal resistive heating [7]. The intrinsic link between higher C-rates and increased, often non-linear, heat generation necessitates a thorough understanding of the cell and pack thermal response across a spectrum of operating conditions for safe and optimized performance [8,9]. Recent work by Velumani et al. has provided a comprehensive review and perspective on advanced thermal characterization techniques under extreme dynamic loads [10]. Conventional battery pack designs predominantly utilize a single-terminal configuration, which can lead to current crowding and localized hot spots, particularly near terminal regions, thereby fostering thermal non-uniformity [11,12,13]. While external thermal management solutions like air/liquid cooling or phase change materials (PCMs) are common [14,15], they often add system complexity, weight, and cost. Recent advancements in external, such as the novel serpentine microchannel liquid cooling proposed by H. Chen et al. [16], and the high-performance PCM–graphite foam composites developed by Park and Kim [17] demonstrate ongoing efforts to enhance heat dissipation, yet the pursuit of intrinsic thermal control through architectural modifications remains a key research focus.

To address these challenges from a design perspective, research exploring modifications to internal architecture and power delivery pathways, such as the work by Zhao et al. on customized tab thickness [18], has gained traction. Building on this, strategies involving multi-terminal battery architectures are emerging as a promising avenue for intrinsic thermal regulation. These architectures, by offering multiple current entry and exit points, allow for dynamic redistribution of current density within the battery module, as explored in recent multi-tab cell designs by X. Li et al. [19] and in reviews on reconfigurable battery modules [20,21] and preventing excessive localized temperature rise and promoting a more uniform thermal profile across the pack [22]. Such approaches aim to achieve enhanced thermal management of Li-ion cylindrical cells, like the 18,650 s used in many packs, by influencing the thermal characteristics during high discharge rates without relying solely on external cooling [23]. For instance, recent simulations by Cao et al. demonstrated potential reductions in peak cell temperatures [24]. Furthermore, as Zhang et al. discussed, current distribution plays a critical role in mitigating thermal runaway in cylindrical cell arrays [25], a concern that multi-terminal designs could potentially address [26]. Intelligent control algorithms, potentially leveraging artificial intelligence as explored by researchers like Karnehm and Abed for adaptive thermal management in BMS, are also being investigated to optimize switching strategies in real-time based on thermal feedback [27,28].

The present study experimentally investigates a novel dual-terminal battery module architecture featuring a software-enabled, toggle-based switching mechanism. This system is designed to dynamically alter current paths during charging/discharging, thereby facilitating more uniform thermal energy dissipation across the battery pack. We evaluate this system under a comprehensive range of discharge rates (0.5C to 1.5C) using a 16-cell cylindrical Li-ion (LiFePO₄) module instrumented with embedded thermal cameras across four defined regions and zones within the battery module (BM). Temperature profiles, hot spot formation, and threshold temperature crossings are systematically compared between the proposed software-enabled dual-terminal switching (Se-DTS) system and a conventional single-terminal configuration. The analysis incorporates defined thermal metrics (P_rn, P_tn±, $\Delta P_r$, $\Delta P_Z$) and correlates dynamic switching events with observed thermal behavior.

2. Proposed Battery Pack Technology and Experiments

The proposed battery pack technology (BPT) focuses on BM architecture, addressing the major limitations observed in the conventional design and their mitigation through the proposed system.

2.1. Proposed Novel Architecture of Battery Module and Battery Pack

This study investigated the thermal characteristics of a novel Se-DTS battery architecture under various discharge rates, relying on natural convection for heat dissipation. The core of the system is a custom-designed BM constructed using LiFePO₄ IFR18650 cylindrical cells (Orange A-grade, 3C discharge rate) arranged in a nSnP (n = 4) configuration, with overall dimensions of 200 mm × 200 mm × 90 mm (Figure 1a). This BM features two distinct terminal pairs: Terminal 1 (T1) with connections at cell (1,1) (cell (i = series, j = parallel)) (positive) and cell (4,1) (negative) cell positions, and Terminal 2 (T2) with connections at cell (1,4) (positive) and cell (4,4) (negative) positions. A 22 mm spacing between cells was maintained for distinct thermal observation. Each terminal delivers a nominal 12.8 V and 6 Ah. The BM was housed within a custom acrylic casing (200 mm × 200 mm × 90 mm). To manage the dual-terminal functionality, two 4 s 24 V battery management systems (BMS), rated for 40 A discharge, were integrated, with BMS_1 connected to T1 and BMS_2 connected to T2. A relay module, controlled by a microcontroller, managed the switching between these terminals, ensuring only one was active for discharging at any time (Figure 1b). The detailed specifications of the individual test cell and the assembled BM are provided in

Table 1. Specifications of the test cell and lumped cell pack model.

	Test Cell	Lumped Cell
Chemical composition	LiFePo4	LiFePo4
Configuration	1S1P	4S4P
Nominal discharge capacity (Ah)	1.5	6.0
Nominal voltage (V)	3.2	12.8
Standard charging rate (Ah)	1.5	6.0
Maximum continuous discharge (A)	4.5	18
Discharge cut-off voltage (V)	2.5–2.8	10–11.2
Cell/Pack weight (g)	39	~800
Cell/Pack height (mm)	65	95
Cell diameter/pack dimensions (mm)	18	200 × 200 × 90
C-rate (C)	0.5–3	0.5–3
Manufacturer	Orange	Orange

.

2.2. Experimental Set-Up and Procedure

The actual set-up of the experiment and the procedure to conduct the experimental test cases are discussed in this section, focusing on the sub-systems connection and the software-enabled program conditions to test Se-DTS.

2.2.1. Experimental Setup

A LiFePO₄-based BM (Figure 1a) and BPT (Figure 1b) were integrated in the experimental setup as shown in Figure 2. The experiment was designed to monitor the BM’s thermal behavior under natural convection, with all tests conducted at an ambient room temperature of 25 °C. A comprehensive thermal data acquisition system was employed, featuring an InfiRay P2 Pro thermal camera having 256 × 192 IR resolution with a 1112 °F high-temperature testing camera, positioned 150 mm above the BM’s top surface for capturing thermal contour maps. Data from the thermal camera were logged via InfiRay P2 Software 1.1.3.240316 version. A Semco S1 BCDS 99 V 20 A 1CH battery charge/discharger and analyzer, controlled by Semco Lithium Battery Charging System_4.3 software, served as the programmable load to subject the BM to constant current discharge cycles at various C-rates (e.g., 0.5C, 0.75C, 1C, 1.25C, and 1.5C). A relay is controlled through a terminal switching program-enabled microcontroller.

2.2.2. Experimental Procedure

The experiment focuses on characterizing heat generation and temperature distribution during discharge cycles. Prior to each test, the BM was fully charged using a standard CC-CV protocol. Test cases involved both single-terminal (1T) operation and dual-terminal (2T) operation with switching. In the 2T configuration, the switch between T1 and T2 was typically triggered at 50% depth of discharge (DoD), though other switching points based on DoD could be implemented as per specific test protocols. Thermal recording was collected continuously throughout each discharge to analyze the spatiotemporal temperature evolution of the regions (region1 = R1, region2 = R2, region3 = R3, region4 = R4) and zonal parts (T1− = Zone1/Z1, T1+ = Zone2/Z2, T2− = Zone3/Z3, T2+ = Zone4/Z4) of the BM as shown in Figure 3a,b. A consistent test case nomenclature (e.g., D_1C_1T_0S for discharge, for 1 C-rate, single terminal, zero switches; D_1C_2T_1S for discharge, for 1 C-rate, dual terminal, one switch) was adopted, with parameters detailed of features and states in a format similar to

Table 2. Specifications of the test case: features and states.

Battery State	C-Rate	Con-V (V)	Capacity (Ah)	Limited-V (V)	Limited-C (A)	Convection	Terminal Count	Terminal Switch Count
Con-C and V Charge	1	14.8	6	--	0.6	Natural	1, 2	0, 1
Con-C Discharge	0.5, 0.75, 1, 1.25, 1.5	--	3, 4.5, 6, 7.5, 9	11.2	--	Natural	1, 2	0, 1

, adapted for discharge rate and terminal switching under natural convection.

2.3. Theoretical Framework of Numerical Simulation: Heat Generation and Transfer

The thermal analysis within the battery module was based on fundamental principles of heat transfer and simplified models for battery heat generation [29], suitable for the scope of this study.

Heat Generation Model: The primary source of heat generation ($Q_{gen}$, W) within each cell during discharge was modelled using Joule’s Law, representing the irreversible heat due to internal resistance (R_i, Ω).

$$Q_{gen}=I^2\ast R_i$$

(1)

where $I$ is the current (A) flowing through the cell. The internal resistance (R_i) was considered as an effective value, potentially dependent on state of charge (SOC) and temperature, encompassing ohmic and polarization effects. This approach focuses on the dominant irreversible heating component. For use in spatial heat transfer calculations, this total heat generation was converted to a volumetric rate ($Q_{gen\text{\_}per\text{\_}vol}$, W/m³) based on the cell volume ($V_{cell}$, m³).

$$Q_{gen\text{\_}per\text{\_}vol}={\displaystyle \frac{Q_{gen}}{V_{cell}}}$$

(2)

Heat Transfer Model: The temperature distribution (T_(x,y,z,t)) within the battery module over time (t) is governed by the heat diffusion equation (conservation of energy).

$$\rho Cp{\displaystyle \frac{\partial T}{\partial t}}=\nabla \cdot (k\nabla T)+Qgen\text{\_}per\text{\_}vol$$

(3)

where $\rho$ is the density (2700 kg/m³), $Cp$ is the specific heat capacity (900 J/kg·K), $k$ is the thermal conductivity (1 W/m·K) of the cell materials (potentially treated as effective properties), and $Q_{gen\text{\_}per\text{\_}vol}$ is the volumetric heat generation rate from Equation (3).

Heat Dissipation (Boundary Condition): Heat transfer ($Q_t$, W) from the external surfaces of the cells or module to the ambient environment was modelled using Newton’s Law of Cooling.

$$Q_t=hA(T_{cell}-T_{amb})$$

(4)

where $h$ is the convective heat transfer coefficient (5 W/m²·K), $A$ is the surface area (0.224 m²), $T_{cell}$ is the surface temperature of the cell/module (K or 25 °C), and $T_{amb}$ is the ambient temperature (K or 25 °C). This equation defined the convective boundary conditions applied in the numerical simulation.

Lumped Parameter Energy Balance: For understanding the overall thermal behavior of a single cell treated as a lumped mass, the energy balance can be expressed as follows:

$$Q_{gen}=mc{\displaystyle \frac{\Delta T}{\Delta t}}+hA(T_{cell}-T_{air})$$

(5)

Here, $m$ is the cell mass (0.18 kg), c is the specific heat capacity (J/kg·K), $∆T$ represents the temperature change over a time interval, and the second term represents the convective heat loss to the surrounding air assuming ($T_{air}$ = $T_{amb}$). This equation relates the heat generated to the energy stored within the cell and the energy dissipated to the environment.

Analytical Temperature Approximation: Under simplifying assumptions of a constant heat generation rate ($Q_{gen}$), constant material properties ($m_{cell}$, C), a constant convective coefficient (h), and a uniform cell temperature, an analytical approximation for the cell temperature ($T_{cell}$) evolution over time (t) can be derived:

$$T_{cell}={\displaystyle \frac{Q_{gen}t+hA{T}_{amb}}{m_{cell}C+hA}}$$

(6)

Assuming Equation (5) as provided is intended, it offers a simplified analytical estimate for comparison or validation under specific conditions.

The results from the mathematical model having a dimension of 132 * 132 * 65 mm under natural convention simulation model are shown in Figure 4. Figure 4a presents the heat generation (W) in the battery module (BM), comparing simulated and experimental results. The experiments were conducted for both the STBM and the Se-DTS configuration. The comparison illustrates the simulated mathematical results alongside the STBM and Se-DTS models with respect to various C-rates. As the results illustrate, the simulated values are consistently lower than the real-time STBM experimental values, which can be attributed to several factors. First, the thermal properties of the cell materials (especially electrolyte and separator) are considered homogeneous in nature for simplification and were taken from the literature, which introduces uncertainty. Second, heat loss through the electrical interconnects and supporting fixtures was not fully modelled, whereas they contribute to additional dissipation in the experiment. Finally, slight variations in ambient airflow conditions can also influence the experimental measurements. However, as expected, the progression of heat generation follows the anticipated pattern, increasing from smaller to higher C-rates [30]. The results from the proposed Se-DTS configuration demonstrate superior performance in generating lower heat during the discharge cycles at higher C-rates. Specifically, heat generation at 1C, 1.25C, and 1.5C was reduced by 18.80%, 18.18%, and 12.89%, respectively. At lower C-rates, the performance is not better than STBM but remains closely comparable. Overall, the observed performance pattern indicates peak effectiveness at 1C, with performance gradually declining at both higher and lower C-rates. This suggests that the optimal operating region for Se-DTS under the 1S switching condition is at 1C-rate.

The average temperature increase, shown in Figure 4b, follows a pattern similar to the heat generation rate, with the mathematical model simulation values being lower than the experimental values due to the simplified modelling approach. The Se-DTS configuration demonstrates better performance at 1.25C and 1.5C, reducing the average temperature by 26.25% and 16.28%, respectively, compared to the STBM, while remaining closely comparable at lower C-rates. Figure 4c highlights the peak temperature patterns across varying C-rates, where Se-DTS outperforms STBM at higher C-rates of 1C, 1.25C, and 1.5C by 19.33%, 17.83%, and 12.72%, respectively. Analysis of the graphs clearly shows that the Se-DTS architecture achieves the objectives of reducing peak temperature (Figure 4a), improving thermal uniformity (Figure 4b), and controlling hot spot formation (Figure 4c) in the BM when compared to the STBM architecture.

3. Results

This section presents a comprehensive thermal analysis of lithium-ion battery discharge behavior under varying C-rates (0.5C, 0.75C, 1C, 1.25C, and 1.5C), and, as the charging process is considered to have a constant charging rate and no-load demand, the experimental test cases are focused on discharge cycles under various C-rates. The thermal response of the battery pack was investigated using time-sequenced temperature contour plots and temperature evolution graphs mapped across four distinct regions and four zones of the battery structure. Each test case captures the progression of temperature during discharge, with specific emphasis on identifying peak regional (P_rn) and zonal (P_Tn±) temperatures, critical switching events, and instances where local temperatures exceeded the predefined threshold (T_h), which was set at 5 °C above the initial temperature to simulate practical thermal safety limits [31]. Contour visualizations provide insight into spatial temperature distributions, while region-wise and zone-wise plots elucidate the temporal evolution of temperature gradients across the pack. Particular attention is given to the performance of terminal zones (Z1, Z2) and region R1, as they consistently exhibit the earliest and highest thermal activity due to current inflow positions, as seen in the traditional single-terminal approach. Switching instances, denoted by black dotted lines, and thermal threshold crossings, marked by colored dotted lines, further enrich the interpretation of thermal dynamics. The findings discussed herein establish a critical understanding of how discharge rates influence heat generation, distribution uniformity, and overall pack safety, laying the foundation for optimizing thermal design and control strategies in high-performance battery systems.

3.1. Thermal Characteristics of Conventional Battery Pack Under Discharge Rate: 0.5C Condition

As discussed in Equation (6) in Section 2.3, heat generation within the cell during the charging and discharging process is the primary source of thermal behavior in battery packs. Numerous analytical and experimental studies have concluded that cells within a pack exhibit non-uniform thermal characteristics depending on their position, which significantly affects the heat generation pattern [32]. Therefore, understanding the spatial distribution of heat generation and identifying regions with higher-than-expected heat density are crucial steps in designing an effective cooling strategy for battery packs. This highlights the importance of investigating heat generation patterns across different regions, particularly when evaluating the feasibility and stability of a multi-terminal approach under varying operating conditions.

At the lowest discharge rate of 0.5C, both terminal configurations exhibited minimal thermal stress. For the D_0.5C_1T (0S) case (single terminal, no switching) shown in Figure 5, mapping it to the progression of the temperature graph, shown in test case a and b, the temperature rise was gradual and slight throughout the discharge cycle (0–100% DoD). Peak temperatures remained low, with the highest regional average observed in region 2 (P_r2 = 29.3 °C) and the highest zonal peak in Z1 (29.2 °C), as detailed in Figure 6a and Figure 7a. Crucially, the pattern that emerged showcases R2 and R3 having higher temperature values compared to R1, indicating that inner regions have a higher thermal density; as the low C-rate runs for a longer duration, natural convection heat accumulates in the inner regions of the pack. Temperatures in all measured regions and zones stayed consistently below the 30 °C threshold (T_h), resulting in zero duration above T_h. Thermal gradients were minimal ($∆$P_r = 1.3 °C, $∆$P_Z = 1.1 °C), indicating relatively uniform heat distribution, corroborated by the heat maps showing diffuse, low temperatures. Initial temperature rise rates were slightly negative in some regions (e.g., R1: −0.020 °C/min,

Table 3. Heat dissipation, peak temperature, and temperature difference of battery pack at 0.5C discharge rate.

Test Case	Switch	R1 °C/min	R2 °C/min	R3 °C/min	R4 °C/min	△Pr	△Pz	Pr	Pz
D_0.5C_1T	0	−0.020	−0.032	−0.038	−0.042	1.3	1.3	29.3	29.2
D_0.5C_2T	0	0.142	0.088	0.064	0.110	2.1	1.9	27.2	27.5
	1	−0.088	−0.063	−0.031	−0.034	1	1.9	29	29.9

), possibly reflecting endothermic mixing effects or thermal equilibration, before becoming marginally positive.

In the D_0.5C_2T (1S) case (dual terminal with switching) as shown in Figure 7, the overall thermal behavior was similar. The terminal switching event (S1) occurred around 50% DoD (~3500 s) but induced no significant thermal perturbation due to the low temperature rise rate. Peak temperatures were comparable to the 1T case, with the highest regional average peak at 29 °C (R2) and the highest zonal peak at 29.9 °C (Z4, derived from plot annotations) as shown in Figure 6b and Figure 7b. As with the 1T case, the 30 °C threshold was never breached. Thermal uniformity remained high ($∆$Pr = <2.2 °C, $∆$P_Z = <1.9 °C, Table 3), and temperature rise rates were very low across all regions. The primary observable effect of switching was a subtle shift in the slightly warmer areas on the heat maps around 50% DoD.

Comparison (0.5C): At this low C-rate, both 1T and 2T configurations performed similarly from a thermal perspective, characterized by minimal heating, high uniformity, and no threshold exceedance. The terminal switching in the D_0.5C_2T (1S) case had a negligible impact on the overall thermal profile or peak temperatures, demonstrating that, under low-load conditions, the terminal configuration strategy is less critical for thermal management.

3.2. Thermal Characteristics of Conventional Battery Pack Under Discharge Rate: 0.75C Condition

Increasing the discharge rate to 0.75C resulted in a more pronounced temperature increase and the emergence of notable differences between terminal configurations as shown in Figure 8. In the D_0.75C_1T (0S) configuration, heat generation became more significant, particularly beyond 25% DoD. The peak regional average temperature reached 30 °C (R2, P_r2), while the peak zonal temperature was notably higher at 32.3 °C (Z1) as shown in Figure 9a and Figure 10a. This led to the T_h being breached. Specifically, Z1 > T_h for a substantial duration of 2520 s, and Z2 > T_h for 1610 s. Regional average temperatures did not register sustained durations above T_h (

Table 4. Heat dissipation, peak temperature, and temperature difference of battery pack at 0.75C discharge rate.

Test Case	Switch	R1 °C/min	R2 °C/min	R3 °C/min	R4 °C/min	△Pr	△Pz	Pr	Pz
D_0.75C_1T	0	0.077	0.037	0.004	0.037	2.9	4.7	30	32.3
D_0.75C_2T	0	0.153	0.092	0.045	0.089	2.4	3.3	29.7	31.9
	1	−0.010	−0.032	−0.044	−0.044	2	4	28.7	30.2

). Thermal gradients increased compared to 0.5C, particularly regionally ($∆$P_r = 2.9 °C), while zonal peak differences were higher ($∆$P_z = 4.7 °C). Temperature rise rates were consistently positive and higher than at 0.5C (e.g., R1: 0.077 °C/min, Table 4).

The D_0.75C_2T (1S) configuration also showed significant heating, clearly modulated by the terminal switching event (S1) near 50% DoD (~2300 s). Before switching, temperatures rose notably, particularly in areas corresponding to Z4 (P_r4 ≈ 29.7 °C). After switching, the thermal focus shifted, causing temperatures in Z1 and Z2 to increase further, reaching a peak zonal temperature of 33.1 °C (Z2, T1+). The peak regional average temperature recorded was 30.4 °C (Regions 1 and 2) as shown in Figure 9b and Figure 10b. Consequently, the 30 °C threshold was exceeded in all four zones (T_z1 > Th to T_z4 > Th), with durations distributed: Z1 (1190 s), Z2 (1260 s), Z3 (490 s), and Z4 (1680 s) (Table 4). The switching action reduces zonal non-uniformity ($∆$P_Z = 4.0 °C) compared to the 1T case, while the regional gradient was slightly lower ($∆$P_r = 2.9 °C).

Comparison (0.75C): Both configurations experienced significant heating and breached the 30 °C threshold. The D_0.75C_1T (0S) case developed a sustained hot spot leading to long threshold exceedance durations primarily in Zones 1 and 2. The D_0.75C_2T (1S) case, via switching, distributed the thermal stress across all four zones, resulting in a slightly lower peak zonal temperature (31.9 °C vs. 32.3 °C) and considerably lower zonal temperature gradients ($∆$P_Z = 4.7 °C vs. 3.3 °C). Switching at this C-rate effectively redistributed heat but did not prevent threshold exceedance and increased spatial uniformity.

3.3. Thermal Characteristics of Conventional Battery Pack Under Discharge Rate: 1C Condition

At a 1C discharge rate, temperatures exceeded the threshold, with the terminal configuration influencing the pattern of heat distribution and gradients as shown in Figure 11. The D_1C_1T (0S) case exhibited a steady temperature increase at regional level, reaching moderate peak levels. The highest regional average peak was observed in R2 (P_r1 = 31.4 °C), just above T_h, while Z1 recorded a peak of 36.9 °C (P_T1+,

Table 5. Heat dissipation, peak temperature, and temperature difference of battery pack at 1C discharge rate.

Test Case	Switch	R1 °C/min	R2 °C/min	R3 °C/min	R4 °C/min	△Pr	△Pz	Pr	Pz
D_1C_1T	0	0.066	0.064	0.041	0.032	3.4	8.9	31.4	36.9
D_1C_2T	0	0.09	0.087	0.10	0.090	1.7	3.3	29.6	31.1
	1	0.13	0.10	0.044	0.007	2.2	4.9	31.8	34.6

), exceeding the threshold by >6 °C. the duration above T_h was recorded as >3000 s for all zones, suggesting the exceedance was extreme by meeting the averaging criteria for duration calculation as shown in Figure 12a and Figure 13a. Thermal uniformity difference started to display relatively high ($\Delta P_R$ = 3.4 °C, $\Delta P_Z$ = 8.9 °C, Table 5).

In the D_1C_2T (1S) configuration, the temperature profile was again clearly influenced by the switching event (S1) at ~50% DoD (~1800 s). Before switching, Region 1 showed elevated temperatures (P_r3 ≈ 29.6 °C), nearly crossing the threshold. After switching, the heat concentration shifted, leading R1 to peak at 31.8 °C (P_r1). Zonal peaks reached 31.1 °C (Z3, P_T2-) before switching and 34.6 °C (Z2, P_T1+, Table 5) after switching. The 30 °C threshold was definitively crossed, with durations recorded for region 1 at 2350 s, region 3 at 2550 s, and region 2 at 2900 s; similarly, in zonal parts Z3, Z4, Z2, and Z1, the T_h was crossed at 900, 1090, 1920, and 2200 s as shown in Figure 12b and Figure 13b. In comparison, STBM crosses R1 > T_h at 2200 s and T_r3 at 2900 s; similarly, zonal level Z2 crosses at 470 s, and Z1 crosses at 1430 s. This clearly indicates that switching has a noticeable influence on the thermal distribution in the BM.

Comparison (1C): Both configurations operated near and above T_h. The D_1C_2T (1S) case maintained better thermal uniformity but experienced a zonal peak above T_h. The D_1C_1T (0S) case exhibited a clearly extreme scenario distributed by the non-switching event and displayed significantly higher thermal gradients ($\Delta P_r$ and $\Delta P_Z$). The peak regional temperature was slightly higher in the 2T case (31.4 °C vs. 31.8 °C). At 1C, switching began to show a more pronounced effect on temperature distribution and gradients ($∆$P_Z 8.9 vs. 3.3 °C).

3.4. Thermal Characteristics of Conventional Battery Pack Under Discharge Rate: 1.25C Condition

Discharge at 1.25C induced significant thermal challenges in both configurations as shown in Figure 14. The D_1.25C_1T (0S) case showed a rapid and substantial temperature rise, especially after 25% DoD. High peak temperatures were reached, with regional averages peaking at 33 °C (R1 and R2,

Table 6. Heat dissipation, peak temperature, and temperature difference of battery pack at 1.25C discharge rate.

Test Case	Switch	R1 °C/min	R2 °C/min	R3 °C/min	R4 °C/min	△Pr	△Pz	Pr	Pz
D_1.25C_1T	0	0.085	0.041	−0.017	0.009	3.7	11.7	33	40.7
D_1.25C_2T	0	0.289	0.111	0.044	0.072	2.5	8.9	30.4	37.0
	1	0.052	0.060	0.120	0.120	2.7	9.4	30.9	37.9

). A severe hot spot developed, leading to a zonal peak temperature of 40.7 °C (Z2, P_T1+). The 30 °C threshold (T_h) was significantly breached early in the discharge as shown in Figure 15a and Figure 16a. Sustained durations above T_h were recorded for region 1 (1050 s, R1 > T_h) and region 2 (420 s, R2 > T_h). Zonal exceedances were extensive: Zone 1 (2310 s, Z1 > T_h), Zone 2 (2590 s), and Zone 3 (1050 s). Marked thermal non-uniformity was observed, with significant regional ($∆$P_r = 3.7 °C) and particularly zonal gradients ($∆$P_Z = 11.7 °C, and >7 °C difference visible between peak Z2 and others in the plot). Temperature rise rates were considerably elevated (e.g., R1: 0.085 °C/min, R2: 0.041 °C/min, Table 6).

The D_1.25C_2T (1S) configuration also experienced rapid heating but comparatively less than the STBM configuration, with the switching event (S1) at ~50% DoD (~1300 s) clearly redistributing the intense heat. Before switching, Zones 1 and 2 were hottest, with peaks around 37 °C (P_T1+). After switching, the thermal load shifted, and Zones 3 and 4 became the hottest areas, peaking near 37.9 °C (P_T2+, Table 6). The highest recorded regional average peak was 30.9 °C (R2), while the maximum zonal peak reached 37.9 °C (Z4). The 30 °C threshold was crossed early and extensively across all four zones (T_z1 > Th to T_z4 > Th) as shown in Figure 15b and Figure 16b. Durations above T_h were substantial and distributed by switching: Z1 (840 s), Z2 (1890 s), Z3 (980 s), and Z4 (1120 s). Switching reduced thermal non-uniformity, leading to zonal gradients ($\Delta P_Z$ = 9.4 °C) and a reduction in regional gradients ($\Delta P_r$ = 2.7 °C). Temperature rise rates were higher in the first half and were reduced after switching (e.g., R1: 0.289 vs. 0.052 °C/min, Table 6).

Comparison (1.25C): At this high C-rate, both configurations faced severe thermal stress with prolonged threshold exceedances. The D_1.25C_1T (0S) case resulted in a significantly higher localized peak temperature (40.7 °C). The D_1.25C_2T (1S) case, through switching, mitigated this absolute peak (36.6 °C) but distributed the heat across all zones, leading to lower overall spatial temperature differences ($\Delta P_Z$ = 8.9 °C vs. ~11.7 °C). Switching provided a benefit in limiting the maximum single-point temperature and reduced hot spot generation.

3.5. Thermal Characteristics of Conventional Battery Pack Under Discharge Rate: 1.5C Condition

The highest tested discharge rate of 1.5C resulted in extreme heating and potentially hazardous conditions for both terminal configurations as shown in Figure 17. In the D_1.5C_1T (0S) case, the temperature rise was very rapid and substantial from early in the discharge. This configuration produced the highest temperatures overall, with regional averages peaking at 33.5 °C (R2, P_r) and an extreme zonal peak temperature of 42.3 °C (Z2, P_T1+,

Table 7. Heat dissipation, peak temperature, and temperature difference of battery pack at 1.5C discharge rate.

Test Case	Switch	R1 °C/min	R2 °C/min	R3 °C/min	R4 °C/min	△Pr	△Pz	Pr	Pz
D_1.5C_1T	0	0.182	0.137	0.079	0.082	4.2	12.9	33.5	42.3
D_1.5C_2T	0	0.339	0.200	0.122	0.111	3.9	10.9	32.2	39.2
	1	0.006	0.011	0.100	0.161	2.8	10.1	32	39.8

). The 30 °C threshold (T_h) was breached very quickly and for extended durations: R1 > T_h for 910 s, and R2 > T_h for 560 s. Zonal durations above T_h were significant: Zone 1 (1890 s, Z1 > T_h), Z2 (2100 s), and Z3 (910 s) as shown in Figure 18a and Figure 19a. Extreme thermal non-uniformity characterized this case, with high regional ($\Delta P_r$ = 4.2 °C) and zonal gradients ($\Delta P_Z$ = 12.9 °C based on table averages, but plots show >12 °C difference between peak Z2 and others). Temperature rise rates were the highest recorded (e.g., R1: 0.182 °C/min, R2: 0.137 °C/min, Table 7).

The D_1.5C_2T (1S) configuration also exhibited extremely rapid heating. The terminal switching (S1) at ~50% DoD (~1100 s) occurred amidst intense heat generation and served to redistribute the thermal load. Before switching, Z1 and Z2 were the hottest areas, peaking around 39.2 °C (P_T1+ from plot). After switching, the heat focus moved to Z3 and Z4, reaching a peak zonal temperature of approximately 40 °C (P_T2+, Table 7 reports 39.8 °C). The highest regional average peak noted was 32.2 °C (R1). The 30 °C threshold was exceeded, with R1 > T_h recording a duration of 140 s above T_h. All zones (T_z1 > T_h to T_z4 > T_h) experienced significant durations above the threshold: Z1 (280 s), Z2 (1890 s), Z3 (630 s), and Z4 (700 s). Thermal gradients were reduced to $\Delta P_Z$ = 10.1 °C, representing the lower zonal difference as shown in Figure 18b and Figure 19b. Regional gradients were also reduced ($\Delta P_r$ = 2.8 °C, Table 7). Temperature rise rates were high, with region 1 showing a rate of 0.339 °C/min.

Comparison (1.5C): Both configurations displayed dangerous levels of heating at 1.5C. The D_1.5C_1T (0S) configuration generated the highest absolute peak temperature (42.3 °C) due to localized heat accumulation. The D_1.5C_2T (1S) configuration, via switching, limited the maximum peak to ~40 °C; distributed heat across the cell, leading to the uniform zonal temperature difference ($\Delta P_Z$ = 10.1 °C); and was subjected to a reduction in hot spot generation for a prolonged duration above the threshold. At this extreme C-rate, while switching moderated the absolute peak, it created severe spatial temperature gradients and widespread thermal distribution.

3.6. Proof of Stability Using Discharge Cycle Voltage at Varying C-Rates

In this section, system stability was assessed using discharge cycle voltage profiles at various C-rates for both STBM and Se-DTS configurations. Figure 20 outlines the voltage characteristics, with STBM curves (solid lines) serving as the benchmark for 0.5C, 0.75C, 1C, 1.25C, and 1.5C, corresponding to discharge times of 120, 80, 60, 48, and 40 min and continuous discharge currents of 3, 4.5, 6, 7.5, and 9 Ah, respectively. The maximum charge voltage was 13.6 V, nominal voltage was 12.4 V, and the cut-off limit was 11.2 V. STBM discharge cycles showed no anomalies and were used as the reference for evaluating the proposed Se-DTS model.

Meanwhile, the Se-DTS discharge voltage curves (dotted lines) illustrate the 1S switching condition at varying C-rates, initiated at 50% DoD. Analysis of the graph shows that the overall BM performance is independent of the switching mechanism, with only minor voltage fluctuations observed. Each switching event observed at varying C-rates produces an average voltage drop of approximately 0.05V ($∆v$) during the terminal transition in the discharge cycle. Hence, it can be concluded that integrating dual-terminal design in BM architecture and using a software-enabled terminal switching mechanism does not impact system performance.

4. Discussion

In this study, a novel architectural design and well-structured test cases are proposed and explored to investigate the thermal aspects of a BM model built using a dual-terminal approach, controlled via a software-enabled switching mechanism under varying C-rates. The major findings are summarized below.

• Peak value in the BM: Our findings unequivocally demonstrate that increasing the discharge C-rate significantly elevates cell temperatures and accelerates temperature rise rates. Figure 21 illustrates the thermal distribution at the regional level for various C-rates vs. heat generation for simulated v/s experimental results. Figure 21a illustrates the simulation results of the BM with respect to regions at different C-rates, where the simplified model assumes uniform heat generation (W) that decreases proportionally from the terminal-positioned region to the remaining regions. Under the STBM switching condition, i.e., the conventional single-terminal model, for C-rates of 0.5C, 0.75C, 1C, 1.25C, and 1.5C discharge cycles, P_r values of 29.3, 30, 31.4, 32.44, 32.9, and 33.5 °C at the regional level were observed, respectively, and at the zonal level, P_r values of 29.2, 32.3, 38.9, 40.7, and 42.3 °C were observed, respectively. In comparison, the Se-DTS set-up showed an improvement in the regional level by up to 6.08% at 1.25C and 3.88% at 1.5C, while in the zonal level, an improvement of 11.31% was observed at 1C, 6.88% was observed at 1.25C, and 6.38% was observed at 1.5C, while at the remaining C-rates, similar characteristics were observed. Using Se-DTS, peak reduction in the regional and zonal levels can be improved, and it performs better in comparison to the STBM configuration.
• Set threshold temperature crossover duration: The threshold crossover occurs at different times at both regional and zonal levels, providing insights into hot spot generation locations and their behavior. At 0.5C, both STBM and Se-DTS behave ideally, maintaining the set threshold throughout the discharge cycles. At 0.75C, both STBM and Se-DTS remain ideal at the regional level. However, at the zonal level, T_z1 and T_z2 of the STBM configuration cross the threshold at 2120 s and 2500 s, respectively. This leads to the formation of hot spots in the pack, resulting in approximately 55.88% of the remaining discharge duration being above the desired temperature range. In contrast, for Se-DTS, T_z4 crosses earlier than STBM—by about 20%—but it does not remain above the threshold for the rest of the cycle. Instead, it retraces after reaching 50% DoD, while T_z2 crosses at 3450 s. This constrains and manages hot spot generation within the pack. From 1C onwards, the Se-DTS performance begins to clearly demonstrate superiority over STBM. In the STBM configuration, T_R1 crosses the threshold at 2200 s, and T_z2 crosses the threshold at 470 s. In comparison, Se-DTS shows T_R1 crossing at 2350 s and T_z3 crossing at 900 s; after a terminal switch, T_z2 crosses at 1920 s. This indicates an improvement of 6.82% at the regional level and 91.49% at the zonal level. At 1.25C, STBM crosses the threshold at 1350 s (T_R1), while Se-DTS maintains an ideal temperature range. At the zonal level, STBM crosses at 180 s (T_z4), whereas Se-DTS crosses at 310 s (T_z2), showing improvements of 100% at the regional level and 72.22% at the zonal level. Finally, at 1.5C, Se-DTS demonstrates an 80% improvement at the zonal level. This clearly highlights the superiority of Se-DTS over STBM.
• Thermal distribution in the battery pack module: Thermal distribution was assessed by measuring the difference between the maximum and minimum heat generation (W) across all four regions at each C-rate, as shown in Figure 21b,c. In the STBM configuration, the calculated $∆$T_d values were 1.58, 2.24, 3.47, 4.03, and 3.36 W at 0.5C, 0.75C, 1C, 1.25C, and 1.5C, respectively. In comparison, the corresponding Se-DTS values were 0.90, 1.57, 1.35, 0.56, and 1.12 W. These results demonstrate a consistent reduction in heat generation in the Se-DTS configuration, indicating improved thermal uniformity and confirming its superior performance over STBM.

Figure 21. Region-wise thermal distribution for various C-rates vs. heat generation. (a) Simulated model, (b) single terminal BM, (c) multi-terminal BM.

Figure 21. Region-wise thermal distribution for various C-rates vs. heat generation. (a) Simulated model, (b) single terminal BM, (c) multi-terminal BM.

5. Conclusions

In summary, the Se-DTS approach proves to be stable and effective in sustaining and maintaining thermal characteristics much better than the STBM approach in reducing and controlling thermal peak values and hot spot generation and distribution in the BM. This study presents a detailed thermal characterization of lithium-ion battery packs subjected to various discharge rates, ranging from 0.5C to 1.5C, under two different configurations—single terminal (1T) and dual terminal (2T). The results clearly demonstrate the progressive impact of increased C-rates on temperature rise, threshold crossings, and spatial thermal gradients within the BM. Key findings show that, as the discharge rate increases, the system transitions from a thermally stable state to a more dynamic thermal environment that requires active management.

At low C-rates (0.5C and 0.75C), the thermal response remains within acceptable limits for both configurations, although the 2T setup consistently provides more uniform temperature distribution and lower peak temperatures. As the discharge rate increases to 1C and beyond, the benefit of the 2T configuration becomes increasingly evident. The dual-terminal design successfully delays threshold temperature crossings, reduces peak zone and region temperatures, and minimizes the thermal differential across the pack ($\Delta P_r$ and $\Delta P_Z$). This is made possible through the activation of a software-controlled switching mechanism, which dynamically redistributes current pathways and helps mitigate heat accumulation near terminal zones.

The 1T configuration consistently shows a more aggressive temperature rise, with earlier threshold breaches and larger thermal gradients, particularly around terminal-adjacent zones (Z1, Z2) and Region 1 (R1). On the other hand, the 2T configuration exhibits superior thermal resilience, demonstrating smoother temperature slopes, later threshold crossings, and improved thermal balance across all regions. The data also show that, as the discharge rate increases, the frequency and timing of switching events become critical in managing thermal loads efficiently.

Overall, this research confirms that a multi-terminal battery architecture with an intelligent switching control offers a scalable, efficient, and passive method to enhance the thermal performance of Li-ion battery systems. By enabling dynamic control over current pathways, the proposed design mitigates hot-spot formation and extends the operational safety window, making it highly suitable for high-demand applications where thermal management is a limiting factor.

As a follow-up to our previous work on the multi-terminal approach in the BM with a constant C-rate and multiple switching mechanisms within the same discharge cycle, we extend this method to study scenarios where the switch count is kept constant while varying the C-rate. This study focuses on exploring the thermal characteristics and stability of a multi-terminal switching strategy at different C-rates under controlled laboratory conditions. These controlled conditions allowed us to isolate and better understand the behavior of the proposed architecture. However, we acknowledge that real-world EV operations involve more complex and dynamic charging/discharge patterns, which were beyond the scope of this work. Another important aspect not addressed in this study is the potential for long-term cell aging due to frequent switching and altered current paths—factors such as impedance growth or lithium plating could become significant over time. In future studies, we aim to investigate long-term system performance. Such investigations could provide deeper insights into the full potential and limitations of the proposed design under real-world conditions. Furthermore, future work may extend this study by integrating the switching logic into real-time battery management systems (BMS) and validating performance under diverse ambient and load conditions.