Novel Hybrid Thermal Management System for High-Power Lithium-Ion Module for Electric Vehicles: Fast Charging Applications

: Lithium-ion capacitors (LiC) are hybrid energy storage systems (ESS) combining the advantages of lithium-ion batteries and electric double-layer capacitors, including longer lifetime, high power, and energy densities. LiCs are popular for high-power applications where fast charge and discharge driving proﬁles are demanded from electric vehicles (EV). However, LiCs generate excess heat when they are exposed to fast charging/discharging proﬁles. Therefore, a robust thermal management system (TMS) is crucial, in order to ensure reliable operation. In this study, a novel hybrid TMS based on air-cooling system assisted phase change materials (PCM), heat pipes, and a heat sink is proposed for an LiC module under a 150 A continuous current proﬁle. A very thin aluminum heat sink and ﬂat copper heat pipes were added to the PCM to increase its thermal conductivity. An experimental test bench of the proposed TMS was developed, and the temperature distribution of the module for each of the individual LiC cells was studied. The maximum temperature of the module under natural convection, when there was not any cooling system, reached almost 59.8 ◦ C. The experimental results showed that after using the proposed hybrid TMS, the hottest cell reached 36.18 ◦ C while the coldest cell reached 35.54 ◦ C. Therefore, 39.5% improvement could be seen during the whole charge and discharge process after 3000 s. Moreover, the temperature difference within the module, of four LiCs, was around 0.64 ◦ C, which was exceptional.


Introduction
Nowadays, global warming and the usage of fossil fuels are at the worldwide center of attention [1]. The transportation industry is one of the main sectors which is highly connected to fossil fuels [2]. Therefore, electric vehicles (EVs) and hybrid electric vehicles (HEVs) are the most appropriate substitutes, with very low fuel consumption and emission [3]. Lithium-ion (Li-ion) batteries are classified as a reliable option for EVs and HEVs due to their features [4]. Nevertheless, their chemistry is not suitable for high-current profiles [5]. Therefore, lithium-ion capacitors (LiCs) have been introduced to the market [6], which have the benefits of Li-ion batteries and electric double-layer capacitors (EDLC) [7]. Thus, LiCs have the high power and high energy densities, which makes the LiC technology a suitable applicant for high-current profiles [8]. However, despite great potential, LiCs generate heat during the process in high-current profiles [9]. Thermal runaway and nontemperature uniformity damage the LiCs and affect their performance and lifetime [10]. Thus, their heat generation needs to be controlled by a proper thermal management system (TMS) [11,12]. TMS is an inevitable part of each electric and hybrid vehicle to ensure safety and reliability [13]. Many studies evaluated the effect of different thermal management keep the absolute temperature of the module below 40 • C, and at the same time, decrease the temperature difference between the coldest cell and the hottest cell.

Design of Hybrid HPA-TMS
As shown in Figure 1, the proposed hybrid HPA-TMS consisted of four prismatic lithium-ion capacitors (LiC), a paraffin phase change material (PCM), six flat heat pipes, and a fan for air cooling. A data logger and K-type thermocouples were responsible for monitoring the temperature evolution during the cycling profile. The battery tester and battery cycler were responsible for cycling the module.
World Electr. Veh. J. 2022, 13, x FOR PEER REVIEW 3 of 15 and to avoid hotspots on individual cells. The LiC should work under 40 °C regarding the thermal limits announced by the manufacturer (JSR Company, ULTIMO). Therefore, the main aim was to keep the absolute temperature of the module below 40 °C, and at the same time, decrease the temperature difference between the coldest cell and the hottest cell.

Design of Hybrid HPA-TMS
As shown in Figure 1, the proposed hybrid HPA-TMS consisted of four prismatic lithium-ion capacitors (LiC), a paraffin phase change material (PCM), six flat heat pipes, and a fan for air cooling. A data logger and K-type thermocouples were responsible for monitoring the temperature evolution during the cycling profile. The battery tester and battery cycler were responsible for cycling the module. A computer was linked to the data logger and the tester, to get the voltage, current, and temperature of the cells, based on a continuous 150 A charge and discharge current rate. Before the commencement of the tests, the PCM was mounted inside the climate chamber, with a temperature more than the melting temperature of the PCM. Then, the melted PCM was poured inside the PVC container, with very low thermal conductivity (around 0.2 W/m.K) between the LiC cells. In addition, six flat heat pipes were stuck to the thin aluminum heat sink through a very tiny layer of thermal interface material (TIM). In this study, thermal grease with thermal conductivity of 8 W/m.K was selected to be the TIM. There were two LiC cells (dual-cells) on each side of the container, so that the module consisted of four LiCs. Each side of the dual-cells was covered with a 10 mm layer of PCM to absorb the heat loss of the module, and to release it to the ambient through the heat pipes and the aluminum heat sink. The process of absorbing/rejecting the heat loss of the module was a perfect method of making thermal insulation to avoid thermal runaway.
The flat heat pipes and the aluminum heat sink were the added secondary materials for increasing the very low thermal conductivity of the PCM. In this respect, a very thin layer of the aluminum heat sink was added between the dual-cells and the PCM, forming A computer was linked to the data logger and the tester, to get the voltage, current, and temperature of the cells, based on a continuous 150 A charge and discharge current rate. Before the commencement of the tests, the PCM was mounted inside the climate chamber, with a temperature more than the melting temperature of the PCM. Then, the melted PCM was poured inside the PVC container, with very low thermal conductivity (around 0.2 W/m·K) between the LiC cells. In addition, six flat heat pipes were stuck to the thin aluminum heat sink through a very tiny layer of thermal interface material (TIM). In this study, thermal grease with thermal conductivity of 8 W/m.K was selected to be the TIM. There were two LiC cells (dual-cells) on each side of the container, so that the module consisted of four LiCs. Each side of the dual-cells was covered with a 10 mm layer of PCM to absorb the heat loss of the module, and to release it to the ambient through the heat pipes and the aluminum heat sink. The process of absorbing/rejecting the heat loss of the module was a perfect method of making thermal insulation to avoid thermal runaway.
The flat heat pipes and the aluminum heat sink were the added secondary materials for increasing the very low thermal conductivity of the PCM. In this respect, a very thin layer of the aluminum heat sink was added between the dual-cells and the PCM, forming a sandwich structure. The top surface of the heat pipes was exposed to the airflow coming from the fan. Since the air loop was arranged outside of the LiC module, the complexity of the system was reduced greatly, which bespoke the feasibility of the proposed system as compared with the traditional system of having the air-cooling system inside the module.
The next stage was to again mount the PVC container full of melted PCM inside the climate chamber, at a constant temperature of 23 • C. This was to help the PCM solidify and reach 23 • C, which was the initial temperature for starting the experimental tests. After solidifying the PCM, the voltage and current cable were connected to the battery tester. The thermocouples were also connected to the LiC cells from one side and to the data logger from the other side. The test bench of the hybrid HPA-TMS was then ready for commencement of the cycling tests. The test profile was a continuous 150 A fast charging and discharging without any pause. Table 1 lists the specifications of the employed LiC cells, the flat heat pipes, the heat sink, the PCM, and the PVC container. Figure 2 exhibits the whole process of the generated heat absorption and removal by the proposed hybrid HPA-TMS. Temperature is one of the most vital parameters that affect the performance of LiCs [42,43] by causing irreversible effects, including accelerated aging [44], impurities production [45], and evaporation of solvent [46][47][48]. LiCs generate more heat with smaller ESR, since they have been used in high power applications under high current rates [49]. Therefore, the lifetime and performance of LiCs significantly depend on their working temperature. In this context, a robust TMS is quite essential to control the thermal behavior of LiCs [8]. As can be seen from Figure 2, the heat was generated inside the LiC cells in the module due to internal resistance and entropy change in the system, which were counted as reversible and irreversible heat sources. Then, the generated heat was absorbed by the PCM, but could be fully rejected to the ambient due to the low thermal conductivity of the paraffin PCM. Therefore, the combination of the heat pipes and the heat sink absorbed the ingested heat from the PCM. The heat sink and heat pipes together increased the heat dissipation rate from the module to the ambient. As can be seen from Figure 2, the heat was generated inside the LiC cells in the module due to internal resistance and entropy change in the system, which were counted as reversible and irreversible heat sources. Then, the generated heat was absorbed by the PCM, but could be fully rejected to the ambient due to the low thermal conductivity of the paraffin PCM. Therefore, the combination of the heat pipes and the heat sink absorbed the ingested heat from the PCM. The heat sink and heat pipes together increased the heat dissipation rate from the module to the ambient.
Generally, each heat pipe had three sections including the evaporator section, the condenser section, and the adiabatic section. The stored heat inside the PCM was transferred to the aluminum heat sink, which was absorbed quickly by the heat pipes due to their evaporator section. Then, the heat loss was moved to the condenser section of the heat pipe, and escaped to the ambient through the airflow of the fan. The efficiency of the whole process of heat loss absorption and rejection was enhanced by this method, which improved the performance of the proposed thermal management system.

Heat Transfer Model
In this work, uniform heat generation was considered for the LiC cell during charging and discharging continuous current rate [50]: where , , , , and denote the generated heat, resistance, current rate, temperature, entropy change, and volume of the LiC. The term represents the heat loss caused by the electrochemical reaction inside the LiC, which can be considered as a constant value since its products were very low compared to the heat loss due to the resistance.
To model the heat transfer, the energy equation was utilized to get the temperature information of the LiC [51,52]: Generally, each heat pipe had three sections including the evaporator section, the condenser section, and the adiabatic section. The stored heat inside the PCM was transferred to the aluminum heat sink, which was absorbed quickly by the heat pipes due to their evaporator section. Then, the heat loss was moved to the condenser section of the heat pipe, and escaped to the ambient through the airflow of the fan. The efficiency of the whole process of heat loss absorption and rejection was enhanced by this method, which improved the performance of the proposed thermal management system.

Heat Transfer Model
In this work, uniform heat generation was considered for the LiC cell during charging and discharging continuous current rate [50]: where Q gen , R, I, T LiC , ∂U ∂T and V LiC denote the generated heat, resistance, current rate, temperature, entropy change, and volume of the LiC. The term IT LiC ∂U ∂T represents the heat loss caused by the electrochemical reaction inside the LiC, which can be considered as a constant value since its products were very low compared to the heat loss due to the resistance.
To model the heat transfer, the energy equation was utilized to get the temperature information of the LiC [51,52]: where ρ denotes density, C p is the specific heat, and k represents the thermal conductivity.
For the PCM, the enthalpy-porosity method was used for the heat transfer and melting process. The momentum and continuity equations for the Newtonian incompressible molten PCM, which was considered as a laminar flow, is expressed as [53]: where ρ PCM and → υ are the density and velocity vector of the PCM. Moreover, the energy equation can be expressed as [54]: where k PCM and H denote the thermal conductivity and the enthalpy of the PCM. In addition, β represents the liquid volume fraction of the PCM, and ε = 0.001. The constant ε is a decimal that helps the equation avoid being divided by zero. Lastly, A mushy denotes the constant of the mushy zone that was happening between the liquid and solid phases of the PCM. This constant reflects the melting behavior of the PCM, which was considered as 10 5 .
For the enthalpy of the PCM, the sensible heat and the latent heat of fusion have been summed as follows: where h re f , T re f , C PCM and L represent the reference enthalpy, reference temperature, specific heat of the PCM, and latent heat of fusion, respectively. Therefore, the liquid fraction of the PCM (β) was calculated as follows: where T s and T L are the solid phase and liquid phase temperatures. In the equation, 0 means fully solid PCM and 1 means fully liquid PCM. The parameters of the air-cooling system show that the airflow is laminar. Therefore, the energy, continuity, and momentum equations for the air cooling can be expressed as follows [21]: where ∂ρ air and → υ air denote the density and flow rate of air, and: where C air , k air , and T air represent the specific heat, thermal conductivity, and temperature of the air. For the heat pipe, the whole system of the used heat pipes was considered as a uniform thermal conductor, that had an equivalent thermal conductivity, which can be expressed as: where k hp denotes the equivalent thermal conductivity of the heat pipes. Q hp , A hp and ∆T hp are the amount of heat that can be transferred by the heat pipes, the cross-sectional area, and the temperature difference between the condenser and evaporator sections, respectively. Also, l hp is the equivalent length of the heat pipe, that can be expressed as: where l c , l e and l a denote the length of the condenser section, the evaporator section, and the adiabatic section, respectively. Therefore, the heat transfer equation of the heat pipe can be written as:

Initial Conditions and Boundaries
The initial temperature of T test = 23 • C was considered for the experimental tests at t = 0 s. Therefore, T test = T LiC = T PCM = T hp = T air = T amb = T hs , in which T LiC , T PCM , T hp , T air , T amb , T hs denote the temperature of LiC, PCM, heat pipe, airflow, ambient, and heat sink, respectively. There was an interface between the LiC and the PCM in the system. Thus, the boundary condition of this interface was considered as: For the interface between the PCM and the heat pipes and heat sink: For the interface between the heat sink, the heat pipe, and the air-cooling system: Finally, for the interface between the heat pipe and the air-cooling system: where k LiC , k hp , k hs , k air , and k PCM denote the thermal conductivity of the LiC, the heat pipes, the heat sink, the airflow, and the PCM. Moreover, ∂T ∂n is the temperature gradient. The thermal resistance was negligible where each boundary was supposed to be an ideal boundary condition. Moreover, the boundary condition between the LiC module and the ambient is expressed as: where h denotes the heat transfer coefficient, which was considered as 13 W/m 2 ·K in this work. Such a low heat transfer coefficient was assumed to be natural convection, since the tests were performed in a room with a constant temperature of 23 • C.

Results and Discussion
The experimental tests of the proposed HPA-TMS were conducted under a 150 A continuous cycling profile, where the current rate corresponded to the charge and discharge current that was applied to the cell. Figure 3 shows the current and voltage applied to the module. Using the electro-thermal model that had been developed in our previous works [55,56], the heat loss of the system under such a harsh driving cycle was calculated. The generated heat under the continuous 150 A current rate is exhibited in Figure 4. The generated heat of the LiC regarding its dimensions for a 150 A fast charging and fast discharging profile is listed in Table 2, which shows that the generated heat of the LiC was 3.5 times larger than the normal heat generation condition of LiBs [57].
continuous cycling profile, where the current rate corresponded to the charge and discharge current that was applied to the cell. Figure 3 shows the current and voltage applied to the module. Using the electro-thermal model that had been developed in our previous works [55,56], the heat loss of the system under such a harsh driving cycle was calculated. The generated heat under the continuous 150 A current rate is exhibited in Figure 4. The generated heat of the LiC regarding its dimensions for a 150 A fast charging and fast discharging profile is listed in Table 2, which shows that the generated heat of the LiC was 3.5 times larger than the normal heat generation condition of LiBs [57].    continuous cycling profile, where the current rate corresponded to the charge and discharge current that was applied to the cell. Figure 3 shows the current and voltage applied to the module. Using the electro-thermal model that had been developed in our previous works [55,56], the heat loss of the system under such a harsh driving cycle was calculated. The generated heat under the continuous 150 A current rate is exhibited in Figure 4. The generated heat of the LiC regarding its dimensions for a 150 A fast charging and fast discharging profile is listed in Table 2, which shows that the generated heat of the LiC was 3.5 times larger than the normal heat generation condition of LiBs [57].     To cycle the LiC module under a 150 A intense driving profile, a battery tester was used to acquire the data, by using a computer linked to the tester. This experimental test bench monitored the module's current, voltage, and temperature, along with other parameters during testing. The battery tester handled the test's definition, and the current rate's value was implemented according to the cell requirements [58]. The cell was located inside the climate chamber to stabilize its initial temperature at 23 • C before the experimental test to investigate the cell's temperature evolution. The charging and discharging of the cell was carried out utilizing the testers in which the cell was cycled by the continuous current rate of 150 A with a 3000 s current profile.
To calculate the error of the experiments, and to check the accuracy of the results, the Schultz and Cole method was employed for the uncertainty analysis, which can be expressed as [59,60]: where U V I denotes the error of every single factor and U R stands for total errors. The maximum uncertainty is expressed in percentage, which was calculated as 1.17%, which was perfect. Four K-type thermocouples were used, such that one thermocouple was pasted on the top surface of each individual cell, to measure the maximum temperature of each cell, as can be seen in Figure 2b. The temperature difference between each cell of the module was determined by the surface temperatures. In this regard, the maximum temperature means the highest temperature value that a cell was able to obtain on its top surface, while the temperature difference refers to the temperature uniformity. Figure 5 exhibits the temperature evolution of the LiC module, based on the temperature of four cells and the ambient temperature that was monitored by the thermocouples and the data logger connected to the computer. To cycle the LiC module under a 150 A intense driving profile, a battery tester was used to acquire the data, by using a computer linked to the tester. This experimental test bench monitored the module's current, voltage, and temperature, along with other parameters during testing. The battery tester handled the test's definition, and the current rate's value was implemented according to the cell requirements [58]. The cell was located inside the climate chamber to stabilize its initial temperature at 23 °C before the experimental test to investigate the cell's temperature evolution. The charging and discharging of the cell was carried out utilizing the testers in which the cell was cycled by the continuous current rate of 150 A with a 3000 s current profile.
To calculate the error of the experiments, and to check the accuracy of the results, the Schultz and Cole method was employed for the uncertainty analysis, which can be expressed as [59,60]: where denotes the error of every single factor and stands for total errors. The maximum uncertainty is expressed in percentage, which was calculated as 1.17%, which was perfect.
Four K-type thermocouples were used, such that one thermocouple was pasted on the top surface of each individual cell, to measure the maximum temperature of each cell, as can be seen in Figure 2b. The temperature difference between each cell of the module was determined by the surface temperatures. In this regard, the maximum temperature means the highest temperature value that a cell was able to obtain on its top surface, while the temperature difference refers to the temperature uniformity. Figure 5 exhibits the temperature evolution of the LiC module, based on the temperature of four cells and the ambient temperature that was monitored by the thermocouples and the data logger connected to the computer. The ambient temperature was also recorded to check if the room temperature was kept constant at its initial temperature of 23 °C. Table 3 lists the maximum temperature of each individual cell within the module after cooling down by the hybrid HPA-TMS during 3000 s. As can be seen, the hottest cell reached 36.18 °C while the coldest cell reaches The ambient temperature was also recorded to check if the room temperature was kept constant at its initial temperature of 23 • C. Table 3 lists the maximum temperature of each individual cell within the module after cooling down by the hybrid HPA-TMS during 3000 s. As can be seen, the hottest cell reached 36.18 • C while the coldest cell reaches 35.54 • C. Therefore, during the whole charge and discharge process, the temperature difference within the module of four LiCs was around 0.64 • C, which was exceptional. LiC structure is made for heavy-duty applications, where high power is demanded. The performance of the proposed TMS is more evident when comparing the results with the thermal behavior of the module before using any TMS. The maximum temperature of the module is achieved under natural convection when there is not any TMS. The results show that, in this case, the maximum temperature of the module exceeded 59.8 • C, which is harmful for the lifetime of the module since the cycle life of the LiCs would be diminished. Comparing the temperature results of the system under natural convection, and the system after using the HPA-TMS, shows a reduction of 39.5% in the maximum temperature of the module. Therefore, under a high-power driving profile of 150 A charge and discharge without any rest, the proposed TMS works perfectly to ensure the safe and reliable operation of such technology. The temperature evolution of the LiC, when only PCM is used as the TMS, has been reported in our previous work [61], and is not reported here to avoid redundancy. Likewise, the thermal behavior of the LiC using the same heat pipes has been reported in our previous work [62]. The results show that the maximum temperature of the LiC, employing only the PCM or only the heat pipe TMSs after 1400 s, reaches 40.8 • C and 48.8 • C, respectively. These results are for a single LiC cell. Nevertheless, for a module of four cells, the maximum temperature using only the PCM, or only the heat pipe, would be far higher. On the other hand, in the hybrid HPA-TMS, the maximum temperature of the hottest point in the module after 1400 s reached 33.6 • C, which was a 17.6% enhancement compared with the PCM, and a 31.1% improvement compared with the heat pipe TMS. Moreover, the hybrid HPA-TMS results prove that even with the double time, after 3000 s, the proposed hybrid system can fully control the maximum temperature of the module.
Another critical challenge for LiCs is the temperature non-uniformity that is the main reason for electrical imbalance and SoC mismatch among the module cells [63]. Therefore, special attention was given to designing a robust TMS to manage the temperature rise of the LiCs during high charging and discharging.
At the beginning of the test, the temperature of the module (all the LiC cells) rose sharply due to the electrochemical reactions inside the LiCs. However, the temperature difference within the cells of the module remained almost the same below 1 • C. In the case of testing the module with the same load profile under natural convection when there was not any TMS, the temperature difference between neighboring cells in the module was more than 7 • C. This trend proved that more than 85.7% improvement is achieved for the temperature uniformity when the hybrid HPA-TMS is employed.
In addition, the temperature of the LiC cells was recorded using IR images thanks to the thermal camera having an error of less than ±0.2 • C. As is illustrated in Figure 6, the maximum temperature of the LiC at the end of the cycling profile (3000 s) was around 36.9 • C, while the current cables reached more than 42 • C. The proposed HPA-TMS proves that combining the heat absorption ability of the PCM with a secondary material that can reject the stored heat of the PCM leads to the high performance of the TMS. This is because the very low thermal conductivity of the paraffin PCM is not able to release the absorbed heat. In this scenario, the PCM gets hotter and cannot reject the stored heat after reaching its evaporation point. Thus, at this point, the PCM is not able to receive the generated heat of the LiC module, which will result in accumulated temperature inside the system and thermal runaway. On the other hand, if a secondary passive system is used to receive the stored heat of the PCM, the heat can be rejected to the environment easily. However, passive systems alone are not good candidates in high power applications where the heat convection coefficient is very low, and the flow is natural convection. In such conditions, the temperature of the module can be reduced, but not efficiently below the safe limit. Therefore, an active cooling system is required to increase the thermal performance and reject the accumulated heat of the system.
The used air-cooling system increases the heat transfer rate by enhancing the heat convection coefficient, and thereby enhancing the thermal performance. The module without any cooling system under natural convection reaches almost 59.8 °C; its lifetime will be reduced in such operating temperatures [62]. Comparing the results of the natural convection and the hybrid HPA-TMS shows that the proposed hybrid system has 39.5% efficiency, which is ideal.
The temperature of the heat pipes and the heat sink was monitored employing IR thermal camera images. Figure 7 depicts the temperature distribution of the system where the focus is on the heat pipes and the heat sink. As can be seen, the maximum temperature of the heat pipes after a 3000 s cycling profile reached almost 29.6 °C, meaning that a 6.6 °C increase happened inside the heat pipes.  On the other hand, if a secondary passive system is used to receive the stored heat of the PCM, the heat can be rejected to the environment easily. However, passive systems alone are not good candidates in high power applications where the heat convection coefficient is very low, and the flow is natural convection. In such conditions, the temperature of the module can be reduced, but not efficiently below the safe limit. Therefore, an active cooling system is required to increase the thermal performance and reject the accumulated heat of the system.
The used air-cooling system increases the heat transfer rate by enhancing the heat convection coefficient, and thereby enhancing the thermal performance. The module without any cooling system under natural convection reaches almost 59.8 • C; its lifetime will be reduced in such operating temperatures [62]. Comparing the results of the natural convection and the hybrid HPA-TMS shows that the proposed hybrid system has 39.5% efficiency, which is ideal.
The temperature of the heat pipes and the heat sink was monitored employing IR thermal camera images. Figure 7 depicts the temperature distribution of the system where the focus is on the heat pipes and the heat sink. As can be seen, the maximum temperature of the heat pipes after a 3000 s cycling profile reached almost 29.6 • C, meaning that a 6.6 • C increase happened inside the heat pipes. On the other hand, if a secondary passive system is used to receive the stored heat of the PCM, the heat can be rejected to the environment easily. However, passive systems alone are not good candidates in high power applications where the heat convection coefficient is very low, and the flow is natural convection. In such conditions, the temperature of the module can be reduced, but not efficiently below the safe limit. Therefore, an active cooling system is required to increase the thermal performance and reject the accumulated heat of the system.
The used air-cooling system increases the heat transfer rate by enhancing the heat convection coefficient, and thereby enhancing the thermal performance. The module without any cooling system under natural convection reaches almost 59.8 °C; its lifetime will be reduced in such operating temperatures [62]. Comparing the results of the natural convection and the hybrid HPA-TMS shows that the proposed hybrid system has 39.5% efficiency, which is ideal.
The temperature of the heat pipes and the heat sink was monitored employing IR thermal camera images. Figure 7 depicts the temperature distribution of the system where the focus is on the heat pipes and the heat sink. As can be seen, the maximum temperature of the heat pipes after a 3000 s cycling profile reached almost 29.6 °C, meaning that a 6.6 °C increase happened inside the heat pipes.  The results prove that the maximum temperature of the module was negatively correlated with the temperature uniformity within the LiC module, as high thermal performance leads to a high dissipation rate, which resulted in higher temperature uniformity due to higher temperature difference between the neighboring cells in the module.

Conclusions
In this study, a novel hybrid TMS based on air cooling system assisted phase change materials (PCM), six flat heat pipes, and a thin aluminum heat sink was proposed for an LiC module under a 150 A continuous current profile. The heat sink and heat pipes were added to the PCM to increase its low thermal conductivity. The experimental test bench of the proposed system was developed, and the temperature distribution of the module for each of the individual LiC cells was studied. The experiments produced the following results: • The maximum temperature of the module under natural convection when there is not any cooling system reached almost 59.8 • C. • After using the proposed hybrid TMS, the hottest cell reached 36.18 • C while the coldest cell reached 35.54 • C. Therefore, a 39.5% improvement was seen during the whole charge and discharge process after 3000 s. • Using only PCM, the maximum temperature of a single cell reached 40.8 • C. Therefore, the HPA-TMS had a 17.6% better performance than the only PCM. • Using only heat pipe, the maximum temperature of a single cell reached 48.8 • C. Therefore, the HPA-TMS had 31.1% better performance than the only heat pipe.
The temperature difference within the module of four LiCs was around 0.64 • C, which was exceptional.