- Email: [email protected]
EV lithium-ion battery
Applied Thermal Engineering 63 (2014) 551e558
Contents lists available at ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
Experimental investigation on the feasibility of heat pipe cooling for HEV/EV lithium-ion battery Thanh-Ha Tran a, b, c, *, Souad Harmand a, b, Bernard Desmet a, b, Sebastien Filangi c a
Université Lille Nord de France, 59000 Lille, France UVHC, TEMPO, 59313 Valenciennes, France c PSA PEUGEOT CITROEN, France b
h i g h l i g h t s Constant heat flux was applied to a flat heat pipe cooling system. Its thermal performance was compared with that of a heat sink under several cooling conditions. The influence of the inclination was evaluated. The heat pipe transient behaviour was also studied under variable input power. Heat pipe was found to be an effective and low-energy solution for HEV/EV battery cooling.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 June 2013 Accepted 24 November 2013 Available online 2 December 2013
In this paper, the use of flat heat pipe as an effective and low-energy device to mitigate the temperature of a battery module designed for a HEV application was investigated. For this purpose, nominal heat flux generated by a battery module was reproduced and applied to a flat heat pipe cooling system. The thermal performance of the flat heat pipe cooling system was compared with that of a conventional heat sink under various cooling conditions and under several inclined positions. The results show that adding heat pipe reduced the thermal resistance of a common heat sink of 30% under natural convection and 20% under low air velocity cooling. Consequently, the cell temperature was kept below 50 C, which cannot be achieved using heat sink. According to the space allocated for the battery pack in the vehicle, flat heat pipe can be used in vertical or horizontal position. Furthermore, flat heat pipe works efficiently under different grade road conditions. The transient behaviour of the flat heat pipe was also studied under high frequency and large amplitude variable input power. The flat heat pipe was found to handle more efficiently instant increases of the heat flux than the conventional heat sink. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Heat pipe Transient input power Cooling system Lithium-ion battery Hybrid electric and electric vehicle
1. Introduction Lithium-ion batteries possess a high energy density compared with other secondary batteries; consequently, they are highly recommended as power sources for hybrid and electric vehicles (HEV/ EV) to provide longer driving range and faster acceleration. However, lithium-ion batteries have relatively poor performance at low temperature and may suffer thermal runaway at high temperature. For security reasons, a battery thermal management system (BTMS) always includes an internal switch which is opened if the battery is
* Corresponding author. Université Lille Nord de France, 59000 Lille, France. Tel.: þ33 620005163. E-mail addresses: [email protected], [email protected] (T.-H. Tran). 1359-4311/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.applthermaleng.2013.11.048
operated outside of its operating temperature range. This can prevent fire or explosion risks but the battery would be temporarily unavailable. Fuel economy of HEV and the driving range of EV are consequently affected. Furthermore, lithium-ion batteries calendar life and cycle life degrade quickly if kept or used at high temperature. The goal of a cooling system is to keep the battery pack within its operating temperature range. The temperature distribution within the pack should be even because temperature gradient could lead to different ageing levels between cells and therefore, different charge/ discharge behaviours for each cell. For a Li-ion battery, the desired operating temperature range is between 25 C and 50 C and the temperature gradient should be less than 5 C [1]. In addition, the cooling system has to meet the requirements of the vehicle such as: reliable, compact, lightweight, easily accessible for maintenance, low cost, and low power parasitic consumption.
552
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Nomenclature A Cpcell Cpresin Cpwall dUo/dT dt hwall I i ma mb mcell mwall P qcell/a qcell/b qa/a qa/b
cooling wall area [m2] cell specific heat capacity [kJ kg1 K1] resin specific heat capacity [kJ kg1 K1] aluminium shell specific heat capacity [kJ kg1 K1] cell entropy coefficient [V K1] time step [s] overall equivalent heat transfer coefficient [W m2 K1] cell current intensity [A] number of the surrounding nodes [e] mass of the matrix node a [kg] mass of the matrix node b [kg] cell mass [kg] aluminium shell mass [kg] heat input used for experiment [W] heat transfer rate from cell to its corresponding resin matrix node a [W] heat transfer rate from cell to its corresponding resin matrix node b [W] heat transfer rate between 2 resin matrix nodes a [W] heat transfer rate between resin matrix nodes a and b [W]
Up to now, battery cooling systems may use air, liquid (water/ oil/refrigerant), phase change materials (PCM), or a combination of these methods. Each solution has its advantages and weaknesses. Air cooling solution can be passive (i.e., only the ambient environment is used) or active (i.e., a built-in source provides heating and/or cooling). The obvious benefit of air-cooled systems is the elimination of on-board chiller unit and coolant pump; leading to savings in parasitic power consumption and weight. However, air cooling system may use fan. If a high flow rate is needed, air cooling may cause noise problem in the vehicle. Furthermore, due to its poor thermal conductivity and thermal capacity, air convection can’t be sufficient for heat dissipation from battery under stressful and abuse conditions. The non-uniform distribution of temperature within battery pack is also inevitable [2]. Compared with air cooling, liquid cooling offers higher cooling capacity at similar parasitic pump/fan power [3] but is heavier, and costlier due to the use of pump, tank, heat exchanger. Maintenance and repair of liquid cooling systems are also complicated and costly. PCM systems have high thermal energy storage capacity thanks to the use of latent heat and therefore can effectively limit the temperature rise and the temperature uniformity of the battery [4]. However, the weak point that has limited widespread use of PCM system is its insufficient long term thermal stability. It is well known that heat pipe has very high thermal conductivity and can maintain homogeneously the evaporator surface at nearly constant temperature. This device also has flexible geometry which can fit variable area spaces. Moreover, heat pipe has infinite life and does not require any maintenance. Theses attractive characteristics make heat pipe a promising candidate for HEV/EV battery cooling. Previously, Mahefkey et al. [5] has judged heat pipe to be suitable to mitigate the temperature of NieCd. Similar conclusion has been drawn by Zhang et al. [6] for Ni-MH battery. Concerning Liion battery, Wu et al. [7] have reported that the cell temperature could be significantly reduced using heat pipe with aluminium fin on the condenser section, especially with the help of a cooling fan at the condenser section. More recently, Rao et al. [8] have investigated experimentally the cooling performance of tube heat pipes with
qb/a
heat transfer rate between resin matrix nodes b and a [W] qb/b heat transfer rate between 2 resin matrix nodes b [W] qb/wall heat transfer rate from resin matrix node b to the module wall [W] qwall/wallheat transfer rate between 2 module wall nodes [W] R1 thermal resistance of the experimented system [ C W1] R2 thermal resistance of the full size system [ C W1] Ta temperature of the matrix node a [ C] Tamb ambient temperature [ C] Tb temperature of the matrix node b [ C] Tcell cell temperature [ C] Twall temperature of aluminium wall [ C] Uo open-circuit potential [V] U measured cell potential [V] Greek symbols: radial thermal conductivity of the cell [W m1 K1] thermal conductivity of the resin matrix [W m1 K1] thermal conductivity of the aluminium shell [W m1 K1] F heat generated by cell [W]
lcell lresin lwall
condenser sections cooled by a water module. The battery maximum temperature has been controlled below 50 C when the heat generation rate was lower than 50 W. Coupled with the desired battery temperature gradient, the heat generation rate should not exceed 30 W. In other words, with well-designed heat pipes, the temperature rise and temperature difference of power batteries can be effectively controlled within desired range. However, only tube heat pipe had been experimented and the behaviour of heat pipe under high frequency variable power input had not been studied. In this work, the use of flat heat pipe as an effective and lowenergy cooling solution for a HEV application lithium-ion battery was investigated. Both constant and transient heat input conditions were experimented. The thermal performance of heat pipe cooling was compared with that of a conventional heat sink in order to highlight the thermal performance enhancement of the heat pipe. 2. Experimental set-up 2.1. The battery module description We considered the use of flat heat pipe in order to mitigate the temperature of a 14 cylindrical cells battery module designed for a specific HEV application. Cells (7 Ah of capacity, 38 mm in diameter, 142 mm in height) were implemented in a thermally conductive and electrically insulating resin matrix. The matrix helps keep cells in place and increase the mechanical rigidity of the module. Moreover, the matrix enhances the heat transfer between the cells and the aluminium module walls, as well as the electric insulation between the cells. During charge and discharge, heat generated by cells can be considered to be the sum of the resistive heat and the entropic heat [9e 11]. Consequently, the global heat generated can be determined by:
f ¼ IðU0 UÞ IT
dU0 dT
(1)
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
553
achieved by applying cooling devices to the two larger ones (162 mm in height by 291 mm in large) as shown in Fig. 2. 2.2. Experimental set-up Instead of using directly the battery module as heat source for experiment, a flat heater was used to reproduce the nominal heat flux and an induction heating machine was employed to generate the transient heat flux. Two laboratory size cooling systems were investigated. The first system was a simple copper heat sink. The heat sink base surface measured 39 mm in length by 16 mm in width. In the second cooling system, the previous heat sink was glued with thermal grease on the condenser surface of a flat heat pipe. The structure of the heat pipe is presented in Fig. 3. The copper envelope of the heat pipe was under partial vacuum. The capillary structure was made of sintered copper powder. The effective permeability of the capillary structure was 7 109 m2 and its hydraulic diameter was 40 mm. The heat pipe uses water as working fluid. The peripheral surfaces of both the systems were insulated; hence, heat was dissipated in to the surround air only through the fins. Temperatures at different locations were measured for each system using three K-type thermocouples which have uncertainty of þ/1.5 C (Fig. 4).
Fig. 1. Heat profile generated from cells.
Variable heat rejected by cells in the battery module during standard driving cycle was calculated according to Eq. (1). The nominal value has been found to be slightly over 100 W (Fig. 1). In order to maintain the cells within their optimum temperature range, heat generated needs to be evacuated through the module walls. The two walls corresponding to the two ends of the cells could not be used for cooling purpose. Indeed, one of those was intended for bus bars and electric module installation and the other was used for degassing system implement. Among the 4 remaining walls of the battery module, optimal thermal performance could be
2.2.1. Tests under constant input power We first performed experiments under constant input power condition. After stabilization, the nominal heat flux arrived to the module cooling walls corresponds to the nominal heat generated
Resin matrix
Cell
Cooling systems Upper end of cells: Bus bars and electric module installation
Lower end of cells: Degassing cover & channel
Aluminium shell Fig. 2. Battery module.
6 mm
16 mm
39 mm Copper envelope
Condenser Capillary structure
Evaporator Fig. 3. Schematic of the experimented heat pipe.
Water filling tube
554
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Heat pipe Heat sink
Heat sink
Insulation
Heat input
K-type thermocouples Heat input
Fig. 4. Schematic of heat sink and heat pipe cooling systems.
by cells divided to the cooling walls surface area and equals to 0.115 W/cm2. We wanted to test the cooling systems under the same heat flux as that arrived to the module cooling wall; consequently, the total heat input P of 0.72 W was reproduced using a 23 U flat heater and applied to both the cooling systems. Their thermal performance was evaluated under various cooling conditions and under several inclined positions. Under forced convection conditions, air was blown toward the fins using a flow channel. Fig. 5a shows air blown over the fins for cooling system in vertical position. For cooling system in horizontal position, air flowed through the fins as shown in Fig. 5b. Measurements of the air velocity were made along the fin array using a hot-wire anemometer with measuring range from 0.15 to 30 m/s (Fig. 5a and b). The accuracy of the anemometer was 3% of reading þ0.05 m/s for air flow ranging from 0.15 to 3 m/s and 3% of reading þ0.2 m/s for air flow ranging from 3.1 to 30 m/s. Three air velocities were experimented (0.2 m/s, 0.5 m/s and 1.5 m/s). 2.2.2. Tests under transient input power Finally, the transient behaviour of heat pipe was studied under variable input power produced using an induction heating machine with a reactive power at inductor of 32 kVAR. The output power of the induction generator could be piloted between 10% and 100% of its maximal capacity using a 0-10 V controller. A specific work coil was designed according to the shape of the conductive work surface to be heated. In induction heating, heat absorbed by the work surface depends on the distance between the latter and the coil. According to the power range of this application, the coil was placed at
a distance of 10 mm to the surface to be heated of each cooling system as shown in Fig. 6 (cooling system was placed in horizontal position, insulation and thermocouples are not represented). A thermal infrared camera was used in order to verify that the magnetic field generated by the current in the coil did not affect the thermocouples measurement accuracy. In this regard, work surfaces on which thermocouples were fixed using aluminium tape were coated with 0.95 emissivity black paint. It was found that in our experiment, temperatures indicated by thermocouples corresponded to that given by thermal camera, whether the induction heating system was turned on or shut off. Each test condition was experimented three times and the reproducibility has been found to be good. The temperature deviation between tests under iso conditions was lower than 2.0 C. For all of the tests, the temperatures measured at different locations of the heater/cooling system interface were very close to each other and the margin was less than 1.0 C. Hence, only the mean temperature was reported in this paper. 3. Results and discussion 3.1. AMESim model of the battery module In a configuration such as shown in Fig. 2, cooling devices are not in direct contact with cells but the battery module walls. Consequently, temperature measured at the heater/cooling system interface in our experiment corresponds to the battery module walls temperature and there will be a temperature gradient
a
Flat heater
Air velocity measurements
Fan
b Fig. 5. a e Schematic of the air channel with cooling device in vertical position. b e Flow direction for cooling device in horizontal position (Insulation and thermocouples are not represented).
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
555
Induction heating coil
Fig. 6. Schematic of cooling systems under induction heating coil.
Overall equivalent heat transfer coefficient (hwall)
Adiabatic wall
Adiabatic wall
Nodes b
Nodes a
Overall equivalent heat transfer coefficient (hwall)
Resin matrix (ma, mb, Cpresin, λresin)
Cells (mcell, Cpcell, λcell, Φ)
Aluminum Shell (mwall, Cpwall, λwall)
Fig. 7. Battery module scheme and data used for AMESim modelling.
between module walls and cells. In order to estimate cell temperature at different locations of the module, a numerical model of the module was developed using AMESim software [12]. This model took into account the module geometry, the mass, the thermal characteristics of different components (cell, resin matrix, aluminium shell) and finally the heat generated from cells. It was assumed that the temperature is homogeneous within each cell. A 28 elements mesh was generated for 2D modelling of the resin matrix. Each cell was surrounded by 2 resin matrix nodes: a and b (Fig. 7). The aluminium shell was modelled using 14 elements corresponding to the 14 exterior resin matrix nodes b. The two smaller walls of the module were considered adiabatic. In order to model cooling systems applied to the two larger walls, whether heat sink or heat pipe system was used, an overall equivalent heat transfer coefficient was employed. Hence, the cell energy balance was expressed as below:
dT mcell Cpcell cell ¼ f qcell/a qcell/b dt
(2)
The energy balances for resin matrix nodes a and b are the following:
ma Cpresin
X dTa ¼ qcell/a qa/ai qa/b dt i
ma Cpre sin
X dTb ¼ qcell/b qb/a qb/bi qb/wall dt
(3) (4)
i
Finally, the energy balance for the wall is:
mwall Cpwall
X dTwall ¼ qb/wall qwall/walli dt i
hwall AðTwall Tamb Þ
(5)
The overall equivalent heat transfer coefficient could be determined by fitting the steady state wall temperature after
stabilization of the model with the heater/cooling system interface temperatures measured experimentally. The cell temperature at different locations of the module could also be evaluated using this AMESim model. In order to quantify the thermal performance of the cooling systems used in our experiment, we calculated the thermal resistance according to:
R1 ¼
ðT Tamb Þ P
(6)
The tested cooling systems were laboratory small size prototypes. The thermal resistance of the full size cooling system to be implemented in the battery module could be calculated from the overall equivalent heat transfer coefficient given by the AMESim model:
R2 ¼
1 hwall A
(7)
3.2. Tests in horizontal position In horizontal position, under constant input power condition and natural convection, heat pipe cooling was found to be far more effective than heat sink cooling. After stabilization, the heater/heat sink interface temperature reached 46.0 C whereas that of the heater/heat pipe interface was kept at 38.0 C (Fig. 8). The thermal resistance of both the systems was reported in Table 1. Adding heat pipe to a common heat sink reduced the thermal resistance of the cooling system of about 30%. According to these experimental temperatures, respective overall equivalent heat transfer coefficient was computed for each cooling system thank to the AMESim model aforementioned. Consequently, cell temperature at different locations of the module could be determined. Results were reported in Table 1. The
556
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Fig. 8. Heater/cooling system interface temperatures under natural and forced convection.
Table 1 Natural convection: Temperatures after stabilization for heat sink and heat pipe cooling.
Module wall temperature (measured experimentally) Thermal resistance R1 (calculated from experimental data) Coolest cell temperature (estimated using AMESim model) Hottest cell temperature (estimated using AMESim model) Thermal resistance R2 (estimated using AMESim model) Stabilization time (estimated using AMESim model) Cell temperature goes over 50 C after (estimated using AMESim model)
Heat sink
Heat pipe
46.0 C 36 C/W
38.0 C 25 C/W
55.7 C
48.4 C
57.3 C
49.5 C
0.480 C/W 0.332 C/W 3500 s 800 s
2500 s e
Fig. 9. Heater/heat pipe interface temperatures in horizontal and vertical positions.
when the temperature was lower or equal to 29 C, contrarily to what is intended. The partial vacuum created while producing this heat pipe was certainly higher than what it should have been. According to the temperature measured under the first cooling condition, the operating temperature of the experimented heat pipe was certainly between 29 C and 33 C. Even though the prototype heat pipe did not work at temperature close to that of the ambient, this does not call into question the benefits of using heat pipe under natural convection and under the first cooling condition. Overall, heat pipe can be an effective and low-energy solution for lithium-ion batteries cooling. The suppression/reduction of the parasitic power consumption in battery thermal management helps increase the HEV/EV autonomy and helps reduce it running cost and thus contributes to the wider use of HEV/EV. 3.3. Tests in vertical and inclined positions
temperature gradient within the pack was mainly due to the larger volume of the resin matrix nodes surrounding the cells in the 4 corners. These cells were the coolest and the hottest were those in the centre of the module. We noticed that cell temperatures exceed the upper operating limit which (50 C) with heat sink cooling while all cells were kept below 50 C with heat pipe cooling. Three forced convection configurations were also performed; 20.0 C air was blown at different velocities, from 0.2 m/s to 0.5 m/s and 1.5 m/s. We remarked that under the first cooling condition, the thermal performance of the heat pipe was better than that of the heat sink (20% reduction of the thermal resistance). Unfortunately, the results under the second and the third cooling conditions showed that heat pipe did not perform better than heat sink (Table 2). The flat heat pipe was aimed to be able to work even under ambient temperature (w20 C). However, the above experimental results make us believe that the phase change did not occur
Heat pipe was tested in vertical position and no significant evolution in cooling performance was noticed compared with that in horizontal position (Fig. 9). This result means that the flat heat pipe tested can be used in both vertical and horizontal positions, which leaves a degree of freedom open to best match the battery pack with the space allocated for it in the vehicle. The evaluation of the battery cooling system performance under uphill and downhill drive conditions is necessary for a vehicle application. Common roads are usually graded below 10% (5.71 ). Inclines of over 20% (11.31 ) are rare and maximum known incline on public streets is 37% (20.31 ). Hence, in this part, heat pipe in horizontal and vertical positions were inclined with an angle of 10 and 20 . Fig. 10 shows a very slight evolution of the heater/heat pipe interface temperature between various incline positions. The temperature deviation was found to be less or equal to 2.0 C. This
Table 2 Forced convection: Temperatures after stabilization for heat sink and heat pipe cooling. Heat sink
Module wall temperature Thermal resistance R1 Coolest cell temperature Hottest cell temperature Thermal resistance R2 Stabilization time
Heat pipe
0.2 m/s
0.5 m/s
1.5 m/s
0.2 m/s
0.5 m/s
1.5 m/s
37.0 C 24 C/W 47.5 C 48.4 C 0.313 C/W 3000 s
29.0 C 13 C/W 39.5 C 40.4 C 0.166 C/W 1700 s
24.0 C 5.6 C/W 34.5 C 35.4 C 7.37 102 C/W 1300 s
33.5 C 19 C/W 43.9 C 44.8 C 0.249 C/W 2500 s
29.0 C 13 C/W 39.5 C 40.5 C 0.166 C/W 1700 s
25.0 C 6.9 C/W 35.5 C 36.4 C 9.30 102 C/W 1300 s
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
557
Fig. 10. Heater/heat pipe interface temperatures in various inclined positions under natural convection.
means flat heat pipe works efficiently under different grade road conditions. In Ref. [13], the authors had also found out that thanks to the capillary system, when the working fluid was not overcharged, the orientation of the heat pipe had little effect on its thermal performance. 3.4. Tests under transient condition In real life driving conditions, battery power solicitation profiles are always dynamic. Consequently, heat profiles generated from batteries are of large amplitude and high frequency, unlike those produced by other electronic and electric devices. Fig. 1 is a typical example of how heat generated by cells has highly variable amplitudes and short durations. For battery module with cells inserted in a resin matrix and cooling systems applied to module walls such as shown in Fig. 2, transient heat arrived to cooling devices was computed using the AMESim model presented in previous section. We remarked in Fig. 11 that the heat profile at cooling wall has slight fluctuation and very smooth shape. This is due to large heat capacity of the cells and the resin matrix. Consequently, the influence of the variation of heat flux on the heat pipe behaviour would be very slim. However, in case of larger variation in cell heat loss or lower module heat capacity or direct contact between cooling device and cells, heat arrived to cooling system may vary rapidly and have larger amplitude; hence, the transient behaviour of the heat pipe would be more prominent and investigation of this transient behaviour is necessary. Fig. 11 shows the heat profile at cooling
Fig. 11. Heat profile at module wall and at cell wall.
system/cell wall interface, obtained by AMESim modelling for configuration where cooling device is in direct contact with cells. The corresponding heat flux was reproduced repeatedly using induction heating machine in order to study the transient behaviour of the flat heat pipe. Fig. 12a shows the work surface temperature for both cooling systems. The heat flux applied was also added in this figure. We can see that it consists of six identical cycles of 1460 s duration each. Fig. 12b zooms in on the last cycle of Fig. 12a. We remark that the work surface temperature in heat pipe cooling case was easily lower than that in heat sink cooling case. After stabilization, in heat pipe cooling case, the mean temperature obtained during a cycle of variable heat flux (38.4 C) corresponded to the steady state temperature measured under nominal constant input heat flux in the previous section (38.0 C). On the contrary, in heat sink cooling case, the mean temperature obtained during a cycle of variable heat flux (50.4 C) was about 4 C higher than the temperature measured under nominal constant input heat flux (46.0 C). Furthermore, the temperature fluctuation within cycle was also smaller in heat pipe cooling case. These points make us believe that in transient regime, heat pipe handled more efficiently instant increases of the heat flux than conventional heat sink. Consequently, it can keep the temperature close to that obtained with a nominal heat flux with slight fluctuation (Table 3). 3.5. Further discussions Up to now, the weak point that has limited widespread use of heat pipe system is its high cost due to the complicated fabrication process and the use of copper, an expensive metal, as wick and wall material. However, recent researches on aluminium heat pipe manufacturing [14,15] have revealed efficient and reliable way to decrease the heat pipe cost. Furthermore, the use of aluminium also helps reduce the weight of the cooling system, which is highly appreciable in HEV/EV application. In order to validate the use of heat pipe in automotive application, aside from the impact of the inclination due different grade roads, investigation on the influence of the vehicle shock and vibration to the heat pipe thermal performance is unavoidable. In our study, the flat heat pipe was not tested under vibration environment; however, the influence of vibration on the heat pipe heat transfer has been investigated in some relevant studies. Connors and Zunner [16] have investigated the thermal performance of a flat heat pipe with copper powder wick structure. The flat heat pipe has demonstrated no degradation of the thermal performance under military vehicle shock and vibration conditions. More recently, Guo et al. [17] have investigated the influence of mechanical vibration on the heat transfer characteristics for rectangular microgrooves. It has been found that the vibration movement enlarge the wetting area
558
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Fig. 12. a) Heater/cooling system interface temperatures under transient heat flux (during 6 cycles, in horizontal position, under natural convection). b) Heater/cooling system interface temperatures under transient heat flux (zoom-in on the last cycle of a).
Table 3 Transient heat flux: Steady state work surface temperature for heat sink and heat pipe cooling. Work surface temperature
Heat sink
Heat pipe
Mean value for 1 cycle Maximum value for 1 cycle Minimum value for 1 cycle Temperature variation for 1 cycle
50.4 C 52.4 C 48.0 C 4.3 C
38.4 C 39.9 C 36.5 C 3.4 C
and intensified consequently the heat transfer of the microgroove. Consequently, the heat pipe seems to be able to work under vibration environment without any thermal performance degradation. 4. Conclusion In order to extend the battery life and decrease HEV/EV cost, battery should be operated within its optimum temperature range with low power parasitic consumption. The use of flat heat pipe as an effective and low-energy cooling device for a HEV application lithium-ion battery was considered under different conditions. According to experimental results, the following conclusions could be drawn: (1) Flat heat pipe can be an effective and low-energy solution for battery cooling. Adding heat pipe reduced the thermal resistance of a common heat sink of 30% under natural convection and 20% under low air velocity cooling. Consequently, the cell temperature was kept below 50 C, which cannot be achieved using heat sink. (2) According to the space allocated for the battery pack in the vehicle, flat heat pipe can be used in vertical or horizontal position. Furthermore, flat heat pipe works efficiently under different grade road conditions. (3) Under high frequency and large amplitude variable heat generated condition, heat pipe handled more efficiently instant increases of the heat flux than conventional heat sink and minimized the temperature fluctuation during transient cycle.
References [1] A.A. Pesaran, Battery thermal models for hybrid vehicle simulations, J. Power Sources 110 (2002) 377e382. [2] A.A. Pesaran, An approach for designing thermal management systems for electric and hybrid vehicle battery packs, in: Fourth Vehicle Thermal Management Systems Conference and Exhibition London, National Renewable Energy Laboratory, 1999. Retrieved from, www.nrel.gov/docs/fy99osti/25992. pdf (accessed 13.11.13). [3] A.A. Pesaran, Battery thermal management in EVs and HEVs: issues and solution, in: Advanced Automotive Battery Conference, National Renewable Energy Laboratory, 2001. Retrieved from, www.nrel.gov/vehiclesandfuels/ energystorage/pdfs/aabc_lv.pdf (accessed 13.11.13). [4] R. Sabbah, R. Kizilel, J.R. Selman, S. Al-Hallaj, Active (air-cooled) vs. passive (phase change material) thermal management of high power lithium-ion packs: limitation of temperature rise and uniformity of temperature distribution, J. Power Sources 182 (2008) 630e638. [5] E.T. Mahefkey, M.M. Kreitman, An intercell planar heat pipe for removal of heat during the cycling of a high rate nickel cadmium battery, J. Electrochem. Soc. 118 (1971) 1382e1386. [6] G. Zhang, Z. Wu, Z. Rao, L. Fu, Experimental investigation on heat pipe cooling effect for power battery, Chem. Ind. Eng. Prog. 28 (2009) 1165e1168 (in Chinese). [7] M.S. Wu, K.H. Liu, Y.Y. Wang, C.C. Wan, Heat dissipation design for lithium-ion batteries, J. Power Sources 109 (2002) 160e166. [8] Z. Rao, S. Wang, M. Wu, Z. Lin, F. Li, Experimental investigation on thermal management of electric vehicle battery with heat pipe, Energ. Convers. Manage. 65 (2013) 92e97. [9] D. Bernardi, E. Pawlikowski, J. Newman, A general energy balance for battery system, J. Electrochem. Soc. 132 (1985) 5e12. [10] L. Rao, J. Newman, Heat generation rate and general energy balance for insertion battery systems, J. Electrochem. Soc. 144 (1997) 2697e2704. [11] T.H. Tran, S. Harmand, B. Desmet, F. Pailhoux, Thermal characterization of LiNi0.8Co0.15Al0.05O2/graphite battery, J. Electrochem. Soc. 160 (2013) 775e780. [12] www.lmsintl.com/LMS-Imagine-Lab-AMESim (LMS International, (accessed 13.11.13). [13] Y. Xuan, Y. Hong, Q. Li, Investigation on transient behaviors of flat plate heat pipes, Exp. Therm. Fluid Sci. 28 (2004) 249e255. [14] Y.T. Chen, S.W. Kang, Y.H. Hung, C.H. Huang, K.C. Chien, Feasibility study of an aluminum vapor chamber with radial grooved and sintered powders wick structures, Appl. Therm. Eng. 51 (2013) 864e870. [15] M. Ameli, B. Agnewa, P.S. Leung, B. Ng, C.J. Sutcliffe, J. Singh, R. McGlen, A novel method for manufacturing sintered aluminium heat pipes (SAHP), Appl. Therm. Eng. 52 (2013) 498e504. [16] M.J. Connors, J.A. Zunner, The use of vapor chambers and heat pipes for cooling military embedded electronic devices, in: Military Communications Conference, IEEE, 2009, pp. 1474e1480. [17] C. Guo, X. Hu, W. Cao, D. Xu, D. Tang, Effect of mechanical vibration on flow and heat transfer characteristics in rectangular microgrooves, Appl. Therm. Eng. 52 (2013) 385e393.
Contents lists available at ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
Experimental investigation on the feasibility of heat pipe cooling for HEV/EV lithium-ion battery Thanh-Ha Tran a, b, c, *, Souad Harmand a, b, Bernard Desmet a, b, Sebastien Filangi c a
Université Lille Nord de France, 59000 Lille, France UVHC, TEMPO, 59313 Valenciennes, France c PSA PEUGEOT CITROEN, France b
h i g h l i g h t s Constant heat flux was applied to a flat heat pipe cooling system. Its thermal performance was compared with that of a heat sink under several cooling conditions. The influence of the inclination was evaluated. The heat pipe transient behaviour was also studied under variable input power. Heat pipe was found to be an effective and low-energy solution for HEV/EV battery cooling.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 June 2013 Accepted 24 November 2013 Available online 2 December 2013
In this paper, the use of flat heat pipe as an effective and low-energy device to mitigate the temperature of a battery module designed for a HEV application was investigated. For this purpose, nominal heat flux generated by a battery module was reproduced and applied to a flat heat pipe cooling system. The thermal performance of the flat heat pipe cooling system was compared with that of a conventional heat sink under various cooling conditions and under several inclined positions. The results show that adding heat pipe reduced the thermal resistance of a common heat sink of 30% under natural convection and 20% under low air velocity cooling. Consequently, the cell temperature was kept below 50 C, which cannot be achieved using heat sink. According to the space allocated for the battery pack in the vehicle, flat heat pipe can be used in vertical or horizontal position. Furthermore, flat heat pipe works efficiently under different grade road conditions. The transient behaviour of the flat heat pipe was also studied under high frequency and large amplitude variable input power. The flat heat pipe was found to handle more efficiently instant increases of the heat flux than the conventional heat sink. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Heat pipe Transient input power Cooling system Lithium-ion battery Hybrid electric and electric vehicle
1. Introduction Lithium-ion batteries possess a high energy density compared with other secondary batteries; consequently, they are highly recommended as power sources for hybrid and electric vehicles (HEV/ EV) to provide longer driving range and faster acceleration. However, lithium-ion batteries have relatively poor performance at low temperature and may suffer thermal runaway at high temperature. For security reasons, a battery thermal management system (BTMS) always includes an internal switch which is opened if the battery is
* Corresponding author. Université Lille Nord de France, 59000 Lille, France. Tel.: þ33 620005163. E-mail addresses: [email protected], [email protected] (T.-H. Tran). 1359-4311/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.applthermaleng.2013.11.048
operated outside of its operating temperature range. This can prevent fire or explosion risks but the battery would be temporarily unavailable. Fuel economy of HEV and the driving range of EV are consequently affected. Furthermore, lithium-ion batteries calendar life and cycle life degrade quickly if kept or used at high temperature. The goal of a cooling system is to keep the battery pack within its operating temperature range. The temperature distribution within the pack should be even because temperature gradient could lead to different ageing levels between cells and therefore, different charge/ discharge behaviours for each cell. For a Li-ion battery, the desired operating temperature range is between 25 C and 50 C and the temperature gradient should be less than 5 C [1]. In addition, the cooling system has to meet the requirements of the vehicle such as: reliable, compact, lightweight, easily accessible for maintenance, low cost, and low power parasitic consumption.
552
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Nomenclature A Cpcell Cpresin Cpwall dUo/dT dt hwall I i ma mb mcell mwall P qcell/a qcell/b qa/a qa/b
cooling wall area [m2] cell specific heat capacity [kJ kg1 K1] resin specific heat capacity [kJ kg1 K1] aluminium shell specific heat capacity [kJ kg1 K1] cell entropy coefficient [V K1] time step [s] overall equivalent heat transfer coefficient [W m2 K1] cell current intensity [A] number of the surrounding nodes [e] mass of the matrix node a [kg] mass of the matrix node b [kg] cell mass [kg] aluminium shell mass [kg] heat input used for experiment [W] heat transfer rate from cell to its corresponding resin matrix node a [W] heat transfer rate from cell to its corresponding resin matrix node b [W] heat transfer rate between 2 resin matrix nodes a [W] heat transfer rate between resin matrix nodes a and b [W]
Up to now, battery cooling systems may use air, liquid (water/ oil/refrigerant), phase change materials (PCM), or a combination of these methods. Each solution has its advantages and weaknesses. Air cooling solution can be passive (i.e., only the ambient environment is used) or active (i.e., a built-in source provides heating and/or cooling). The obvious benefit of air-cooled systems is the elimination of on-board chiller unit and coolant pump; leading to savings in parasitic power consumption and weight. However, air cooling system may use fan. If a high flow rate is needed, air cooling may cause noise problem in the vehicle. Furthermore, due to its poor thermal conductivity and thermal capacity, air convection can’t be sufficient for heat dissipation from battery under stressful and abuse conditions. The non-uniform distribution of temperature within battery pack is also inevitable [2]. Compared with air cooling, liquid cooling offers higher cooling capacity at similar parasitic pump/fan power [3] but is heavier, and costlier due to the use of pump, tank, heat exchanger. Maintenance and repair of liquid cooling systems are also complicated and costly. PCM systems have high thermal energy storage capacity thanks to the use of latent heat and therefore can effectively limit the temperature rise and the temperature uniformity of the battery [4]. However, the weak point that has limited widespread use of PCM system is its insufficient long term thermal stability. It is well known that heat pipe has very high thermal conductivity and can maintain homogeneously the evaporator surface at nearly constant temperature. This device also has flexible geometry which can fit variable area spaces. Moreover, heat pipe has infinite life and does not require any maintenance. Theses attractive characteristics make heat pipe a promising candidate for HEV/EV battery cooling. Previously, Mahefkey et al. [5] has judged heat pipe to be suitable to mitigate the temperature of NieCd. Similar conclusion has been drawn by Zhang et al. [6] for Ni-MH battery. Concerning Liion battery, Wu et al. [7] have reported that the cell temperature could be significantly reduced using heat pipe with aluminium fin on the condenser section, especially with the help of a cooling fan at the condenser section. More recently, Rao et al. [8] have investigated experimentally the cooling performance of tube heat pipes with
qb/a
heat transfer rate between resin matrix nodes b and a [W] qb/b heat transfer rate between 2 resin matrix nodes b [W] qb/wall heat transfer rate from resin matrix node b to the module wall [W] qwall/wallheat transfer rate between 2 module wall nodes [W] R1 thermal resistance of the experimented system [ C W1] R2 thermal resistance of the full size system [ C W1] Ta temperature of the matrix node a [ C] Tamb ambient temperature [ C] Tb temperature of the matrix node b [ C] Tcell cell temperature [ C] Twall temperature of aluminium wall [ C] Uo open-circuit potential [V] U measured cell potential [V] Greek symbols: radial thermal conductivity of the cell [W m1 K1] thermal conductivity of the resin matrix [W m1 K1] thermal conductivity of the aluminium shell [W m1 K1] F heat generated by cell [W]
lcell lresin lwall
condenser sections cooled by a water module. The battery maximum temperature has been controlled below 50 C when the heat generation rate was lower than 50 W. Coupled with the desired battery temperature gradient, the heat generation rate should not exceed 30 W. In other words, with well-designed heat pipes, the temperature rise and temperature difference of power batteries can be effectively controlled within desired range. However, only tube heat pipe had been experimented and the behaviour of heat pipe under high frequency variable power input had not been studied. In this work, the use of flat heat pipe as an effective and lowenergy cooling solution for a HEV application lithium-ion battery was investigated. Both constant and transient heat input conditions were experimented. The thermal performance of heat pipe cooling was compared with that of a conventional heat sink in order to highlight the thermal performance enhancement of the heat pipe. 2. Experimental set-up 2.1. The battery module description We considered the use of flat heat pipe in order to mitigate the temperature of a 14 cylindrical cells battery module designed for a specific HEV application. Cells (7 Ah of capacity, 38 mm in diameter, 142 mm in height) were implemented in a thermally conductive and electrically insulating resin matrix. The matrix helps keep cells in place and increase the mechanical rigidity of the module. Moreover, the matrix enhances the heat transfer between the cells and the aluminium module walls, as well as the electric insulation between the cells. During charge and discharge, heat generated by cells can be considered to be the sum of the resistive heat and the entropic heat [9e 11]. Consequently, the global heat generated can be determined by:
f ¼ IðU0 UÞ IT
dU0 dT
(1)
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
553
achieved by applying cooling devices to the two larger ones (162 mm in height by 291 mm in large) as shown in Fig. 2. 2.2. Experimental set-up Instead of using directly the battery module as heat source for experiment, a flat heater was used to reproduce the nominal heat flux and an induction heating machine was employed to generate the transient heat flux. Two laboratory size cooling systems were investigated. The first system was a simple copper heat sink. The heat sink base surface measured 39 mm in length by 16 mm in width. In the second cooling system, the previous heat sink was glued with thermal grease on the condenser surface of a flat heat pipe. The structure of the heat pipe is presented in Fig. 3. The copper envelope of the heat pipe was under partial vacuum. The capillary structure was made of sintered copper powder. The effective permeability of the capillary structure was 7 109 m2 and its hydraulic diameter was 40 mm. The heat pipe uses water as working fluid. The peripheral surfaces of both the systems were insulated; hence, heat was dissipated in to the surround air only through the fins. Temperatures at different locations were measured for each system using three K-type thermocouples which have uncertainty of þ/1.5 C (Fig. 4).
Fig. 1. Heat profile generated from cells.
Variable heat rejected by cells in the battery module during standard driving cycle was calculated according to Eq. (1). The nominal value has been found to be slightly over 100 W (Fig. 1). In order to maintain the cells within their optimum temperature range, heat generated needs to be evacuated through the module walls. The two walls corresponding to the two ends of the cells could not be used for cooling purpose. Indeed, one of those was intended for bus bars and electric module installation and the other was used for degassing system implement. Among the 4 remaining walls of the battery module, optimal thermal performance could be
2.2.1. Tests under constant input power We first performed experiments under constant input power condition. After stabilization, the nominal heat flux arrived to the module cooling walls corresponds to the nominal heat generated
Resin matrix
Cell
Cooling systems Upper end of cells: Bus bars and electric module installation
Lower end of cells: Degassing cover & channel
Aluminium shell Fig. 2. Battery module.
6 mm
16 mm
39 mm Copper envelope
Condenser Capillary structure
Evaporator Fig. 3. Schematic of the experimented heat pipe.
Water filling tube
554
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Heat pipe Heat sink
Heat sink
Insulation
Heat input
K-type thermocouples Heat input
Fig. 4. Schematic of heat sink and heat pipe cooling systems.
by cells divided to the cooling walls surface area and equals to 0.115 W/cm2. We wanted to test the cooling systems under the same heat flux as that arrived to the module cooling wall; consequently, the total heat input P of 0.72 W was reproduced using a 23 U flat heater and applied to both the cooling systems. Their thermal performance was evaluated under various cooling conditions and under several inclined positions. Under forced convection conditions, air was blown toward the fins using a flow channel. Fig. 5a shows air blown over the fins for cooling system in vertical position. For cooling system in horizontal position, air flowed through the fins as shown in Fig. 5b. Measurements of the air velocity were made along the fin array using a hot-wire anemometer with measuring range from 0.15 to 30 m/s (Fig. 5a and b). The accuracy of the anemometer was 3% of reading þ0.05 m/s for air flow ranging from 0.15 to 3 m/s and 3% of reading þ0.2 m/s for air flow ranging from 3.1 to 30 m/s. Three air velocities were experimented (0.2 m/s, 0.5 m/s and 1.5 m/s). 2.2.2. Tests under transient input power Finally, the transient behaviour of heat pipe was studied under variable input power produced using an induction heating machine with a reactive power at inductor of 32 kVAR. The output power of the induction generator could be piloted between 10% and 100% of its maximal capacity using a 0-10 V controller. A specific work coil was designed according to the shape of the conductive work surface to be heated. In induction heating, heat absorbed by the work surface depends on the distance between the latter and the coil. According to the power range of this application, the coil was placed at
a distance of 10 mm to the surface to be heated of each cooling system as shown in Fig. 6 (cooling system was placed in horizontal position, insulation and thermocouples are not represented). A thermal infrared camera was used in order to verify that the magnetic field generated by the current in the coil did not affect the thermocouples measurement accuracy. In this regard, work surfaces on which thermocouples were fixed using aluminium tape were coated with 0.95 emissivity black paint. It was found that in our experiment, temperatures indicated by thermocouples corresponded to that given by thermal camera, whether the induction heating system was turned on or shut off. Each test condition was experimented three times and the reproducibility has been found to be good. The temperature deviation between tests under iso conditions was lower than 2.0 C. For all of the tests, the temperatures measured at different locations of the heater/cooling system interface were very close to each other and the margin was less than 1.0 C. Hence, only the mean temperature was reported in this paper. 3. Results and discussion 3.1. AMESim model of the battery module In a configuration such as shown in Fig. 2, cooling devices are not in direct contact with cells but the battery module walls. Consequently, temperature measured at the heater/cooling system interface in our experiment corresponds to the battery module walls temperature and there will be a temperature gradient
a
Flat heater
Air velocity measurements
Fan
b Fig. 5. a e Schematic of the air channel with cooling device in vertical position. b e Flow direction for cooling device in horizontal position (Insulation and thermocouples are not represented).
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
555
Induction heating coil
Fig. 6. Schematic of cooling systems under induction heating coil.
Overall equivalent heat transfer coefficient (hwall)
Adiabatic wall
Adiabatic wall
Nodes b
Nodes a
Overall equivalent heat transfer coefficient (hwall)
Resin matrix (ma, mb, Cpresin, λresin)
Cells (mcell, Cpcell, λcell, Φ)
Aluminum Shell (mwall, Cpwall, λwall)
Fig. 7. Battery module scheme and data used for AMESim modelling.
between module walls and cells. In order to estimate cell temperature at different locations of the module, a numerical model of the module was developed using AMESim software [12]. This model took into account the module geometry, the mass, the thermal characteristics of different components (cell, resin matrix, aluminium shell) and finally the heat generated from cells. It was assumed that the temperature is homogeneous within each cell. A 28 elements mesh was generated for 2D modelling of the resin matrix. Each cell was surrounded by 2 resin matrix nodes: a and b (Fig. 7). The aluminium shell was modelled using 14 elements corresponding to the 14 exterior resin matrix nodes b. The two smaller walls of the module were considered adiabatic. In order to model cooling systems applied to the two larger walls, whether heat sink or heat pipe system was used, an overall equivalent heat transfer coefficient was employed. Hence, the cell energy balance was expressed as below:
dT mcell Cpcell cell ¼ f qcell/a qcell/b dt
(2)
The energy balances for resin matrix nodes a and b are the following:
ma Cpresin
X dTa ¼ qcell/a qa/ai qa/b dt i
ma Cpre sin
X dTb ¼ qcell/b qb/a qb/bi qb/wall dt
(3) (4)
i
Finally, the energy balance for the wall is:
mwall Cpwall
X dTwall ¼ qb/wall qwall/walli dt i
hwall AðTwall Tamb Þ
(5)
The overall equivalent heat transfer coefficient could be determined by fitting the steady state wall temperature after
stabilization of the model with the heater/cooling system interface temperatures measured experimentally. The cell temperature at different locations of the module could also be evaluated using this AMESim model. In order to quantify the thermal performance of the cooling systems used in our experiment, we calculated the thermal resistance according to:
R1 ¼
ðT Tamb Þ P
(6)
The tested cooling systems were laboratory small size prototypes. The thermal resistance of the full size cooling system to be implemented in the battery module could be calculated from the overall equivalent heat transfer coefficient given by the AMESim model:
R2 ¼
1 hwall A
(7)
3.2. Tests in horizontal position In horizontal position, under constant input power condition and natural convection, heat pipe cooling was found to be far more effective than heat sink cooling. After stabilization, the heater/heat sink interface temperature reached 46.0 C whereas that of the heater/heat pipe interface was kept at 38.0 C (Fig. 8). The thermal resistance of both the systems was reported in Table 1. Adding heat pipe to a common heat sink reduced the thermal resistance of the cooling system of about 30%. According to these experimental temperatures, respective overall equivalent heat transfer coefficient was computed for each cooling system thank to the AMESim model aforementioned. Consequently, cell temperature at different locations of the module could be determined. Results were reported in Table 1. The
556
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Fig. 8. Heater/cooling system interface temperatures under natural and forced convection.
Table 1 Natural convection: Temperatures after stabilization for heat sink and heat pipe cooling.
Module wall temperature (measured experimentally) Thermal resistance R1 (calculated from experimental data) Coolest cell temperature (estimated using AMESim model) Hottest cell temperature (estimated using AMESim model) Thermal resistance R2 (estimated using AMESim model) Stabilization time (estimated using AMESim model) Cell temperature goes over 50 C after (estimated using AMESim model)
Heat sink
Heat pipe
46.0 C 36 C/W
38.0 C 25 C/W
55.7 C
48.4 C
57.3 C
49.5 C
0.480 C/W 0.332 C/W 3500 s 800 s
2500 s e
Fig. 9. Heater/heat pipe interface temperatures in horizontal and vertical positions.
when the temperature was lower or equal to 29 C, contrarily to what is intended. The partial vacuum created while producing this heat pipe was certainly higher than what it should have been. According to the temperature measured under the first cooling condition, the operating temperature of the experimented heat pipe was certainly between 29 C and 33 C. Even though the prototype heat pipe did not work at temperature close to that of the ambient, this does not call into question the benefits of using heat pipe under natural convection and under the first cooling condition. Overall, heat pipe can be an effective and low-energy solution for lithium-ion batteries cooling. The suppression/reduction of the parasitic power consumption in battery thermal management helps increase the HEV/EV autonomy and helps reduce it running cost and thus contributes to the wider use of HEV/EV. 3.3. Tests in vertical and inclined positions
temperature gradient within the pack was mainly due to the larger volume of the resin matrix nodes surrounding the cells in the 4 corners. These cells were the coolest and the hottest were those in the centre of the module. We noticed that cell temperatures exceed the upper operating limit which (50 C) with heat sink cooling while all cells were kept below 50 C with heat pipe cooling. Three forced convection configurations were also performed; 20.0 C air was blown at different velocities, from 0.2 m/s to 0.5 m/s and 1.5 m/s. We remarked that under the first cooling condition, the thermal performance of the heat pipe was better than that of the heat sink (20% reduction of the thermal resistance). Unfortunately, the results under the second and the third cooling conditions showed that heat pipe did not perform better than heat sink (Table 2). The flat heat pipe was aimed to be able to work even under ambient temperature (w20 C). However, the above experimental results make us believe that the phase change did not occur
Heat pipe was tested in vertical position and no significant evolution in cooling performance was noticed compared with that in horizontal position (Fig. 9). This result means that the flat heat pipe tested can be used in both vertical and horizontal positions, which leaves a degree of freedom open to best match the battery pack with the space allocated for it in the vehicle. The evaluation of the battery cooling system performance under uphill and downhill drive conditions is necessary for a vehicle application. Common roads are usually graded below 10% (5.71 ). Inclines of over 20% (11.31 ) are rare and maximum known incline on public streets is 37% (20.31 ). Hence, in this part, heat pipe in horizontal and vertical positions were inclined with an angle of 10 and 20 . Fig. 10 shows a very slight evolution of the heater/heat pipe interface temperature between various incline positions. The temperature deviation was found to be less or equal to 2.0 C. This
Table 2 Forced convection: Temperatures after stabilization for heat sink and heat pipe cooling. Heat sink
Module wall temperature Thermal resistance R1 Coolest cell temperature Hottest cell temperature Thermal resistance R2 Stabilization time
Heat pipe
0.2 m/s
0.5 m/s
1.5 m/s
0.2 m/s
0.5 m/s
1.5 m/s
37.0 C 24 C/W 47.5 C 48.4 C 0.313 C/W 3000 s
29.0 C 13 C/W 39.5 C 40.4 C 0.166 C/W 1700 s
24.0 C 5.6 C/W 34.5 C 35.4 C 7.37 102 C/W 1300 s
33.5 C 19 C/W 43.9 C 44.8 C 0.249 C/W 2500 s
29.0 C 13 C/W 39.5 C 40.5 C 0.166 C/W 1700 s
25.0 C 6.9 C/W 35.5 C 36.4 C 9.30 102 C/W 1300 s
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
557
Fig. 10. Heater/heat pipe interface temperatures in various inclined positions under natural convection.
means flat heat pipe works efficiently under different grade road conditions. In Ref. [13], the authors had also found out that thanks to the capillary system, when the working fluid was not overcharged, the orientation of the heat pipe had little effect on its thermal performance. 3.4. Tests under transient condition In real life driving conditions, battery power solicitation profiles are always dynamic. Consequently, heat profiles generated from batteries are of large amplitude and high frequency, unlike those produced by other electronic and electric devices. Fig. 1 is a typical example of how heat generated by cells has highly variable amplitudes and short durations. For battery module with cells inserted in a resin matrix and cooling systems applied to module walls such as shown in Fig. 2, transient heat arrived to cooling devices was computed using the AMESim model presented in previous section. We remarked in Fig. 11 that the heat profile at cooling wall has slight fluctuation and very smooth shape. This is due to large heat capacity of the cells and the resin matrix. Consequently, the influence of the variation of heat flux on the heat pipe behaviour would be very slim. However, in case of larger variation in cell heat loss or lower module heat capacity or direct contact between cooling device and cells, heat arrived to cooling system may vary rapidly and have larger amplitude; hence, the transient behaviour of the heat pipe would be more prominent and investigation of this transient behaviour is necessary. Fig. 11 shows the heat profile at cooling
Fig. 11. Heat profile at module wall and at cell wall.
system/cell wall interface, obtained by AMESim modelling for configuration where cooling device is in direct contact with cells. The corresponding heat flux was reproduced repeatedly using induction heating machine in order to study the transient behaviour of the flat heat pipe. Fig. 12a shows the work surface temperature for both cooling systems. The heat flux applied was also added in this figure. We can see that it consists of six identical cycles of 1460 s duration each. Fig. 12b zooms in on the last cycle of Fig. 12a. We remark that the work surface temperature in heat pipe cooling case was easily lower than that in heat sink cooling case. After stabilization, in heat pipe cooling case, the mean temperature obtained during a cycle of variable heat flux (38.4 C) corresponded to the steady state temperature measured under nominal constant input heat flux in the previous section (38.0 C). On the contrary, in heat sink cooling case, the mean temperature obtained during a cycle of variable heat flux (50.4 C) was about 4 C higher than the temperature measured under nominal constant input heat flux (46.0 C). Furthermore, the temperature fluctuation within cycle was also smaller in heat pipe cooling case. These points make us believe that in transient regime, heat pipe handled more efficiently instant increases of the heat flux than conventional heat sink. Consequently, it can keep the temperature close to that obtained with a nominal heat flux with slight fluctuation (Table 3). 3.5. Further discussions Up to now, the weak point that has limited widespread use of heat pipe system is its high cost due to the complicated fabrication process and the use of copper, an expensive metal, as wick and wall material. However, recent researches on aluminium heat pipe manufacturing [14,15] have revealed efficient and reliable way to decrease the heat pipe cost. Furthermore, the use of aluminium also helps reduce the weight of the cooling system, which is highly appreciable in HEV/EV application. In order to validate the use of heat pipe in automotive application, aside from the impact of the inclination due different grade roads, investigation on the influence of the vehicle shock and vibration to the heat pipe thermal performance is unavoidable. In our study, the flat heat pipe was not tested under vibration environment; however, the influence of vibration on the heat pipe heat transfer has been investigated in some relevant studies. Connors and Zunner [16] have investigated the thermal performance of a flat heat pipe with copper powder wick structure. The flat heat pipe has demonstrated no degradation of the thermal performance under military vehicle shock and vibration conditions. More recently, Guo et al. [17] have investigated the influence of mechanical vibration on the heat transfer characteristics for rectangular microgrooves. It has been found that the vibration movement enlarge the wetting area
558
T.-H. Tran et al. / Applied Thermal Engineering 63 (2014) 551e558
Fig. 12. a) Heater/cooling system interface temperatures under transient heat flux (during 6 cycles, in horizontal position, under natural convection). b) Heater/cooling system interface temperatures under transient heat flux (zoom-in on the last cycle of a).
Table 3 Transient heat flux: Steady state work surface temperature for heat sink and heat pipe cooling. Work surface temperature
Heat sink
Heat pipe
Mean value for 1 cycle Maximum value for 1 cycle Minimum value for 1 cycle Temperature variation for 1 cycle
50.4 C 52.4 C 48.0 C 4.3 C
38.4 C 39.9 C 36.5 C 3.4 C
and intensified consequently the heat transfer of the microgroove. Consequently, the heat pipe seems to be able to work under vibration environment without any thermal performance degradation. 4. Conclusion In order to extend the battery life and decrease HEV/EV cost, battery should be operated within its optimum temperature range with low power parasitic consumption. The use of flat heat pipe as an effective and low-energy cooling device for a HEV application lithium-ion battery was considered under different conditions. According to experimental results, the following conclusions could be drawn: (1) Flat heat pipe can be an effective and low-energy solution for battery cooling. Adding heat pipe reduced the thermal resistance of a common heat sink of 30% under natural convection and 20% under low air velocity cooling. Consequently, the cell temperature was kept below 50 C, which cannot be achieved using heat sink. (2) According to the space allocated for the battery pack in the vehicle, flat heat pipe can be used in vertical or horizontal position. Furthermore, flat heat pipe works efficiently under different grade road conditions. (3) Under high frequency and large amplitude variable heat generated condition, heat pipe handled more efficiently instant increases of the heat flux than conventional heat sink and minimized the temperature fluctuation during transient cycle.
References [1] A.A. Pesaran, Battery thermal models for hybrid vehicle simulations, J. Power Sources 110 (2002) 377e382. [2] A.A. Pesaran, An approach for designing thermal management systems for electric and hybrid vehicle battery packs, in: Fourth Vehicle Thermal Management Systems Conference and Exhibition London, National Renewable Energy Laboratory, 1999. Retrieved from, www.nrel.gov/docs/fy99osti/25992. pdf (accessed 13.11.13). [3] A.A. Pesaran, Battery thermal management in EVs and HEVs: issues and solution, in: Advanced Automotive Battery Conference, National Renewable Energy Laboratory, 2001. Retrieved from, www.nrel.gov/vehiclesandfuels/ energystorage/pdfs/aabc_lv.pdf (accessed 13.11.13). [4] R. Sabbah, R. Kizilel, J.R. Selman, S. Al-Hallaj, Active (air-cooled) vs. passive (phase change material) thermal management of high power lithium-ion packs: limitation of temperature rise and uniformity of temperature distribution, J. Power Sources 182 (2008) 630e638. [5] E.T. Mahefkey, M.M. Kreitman, An intercell planar heat pipe for removal of heat during the cycling of a high rate nickel cadmium battery, J. Electrochem. Soc. 118 (1971) 1382e1386. [6] G. Zhang, Z. Wu, Z. Rao, L. Fu, Experimental investigation on heat pipe cooling effect for power battery, Chem. Ind. Eng. Prog. 28 (2009) 1165e1168 (in Chinese). [7] M.S. Wu, K.H. Liu, Y.Y. Wang, C.C. Wan, Heat dissipation design for lithium-ion batteries, J. Power Sources 109 (2002) 160e166. [8] Z. Rao, S. Wang, M. Wu, Z. Lin, F. Li, Experimental investigation on thermal management of electric vehicle battery with heat pipe, Energ. Convers. Manage. 65 (2013) 92e97. [9] D. Bernardi, E. Pawlikowski, J. Newman, A general energy balance for battery system, J. Electrochem. Soc. 132 (1985) 5e12. [10] L. Rao, J. Newman, Heat generation rate and general energy balance for insertion battery systems, J. Electrochem. Soc. 144 (1997) 2697e2704. [11] T.H. Tran, S. Harmand, B. Desmet, F. Pailhoux, Thermal characterization of LiNi0.8Co0.15Al0.05O2/graphite battery, J. Electrochem. Soc. 160 (2013) 775e780. [12] www.lmsintl.com/LMS-Imagine-Lab-AMESim (LMS International, (accessed 13.11.13). [13] Y. Xuan, Y. Hong, Q. Li, Investigation on transient behaviors of flat plate heat pipes, Exp. Therm. Fluid Sci. 28 (2004) 249e255. [14] Y.T. Chen, S.W. Kang, Y.H. Hung, C.H. Huang, K.C. Chien, Feasibility study of an aluminum vapor chamber with radial grooved and sintered powders wick structures, Appl. Therm. Eng. 51 (2013) 864e870. [15] M. Ameli, B. Agnewa, P.S. Leung, B. Ng, C.J. Sutcliffe, J. Singh, R. McGlen, A novel method for manufacturing sintered aluminium heat pipes (SAHP), Appl. Therm. Eng. 52 (2013) 498e504. [16] M.J. Connors, J.A. Zunner, The use of vapor chambers and heat pipes for cooling military embedded electronic devices, in: Military Communications Conference, IEEE, 2009, pp. 1474e1480. [17] C. Guo, X. Hu, W. Cao, D. Xu, D. Tang, Effect of mechanical vibration on flow and heat transfer characteristics in rectangular microgrooves, Appl. Therm. Eng. 52 (2013) 385e393.