Development and experimental analysis of a hybrid cooling concept for electric vehicle battery packs

Development and experimental analysis of a hybrid cooling concept for electric vehicle battery packs

Journal of Energy Storage 25 (2019) 100906 Contents lists available at ScienceDirect Journal of Energy Storage journal homepage:

2MB Sizes 0 Downloads 78 Views

Journal of Energy Storage 25 (2019) 100906

Contents lists available at ScienceDirect

Journal of Energy Storage journal homepage:

Development and experimental analysis of a hybrid cooling concept for electric vehicle battery packs


Yuyang Wei, Martin Agelin-Chaab

Automotive Center of Excellence, Faculty of Engineering and Applied Science, University of Ontario Institute of Technology, 2000 Simcoe Street North, Oshawa, Ontario, L1H 7K4, Canada


The electric vehicle (EV) technology is one of the most promising pathways to a future of green transportation. One of the keys to EV development is the electric battery. At the moment, Li-ion battery has been the most popular choice in the automotive industry for a variety of advantages. However, battery performance is strongly related to its working temperature, and the health of battery pack in the long-term is greatly affected by temperature uniformity. A novel hybrid cooling concept for battery applications is proposed and experimentally studied in this paper. The concept can utilize any combination of conductive, convective, and evaporative phase change cooling effects. This can be achieved with no extra power from a normal air cooling method using capillary effect as the driving force of the water coolant. But it can attain a higher cooling efficiency and better temperature uniformity. Furthermore, the A/C condensate can be recycled for use as the water coolant, which also adds no extra weight to the vehicle. The air and water coolants can be released to the ambient after usage without harming the environment. The results show that the proposed concept can improve both the cooling efficiency and temperature uniformity by more than 70% compared to the no-cooling baseline. Additionally, it has a 20% improvement in cooling efficiency and a 56% improvement in the uniformity compared to air cooling.

1. Introduction Electric vehicles (EVs) have seen an increasing demand in the past decade. Two main types of EV are mostly found in the market: hybrid (HEV) and full-electric plug-in electric vehicle (PEV). Irrespective of the type the EV, the battery technology is the key part of the development and marketing of the EV. Lithium-ion battery (LIB) is at the moment the most popular choice for EV applications for a number of reasons: high energy density, long cycle life, low self-discharge rate, and high efficiency [1]. However, the performance and health of a battery are strongly related to its working temperature [2–4]. The optimal working temperature for common LIBs is reported to be between 25 °C and 40 °C [2,5]. Additionally, the temperature non-uniformity over the pack, which is calculated as the difference between the highest and the lowest cell temperature is recommended not to exceed 5 °C [6]. However, the actual working temperature and non-uniformity can be much higher in cases of high current draining and thermal runaway. The charge and discharge current load on a rechargeable battery cell is measured as Crate. C-rate is the measurement of current charged into or discharged from an electric power unit, where 1C stands for a current rate reading that is same as the magnitude of the battery nominal capacity. A high Crate in both charge and discharge increases the load on the battery and hence the heat generation. For EVs, the situations such as accelerating and high load on air-conditioning can make the battery load much

higher than 1C. This increases the thermal instability and hence battery degradation or even safety issues. Therefore, battery thermal management is critical for EV applications. Three cooling methods are widely studied when it comes to thermal management of EV batteries: (1) air-cooling, (2) liquid cooling, and (3) phase-change material (PCM). Air-cooling is the most traditional cooling method in industries. It is popular due to the fact that it has a simple and light-weight structure, low running power, low development, and maintenance cost, and is environmentally friendly. However, it has low thermal conductivity and heat capacity and this limits its cooling efficiency. In addition, the air-cooling efficiency is highly dependent on the ambient temperature. Furthermore, the prediction of the airflow behaviors within a battery pack can be difficult. For example, a pack of cylindrical batteries cooled by unidirectional airflow was reported to have a higher cooling efficiency using in-line layout than the staggered layout [1]. At the same time some studies made opposite conclusions [7], and so the design criteria for the air-cooling systems for different battery sizes and pack dimensions may not be consistent. Other designs are similar to each other by having converging inlet and diverging outlets located at the top and bottom of the battery pack to provide uniform air pressure over the energy unit [8–11]. Even though these duct designs had improved the cooling efficiency and the uniformity, they are geometrically complex. Complex external and internal duct designs can significantly restrict the energy density because

Corresponding author. E-mail address: [email protected] (M. Agelin-Chaab). Received 14 June 2019; Received in revised form 14 August 2019; Accepted 17 August 2019 Available online 30 August 2019 2352-152X/ © 2019 Elsevier Ltd. All rights reserved.

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

Nomenclature A ai C Cp

surface area (m2) specific area (m2) correction coefficient constant kJ heat capacity ( kg·K )

D Ei F g

diffusion coefficient ( s ) equlibrium potential (V) Faraday constant (9.648456e + 04 C/mol) m gravity acceleration ( s2 )


R electric resisitance (ohm) m sorptivity ( s0.5 ) S T temperature (K) t time (s) V battery working voltage (V) Vo open circuit voltage (V) kg water evaporation mass transfer coefficient ( m2Pa·s or β ηi overpotential between adjacent anode and catode layers (V) m2 mean kinematic viscosity ( s ) μ



specific enthalpy of air ( g )


effective convective heat transfer coefficient ( 2 ) hc mK hvs J specific enthalpy of water vapor at water surface temperature ( g ) I infiltration (m) m infiltration rate ( s ) i Ii current (A) ii current generationby a specific area of a layer of battery active material (A) mol specific reaction flux ( m2s ) ji keff kT L m p Q

density (


electric conductivity ( m )

φ ϕs ϕe



) S


moisture mass fraction ( kg ) electric potential (V) electrolyte potential (V)

Subscripts a b cc eff ma SEI s w


electrochemical reaction rate constant ( s )

effecitve thermal conductivity characteristic length (m) mass (kg) pressure (Pa) heat generation (W)



kg m2s

W ( mK )

ambient air battery connector contact effective moisturized air solid electrolyte interface saturated steam water

thermal runaway [18]. However, PCM can also be heavy, expensive, and not environmentally friendly. Also, the heat taken out by PCM may also need to be extracted by other cooling methods because the total heat capacity of a given mass of PCM is limited [19,20]. Another PCMlike cooling approach is the heat pipe, which contains complex inner structures and special chemicals with low boiling points, making it too expensive for automotive applications. For all these mentioned studies reviewed from the literature, it can be concluded that (1) the air-cooling method is usually simple in structure, light-weight, energy-saving, but has a poor cooling performance. Improvement in cooling performance could sacrifice the simplicity and the battery pack energy density; (2) liquid cooling method performs better than air, but it needs a circulation systems which adds an extra layer of complexity, weight, and energy consumption; (3) the PCM method can provide the best cooling performance both in terms of the temperature rise and uniformity, however, because the thermal capacity is dependent on mass, PCM systems are usually heavy or require additional cooling systems to extract the heat from the PCM pack; and (4) coolants chosen for the liquid cooling and the PCM systems may not be environmentally friendly. The above conclusions motivated a concept of hybrid EV battery cooling that combines at least two cooling methods directly working on battery cell surface. The goal is to improve system simplicity, weight reduction, energy saving, and environmental friendliness while improving the cooling efficiency and uniformity.

of the large cell-to-cell gaps. Cylindrical battery packs under unidirectional air cooling suffer higher temperature non-uniformity because aircooling is not effective on the leeward side of the cells, and the heat can accumulate at the pack outlet. Therefore, some investigations were made on active air-cooling systems with reciprocating flows to improve uniformity and showed that the uniformity is positively correlated to the frequency of flow reciprocation [11–14,33]. However, the highfrequency reciprocating flow systems are expensive to implement and can decrease system reliability dramatically [7]. It should also be noted that reciprocating flow concepts are mainly for improving uniformity but not cooling efficiency because the air thermal properties do not change. In conclusion, the air-cooling method can work properly under low battery discharge rates, but it is not effective enough for intensive working conditions such as racing and highly loaded air conditioning in hot summer days [8–11]. Liquids are generally more effective for cooling because of their much higher thermal conductivity and heat capacity. However, normal liquid-cooling systems are complex in structure and design, heavy, costly and have higher running power. This is because of requirements of a recirculating system including pumping, piping and flowrate control. Some existing liquid cooling methods employ glycol or water as the coolant, however, different coolant materials can be employed. When it comes to liquid cooling, pouch or prismatic cells have been studied more extensively than cylindrical cells since they are simpler and more compact in the design of cooling plates [15,16]. The coolant flow rate for normal EV battery pack can be as high as 10 L/min and hence energy consumption by the pump can be high, resulting in a dramatic drop in EV driving range [17]. On the other hand, PCM is a relatively new approach for thermal management of EV battery. Essentially, PCM cooling is a purely passive system and require no extra energy to work. With the ability to trap latent heat while melting, PCMs theoretically have the highest cooling efficiency. Additionally, PCM cooling comes with the best uniformity since the heat source (the cells) can be fully buried by the coolant (PCM). In fact, PCM is superior in cooling performance and can avoid

2. Description of the proposed hybrid cooling concept The design of the proposed system is based on a simple air-cooling duct in which there are a series of hydrophilic fiber channels. A water coolant is driven through the channels purely by capillary forces, which can be exposed to the airflow to extract the latent heat by water evaporation from the battery. Therefore, the system can utilize enhanced water vaporization by air convection to achieve an effective cooling. Fig. 1 demonstrates the concept of a single cell. 2

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

Fig. 2. Schematic of a conceptual system configuration of the proposed hybrid cooling system.

reversible heat by the chemical reactions, connector resistance, and internal resistances [1,9,21]. The equations for these heat sources are complex as the internal resistances include electronic conductive resistance (resistance of active materials in positive and negative electrodes and current collector), ion conductive resistance (electrolyte resistance), and the resistance from the solid electrolyte interface (SEI) film on the negative electrode [1]. Therefore, the general thermal equation of heat conduction within a lumped battery model can be expressed as [1,9,21]:

Fig. 1. Illustration of the proposed hybrid cooling concept.

The proposed design is applicable to any type of battery (cylindrical, prismatic, or pouch type in different sizes or dimensions) or electric energy device that requires cooling. A battery mount forms a duct with an inlet and an outlet, allowing convective coolant to flow through. The hydrophilic or super-hydrophilic fiber material is also positioned and retained by this mount. The fiber can be either contacted or contactless to the battery surface, served as fiber channel for the conductive and evaporative water coolant. For the case with the fiber channel in contact with the battery surface, a super-hydrophilic fiber material with high thermal conductivity is preferred. The energy unit mounts in this study has the air flow duct placed horizontally, so the air will flow perpendicularly to the center line of the energy unit. In a variation of the design, multiple air ducts can be placed vertically, so that the air can flow parallel to the battery center line. The reservoir for the water coolant continuously provides water coolant to the fiber channels. The reservoir is located under the energy packs so that the water is driven by capillary effect only. This also naturally constrain the water from leaking out from the fiber because of the surface tension between water and the fiber. Additionally, after the fiber strip is saturated wet, the water flow rate will be purely controlled by the rate of water vaporization, eliminating the requirement of the flow rate control system. Fig. 2 shows the schematic overview of one of the possible system configuration. Three cooling stages can be designed, for example, a pure air-cooling mode can be assigned as first-stage cooling intended for the small-load driving mode, a second-stage water-cooling is suitable for a further temperature control in a medium-load driving mode, and a final-stage hybrid cooling mode is able to cool down the batteries in extreme-load driving mode. Condensation from the air-conditioner (A/C) can be recycled as the water coolant as well. The low-temperature condensate from the A/C can enhance the cooling efficiency. A low-power pump might be required only for transporting the A/C condensate to the reservoir, while the working fluid (or water) in the cooling system is driven purely by capillary effect, with no extra power required from the batteries. The water coolant may be specifically applied to locations with most heat accumulation for optimal pack-level temperature uniformity. The air and the water coolant can be released to the ambient after use without harming the environment.


T = t

·kT T + Qp + Qrea + Qc + Qohm


where kT is the effective thermal conductivity of the active battery materials, Qp, Qrea, Qc, Qohm are heat generation from polarization, reversible heat by chemical reactions, electric contacts, and internal resistance, respectively:

Qp = ai ji F



Qrea = ai ji FT

dEi dT


(I ) . Q = c

Qohm =

2 Rcc i Acc



eff , i


s, i


+ keff , i


2keff , i RT (1

t+ )


e, i


x 1+

dlnf± dlnci

lnci x

e, i


+ ai ji F


(5) ai is the specific surface area, ji is the specific reaction flux, F is the Faraday constant (9.648456e + 04 C/mol) . The product of these three parameters is the current ii produced by the specific area ai: (6)

ii = ai ji F

However, a simplified and commonly used equation for the heat generation was given as [22,23]:

Qtotal = Ii (Vo


Ii T

dVo dT


where Q is the heat generation in the battery cell, Vo is the open circuit voltage, V is the cell working voltage, I is the applied current and T is the temperature of the cell. The first term on the right-hand-side of the equation is essentially an integral form of Eq. (2), but it stands for the irreversible heat generation by the overpotential of the whole circuit, which concludes all the heat generation by polarization and internal resistance. The second term is actually the integration of Eq. (3) and stands for the reversible entropic heat generation by the total chemical

3. Working principles 3.1. Battery heat generation The heat generation for Li-ion batteries comes from polarization, 3

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

reaction within a cell [22,23].

Table 1 Battery specifications [28].

3.2. Water evaporation The mass flow rate of water evaporation for a unit area (kg s can be calculated as:

m wev = (ps

pa )


m 2) (8)

where β is the convective mass transfer coefficient, and it is related to the airflow velocity; ps and pa are the saturated steam partial pressure at water temperature and the vapor partial pressure at the ambient air temperature. However, it was argued that the water evaporation rate is nearly impossible to be predicted due to its complex and stochastic nature [24]. Eq. (8) can also be written in the form of the humidity ratio φ as [24,25]:

m wev = (



L 1

D H4

= C. (Re )a . (Ar )b . (Sc )d .


Supplier Chemistry Diameter Height Weight Nominal capacity Nominal voltage Charge cut-off voltage (overvoltage limit) Discharge cut-off voltage (undervoltage limit) Maximum continuous discharge current Maximum pulse discharge current Calculated internal resistance

Efest LiMn2O4 (IMR/LMO) 26.5 ± 0.2 mm 65.98 ± 0.2 mm 92 ± 1 g 5200 mAh 3.7 V 4.1 ± 0.1 V 2.5 ± 0.1 V 15 A 40 A 0.15 hm

dT = Qg dt



Qmar = hc Ab (Tb

Tma )


mb c pb


where the transfer coefficient β is hard to predict as stated above, but the equation for the coefficient is given by the Sherwood criterion (Sh) [24]:

Sh =


where (mb c pb dt ) denotes the energy change with respect to time of a battery cell, Qmar is the heat removed by moisturized air convection. T airflow will be cooled before hitting the battery surface by the wetted fiber channels. The heat taken out from the air by the water is provided by the equation in El-Ladan's study [26]: dT


where Re, Ar, and Sc are the Reynolds, Archimedes, and Schmidt numbers, respectively.

ma dha = [hc (Ta

Tw ) + (


a ) h vs ] dA


where ma is the mass of airflow, A is the surface area of the fiber channels exposing to the airflow. From Eq. (9), it can be concluded that lower relative humidity can contribute to a higher water vaporization rate, which can trap larger amount of latent heat. Additionally, it is pointed out that a lower humidity decreases water droplet temperature [32], and hence further increase the convective cooling effect according to Eq. (13).

3.3. Energy governing equations In the proposed hybrid cooling method, a combination of convective cooling by air, and latent heat trapped by water vaporization is directly applied to the cells. As the fiber channel makes no contact with the battery surface, energy governing equations similar to El-Ladan's model [26] can be used:

Fig. 3. Experimental setup for temperature rise and uniformity measurements of the proposed thermal management concept. (a) A schematic and (b) a picture. 4

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

mcap = mev

Table 2 Thermal camera specifications [29]. Properties


Detector type Thermal resolution Spectral range Field of view Thermal sensitivity Temperature range Accuracy Measurement mode

FPA, uncooled ASi Microbolometer 384 × 288 / 640 × 512 pixels and 1024 × 768 XGA HD 8–14 µm 17.5 × 13° ≤ 80 mk @ f/1.60 Hz, 300 K −20 to 100 °C ± 2% of reading Spot (Sp), area (Ar), isotherm, profile, auto hot spot, auto alarm Variable from 0.01 to 0.99 in 0.01 increments

Emissivity correction Measurement features

4. Experimental setup and procedure 4.1. Experimental setup Fig. 3 shows the schematics and the pictures of the experimental setups of the proposed concept. A row of eight Li-ion batteries (Efest IMR 26650 5200 mAh) with the specifications shown in Table 1 was connected in series inside an air duct made from acrylic and a thin transparent plastic film. This number of cells was used due to the limitations of the power capacity of the power supply. In this set up the fiber channels were made out of paper towels in this case after preliminary tests showed that it was the cheapest and most effective option for the lab testing. A power supply (Extech Instrument 382275 30 V 20A DC power supply) connected with a Li-ion battery management circuit board and a multifunctional high-power load (TDI RBL488) were used for the charge and discharge, respectively. The battery, charger, and the discharger were connected in parallel as the discharger could also be used as a multimeter to monitor different working parameters of the battery in this case. The thermal characteristics were measured by an infrared thermal camera (SPI IRXP-5000, specifications shown in Table 2) which was calibrated with thermocouples placed at different battery locations in our previous study [34]. The images from the thermal camera were recorded by a screen recorder with a built-in timer. For optical access of the infrared camera, one sidewall of the air duct was replaced by a plastic film. A small model wind tunnel was custom made from an axial fan with an outlet size that matches the battery pack inlet. A multifunction anemometer (Proster TL107) with temperature range from −10 to 40 °C was used to measure the air velocity and ambient temperatures. The accuracy of the anemometer reading was within ± 0.1 °C. A weight scale with 0.001 g accuracy was used to estimate the infiltration rate of the water coolant and water vaporization rate.

All correction based on distance, relative humidity, atmospheric transmission & external optics

The water coolant was purely driven by the capillary effect. The flow rate of the liquid within a fiber material can be simply expressed by taking derivatives of the cumulative infiltration I (i.e. distance the liquid travel through the fiber material) over a short time t [27]: (14)


dI 0.5S =i= . dt t


where mcap , mev , and mpr are the mass flow rate by capillary pressure, evaporation, and hydrostatic pressure, and the hydrostatic pressure can be zero if the fiber locates over the liquid free surface.

3.4. Capillary action within fiber channels

I=S t

mpr .


where S is the so-called sorptivity. The equations above are good for horizontal infiltration where the capillary effect is the only force acting on the liquid. For vertical or angled infiltration where gravity is significant, a parameter A1 can be added to Eq. (14) [27]. The estimation of capillary flowrate within fibers can be more complex if evaporation of liquid is accounted for in the calculation. From the equations above, we can see that the infiltration rate by the capillary effect decreases non-linearly over time. However, it is also known that the capillary infiltration rate can decrease to 0, with the fiber moisture saturation increasing to 1. Therefore, if the liquid vaporization cannot go over the infiltration rate at one region of a fiber channel with the highest capillary pressure, the coolant compensation rate can be determined by the vaporization rate instead of the capillary effect [25]:

4.2. Experimental procedure Tests were designed to compare the cooling performance between air cooling only and the hybrid cooling. A test under no cooling effect was made as a baseline. The pack was designed based on Yang's study

Fig. 4. Schematic and dimensions of the tested battery pack. (a) Side view and (b) top view.


Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

Fig. 5. Temperature contours at the end of 1.15C discharge and plots of average surface temperature for each cell versus time for the no-cooling baseline.

[1] with critical dimensions shown in Fig. 4. Eight batteries were connected in series with the inline layout in the same orientation of airflow. The fiber strip was made by twisting the paper towel material to an average diameter of 2 mm. For the hybrid cooling test, there were 35 fiber strips evenly distributed along the pack's length by placing 5 strips between each cell in the pattern shown in Fig. 4. All the experiments were conducted at a room temperature of 21 °C. The axial fan generated a consistent airflow velocity of 0.7 m/s. For the hybrid cooling test, the tap water with temperature measured as 21 °C was driven into the twisted fiber channel by capillary effect. The water coolant flowrate for the hybrid cooling test was purely determined by the capillary effect which will be discussed later. The battery pack was balanced charged to 32 V with constant current and constant voltage (CC/CV) charging mode starting with a 1.0 A (<0.2C) current. The charging processes were cut off when the charging current dropped below 0.1 A (≅ 0.02C). By maintaining a low charging current, the battery temperatures were stable at room temperature, hence no degradation occurred during the charging period. The discharge cut off voltage (under-voltage) limit was set to 21.7 V as recommended by the battery specification. The batteries were then discharged with a constant current (CC) of 6.0 A (≅ 1.15C). The discharging rate was set so that the discharging period is long enough for effective observation, meanwhile the surface temperatures show obvious diversity to cover equipment tolerances, errors, and noise readings. The infrared camera was set to start recording as soon as the circuit was closed. Additionally, a timer built-in screen recorder software was set to start timing simultaneously. An aerial infrared sensing region (Ar, as shown by the yellow regions in Fig. 4) was mapped on each battery, in the pattern shown in Fig. 4. The areal sensing method can automatically recognize the location of the highest and the lowest surface temperature and makes the

most accurate measurement. Real-time maximum/minimum, and average temperatures are automatically presented in the infrared camera recording software by the following equation:

Tavg =

n (Ti Ai ) i=1 n A i=1 i


For the water infiltration and vaporization rate measurements, the weight of 10 dry fiber strips was measured by the weight scale first, then the tips of these strips were dipped into the water and the camera recording the infiltration process until the whole strip length is wet. The infiltration length was measured by scaling the image from the camera record. The weight of these 10 wet fiber strips was measured after the infiltration ends. The water mass flow rate during infiltration is simply calculated as the difference between the dry and wet weight divided by the infiltration time. A similar measurement method was used in the water vaporization rate measurements. The water vaporization rate was measured by dividing the weight difference, between the saturated wetted and the post-vaporized fibers, by the vaporization time. The uncertainty in this study mainly came from the equipment (thermal camera) and was calculated based on Moffat's theory [35]. The total uncertainty was estimated to be about ± 5.9%. 5. Result and discussion 5.1. Working temperature 5.1.1. No-cooling and air-cooling Fig. 5 shows the temperature contours at the end of discharge and plots of average surface temperature for each cell versus time for the no-cooling conditions. One can obviously observe that the cell in the 6

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

Fig. 6. Temperature contours at the end of discharge and plots of average surface temperature for each cell versus time for the air cooling method.

middle of the pack has the highest surface temperature. The temperature diminishes from the mid-pack to both the inlet and outlet locations. This makes sense since the cells at both ends have no heat accumulation from an adjacent cell, while the cell in the middle suffers the most heat transferred from both sides of the battery pack. The plot shows results that are consistent with this conclusion. During the entire discharging period, Ar4 located in the middle of the pack shows the most aggressive temperature rise, and it records a final temperature of 60 °C. Ar3, Ar5, and Ar6 have rising temperature trends that are almost identical. The final reading for these three cells is about 57 °C. The cells Ar2 and Ar7 at the locations indicated share a similar temperature trend, and the final temperature is recorded as 55 °C and 52 °C, respectively. The cells Ar1 and Ar8 at the locations indicated have the lowest rate of temperature increase because they are furthest to the central location and has the most heat cumulated. The final temperatures at these locations are identical for both cells and are recorded as 47 °C. Therefore, all the cells in this no-cooling test under 1.15C discharging rate have final temperatures that are over the limit of 40 °C. In fact, the temperatures start exceeding this limit after only 14 min of discharge. Fig. 6 shows the temperature contours at the end of discharge and plots of the average surface temperature for each cell versus time for the air-cooling conditions. Compared to Fig. 5, the heat accumulation region (yellow) has shifted towards the pack outlet because the heat can only be transferred from the inlet to outlet under unidirectional airflow. For each cell location except for Ar1, the hot spots, indicated by the red triangle indexes, occurs at the leeward side due to the wake there. This conclusion is consistent with previous studies [1]. The cell with the highest temperature for this cooling case is Ar2 instead of Ar4 for the no-cooling case. The rest of the cells have nearly identical temperature trends and final temperature readings with the final temperatures around 38–°C. For the first 5 min of discharge, cell Ar1 had a rate of

temperature rise similar to Ar8 that is located at the pack inlet, because there is no heat built up in the pack. The temperature rose sharply after 5 min and reached to a final reading of 38 °C, which is close to the value at Ar6. Cells Ar7 and Ar8 located around the pack inlet have the working temperature lower than the rest, and the final temperature for Ar8 is recorded as 30 °C. Air cooling successfully lowers the working temperature for each cell under the limit of 40 °C. However, the temperatures for 6 out of the 8 cells are very close to the limit of 40 °C. This strongly suggests that the temperature will exceed this limit even at a slightly higher discharge C-rate. 5.1.2. Hybrid cooling Water flow rate. Fig. 7 illustrates the average time needed for the water coolant to infiltrate the height of the battery pack (55 mm) within a fiber strip described earlier. The plot is consistent with the mathematical property of Eq. (14), which predicts that the infiltration rate decreases non-linearly over time. It shows that water within a paper fiber strip can flow through the entire battery pack height within 70 s. This means by using a more hydrophilic fiber material, the hybrid cooling system can reach fully working condition within a minute after deployment. The mean mass of water infiltrated was recorded as 0.061 g/s. Since there are 35 fiber strips installed on the battery pack, the total mass flow rate during the infiltration process was estimated as 2.135 g/s. The water vaporization rate in a fiber strip under room temperature (21 °C) and 0.7 m/s wind speed was measured on average as 1.70e 04 g/s (1.02e 05 L/min ), which was much lower than those presented in other studies [3,30,31]. Hence, it is reasonable to assume that the capillary pressure is high enough to compensate for the water lost by vaporization. Therefore, the fiber channel strip design is suitable for even higher vaporization rates. Additionally, since the latent heat 7

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

Fig. 7. Water coolant infiltration over time for a single fiber strip used in the hybrid cooling.

trapped by water vaporization is 2260 J/g, the 35 fiber strips can transfer at least 13.5 W of heat by the vaporization.

of the cells, because of the same reason previously discussed. After 20 min, the heat built up in the pack is transferred to the outlet, and hence Ar2 shows a sharp temperature rise until at the end of discharge. Ar1, Ar3, Ar5, Ar6, and Ar7 have similar temperature trends. The final temperatures for all of them were recorded as 0 °C. It can be observed that Ar1 at the outlet position comes with the lowest temperature for the first 15 min, but its final temperature ranks the third highest among all the cells. This is consistent with the temperature trend of Ar2. Ar8 which is located at the inlet without heat transfer from other cells presents the lowest rate of temperature rise and the lowest final temperature (28 °C). The hybrid cooling method successfully retains Hybrid cooling results. Fig. 8 shows the temperature contours at the end of discharge and plots of average surface temperature for each cell versus time for the hybrid cooling conditions. From the contour, it is hard to tell which cell has the highest working temperature, and the color distributes more evenly compared to the two previous tests. The indexes for the hot spots are more random. From the plot, both Ar2 and Ar4 have the highest final temperature of 32 °C. The rate of Ar2 temperature rise for the first 20 min of discharge is same as the majority

Fig. 8. Temperature contours at the end of discharge and plots of average surface temperature for each cell versus time for the hybrid cooing method. 8

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

Fig. 9. Average pack temperature comparison among the three cooling types: no cooling, air cooling, and hybrid cooling.

the working temperature for all the cells under the 40 °C limit, and the highest temperature is still 8 °C lower than this limit, making more room for further discharging and higher discharge C-rate. Fig. 9 shows the pack level average temperature during discharges under the three conditions shown. The temperature in the no cooling test reaches 54 °C which is 14 °C over the limit. By simply using the air cooling method, the final temperature was lower by 50% to 37 °C. The proposed hybrid cooling method can further lower the final temperature to only 30 °C. This is a 71% improvement compared to the nocooling baseline.

out in this comparison. The non-uniformity in the hybrid cooling test retains under 4 °C. The final reading is shown as only 3.4 °C. Therefore, the proposed hybrid cooling method improves the temperature uniformity by 72.4% compared to the baseline. 6. Conclusion A new hybrid cooling concept for battery applications is proposed and experimentally tested in this study. The concept utilizes any combination of conductive, convective, and evaporative phase change cooling effects. The results show a higher cooling efficiency and much better temperature uniformity over the battery pack. The concept was able to maintain both the maximum working temperature and nonuniformity much lower than the recommended limits. The proposed concept provides more than 70% improvement in both the cooling efficiency and the temperature uniformity compared to the no-cooling baseline, and it is a 20% improvement in the cooling efficiency and a 56% improvement in the temperature uniformity compared to the air cooling. It must be noted that these results were achieved under the above model experiments. Therefore, there is a great potential for improvement for the proposed hybrid cooling concept. The innovative nature of the concept is that it achieves these

5.2. Temperature uniformity Fig. 10 shows the trends of pack-level temperature non-uniformity for the three cooling conditions. It was mentioned previously that the maximum pack-level non-uniformity should not exceed 5 °C [6]. The non-uniformity of no-cooling condition exceeds this limit after only 9 min of discharge, and it ends up with 12.3 °C. Even though the air cooling can lower the final working temperature by 50%, it does not work well from the temperature uniformity perspective. The final nonuniformity is shown as 10 °C, which is only a 16.7% improvement which is double the 5 °C limit. In fact, the hybrid cooling method stands

Fig. 10. Temperature non-uniformity comparison among the three cooling types: no cooling, air cooling, and hybrid cooling. 9

Journal of Energy Storage 25 (2019) 100906

Y. Wei and M. Agelin-Chaab

significant improvements with nearly no extra power from a normal air cooling method by using capillary effect as the driving force and vaporization as the flow rate control for the water coolant. The concept may recycle A/C condensate so the water coolant can be instantly consumed as it is generated and hence adds negligible weight to the vehicle. The air and water coolant after use can be released to the ambient environment without harming the environment since no chemical is used. Although due to some limitations of experimental setup, this packlevel study was not able to compare directly the cooling performance with a liquid cooling system, which is more conventional and has existing commercial application. However, the cell-level performance which was studied and published earlier [34], showed positive results for the new hybrid concept. Future studies will focus on pack-level experimental comparison between the liquid-cooling and the proposed concept. Studies involving analysis of the battery cooling performance under various charging/discharging rates and ambient factors including temperature, moisture level, and air flow velocity are in progress.

[13] A. Sasmito, E. Birgersson, A. Mujumdar, A novel flow reversal concept for improved thermal management in polymer electrolyte fuel cell stacks, Int. J. Therm. Sci. 54 (2012) 242–252. [14] R. Mahamud, C. Park, Reciprocating air flow for Li-ion battery thermal management to improve temperature uniformity, J. Power Sources 196 (13) (2011) 5685–5696. [15] S. Panchal, R. Khasow, I. Dincer, M. Agelin-Chaab, R. Fraser, M. Fowler, Thermal design and simulation of mini-channel cold plate for water-cooled large sized prismatic lithium-ion battery, Appl. Therm. Eng. 122 (2017) 80–90. [16] S. Panchal, S. Mathewson, R. Fraser, R. Culham, M. Fowler, Thermal management of lithium-ion pouch cell with indirect liquid cooling using dual cold plates approach, SAE Int. J. Altern. Powertrains 4 (2015) 293–307. [17] F. He, A. Akram AMS, Y. Roosien, W. Tao, B. Geist, K. Singh, Reduced-Order Thermal Modeling of Liquid-Cooled Lithium-Ion Battery Pack for EVs and HEVs, Chrysler Tech Center, 2017 1000 Chrysler Dr, Auburn Hills, MI, USA. [18] R. Kizilel, R. Sabbah, J.R. Selman, S. Al-Hallaj, An alternative cooling system to enhance the safety of Li-ion battery packs, J. Power Sources 194 (2) (2009) 1105–1112. [19] Z. Ling, F. Wang, X. Fang, X. Gao, Z. Zhang, A hybrid thermal management system for lithium-ion batteries combining phase change materials with forced-air cooling, Appl. Energy 148 (2015) 403–409. [20] H. Fathabadi, High thermal performance lithium-ion battery pack including hybrid active-passive thermal management system for using in hybrid/electric vehicles, Energy 70 (2014) 529–538. [21] K. Smith, C. Wang, Power and thermal characterization of a lithium-ion battery pack for hybrid-electric vehicles, J. Power Sources 160 (1) (2006) 662–673. [22] Y. Abdul-Quadir, T. Laurila, J. Karppinen, K. Jalkanen, K. Vuorilehto, L. Skogström, M. Paulasto-Kröckel, Heat generation in high power prismatic Li-ion battery cell with LiMnNiCoO2 cathode material, Int. J. Energy Res. 38 (11) (2014) 1424–1437. [23] G. Liu, M. Ouyang, L. Lu, J. Li, X. Han, Analysis of the heat generation of lithium-ion battery during charging and discharging considering different influencing factors, J. Therm. Anal. Calorim. 116 (2) (2014) 1001–1010. [24] R. Turza, B. Füri, Experimental measurements of the water evaporation rate of a physical model, Slovak J. Civ. Eng. 25 (1) (2017) 19–24. [25] G. Leroux, M. Mendes, L. Stephan, N., .L.. Pierrès, E. Wurtz, An innovative cooling system based on evaporation from a porous tank, 14th Conference of International Building Performance Simulation Association, Hyderabad, India, December 7–9, 2015, 2015, pp. 2453–2460. [26] A.D. El-Ladan, O.C.L. Haas, Fan-pad evaporative battery cooling for hybrid electric vehicle thermal management, IET International Conference on Resilience of Transmission and Distribution Networks (RTDN) 2015, Birmingham, 2015, pp. 1–7. [27] J.R. Philip, The theory of infiltration: 4. Sorptivity and algebraic infiltration equations, Soil Sci. 84 (1957) 257–264. [28], Efest IMR 26650 5000 mAh 45A Flat Top Battery - Batteries – EFEST, (2017) [online] Available at: article&at=read&did=448. [29] SPI Corp, XP 5000 Radiometric Infrared Camera, (2017) [online] Available at: [30] C. Lan, J. Xu, Y. Qiao, Y. Ma, Thermal management for high power lithium-ion battery by mini-channel aluminum tubes, Appl. Therm. Eng. 101 (2016) 284–292. [31] J. Zhao, Z. Rao, Y. Li, Thermal performance of mini-channel liquid cooled cylinder based battery thermal management for cylindrical lithium-ion power battery, Energy Convers. Manage. 103 (2015) 157–165. [32] A. Fujita, R. Kurose, S. Komori, Experimental study on effect of relative humidity on heat transfer of an evaporating water droplet in air flow, Int. J. Multiphase Flow 36 (3) (2010) 244–247. [33], U. S. Environmental Protection Agency and U.S. Department of Energy, 2015-01-06. Retrieved 2015-11-06. [34] Y. Wei, M. Agelin-Chaab, Experimental investigation of a novel hybrid cooling method for lithium-ion batteries, Appl. Therm. Eng. 136 (2018) 375–387. [35] R. Moffat, Describing the uncertainties in experimental results, Exp. Therm. Fluid Sci. 1 (1) (1988) 3–17.

Acknowledgment Financial support of this work by the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged. References [1] N. Yang, X. Zhang, G. Li, D. Hua, Assessment of the forced air-cooling performance for cylindrical lithium-ion battery packs: a comparative analysis between aligned and staggered cell arrangements, Appl. Therm. Eng. 80 (2015) 55–65. [2] Z. Rao, S. Wang, A review of power battery thermal energy management, Renew. Sustain. Energy Rev. 15 (9) (2011) 4554–4571. [3] J. Xu, C. Lan, Y. Qiao, Y. Ma, Prevent thermal runaway of lithium-ion batteries with mini-channel cooling, Appl. Therm. Eng. 110 (2017) 883–890. [4] W. Gu, C. Wang, Thermal-electrochemical modeling of battery systems, J. Electrochem. Soc. 147 (8) (2000) 2910. [5] J. Jaguemont, L. Boulon, Y. Dubé, A comprehensive review of lithium-ion batteries used in hybrid and electric vehicles at cold temperatures, Appl. Energy 164 (2016) 99–114. [6] S. Shahid, M. Agelin-Chaab, Analysis of cooling effectiveness and temperature uniformity in a battery pack for cylindrical batteries, Energies 10 (8) (2017) 1157. [7] W. Tong, K. Somasundaram, E. Birgersson, A.S. Mujumdar, C. Yap, Thermo-electrochemical model for forced convection air cooling of a lithium-ion battery module, Appl. Therm. Eng. 99 (2016) 672–682. [8] H. Park, A design of air flow configuration for cooling lithium-ion battery in hybrid electric vehicles, J. Power Sources 239 (2013) 30–36. [9] L.H. Saw, Y. Ye, A.A.O. Tay, W.T. Chong, S.H. Kuan, M.C. Yew, Computational fluid dynamic and thermal analysis of lithium-ion battery pack with air cooling, Appl. Energy 177 (2016) 783–792. [10] H. Sun, R. Dixon, Development of cooling strategy for an air-cooled lithium-ion battery pack, J. Power Sources 272 (2014) 404–414. [11] Z. Liu, Y. Wang, J. Zhang, Shortcut computation for the thermal management of a large air-cooled battery pack, Appl. Therm. Eng. 66 (1–2) (2014) 445–452. [12] L. Fan, J.M. Khodadadi, A.A. Pesaran, A parametric study on thermal management of an air-cooled lithium-ion battery module for plug-in hybrid electric vehicles, J. Power Sources 238 (2013) 301–312.