Temperature Gradient Reduction in High-Power

Battery Systems using Prismatic Cells combined with

Phase-Change Sheets and Graphite Foils
M. Gepp, H. Reisenweber, V.R.H. Lorentz, M. März
Fraunhofer Institute for Integrated Systems and Device Technology IISB
91058 Erlangen, Germany
[email protected]

Abstract — In lithium-ion battery systems thermal manage- matic cell and PCM were performed. A reduction of the maxi-
ment has an important influence on performance, safety and mum cell temperature was achieved while decreasing the tem-
service life. Especially in automotive applications short-term perature differences by about 10%. On module level, Khateeb
peak loads thermally stress the battery cells. Temperature inho- et al. investigated passive cooling approaches for a lithium-ion
mogeneities arise, effecting ageing rates and electrical properties battery with cylindrical cells used in electric scooter applica-
with the consequences of a reduced system lifetime and diverging tions [3]. Different heat dissipation modes including natural
state-of-charge of the battery cells. By integration of Phase- convection, heat conduction, PCM and PCM filled aluminum
Change-Material together with high thermally conductive Pyro- foam were compared. By use of the latter encapsulated PCM, a
lytic Graphite Sheets and graphite gap filler pads, thermal peak reduction of the maximum temperature by 50% and a uniform
shaving and temperature homogenization are implemented to temperature distribution were achieved. Similar results were
reduce temporal and spatial temperature gradients. In this paper
shown by Ling et al. [4], who investigated passive PCM cool-
a battery module with prismatic cells and advanced thermal ma-
terials was investigated on test-bench. Measurements of tempera-
ing systems with and without forced convective cooling. Re-
ture rise, differences and distribution at different boundary con- garding the passive system, insufficient convection led to ac-
ditions were performed to evaluate the designed concept. The cumulation of heat within PCM and cells, whereby the speci-
measured temperature profiles were captured in all three spatial fied temperature range of the cells was exceeded. However, air
directions within the battery module. As a result, without in- cooling enabled permanent cycling of the battery module.
creasing the volumetric overhead, the maximum rise in tempera- To overcome the disadvantage of a low thermal conductivi-
ture was reduced by 13%, while the additional Phase Change
ty, Lin et al. combined PCM with graphite sheets [5]. By at-
Material has 5% of the cells mass. Furthermore the temperature
taching the sheets between prismatic battery cells, the heat
difference on module level was kept below 5K at all conditions
above zero degrees ambient temperature even at the maximum
transfer to the PCM aligned at the lateral sides was improved.
specified discharge current and continuous cycling. Furthermore a maximum temperature difference below 5K was
achieved. Another approach to increase heat conductivity of
Keywords: Prismatic Cell Battery System, Phase Change Phase Change Materials was implemented by Wang et al. by
Material, Temperature Homogenization, Thermal Peak Shaving, adding aluminum foam [6]. A comparison of pure paraffin with
Graphite Foils aluminum/paraffin composite showed differences in tempera-
ture uniformity and the duration of the PCM melting process.
Acronym Term
Babapoor et al. combined PCM with carbon fibres to enhance
PCM Phase Change Material
PGS Pyrolytic Graphite Sheet the heat conductivity [7]. By studying the influence of the fibre
TSS Thermal Storage Material length and weight percentage on thermal performance, they
realized a drop in temperature rise by 45%.
I. INTRODUCTION
A space-saving strategy using Phase Change Materials was
As safety aspects, temperature dependent performance and developed by Javini et al. [8]. In their battery module the cells
ageing rate demand an efficient and optimized thermal man- were alternately separated by cooling plates and foam, the lat-
agement system [1], the integration of Phase Change Materials ter compensated cell swelling and provided initial pressure. By
(PCM) with their particular thermal characteristics is widely replacing the conventional foam by foam soaked with PCM, a
investigated within lithium ion battery applications. In general, reduction of the peak temperature by 8K was accomplished.
the aspired benefit is an improved thermal behavior in terms of The thermal management system designed by Rao et al. uses
temporal temperature homogenization and thermal peak shav- benefits of both PCM and liquid cooling [9]. By surrounding
ing. However, possible adverse consequences are reduced the cells with Phase Change Material and multiple mini-
gravimetric and volumetric energy and power densities, heat channel cooling pipes, high specific heat and low thermal re-
accumulation and thermal inertia due to limited heat conduc- sistance of liquid cooling were combined. An optimum config-
tion capabilities. uration was investigated on simulation level, resulting in a
Various approaches of PCM integration in battery systems maximum temperature difference of less than 6K.
are found in literature. In [2], numerical studies with a pris-

978-1-5090-3474-1/16/$31.00 ©2016 IEEE 5519

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on December 20,2024 at 07:57:28 UTC from IEEE Xplore. Restrictions apply.
 In this paper, the battery module with prismatic cells, PCM Taking all factors into account, two integration possibilities
and liquid cooling developed in [10], is evaluated on test are derived: with PGS relative to PCM on same or opposing
bench. By applying various temperature sensors the tempera- side of the cell. The latter variant is selected due to several rea-
ture distribution is accurately captured. The influence of sons. No additional thermal gap filler or electrical insulation
boundary conditions such as ambient temperature and several foil are needed. With a direct connection of PGS to aluminum
charge and discharge current rates is shown. The results are sheets as a supporting structure, mechanical damages to the
evaluated with focus on temperature homogenization and thin graphite sheets are excluded. Sharp radii of PCM sheet
thermal peak shaving. According to the above referenced edges otherwise would induce mechanical stress. Besides, ad-
works on temperature homogenization in battery modules, the ditional constructive effort would be necessary to connect the
goal for a maximum temperature difference is set to 5K. opposing PGS layer to the cooling plate.
As the phase change effect is in focus of this investigation
II. BATTERY MODULE DESIGN AND TEST SETUP and the liquid cooling performance is considered as sufficient,
the relatively higher thermal resistance of this variant is accept-
A. Battery Module Design ed with the benefit of lower construction efforts and a robust
As shown in Fig. 1 the cooling sheets are preassembled module design. A more detailed description about the thermal
with Thermal Storage Sheets (TSS), Pyrolytic Graphite Sheets concept can be found in [10]. Table I gives a general overview
(PGS) and Graphite Pads. of the material properties. TSS, PGS and Graphite Pads are
produced by Panasonic.
Cooling Sheet
Cells
TABLE I. DESIGN PARAMETERS

Component Parameter Value

TSS Thickness 1 mm
Graphite TSS Isotropic thermal conductivity 0.2 W/m·K
Sheets PGS Thermal Storage
Sheets TSS TSS Heat capacity at phase change 13.8 kJ/K
Battery Cells Capacity 20 Ah
Battery Cells Heat Capacity 10.4 kJ/K
Gap Filler
Pad PGS Thickness 70 µm
Liquid Cooling PGS In-plane thermal conductivity 1850 W/m·K
Plate
Graphite Pad Thickness 0.5 mm

Fig. 1: Cooling sheets assembly in exploded view Graphite Pad Through-plane thermal conductivity 10 W/m·K

By the integration of PCM, two benefits in thermal man-

agement are aimed. With respect to temporal homogenization, B. Test Set-Up
material with a phase change temperature at the upper specified In Fig. 2 the experimental set-up is schematically illustrated
temperature range of the cells is selected to smooth thermal in a block diagram. Basically the module is in a temperature
peaks. In addition, due to the sensible heat outside the melting chamber, connected to an electrical source/sink and a thermo-
range, the overall temperature rise is further reduced. The sec- stat. The voltages and temperatures are monitored and captured
ond aim is spatial homogenization. By separating the cells with with data loggers. Flat Negative Temperature Coefficient sen-
the low thermal conductive PCM sheets the parabolic tempera- sors (NTC) are attached between the cells, at the bus-bars and
ture distribution through the module with a hot spot in the cen- on the cooling plate. The symmetrical design of the battery
ter is smoothed. The internal structure of a prismatic cell in- module is taken into account when positioning the sensors.
cluding foils, plastic elements or air gaps has influence on its
cooling characteristics. With infrared thermography the lateral
surfaces of the cell were identified as preferred cooling inter- Temperature Data Logger
Chamber
faces. Thus, sheet shaped PCM is placed in this area.
+ NTC Sensors
Furthermore several boundary conditions have to be con- Source Battery
←
Thermostat
sidered. As pointed out in the introduction, heat has to be ex- Sink Module
→
- Voltage Sense
tracted from the PCM. Therefore PGS heat spreaders are con-
nected with the PCM sheets. For reasons of costs, a single pla- USB
PC Data Logger
nar cooling plate positioned at the bottom of the cells is fore-
seen requiring bended L-shaped cooling sheets for heat spread-
ing. In addition, an electrical insulation and a thermal connec- Fig. 2: Schematic representation of experimental set-up
tion between cells and cooling sheets with gap filler material is
necessary. The conformable PCM would be suitable, but will
increase the thermal resistance.

5520
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on December 20,2024 at 07:57:28 UTC from IEEE Xplore. Restrictions apply.
 The gathered data are processed via PTC Mathcad with a module active cooling is essential, hence the further tests are
calibration table to compensate measurement errors like pro- conducted with activated liquid cooling.
duction tolerances or self-heating of the NTCs. Fig. 3 shows
the module without housing in the temperature chamber.

Power cables Temperature sensors busbars Voltage measurement

Fluid cooling Temperature sensors cells Fig. 5: Cycle test with activated and deactivated liquid cooling

Fig. 3: Battery module in temperature chamber A. Thermal Peak Shaving

Single discharge processes are performed with 4C, 8C and
III. RESULTS AND DISCUSSION the maximum specified 10C current rates. In Fig. 6 the 10C
discharge curve is exemplarily pictured. A local change in
The following test results are grouped into the categories slope can be identified at roughly 36°C, caused by the phase
thermal peak shaving and spatial temperature homogenization. change of the TSS. This effect becomes more apparent when
All results are referred to the coordinate system shown in Fig. active cooling is disabled. While discharging, the encapsulated
4. The origin of the coordinate system is located in the module PCM changed to liquid phase. The fact that the rise in tempera-
center of symmetry. Temperature distribution profiles are pol- ture at high discharge currents is only marginally influenced by
ynomial interpolated with the position of the sensors highlight- activated or deactivated liquid cooling, shows that the system is
ed in the following diagrams. mainly dependent on the thermal capacity at high loads.
Against this background, the use of PCM for compensation of
peak loads seems reasonable.

Fig. 4: Coordinate system located in module center

Fig. 5 shows voltage and temperature evolution at activated

and deactivated liquid cooling with an initial temperature of
25°C. The discharge process without liquid cooling was Fig. 6: Discharging with activated and deactivated liquid cooling at 10C
stopped at the 3rd cycle, as one bus-bar temperature exceeded current rate
the maximum cell temperature specified by the manufacturer,
To evaluate the impact of the TSS phase change, cycle test
while at activated cooling a state of equilibrium has been
with 6C/2.5C discharge/charge currents were performed with
reached after the 1st cycle. A temperature difference of 15K is
different initial temperatures. While at an initial ambient and
obtained between the maximum cell temperature and the cool-
cooling temperature of 25°C the phase change range is not
ant temperature of 25°C at 6C discharging. To conclude, as
reached, the solid phase range is exceeded for 30°C and 35°C,
assumed in Chapter I. for continuous operation of the battery
respectively. Fig. 7 shows a declining temperature rise for in-
creasing initial temperatures, due to the phase change effect.

5521
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on December 20,2024 at 07:57:28 UTC from IEEE Xplore. Restrictions apply.
 The curves begin diverging when the melting range of the TSS
is reached. The maximum temperature rise of the curves differs
by 23%, from which 10% are attributed to a reduction of the
temperature dependent internal resistance according to cell
manufacturer’s measurements and 13% are connected to the
melting enthalpy of the PCM. The slight variation in slope of
temperature and voltage curves at charging is connected to
thermodynamic processes in the cell.

Fig. 8: Temperature distribution along x-axis at 10C discharge with cooling

Tinitial
Tinitial
Tinitial

Fig. 7: Demonstration of the phase change effect

B. Spatial Temperature Homogenization

Table II. shows, dependent on discharge current rate, the
maximum temperature rise Trise,max and the maximum tempera-
ture difference within the module ∆Tmax in all three spatial
directions. A maximum temperature difference, which is corre-
lating with the temperature rise, of 5K was measured. The
highest differences occurred in y-direction while the differ-
ences in z-direction were mainly dependent on the active cool-
ing. Fig. 9: Temperature distribution along z-axis at 10C discharge with cooling

TABLE II. MEASUREMENT DATA FOR SINGLE DISCHARGE OPERATIONS The temperature curve in y-direction is pictured in Fig. 10.
Due to the thermal separation of the cells, the temperature
Discharge
current rate
Trise,max (K) ∆Tmax,x (K) ∆Tmax,y (K) ∆Tmax,z (K) along the y-axis is nearly constant. The temperature differences
from Table II. are to be explained with a temperature drop in
4C 8.5 0.7 2.3 1.5
the rear region at cell 12 “c12” of the module, as the front plate
8C 16.8 1.5 4.4 3.3 has a different mechanical structure and is also in thermal con-
nection with the cooling plate. An amendment of this front
10C 21.2 2.1 5.0 3.6
plate design would have improvement potential in terms of
temperature homogenization.
The temperature profiles at different time steps during the
discharge process are represented at the following graphs. Fig.
8 shows an yz-plane symmetrical temperature distribution with
a hot spot at the center of the cell surfaces. As expected, a tem-
perature gradient along the cooling sheets in vertical direction
is observed (Fig. 9).

5522
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on December 20,2024 at 07:57:28 UTC from IEEE Xplore. Restrictions apply.
 Temperature differences between cell surface and the laser
welded aluminum bus-bars were observed and are visualized in
Fig. 13. These differences are dependent on current rate and
decline during the discharge process as the cells are heating up.
This is associated with the ohmic resistances of the bus-bars
and current collectors, the weld section and insufficient cooling
possibilities.

Fig. 10: Temperature distributions along y-axis at 10C discharge with cooling

Concerning the temperature distribution in y-direction

while cycling, Fig. 11 shows an even temperature curve. The
roughly symmetrical parabolic distribution without cooling is
depicted in Fig. 12.
Fig. 13: Temperature differences between cell and bus-bar while discharging
at different current rates

The integration of additional thermal mass in form of PCM

implies adverse effects for low temperatures behavior. Due to a
sharply reduced usable electrical capacity of Li ion batteries at
subzero temperatures, for a sufficient functional capability a
preconditioning of the battery module is required. The reduced
usable electrical capacity is indicated in Fig. 14 by shorter cy-
cles (2.5C charge, 6C discharge) at lower temperatures, as the
voltage limits are reached earlier.
Cycle tests at low ambient temperatures at -25°C were con-
ducted to demonstrate preconditioning by the fluid. As it can
be seen, preconditioning with internal heat generation by cy-
cling and preconditioning by fluid with heating power 2.6kW
are comparable in terms of heat up rate. But it has to be noted,
that preconditioning with the fluid induces significantly higher
temperature differences in z-direction.
Fig. 11: Temperature distribution along y-axis while cycling with cooling

Fig. 14: Cycling at low ambient temperature

Fig. 12: Temperature distribution along y-axis while cycling without cooling
for cycle 1 to 3

5523
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on December 20,2024 at 07:57:28 UTC from IEEE Xplore. Restrictions apply.
 IV. CONCLUSION REFERENCES
A liquid cooled prismatic cell battery module with TSS [1] M. Giegerich, S. Koffel, R. Filimon, J.L. Grosch, T. Fühner, M.M.
Phase-Change-Material, Pyrolytic Graphite Sheets PGS and Wenger, M. Gepp, V.R.H. Lorentz, „Electrothermal modeling and
characterization of high capacity lithium-ion battery systems for mobile
graphite gap filler pads was investigated on a test bench at dif- and stationary applications”, IEEE Industrial Electronics Society 39 pp.
ferent ambient conditions. Single discharge operation and cycle 6721, 2013
loads were applied and temperature rise and profile were cap- [2] N. Javani; I. Dincer; G.F. Naterer; B.S. Yilbas, “Heat transfer and
tured to evaluate the battery module in terms of temperature thermal management with PCMs in a Li-ion battery cell for electric
homogenization and thermal peak shaving. The goal of a max- vehicles”, International Journal of Heat and Mass Transfer 72 pp. 690,
imum temperature difference below 5 K, without increasing the 2014
volumetric overhead, was achieved with potential for further [3] S.A. Khateeb; S. Amiruddin; M. Farid; J.R. Selman; S. Al-Hallaj,
“Thermal maangement of Li-ion battery with phase change material for
improvement. In addition, the maximum temperature rise was electric scooters: Experimental validation”, Journal of Power Sources
reduced; this was attributed to the phase change enthalpy of 142 pp. 345, 2005
TSS. The influence of liquid cooling on temperature rise and [4] Z. Ling; F. Wang; X. Fang; X. Gao; Z. Zhang, “A hybird thermal
differences was shown for different ambient conditions. As a management system for lithium ion batteries combining phase change
next step, temperature homogenization on module level is in- materials with forced-air cooling“, Applied Energy 148 pp. 403, 2015
vestigated and evaluated in terms of ageing for a different cell [5] C. Lin; S. Xu; G, Chang; J. Liu, “Experiment and simulation of a
type and chemistry with focus on low volumetric overhead and LiFePO 4 battery pack with a passive thermal management system using
composite phase change material and graphite sheets”, Journal of Power
a lightweight design. Sources 275 pp. 742, 2015
[6] T. Wang; K.J. Tseng; J. Zhao, “Development of efficient air-cooling
ACKNOWLEDGMENT strategies for lithium-ion battery module based on empirical heat source
model”, Applied Thermal Engineering 90 pp. 521, 2015
The research leading to these results has received funding
[7] A. Babapoor; M. Azizi; G. Karimi, “Thermal management of a Li-ion
from the European Union’s Seventh Framework Program for battery using carbon fiber-PCM composites”, Applied Thermal
research, technological development and demonstration under Engineering 82 pp. 281, 2015
grant agreement no 608770 (“eDAS”). [8] N. Javani; I. Dincer; G.F. Naterer; G.L. Rohrauer, “Modeling of passive
thermal management for electric vehicle battery packs with PCM
The authors acknowledge the Panasonic Device Solution between cells”, Applied Thermal Engineering 73 pp. 307, 2014
Business Division for the supply of PGS, TSS and Graphite- [9] Z. Rao; Q. Wang; C. Huang, “Investigation of the thermal performance
Pad products and their technical support and collaboration. of phase change material/mini-channel coupled battery thermal
management system”, Applied Energy 164 pp. 659, 2016
[10] M. Gepp; R. Filimon; S. Koffel; V.R.H Lorentz; M. März, “Advanced
thermal management for temperature homogenization in high-power
lithium-ion battery systems based on prismatic cells” IEEE International
Symposium for Industrial Electronics pp. 1230, Jun. 2015

5524
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on December 20,2024 at 07:57:28 UTC from IEEE Xplore. Restrictions apply.

Powered by TCPDF (www.tcpdf.org)