Battery Thermal Management System

Design Modeling
Gi-Heon Kim, Ph.D
Ahmad Pesaran, Ph.D (ahmad_pesaran@nrel.gov)
National Renewable Energy Laboratory, Golden, Colorado, U.S.A.

EVS 22
October 23-28, 2006
Yokohama, Japan

With support from

High Power Energy Storage Program (Tien Duong and Dave Howell)
Office of FreedomCAR and Vehicle Technologies
Energy Efficiency and Renewable Energy Office
U.S. Department of Energy
NREL/PR-540-40848
Motivations & Objectives
Motivations
ƒ Battery thermal management is critical in achieving performance and
extended life of batteries in electric and hybrid vehicles.
ƒ Appropriate models that can predict thermal behaviors of batteries
shorten the development process for improving battery system design.

Objectives of this Study

ƒ To investigate the impact of cooling strategies with different coolant
systems; air and direct/indirect liquid cooling
ƒ To evaluate system thermal responses and their sensitivities as a
function of controllable system parameters
ƒ To provide battery thermal management system design insight by
identifying analyses and approaches that engineers should consider
when they design a battery thermal management system for electric
and hybrid vehicles

3
 Battery Thermal Management
Modeling at NREL

Cell Characteristics
• Shape: Prismatic/Cylinder/Oval etc
• Materials/Chemistries
• Size/Dimensions/Capacity
• Thermal/Current Paths inside a Cell
Module Cooling Strategy Operating Conditions
• Passive control with phase change • Vehicle Driving Cycles
• Coolant type: Air/Liquid • Control Strategy
• Direct Contact/Jacket Cooling Design Process • Ambient Temperature
• Serial/Parallel Cooling • etc
• Terminal/Side Cooling 3D
System
• Module Shape/Dimensions Component
Analysis
• Coolant Path inside a Module Analysis
• Coolant Flow Rate
• etc

Battery Thermal Responses

• Temperature History Cells/Module/Pack
• Temperature Distribution in a Cell
• Cell-to-Cell Temperature Imbalance in a Module
• Battery Performance Prediction
• Pressure Prop and Parasitic Power
• etc. 4
Approach
ƒ NREL has developed a 3-D electro-thermal model to predict
thermal response of real cells – focus on cell internal
temperature (EVS-21)
ƒ In this study, focus is mostly external to cell – fluid side
– Air cooling
– Liquid cooling
• Direct
• Indirect
ƒ The battery management system response were evaluated
using
– Fully developed laminar channel flow analysis
– Computational Fluid Dynamic (CFD) analysis
ƒ The system responses oof interest were
– Coolant temperature change (outlet - inlet)
– Temperature difference between cell surface and bulk coolant
– Pressure drop in coolant channel (fluid power requirements)

5
Typical Parallel Cooling Analysis
(all the cells are treated the same)
0.5Dh Cell Specifics
ƒ Dimensions
.
. Q Dcell ƒ Heat Generation
mc
Heat Transfer Fluid
Lcell Properties
T ΔT2 ƒ Density
ƒ Specific Heat
ΔT1 ƒ Thermal Conductivity
ƒ Viscosity
x
P Control Parameters
ƒ Coolant Mass-flow Rate (mc)
ΔP ƒ Coolant Channel Dimension (Dh)

System Responses of Interest

x
ΔP : Channel Pressure Loss
cell surface temperature
coolant mean temperature ΔT1 : Coolant Temperature Change
pressure ΔT2 : Coolant-Cell Temperature Difference
6
 Example Study

Fully Developed Laminar Flow Analysis

Annular Channel Cooling Cell Specifics
If coolant channel gap is small enough
Q=2W Dcell=5cm
compared with cell diameter,

c f Re = 24 Lcell=10cm

Nu = 5.385 Heat Transfer Fluid Specifics

VDh hDh Direct Indirect
Water/Glycol
Nu =
Mineral Oil
where, Re = , Air Cooling Cooling
ν k Liquid Liquid

?ρ(kg/m3) 1.225 924.1 1069

hydraulic diameter mean velocity
Dh = 4 Ac / p V= m & / (ρAc) cp (J/kg K) 1006.43 1900 3323
where where
k (W/m K) 0.0242 0.13 0.3892
Ac : area section of the channel ρ : coolant density
p :perimeter of the channel m& : coolant mass flow rate ν (m2/s) 1.4607e-5 5.6e-5 2.582e-6

System Responses Control Parameters

m& : Coolant Mass Flow Rate
ΔP = 4τoL/Dh ΔT1 = Q/( m
& c p) ΔT2 = Q/(πDcellLh)
Dh : Hydraulic Diameter of Coolant
Channel
7
 m& cν ∂ΔP m& ν
Channel Pressure Loss ΔP ΔP ~
Dh
3
∂Dh
~ − c4
Dh
4
Mass-rate specific pressure Square-of-mass-rate
6
specific power
10 10
(a) Air (b) Air
Mineral Oil 5 Mineral Oil
10
3 Water/Glycol Water/Glycol
10
4

Power/mc2 [W/(kgs-1)2]
10
ΔP/mc [kPa/(kgs-1)]

2
10 3
10

2
10

.
1
10
.

1
10
0
10
0
10

-1 -1
10 10
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Dh (mm) Dh (mm)

m& c ν
2
ƒ Large difference in kinematic viscosity Æ, ΔP varies in very different ranges Wf ~
ƒ ΔP inversely proportional to Dh3 at Dcell >> Dh and laminar flow ρDh 3
ƒ The channel pressure loss changes are very sensitive to Dh when it is small.
ƒ Due to the much smaller fluid density and consequently larger volumetric flow rate at
given mass flow rate, the air cooling system requires much higher flow power
8
 Coolant Temperature Increase ΔT1
p
4 ƒ To achieve the temperature uniformity
Air
over a cell, it is preferred to keep
3.5 ΔT1 Mineral Oil
coolant temperature change in the
Water/Glycol

3 channel as small as possible.

ƒ ΔT1 is inversely proportional to coolant
heat capacity flow rate.
Temperature (oC)

2.5

2 ƒ Therefore, increasing mass flow rate is

not as effective for reducing coolant
1.5 temperature change in large flow rate
cooling as it is in a small flow rate
1
cooling.
0.5 ƒ A little change of flow rate can greatly
affect the coolant temperature change
4 and consequently cell temperatures at
0
0.5 1 1.5 2 2.5 3 3.5
Mass-Flow Rate (g/s)
small coolant flow rate (especially for
1 ∂ΔT1 1 1 air system having small cp).
ΔT1 ~ ~−
m& c c p ∂m& c p m& c 2
9
 Temperature Difference ΔT2
between Coolant Bulk and Cell Surface Heat Transfer Coefficient
Temperature Difference Between Coolant & Cell Surface
8 1000
Air Air
900 Mineral Oil
ΔT2
7 Mineral Oil
Water/Glycol 800
h Water/Glycol

6
700
h at direct contact surface
Temperature (oC)

5 600
for water/glycol system

h (W/m 2K)
4 500

400
3

300
2
200
1
100

0 0
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Dh (mm) Dh (mm)

ƒ ΔT2 varies linearly with Dh with slope being proportional to 1/k .

ƒ ΔT2 rapidly increases with Dh in air cooling due to its small thermal conductivity.
ƒ The heat transfer coefficient (h) evaluated at cell surface for water/glycol jacket cooling is
greatly reduced and not sensitive to channel height due to added thermal resistances
between coolant and cell surface.
ƒ High h system reduces ΔT2, and removes heat fast from small temperature difference. 10
 Maximum Cell Surface Temperature
relative to Coolant Inlet ΔTmax (= ΔT1+ ΔT2)
DTmax (oC) DTmax (oC) DTmax (oC)
4 4 4
5

1
7

1
2.6

1.2
5.4

1.4

1.6
4.8
3
3.2
3.4

5.8
1.6

0.8
2.2

1. 2
3.6

0.6
4.2
2

4.4

3.5 3.5 3.5

(g/s) (g/s)

4. 6
3. 8
1.8

7.2
2.8

8. 2
6. 4

7. 6
6. 2
4

5. 6

7.4
7. 8
6.8
2.4

3 6.6 3 3

1. 8
rate

5.2

8.4

4
1.
1
1

1. 2

1. 6
2.5 2.5 2.5
5

1.4
6
flow rate

8
7

2
2.6

5.4
4.8

1.
0. 8
3
3.2
3.4

5.8
flow

2.2

3.6

8. 8 8.6
4.2
2

4.4
4.6

2 2 2
4 3. 8

7. 4 7. 2
2. 8
massmass

8.2

1.4
6.6 6. 4

7. 8.6

8
6.2

1.
5.6

1.2
7
6.8
2.4

1.5 1.5 1 2 1.5

1.2 1.6
5.2

1.4
8.4

9.2

1.4 1.6
9
5

8
7
4

2 2.2
8

3. 3.4.6 8
3

5.

1 5. 1 1. 8 2.4
4.

3 2 1
. 8 4.42.44.6 8.6 1.4 1.6 1. 6 1.8
3 7.2 8.2 8.8 9. 69.8 8 2 2.2 2.6 1.6
4
.2 5. 6 6.62.66.64.8 7.4 77. 6
.8 .4 .8
1. 2.4 2.8 3 1.8 2
0.5 5
5 6 7 8 8.4 9 9.2 9 1100 .21100.4.6
1110 0.5 2.2 2.4 2. 6
2.8 3 3.2 3.4 0.5 1.8 2 2
22
2.2
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Air Dh (mm)
Dh(mm)
Mineral Oil Dh (mm)
Dh(mm) Water/Glycol Dh (mm) Dh(mm)

ƒ ΔTmax in the air system are much higher compared with other fluid
systems due to its small heat capacity and thermal conductivity.
ƒ The air system: ΔTmax is dominated by and sensitive to Dh.
ƒ The water/glycol jacket cooling system: ΔTmax is not very sensitive to
Dh, and ΔTmax is not a strict limiting thermal design factor in a
water/glycol system.
11
Optimizing Operation Parameters
Δ Tmax=3.43 oC
- Air Cooling System

Δ Tmax=4.5 oC Cell Specifics

Re=2300 Q=2W Dcell=5cm

Lcell=10cm

Confining Factors
Colored Zone satisfies
Δ T1=1.5 oC Re < 2300
ΔP < 110 Pa
Δ P=110 Pa ΔT1+ ΔT2 < 4.5 oC
ΔT1 < 1.5 oC
Δ P=37 Pa

: maximum heat transfer coefficient (h) operating point

: lowest max temperature (ΔT1+ ΔT2 ) operating point
: minimum pressure loss (ΔP) operating point 12
 Trend Validation: Module Liquid Cooling
Experiment Tested Module Cooling System
ƒ 12 li-ion cylindrical cells
ƒ Indirect jacket cooling
ƒ Coolant channels not completely
fully developed
ƒ Discharged at constant C rate
Coolant Tem perature Inc reas e
0.9

0.8

0.7

0.6
Temperature ( C)
o

0.5

ƒ Coolant (water) temperature

0.4
prediction
0.90 0.3 increase is inversely proportional
0.2
to the coolant mass flow rate.
(oC)

0.80

ƒ Coolant temperature change at

0.1
Coolant Temperature Rise

0.70
Coolant Temperature Rise ( C)

high mass flow rates is not as

0 0.5 1 1.5 2 2.5 3 3.5 4
o

0.60 M as s -F low R ate (g/s )

0.50 sensitive to mass flow rate as it is

0.40 at low mass flow rates.
0.30
ƒ Note that the magnitude of
0.20
experiment coolant temperature change at
0.10
given heat removal rate is
0.00
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045
relatively small due to large heat
Mass Flow Rate (kg/s)
Mass Flow Rate (kg/s) capacity of water/glycol.
13
 Computational Fluid Dynamic (CFD) Evaluation
Direct Air Cooling
9 Heat Generation: 2 W per Cell
9 Channel Inlet Air Temperature: 35oC
9 Air Mass Flow Rate: 1.33 g/s per Cell
9 Channel Gap Height: 1.1 mm (Dh= 2.2 mm)

Geometry & Mesh Temperature Distribution

Coolant Channel

Can Cell Core

NOTE: Radial direction length is exaggerated.

Terminal Current Collector Terminal

14
 CFD Predictions:
Temperatures, h and Surface Heat Flux Profiles
120

100
h ΔTmax Disagreement

Heat Transfer Coefficient (W/m2K)

39 80

38.5
ƒ CFD captures entrance effects.
60

38 40

ƒ CFD model addresses battery

ΔTmax internal heat flow and captures
Temperature (oC)

37.5 20

37
axially decreasing heat flux from
0
0 1 2 3 4 5 6 7 8 9 10

Heat Flux
A i l Di ( )

cell to air.
250
36.5

ƒ Internal heat flow through high

200
36
Heat Flux (W/m 2)

ΔT1 conductivity material distributed

Air Mean Temperature
35.5 150
Cell Surface Temperature

inside a cell (such as container

Cell Temperature at Axis
35 100
0 1 2 3 4 5 6 7 8 9 10
Axial Distance (cm)
can) makes the axial gradient of
50

cell surface temperature smaller

than that of air temperature.
0
0 1 2 3 4 5 6 7 8 9 10
Axial Distance (cm)

ΔP [Pa] ΔT [ C] 1 ΔT [ C]
o ΔT [ C] h [W/m K] ƒ This result strongly implies that
max
o
cell
o 2

channel coolant maximum maximum mean heat

pressure temperature cell surface cell internal transfer
capturing the internal heat flow
loss change temperature temperature coefficient paths and thermal resistances
to inlet air to inlet air
inside a cell are important for the
59.24 improved prediction of
Fully Developed Flow Relations
with Constant Heat Flux
109.1 1.49 3.64 N/A
cell/battery thermal behaviors.
CFD 114.2 1.50 2.89 3.41 60.96
15
 Air vs Water/Glycol CFD air water/glycol
Performance comparison analyses were jacket
Dh= 2.2 mm Dh= 2.2 mm
made between Air Cooling System and m& c= 1.33 g/s m& c= 1.33 g/s
Water/Glycol Jacket Cooling System in
order to contrast the characteristics of each
system.
Dcell = 50 mm

Lcell = 200 mm

water/glycol
air
jacket
aluminum can aluminum can 9 Heat Generation: 4 W per Cell
9 Channel Inlet Air Temperature: 35oC
cell core cell core
9 Coolant Mass Flow Rate: 1.33 g/s
9 Channel Gap Height: 1.1 mm
9 5 cm diameter, 20 cm height cell

16
Concluding Remarks
Air Cooling System
ƒ The heat transfer coefficient (h) of an air cooling system is lower than that of
liquid cooling systems.
ƒ Due to the small heat capacity of air, it is difficult to accomplish temperature
uniformity inside a cell or between the cells in a module.
ƒ The temperature difference between coolant air and cell surface is sensitive
to variations of channel height due to the small heat conductivity of air.
ƒ Heat transfer coefficient is inversely proportional to Dh, while friction pressure
loss in channel is inversely proportional to Dh3.
ƒ Increasing heat transfer coefficient by reducing channel thickness is limited
by the required blower power.
ƒ The simplicity of an air cooling system is an advantage over a liquid coolant
system.
ƒ Air cooling could have less mass, has no potential for leaks, needs fewer
components, and could cost less.

18
Concluding Remarks
Liquid Cooling System
ƒ A water/glycol solution for a jacket cooling has much lower viscosity than
dielectric mineral oil for direct cooling. Increasing the coolant flow rate in
water/glycol system is not as severely restricted by the pump power as in a
mineral oil system.
ƒ Cell/module temperature uniformity can be effectively achieved in liquid
cooling system due to the large heat capacity of liquid coolant.
ƒ Water/glycol solutions generally have a higher thermal conductivity than oil.
However, the effective heat transfer coefficient at the cell surface is greatly
reduced due to the added jacket wall and air gap layer.
ƒ Because of the added thermal resistances, h is not as sensitive to the
variation of channel height in indirect liquid cooling systems.
ƒ Liquid cooling systems are more effective in heat transfer and take up less
volume.
ƒ However, the added complexity and cost may outweigh the merits.
ƒ Maintenance and repair of a liquid cooled pack is more involved and costlier.
Indirect liquid cooling, with jackets, is easer to handle than direct liquid
cooling.
19
Concluding Remarks
General
ƒ Selection between air or liquid cooling depends on applications.
ƒ Trade-off between performance, application, and cost must be
considered
ƒ It is recommended in liquid cooling system to use a small gap coolant
channel to reduce system weight and volume by minimizing the
amount of liquid coolant in a system operated at a given coolant flow
rate.
ƒ Capturing the internal heat flow paths and thermal resistances inside
a cell using a sophisticated three-dimensional cell model is important
for the improved prediction of cell/battery thermal behaviors.
ƒ With the model we can look at turbulent flows, mixed
conduction/convection, and phase change materials.

20