Modeling the Cost and Performance of
Lithium-Ion Batteries for Electric-Drive
Vehicles
Draft Report
&EPA
United States
Environmental Protection
Agency
-------
Modeling the Cost and Performance of
Lithium-Ion Batteries for Electric-Drive
Vehicles
Draft Report
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
Prepared for EPA by
Argonne National Laboratory
Contract No. DE-AC02-06CH11357
NOTICE
This technical report does not necessarily represent final EPA decisions or
positions. It is intended to present technical analysis of issues using data
that are currently available. The purpose in the release of such reports is to
facilitate the exchange of technical information and to inform the public of
technical developments.
&EPA
United States
Environmental Protection
Agency
EPA-420-D-12-004
August 2012
-------
TABLE OF CONTENTS
LIST OF FIGURES vii
LIST OF TABLES ix
ABBREVIATIONS x
LIST OF SYMBOLS xii
ACKNOWLEDGEMENTS xv
EXECUTIVE SUMMARY xvi
1. Introduction 1
2. Battery and Cell Design Format 2
2.1 Cell Design 3
2.2 Module Design 4
2.3 Battery Pack Design 5
3. Modeling of Battery Design and Performance 6
3.1 Criteria for Power, Energy, and Life 6
3.2 Voltage at Maximum Power 8
3.3 Governing Equations 14
3.4 Calculation of the ASI 15
3.4.1 Current Collection Resistance 17
3.4.2 Potential and Current Distribution 19
3.4.3 Determination of Module Terminal Size 21
3.5 Calculation of Battery Dimensions 21
3.5.1 Cell Dimensions 22
3.5.2 Module Dimensions 22
in
-------
3.5.3 Battery Pack Dimensions 23
3.6 Additional Considerations 23
3.6.1 Maximum Electrode Thickness 23
3.6.2 Accounting for Parallel Cell Arrangements 24
4. Modeling of Battery Pack Manufacturing Cost 25
4.1 Approach 25
4.2 Materials Costs and Purchased Items 26
4.2.1 Battery Specific Materials Cost 26
4.2.2 Purchased Items Cost 30
4.3 Baseline Manufacturing Plant 31
4.3.1 Receiving and Shipping 32
4.3.2 Electrode Materials Preparation 34
4.3.3 Electrode Coating 34
4.3.4 Calendering 35
4.3.5 Inter-Process Materials Handling 35
4.3.6 Electrode Slitting 36
4.3.7 Final Electrode Drying 36
4.3.8 Control Laboratory 36
4.3.9 Cell Stacking 37
4.3.10 Current Collector Welding 37
4.3.11 Enclosing Cell in Container 37
4.3.12 Electrolyte Filling and Cell Sealing 38
4.3.13 Dry Room Management 38
IV
-------
4.3.14 Formation Cycling, Final Cell Sealing, etc 38
4.3.15 Module and Battery Assembly 39
4.3.16 Rejected Cell and Scrap Recycle 40
4.3.17 Baseline Plant Summary 41
4.4 Adjustment of Costs for Rates 41
4.5 Plant Investment Costs 43
4.6 Unit Costs for Battery Pack 44
4.6.1 Variable Costs 44
4.6.2 Fixed Expenses 45
4.6.3 Profits 46
4.6.4 Battery Pack Warranty Costs 46
4.7 Summary of Baseline Battery Cost 46
5. Description of Spreadsheet Model and Instructions for Use 49
5.1 Background 49
5.2 Instructions 49
5.2.1 Enabling Calculation 49
5.2.2 System Selection Worksheet 50
5.2.3 Battery Design Worksheet 50
5.2.4 Remaining Worksheets 54
5.3 Battery Design Format Requirements 57
5.4 Troubleshooting and General Advice 57
6. Illustrated Results 58
6.1 Number of Cells in Series 58
-------
6.2 Battery Packs for HEVs 58
6.3 Battery Packs for EVs 61
6.4 Parallel-Connected Cell Groups and Electrode Thickness 63
6.5 Effects of Manufacturing Scale on the Price of the Pack 64
7. Future Work 67
7.1 Thermal Modeling 67
7.2 Optimum Battery Voltage for Minimum Drivetrain Cost 67
7.3 Multipurpose Battery Manufacturing Plants 68
7.4 Stand-Alone Graphical User Interface for Model 68
References 69
VI
-------
LIST OF FIGURES
2.1 Prismatic cell and module design for battery packs 2
2.2 Cell sandwich inside of prismatic pouch cells 3
2.3 Coated current collector foil for prismatic electrodes 4
3.1. Summary flow of the design model 7
3.2 a) Required change in [V/U] to maintain rated power with increases in internal
resistance over the life of the battery, b) Increase in current due to lowered [V/U] 10
3.3 Change in heat rejection requirement from increases in resistance for batteries with
different designed voltages at maximum power 11
3.4 Efficiencies for batteries designed to achieve maximum power at different fractions of
their open-circuit voltage 13
3.5 The change in current and potential within the positive and negative foils. The current
collection design results in a uniform current distribution along the length of the foil ....20
4.1 Metal ingot cost contribution to the current collector foils over a 20 year period 30
4.2 Baseline lithium-ion battery manufacturing plant schematic diagram 32
4.3 Breakdown of installed capital equipment costs for the baseline plant 43
4.4 Breakdown of unit costs for baseline battery with total price to OEM of $2428 48
5.1 Iteration must be enabled for the spreadsheet model to function 49
5.2 The specific cell chemistry for the battery design is selected on the System Selection
worksheet 50
5.3 System Selection worksheet 51
5.4 Top portion of Battery Design worksheet 52
5.5 Middle portion of Battery Design worksheet 53
5.6 Bottom portion of Battery Design worksheet 55
5.7 Summary of Results worksheet 56
Vll
-------
6.1 The effect of the number of cells for LMO-Gr, 60-kW, PHEV20 packs with 7.14 kWh
total energy (70% useable) 58
6.2 Cell area-specific impedance (AST) for various cell chemistries for 25-kW HEV
battery packs delivering full power at about 238 A for 2 sec at 80% OCV 59
6.3 Volume, voltage and maximum current for 1.6-kWh lithium-ion battery packs as a
function of pack power for packs delivering full power 60
6.4 Battery pack price to OEMs at 100,000 packs per year manufacturing rate for 1.6-kWh
lithium-ion battery packs as a function of pack power 61
6.5 Weight and volume of electric vehicle battery packs with LFP, LMO, and NMC
electrodes versus graphite designed to deliver 150 kW of power at about 425 A 62
6.6 Battery pack price to OEM for LFP-G, LMC-G and NMC-G battery packs 63
6.7 Battery pack cost as a function of number of parallel cells and for different maximum
electrode thicknesses 64
6.8 The effects of manufacturing rate on the price calculated by the model for battery
packs of various cell chemistries, power capabilities and vehicle types 65
Vlll
-------
LIST OF TABLES
3.1 Criteria for designing batteries for a specific end-use application 6
4.1 Details of stated costs for cathodes, anodes, electrolyte, and separator 27
4.2 Cost equations for purchased items 31
4.3 Summary table of the baseline plant 33
4.4 Materials yields during electrode and cell fabrication 40
4.5 The effect of processing rate (R) on cost for various scale factors 42
4.6 Battery pack manufacturing investment costs 44
4.7 Unit cost of battery pack 45
4.8 Summary of results for cost of baseline battery and that of similar batteries with
double the power and double the capacity of the baseline battery 47
IX
-------
ABBREVIATIONS
AST area specific impedance
BOL beginning of life
DMC dimethyl carbonate
EC ethylene carbonate
EMC ethyl methyl carbonate
EOL end of life
EV electric vehicle
Gr graphite
GSA General, Sales, and Administration
HEV hybrid electric vehicle
HEV-HP high-power assist hybrid electric vehicle
LCO lithium cobalt oxide
LFP lithium iron phosphate
Li lithium
Li-ion lithium-ion
LMO lithium manganese spinel
LMR lithium and manganese rich
LTO lithium titanate spinel
microHEV micro or mild power assist hybrid electric vehicle
NCA lithium nickel cobalt aluminum oxide
NMC lithium nickel manganese cobalt oxide
NMP N-Methyl-2-pyrrolidone
-------
OCV open-circuit voltage
OEM original equipment manufacturer
PE polyethylene
PEP polyethylene terephthalate
PHEV plug-in hybrid electric vehicle
PP polypropylene
R&D research and development
SOC state of charge
USGS United States Geological Survey
XI
-------
LIST OF SYMBOLS
Section 3 and 5
ratio of interfacial area to electrode volume, cm"1
a
Apos
Aterm
AS I energy
AS I pOWer
C
Cp
E
F
Hj
H/W
i0
/
Iiim°mc
/„
P
Pba
q
area of the positive electrode, cm2
area of the terminal, cm2
area specific impedance for energy, ohm cm2
area specific impedance for power, ohm cm2
cell capacity, Ah.
specific heat capacity, J/g K
total energy, Wh
Faraday constant, 96485.3 C/mol
height of j, cm
aspect ratio of pouch cell
exchange current density related to the interfacial area, A/cm2
average current density, A/cm2
ionic limiting current density, A/cm2
local current density, A/cm2
thickness of j, cm
mass of j, g
number of j
battery power, W
maximum designed battery power, W
heating rate, W
Xll
-------
Q specific capacity of the electrode, mAh/g
re C-rate, h"1
rc,um limiting C-rate, h"1
TJ radius of j, cm
R universal gas constant, 8.3144 J/mol K
RJ resistance of j, ohm
T temperature, K
t time, s
UOCV.P open-circuit voltage at SOC for power, V
UOCV.E open-circuit voltage at SOC for energy, V
Vceii cell voltage, V
v square root of dimensionless exchange current
[V/U] fraction of the open-circuit voltage
Wj width of j
Xcomp compression factor
a constant, ohm cm3
P constant, ohm cm2
sact volume fraction of active material
i,k metal potential of foil k, V
PJ density of j, g/cm3
GJ conductivity of j, S/cm
Xlll
-------
Section 4
C capital cost of an installed equipment for the designed battery, $
C0 capital cost of an installed equipment for the baseline plant battery, $
p power factor
R designed battery processing rate for specific process step
R0 baseline plant processing rate for specific process step
XIV
-------
ACKNOWLEDGEMENTS
Support from the Vehicle Technologies Program, Hybrid and Electric Systems, initially under
Tien Duong and now David Howell, at the U.S. Department of Energy, Office of Energy
Efficiency and Renewable Energy, is gratefully acknowledged. The submitted manuscript has
been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory
("Argonne"). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated
under contract no. DE-AC02-06CH11357. The U.S. Government retains for itself, and others
acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to
reproduce, prepare derivative works, distribute copies to the public, and perform publicly and
display publicly, by or on behalf of the Government. We especially thank Danilo Santini of
Argonne's Transportation R&D Center for his support and suggestions in carrying out this study.
Ralph Brodd reviewed our baseline plant and made several suggestions which we have
incorporated in the present design. Fritz Kalhammer and Haresh Kamath of the Electric Power
Research Institute have reviewed our work over several years and made suggestions that resulted
in improvements. The work was done under the direction of Dennis Dees and Gary Henriksen of
Electrochemical Energy Storage who provided guidance in carrying out the work and preparing
this manuscript.
XV
-------
EXECUTIVE SUMMARY
This report details a design and cost model developed at Argonne National Laboratory for
lithium-ion battery packs used in automotive transportation. The model designs the battery for a
specified power, energy, and degree of powertrain electrification. The cost of the designed
battery is then calculated by accounting for every step in the lithium-ion battery manufacturing
process. The assumed annual production level directly affects each process step. The total cost to
the original equipment manufacturer calculated by the model includes the materials,
manufacturing, and warranty costs for a battery produced in the year 2020. At the time this report
is written, this calculation is the only publically available model that performs a bottom-up
lithium-ion battery design and cost calculation.
The purpose of the report is to document the equations and assumptions from which the model
has been created. A user of the model will be able to recreate the calculations and perhaps more
importantly, understand the driving forces for the results. Instructions for use and an illustration
of model results are also presented. Almost every variable in the calculation may be changed by
the user to represent a system different from the default values pre-entered into the program.
The model will be openly distributed to the public in the year 2011. Currently, the calculations
are based in a Microsoft® Office Excel spreadsheet. Instructions are provided for use; however,
the format is admittedly not user-friendly. Future work is proposed to create a more intuitive
user-interface with tools to prevent inappropriate values from being entered. A version that is
more user-friendly will allow for wider adoption of the model.
XVI
-------
1. Introduction
The recent penetration of lithium-ion (Li-ion) batteries into the vehicle market has prompted
interest in projecting and understanding the costs of this family of chemistries being used to
electrify the automotive powertrain. The model described here-in is a calculation method that
was developed at Argonne for estimating the manufacturing cost and performance of Li-ion
batteries for electric-drive vehicles including hybrid-electrics (HEV), plug-in hybrids (PHEV),
and pure electrics (EV). To date, a number of cost models of various levels of detail have been
published in different forms.1"10 Most available cost models for batteries are based on a single
linear equation dependent on the power and energy required such as $/kWh = a + b*kW. While a
simplistic relation like this is attractive, it is only approximate and provides little guidance in
reducing costs. The cost of a battery will change depending upon the materials chemistry, battery
design, and manufacturing process.11'12 Therefore, it is necessary to account for all three areas
with a bottom-up cost model. Other bottom-up cost models exist but are not available to the
general public and have not been explicitly detailed in an open document. The motivation for this
work is based on a need for a battery cost model that meets the following requirements:
1. Open and available to the entire community
2. Transparent in the assumptions made and method of calculation
3. Capable of designing a battery specifically for the requirements of an application
4. Based on a bottom-up calculation approach to account for every cost factor
The model is the product of long-term research and development at Argonne National
Laboratory. Over a period of years, Argonne has developed methods to design Li-ion batteries
for electric-drive vehicles based on modeling with Microsoft® Office Excel spreadsheets.11"18
These design models provided all the data needed to estimate the annual materials requirements
for manufacturing the batteries being designed. This facilitated the next step, which was to
extend the effort to include modeling of the manufacturing costs of the batteries. In the following
sections of this document, a model is presented that meets the above criteria and may be used to
analyze the effect of battery design and materials properties on the cost of the final battery pack.
Use of the model requires some basic knowledge of battery packs; however, a user does not need
to be an expert. For instance, the number of cells and thus battery pack voltage must be specified
by the user. However, default values are available for more specific requirements such as
experimentally measured values. In this way, a person with reasonable knowledge of batteries
may be able to conduct cost comparisons and "what if studies.
The battery pack design and cost calculated by the model represent projections of a 2020
production year and a specified level of annual battery production, 10,000-1,000,000. As the goal
is to predict the future cost of manufacturing batteries, a mature manufacturing process is
assumed. The model designs a manufacturing plant with the sole purpose of producing the
battery being modeled. The assumed battery design and manufacturing facility are based on
common practice today but also assume some problems have been solved to result in a more
efficient production process and a more energy dense battery. Our proposed solutions do not
have to be the same methods used in the future by industry. We simply assume the leading
battery manufacturers, those having successful operations in the year 2020, will reach these ends
by some means.
-------
2. Battery and Cell Design Format
Various cell and battery design concepts are under development at battery manufacturers. The
exact design of the battery does not have an important effect on the cost for a set cell chemistry
system; the amounts of electrode materials and the number, capacity and electrode area of the
cells, are the determining cost factors. The most common cell designs for batteries nearing large-
scale production are cylindrical wound cells, flat wound cells, and prismatic cells with flat plates.
Cylindrical cells probably have a slight advantage for the assembly of the electrode-separator
unit because of the ease of making a cylindrical winding. For the different cell designs, there are
small differences in the weights of the terminal extensions and the procedures for connecting
these extensions to the current collector sheets, with a small advantage for flat plate cells. The
flat-wound and flat-plate cells form a more compact module and have better heat rejection
capabilities than the cylindrical cells; although, each cell design can be adequately cooled for
most applications. These small differences would have minor effects on the cost of batteries
produced in high volume in a mature, automated production plant. We conclude that the cost
calculations would be relevant for batteries differing considerably from the selected design
approach.
To provide a specific design for the calculations, a prismatic cell in a stiff-pouch container and a
double-seamed module container was selected (Fig. 2.1). For this design, calculation of the
current collector and terminal resistances are easily done. The electrical performance is near
optimum in a compact, light-weight configuration. It is unlikely that we have selected the most
viable design in this short study; there may be serious flaws in some details. However, the
overall performance and low cost for the selected design will be challenging to match and will
only be met by the most successful manufacturers, those that will dominate the market.
Heat Transfer Surfaces
on Top and Bottom of
Container in Contact
with Cell Edge Seals
Double-Seamed
Module Closure
Polymer Seal of
Cell Container
to Terminal
Ultrasonic Welds
of Terminal to
Collector Foils
Cell with Stiff, Multi-
Layer Container
Cell Cross-Section Battery Module
Figure 2.1 Prismatic cell and module design for battery packs
-------
2.1 Cell Design
The prismatic cell of this design embodies individual positive and negative electrodes consisting
of current collector foils coated with electrode materials on both sides. The current collectors are
usually solid copper and aluminum foils for the positive and negative electrodes respectively. An
illustration of a segment of the cell is detailed in Figure 2.2. Each electrode is made up of active
material particles held together by a polymeric binder. A conductive additive, carbon black
and/or graphite, is added to the positive electrode and sometimes to the negative electrode. The
electrodes and separator each have porosity that is filled with the electrolyte solution. During
discharge, the Li-ions move from the negative electrode particles into the electrolyte, across the
separator, and then insert into the particles composing the opposite positive electrode. The
electrons simultaneously leave the cell through the current collection system and then enter
through the opposite side after doing external work. The materials currently used in Li-ion cells
are based on an intercalation mechanism. In this process, the Li-ion is inserted into or removed
from the crystal structure of the active material. The oxidation state of the active material, or
host, is concurrently changed by gaining or losing an electron. Other electrode materials based
on conversion reactions or electrodeposition could be implemented into the model if desired.
n:
positive
electrode
n:
n:
n:
separator
Figure 2.2 Cell sandwich inside of prismatic pouch cells.
The electrodes are efficiently prepared by coating wide sheets of foil (up to 2-meters in width)
with uncoated strips running the lengths of the foil being coated. The individual electrodes can
be cut from these sheets with little waste of electrode coating material or foil (Fig. 2.3). The
separator for these cells can be handled as a single sheet that is folded back and forth as the
electrodes are inserted. The electrodes are inserted so that all of the positive tabs extend beyond
the separator sheet in one direction and the negative tabs extend in the opposite direction. The
design model selects the number of electrodes for a set cell thickness determined by the type of
cell: HEV, 6 mm; PHEV, 12 mm; EV, 14 mm. The cell terminals are formed from flat stock to
be almost as wide as the entire cell. They are bent to the shape shown in Fig. 2.1 and
ultrasonically welded to the current collector tabs. The cell stack is then sealed between the two
halves of the cell container. The cell housing material is a tri-layer consisting of an outer layer of
-------
polyethylene terephthalate (PEP) for strength, a middle layer of 0.1-mm aluminum for stiffness
and impermeability to moisture and electrolyte solvent vapors and an inner layer of
polypropylene (PP) for sealing by heating.19'20 The two halves of the cell container are pre-
shaped to facilitate assembly. The aluminum foil in the cell container material provides stiffness
and it may be increased in thickness to assist in conducting heat to the module container.
•
1
• .. 1
•
. • .
•
. • .
1
••
•
. -.
^
•
.- n
• • ••
r
••
i
•
. . -.
•
' •. .
•
.
;
•
*^
:X
H
Electrode
Coating
~ Uncoated Area
for CC Tabs
Lines Showing
Places to Cut for
Electrodes
Figure 2.3 Coated current collector foil for prismatic electrodes
2.2 Module Design
The model designs a module casing of 0.5 mm aluminum that is sealed by double seaming, a
process that is well established and inexpensive because it is automated, rapid, and uses low-cost
capital equipment. The sealing of the module provides an additional barrier to the loss of
electrolyte solvent from the cells and the entrance of water vapor. These deleterious transfers
through the seals of pouch cells may shorten their lives to less than the desired fifteen years.19
The cells are placed on their sides in the module and the terminals of adjacent cells are connected
either mechanically with small bolts and flat springs to maintain contact or by laser welding.
Space is provided within the module casing on the left side, as sketched in Fig. 2.1, for an
electronics package that includes cell monitoring for malfunctions (temperature and voltage) and
for state-of-charge (SOC) control. The SOC control is activated whenever the battery is at rest
and it diverts charge from the cells at highest voltage to those at lowest voltage.
In the model, we assume the module is cooled only by air flowing over the upper and lower
surfaces of the module. This method of cooling is most practical if the amount of heat generated
is kept to moderate levels by defining full power as that provided at 80 % or higher fraction of
the open-circuit voltage. To enhance heat flow from the cells to the cooled walls of the module
casing, the edge seals, which include two layers of 0.1 mm thick aluminum foil (one from each
half of the cell container), can be pressed against the walls of the casing by springs shaped as
hoops. The heat transfer can be further enhanced by increasing the thickness of the aluminum
-------
layer in the cell pouches or by adding an aluminum plate between the cells. The modules can
also be cooled internally by a dielectric fluid such as a transformer coolant. For that approach,
the fluid would be brought to the module through metal tubing or be supplied to the battery
jacket and enter the modules through slots. Corrugated plates inserted between the cells would
provide flow channels for the cooling fluid. In the current model, only air cooling on the outside
of the module is provided. Further discussion of thermal issues may be found in Section 7.
2.3 Battery Pack Design
The model designs the battery pack in sufficient detail to provide a good estimate of the total
weight and volume of the pack and the dimensions of the battery jacket so that its cost can be
estimated. The modules are arranged within the battery jacket either in a single row, with the
terminals facing the same side of the pack, or in an even number of rows with the terminals in
one row facing the terminal of an adjacent row. The terminals are laid out on the module so as
not to interfere with those on the opposite row of modules, thus conserving space in the battery
pack. The modules in a row are interconnected, negative to positive terminals, by copper
connectors. The modules casings are compressed together by two steel sheets bound with steel
straps at the front and back of the battery pack. The compression is necessary to ensure intimate
contact between the active layers that make up the pouch cells that are tightly fit into the
modules. The compressive force also serves to add structural support to the module casing.
The battery jacket consists of a sheet of aluminum on each side of a 10-mm thick layer of ridged,
light-weight high-efficiency insulation. The thickness of each of the aluminum layers is selected
by the modeling program to be 1- to 2-mm thick, depending on the total volume of the modules.
The insulation slows the interaction of the battery with the external environment that cools the
battery in the winter and heats it in the hot summer weather.14
-------
3 Modeling of Battery Design and Performance
The design portion of the model calculates the physical properties of a battery based on user-
defined performance requirements and minimal experimental data. An illustration of the model is
shown in Figure 3.1. The user is asked to enter a number of design parameters such as the battery
power, number of cells and modules, and target voltage at maximum power, etc. In addition, the
user must enter one of the following three measures of energy: battery pack energy, cell capacity,
or vehicle electric range. Defining one of these values will determine the value of the other two.
An iterative procedure then solves for the user defined energy parameter (energy, capacity, or
range) and remaining battery properties by varying the cell capacity and electrode thickness. The
result is the dimensions, mass, volume, and materials requirements for the cells, modules, and
battery pack.
The model has been designed to allow the user to enter as many customized values as desired. In
this way, the model allows flexibility in the battery chemistries studies and some of the cell,
module, and battery design aspects. Hence, the focus of this report is on the method of
calculation and not the exact values chosen for a specific capacity or cell thickness. However, the
default cell design parameters as well as experimental data measured at Argonne National
Laboratory, for a number of different battery chemistries, are available for use within the model.
There are five governing equations for battery performance that calculate the current density,
battery energy, electrode area, electrode thickness, and resistance. The voltage at maximum
power and the area specific impedance (AS I) are two important parameters in the design model
for calculating the battery performance. Most of the discussion will be spent on these two
properties.
3.1 Criteria for Power, Energy, and Life
In order to fully specify a battery design, the user of the design model must supply criteria for
power, energy, and life. These criteria will depend on the application for which the battery will
be used. While the user may change some of the settings as they prefer, we list our suggestions in
Table 3.1. The battery type is defined by the end-use application. Hybrid electric vehicles
(HEVs), plug-in hybrid electric vehicles (PHEVs), and electric vehicles (EVs) have increasing
levels of electrification of the vehicle drivetrain. The model will use Table 3.1 or the user's
explicit inputs to size the battery correctly for the chosen application.
Table 3.1 Criteria for designing batteries for a specific end-use application
Battery Type
SOC for Rated Power, %
Power Duration, sec
SOC Range for Useable Energy, %
Cell Thickness, mm
microHEV
50
2
40-65
6
HEV-HP
50
10
40-65
6
PHEV
25
10
25-95
10
EV
25
10
15-95
14
-------
Battery Design Model
Pack Requirements Key Constraints
• power • max electrode thickness
• energy or range • target cell potential, V, at
• number of cells peak power
• cell/module format
Iterative Spreadsheet
Solves for cell capacity
and designs battery pack
by varying:
1. Cell area
2. Electrode thickness
3. Internal resistance
Cell Chemistry
Measured Properties
• pulse power ASI
• discharge ASI
• mAh/g, g/cm3
• electrode porosity
* SOC window
• physical properties
ASI = area specific impedance
Calculated Battery
Pack Properties
• dimensions
• volume & mass
• specific energy,
power
• materials required
Figure 3.1 Summary flow of the design model
-------
The microHEV is a micro or mild-hybrid that provides a moderate power level, -25 kW, for two
seconds. This design is best suited for cell chemistries capable of very high power-to-energy
(P/E) ratios. The HEV-HP is a power-assist hybrid that provides the rated power for a full 10
second pulse. The power for both HEV applications is rated at 50 % state-of-charge (SOC). The
energy available for discharge and charging is 25 % of the total energy to ensure long cycle life.
As the capacity of the HEV cells is typically small, a cell thickness of 6 mm is used. The PHEV
utilizes a much larger portion of the total energy, 70 %. At the end of discharge, the PHEV
battery is operated in a charge sustaining mode. Therefore, the power rating for the battery is
determined at 25 % SOC. PHEV cells should be much larger than HEV cells and thus a cell
thickness of 10 mm is assumed. Finally, EV batteries use 80 % of their total energy with their
power rated near the end of discharge. Cell thicknesses are set to 14 mm to accommodate a high
capacity design. As noted later in the report, selecting a parallel arrangement of cells
automatically assumes a cell thickness of 6 mm regardless of the end-use application.
Additionally, the use of negative electrodes operating at potentials high above the lithium metal
potential may extend the upper end of the available SOC range from 95 to 100 %. The lithium
titanate spinel, Li4Ti5Oi2 (LTO), negative electrode is an example of intercalation electrode with
almost no risk of plating lithium metal during a charge pulse. On this basis, the available energy
for LTO-based Li-ion chemistries is suggested to be 75 % for PHEVs and 85 % for EVs.
Accounting for capacity and power fade in the battery requires the user to design the battery with
the appropriate excess energy and power at the beginning-of-life (BOL). Defining the voltage at
which maximum power is achieved at BOL is one way to set the allowable power fade over the
life of the battery. This is discussed in detail in the following section. Capacity or energy fade
must be accounted for by over-sizing the battery at BOL. If the user has a certain fade
requirement, then the BOL energy may be increased to meet the end-of-life (EOL) target. The
design model does not attempt to predict fade rates or even suggest an allowable fade for a
specific application. It is our view that many aspects of materials chemistry, cell design, and
battery use directly affect the decay of the battery pack. Hence, we allow the user to make
accommodations for decay as he or she believes is necessary.
3.2 Voltage at Maximum Power
The voltage at which a cell reaches the designed maximum power is one of the most important
factors in the design of a battery. However, this specification is one of the least discussed aspects
of battery design. The voltage at maximum power, Vceii, is a measure of the largest polarization
the cell will undergo during operation at the BOL. This initial value has a direct effect on round-
trip battery efficiency, heat removal requirements, cold-cranking power, and allowable power
fade. A basic calculation shows the maximum achievable power for a battery at BOL is at 50 %
of the open-circuit voltage (OCV). Operating at these conditions would result in an inefficient
battery and require a significant cooling system to reject heat. More importantly, the battery will
never be able to reach this power level after any increase in impedance occurs. With all certainty,
the impedance of a battery will rise with time and the power rating of battery will no longer be
accurate. The remainder of this section presents a discussion for setting the BOL voltage at
maximum power at no less than 80 % of the open-circuit voltage. Defining the voltage as a
fraction of the OCV, [V/U], allows for direct calculation of all the necessary battery properties
(see for example Eq. 3.6 or 3.8 in the section 3.3).
-------
The allowable increase in battery resistance over the life of the battery is a function of the
designed voltage for maximum power. In general, designing the battery to achieve maximum
power at a higher [V/U] allows for larger resistance or impedance increases over the lifetime of
the battery. Figure 3.2 created from Eq 3.1 displays how the voltage at maximum power will
change to meet the designed power as the internal resistance of the battery increases. Clearly,
achieving BOL power at a high fraction of the OCV allows for greater degradation within the
usable lifetime of a battery. If the minimum voltage is 55 % of the OCV, the allowable increase
in resistance for batteries designed for BOL max power at 70, 80, and 90 % OCV is 18, 55, and
175 %. The consequence of achieving the power at lower and lower fractions of the open-circuit
voltage is that both electric current and heat generation will increase over the lifetime of the
battery, Figure 3.2b and Figure 3.3. The proper design of a battery will account for the changes
over the entire lifetime and not just desired behavior at BOL.
V_
u
y
(3.1)
The level of heat production is significantly different at BOL for batteries designed to meet
maximum power at differing fractions of the open-circuit voltage. We may compare the
differences in designed [V/U] by assuming the resistive heating (joule heating) is the most
significant factor in determining the heat generation, Eq. 3.2. We also reasonably assume the ASI
will not change significantly in the range of current densities and electrode thickness we vary in
the comparisons. From this point, we can analyze the difference in heat generation from different
designed [V/U] values, Eq 3.3.
U2 1-
qj=(Aposl)2RJ=-
U
R;
(3.2)
42
y
u
1-
u
(3.3)
The ratio of resistances may be found by equating the power for the two cases. Then the
resistances, and areas if the ASIs are equivalent, are determined solely by the fraction of the
open-circuit voltage at which they achieve maximum power, Eq 3.4. Then substitution will give
the ratio of heat production at maximum power for the two cases, Eq. 3.5. A battery that achieves
maximum power at 80 % of OCV will have a heat production at maximum power that is 2.3
times higher than one designed at [V/U] = 90%. A battery producing power at 70 % of the OCV
will have 3.9 time higher heat generation than at [V/U] = 90 %.
-------
I
o
O.
re
O
O
H—
O
c
o
o
re
Designed
BOL [V/U] at
max Power
50
80
70
60
50
40
Designed
BOL [V/U] at
max Power
0 25 50 75 100 125 150 175 200
Increase in Resistance, %
Figure 3.2 a) Required change in [V/U] to maintain rated power with increases in internal
resistance over the life of the battery, b) Increase in current due to lowered [V/U].
10
-------
Designed
BOL [V/U] at
max Power
90%
80%
0 25 50 75 100 125 150 175 200
Increase in Resistance, %
Figure 3.3 Change in heat rejection requirement from increases in resistance for batteries with
different designed voltages at maximum power.
>
4
A<\I A
rl>J1 powerl'rlpos2
\ A^I power2Aposl
u_
}j
,('-
1-
"y"
u _
"y"
u _
j
J
(3.4)
"y"
u _
"y"
u _
(
i-
i
T
2[
"y"
_f/_
"y"
c/_
^
2J
]
J
(3.5)
11
-------
Two straightforward design changes will enable operating a battery at 90 % of OCV compared to
80 % while maintaining the same power output. First, a second identical battery may be
connected in parallel to the original battery. This will lower the resistance of the battery pack by
one half but will also double the energy and cost of the battery. A more realistic approach is to
reduce the electrode thickness by coating a larger separator area. The capacity of the cell is
maintained while minimizing increases in cost from a larger separator, current collector and
packaging area. This approach is feasible as long as the reduced electrode thickness is above the
increase in AST, > 20 microns as discussed in detail below.
The efficiency of a battery defines the heat rejection requirements and may be measured or
calculated. Measurement of round-trip efficiency of a battery is best performed by using a
calorimeter to measure the heat given off during the cycling of the battery. The calorimeter
removes the requirement of knowing the exact SOC of a battery during the entire drive cycle.
Calculation of the round-trip efficiency of a battery requires a detailed transient battery model
within a vehicle simulation program to exercise the battery over the many acceleration and
deceleration periods that occur during a drive cycle. The interesting result is that the same battery
will have different power ratings depending on what level of round-trip efficiency the user is
willing to accept.
Figure 3.4 shows the efficiency of a battery as a function of the designed potential at which the
battery reaches maximum power. The figure is created using Equations 3.1 and 3.4 above. Each
line may be considered a different drive cycle, or duty load, for a battery with the same energy
but different impedance (changing separator area). The straight, solid black line represents the
efficiency of the battery operated only at maximum power, P/Pnwx = 1. In example, a battery
designed at [V/U] =0.8 will have 80 % efficiency for a single discharge pulse at maximum
power. Likewise, a battery designed at 0.9 will be 90 % efficient at maximum power. Batteries
are normally operated in the area above the line of the maximum power. Therefore, the other
curves represent the efficiency of discharging a battery at power levels below maximum power
(typical driving conditions). Consider two batteries each designed for a maximum power of 100
kW although one achieves this power at a [V/U] = 0.9 and the other at 0.7. If the two batteries
are discharged at 45 kW, P/PmaX = 0.45, the battery designed at [V/U] = 0.9 will be 6.4 % more
efficient. This is significantly less than the 20 % efficiency improvement realized when operated
at maximum power. The efficiency penalty is reduced as the battery operates less and less near
the designed maximum power.
12
-------
100%
50%
0.5
0.6
0.7
0.8
0.9
Designed [V/U] at Max Power
Figure 3.4 Efficiencies for batteries designed to achieve maximum power at different fractions
of their open-circuit voltage. Comparative efficiency lines are shown for equivalent power
demands over a period of battery operation.
13
-------
3.3 Governing Equations
The five coupled, algebraic equations that govern the battery design are presented in this section.
While these equations are perhaps the most important, many other equations are used to fully
define the battery mass and volume. These other equations will be specified where necessary in
the following subsections.
The user of the model specifies the required maximum power, Pbatt, of the battery. This power is
translated to a current density, /, in Eq 3.6 using the area of the positive electrode, Apos, the
number of cells, NCeii, the open-circuit voltage at the SOC for power, £/OCv,p, and the fraction of
the open-circuit voltage at which maximum power is achieved, [V/U].
/=-
batt
u
(3.6)
The relationship between capacity and battery energy is described by Equation 3.7. Formally, the
energy of a battery is the product of the capacity and the average voltage at which the energy is
obtained. The average cell voltage is approximated in Eq. 3.7 by subtracting the polarization
from discharging the battery at a C/3 rate from the open-circuit voltage at the SOC for energy,
£/OCV,E- The energy for all batteries designed by the design model is calculated at a C/3 rate and
the average open-circuit voltage at 50 % SOC. The remaining necessary values are the capacity
of the cell, C, AST for energy, ASIenergy, number of cells, and area of positive electrode. Either the
battery energy or capacity may be specified. The energy may alternatively be determined from a
stated range, fraction of total energy available, and energy usage rate for the vehicle (Wh/mile).
NcellQ
C
energy
(3.7)
pos
The area of the positive electrode in Eq. 3.8 is determined largely by the area specific impedance
for power, ASIpower, and resulting voltage drop. The voltage of cell at max power, Vceii,p, is found
from the product [V/U] L^OCv,p. In general, the area of the cell will increase if the AST for power
increases. The areas of the negative electrode and separator are determined from the area of the
positive electrode. The negative electrode is taken to be 1 mm larger than the positive electrode
in both height and width to alleviate concerns of lithium plating during charge pulses. The
separator area is slightly larger than the negative electrode to prevent the electrical shorting of
the two electrodes.
P
power rbatt
u
(3.8)
The positive electrode thickness, Lpos, in Eq. 3.9 is determined from the capacity of the cell, C,
specific capacity of the electrode, Q, volume fraction of active material, sact, bulk density of the
14
-------
active material, p, and the positive electrode area. The negative electrode thickness is determined
by its specific reversible-capacity and the designed excess-capacity to prevent lithium plating
during charging. We have chosen a ratio of 1.25 negative to positive reversible-capacity (N/P
ratio) for the default value for the cells with graphite negative electrodes. LTO negative electrode
based cells are designed at a 1.1 N/P ratio because of the previously mentioned minimal
possibility of lithium deposition. The maximum allowable electrode thickness is a user defined
value. The calculation for the electrode area changes when the designed thickness is greater than
the maximum allowed (Section 3.6.1).
. (3.9)
Finally, the AST for power (and for energy to a lesser extent) is calculated using an expression
that is based on the electrode thicknesses, the current density, and the C-rate. The exact
expression will be discussed in the next session. The AST in Eq 3.10 shows the basic
dependencies with a and P being constant valued parameters.
pos
3.4 Calculation of the ASI
In most battery design scenarios, the ASI directly determines the electrode thickness to meet a
specified power-to-energy (P/E) ratio. From this electrode thickness, the area of the electrode is
set to meet the capacity requirements. Clearly, the ASI plays a significant role in the design of a
battery and particularly in the case of the P/E ratios required by automotive applications.
However, the ASI is not an inherent constant of a specific battery chemistry or cell design. The
measured value of the ASI is a complex combination of resistances within the battery resulting
from the physical processes occurring at different length and time scales. Consequently, the
measured value is a function of many factors (state of charge, pulse length, current density, C-
rate, particle size, transport and kinetic parameters, etc). The calculation used for the ASI in this
battery design model has been discussed in detail elsewhere.18 The physical meaning of the
equation will be discussed but those interested in the derivation are directed to the separate
publication. We note that the ASI described here is slightly different than the one addressed in
the paper. The thermodynamic component is removed that originated from the change in open-
circuit potential with concentration for the intercalation materials. Equation 3.11 contains the
definition of the ASI used in this document. 7ti is a positive valued current density for a
discharge pulse. 7t2 is equal to zero as it is during the relaxation period after the pulse. Time 0, tO,
is the time just before a current pulse begins, time 1, tl, is the time just before the current pulse
ends, and time 2, t2, is the time long after the current pulse when the cell is at open-circuit and
the concentration gradients have relaxed. Therefore, this ASI measurement is not troubled by
accounting for a change in open-circuit voltage with the passage of current. In general, the ASI
measured with this definition is similar, although smaller, in value to those produced using the
more standard definition used elsewhere.
15
-------
(3.11)
The AST for the electrochemical charge and discharge process is referred here-in as ASIechem- Our
calculation approach for both the AST for power and for energy involves adding three
components together to reach lheASIechem, Eq. 3.12. The first two factors include impedance that
arises from the interfacial charge transfer and transport. The third factor is a lumped parameter
used to capture the remaining impedance.
" + ASI^f + ASIcwst (3.12)
The interfacial impedance for positive and negative electrodes both contain the charge transfer
resistance component RTI(i0aLF) as shown in Eq. 3.13 and 3.14. Here, i0 is the exchange current
density related to the interfacial area and a is the ratio of interfacial area to electrode volume. An
approximation often used for a relates the parameter to the volume fraction of the active material
and the particle radius, a = 3sact/rp. The variables i0 and a should be specified to relate to the
same area as they are often not independently determined. R and T correspond to the universal
gas constant and absolute temperature respectively. F is Faraday's constant. The influence of the
interfacial impedance is that the ASIechem increases as the electrode thickness is reduced. This
behavior is typically observed at electrode thicknesses less than 30 microns for common Li-ion
battery materials.
= ~^~: (3-13)
-0.5
I /- \ /" \ ^ I
RT
1 —
T ionu
Mim
(3.14)
The positive electrode interfacial impedance also includes two factors that account for the
physical limitations that occur from depleting the concentration of the reactants within the
porous electrode. The I{™'€ term is the limiting ionic current for lithium cation transport through
the porous separator. The rc lim term is the limiting C-rate for solid state diffusion of lithium in
the active materials. The C-rate may be related to the current density with Eq. 3.15. Here, the
specific capacity, Q, the active material density, p, active material volume fraction, eact, and the
electrode thickness, L, are used. If either the limiting C-rate or limiting ionic current are
approached, the AST will begin to approach an infinite value. The parameters required for the
AST expression are fit to experimental measurements. The AST values are corrected for the
interfacial contributions present during measurement so that the correct AST may be determined
at different electrode thicknesses.
16
-------
I = rcQp£aaLpos (3.15)
The cell AST for energy, ASIenergy, and power, ASIpower, are determined by adding the ASIechem to
that of the current collectors, ASICC, as discussed in the next subsection. The difference between
ASIenergy and ASIpower is that the limiting currents are not important during the C/3 discharge for
energy and the ASIconst is a different value for two cases. ASIenergy will always be higher than
ASIpOWer if a battery is operated far from the limiting current. The higher impedance is due to the
formation of significant concentration polarizations during the longer time scale of the energy
discharge. A reasonable rule-of-thumb is that the ASIconst for energy is 2.2x the value for power
in layered oxide materials such as
3.4.1 Current Collection Resistance
The resistance from the conductors used to collect the current must be accounted for as they can
contribute significant ohmic drop to the battery. The AST used to calculate the required cell
separator area, ASIpower, is larger than the AST for the electrochemical charge and discharge
processes, ASIechem,p, as shown in Equation 3.16. The ASIechem value is typically measured from
experiments and must be added to the external resistances that arise from the materials used to
conduct the electric current. These resistances come from current collection in the cell and also
those on the module and battery pack level.
ASIpower - ASI h p + ASICC
Ncells
The current collector foil impedance, ASICC, is determined from an analytical expression, Eq.
3.17, which accounts for the coated and uncoated region of the foil, labeled act for active and tab
respectively. The resistance factor, R/, and the resistance of the current collector foils, Rcc, are
also shown for clarity in Eq 3.18 and 3.19. The factor of 2 in the ^/term is due to assuming half
of the foil thickness carries the current produced on one side of the foil. While all of the current
passes through the tab region, the magnitude of the current varies along the height of the coated
foil as the reaction area continually contributes current to the foil. An equivalent length for the
resistance calculation may be determined so that multiplication by the total current for a cell will
give the correct ohmic drop. This equivalent length is H/3 if the current density is relatively
constant over the entire area. The derivation of this equivalent length as well as an in-depth
discussion of the voltage and current distribution in the foils may be found in subsection 3.4.2.
Also in the later subsection, the assumption of constant current density is verified with numerical
modeling.
(3.17)
Rf=\ 7 + 7 1 (3-18)
V foil,neg foil,neg foil, po s foil, pos
17
-------
(3.19)
The cell terminals are ultrasonically welded to the ends of the current collector foil tabs. While
the welding removes this contact resistance, the AST of the terminal must be included in the total
cell resistance. The AST of the cell terminals, ASI^ , is the summation of the positive and
negative cell terminals as shown in Eq 3.20. The dimensions for these terminals are set by the
calculated width of the cell and the user defined terminal thickness and height.
'- Apos (3.20)
term,neg term,pos j term term
The AST for connection losses is the last term in the AST summation stated in Eq. 3.16. This AST
value is calculated by multiplying the ratio of cell positive electrode area to number of cells by
the summation of the resistances, ^cnct, for cell terminals, module terminals, module
interconnects, and batteries terminals. In this way, each cell shares in the burden of overcoming
the system losses from carrying the electric current. The calculation of ^Cnct is detailed in Eq.
3.21 with the individual sources of connection losses shown. The voltage drop resulting from
cell-to-cell contact resistance, Rccen"a , is taken to be 10~4£/OCv,E in Eq. 3.22, a small fraction of the
open-circuit voltage. A battery manufacturer would only tolerate a minimal voltage drop from
cell-to-cell contact. One connection method is to physically press the two cell terminals together.
This resistance could be lowered by increasing the physical pressure and contact area, or by laser
welding the terminals together. Regardless, the value used in the model is left to the choice of the
user to leave as is or to change to a different value.
d( + R^n (3-21)
lfa = 1Q-4 N**U™.B (3.22)
The module terminal resistance, R™^, calculation in Eq. 3.23 is shown as an example of how the
terminal and interconnect resistances are calculated for the module and battery pack. The size of
the terminals and thus their resistance are determined from a calculation based on a pre-
determined allowable rate of temperature rise for the conductor. This approach is explained in
more detail in subsection 3.4.3.
ff
nmod term »rmod /o OQ\
-"term = 7 ^ Nterm (3-23)
18
-------
3.4.2 Potential and Current Distribution in the Current Collection Foils
The designed current collection system was evaluated using a numerical simulation package.
Equations 3.24-3.26 were solved for a steady state, isothermal, and 1-D simulation. Here, the
conductivity, 03, is the effective conductivity of Vi of the foil (the other half carries the current
from the opposite side). The bulk conductivity value, o]0, is multiplied by the thickness of the
conductor, L/2, to lower the dimension of transport.
)-gW(*))
\ • )
(3.25)
(3.26)
The boundary conditions were set for both ends of each foil. The tab ends of the foils were set to
a specified voltage and the opposite ends of the foils were restricted to a no flux condition. The
simulation was performed using the foils defined in our battery design: 12 micron thick copper
foil and 20 micron thick aluminum. The cell length was 20 cm, the ASIechem was 30 ohm cm2, and
the Uocv and Vceii were set to 3.72 and 3.57 V respectively. Figure 3.5 shows the current and
potential distribution in the foils and in the cell resulting from the simulation. The cell potential
along the length of the foil varies only by 1.5 mV from maximum to minimum difference. The
0.4 % variation in voltage results in a 0.9 % variation in current density. This verifies the current
density is uniform along the length of the foil. This is also obvious from the linear relationship of
current with foil height in Fig. 3.5. The assumption of constant current density was tested in cell
heights up to 100 cm and found to be satisfactory. The assumption should be reasonable as long
as the ASIechem is at least twice the value of ASICC. The simulated resistance of the foils is found
to raise the ASIechem by 0.7 for an ASIpower of 30.7 ohm cm2. Additionally, the numerical result
verified that H/3 is the correct equivalent length to represent the ASICC for the cell. This may also
be found analytically, Eq. 3.27-3.29, if you assume an even current distribution as we have
shown is a reasonable assumption.
(3.27)
I x2 }
*!.»»=— T-rf« (3-28)
°ne& 12 }
^, -4>, y 772f i i
1 _;£ '-Ldx = -z*L + + | (3.29)
ln 73 I (Jpos (Jneg
19
-------
3.5750
„ 3.5740
.2
"^
c
0)
+•>
o
Q.
"55
o
c
(0
3.5730
3.5720
O 3.5710
0)
"^
'(75
o
Q.
o
o>
Q.
0)
i_
O
3.5700
3.5690
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0.0000
•Cell
•Positive foil
•Negative foil
0.2
0.4
0.6
0.8
x/H
a>
-0.0040 '•*=
O)
0)
-0.0050
-0.0060
0.00492
Negative foil
Positive foil
current density
0.00486
Figure 3.5 The change in current and potential within the positive and negative foils. The current
collection design results in a uniform current distribution along the length of the foil.
20
-------
An analogous problem has been solved by Euler and Nonnemacher and then communicated
repeatedly by Newman et al.21' 22 The analytical solution they presented may be used after a
slight alteration to dimensionalize the current density to the geometry of our concern, Eq. 3.30
and 3.31. This solution was reached assuming linear polarization behavior and is valid for cases
where the current density varies along the height of the current collector foil. Thus, this approach
is a more general solution than the one we use in the design model.
AST
echem
+ ASI =•
ltpos
H
1 + -
pos
- + -
neg
v sinh v
(3.30)
v = -
H2
1
ASIechem a
pos
(3.31)
3.4.3 Determination of Module Terminal, Battery Terminal, and Module Inter-connect Size
An important factor for setting the resistances of a module terminal, battery terminal, or module
interconnect is the allowable rate of temperature rise in the conductor at full power. We set the
acceptable rate of temperature rise, dT/dt, at 0.2 °C/sec or a 2 °C rise for a 10-sec power burst
under adiabatic conditions. The heating rate, q, is then used to determine the mass, m, of the
terminal required for the designed battery in Eq. 3.32. Since the heating rate may also be
determined by Eq. 3.33, we may determine the cylindrical terminal radius and mass by assuming
a length, //term- In this way, the size of the module terminal is redesigned during each simulation
to meet the specified power requirements and allowable temperature rise, Eq. 3.34. The mass of
the conductor is found to be inversely proportional to the allowable temperature rise.
r dT
•• mC „
" dt
(3.32)
H,
-KJ2
(3.33)
\errn ~
* term term p 7,
dT_
dt
-1/2
(3.34)
3.5 Calculation of Battery Dimensions
The goal of the model is to quantify how the various components of a specific battery design
sum to make the mass and volume of the battery pack. In this way, a true energy and power
density can be calculated as well as the exact materials requirement to meet this design.
21
-------
Summing the mass of the components is relatively straight forward. Determining the total
volume that contains the components and required free volume is not as obvious. The exact
calculations used in the design model are detailed below for the cell, module, and battery pack.
3.5.1 Cell Dimensions
The number of layers in each cell is approximated in Eq. 3.35 by accounting for the compression
factor, .Xcomp, and the individual thicknesses of the current collector foils, Lf0u, electrodes, Lpos
and Lneg, separator, Lsep, and container, Lcont. Xcomp is usually taken to be 0.97. The Li-ion battery
chemistries this model was designed for are assumed to undergo negligible volume change on
the cell level. No effort was made to address possible changes in electrode volume upon cell
discharge or charge.
\ • I
layers comp mg
mg pos ~, .
foil "•" ^foil """ ^\^sep "•" ^neg """ ^ pos >
The N layers approximation is necessary as the cell thickness is a user defined parameter. The
aspect ratio of the cell is also user defined; therefore, solving for the width also determines the
height of the cell as seen in Eq. 3.36. The width is calculated from the number of layers and the
aspect ratio, H/W. The factor of 2 enters the denominator as both sides of the foil are assumed to
be coated.
(3-36)
Having determined the width and height of the electrode, the rest of the cell dimensions are
relatively straightforward, Eq. 3.37 and 3.38. The width of the cell, Wceii, is 8 mm wider than the
positive electrode to allow for the larger separator area and pouch seals. The pouch seals are
folded up, pressing along the inside wall of the module casing. The height of the cell, Hceii, is the
height of the positive electrode in addition to the distance for the terminals and connections to
the foil tab, LtermyCnt. Our assumed design requires 15 mm for this distance. The volume of the cell
is the product of the three dimensions.
WaU=Wpn+S (3.37)
Hcell = H pos + ^Lterm,cnt (3.38)
3.5.2 Module Dimensions
The module dimensions are defined by Eq. 3.39-3.41. The height and length of the module are
both just 2 mm wider than the cell dimension. The width of the module is related to the total
thickness from all of the cells with allowance for a SOC controller at one end.
mod
= Wcell + 2 (3.39)
22
-------
Lmod = Hcdl + 2 (3.40)
Wmod=4e/;(A^/mod+l) + l (3.41)
3.5.3 Battery Pack Dimensions
The battery pack volume includes all of the modules, spacing for connections between modules,
channel for the cooling air to flow, Hair, thickness of the module compression plates, Lcomp, and
the battery pack jacket, Ljack (Eq. 3.42-3.44). Ljack includes a 10 mm thick insulation layer
sandwiched between two aluminum walls for the container. The thickness of the aluminum wall
increases from 1 to 1.5 to 2 mm as the battery volume increases from < 20 L to < 40 L to larger
dimensions. The layout of the modules, number per row, Nmod/r0w, and number of rows, Nrow, is
also included. The final volume of the battery is the product of the three dimensions. The space
left for connections between modules, Lgap, is a function of the number of rows of modules. Lgap
is equal to 8, 10, or 20 depending if there is one, two, or four rows of modules. Three rows of
modules are not allowed as the positive and negative terminal for the battery would be on
opposite ends and thus not very practical. A number greater than four rows of modules is deemed
unnecessary.
Hhatt = Hmod + 2Halr + 2Ljack (3.42)
^batt = N mod/ row "mod + "air + ^^comp + ^ jack (3.43)
Wbatt=NrowLmod+Lgap+2Ljack (3.44)
3.6 Additional Considerations
A few situations may arise that require a change in the calculation method. These situations are
addressed in the subsections below. The inclusions of these calculations into the model allow for
a more realistic depiction of limitations often encountered by cell manufactures.
3.6.1 Maximum Electrode Thickness
A practical limitation exists for the maximum achievable electrode thickness. This limitation
may be set by manufacturing capabilities, ionic and electronic current transport within the porous
electrode, or aging characteristics related to adhesion to the current collector. When the
maximum electrode thickness, LmaX, has been reached on either the positive or negative electrode,
the electrode area equation is modified as shown in Eq 3.45. The electrode thickness, Ltgt, is the
largest electrode thickness, negative or positive, calculated at the targeted fraction of the OCV
[V/U].
ALr=SLAos (3.45)
23
-------
The area of the electrode is now determined by the cell capacity requirement to meet the battery
energy demands and not the target voltage at maximum power. As a consequence, the battery
pack will operate at a higher [V/U] than originally selected by the battery designer. The new
[V/U] may then be calculated from Eq. 3.46 which is the solution to the quadratic found in Eq.
3.47.
V_
~u
2C7,
P A <\7
TT UTT \2 A batt u power
UceU+J(UceU) -4—
(3.46)
••ucell-
batt
power
cell cells pgs
(3.47)
The maximum electrode thickness has a large impact on the energy density and cost of cells
designed for high energy and range. Nelson et al. demonstrated this concept in 2009 assuming a
100 micron maximum electrode thickness.11'17 In 2010, Santini et al. relaxed this assumption to
300 micons; although, the thickest electrode discussed in the paper was a 225 micron graphite
electrode in the LMO-Gr EV with 100 mile range.12 In conversations with manufactures, 100
microns appears to be the general electrode thickness used for EV type cells at the present time.
However, Santini et al. has shown substantial increases in energy density and decreases in cost if
larger electrode thicknesses may be utilized. The challenges to achieving thick electrodes, in
addition to those already mentioned, relate to fast charging while avoiding lithium metal
deposition, removing gases formed during formation cycling, wetting the full porosity of the
electrode, achieving defect free coatings, and drying the thick electrode at high rates. Our
opinion is that the successful cell manufacturers will engineer ways to overcome these
challenges to increase energy density and lower cost.
3.6.2 Accounting for Parallel Cell Arrangements
The user of the design model may wish to use a parallel arrangement of cells within the larger
series arrangement of the battery pack. Several motivations exist for a parallel cell arrangement.
For example, a battery supplier may wish to only produce cells of a specific capacity. The
manufacturer may only have the equipment to produce a certain size cell or they may encounter
engineering design problems for very large cells ( > 60 Ah). Thus, a cell group composed of
parallel connected cells may be necessary to meet the energy requirements while staying within
battery pack voltage and current requirements.
When the user chooses to have cells connected in parallel, the design model calculation includes
the appropriate factors necessary to account for changes in the resistance, volume, and mass of
the battery. The end result is a lower energy density for the battery from including the cell group
interconnects and additional inactive material for each cell. Furthermore, the thickness of the
cells is reduced to 6 mm to better suit the smaller cell format.
24
-------
4. Modeling of Battery Pack Manufacturing Cost
4.1 Approach
The manufactured cost of a battery pack is calculated with input from the design information
generated in modeling the cell and battery pack performance. The design modeling determines
the annual materials and purchased items requirements. The manufacturing cost is then added to
these materials costs, along with a warranty cost, to reach the unit cost of a single battery pack.
The manufacturing costs for the designed battery are scaled from a baseline plant. The baseline
plant was designed for a battery of intermediate size and production scale so as to establish a
center-point for other designs. The baseline plant accounts for the size, speed, number of units,
direct labor, and depreciation of the capital cost for each processing step. These costs are
adjusted to meet the requirements for a plant producing the battery under study. The process
expenses are summed with the additional costs of operating the manufacturing facility. These
costs include launch costs, working capital, variable overhead, general, sales, administration
(GSA), research and development, depreciation, and profit.
In this analysis, all costs are evaluated for 2020 when large battery manufacturing plants are
built. All dollar values are brought back to 2010 with allowance for inflation. Some materials
and battery manufacturing costs are lower than recent values, where we judged that processing
improvements for high volume production of materials would reduce costs.
The baseline manufacturing plant was calculated for an annual production rate of 100,000
batteries. The cost model accounts for different scales of manufacture by recalculating the costs
of each individual step in the manufacturing process. The changes in capital and operating costs
will change the calculated unit cost of the battery pack. However, the parameters were
determined to provide reasonable estimates for manufacturing rates of 20-500 % of the baseline
rate. Thus, for a plant that is far different in size from the baseline plant, for instance a pilot plant
having an annual production of only 5,000 battery packs per year, the estimate from this study
would be expected to be less accurate than if determined in a study dedicated to that purpose.
To simplify the cost calculations, it was assumed that all hardware items for the cells, modules
and battery will be purchased from a vendor specializing in similar products. The costs for these
items were estimated to be a fixed value plus an additional value proportional to the weight of
the item, which is calculated during the battery design. In mature manufacturing plants in 2020,
toward which this study is directed, some items which are assumed to be purchased in this study
might actually be internally manufactured from raw materials. This would increase the number
of processing steps needed in our manufacturing simulation and thus complicate the cost
calculations. Assuming that some parts would be purchased if they would actually be produced
from raw materials would tend to underestimate capital and labor costs and overestimate
purchased items expenses. However, the net effect would be a very small change to the overall
unit cost of the battery pack.
25
-------
4.2 Materials Costs and Purchased Items
The end battery pack cost depends significantly on the cost of both the active and inactive
materials that compose the design. In this subsection, the assumed material costs and the rational
behind them are presented. While we state suggested materials costs, the user of the cost model
may enter any value that they desire.
4.2.1 Battery Specific Materials Cost
The largest contributions to the materials cost of the battery are from the following components:
positive and negative electrode active material, separator, electrolyte, and current collector foils.
The choice of the materials often defines the size and performance of the battery as well as the
cost. Many different variations of materials are possible in the Li-ion family of chemistries.
However, we have chosen to focus on the different available positive electrode materials with
less attention on the negative electrode. This reflects the current research and manufacturing
activities. The separator and the electrolyte are also both active areas of development. However,
the following battery designs are based on a single electrolyte and separator combination.
Including the cost and effect of additives and enhanced separators is beyond the scope of this
work. The user is always able to modify the dimensions, cost, and ASI that may be required to
account for changes in these materials.
The price of specific battery materials is of some debate. The values presented in Table 4.1
compare our suggested costs to those reported recently in the open literature. Our values, as well
as the others in the table, are derived from conversations with material, cell, and original
equipment manufacturers. The sources are commonly anonymous and the accuracy of the values
is generally unknown. We present the comparison of published values so that the user of the cost
model may appreciate the accepted range of values for commonly used materials.
4.2.1.2 Positive Electrode Active Materials
The cost of positive electrode materials is driven to a large extent by the cost of the raw materials
from which it is made. The archetype Li-ion positive electrode material, lithium cobalt oxide
(LCO), was the original material commercialized in Li-ion batteries for consumer electronics.
LCO has many excellent characteristics but is not considered a viable choice for use in Li-ion
batteries for automotive applications. One of the largest drawbacks of LCO, other than safety
concerns, is the cost of the cobalt. While tolerable in the consumer electronics market, the cost is
too high for use in an automobile battery. Many other materials are in a commercially viable
state of development and are currently utilized in Li-ion batteries produced today (Table 4.1).3'6
The relative advantages and disadvantages of each material will not be discussed here.
The amount of cobalt and nickel, as well as ease of manufacture, controls the end price for a
positive electrode material. For example, the NMC-441 is less expensive than the NMC-333 as
the cobalt quantity is significantly reduced. The market price for cobalt and nickel metal varies
dramatically from year to year. Reducing the quantities of these materials will reduce the
26
-------
o.
(U
05
OS
S3
O O in
(N en (N
o
en
HH O
H ^
£ ^
8 S
en
en
cs
(N
•x
cs cs
(N
0
en
(N
oo
en C>
_o
%-^
en ^H
en -xt-
en -xt-
^H
o
"S
1
1-1
(^
u
N
4
S
(5
1
X
o
JZ
c3
O
•73
(•H
^^
C^
'"^
o
•a
1
=«
0
o
y
J
-------
positive electrode material price and price volatility. Researchers at TIAX LLC have treated this
variation and shown the significant effect on end battery cost.10 The average metal prices for the
last 20 years is 48 $/kg and 15 $/kg for cobalt and nickel respectively. These numbers are based
on historical prices for the metals as collected by the United States Geological Survey (USGS).23
The metal prices are best used as indicators for how the intercalation material cost will relate
when compared to one another. The fact these materials are not earth abundant means they will
not benefit as much as other materials from increased scales of production.
In general, earth abundant elements should be the dominate transition metal used if a low
positive electrode cost is desired. Both iron and manganese are abundant and inexpensive
transition metals for intercalation materials. Comparison of the iron phosphate, LFP, to
manganese spinel, LMO, reveals how processing costs contribute to the end price of a material.
LMO is relatively easy to manufacture. In contrast, LFP requires a reducing atmosphere and a
carbon coating step to reach the end product. The increased complexity in the manufacturing
process is realized in the price. However, one could argue that the manufacturing cost will
decrease with increased knowledge from larger scales of production.
4.2.1.2 Negative Electrode Active Materials
While multiple negative electrode materials exist for Li-ion batteries, carbon materials in the
form of graphite and/or hard carbon are still used in the vast majority of commercial cells.
Graphite offers the greatest energy density while hard carbon is said to enable high rate
capability with decreased risk of lithium plating (an undesired side reaction) during high charge
rates. We have chosen synthetic graphite as a generic carbon electrode in our model. The price of
graphite is much better understood than that of most of the positive electrode materials.
However, significant differences in cost and performance will exist between synthetic, natural,
and coated-natural graphite. The method of production and necessary heat-treatment will control
the end cost. Graphite, although in different purity grades or micro structure forms, is used in
many industries. This is in stark contrast to the positive electrode materials.
The lithium titanate electrode, LTO, offers an interesting option compared to graphite. Unlike
graphite, LTO operates within the stability window of the electrolyte. The higher electrode
potential dramatically reduces or eliminates the formation of the solid electrolyte interphase
(SEI). As a result, nanoparticle-based LTO may be implemented without concerns of increased
side reactions with the electrolyte. The reduced dimensions increase the available surface area
for reaction while simultaneously shortening the diffusion length. Both of these factors combined
with the lack of SEI dramatically reduce the impedance of the electrode.
4.2.1.3 Electrolyte and Separator
The electrolyte used in this model is based on a lithium hexafluorophosphate salt, LiPFe,
dissolved in a carbonate based solvent system. The carbonate solvent system is a blend of
ethylene carbonate, EC, and a linear carbonate such as ethyl methyl carbonate, EMC, or dimethyl
carbonate, DMC. Other chemical additives may be used to lower the capacity and power fade of
the battery over time. Polymers may be added to the electrolyte as either a minor or major
28
-------
component. This is not discussed in any further detail in this work. The price of $16/L, about
$19/kg, is only for the base electrolyte (i.e. no additional additives).
The separator is typically a porous membrane based on polypropylene (PP) and sometimes
includes a polyethylene (PE) middle layer. PP and PE are very inexpensive raw materials and
thus the suggested cost of $2/m2 is in large part due to the manufacturing process required to
form the porous network in the membrane. As competition and scale of manufacture increase,
the prices of the separator may fall closer to $l/m . However, the cost of improved technology
may offset some of this cost reduction, so we have retained our cost estimate of $2/m .
As safety is a major concern for Li-ion batteries, the separator plays a key role in isolating the
oxidant from the fuel. If the two charged electrodes contact each other (short), then a run-away
reaction is possible. Separators have been designed to "shut-down" or melt at key temperatures.
The middle PE layer is the shut-down feature in our proposed separator. Ceramic coatings have
also been used to ensure structural integrity. Many other approaches are being developed to
increase the safety of Li-ion batteries. The user of the cost model should account for the
increased technology in the price and dimensions of the separator as needed.
4.2.1.3 Current Collector Foils
The current collector foils are based on copper metal for the negative electrode and aluminum for
the positive electrode. However, the LTO anode material, because of its high voltage relative to
lithium, enables the use of aluminum as the negative electrode current collector. The price of
these foils is based on raw materials and manufacturing costs. The foils are 12 microns and 20
microns thick for the copper and aluminum current collectors respectively. The foils used in
batteries have additional requirements beyond the cheapest product available. Surface treatments
are often necessary to promote adhesion of the electrode to the foil surface. In addition, alloying
of the foil may be necessary to achieve the required material properties for long life.
The raw material contributions to the foil price will vary with the volatility of the market price
for the metals. Figure 4.1 displays the metal ingot price contribution on a $/m2 basis. These
numbers are based on historical prices for the metals as collected by the USGS.23
The values for both aluminum and copper tend to vary significantly over the time period
examined. The price for copper is more volatile and always more expensive than aluminum.
Analysis of Figure 4.1 reminds the user of the cost model that cost quotes are only valid for a
short period. As the market price for raw materials changes, so will the price for the finished
product.
r\
Conversations with manufacturers and suppliers lead us to take a price of 3.00 and 0.80 $/m for
battery grade copper and aluminum foil respectively. We point out that the current metal ingot
price is only a small contribution to the end foil price being about 15 % of the aluminum foil
price and 25 % of the copper foil price. Thus, a doubling of the ingot prices would only
moderately increase the foil prices. The aluminum foil is produced by rolling of thicker stock
foils into thinner and thinner sheets. On the other hand, copper foil is more likely to be produced
through an electrodeposition process.
29
-------
-------
a fixed amount plus an additional factor, which is proportional to their mass, mi. The cell
negative terminal and parallel cell group connection are both made from nickel plated copper
sheet and thus have the same cost equation.
Table 4.2 Cost equations for purchased items
Component, /
SOC controller
Cell positive terminal
Cell negative terminal
Cell container
Parallel cell group connection
Module terminals
Balance of module (casing)
Battery enclosure & cooling
Cost Equation, $/unit
1.80 + 0.005C
0.10 + 4mr-
0.10 + 5m,-
0.10 + 5m,-
0.10 + 5mr-
0.75 + 4nn
1.00 + 3m,-
50.00 + 7m,-
Cost per unit
cell or parallel cell group
cell
cell
cell
parallel cell group
module
module
battery pack
4.3 Baseline Manufacturing Plant
The baseline plant is designed to produce 100,000 NCA-Gr baseline battery packs per year. The
baseline battery pack produced by the plant has sixty, 40-Ah capacity cells, providing a total
pack power of 50 kW and total energy of 8.7 kWh. The battery will power 24.4 miles of vehicle
travel at 70% of the pack energy and 250 Wh/mile. The schematic diagram of the plant (Fig. 4.2)
is designed to illustrate the flow of materials through the plant and the relative floor areas for the
processing steps rather than representing a realistic plant layout. The overall manufacturing rate
of 100,000 battery packs per year is achieved by operating for three shifts at the equivalent of
300 days per year of fully effective production. There will be more than 300 days of operation,
but some days will have less than 100% effectiveness. The exceptions to three-shift operation are
the Receiving and Shipping sections, which are active for only two shifts per day. The cost
factors for the individual manufacturing steps in the baseline plant are summarized in Table 4.3
and discussed in detail in the sections that follow. Most of the operations are carried out with
normal factory atmosphere, but the cell assembly process steps are completed in a dry room
atmosphere.
31
-------
Receiving
Electrode Materials
Preparation
> Positive
Negative
Electrode
Coating
Positive itz>
Negative
Solvent r.v.--1.-"-- Solvent
Evaporation
Positive i
Negative
Recovery
Calendering
Electrode
Slitting
Vacuum
Drying
Shipping
Cell and
Scrap
Recycling
Battery Pack
Assembly
and Testing
Control
Laboratory
Module
Assembly
Charge-
Retention
Testing
Formation
Cycling
Final
Cell
Sealing
Air
Locks
Materials
Handling
Cell Stacking
Enclosing
Cell in
Container
Current
Collector
Welding
Electrolyte Riling
and Cell Closing
Assembly Route
Dry Room
Outdoor dry-room air
processing equipment
The areas in this diagram for each processing step are approximately proportional to
the estimated plant areas in the baseline plant.
Figure 4.2 Baseline lithium-ion battery manufacturing plant schematic diagram. Manufacturing
rate: 100,000 NCA-Gr battery packs per year, 50-kW pack power, 40-Ah capacity, 60 cells per
battery. Operating year: 300 days with three 8-h shifts (two shifts for receiving and shipping)
4.3.1 Receiving and Shipping
These operations require the moving equipment and storage facilities common to any such
factory facilities. The Receiving section handles slightly less than 6,000,000 kg of materials per
year and also has facilities to handle and store some of the electrode materials in a dry
atmosphere. The Shipping section is required to enclose the battery packs in crates, which
requires some automated equipment and more labor than is required for Receiving. Shipping also
handles about 400,000 kg of scrap each year, which is broken down and prepared for shipping in
the Rejected Cell and Scrap Recycle section. The estimated resources needed for the Receiving
and Shipping sections are shown in the table below.
Receiving
Off-loading
Moving
Storage
Shipping
Rate Factor
870,000 kWh/y
870,000 kWh/y
Direct Labor
3 per shift
6 per shift
Capital Equip.*
3.60 mil$ total
0.60
1.20
1.80
5.0 mil$ total
Plant Area, m2
600
600
*Total cost including installation
32
-------
re S ^
£ £ °
< Q.
^
C CM
1 E §
Q. CD
CD CD
O O
O O
O O
00 00 CD
OOO
OOO
o o in
in in -i-
CD
o
o
T—
CD
O
O
in
T—
CD
O
O
O
CD
CD CD CD
OOO
OOO
OOO
CM CM CM
OOOOCDCD -sJ-OOCDCDCDCDCDCD
OOOOOOOOOOOO CM e/)
•;= 5_ o
'^ 0 ^_
FT CO o
OOOOLOOOOOOOOLO m DC •*-
OOOOts-OOOOOOOCM £ -- ^ j
1— " o" CD £ ci
CD 00 CO
3=! d °
CO <- O
0)
E
Q
'B
LLJ
d
re
O
O CD
re d
LL
Q.
o
CD
CM CM
s, s,w-
O
00 00
d d
[s.
d
- c c c
o o
o o
o
o
o
o
o
o
o
o
o
o
o
in
o
o
o
CD
o
in
OOOO OOOLOOOOO
OOOO OOLOts-OOLOO
CM CO
in
en
CO
co° o
DC O
O
CO
d
T3
O
CO
o
Q)O
CO
CO
_Q
CO CD O
3 & °
. o co"
O
1o
O
£
re
o
re
LL
o
I
in in
d d
in in •*
odd
in in is. in in in
d d d d d d
rs.ts.LOLO •^Is.Ln^LOLnis.LO
dddd dddddddd
o o
o o
CD CD
-,_ -,_
CM CM
O
O
00
00
CM
O
O
00
00
CM
O
O
•*
T
O
o
CM
fs.
O
O
CM
fs.
O
O
00
00
CM
O
O
00
00
CM
O
O
•*
•sj-
T
O
o
00
00
CM
OOOO OOOOOOOO
OOOO OOOOOOOO
CMCMOOCM •sJ-CD-sJ-CDCMCMOOO
COCOOOCO •=!- [s. Tj- T— COCOCDOO
^J"^J"CM^J" T — LO i — CM ^J" ^J" CO CM
O
O
•*
CM
CD
•* CM -1-1-
CDCD-!l-CDCMOOC\lcr)CDCDLOCD
^
O
<
-0 2
£ CD
"o 'S
C "~
O CO
O CD
. O
0 ffl
£:£
T3
CD
CO
i'-g
&•§
o >-
C jz"
° <
£v
CD
ti
(U
CD
1
re
OQ
"re
CO CO
'i_ 'i_
CD CD
"CO "CO
E E
CD CD
O O
CO CO
D) D)
_^ _^
CO CO
CD CD
O O
D)
«
CO CO CO
CO CO CO
CD CD CD
OOO
CM in
to en
CM 00
CO
00
en
oo"
in -i-
CO CM
00 O
en en
in in
CO CO
00 00
en en
oo" oo"
in in
CO CO
00 00
en en
oo" oo"
in o
CO CM
en en
oo"
en en
00 00
in in
CD" CD"
en en
00 00
in in
CD" CD"
o en
o oo
CM in
CD"
en en
00 00
in in
CD" CD"
o o
o o
o o
CD"
en o
00 CM
in en
CD"
0
Q.
CO
CD
CD
"o
g
"CO
1
E
_0
T3
rrl
Q.
CO
o^
_0
CO
^
modules
.!= D)
=0 .E
| 1
o
o
CD
CD ±±
p =i
.i-E
^ CO
CO T3
'F O
O
—
fl
ca
H
CO
CD
Q_
O
!E
CO
6
D)
c
'>
'CD
o
CD
DC
Materials preparation
Positive electrode
Negative electrode
Electrode coating
Positive electrode
Negative electrode
Solvent recovery
Calendering
Positive electrode
Negative electrode
Materials handling*
a,
!
co
gra
2 ° "
T3 O O
« ± 2
o
«
i
E
S«
LU > O O
coJ5
S
8
33
-------
4.3.2 Electrode Materials Preparation and Delivery to Coating
The electrode materials, which consist of active material, carbon (if necessary), binder and
binder solvent, are well mixed in small batches in portable tanks. At the design production rate
in the baseline plant, each shift requires three tanks each holding about 1000 liters of positive
electrode material mix and three tanks each holding about 900 liters of negative electrode
material mix. The section must be capable of exceeding this design rate of production by at least
25% to catch up in case of unscheduled downtime in Materials Preparation or in some of the
section immediately following that section. The tanks of prepared materials are moved to the
Coating section and pressurized to push the coating paste into the coating mechanism. The
estimated resources needed are the following:
Materials Prep.
Positive
Materials
Storage tanks
Mixing tanks
Moving equip.
Negative
Materials
Rate Factor
1,7 10,000 kg/y
active material
1,2 10,000 kg/y
active material
Direct Labor
3 per shift
3 per shift
Capital Equip.*
4.0 mil$ total
2.00 mil$
1.00
1.00
4.0 mil$ total
Plant Area, m
400
400
*Total cost including installation
4.3.3 Electrode Coating on Current-Collector Foil
The positive and negative electrode structures are formed by coating both sides of the current
collector foil. In the baseline plant, the coating lines are 1.5 meter wide continuous roll-to-roll
coating processes carried out at a line speed of 10 m/min. The first set of coating and drying
stations coats one side of the current collector foil, drives off the solvent in a heated oven, and
turns the foil over while transferring it to a second set of stations. The second set of coating and
drying stations applies and dries the remaining coating before the coated foil is wound into a
large roll at the end of the line. The negative and positive coating lines are very similar.
However, some of the negative material is coated only on one side to provide the electrodes at
the end of the cell stacks. For the baseline plant, a total of 8,170,000 m2/y of coating (annual cell
area) is required for the positive electrode (slightly more for the negative electrode), which
allows for the 5% loss of cells expected to fail testing and inspection. A larger area of foil than
the coated area must be fed to the coaters to allow for the part of the foil that is not coated so as
to provide tabs for welding to the terminals (about 10%) and to allow for trimming losses during
electrode slitting (8%). Also, about 30% excess coating capacity must be provided to allow for
unscheduled downtime. Only one coating line is needed for each electrode type to meet these
needs. If one coating line breaks down, the other coating line may change over temporarily to
coat the other electrode material.
The oven sections of the coating line are designed to dry coatings about 100 microns thick at the
coating speed of 10 m/min. A thicker coating will require longer ovens at additional capital cost
which is provided in the adjustment of costs discussed in section 4.4. For the same annual area
throughput, a coating line that coats both sides with a 300-micron coating would cost $7,500,000
34
-------
rather than the $6,000,000 cost for the 100-micron coater. The binder solvent for the positive
electrode in the baseline plant is NMP, which must be recovered by condensation and recycled.
About 0.5% of the binder solvent is combusted with a thermal oxidizer and must be replaced. For
the negative electrode the binder is water, which need not be recovered. The estimated resources
to meet these needs are the following:
Electrode Coating
Positive Electrode
Uncoated area
Width of coater
Coating speed
Number of coaters
Maximum rate
Excess capacity
Negative Electrode
Solvent Recovery &
Oxidation
Rate Factor
8,170,000
r\
m /y cell area
8,170,000
m2/y cell area
1,527,000 kg
NMP/y
Direct Labor
4 per shift
4 per shift
2 per shift
Capital Equip.*
6.0 mil$ total
18%
1.5m
10 m/min
One
13,000,000 m2/y
30%
6.0 mil$ total
3.0 mil$ total
Plant Area, m2
500
500
150
*Total cost including installation
4.3.4 Calendering
The materials leaving the coating lines may be stored on large rolls (see next section). However,
typically the materials leaving the coaters would go directly to the calendering process in which
the coatings are compressed by rolling to meet the specified void volume fraction, which will
later be filled with electrolyte. The calendering equipment must match the output of the coating
equipment producing 8,170,000 m2/y of cell area with a maximum rate of 13,000,000 m2 of foil
per year to meet contingencies as in coating. The estimated resources to meet these needs are the
following:
Calendering
Positive Electrode
Negative Electrode
Rate Factor
8,170,000
r\
m /y cell area
8,170,000
m2/y cell area
Direct Labor
1 per shift
1 per shift
Capital Equip.*
1.0mil$ total
1.0mil$ total
Plant Area, m2
150
150
*Total cost including installation
4.3.5 Inter-Process Materials Handling
For all processes (Fig. 4.2), work in progress must be transported and occasionally stored to
permit nearly-continuous operation of the equipment. Storage areas must be provided both inside
and outside of the dry room. Raw materials must also be moved to the processing sites, which for
those in the dry room means through a separate air lock for materials transfer. One-third of the
35
-------
total space for Inter-Process Materials Handling is within the dry-room for the baseline plant and
also for all other plants. The estimated resources to meet these needs are the following:
Materials Handling
Rate Factor
8,170,000
m2/y cell area
Direct Labor
4 per shift
Capital Equip.*
1.5mil$total
Plant Area, m2
600
*Total cost including installation
4.3.6 Electrode Slitting
The coated electrode foils are slit into strips between the coated sections and then into individual
electrodes as shown in Fig. 2.2. The estimated scrap loss of foil for this process is about 8%. The
estimated resources to meet these needs are the following:
Electrode Slitting
Rate Factor
8,170,000
r\
m /y cell area
Direct Labor
4 per shift
Capital Equip.*
2.0 mil$ total
Plant Area, m2
200
*Total cost including installation
4.3.7 Final Electrode Drying
In the absence of electrolyte, no harm is done by exposing the electrodes to normal factory air;
however, the electrodes must be dried by heating under vacuum prior to cell assembly.
Maintaining extremely low moisture conditions during cell assembly is believed to be very
important in achieving long battery life. The final drying step coupled with dry room conditions
ensures a minimal quantity of moisture will exist in the final product. The pertinent processing
rate in determining the resources necessary for drying is the total amount of active materials
processed per year (other electrode materials are approximately proportional), which for the
baseline plant is 2,950,000 kg/y or 3,275 kg/shift. The individual electrodes exiting from the
electrode slitting process are separated into stacks by polarity, loaded into vacuum drying ovens,
dried for several hours, and unloaded directly into the dry room. The estimated resources to meet
these needs are the following:
Electrode Drying
Dryer capacity
Number of dryers
Maximum rate
Rate Factor
2,950,000 kg/y
active material
Direct Labor
2 per shift
Capital Equip.*
1.6mil$total
600 kg/shift
8
4,320,000 kg/y
Plant Area, m2
200
*Total cost including installation
4.3.8 Control Laboratory
The purpose of the control laboratory is to ensure that the raw materials and the electrodes being
fabricated meet specifications. Laboratory personnel collect or supervise collection of samples
and carry out analyses. The estimated resources to meet these needs are the following:
36
-------
Control Lab
Rate Factor
869,000 kWh/y
Direct Labor
4 per shift
Capital Equip.*
1.5mil$total
Plant Area, m2
200
*Total cost including installation
4.3.9 Cell Stacking
The cells are assembled in four steps, which are carried out in a dry room. The first of these steps
is cell stacking. The primary rate factor that determines the cost for all steps in cell assembly is
the number of cells assembled per year. For cell stacking an additional cost factor is the capacity
of the cells; large cells usually require more electrodes of larger area and thus a more capable,
faster cell stacking machine. The method used to determine the extra costs of stacking equipment
in detailed in Table 4.3. The capacity of the cells is deemed to have only a minor effect on the
other steps in cell assembly and is not taken into account for those steps. The electrodes are
inserted in a folded separator sheet, the positive electrodes tabs protrude on one side and the
negative electrodes tabs on the other. As in other parts of the plant, excess capacity is provided to
allow catching up after unscheduled downtime. The estimated resources to meet these needs for
the baseline plant are the following:
Cell Stacking
Stacking rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
6 per shift
Capital Equip.*
5.0 mil$ total
4 cells/min
5
8,640,000 cells/y
Plant Area, m2
400
*Total cost including installation
4.3.10 Current Collector Welding
The current collector tabs for the negative and positive electrodes are welded to their respective
terminals by ultrasonic welding. This procedure achieves a connection of near-zero resistance
and avoids overheating the electrodes during the welding process. The estimated resources to
meet these needs are the following:
Tab Welding
Cell rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
6 per shift
Capital Equip.*
5.0 mil$ total
4 cells/min
5
8,640,000 cells/y
Plant Area, m2
400
*Total cost including installation
4.3.11 Enclosing Cell in Container
The aluminum foil in the pouch container is sufficiently thick (100 microns default thickness) to
permit the use of stiff, pre-shaped pouch halves. The pouches are assumed to be purchased as
37
-------
finished parts. Each cell is enclosed in these containers, which are then partially sealed prior to
injecting electrolyte. The estimated resources to meet these needs are the following:
Enclosing cells
Cell rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
4 per shift
Capital Equip.*
3.0 mil$ total
4 cells/min
5
8,640,000 cells/y
Plant Area, m2
400
*Total cost including installation
4.3.12 Electrolyte Filling and Cell Sealing
At this station, the cells are evacuated, filled with electrolyte and temporarily sealed. The
estimated resources to meet these needs are the following:
Filling & 1st Seal
Cell rate
Number of units
Maximum rate
Rate Factor
6,320,000 cells/y
Direct Labor
6 per shift
Capital Equip.*
6.0 mil$ total
4 cells/min
5
8,640,000 cells/y
Plant Area, m2
600
*Total cost including installation
4.3.13 Dry Room Management
Excellent dry-room atmosphere is required for lithium-ion cell assembly. A maximum dew point
temperature of -40 °C is maintained in the room. The load on the dry-room drying apparatus is
determined by diffusion of water vapor through the walls, entry of air through the air locks, the
number of workers in the room, and the need to admit some fresh air to limit the build up of
contaminants such as electrolyte solvent vapor. These load factors are approximately a function
of the room area. Because of the importance of the proper functioning of the dry room, two
workers are on duty at all times to monitor its performance. The equipment for circulation and
purification of the dry air will be located outside of the plant building, adjacent to the dry room.
The estimated resources to meet these needs are the following:
Dry Room
Operating
Area
2,000 m2
Direct Labor
2 per shift
Capital Equip.*
20.0 mil$ total
Air Locks, m2
75
*Total cost including installation
4.3.14 Formation Cycling, Final Cell Sealing and Charge Retention Testing
Formation cycling is expensive because it takes considerable time and each cell must be
monitored separately. For plants to be operated in 2020, we expect some improvements from
present day operations because of the urgency to improve and thus save cost. We project that the
entire formation cycling and testing can be done in two shifts. These operations consist of
38
-------
charging the cell, discharging to full depth to measure capacity and impedance, followed by fully
recharging the cells. These tests will be carried out in large temperature controlled cycling units
that test 500 cells simultaneously, monitor each cell and automatically identify failed cells. The
capital cost of the cycling equipment is primarily a function of the annual number of cells to be
tested, but to a lesser extent on the capacity of the cells.
The short-term testing described above does not detect cells that have self-discharge rates that
are slightly above normal, which could lead to catastrophic failures later. To detect such defects,
the cell charge is topped off and the cells are stored for two weeks and then checked for loss of
charge. Most of the test period is spent in large racks in compact arrays, without electronic
monitoring. Incidentally, the two-week long self-discharge testing requires less floor space than
for formation cycling, which lasts only two shifts.
The final cell sealing occurs between the formation cycling and charge-retention storage test.
Gas generated during formation cycling may accumulate in the reservoir space that was created
during the temporary sealing step. This gas is removed by creating the final seal below the
reservoir and trimming off the unwanted portion.
The estimated resources to meet these needs are the following:
Formation Cycling
Cell capacity
Number of cyclers
Cells per cycler
Length of test
Testing capacity
Final Cell Sealing
Charge Retention
Testing rack capacity
Racks per stack
Number of racks
Length of test
Testing capacity
Rate Factor
6,320,000 cells/y
6,320,000 cells/y
6,320,000 cells/y
Direct
Labor
8 per shift
2 per shift
3 per shift
Capital Equip.*
30.0 mil$ total
40 Ah
35
500
2 shifts
7,875,000 cells/y
2.0 mil$ total
4.75 mil$ total
500 cells
5
750
14 days
8,040,000
Plant Area,
m2
1500
300
600
*Total cost including installation
4.3.15 Module and Battery Assembly
Approximately 5% of the cells are expected to fail the formation cycling and charge-retention
tests and these are sent to the Rejected Cell and Scrap Recycle section. The accepted cells
(6,000,000 finished cells per year) are assembled into modules by attaching the terminals
through laser welding or mechanical joining with spring loaded devices. Electronic circuit packs
are attached that occupy about the same volume as a cell. These operations are carried out at four
39
-------
automated stations each capable of handling about 280 cells per hour. For the module design
being cost estimated in this model, the module is enclosed in an air-tight aluminum container by
double seaming. The processing rate that determines the cost of module assembly is the number
of finished cells that must be handled per year.
The finished modules are assembled into battery packs with the aid of automated stations. The
total cost of these stations is dependant mainly on the number of battery packs to be assembled
per year (100,000 for the baseline plant), but to a lesser extent on the number of modules per
pack. After assembly, the packs are moved to testing stations where they are discharged as a final
check of impedance and to lower the state of charge to a level suitable for shipping. The
estimated resources to meet these needs are the following:
Module Assembly
Number of stations
Cells/h/station
Capacity
Battery Pack Assembly
Modules/pack
Number of stations
Packs/h/station
Capacity
Battery Pack Testing
Rate Factor
6,000,000 cells/y
100,000 packs/y
100,000 packs/y
Direct
Labor
6 per shift
4 per shift
2 per shift
Capital Equip.*
6.0 mil$ total
4
280
8,060,000 cells/y
3.0 mil$ total
4
3
6
130,000 packs
3.0 mil$ total
Plant Area,
m2
400
300
300
4.3.16 Rejected Cell and Scrap Recycle
Scrap is generated in preparing the electrodes and by the rejection of 5% of the cells that go
through formation cycling and charge-retention tests. This scrap is gathered and packaged for
shipment for recycling of the materials having value. No credit is taken for the value of the scrap
in this model except that the costs of gathering, sorting, packaging and shipping are understated
by about that value. The main factor in determining the cost of scrap recycle is the number of
cells rejected, which have to be disassembled to recover the scrap, a labor intensive process. The
yields of materials in the various processing steps are shown in Table 4.4.
Table 4.4 Materials yields during electrode and cell fabrication
Material
Positive Electrode
Negative Electrode
Positive Current Coll.
Negative Current Coll.
Separator
Electrolyte
Material
Mixing
99
99
Coating
95
95
99
99
Electrode
Slitting
99
99
92
92
Cell
Stacking
99
99
99
99
98
Electrolyte
Filling
94
Total
92.2
92.2
90.2
90.2
98.0
94.0
40
-------
The estimated resources needed for scrap recycle are the following:
Scrap Recycle
Scrap rate
Rate Factor
6,320,000 cells/y
Direct
Labor
5 per shift
Capital Equip.*
2.5 mil$ total
441 kg/shift
Plant Area, m2
400
4.3.17 Baseline Plant Summary
The processing rates and the primary cost factors for the baseline plant are summarized in Table
4.3. The main cost-determining rate of processing for each step is shown in the second column.
The requirements for direct labor, capital equipment and plant area, which are shown in detail in
the subsections above, are summarized in the table. It is seen that the plant requires a total of 95
workers per shift, $130,450,000 worth of capital equipment, and 10,325 square meters of plant
area to manufacture the baseline battery at a rate of 100,000 battery packs per year.
4.4 Adjustment of Costs for Rates Different than those of the Baseline Rates
As noted in Table 4.3, there are many processing rates that must be considered in addition to the
overall number of battery packs manufactured per year. Each of these rates affects the costs of
one or more steps in the process and may have no effect upon the costs of other steps in the
process. For instance, when the user of the model increases the power of the battery packs
without increasing the number of cells or their capacity, the model increases the area of the cells
and decreases the electrode coatings thicknesses. Such changes would result in an increase in the
cost of the coating equipment, the floor area occupied by the equipment, and in the direct labor
for that step in the process. It would have no effect on the cost of mixing the materials to be
coated because the amounts of these materials per battery back are unchanged under the assumed
conditions.
The general approach to cost estimation of multiplying a known cost by the ratio of processing
rates raised to a power has also been applied to the capital cost of individual items of
24
equipment.
C = C0(R/R0f
(4.1)
Here, C0 is the capital cost of an installed equipment item designed for the baseline processing
rate, ^0. The power factor, p, relates the capital investment cost and the processing rate for the
manufacturing step.
If the value of p were 1.0, it would imply that the cost of the equipment item, or the equipment
items if there are several in parallel, would be directly proportional to the processing rate.
However, the value of p for the cost of equipment is frequently about 0.6 to 0.7 for many
manufacturing process steps because the equipment is larger for the higher processing rates and
its cost is less than if it were directly proportional to the processing rate. For process steps
41
-------
requiring the addition of many identical pieces of equipment for scale up, such as may be true for
formation cycling of battery cells, the value of;? may be as high as 0.9. The value ofp is unlikely
to reach 1.0 because the equipment cost includes installation, for which there is some savings
even in installing multiple units of the same processing capacity. The relationship between cost
and processing rate for two-fold and three-fold rate changes is illustrated in Table 4.5.
Table 4.5 The effect of processing rate (R) on cost for various scale factors
CIC0 = (R/R0f
Scale Factor, p
0.25
0.3
0.4
0.5
0.6
0.7
0.8
0.95
1.0
Cost Ratio, C/C0
R/R0 = 2
1.19
1.23
1.32
1.41
1.52
1.62
1.74
1.93
2.00
R/R0 = 3
1.32
1.39
1.55
1.73
1.93
2.16
2.41
2.84
3.00
Similar equations have been applied for determining the effect of processing rate on the annual
hours of labor and the plant area required for a manufacturing step. In general, the value of p is
low for the labor equation, usually only 0.4 to 0.5, because only a relatively small addition to the
labor crew permits operation of larger equipment or of operating several more units of the same
processing capacity.24 The value of p for the plant area required for a processing step is slightly
less than that for equipment. The floor area required for larger equipment or for more equipment
items of the same size is proportionately less than the increase in the processing rate because of
the more efficient use of the space occupied by the equipment and the savings in aisle area.
The value of the scale factors (i.e. p factors) for labor, capital equipment, and floor area were
estimated for each of the processing steps (Table 4.3). The scale factors selected for the direct
labor requirement are usually only 0.4 to 0.5, which indicates considerable unit cost reduction for
increasing the plant throughput.
For most processing steps, increasing the processing rate beyond that in the baseline plant would
result in a decision to increase automation or use faster equipment to mitigate the costs of higher
levels of throughput. Decreasing the processing rate would have the opposite effect. Some steps
in the process such as cell stacking, welding of current collectors, and formation cycling do not
appear to be easily automated beyond the level intended in the baseline plant and, thus require a
higher value for the scale factor of 0.8. This higher scale factor results in achieving fewer
reductions in the cost per battery pack with increasing production volume. Additionally, a higher
p factor results in a less severe penalty for lower production scale for an individual step in the
process.
42
-------
There are five steps for which the cost of the capital equipment is affected by other factors than
the main processing rate for the process step. These are discussed in the footnotes at the bottom
of Table 4.3. For these steps, the costs that have been adjusted for the changes in the processing
rate from the baseline rate are further adjusted to take into account the other cost factors. The
cost of the coating equipment is adjusted for the amount of solvents to be driven off of the
positive and negative electrodes; thicker coatings need longer, more expensive ovens to drive off
the additional binder solvent or the coater most be operated at lower speeds. The cost of the cell
stacking equipment and that of the formation cycling equipment, for which the main cost factor
in both cases is the number of cells to be fabricated annually, are also adjusted for the capacity of
the cells; larger cells require more expensive equipment. The cost of the capital equipment for
battery assembly is primarily a function of the number of cells in the battery, but it is also a
function of the number of modules that must be interconnected. This dependence is accounted
for in the model with an additional multiplying factor.
A breakdown of the baseline plant capital equipment costs listed in Table 4.3 is illustrated in Fig.
4.3. The largest costs for capital equipment are for formation cycling and testing, cell assembly
in the dry room and electrode coating. These capital costs are likely to be dominant in any
lithium-ion battery plant in the near future.
D Receiving and shipping
• Materials preparation
D Electrode coating
D Calendering
• Materials handling
D Electrode slitting
• VacuLmdrying
D Control laboratory
• Cell assembly i n dry room
• Format! on cydi ng and testi ng
n ModUe and pack assembly
D Rejected cell and scrap recycle
Figure 4.3 Breakdown of installed capital equipment costs for the baseline plant
4.5 Plant Investment Costs
In this model, the calculated investment costs are defined as those directly related with building
and operating the plant (Table 4.6). Other costs that may require investment, such as research
43
-------
and development, are added separately to the unit cost of the battery. The largest investment cost
is for the installed capital equipment. Each cost item for the battery under design is adjusted from
the estimate of the baseline plant. The plant cost is done in a similar way with a cost of $1,500
per square meter ($140/sq. ft) including land and utilities. Launch costs include plant start-up,
employee training and materials that are lost or recycled in early stages of production, beyond
the normal amounts. Launch costs are estimated to be 5 % of annual materials costs plus 10 % of
annual direct labor and variable overhead (Section 4.6). Working capital is needed to cover the
costs of payroll, receivables, and the inventories of raw materials, work in progress and finished
product. These working capital costs are partially offset by bills that are payable. We estimate
the working capital to be 15 % of the annual variable costs.
Table 4.6 Battery pack manufacturing investment costs
Investment Costs
Capital Equipment
Plant Floor Space
Launch Costs
Working Capital
Description
Equipment costs including
installation
Space includes aisles and space
for unfinished processing
inventory plus land and utility
costs.
Plant start-up, training, out-of-
spec product.
Cash to meet payroll,
receivables, inventories of raw
materials and of unfinished and
finished product, minus
payables.
Method of Calculation
Estimates of costs for each
processing step at baseline rates
adjusted for actual rates.
Estimates of costs for each
processing step at baseline rates
adjusted for actual rates.
5% of annual materials cost,
10% of direct labor plus
variable overhead.
15% of annual variable costs.
4.6 Unit Costs for Battery Pack
The unit costs of the battery pack are calculated as summarized in Table 4.7.
4.6.1 Variable Costs
The costs of the materials and purchased items are based on the costs discussed in section 4.2,
and the annual amounts of materials are adjusted for the yields of materials (section 4.3) and
yield of cells. The direct labor is the sum of the labor cost for each step in the process, which are
each calculated for the baseline plant and adjusted for the rate associated with the battery under
study. Variable overhead is the cost of indirect materials and labor, utilities, and plant
maintenance. It is estimated to cost 60 % of direct labor costs.
44
-------
4.6.2 Fixed Expenses
Fixed expenses include General, Sales, and Administration (GSA), research and development,
and depreciation. The cost of GSA includes the plant office, taxes on income and property, cost
of sales and insurance. It is estimated by the model as 25 % of direct overhead plus 35 % of
depreciation. Research and development (R&D) must be carried out to ensure that the battery
packs that are produced in the plant and the means of production continue to be competitive in
the world market with respect to performance and price. The greater the investment in the plant
and its equipment, the greater is the need to be successful in the R&D effort. Thus, the
expenditure has been set at 50 % of the depreciation expense. Depreciation expense provides
funding available for future investment in this plant or another venture to replace deteriorating
plant and equipment. The equipment and plant are depreciated at straight-line rates for 8-year life
(12.5 % per year) and 20-year life (5 % per year).
Table 4.7 Unit cost of battery pack
Variable Costs
Materials and Purchased
Items
Direct Labor
Variable Overhead
Fixed Expenses
General, Sales, and
Administration (GSA)
Research and Development
Depreciation
Profit
Warranty
Description
All materials and purchased
items in finished product and
lost in processing.
Labor costs for operations and
immediate supervision.
Indirect materials, labor,
utilities, plant maintenance
Plant office, taxes on income
and property, cost of sales and
insurance expenses.
On-going research needed to
upgrade product and maintain
competitive position.
Provides funds for new
investments to replace those in
current equipment and plant.
Return on invested capital after
taxes.
Funds set aside for reimbursing
customers for battery pack
failures.
Method of Calculation
Based on prices of materials,
cost equations for purchased
items and yields.
Estimates of costs for each
processing step at baseline rates
adjusted for actual rates.
60% of direct labor cost.
25% of direct labor and variable
overhead plus 35% of
depreciation.
50% of depreciation
12.5% of capital equipment cost
plus 5% of plant floor space
cost.
5% of total investment costs.
5.6% added to price based on
present worth of projected
payments.
45
-------
4.6.3 Profits
The profit goal for this type of venture varies with the financial structure of the company,
especially regarding long-term debt. For the model, the profit is set to provide a 5 % return on
the total investment, which is an approximate average for mature manufacturing as vehicle
battery production is expected to be in 2020.
4.6.4 Battery Pack Warranty Costs
If a battery module or an entire pack fails, the replacement will cost much more than the original
price paid by the OEM. It is important that such events are rare, but provision must be made to
reimburse the vehicle owner, especially in the early years of the projected battery life. The extra
costs of replacing the battery will result from labor for testing and replacing the battery,
inventory costs for stocking replacement batteries, and servicing the battery controller if the new
battery is slightly different than the old battery. It is likely that the battery manufacturer will be
responsible for the cost of the new battery, which we assume will be equal to the cost of the
original battery. The other costs of replacing the battery, to the extent that they are covered by
the warranty, are assumed here to be covered by the automobile manufacturer and the dealer.
The goal for average battery life is 15 years and a warranted life of 10 years, with full
replacement in the first five years and shared cost of replacement for the last five years seems
appropriate. The vehicle owner would pay an increasing share of the cost from between 0 % at 5
years to 100 % at 10 or more years. With these assumptions, the cost to the battery manufacturer
will be equal to the present worth of the future costs of the new battery or modules as provided in
the warranty. The rate of battery failure will vary over the life of the battery with a slightly
higher rate early in life, then a low failure rate followed by a gradually increasing failure rate.
For purposes of calculation we assume a failure rate of 1.0 % per year throughout the warranty
period. With an internal rate of return of 8 % and calculated on a monthly basis, the present value
of the future costs would be about 5.6 % of the price of the battery before adding the warranty
cost.
4.7 Summary of Baseline Battery Cost
The spreadsheet version of the model, which is discussed in more detail in sections 5 and 6,
provides a summary sheet which is illustrated in Table 4.8 for the cost of the baseline battery and
that of two others. This breakdown of the battery costs, with a brief summary of the design
values, illustrates the effects of the cost factors. The second battery has twice the power of the
baseline battery and the third battery has the same power as the baseline battery, but twice the
capacity. The number of cells is the same for each battery. The energy storage is slightly higher
for the battery with double power because the voltage would be slightly higher during the
discharge to determine capacity. The battery with double the capacity has fewer electrodes which
are longer and wider, because the cell thickness in maintained, resulting in higher resistance in
the current-collector structure. The higher impedance lowers the voltage during the discharge
capacity measurements and results in slightly less than twice the energy storage of the baseline
battery.
46
-------
Table 4.8. Summary of results for cost of baseline battery and that of similar batteries
with double the power and double the capacity of the baseline battery
Baseline Double Power Double Capacity
Calculated Battery Parameters
Battery energy storage, kWh 8.7 8.8 17.1
Battery power at 80 % OCV, kW 50.0 100.0 50.0
Required battery power, kW 50.0 100.0 50.0
Capacity, Ah 40 40 80
Number of cells 60 60 60
Battery weight, kg 55.8 69.2 96.9
Battery volume, L 35.9 42.0 58.3
Vehicle electric range, miles 24.3 24.6 47.9
Investment Costs
Capital equipment cost including installation, mil$ 130 149 152
Building, Land and Utilities
Area, m2 10,325 12,193 11,744
Cost, $/m2 1,500 1,500 1,500
Building investment, mil$ 15 18 18
Launch Costs 10 13 15
Working capital, mil$ 28 36 42
Total investment, mil$ 184 216 227
Unit Cost of Battery Pack, $
Variable Cost
Materials and Purchased Items
Cell materials 1,350 1,864 2,232
Cell purchased Items 75 75 90
Module and battery 228 234 276
Total 1,652 2,173 2,598
Direct Labor
Electrode processing 36 52 40
Cell assembly 31 31 31
Formation cycling, testing and sealing 17 17 17
Module and battery assembly 16 16 16
Rejection and recycling 666
Receiving and shipping 8 8 11
Control laboratory 557
Total 119 135 128
Variable Overhead 72 81 77
Total Variable Cost 1,843 2,389 2,803
Fixed Expenses
General, Sales, Administration 107 122 121
Research and Development 85 98 99
Depreciation 171 195 199
Total Fixed Expenses 364 415 419
Profits after taxes 92 108 113
Total unit cost per battery not including warranty, $ 2,299 2,912 3,335
Summary of Unit Costs, $
Materials 1,350 1,864 2,232
Purchased Items 303 309 366
Direct Labor 119 135 128
Variable Overhead 72 81 77
General, Sales, Administration 107 122 121
Research and Development 85 98 99
Depreciation 171 195 199
Profit 92 108 113
Warranty 129 163 187
Price, $ 2,428 3,075 3,522
47
-------
Doubling the power does not add as much cost to the materials and purchased parts as doubling
the cell capacity. Most of the labor costs for the three batteries are similar with the major
difference being for the labor cost for electrode processing. The double power battery requires
greater labor costs principally for coating the larger electrode area. Capital equipment and
depreciation costs are higher for both the high power and high capacity battery packs. The
increases in capital equipment cost for the high-power battery are for coating, calendering,
materials handling and vacuum drying equipment. For the high-capacity battery, the main
additional capital equipment costs are for the materials mixing, binder solvent recovery, cell
staking and formation cycling steps in the process.
Overall, doubling the power of the battery increases the cost by only 27 %. Doubling the
capacity of the cells increases the cost by 45 %, considerably more than for doubling the power.
Alternatively, doubling the number of baseline cells and modules within a larger battery jacket
(two rows of modules instead of one, twice the voltage, energy, and power) would increase the
cost by 83 %.
The summary of unit costs for the baseline battery pack, which is shown at the bottom of Table
4.8, is illustrated in Fig. 4.4. The materials and purchased items are the largest costs for the
battery. For larger levels of production, these costs are even more dominant because the scale
factors for these items are close to one.
56%
12°/c
D Materials
• Purchased Items
D Direct Labor
n Variable Overhead
• General, Sales, Administration
D Research and Development
• Depreciation
n Profit
• Warranty
Figure 4.4 Breakdown of unit costs for baseline battery with total price to OEM of $2428.
48
-------
5 Description of the Spreadsheet Model and Instructions for Use
5.1 Background
®
Historically, the model has been based on Microsoft Office Excel spreadsheets. The flexibility
afforded by a spreadsheet approach has been extremely useful to the development of the
calculations. Until now, the model had been in a constant state of development. Changes to
parameters and equations were made rapidly and frequently. The publication of this report
represents the first time a version of the model will be "frozen" for open distribution to the
public. Advances will continue to be made with the model, such as those discussed in the last
section of this report. However, distributions of the revised model will be made in an orderly
fashion rather than the continuous improvement approach taken over the last number of years.
5.2 Instructions
The following subsections are a brief explanation of how one may operate the spreadsheet based
model. The user is advised to save the original document separately as a back-up copy.
Corruption of the calculation is possible and will likely occur during use by someone unfamiliar
with the model.
5.2.1 Enabling Calculation
This Microsoft® Office Excel workbook requires the use of iteration. To enable this feature, go
to the "Tools" drop-down menu and select "Options." On the calculation tab, check the box next
to "Iteration" and change the maximum number of iterations to 1000 (Figure 5.1). Also ensure
the calculation is set to automatic and not manual. If the iteration is not turned on, the software
will present an error complaining about circular references.
] File Edit View Insert Format T>:.ls Data Window Help
L142
136
137
138
39
40
41
42
43
44
46
•47
148
149
150
151
157
1 53
154
155
156
1 57
158
159
160
161
162
ltd
164
165
166
1 68
A
«
B C D E F
Energy requ rement. Wh/mile 250
Available bat ery energy. % of total 70
Vehicle rang
Cell Caps
Select cap*
Capacity
Battery e
Vehicle r
Capacity at
Capacity ho
Posit™ elec
^ositive elec
Convergence
5»
111
O
e. miles
city Calci.
city, battery
|Ah)
ergy (kWh)
n ge (miles
;/3, Ah
10 DO
lation
energy, or vehicle range, but only one.
10.148
rode thickness 16.4
rode thickness holding 16,4
parameter
0.1
Restart (0/1) 1
^^^^^^ Chart Data
Vein cle range, miles 10
Number of cells in parallel 1.00
Volume 28
Weight 42
Price 2438
4,000 -
2 500 -
9 nnn -
inal^|ioo-/. - ®|:
6
250
70
20.00
20.0
20.429
20.429
' sTi
51.1
0.1
1
20
1.00
35
53
2603
H
250
70
3D no
30.0
30.825
30.825
' 82.0
82.0
01
1
IB
1.00
44
69
3038
^^
**^^
_Jr^Vx"
^*
- 1
Anal
I J
250 250
70 70
40.00 50.00
40.0 50.0
41.343 51.988
41.343 51.988
' 112.6 [ 143.2
11Z6 1 143.2
01 0.1
1 1
40 50
1.00 100
53 63
88 105
3483 3962
>0
)0
)
O) i-
m 0)
-60 § |
||io - B / u | m
K L
Check
= •— i
M
Color | International |
View Calculation | Edit
& Automatic
* tables
| $ %
N
*
Td8 &
0
Save 1 E
| General
=^
P
or Checking
Transition
:M . (3» .
0
A * B i i> 1 s>
R S
-U2!
| Spelling | Security
Custom Lists 1 Chart
r Manual
17 Recalculate before save
17 Iteration
Maximum iterations: J1000
l~ 1904 date syste
ayed
Calc Now (F9) |
Calc Sheet |
Maximum change: 1 0,001
|7 5ave external link values
F Accept labels in formulas
r
OK
Cancel
3=
Figure 5.1 Iteration must be enabled for the spreadsheet model to function.
49
-------
5.2.2 System Selection Worksheet
The cell chemistry is selected by copying the system designated at the top of a column, for
instance NCA-G in cell F3, pasting it into cell E3 (Figure 5.2). Any of the values in row E can be
overridden by entering the desired value in column L. The selection of the cell chemistry also
includes the associated prices at the bottom of the page. These prices can also be overridden by
entering the desired values in column L. A full screen shot of the system selection worksheet is
in Figure 5.3.
loot Data window Nalp
A -J 3- / 1 - -
/jt a ilLK-vt'CT. frr help - - 6> )
Selected System
Positive Electrode
Active material capacity, mAh/g:
Weight %
Active material
Carbon
Binder
Binder solvent
Void, Vol%%
Density, a/cm'
Active material
Carbon
Binder
Negative Electrode
NIP capacity ratio after formation
Active material capacity, mAh/g:
Weight %
Active material
Carbon
Binder
Binder solvent
iVoid, Vol%%
Density, g/cm
Active material
Carbon
Cell Chemistry
Default Values
Override
NCA-G
160
89
6
5
NMP
32
4,78
1.825
1.77
1.25
290
95
0
5
Water
34
2.24
1.95
NMC-G
175
89
6
5
NMP
32
4.65
1.825
1.77
1.25
290
95
0
5
Water
34
2.24
1.95
LFP-G
155
89
6
5
NMP
50
3.45
1.825
1.77
1.25
290
95
0
5
Water
34
2.24
1.95
LMO-LTO
100
89
6
5
NMP
32
4,23
1.825
1.77
1,10
170
89
6
5
Water
40
3.40
1.95
LMO-G
100
89
6
5
NMP
32
4.23
1.825
1.77
1,25
290
95
0
5
Water
34
2.24
1.95
Other
Values
1.10
1.10
1.10
1.10
1.10
Figure 5.2 The specific cell chemistry for the battery design is selected on the System Selection
worksheet. Any value may be overridden by entering a value in column L.
5.2.3 Battery Design Worksheet
The Battery Design worksheet designs five or more batteries for any type of electric-drive
vehicle (Figure 5.4-5.6). The calculated designs are specific for the end batteries requirements
specified by the user. From the result, the amounts of materials and the purchased items required
for manufacture are easily available to be used in the manufacturing cost calculations found on
subsequent worksheets. Although a cell and module format is assumed, the exact format
(prismatic, pouch, can, etc) of the battery does not have a dominant effect on the cost for a set
cell chemistry system. Our experience teaches us that the amounts of electrode materials and the
number, capacity and electrode area of the cells, are the determining cost factors. Nevertheless, a
specific design format was selected and is shown on the Cell Design worksheet to provide a basis
for calculating the entire cell and battery related costs.
50
-------
File Edit View Insert Format loois Data Window Help
\3® AJ eJ ui I *> * <* - I tt W5% , | i Ar,d
E3 - * NCA-G
JSJ.X]
Type a question For help - _ fl X
U =[S]=Si$ % • Ts§ i!Ji«F*l E- •3>- A'|:.iif
B
I
K
L
M
Cell Chemistry
Negative Electrode
N/P capacity ratio after formation
Active material capacity, mAh/g:
Weight %
Active material
Carbon
Binder
Thickness, u,m
Void, Vol%%
Density, g/cm
Electrolyte density, g cnr
Cell Dimensions
Length-to-width ratio of positive electrode
Cell container thickness, jjm
Cell container density, g/cm3
Top of positive electrode to top of terminal, mr
Cell Voltage and Resistance Parameters
OCVatfull power (atSOCfor power rating), V 3.5
Open circuit voltage at 50% SOC, V 3.63
Solid state diffusion limiting C-rate (1Q-s), AW 27
Negative electrode cm3/cm3
Positive electrode crrrVcm3
Electrode system A3I for power, ohrn-crn2
Selected ASI value
At 50% SOC, 2-sec burst
At 50% SOC, 10-sec burst
At bottom of SOC range, 10-sec burst
ASI correction factor for limiting current, %
Electrode system ASI for energy, ohm-cm2
Maximum electrode coating thickness, |irn
Available battery energy, % of total
Selected % energy
HEV and HEV-HP
PHEV
ev
Cell Materials Costs
Positive Electrode, $/kg
Active material
Carbon 1
Binder PVDF
Binder Solvent (NMP)
Negative electrode material, $/kg
Active Material
Carbon Black
Binder PVDF
Binder Solvent
Positive current collector foil, S/m1
Negative current collector foil, $/n
Separators, $/m2
Insulators, $kg
Electrolyte, $/L
NCA-G NMC-G LFP-G LMO-LTO LMO-G Other
160
39
e
§
NMP
32
4 .78
1.825
1.77
1.2S
290
-9!-
5
Water
34
2.24
1.95
175
89
6
5
NMP
32
155 100
as
e
5
NMP
50
4.65 3.45
1 825
1.77
1.25
290
95
0
5
Water
34
2.24
1.95
1.10 1.10
Aluminum
20
Copper
12
20
SO
P. 46
1.20
1.3
160
2.12
15
3.551
3.680
27
74000
8900
18
236
30
3
51.9
300
25
70
8D
NCA-G
Aluminum
20
Copper
12
20
60
0.46
1.20
1.825
1.77
1.25
290
95
0
5
Water
34
2.24
1.95
1.10
Aluminum
20
Copper
12
20
50
0.46
1.20
1.3
150
2.12
15
3.663
3.750
27
74000
S9DO
21
26.6
33
3
58.5
300
25
1.3
150
2.12
19
3.350
3.350
120
74000
420000
89
6
5
NMP
32
4.23
1 825
1.77
1.10
170
89
6
5
Water
40
3.40
1.95
1.10
Aluminum
20
Aluminum
20
20
50
0.46
1.20
1.3
150
2.12
15
2408
2.514
120
500000
49200
20 6
25
32
1.5
55.0
300
25
e
10
1.5
13.3
300
25
70 7D 75
80 8D 80
NMC-G LFP-G LMO-LTO
100
89
6
5
NMP
32
4.23
1.825
1.77
1.25
290
95
0
5
Water
34
2.24
1.95
1.10
Aluminum
20
Copper
12
20
50
0.48
1.20
1.3
150
2.12
IS
3.819
3.954
120
74000
49200
16
20
25
2
44.0
300
25
70
BO
LMO-G
Values
Override Values
36.00
6.80
10.00
3.20
1900
680
10 Op
0.00
0.80
3.00
2.00
500
16.00
3300
6.80
10.00
3.20
19.00
68D
10.00
0.00
0.80
300
20D
500
16.00
20.00
6.80
10.00
3.20
19.00
6.8D
10.00
0.00
0.80
3.00
2.00
5.00
16.00
1000
6. BO
10.00
3.20
12.00
6.80
10.00
0.00
O.BO
O.BO
2.00
5.00
16.00
10.00
6.80
10.00
3.20
19.00
6.80
10.00
0.00
0.80
3.00
2.00
5.00
16.00
Baseline
p
14 4 . >i \System Selection/ Battery Design /' 5t
:Dtaw- ^ AutoShapes- \ ~» n O .Si -4 O ffl SI
of Results ,( ManufactuririQ '_osr '_.ali-ulatiuns / 'lost Input
ell Design / Plant Schematic
LtJjJT1
' A ' = 3 g
Figure 5.3 System Selection worksheet
51
-------
S] File Edit View
F4S
A
Insert Format
ft PHEV
B
Tools Data Window Help
"J • I l^ 1£5% - g Anal - 10
C
D E
B I U
F
-g
G
$ % • T.8
H
1 | Program for Calculating Performance and Materials Requiremen
2 LirliO. SOCoO. 1 5AI0.05O2-Graphite
4 System Chemistry Input
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Finished Cell Materials
Positive Electrode, g
Weight %
Active material 89
Carbon 1
Binder
Void
Total
SF6B
Vol. %
Negative Electrode, g
Active material
Carbon
Binder
Void
Total
Vol. %
r™
34
Balance of Cell
Positive foil,
m2
Negative foil, m2
Al
Cu
Separator, m2
Electrolyte ,
L
I
Density
4.78
1.825
1.77
100 2.749
Weight% Density
J95 2.24
0 1.95
5 1.10
100
Thick., [.im
20
12
20
1.406
Density
2.70
8.92
046
1.20
Positive terminal assembly, g
Negative terr
Cell contains
Cell mass, g
ninal assembly, g
r, g
Length-to-width ratio for positive electrode
Cell thickness, rnm
Thickness of terminal material, rnm
Top of positive electrode to top of terminal, mm
Cell Capacity Parameters
Positive active material capacity, mAh/g:
Positive elec
Negative act
Negative ele
trode capaci
ve material c
;trode capac
y, Ah/cm3
apacity, mAh/g:
ty, Ah/cm3
Negative-to-positive capacity ratio after formation
Cell Voltage and Resistance Parameters
OCV at full power, V
Open circuit
Electrode sy
Excess negs
Maximum all
Cell terminal
voltage average for discharge, V
stem ASI for
live area, %
energy, ohm-cm2
nwable electrode coating t
contact voltage loss, % of
2.12
lickness, ^m
cell OCV
Rate of terminal temperature rise at full power, °C/sec
Target % OCV at full power
% OCV at full power adjusted for thickness limit
47 I Battery Input Parameters
48 Vehicle type (microHEV, HEV-HP, PHEV, EV)
49 Duration of power burst (10 or 2), s
50 I Battery power, kW
51 1 Number of cells per module
52 Number of cells in parallel
53 Number of modules in row
54 Number of rows of modules
55 Number of modules per battery
56 Cells per battery
57 Battery pack insulation th ckness, mm
58 I Battery jacket total thickness, mm
59 Number of batteries manufactured per year
60
61
62
63
64
65
66
Cell Chemistry Input
Battery Performance and Design Input
Battery 1
63.43
4. 23
3.56
-
71.27
45.35
-
2.39
-
47.74
0.900
0.946
1.719
0.0395
7.2
23.8
12.5
376
1.30
10.0
1.00
15
160
0.392
290
0.387
1.25
3.551
3.680
51.9
3.68
300
0.01
0.2
80
80.0
PHEV
10
60
16
1
3
2
6
96
10
12
100,000
Battery 2
127.68
8.61
7.17
-
143.46
90.85
Battery 3
192.65
12.99
10.82
-
216.47
136.46
-
4.78
-
95.63
0.573
0.609
1.100
0.0541
8.5
28.0
15.9
463
1.30
10.0
1.00
15
160
0.392
290
0387
1.25
3.551
3.680
51 9
3.17
300
0.01
0.2
80
80.0
PHEV
10
60
16
1
3
2
6
96
10
13
100,000
7.18
-
143.64
0.530
0.570
1.024
0.0747
10.0
S3.1
20.6
613
1.30
10.0
1.00
15
160
0.392
290
0387
1.25
3.551
3.680
51.9
2.70
300
0.01
0.2
80
80.0
PHEV
10
60
16
1
3
2
6
96
10
13
100,000
Type a question For help ^ _ ff X
•r- ^- ^ g • *i» J
i
:s
Battery 4
258.39
17.42
14.52
-
290.33
182.45
-
9.60
-
192.05
0.512
0.556
0.992
0.0960
11.4
37.8
25.4
769
1 30
100
1.00
15
160
0.392
290
0.387
1.25
3.551
3.680
51.9
2.38
300
0 01
0.2
80
80.0
PHEV
10
60
16
1
3
2
6
96
10
13
100,000
J 7-
^m :
Battery 5
324.92
21.90
18.25
-
365.08
228.91
-
12.05
-
24096
0503
0.552
0.976
0 1176
12.7
30.2
927 1
1.30
10.0
1 00
15
160 H
0.392
290
0.387
•••
3.551
3.680
51.9
2.15
300
0.01
0.2
80
80.0
PHEV
10
60
16
1
3
2
6
96
10
14
100,000
Program for Calculating Performance and Materials Requirements
LiNiO. SOCoO.1 5AIO. 0502-Gi aphite
Calculated Cell Parameters
Capacity, Ah
67 Cell group capacity
68 Cell capacity
Draw
- ,j flutoShapes- \ \ n O iJ 4 O
v of Rir-r-dts / Mani.iF.artuMrn] Last L.I!
[HHi<3»-^-£- = s=2
Battery 1
10.1
10.1
ulations / Cost
J J,
Battery 2
Battery 3
20.4
20.4
npuC / Ceil Design
30.8
30.8
/ Plant Schemat
Battery 4
41.3
41.3
C /
Battery 5
52.0
52.0 ^j
iii-tir
Figure 5.4 Top portion of Battery Design worksheet.
52
-------
File Edit View Insert Format lools Data Window Help
- 10 - B I U P=
Type a question For help
ff x
ul
64
65
66
_6L
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
A B
C
Calculated Cell Parameters
Capacity, Ah
Cell group capacity
Cell capacity
ASI Calculation
Limiting current density, mA/cm2
Limiting C-rate, A/Ah
D
Electrode system ASI for power, ohm-cm2
Current collector resistance parameter, ohms
Current collector ASI, ohms-cm2
Total cell terminal ASI, ohms-cm2
Cell and battery terminal connections,
ohms
E F
Total cell hardware and battery resistance, ohrn-cm2
Total cell ASI for power, ohm-cm2
Total cell ASI for energy (C/3 rate), ohm-cm2
Electrode Thickness calculation
Thickness of positive electrode at target %OCV, (im
Thickness of negative electrode at target % OCV, p
Positive electrode thickness at adjusted % OCV, |irt
in
Negative electrode thickness at adjusted % OCV, |j.m
Cell Area Calculation
Area determined at target % OCV
Area limited by max. allowed electrode thickness,
Cell area based on total ASI for power
Cell Dimensions
cm2
Number of bicell layers (97% packing density)
Width of positive electrode, rnm
Length of positive electrode, mm
Length of current collector tabs, mm
Width of terminals, mm
Length of terminal material, mm
Width of cell, mrn
Length of cell, mm
Volume of cell, cm3
Module Parameters
Weight of each cell group interconnec
Module state-of-charge regulator asse
Terminal heating factor, W/g
Terminal resistance factor, A-ohms/ctr
Module terminals, if more than one mo
Module terminal resistance both termir
Module wall thickness (aluminum), mm
Balance of module materials, g
Module length, mm
Module width, mm
Module height, mm
Module volume, L
Module weight, kg
t (copper), g
nbly, g
dule (each 2. 0-cm long), g
als, ohrns
Calculated Battery Parameters
Total battery energy storage, kWh
Useable battery energy storage, kWh
OCV at full power, V
Nominal battery voltage (OCV at 50% SOC)
Maximum current at full power, A
Maximum current density at full power
C-rate at full power, A/Ah
mA/cm2
Air flow space above and below modules and at end of jacket, mm
Thickness of module compression plates (steel), mm
Battery pack length, mm
Battery pack width, mm
Battery pack height, rnm
Battery volume, L
Weight of module inter-connects (each 3-cm long), g
Battery terminals (each 3 0-cm long), g
Weight of module compression plates and steel straps, g
Resistance of module interconnects if more than one module, ohms
Resistance of battery terminals
Battery jacket weight parameter, g/cm2
Battery 1
G
H
Battery 2 Battery 3
10.1 20.4 30.8
10.1 20.4 30.8
85
27
50.2
0.005409
0.410
0.228
0.000447
0.712
50.9
65.8
16.4
20.8
16.4
20.8
15,759
1,092
15,759
64
97
126
18
89
30
105
156
164
0.0
128
0.077
0.00107
15
0.0000195
0.5
168
158
85
85
27 27
32,2 30.2
0.005409
0.005409
I J
Battery 4
Battery 5
41.3 52.0
41.3 52.0
85 85
27 27
29.3 28.8
0005409
0.528 0.696 0.868
0.126 0.100 0.086
0.000447
0.701
32.9
56.9
51.1
64.6
51.1
64.6
10,205
2,198
10,205
0.000447
0.841
31.0
55.4
82.0
103.6
82.0
103.6
9,601
3.317
9,601
31
112
146
18
104
30
120
176
212
0.0
21
132
171
18
124
30
0.000447
0.997
30.3
54.8
112.6
142.3
112.6
142.3
9,376
4,449
9,376
16
149
194
18
141
30
0.005409
1.040
0.076
0.000447
1.159
29.9
54.6
143.2
181.0
143.2
181.0
9,273
5,594
9,273
13
165
215
18
157
30
140 157 173
201 224 245
282 352 424
0.0 0.0 0.0
128 128 128 128
0077 0.077 0.077
0.00107 0.00107 0.00107
15
0.0000195
0.5
198
178
0.077
0.00107
15 15 15
0.0000195 ' 0 0000195 0.0000195
0.5 0.5 0.5
237 275 31 1
203 226 247
171 171 171 171 171
107 122 142 159 175
2.90 3.73 4.93 615 7.39
6.32
3.57
2.50
340.9
353.3
220
13.96
21.7
7.74 10.18 12.71
7.14
5.00
340.9
353.3
220
21.56
10.8
10.71
7.50
340.9
353.3
220
22.92
7.1
14.29
10.00
340.9
15.29
17.86
12.50
340.9
353.3 353.3
220
23.47
5.3
6.8 6.8 6.8 6.8
1
_
220
23.72
4.2
6.8
1.5 1.5 1.5 1.5 1.5
547 549 549 549 551
350 392 443 488 531
145
27.7
55
22
921
0.0001460
0.0000292
0.572
162 181 199
34.9 44.1 53.2
55
22
1166
0.0001460
0.0000292
0.842
55
22
1515
0.0001460
0.0000292
0842
55
22
217
63.4
55
22
1868 2223
0 0001460 0.0001460
0.0000292 0.0000292
0.842 1.112
i >ir
Figure 5.5 Middle portion of Battery Design worksheet
53
-------
The Battery Design worksheet automatically receives input from the System Selection
worksheet. These values are shown in purple (Figures 5.4 and 5.6) and must not be altered on the
Battery Design worksheet. As explained above, cell chemistry values may be adjusted on the
System Selection worksheet. The operator provides battery design input in the aqua colored cells
(Figures 5.4 and 5.6). It is expected that the default values for the cell design (lines 27 to 30) and
the thickness of the battery insulation (line 57) should serve well for most batteries. The battery
input parameters on lines 50 to 54 (Figure 5.4) and lines 141 to 143 (Figure 5.6) are the only
input values that the operator is required to provide to study a group of batteries. An important
variable is the type of vehicle battery, (microHEV, HEV-HP, PHEV, or EV) on line 48 in Figure
5.4. One performs the selection by typing the name of the vehicle battery type in cell F48. While
the correct spelling is important, capitalization is not. This selection automatically determines the
state of charge at which full power is designated (thus, the open-circuit voltage and AST for full
power) and the length of the power burst (2 seconds for microHEV and 10 seconds for all
others).
The cell capacity (lines 141 to 143 in Figure 5.6) can be set in any of three ways: (1) directly
specifying the capacity (Ah) on line 141, (2) specifying the total battery energy on line 142 or (3)
specifying the electric range of the vehicle (miles). Only one of the three lines should be filled in
and the others should be blank. The model will follow the directions of the top-most line with
non-zero values.
The number of batteries manufactured per year is selected on line 59 in Figure 5.4. Changing this
value from the default value of 100,000, which is the manufacturing rate for the baseline plant,
will change the manufacturing cost.
If it is desired to study more than five batteries in the same workbook it is only necessary to add
additional columns by copying the battery 5 column to the right as many times as desired. Care
should be taken that the appropriate values are maintained when the cells are copied over. The
aqua colored cells are typically the source of any problems. The same column additions most
also be done for the worksheets on Summary of Results and Manufacturing Cost Calculations.
5.2.4 Remaining Worksheets
The cost calculations are done on the Manufacturing Cost worksheet and the results for the
model are shown on the Summary of Results worksheet (Figure 5.7). No parameters need to be
entered on these worksheets by the operator; all of the input for these worksheets is from the
Battery Design and the Cost Input worksheets. Tables for presentations or for preparing graphs
of the data can be assembled at the bottom of either the Battery Design or the Summary of
Results worksheet. These tables can be transferred to a blank worksheet for more complex
studies. For instance, results for different cell chemistries can be copied and pasted (special
paste, values and numbers formats) to a blank worksheet. Preparing a chart has been illustrated
in the initial configuration by graphing the data tabulated at the bottom of the Battery Design
worksheet. The X-axis description will have to be adjusted if the cell capacity is selected by one
of the alternative methods. On the last two worksheets, the cell design and the baseline plant are
sketched.
54
-------
^j File Edit View Insert Format loois Data Window Help
;jL3aiiJifJiii*J-f"- tt 125°'- - §; «™'
C153 - *
JSJJ<1
Type a question For help -r _ fl X
! # SP 5 - a. - A -11 ,-H. B
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
A
B C
D
E
F
Module weight, kg 6.32
Calculated Battery Parameters
To
UE
Ol
N.
M
M
C-
Ai
TT
B;
Be
Bs
:al battery
eable batt
:V at full p
rninal batt
aximum cu
sximum cu
rate at full
- flow spat
ckness o
ttery pack
ttery pack
ttery pack
ttery volur
energy storage, kWh
ery energy storage, kWh
ower, V
sry voltage (OCV at 50% SOC)
rrent at full power, A
rrent density at full power
power, A/Ah
mA/cm2
e above and below modules and at end of jacket, mm
module corr
length, mm
width, mm
height, mm
ne, L
pression plates (steel), mm
Weight of module inter-connects (each 3-cm long), g
Battery terminals (each 3.0-cm long), g
Weight of module compression plates and steel straps, g
130 Resistance of module interconnects if more than or
131 Resistance of battery terminals
132 Battery jacket weight parameter, g/cm2
133|Battery jacket weight, kg
134
135
136
137
138
139
140
141
B£
V
Er
A\
Ve
ttery weight, kg
;hicle Electric Range
ergy requirement, Wh/mile
ailable battery energy,
hide range, miles
% of total
Cell Capacity Calculation
Se ect capacity, battery energy, or vehicle rang
Capacity (Ah)
e module, ohms
e, but only one.
142 Battery energy (kWh)
143 Vehicle range (miles)
144J Capacity atC/3, Ah
145 Ct
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
187
168
169
170
172
173
177
178
179
180
181
1S2
Ready
Pi
pi
C<
_
apacity holding
sitive electrode thickness
sitive electrode thicknt
mergence parameter
3?
LU
O
2
-------
^j File Edit View Insert Format lools Data Window Help
i L) L5 U ui Ul C* ul I 1 * <* - I tt 13°* - § i «™'
A3 ~ «
Type a question For help
s % • tig £°, if if i EE -
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
A
,
BCDEF G H 1 J K
Summary of Results
LiNi0.80Co0.15AI0.05O2-Graphite
Battery 1 Battery 2 Battery 3 Battery 4
Calculated Battery Parameters
Battery energy storage, kWh
Battery power at 80 % OCV, kW
Required battery power, kW
Capacity, Ah
Number of cells
Battery weight, kg
Battery volume, L
Vehicle electric range, miles
Investment Costs
Capital equipment cost including installation mil$
Building, Land and Utilities
Area, m2
Cost. $/m2
Building investment, mill
3.6
60.0
60.0
10
96
423
27.7
10.0
149
12,128
1,500
18
7.1
60'. 0
60.0
20
96
53.4
34.9
20.0
157
12,209
1.500
18
10.7
60.0
60.0
31
96
69.3
44.1
30.0
168
12,885
1.500
19
Launch Costs
Rate: 5
D/o of direct annual materials + 10% of other annual costs
Total, million$
Working capital (30% of annual variable costs), mil$
Total investment, mil$
Unit Cost of Battery Pa
Variable Cost
Materials and Purchased Items
Cell materials
Cell purchased Items
Module and battery
Total
Direct Labor
Electrode processing
Cell assembly
Format on cycling, testing ar
:k, $
d sealing
Module and battery assembly
Rejection and recycling
Receiving and sh pping
Control laboratory
Total
Variable Overhead
Total Variable Cost
Fixed Expenses
General, Sales, Admi
Research and Develc
Depreciation
nistration
pment
Total Fixed Expenses
Profits after taxes
10
27
204
1,191
93
272
1,557
44
41
11
20
9
5
3
144
Total unit cost per battery not including warranty, $
Summary of Unit Costs, $
Materials
Purchased
Items
Direct Labor
Variable Overhead
General, Sales, Administration
Research and Development
Depreciation
Profit
Warranty
Price to OEM, $
Price to OEM for Modules for One Pack, $
Chart Values
Range
Thickness
87
1,788
126
97
195
418
102
2,308
1,191
365
144
87
126
97
135
102
129
2,438
2,330
108
10.00
21
Weight 42.3
11
29
215
1,296
100.9
295.7
13
34
235
1,642
108 .1
310.2
1,693 2,060
39
41
22
20
9
7
5
143
86
1,921
129
103
205
437
107
2,465
1,296
397
143
86
129
103
205
107
40
41
22
20
9
S
6
146
88
2,294
135
110
220
465
117
2,877
1,642
418
146
88
135
110
220
117
138 161
2,603
2,476
127
20.00
65
3,038
2,901
136
30.00
104
53.4 69.3
14.3
Battery 5
17J
60.0 60.0
60.0
41
96
85.7
53.2
40.0
178
13,510
1 .500
20
15
40
253
60.0
52
96
105.0
63.4
50.0
187
14,081
1.500
21
17
47
272
2,007 2,380
114.8 121.2
324.2 356.8
2,446
41
41
22
20
9
10
7
150
90
2,686
141
116
233
491
127
3,303
2,007
439
150
90
141
116
233
127
185
2,858
43
41
22
20
9
11
7
153
92
3,103
147
122
244
513
136
3,752
2,380
478
153
92
147
122
244
136
210
3,488 3,962
3,343
145
40.00
142
85.7
3,788
174
50.00
I
181]
105.0
I I h/olume 27.7 34.9 44.1 53.2 63.4
Cost 2438 2603 3038 3488 3962 zl
D,™-
Ready
flutoShapes- \ \aO6i-4ClllH
Figure 5.7 Summary of Results worksheet
56
-------
5.3 Battery Design Format Requirements
As the battery design is based off an assumed format (Section 2), certain design requirements are
necessary to ensure the modeled battery is physically realistic. For a battery based on a single
module, no additional restrictions are required. However, most batteries for transportation
applications will use multiple modules. For these applications, the following rules should be
followed by the user of the model. An even number of cells should compose each module and an
even number of modules should compose each battery pack. These two requirements are to
ensure the electrical connections are in the appropriate places in the final battery pack. Finally,
the dimensions of the resulting battery pack should be examined. Some final designs may benefit
from changing the cell aspect ratio, HAV, to fit the end-use application. One example would be,
to change the height of the battery pack. Also, for a set number of cells in the pack, changing the
number of modules, thus cells per module, allows for adjustment of the pack dimensions.
5.4 Troubleshooting and General Advice
The spreadsheet iterates to find the solution and this sometimes causes error messages to appear
after an entry is changed. These errors can usually be removed by first correcting any erroneous
entries (non-numeric, two decimal points, etc.). Then the cells may be reset to default values by
entering a "0" (i.e. zero) in the restart cell, F149 in Figure 5.6. Finally, entering a "1" in F149
restarts the iteration process leading to a successfully converged answer.
At some point, a user will ask the model to design a battery that is outside the bounds of what is
allowable for the selected cell chemistry. The most common error is when too large of a P/E ratio
is requested. Two different physical limitations are approached with increasing P/E ratio. First,
the electrode thickness is shrinking. At some point, the value will become unrealistic and
eventually approach 0 crashing the calculation. At the same time, the C-rate for the active
material is approaching the limiting C-rate defined in the Cell Chemistry Worksheet. As this
value is approached, the AST will increase to larger and larger values, which thus demands
smaller and smaller electrode thicknesses. Eventually, the calculation will crash.
Common sense approaches to resolve these issues are to use lower designed power or higher
designed energy. The C-rate and electrode thickness are easily viewed in the model output.
These are found on the Battery Design worksheet in row 120 for the C-rate and rows 81 and 82
for the electrode thickness. Therefore, the user may try designs of increasing P/E ratios and
watch to see how the electrode thickness and C-rate is changing. Different cell chemistries will
have different sensitivities to the P/E ratio depending on the defined limiting C-rate and
calculated AST for power. What is possible with the LMO-G system will not always be possible
with the NCA-G system. P/E ratios that satisfy the expression in Eq. 5.1 generally result in
successful battery designs. Higher P/E ratios are allowable in some situations. Note that selecting
the microHEV design doubles the allowable C-rate since only two second pulses are used. The
limiting C-rate, rc,iim, may be found in cell E48 on the System Selection worksheet and is carried
over to row 71 in the Battery Design worksheet.
-<^ (5.1)
E 1.35
57
-------
6. Illustrated Results
The model may be used to study the effects of battery parameters on the performance and the
manufactured cost of the designed battery packs. A few examples are given below for the effects
of various parameters on battery pack volume, weight and cost.
6.1 Number of Cells in Series
For a set battery pack power, the number of cells in the pack has substantial effects on the price
of the pack, the pack voltage and the maximum current. These effects are illustrated (Figure 6.1)
for LMO-Gr PHEV20 batteries (providing 20-mile electric range) with 60-kW power. The price
of the battery increases by 26% in changing the number of series-connected cells in the pack
from 32 to 96. The change in the maximum current would also affect the cost of the motor and
the electronic converter and controller, but in the opposite direction. As a result of these
offsetting effects on the total cost of the electric drivetrain, a study needs to be done to determine
the optimum current at maximum power as a function of the total battery pack power and other
parameters (see the Future Work section).
2,500
2,000
Price
Max Current
Voltage
1,000
800
600
400
200
o>
T3
C
(0
g
I
3
O
u
<0
Q.
20 40 60 80 100
Number of Cells per Pack
120
Figure 6.1 The effect of the number of series-connected cells for LMO-Gr, 60-kW, PHEV20
packs with 7.14 kWh total energy (70% useable).
6.2 Battery Packs for HEVs
Lacking a definitive study on the cost of the entire drivetrain to establish the appropriate
relationship for battery current, voltage and power, Equation 6.1 is employed in the HEV
58
-------
examples that follow. Equation 6.1 strikes a balance between very high current and a large
number of cells, to aid in selecting the number of cells for a battery of 25 to 150 kW of pack
power:
Battery Current at full Power (A) = 1.5 x Power (kW) + 200
(6.1)
The number of cells required to match this relationship will differ for each cell chemistry to
match the cell open-circuit voltage and impedance of the cell. Also, the selection of the current
by this means is only approximate because the voltage is not continuous being defined in terms
of a discrete number of cells, which must be an even number for the module design in the model
except for single-module battery packs.
Low-powered (25-kW) battery packs for HEVs need to deliver only about 0.15 to 0.25 kWh of
energy or a total of about 0.6 to 1.0 kWh. Thus, a cell chemistry that is capable of a high P/E
ratio is paramount for these batteries in achieving minimum cost. Some battery chemistries
perform much better under these conditions than others as illustrated in Fig. 6.2. The number of
cells in these single-module HEV battery packs varies from 33 to 52 to provide a maximum
current at full power of 238 + 2 A, in agreement with Eq. 6.1. It is apparent that the NMC-Gr and
NCA-Gr systems studied here are not as well suited to very high P/E requirements as the other
systems.
-*-LFP-Gr
^LMO-Gr
-*-LMO-LTO
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Total Pack Energy (kWh)
1.6 1.8
2.0
Figure 6.2 Cell area-specific impedance (AST) for various cell chemistries for 25-kW HEV
battery packs delivering full power at about 238 A for 2 sec at 80% OCV
59
-------
HEV battery packs of about 0.4 kWh of useable energy may be utilized for high power-assist
applications. High-powered HEV battery packs of 40 to 100 kW may be required to deliver
maximum power for up to 10 seconds. Also, they may be required to deliver moderate power to
drive the vehicle electrically at 30 to 50 mph. This latter requirement would also require more
useable energy than for the standard HEV battery pack. For this illustration of high-powered
HEV battery packs, we have set the energy requirement at 1.6-kWh total energy and 0.4-kWh
useable energy. For a moderate-weight sedan with an average energy requirement of 0.25
kWh/mile, the battery pack could power the vehicle for 1.6 miles before being recharged. The
battery pack could then be recharged in 1.2 minutes at a rate of 20 kW or at a higher rate if
desired.
A comparison of some of the characteristics of LMO-LTO, LFP-Gr, and LMO-Gr battery packs
with power capabilities of 25-kW to 100-kW and with 1.6 kWh total energy is shown in Fig. 6.3.
As the power of the battery pack is increased, we have increased the number of cells so that both
the voltage and the current increase as determined by the relationship of Eq. 6.1. The unevenness
in the plotted lines for pack volume results from changes in the number of modules and the
dimensions of the packs. The volumes for these high-powered battery packs (Fig. 6.3) as well as
their weights (not shown in Fig. 6.3) are all well below the USABC targets for 25-kW battery
packs even though most of the battery packs are much more powerful than 25 kW.
25
20
= 15
o
o
ro
Q.
ro
m
10
-o- LMO-LTO
- A- LFP-Gr
-•-LMO-Gr
- - - Pack OCV
Max Current
- - 500
<,
c
400 2
-• 300
200
3
O
•o
c
ro
-- 100
Q)
O)
ro
*-
o
20 40 60 80
Battery Pack Power (kW)
100
120
Figure 6.3 Volume, voltage and maximum current for 1.6-kWh lithium-ion battery packs as a
function of pack power for packs delivering full power for 10 sec at 80% OCV and 50% SOC.
Battery packs have 10-mm insulation thickness.
60
-------
Many factors affect the costs of manufacturing lithium-ion battery packs. For batteries of high
P/E, the chief factors are cell chemistry and pack power as shown in Fig. 6.4, where the prices
are plotted for the same batteries as shown in Fig. 6.3, with the addition of the NCA-Gr system.
In our attempt to provide high power with a minimum amount of electrode materials with the
NCA-Gr system, the area-specific impedance (AST) became very high as the power was
increased resulting in large cell areas and high costs (Fig 6.4). The LMO-Gr system has a
significant cost advantage over the other systems in this illustration. These battery chemistries do
not have the same probability of meeting the life goal of 15 years, but work is underway
throughout the world to meet this need for all of these systems.
2,500
t— NCA-Gr
fc- LFP-Gr
i- LMO-LTO
3- LMO-Gr
20 40 60 80
Battery Pack Power (kW)
100
120
Figure 6.4. Battery pack price to OEMs at 100,000 packs per year manufacturing rate for 1.6-
kWh lithium-ion battery packs as a function of pack power for the same battery packs as in Fig.
6.3 with the addition of the NCA-Gr system.
6.3 Battery Packs for EVs
Lithium-ion batteries for PHEVs and EVs do not require a high P/E ratio or low AST to meet
their goals. The most important factors for performance are high specific capacity (mAh/g), high
cell voltage, and high electrode density. To compare the performance of EV battery packs made
from various Li-ion chemistries, we designed the packs to provide 150 kW of power at 80%
OCV at a maximum current of 425 A, which is consistent with Eq. 6.1. Each pack consisted of
four modules containing 28, 30, and 32 cells for the cell chemistries LMO-Gr, NMC-Gr, and
LFP-Gr, respectively. These selections resulted in nominal (OCV) pack voltages of 429 to 443 V
61
-------
and within 2 to 13 A of the targeted maximum current of 425 A (Eq. 6.1) for these 150 kW
battery packs. On this basis, the NMC-Gr system has excellent power and energy density (Fig.
6.5). The LMO-Gr system does well, despite only moderate specific capacity because of its high
voltage. The LFP positive electrode has higher capacity than the LMO chemistry but is hindered
by a lower cell voltage and greater electrode porosity resulting from the use of nanostructure
particles.
450
^400
E
I 350
•LFP(155mAh/g)
•LMO(100mAh/g)
NMC(175mAh/g)
Weight
50
100 150
Electric Range, miles
200
250
Figure 6.5 Weight and volume of electric vehicle battery packs with lithium iron phosphate
(LFP), lithium manganese-spinel (LMO) and lithium nickel-manganese-cobalt (NMC) positive
electrodes versus graphite designed to deliver 150 kW of power at about 425 A.
The LMO-Gr system appears to have the lowest price of the three systems in this illustration
(Fig. 6.6). By experimenting with the parameter values, it was found that EV battery packs of the
NMC-Gr type could match the low price of those in the LMO-Gr system. However, that would
require that the specific capacity would be substantially increased and, more importantly, the
cost of the NMC would have to be reduced to about $20/kg from $33/kg, perhaps by reducing
the cobalt content and increasing the manganese level.
62
-------
LFP(155mAh/g, $20/kg)
NMC(175mAh/g, $33/kg)
LMO(100mAh/g, $10/kg)
^ 10,000
'
o
g 8,000
4,000
50
100 150 200
Electric Range, miles
250
Figure 6.6 Battery pack price to OEM for LFP-G, LMC-G and NMC-G battery packs for same
designs as in Fig. 6.5.
6.4 Parallel-Connected Cell Groups and Electrode Thickness Limits
The model also allows the user to create parallel cell groups and to set a maximum electrode
thickness. The effect these two unique design factors have on cost are illustrated below in Figure
6.7 for the LMO-Gr and NCA-Gr systems. In this illustration, the EV battery pack design
parameters are 100 kW of power, 31 kWh of total energy, and 100 mile range. The nominal
battery pack voltage (OCV at 50% SOC) is around 340 V. Two maximum electrode thicknesses
of 100 and 200 microns are shown for the two cell chemistries. The thickness of the positive
electrode is limiting the LMO-Gr chemistry while the thickness of the negative electrode limits
the NCA-Gr chemistry. The LMO-Gr is the least expensive in all cases. However, the difference
between the two chemistries lessens with smaller limiting electrode thickness. The costs will
become even closer for lower designed P/E ratios. In general, thicker electrodes reduce the cost
of the battery pack by lessening the amount of inactive materials used (separator, current
collector, etc). Moving to 300 microns allows for greater savings in the LMO-Gr design but not
the NCA-Gr design. However, a lower P/E ratio design for NCA-Gr would take advantage of
electrode thicknesses greater than 200 microns.
The cell capacity is shown for the NCA-Gr case limited to 100 microns. While the exact values
will change with cell chemistry, they will all be similar. The cell capacity is reduced by one half
as a single cell is added in parallel. This approach is commonly used by cell manufacturers and
63
-------
OEMs that cannot reliably produce or successfully operate cells of high capacity for
transportation applications (> 60 Ah). However, this approach also increases the price of the
battery pack. In this example, the price is increased by ~ $500 when an additional string of cells
is incorporated in a parallel arrangement.
The model calculations show that the lowest cost battery pack will utilize thick electrodes and
large capacity cells. In current practice, these two approaches have yet to be successfully
implemented within the entire community. In the challenge of lowering costs, it is useful to point
out the largest gains come from the initial advances (e.g. moving from 100 to 200 micron limit).
After that point, the benefits are diminishing.
12,000
g 10,000
2
LU
o
S 8,000
8
8 6,000
a.
C3
CD
4,000
2,000
120
NCA 100 micron
NCA200 micron
Cell capacity, Ah
LMO 100 micron
1234
Number of Parallel Cells
Figure 6.7 Battery pack cost as a function of number of parallel cells and for different maximum
electrode thicknesses. The cell capacity is also shown for the NCA-Gr limited to 100 micron
electrode thickness.
6.5 Effects of Manufacturing Scale on the Price of the Pack
The effects of manufacturing scale come into the cost calculation even if the annual number of
packs produced is unchanged, but the design is altered (e.g. power is increased). For a fixed
design, the effect of changing the scale of operations depends on the fraction of the total price
that is made up of materials costs. Unit materials costs change little with scale whereas the costs
per pack for labor, capital and plant area may decline substantially with increasing production
rates, especially at low production rates, Fig. 6.8.
64
-------
9,500
9,000
g 8,500
8,000
2,000
o 1,500
£
LMO-Gr, EV150 mile, 150 kW
NMC-Gr, EV150 mile, 150 kW
100,000 200,000 300,000 400,000 500,000 600,000
Annual Manufacturing Rate (packs/y)
A LMO-Gr,PHEV20 mile, 60 kW
• LMO-Gr, HEV-25 kW
y = 10298x
-0.1367
100,000 200,000 300,000 400,000 500,000 600,000
Annual Manufacturing Rate (packs/y)
Figure 6.8 The effects of manufacturing rate on the price calculated by the model for battery
packs of various cell chemistries, power capabilities and vehicle types
65
-------
The lines in the graphs are for the best-fit power relationships through the data with power
factors of -0.086, -0.091, -0.137, and -0.201 from the top curve to that at the bottom. The least
negative power factor is for the battery pack with the highest fraction of materials cost in the
total pack cost. The more negative power factors result from a decreasing contribution of
materials cost as a fraction of the total pack cost. These power factors for equations of the cost of
a single unit can be converted to factors relating the total annual cost of manufacturing similar to
Eq. 4.1 by adding 1.0 to each power factor. Thus the factors become 0.914, 0.909, 0.863, and
0.799. These large factors show only a small to moderate effect of scale. When the power curves
are compared to the points in each of the graphs of Fig. 6.8, it is apparent that the scale factors
approach one as the scale increases. This is because the model assigns a value of 0.95 for the
active materials and 1.0 for the balance of the materials. As the production level increases and
the materials costs become a larger fraction of the total price of the battery, the scaling power
approaches 1.0 and the effect of scale become very small. Likewise, the effect of scale on battery
price is much larger for HEV batteries than for EVs because materials costs constitute a smaller
portion of the total cost for HEV batteries. Increasing the production rate for HEV batteries will
result in a more dramatic reduction in cost than increasing the production rate for EV batteries.
66
-------
7. Future Work
7.1 Thermal Modeling
The battery design format used in this report provides cooling to the battery pack by blowing air
over the top and bottom of the modules. The heat is transferred through the stiff, aluminum
pouch material to the module walls. An additional aluminum plate may be placed between cells
to enhance heat transfer. This method of cooling will be sufficient for some battery applications,
but other applications might require that the battery pack be over-designed to reduce heat
generation. Over-design would include providing a battery pack of higher power than called for
by the application or providing the power at a very high fraction of the open-circuit voltage
(>85%OCV).
We intend to develop an alternative battery design that provides liquid cooling directly to the
cells inside the modules. Means of distributing a dielectric coolant to the interior of the modules
will be studied. We will carry out thermal modeling studies on the heat generation during vehicle
operation on various driving cycles and on the removal of heat from both air-and liquid-cooled
battery packs. We intend to investigate the feasibility and cost of these options. The results of
these studies will be incorporated into the model in a separate thermal modeling section. This
work will be carried out in a cooperative effort with Argonne's Transportation Technology R&D
Center, which will perform vehicle simulation tests to determine the rate of heat generation in the
battery pack.
7.2 Optimum Battery Voltage for Minimum Drivetrain Cost
For a set cell chemistry and set battery pack power, the cost of the pack increases as the number
of cells and the pack voltage are increased (Fig. 6.1). The additional cost results primarily from
the cost of additional state-of-charge equalization circuits and the additional number of cells
needing formation cycling and testing. The increase in battery pack cost is almost linear with the
increase in the number of cells. As the number of cells is decreased, the pack current at
maximum power becomes very high and the cost of the balance of the drivetrain increases at an
accelerating rate. Thus, there must be a number of cells and an associated pack current at
maximum power for which the total cost of the drivetrain is at a minimum. This minimum would
be for the current at full power for which the slope of the cost curve for the balance of the
drivetrain versus current was equal to the negative of slope of the cost of the pack versus current.
This optimum current will increase with the pack power because the slope of the cost-versus-
current curve increases with increasing power and, therefore, the optimum current will also
increase.
To represent these phenomena in illustrating the model in Section 6, we used an equation (Eq.
6.1) for selecting the current at full power as a function of the battery pack power. This equation
is just an estimate and it does not provide for differences in the optimum current that would
result from differences in cell chemistries, which are known to affect the battery cost versus
power function.
67
-------
A study is needed to determine the cost of electric motors and the electronic control equipment
required for the vehicle and battery pack as a function of power and maximum current capability.
Once appropriate cost curves are established, we will determine equations relating the optimum
battery current to the desired battery power taking into account the battery chemistry. We intend
to do this study with the cooperation of Argonne's Transportation Technology R&D Center.
7.3 Multipurpose Battery Manufacturing Plants
Our cost modeling is based on the concept that a manufacturing plant is constructed to produce a
single type of battery pack at a predetermined level of production. In practice, manufacturers will
have to produce several types of batteries within the same manufacturing plant and the levels of
production may fluctuate. We intend to investigate how this increased flexibility requirement
will affect the various manufacturing costs. For example, the cell filling and sealing equipment
may be required to handle cells of different dimensions. This would most likely increase the
capital cost for this equipment. Alternatively, additional packaging and sealing lines might be
needed. We intend to evaluate combinations of vehicle battery packs that are easily integrated
into the same plant. The manufacturing cost will most likely increase after these considerations
are built into the model.
7.4 Stand-Alone Graphical User Interface for Model
The spreadsheet program described here-in allows versatility in designing the battery and in
calculating the costs, but like all complex spreadsheet programs it is not user-friendly to those
unfamiliar with the details of calculation. In addition, the model is easily corrupted by a poor
choice of input parameters. As a result, the final spreadsheet program will be converted to a
stand-alone user-friendly application, primarily with the efforts of Ira Bloom. Visual Basic for
Applications will be used to hard code in the model calculations and to also create the graphical
interface. The new user interface should allow for a wide distribution of the model while
maintaining the ability to change the vast majority of input parameters. The retention of this
flexibility should make the model a valuable tool for those interested in batteries regardless of
the specific material property or manufacturing cost structure the user seeks to analyze.
68
-------
REFERENCES
1. M. Anderman, F. Kalhammer, and D. MacArthur. "Advanced Batteries for Electric Vehicles:
An Assessment of Performance, Cost, and Availability." California Air Resources Board, June
(2000). Available from http://www.arb.ca.gov/msprog/zevprog/2000review/btap report.doc
(accessed on December 17, 2010).
2. B. Barnett, D. Ofer, C. McCoy, Y. Yang, T. Rhodes, B. Oh, M. Hastbacka, J. Rempel, and S.
Sririramulu "PHEV Battery Cost Assessment," 2009 DOE Merit Review, May (2009). Available
from http://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2009/energy_storage/
es_02_barnett.pdf (accessed on December 11, 2010).
3. B. Barnett, J. Rempel, D. Ofer, B. Oh, S. Sriramulu, J. Sinha, M. Hastbacka, and C. McCoy,
"PHEV Battery Cost Assessment," 2010 DOE Merit Review, June (2010). Available from
http ://www 1 .eere.energy. gov/vehiclesandfuels/pdf s/merit_review_2010/electrochemical
_storage/esOOl_barnett_2010_o.pdf (accessed on December 11, 2010).
4. A. Dinger, R. Martin, X. Mosquet, M. Rabl, D. Rizoulis, M. Russo, and G. Sticher, "Batteries
for Electric Vehicles: Challenges, Opportunities, and the Outlook to 2020." The Boston
Consulting Group, Available from http://www.bcg.com/documents/file36615.pdf (accessed on
December 11,2010).
5. L. Gaines and R. Cuenca. "Costs of Lithium-Ion Batteries for Vehicles," Center for
Transportation Research, Energy Systems Division, Argonne National Laboratory, ANL/ESD-
42, Argonne, IL May (2000).
6. F.R. Kalhammer, B.M. Kopf, D.H. Swan, V.P. Roan, M.P. Walsh, "Status and Prospects for
Zero Emissions Vehicle Technology: Report of the ARE Independent Expert Panel 2007,"
California Air Resources Board, April (2007). Available from http://www.arb.ca.gov/
msprog/zevprog/zevreview/zev_panel_report.pdf (accessed on December 17, 2010).
7. M. A. Kromer and J. B. Heywood, "Electric Powertrains: Opportunities and Challenges in the
U.S. Light-Duty Vehicle Fleet," Sloan Automotive Laboratory, Laboratory for Energy and the
Environment, Massachusetts Institute of Technology, LFEE 2007-03 RP, Cambridge, MA
(2007).
8. P. Mock, "Assessment of Future Li-Ion Battery Production Costs," presented at Plug-in 2009,
Long Beach, CA, (2009).
9. National Research Council of the National Academies, "Transitions to Alternative
Transportation Technologies - Plug-in Hybrid Electric Vehicles," The National Academies
Press, Washington, D.C. (2010). Available from http://www.nap.edu/catalog.php?record_id=
12826 (accessed on December 17, 2010).
10. TIAX LLC, "Cost Assessment for Plug-In Hybrid Vehicles (SOW-4656)," Report to US
DOE Office of Transportation Technology, October (2007).
69
-------
11. P.A. Nelson, DJ. Santini, J. Barnes, "Factors Determining the Manufacturing Costs of
Lithium-Ion Batteries for PHEVs." International Electric Vehicles Symposium EVS-24,
Stavanger, Norway, (2009).
12. DJ. Santini, K.G. Gallagher, P.A. Nelson, "Modeling the Manufacturing Costs of Lithium-
Ion Batteries for HEVs, PHEVs, and EVs," International Electric Vehicles Symposium EVS-25,
Shenzhen, China, (2010).
13. P. Nelson, I. Bloom, K. Amine, G. Henriksen, "Design modeling of lithium-ion battery
performance," Journal of Power Sources, 110, 437 (2002).
14. P. Nelson, D. Dees, K. Amine, G. Henriksen, "Modeling thermal management of lithium-ion
PNGV batteries," Journal of Power Sources, 110, 349 (2002).
15. G.L. Henriksen, K. Amine, J. Liu, and P.A. Nelson, "Materials Cost Evaluation Report for
High-Power Li-Ion HEV Batteries," Electrochemical Technology Program, Chemical
Technology Division, Argonne National Laboratory, ANL-03/05, Argonne, IL (2002).
16. P. Nelson, K. Amine, A. Rousseau, H. Yomoto, "Advanced Lithium-Ion Batteries for Plug-
in Hybrid-Electric Vehicles", International Electric Vehicles Symposium, EVS-23, Anaheim, Ca
(2007).
17. P. Nelson, "Modeling the Manufacturing Costs of Lithium-Ion Batteries for PHEVs,"
presented at Plug-in 2009, Long Beach, CA, (2009).
18. K.G. Gallagher, P.A. Nelson, and D.W. Dees, "Simplified Calculation of the Area Specific
Impedance for Battery Design," Journal of Power Sources, 196, 2289 (2011).
19. A.N. Jansen, K. Amine, and G. L. Henriksen, "Low-Cost Flexible Packaging for High-
Power Li-Ion HEV Batteries," Electrochemical Technology Program, Chemical Technology
Division, Argonne National Laboratory, ANL-04/09, Argonne, IL (2004).
20. P.A Nelson and A.N. Jansen, "Comparative Costs of Flexible Package Cells and Rigid Cells
for Lithium-Ion Hybrid Electric Vehicle Batteries," Electrochemical Technology Program,
Chemical Technology Division, Argonne National Laboratory, ANL-06/43, Argonne, IL (2006).
21. J. Euler and W. Nonnenmacher, "Stromverteilun in Porosen Elektroden," Electrochimica
Acta, 2, 268 (1960).
22. J.S. Newman and C.W. Tobias, "Theoretical Analysis of Current Distributions in Porous
Electrodes," Journal of the Electrochemical Society, 109, 1183 (1962).
23. United States Geological Survery, Commodity Statistics and Information, Available from
http://minerals.usgs.gov/minerals/pubs/commodity/ (accessed on December 17, 2010).
70
-------
24. R.H. Perry, D.W. Green, and J.O. Maloney, Perry's Chemical Engineers' Handbook, Sixth
Edition, pp: 25-68 (1984).
71
------- |