development of computer-aided design tools for automotive ...€¦ · development of computer-aided...

Development of Computer-Aided Design Tools

for Automotive Batteries-CAEBAT

Oct 30, 2012

Taeyoung Han, Vehicle Systems Research Lab, GM R&D, Warren, U.S.A. Gi-Heon Kim, Senior Research Engineer, NREL, Golden, U.S.A. Lewis Collins, Director, Software Development, ANSYS, Inc.

Team members, GM, ANSYS, ESim, NREL

• To speed up the product development cycle for batteries, GM needs to have integrated and seamless Computer Aided Engineering (CAE) tools.

• NREL has been developing several tools for electrochemical and abuse reaction modeling of lithium ion cells, but not all integrated.

• GM requires validated CAE tools for cell and pack level designs for rapid development of cost-effective and robust products for various electric-drive vehicles.

• A multi-year effort is needed to produce the validated CAE tools useful to auto industries. Teaming arrangement is essential as no one organization has all the expertise.

GM, ANSYS, and ESim each bring complimentary and needed capabilities and expertise

• The CAE tools should be flexible and open to incorporate new models and data for future cell chemistries.

• In these economic conditions, support from the US government will facilitate and speed up development of the tools. GM, NREL, ESim, and ANSYS signed 3-year contracts (2011-2014) under DOE’s Vehicle Technologies (Energy Storage Program) to develop validated and integrated battery CAE tools for automotive industries.

Background of CAEBAT

2012 Automotive Simulation World Congress

2

October 30, 2012

3

Battery Thermal Management

Battery Thermal Response

Battery Operating Condition

Battery Simulation

Heat source Electro-Chemistry

CFD Flow/Thermal

Vehicle Driving Schedule

Battery Cell Chemistry

Module Cooling Strategy


3

October 30, 2012

$44 M $16M

$16 M PHEV

HEV

Exploratory

FY2010 Budget: $76 M CHARTER: Advance the development of batteries and other electrochemical energy storage devices to enable a large market penetration of hybrid and electric vehicles.

Program targets focus on enabling market success (increase performance at lower cost while meeting weight, volume, and safety targets.)

2015 GOALS: Reduce production cost of a PHEV battery to $270/kWh Intermediate: By 2012, reduce the production cost of a PHEV battery to $500/kWh.

FY2011 Request: $96M

Vehicle Technologies Program eere.energy.gov

Energy Storage R&D at the U.S. Department of Energy Vehicle Technology Battery R&D Activities


4

October 30, 2012

Cell Level 1) Capacity utilization due to current distribution within the cell,

2) Optimum cell energy capacity in terms of electrical performance, cooling requirements, life, safety, and cost,

3) Cell degradation due to non-uniform utilization and heat generation within the cell,

4) Optimum state-of-charge (SOC) range for maximum life and safety,

5) Power requirements at low-temperature operations,

Pack Level 1) Maintain battery cells at an optimum temperature range of 25 to 35o C

under vehicle environments ranging from -40 to 50o C. Temperature uniformity less than 5o C between the cells and within the cell.

2) Cell expansion and contraction due to cycling and aging.

3) Cooling channel design, cooling concept evaluations. Currently, these evaluations and the life predictions are based on time-consuming hardware build-and-test iterations.

Why Battery CAE Tools are Important


5

October 30, 2012

6 6

CAEBAT Program

Task 4: Develop

Component Models

Task 3: Develop Cell Models

Task 2: Develop Pack Models

Task 1: Open Architecture

Software

CAD Design Layout

Thermal Models

Cost

Abuse

Abuse (Thermal Runaway,

safety)

Life Model

Electrode Design

Thermal Models

CAD Design Cathode Material Modeling

Anode Material Modeling

Electrolyte Modeling

Electrochemical couples

Performance models (voltage, current, power

prediction)

Battery management system

2012 Automotive Simulation World Congress October 30, 2012

Collaborations

Project management

•Project technical direction (Tech monitor)

• Physics based cell aging model for capacity fade and cell life

• Model order reduction methods • Tool verification and validation

• Tests for cell level validations • Math model verification and

validation • Cell level assessment • Set vehicle requirements for cell

and pack design • Pack level simulation strategy • Pack level assessment • Vehicle level validation under

various driving schedules

•Open Architecture Software

• flexible framework for multi-physics models

• cell/pack level simulation capability development

• Process automation & OAS interface


7

October 30, 2012

9

Sub-models Integrated into the Cell

Curr

ent C

olle

cto

r (C

u)

+-

Curr

ent C

olle

cto

r (A

l)

sep

Negative

Ele

ctr

ode

Separa

tor

Positiv

e

Ele

ctr

ode

L

Li+

e-

rLi

xC

6

rLi

yCoO

2

e-

Electrolyte

x

cs,e

ce

cs(r)c

s(r)

cs,e

R1(SOC,TJR)

VOCV(SOC)

C1(SOC,TJR)

Rs(SOC,TJR)

V1

RPTC(TPTC)

+

-

I(t)

V(t)

Newman, Tiedemann, Gu, Kim (NTGK)

),(

),(

TDODgY

TDODfU

Newman, Li transport model (P2D)

Equivalent Circuit Model (ECM)

Cathode Anode

Current Current

ip= Current Vectors

at Cathode plate in= Current Vectors

at Anode plate

J = Current Density

J (t, x, y, T )

Cathode Anode

Current Current

ip= Current Vectors

at Cathode plate in= Current Vectors

at Anode plate

J = Current Density

J (t, x, y, T )

• Electrochemical sub-models relate the local current density to the potential

• Cell model couples sub-models to thermal and electrical fields within the cell, integrates over multiple electrode-pairs


9

October 30, 2012

Porous Electrode Lithium Ion Battery Model

2

p eff,p 2

1

eff,p

2

eff,p

, ,2

, 2

p

2

p2

eff,p p

ε 1

lnκ

Positive Electrode

, ,1

κ

s p s p

s p

p

p

p

c cD t a j

t x

a Fjx

ca Fj

x x

c r t c r tD r

t r rr

x x

2

n eff,n 2

1

eff,n

2

eff,n

, ,2

, 2

n

2

n2

eff,n n

ε 1

lnκ

Negative Electrode:

, ,1

κ

s n s n

s n

n

n

n

c cD t a j

t x

a Fjx

ca Fj

x x

c r t c r tD r

t r rr

x x

2

s eff,s 2

2

eff,s eff,s

ε

lnκ 0

Separator:

κ

c cD

t x

c

x xx x

0.5 0.5 0.5

,max,i , , 1 2

0.5sinh BV: 2

i s s s i s s i ii

Fk c c R c R c U

RTj

x

Ln Ls Lp

Separator

Curr

ent

Co

llec

tor

Curr

ent

Co

llec

tor

Negative

Electrode

Positive

Electrode


10

October 30, 2012

Physics of Li-ion Battery System in Different Length-Scales

Li diffusion in solid phaseInterface physicsParticle deformation & fatigue

Structural stability

Charge balance and transportElectrical network in composite electrodes

Li transport in electrolyte phase

Electronic potential &current distributionHeat generation and

transferElectrolyte wettingPressure distribution

Atomistic Scale

Scale of Particles

Scale of Electrodes Scale of Cells

Scale of SystemSystem operating conditionsEnvironmental conditionsControl strategy

Scale of ModulesThermal/electricalinter-cell configurationThermal managementSafety control

Thermodynamic propertiesLattice stabilityMaterial level kinetic barrierTransport properties

Active Research Present Industry Needs

Energy Storage R&D at the U.S. Department of Energy Modeling Length Scales

Vehicle Technologies Program eere.energy.gov 2012 Automotive Simulation World Congress

11

October 30, 2012

12

Reduced Order Model • Linearized model • Non-linear model

Reduced Order Model • State Variable Model (SVM) • Proper Orthogonal

Decomposition (POD)

Empirical Model •Equivalent Circuit •Current Density (U, Y)

3D field simulation • Navier-Stokes solver • Finite Volume Method

Cell Level •Electric potential field •Current density field •Cell capacity, capacity/power fade •Cell temperature •State of Charge •Coupled with flow and heat transfer for cooling channel design

Electrode Level •Electrochemical kinetics •Solid phase Li diffusion •Li transport in electrolyte •Charge conservation & transport

1-D Field Simulation •Finite Volume Method

Agi

ng

/de

grad

atio

n

Ab

use

Models are available To be developed

Co

st

Approach for Cell Level

Cu

rre

nt C

olle

cto

r (C

u)

+-

Cu

rre

nt C

olle

cto

r (A

l)

sep

Ne

ga

tive

Ele

ctr

od

e

Se

pa

rato

r

Po

sitiv

e

Ele

ctr

od

e

L

Li+

e-

rLi

xC

6

rLi

yCoO

2

e-

Electrolyte

x

cs,e

ce

cs(r)c

s(r)

cs,e


12

October 30, 2012

13

Cell Level Sub-models

Model Pros Cons Potential

application area

Further development

P2D

Detailed electrochemistry modeling approach. Solve particle and electrode level including electrolyte

Potential to simulate local detail and cover wide range of cell chemistry

Many physical input parameters and material properties. Difficult to obtain these from the measurements.

Cell design

ROM. Need MSMD approach for cell level simulations

NTGK

Semi-empirical approach. 2 key functional parameters, U, Y. Solve electrical potential for the current collectors.

Practical 2-D approach. Non-uniform SOC, heat distribution

Unable to predict cell design changes without testing

Cell/Pack integration

Temperature effects. Transient terms. Aging, abuse conditions.

ECM

Empirical parameter fit from test data. Do not solve electric potential field.

Very simple. computationally very fast

Unable to predict cell design changes without testing

Pack optimization

Temperature effects.


13

October 30, 2012

14

MSMD (Multi-Scale Multi-Dimensional)

• Higher level domain does not require to resolve lower level geometry. Electrode domain does not need to capture the particle geometry. Cell domain does not need to conform or capture the electrode layers. •MSMD provides a significant flexibility for the cell domain modeling. • Easy to maintain the software structure and allows for various sub-models.


14

October 30, 2012

15

+ and - are the potential at the positive and the negative current collectors

T, temperatures are from the cell level domain. No energy equations in the electrode and the particle domain

Current is generated only from the particle level and applied for the electrode and the cell domain

Heat source is generated from the particle level (heat of charge transfer and mixing)

Heat source is also added in the electrode region by contact resistance and others

Heat source is also added in the cell region by ohmic resistance in the current collectors

+ and - provide to the electrode to compute potentials in the electrode interfaces


15

October 30, 2012

16

ECM was implemented as a sub-model and validated with other results.

Equivalent Circuit Model implementation

Ah

t

sOCV

Q

dttIsocsoc

tIsocC

VsocCsocRdt

dV

tIsocC

VsocCsocRdt

dV

tIsocRVVsocVtV

3600

)(

)()(

1

)()(

1

)()(

1

)()(

1

)()()()(

0

0

2

2

22

2

1

1

11

1

21

charge0

discharge0I(t)

)(tI

)(tV

1R2R

1C 2C

sR

)( socVocv

)(1 tV )(2 tV

)(tI

3.00

3.10

3.20

3.30

3.40

3.50

3.60

3.70

3.80

3.90

4.00

4.10

4.20

0 2000 4000 6000 8000 10000

Ce

ll P

ote

nti

al,

V

Time, seconds

320 mA discharge/charge curves for a 850-mAh capacity cell

charge

discharge

discharge-Simplorer

EC Model: Chen & Rincon-Mora (2006)


16

October 30, 2012

17

Current Density Model

),(

),(

TDODgY

TDODfU

Negative potential

Positive potential

Temperature 3-D Finite Volume

NTGK Model implementation

U and Y are fitting parameters depend on Depth of Discharge (DOD) and Temperature


17

October 30, 2012

18

•P2D particle resolution study (non-uniform radial grid) •Findings will be built into deliverables, providing automation for non-experts

Researching Simulation Best-Practice

1.10E+04

1.15E+04

1.20E+04

1.25E+04

1.30E+04

1.35E+04

1.40E+04

1.45E+04

1.50E+04

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Li c

on

ce

ntr

ati

on

at

so

lid

-ele

ctr

oly

te

inte

rfac

e C

s_

e, m

ol/

m3

r/R

N=35, p=1.0

N-70,p=1.0

N=30, p=0.9

N=20, p=0.8

N=10, p=0.6

1.10E+04

1.15E+04

1.20E+04

1.25E+04

1.30E+04

1.35E+04

1.40E+04

0.9 0.92 0.94 0.96 0.98 1

Li c

on

ce

ntr

ati

on

at

so

lid

-ele

ctr

oly

te

inte

rfac

e C

s_

e, m

ol/

m3

r/R

N=35, p=1.0

N-70,p=1.0

N=30,p=0.9

N=20, p=0.8

N=10, p=0.6

10C discharge rate


18

October 30, 2012

19

Cell level •verification •validation

Standard battery input parameters • Geometry, • Materials, • Physical parameters,

Other sub-models

FLUENT

Operating conditions • Current (t) • Voltage (t) • Power (t) Output

• Terminal V(t), I(t), P(t) • Non-uniform Q • SOC • Cell temperature • Coolant temperatures •

Electrode sub-models • P2D (Full) • NTG semi-empirical • ECM (equivalent circuit)

Build Interface for other

sub-models

Overview of Cell Level Simulation Tool

The first cell level tool will includes 3 sub-models with a standard battery user interface and open interface for other sub-models. Cell level verification and validation will be performed.


19

October 30, 2012

21

Battery Pack (Chevy Volt)

Coolant loop

Micro-channel fin

Buss bar

Voltage, current, Temperature sensing

• Model coolant flow loop including micro channels. • Model electrical wire connections and circuit including buss bar connection. • Include structural members for cell expansion and contraction.


21

October 30, 2012

22

Battery Pack Thermal Design

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed

Vehicle Driving ProfileVehicle Design

Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

Battery Power Profile

Multi-Scale

Multi-Dimensional

Li-ion Battery Model

Battery ResponsesFeedback

Embedded

Battery Model

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

CD CS

Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

CD CS

Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Cell / Pack level Design/Simulation

Vehicle Simulatorwith a simple battery model

Battery Power ProfileVehicle Driving Cycle

Feed Back

•Cell chemistry •C-rate • SOC • Temperature •Aging


22

October 30, 2012

23

Cell to Pack Level Simulation

50 ~ 150 ODE Eqns.

10 ~ 12 cells per module

8 ~ 10 Modules per pack

1-D Newman Model

More than 1011 degrees of freedom !!!

Pack

Electrodes Cell Module

10 ~ 20 M volumes including cooling channel

Current HPC computers can not handle the detailed modeling approach. Requires Reduced Order Model (ROM) approaches or simpler models for the pack.


23

October 30, 2012

24

Pack Level Strategies

Model Pros Cons Potential

application area

Further development

Co-simulation

CFD-based cell models iteratively coupled to network/circuit pack model

Most accurate, no need for ROM-building, Earliest availability for testing

Computational cost (but can exploit periodicity, sampling, module hierarchy to improve over brute force approach)

Totally new designs, validation

Asynchronous time stepping, Automated user interface

Transient ROM

Build time-varying cell ROM from CFD, then use in pack-level simulation

Most efficient

Linearized about some particular values, e.g. coolant flow rate, state of health, etc.

Drive cycles, parametric studies

ROM algorithms, Storage, Results expansion

System level

model

Construct system level models from the full pack simulation model.

Very fast and efficient

Solution available only at monitored locations. Requires full 3-D pack-level flow/thermal solution.

Battery management system, Drive cycles

Non-linearity due to variable flow rates


24

October 30, 2012

Approach for Pack Level Simulation

Cell Level Model •Reduced Order Models for electrochemistry •Cell level performance including local cooling channels

System Level Model •Construct a “linear” or “non-linear” system simulation model from the full pack simulation model

• Strategy is to offer a range of methods allowing analysts to trade off computational expense vs. resolution

Reduced-Order Models •Reduced order models for flow and thermal analysis at the pack level •Reduced order cell models •Ability to “expand” results

Pack Level Model


25

October 30, 2012

27

Model Credibility

Difference between math results and physical test data

Math Model Validation

1. Determine evaluation criteria 2. Specify an input/uncertainty map 3. Design of validation experiments 4. Data collection and generate model outcomes 5. Compare model and test results 6. Feedback into current validation exercise and feed-

forward into future validation activities

Statistically Based Six-step Process at GM (ASME PTC60 guide & NISS)

Verification: Solving the equation right Validation: Solving the right equation


27

October 30, 2012

28

Test Control and Data Acquisition

System

DC Power Supply

Electronic Load

Charge Current Control

Discharge Current Control

Current monitor

Current monitor

Dis

cha

rge

Ch

arg

e

Voltage

Cell skin and ambient Temperature

Thermal Chamber

IR Camera

Window

Lithium cell

Cell Level Test System with Thermal ChamberCell Level Test & Validation by GM

Cell level validation test for electrical and thermal performance for the cell

Thermocouple, thermal imaging, calorimeter test to measure the total heat generation and non-uniform heat source

Cycle life test for aging model

Test Test Conditions

Cell relaxation test

Environmental Temperature range: -20, -10, 0, 25, 40 Deg C C rate: 0.5C, 1C, 3C and 4C

Static capacity and HPPC test

Environmental Temperature range: 0, 10, 25, 40 Deg C

IR thermal imaging 10 to 90% SOC, room temperature, 1C, 3C and 4C

Cell cyclic life test

Environmental Temperature range: 0, 10, 25, 40 Deg C C rate: 2C, 3C SOC window: 30% to 90%

Cell Calorimeter Same test condition cell cycling test


28

October 30, 2012

29

Cell Cycle Life Test Procedure

Temperature Monitor Point Test Cells in Thermal Chamber 2012 Automotive Simulation World Congress

29

October 30, 2012

30

Pack Level Validation

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

CD CS

Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

CD CS

Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Power Profile

Vehicle Driving Cycle

CAN

Complete pack

Charge/Discharge

Coolant flow rateTemperature

Thermal Chamber

Climatic profile

Thermocouples

Validation Parameters •SOC, Cell Voltage, Cell Temperatures •Coolant flow Temperature in & out •Total heat generation

Pack level test •Driving cycles •Environmental temperatures •Thermocouple location •Coolant flow rate & temperature •Cell chemistry

Pack level validation test data will be carefully selected from the existing GM battery Test database.


30

October 30, 2012

31

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

Battery Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

CD CS

Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Virtual

Battery Design

Vehicle Simulator

Vehicle Speed

-10

0

10

20

30

40

50

60

70

80

90

0 100 200 300 400 500 600

Time (s)

Vehicle Speed


Virtual

0 10 20 30 40 50 60-600

-400

-200

0

200

400

600

800

CD CS

Power Profile

Multi-Scale

Multi-Dimensional



Embedded

Battery Model

Cell / Pack level Design/Simulation

Vehicle Simulatorwith a simple battery model

Battery Power ProfileVehicle Driving Cycle

Feed Back

Integration with Vehicle Simulator

CAEBAT program will be interfaced with the vehicle driving cycle simulator


31

October 30, 2012

Deployment to Industry

32

WorkbenchFLUENT

fluid-thermal-chemical-electrical

Mechanical

structural

Other Field SolversFatigue, crack propagation,

nail penetration

Meshing

Design Modeler

Design Xplorer

Simplorer

system

simulation

Connections to other

CAEBAT tools/data

Connections to

CAD tools

Legend: first produced in this project restricted computer software

Battery Design Scriptsworkflow customization, automation,

embedded knowledge

Material properties Physical input parameters

ROM methods

Electro-

chemistry

EKM model/data mgmt.

Standard ANSYS products

2012 Automotive Simulation World Congress October 30, 2012

33

Curr

ent C

olle

cto

r (C

u)

+-C

urr

ent C

olle

cto

r (A

l)

sep

Negative

Ele

ctr

ode

Separa

tor

Positiv

e

Ele

ctr

ode

L

Li+

e-

rLi

xC

6

rLi

yCoO

2

e-

Electrolyte

x

cs,e

ce

cs(r)c

s(r)

cs,e

Present Barriers

Enabling Batteries for Next Generation Green Vehicle Technology

Battery cost, life, & safety

• Improved battery performance, life, and safety • Improved competitiveness of US auto industries and

support for National Energy Independence

Project Impacts

Battery design cycle is slow & prototype-driven development ($$$)

• Virtual battery development tool will speed up product development process, reduce cost with improved quality, and reduce warranty cost.

• Deliver battery multi-physics software package. • Virtual battery and thermal management design

capability and robust engineering.

Existing engineering tools are impractical for realistic battery design and optimization


33

October 30, 2012

Future Work

• Develop model order reduction methods for the pack level

• Extend cell-level models for aging and abuse

• Cell level verification and validation

• Pack level verification, validation, and demonstration

Define pack level validation requirements to meet the future capability matrix for pack level CAE performance.

Identify suitable existing pack level test in progress or from previous tests (Liquid or Air cooling) performed in GM battery group.

Build up the pack level simulation model including meshing and physical boundary conditions, operating conditions.

• Develop battery-specific graphical user interface for workflow automation

• Build a standard data-exchange interface based on specifications from the OAS Workgroup


34

October 30, 2012

development of computer-aided design tools for automotive ...€¦ · development of computer-aided...

Documents