Viewing the Battery Management System Through Many Lenses Martin Klein LG Chem Power, Inc. 2011 PHM Conference Battery Management System Workshop September 26, 2011
Viewing the Battery Management
System Through Many Lenses
Martin Klein
LG Chem Power, Inc.
2011 PHM Conference
Battery Management System Workshop
September 26, 2011
Many Perspectives
• Protect the cells
• Calculate SOx quickly, accurately
• Inform the driver
• Robust design
• Safety
• Low cost
• Delivered on time
• Small, lightweight
Battery Management System Key Functions
Battery Management System Environment
Battery Management System Environment
Battery Scientist: Protect the Cells!
Cells don’t like– Over-charge– Over-discharge– Over-current– High temps– Operating at Low Temps
– ��������� (power drain)
Functions of Interest• Monitoring: V, I, T• Power limits �, , ���• State of Health• Thermal management
– Fan, pump speed– Heaters– Chillers
• Cell balancing• Isolation detection
Controls Engineer
Multiple Loops In Play:
- Measurements / Monitoring / Reporting- Cell voltage, Battery current
- High Voltage Interrupt Loop
- High Voltage Isolation
- Coolant Temperatures (Inlet/Outlet, Cells)
- Circuit Board Temperatures
- External commands (e.g., contactor control, crash sense)
- Weld Checks, Breaking Element Detection
- Cell Balancing
- SOx-Related Algorithms
- State of charge, power, energy, health
- Diagnostics & Rationality
Commands:
- Contactor Open / Close
Other
- Off-line diagnostics, reflash, test mode
- BMS Power Up, Power Down Sequencing
Highly
synchronized
Accuracy, speed
requirements vary with
application (HEV, PHEV, EV)
SOCA Modeling and HIL ValidationDeep-Dive State-of-Charge Algorithm validation in progress
Exercising SOCA to limits of battery pack operation Considering all modes of operation
RTC
Data Initialization
Algorithm (state) InitializationSOC Estimation
Hysteresis
Filter States
ResistanceCapacity
Bias
SOC Estimation
Hysteresis
Filter StatesResistance
Capacity
Current Bias
Robustness to voltage sensors stuck at low, nominal, highRobustness to temperature sensors stuck at low, nominal, high
Robustness to current sensor stuck at low, nominal, high
SOC Estimation
Hysteresis
Filter StatesResistance
Capacity
Bias
SOC EstimationFilter States
HysteresisResistanceCapacity
Current BiasRobustness to SOC Initial Robustness to Cell BalancingRobustness to Temperature
Robustness to Cell Age
Startup mode
Normal operation
Fault mode
Shut-
down
8
Software TasksTop 3 (by HLF Requirements Count):
- Diagnostics: 22%
- Hardware Interface: 12%
- Cell & Pack Energy Mgmt: 12%
MEMO: Code releases
often proliferate to support
multiple vehicle builds!
Electronics Engineer• Electronics Design & Development
– Engineering Costs
– Part Count, Part Cost
– Change Management
– Hardware/Software Compatibility
– Component availability (wrt global shortages, Japan disaster)
– Package Space, connectors, enclosures
• Validation
– Environment• Temperature, Humidity, Vibration, Shock
• Electromagnetic Compatibility
• Electrical Overstress (e.g., ESD)
– Ability to test
Battery Energy
Control Module
(“Master”)
Battery Power
Slave Module
(“Slave”)
OEM Purchasing / Buyers
• The BMS is pure overhead:
– It does not store energy; yet
– It costs a lot of money,
– It takes us space, and
– It has mass
• “Just another piece of electronics”
Design & Development Strategy
Ce
llP
rote
ctio
n
Co
mp
uta
tio
n
Op
tim
iza
tio
n
EE
HW
De
sig
n
Ro
bu
stn
ess
SW
De
sig
n
Ro
bu
stn
ess
Siz
e &
We
igh
t
Tim
e t
o L
au
nch
Pie
ce-C
ost
Ma
na
ge
me
nt
ED
&D
Co
st
Ma
na
ge
me
nt
Move all but cell protection, battery monitoring to
vehicle-level controller + + + + + + +Repeating, common electronic circuits should be
book-shelved for re-use on multiple programs + + + +Next Generation ASICs to in increase circuit
integration (and reduce part count) + + + + + +
Highly modular SW - + +
Hardware-independent SW - +Make SOx only as complex as necessary for the
intended roles (i.e., HEV, PHEV, EV) + + + + +Make HILs an integrated part of the Validation
Process + + +
BMS Design Strategies for Satisfying Multiple Demands
Design Strategies for Satisfying Multiple Demands
Battery Architecture: Minimize BMS Responsibilities
• Keep – Cell-related functions with battery
• V,I,T monitoring
• SOx algorithms, power limits,
– Communications (e.g., CAN)
– Minimum diagnostics; e.g., HVIL, ∆V, ∆T
– Minimum data recording functions
• Move other functions to vehicle controller– Link-side voltage measurement
– Contactor Control / Breaking Element Detection
– Thermal Controls (Fans, Pumps, etc.)
…and Diagnostics will follow!
Design Strategies for Satisfying Multiple Demands
• Hardware Architectures
– Maximize circuit reuse; minimize or simplify pack-
specific circuits
– Seek to increase manufacturing volumes of
common PCBs; eliminate low-volume HW variants
– Off the shelf commercial ASICs
Design Strategies for Satisfying Multiple Demands
• Software Architecture
– Move cell-specific parameters into calibratable tables
– Core / Apps partition, with book-shelved SW modules
– SOx algorithms as complex as necessary, but no more
– Hardware-in-the-Loop should be fundamental element of validation
Challenges in BMS SW Testing + Development
• Simulating accurate cell voltages
• Automating diagnostics testing
• Next generation ASIC decisions
• Cell balancing and State of Health – what is the optimal cell balancing strategy?
• SOC accuracy @ cold temperatures(-20°C) – For example: Coulomb-counting alone at cold
ignores unreliably high OCV’s.)
• SOx and Power Management are critical BMS functions…but
not the only functions
• To make the BMS affordable and to reduce the time to
market, focus on the functions most critical to enhancing cell
performance and life
• Electronic hardware drives piece cost and a good part of
development costs
• Book-shelf and modularize to increase re-use of circuits and
SW to drive up volumes and drive down costs
• Many challenges remain to satisfying all users
Summary
Thank You