robert ratz bms system benchmark and standardization · pdf filebms system benchmark and...
Post on 27-Mar-2018
413 Views
Preview:
TRANSCRIPT
1 © Ricardo Inc 2015 March 2015
BMS System Benchmark and Standardization
Robert Ratz
2 © Ricardo Inc 2015 March 2015
Delivering excellence through experience
Providing technology, product innovation, engineering solutions and
strategic consulting to the world’s automotive industries since 1915.
Today
1915
Ricardo is a global, world-class, multi-industry consultancy for
engineering, technology, innovation, strategy and environment.
3 © Ricardo Inc 2015 March 2015
As with all analysis, approximations in models, loading and interpretation exist, and real components may vary in geometry and properties from those assumed. Hence analysis results should be used only as a tool for guiding the design process and
for reducing risk of poor performance during development and validation testing. Where analysis results are described as acceptable, it is implied only that the risk of poor performance is considered acceptable to proceed to the next stage of the
design and development process, unless otherwise stated. While the analysis may reduce the amount of testing required to prove a design, it should not be considered a replacement for validation testing unless specifically stated by Ricardo.
Battery Management System
Agenda
• Ricardo Vehicle Benchmarking and Analysis
• BMS System Overview
• Types of Cell Balancing
• BMS System Benchmark Comparison
– Architecture Comparison
– Battery size
– Cell Balancing
– Isolation methods
– Voltage sense & battery balance
– Microcontrollers
• Motivation for a standardized BMS system specification
4 © Ricardo Inc 2015 March 2015
Fourteen xEVs have been benchmarked over the last four years by Ricardo
High Voltage
Battery Pack
DC DC
Converter
DC AC
Inverter
Onboard
Charger
System
High Voltage
Junction Box
High
VoltageHarn
Harnes
HVAC and
Cooling
Systems
Transmission
/ EV Drive
System
Ex
am
ple
Sc
he
ma
tic
s
Ba
ttery
Pack
Th
erm
al
Ba
ttery
Pack E
lectr
ical
Introduction to EV/Hybrid Analysis Benchmarking
xEVs Benchmarked (note driveline weights shown)
5 © Ricardo Inc 2015 March 2015
Benchmark activities include in-depth analysis and studies of all major EV systems
Thermal Schematic Cell Characterization BMS Analysis
Exploded Views Battery Schematic Steel Analysis System Schematic
2D CAD Drawing Circuit Analysis Performance Testing 3D CAD Modeling
Analyses performed (sample)
Introduction to EV/Hybrid Analysis Benchmarking
6 © Ricardo Inc 2015 March 2015
BMS System Overview
BCM (Battery Controller Module)
• Current Sensing
• Voltage sensing
• Temperature sensing
• CAN Communications
• Battery Charger control
• Output drivers for pumps, fans, and heaters
• Contactor control:
– Main Contactors
– pre-charge
– Charger Contactors
• State of Charge algorithms
• State of health algorithms
• Power capability algorithm
• Safety feature
– Performs leakage detection and contains
ground fault interrupt functionality
– Self test of all critical functionality on every
“key-on” event
– Diagnostics of key functions
7 © Ricardo Inc 2015 March 2015
BMS Architectures – Distributed
Distributed BMS Architecture is split into
one main controller and multiple slave
boards
Battery Controller Functions:
• Battery charger control
• Output drivers for pumps, fans, and
heaters
• Typically Contactor control
• State of charge algorithms
• State of health algorithms
• Power capability algorithm
• Safety feature
Slave Board (Voltage & Balance PCB)
• Voltage /Temperature
• Cell Balancing
• Communicates to main controller
Battery module
#1
Battery module
#2
)
Battery module
#3
Battery module
#4
Bat
tery
pac
k sy
stem
Traction voltage
Cell monitoring and current
balancing circuits #1
(Red PCB)
Cell monitoring and current
balancing circuits #2
(Blue PCB)
Cell monitoring and current
balancing circuits #3
(Blue PCB)
Cell monitoring and current
balancing circuits #4
(Green PCB)
CAN
Sys
tem
CA
N b
us
lines
CAN
CAN
CAN
CAN BMS master ECU
Pros
• Highly Scalable and modular
• Less wiring
Cons
• Several PCBs
8 © Ricardo Inc 2015 March 2015
• A Centralized BMS has all of the functions in a
single module/PCB.
• Centralized architectures tend to be on smaller
packs (exception is the Nissan Leaf)
Centralized BMS
Pros
• One PCB and reduced
electronic costs
Cons
• Wiring intensive
• Not scalable
Master CPU
(uPD70F3236BM)
Battery monitor #1
(DS15110)
Ga
lva
nic
is
ola
tio
n
(op
toc
ou
ple
rs)
Battery monitor #24
(DS15110)
. . . . . . . . . . . . . . . . .
22
x D
S1
51
10
in d
ais
y-c
ha
in
Battery
module
(4 cells)
Battery
module
(4 cells)
Daisy-chain
serial
communication
interface
Battery monitor #25
(DS15110)
2x96 cells in
parallel
configuration
Battery
192 cells
CAN
driver
Co
nn
ec
tor
9 © Ricardo Inc 2015 March 2015
Cons
• More Complex
• More expensive
Passive Balancing
• Resistive discharging of higher voltage
cells to the lowest denominator
Pros
• Integrated ASICs
• Least expensive
Active Balancing
• Charges a low cell from a higher cell
Pros
• Quicker to balance
• Capable of nurturing
a weak cell to prolong life
• Recovers energy
Cell Balancing
VDD
CB8
IN8
IN7
Voltage measurement circuit and
cell balancing FET for each cell
3.75V
. . . . .
Cell 1
3.75V
Cell 8
6x same block
IN1
CB1
IN0VSS
Mux ADCSPI
interfaceto the
CPU
30R
30R
L9763
IN2
IN3
IN4
IN5
IN6
5V LDO
5V SMPS
+5V
+5V
CB1
CB8
Cell balancing
control circuits. . . . .
Daisy-chain
interfaceto other
L9763
IN1
CB1
IN0
VSS
Mux Serial
interface
Charge
pump with
switch
matrix
2P25 SF367
IN2
IN3
1.2V
1.2V
CB2
CB3
CB4
VDD
Current
balancing
control logic
ADC
Unknown
interface
Connect to next
2P25 SF367 in
daisy-chain
CB5
CB6
IN4
IN5
IN6
to next 6 cells
to next 6 cells
6 c
ells
in
se
rie
s
CB0
Charge balancing
capacitor
Cons
• Slower Balancing
• Wastes energy to heat
10 © Ricardo Inc 2015 March 2015
BMS Benchmark and Analysis
• Analyzed BMS PCB
• Determined PCB functional partitioning
• Developed system diagrams
2011
Chevrolet
Volt
2011
Nissan
Leaf
2012
Toyota
Prius Plug In
2011
Mitsubishi
iMiEV
2013
Chevrolet Malibu
Eco
2013
Ford
Fusion
Master CPU
(R5F7142)
Battery monitor #1
(HCC03-LLV1018)
Galvanic isolation (optocouplers)
Battery monitor #2
(HCC03-LLV1018)
Battery
module
(6 cells)
Battery
module
(6 cells)
Battery
16 cells
CAN
driver
Co
nn
ec
to
r
Battery monitor #3
(HCC03-LLV1018)
Battery
module
(4 cells)
Battery monitor #6
(HCC03-LLV1018)
Battery monitor #5
(HCC03-LLV1018)
Battery
module
(6 cells)
Battery
module
(6 cells)
Battery
16 cells
Battery monitor #4
(HCC03-LLV1018)
Battery
module
(4 cells)
Daisy-chain
synchronous serial
communication
interface
Ground
fault
detection
+115V
Daisy-chain
synchronous serial
communication
interface
Daisy-chain
synchronous serial
communication
interface
Daisy-chain
synchronous serial
communication
interface
Synchronous serial
communication
interface
Synchronous serial
communication
interface
Battery monitoring
circuits block #1
Ga
lv
an
ic
is
ola
tio
n (o
pto
co
up
le
rs
+
o
pto
MO
S)
Battery
(8 triplets)
Battery pack #1
(30 cell triplets)
CAN
driver 1
Co
mm
un
ic
atio
n in
te
rfa
ce
c
on
ne
cto
r (b
la
ck
)
Battery monitoring
circuits block #2
Battery
(8 triplets)
Battery monitoring
circuits block #3
Battery
(8 triplets)
Battery monitoring
circuits block #4
Battery
(6 triplets)
CAN
CAN
driver 2
CAN
driver 3
CAN
driver 4
CAN
CAN
CAN
Status signals
Control signals
Cell monitoring
circuits #1
Cel
l mon
itori
ng c
ircu
its #
2
Cel
l mon
itori
ng c
ircu
its #
3
Cel
l mon
itori
ng c
ircu
its #
4
Isol
atio
n ba
rrie
r
CAN interface
Isolation barrier
Isolated
interface
IN1
CB
1
IN0
VS
S
Mux
Serial
interface
Integrated
4x FE
T
switches
DS
15110
IN2
IN3
3.75V
3.75V
3.75V
3.75V
430R
430R
CB
2
CB
3
CB
4
VD
D
430R
430R
Current
balancing
control logic
AD
C
UA
RT
interface
Connect to
next DS
15110
in daisy-chain
To next battery
stack in series
To next battery
stack in series
Master CPU
(uPD79F0121A)
Battery monitor #4
(2P25 SF367)
Galvanic isolation (optocouplers)
Battery
module
(6 cells)
Daisy-chain
communication
interface
Battery
module
Daisy-chain
communication
interface
Daisy-chain
communication
interface
Daisy-chain
communication
interface
Ground
fault
detection
Battery monitor #3
(2P25 SF367)
Battery monitor #2
(2P25 SF367)
Battery monitor #1
(2P25 SF367)
Daisy-chain
communication
interface
Battery
module
(6 cells)
Battery
module
(6 cells)
Battery
module
(6 cells)
Battery monitor #5
(2P25 SF367)
Battery
module
(6 cells)
Battery monitor #6
(2P25 SF367)
Battery monitor #7
(2P25 SF367)
Battery monitor #8
(2P25 SF367)
Battery
module
(6 cells)
Battery
module
(6 cells)
Battery
module
(6 cells)
Daisy-chain
communication
interface
Battery
module
Next module Previous
module
V+
S8
C8
C7
Vo
ltag
e m
ea
su
rem
en
t circ
uit a
nd
ce
ll ba
lan
cin
g F
ET
for e
ach
ce
ll
3.7
5V
. . . . .
Ce
ll 1
3.7
5V
Ce
ll 8
6x s
am
e b
lock
C1
S1
V-
Mu
x
DS
AD
C
12
.bit
Co
ntro
l
circ
uits
41
R
41
R
LT
C6
80
2-2
C2
C3
C4
SP
I
inte
rface
Re
fere
nce
vo
ltag
e
To
CP
U
AD7280
Master CPU
(MPC5534MVZ80)
Pack
of 5
cells
. . . . . .
6x
AD7280Pack
of 6
cells
. . . . . .
7x
. . . . . . . . . . .
5x battery pack
in a daisy-chain
. . . . . .
5x monitoring
circuits AD7280
SPI(1)
SPI(7)
VSS(1)
VSS(7)
AGND7
SPI
6x6 cells (36 in total) monitoring and signal
processing in this block
AD7280Pack
of 6
cells
. . . . . .
7x
AD7280Pack
of 6
cells
. . . . . .
7x
. . . . . . . . . . .
4x battery pack
in a daisy-chain
. . . . . .
4x monitoring
circuits AD7280
SPI(8)
SPI(13)
VSS(N)
VSS(0)
AGND13
Ga
lva
nic
is
ola
tio
n
Ad
UM
14
00
+ A
Du
M5
40
3
SPI
5x6 cells and 2x5 cells (40 cells in total) monitoring and signal
processing in this block
SPI
Ga
lva
nic
is
ola
tio
n
Ad
UM
14
00
+ A
Du
M5
40
3
• Analyzed subsystems
• Voltage & Temperature sensing
• Balancing methods & IC’s used
11 © Ricardo Inc 2015 March 2015
Car model Year BMS architecture kWhr
Ford Fusion
Hybrid 2.0
2013 Centralized on the single PCB 1.4
Chevrolet
Malibu ECO
2013 Centralized on the single PCB 0.5
Chevrolet Volt 2011 Pseudo-Distributed 4 PCB
modules + BMS master ECU
16
Mitsubishi
I-MiEV
2012 Distributed 11 PCB modules +
BMS master ECU
16
Nissan Leaf 2011 Centralized on the single PCB 24
Toyota Prius 2012 Distributed 4 PCB modules 4.4
Ricardo
McLaren P1
2013 Distributed 1 BCM + Several
VTBMs
4.8
There Are Two Main Architectures
Centralized
Distributed
Master
controller
Slave PCBs
12 © Ricardo Inc 2015 March 2015
Car model Battery voltage Battery
cells
Ricardo McLaren P1 550V 324
BMW i3 360V 96
Tesla Model S60 300V 5,376
Ford Fusion Hybrid 275V 76
Chevrolet Malibu ECO 115V 32
Chevrolet Volt 360V 288
Mitsubishi I-MiEV 330V 88
Nissan Leaf 360V 192
Toyota Prius PlugIn 207V 56
Battery Comparison
Ford Fusion Hybrid Nissan Leaf Chevy Volt Prius Plug-In Hybrid
13 © Ricardo Inc 2015 March 2015
Car model Cell balancing method Cell balancing
technique
Cell balancing
current1
Ford Fusion Hybrid Passive Resistive 17 mA
Chevrolet Malibu ECO Passive Resistive 17 mA
Chevrolet Volt Passive Resistive 125 mA2
Mitsubishi I-MiEV Passive Resistive 92 mA
Nissan Leaf Passive Resistive 8.7 mA
Toyota Prius Active Charge pump Unknown
Ricardo McLaren P1 Active & Passive Charge transfer &
resistive
1Amp Active
250 mA Passive
Battery Cell Balancing Comparison
1 Estimated cell balancing current given at nominal cell voltage 2 Estimated cell balancing current given at nominal cell voltage for cell triplet
Cell Balancing
14 © Ricardo Inc 2015 March 2015
Isolation Method
AD7280
Car model Isolation Method Chip
Ford Fusion
Hybrid Digital isolators
Analog Devices
ADuM5403
Chevrolet
Malibu ECO Optocouplers NEC - PS9121
Chevrolet Volt Optocouplers Avago ACPL-M43T
Matsushita AQW212S
Mitsubishi I-
MiEV
Digital isolators
Analog Device-
ADuM1402W
Nissan Leaf Optocouplers NEC - PS9114
Toyota Prius ????
Renesas &
Matsushita
PS9121
216C020
Ricardo
McLaren P1
OptoMOSFETs
Digital Isolators NEC PS710E
15 © Ricardo Inc 2015 March 2015
Voltage & Balance ASICs
AD7280
Car model ASIC # cells Communications
from ASIC to Micro
Ford Fusion Hybrid Analog Devices
AD7280 6 Cells SPI Daisy Chain
Chevrolet Malibu ECO Hitachi-
HCC03LLV1018 6 Cells Serial Daisy chain
Chevrolet Volt STMicro
L9763 8 Cells SPI Daisy Chain
Mitsubishi I-MiEV Linear-
LTC6802G-2 12 Cells
SPI Daisy Chain
Nissan Leaf NEC
DS15110 4 Cells
UART interface
Toyota Prius ????
2P25 SF367
4 Cells
Serial Daisy chain
Ricardo McLaren P1 Microcontroller 14 Cells NA
IN1
CB1
IN0
VSS
Mux Serial
interface
Charge
pump with
switch
matrix
2P25 SF367
IN2
IN3
1.2V
1.2V
CB2
CB3
CB4
VDD
Current
balancing
control logic
ADC
Unknown
interface
Connect to next
2P25 SF367 in
daisy-chain
CB5
CB6
IN4
IN5
IN6
to next 6 cells
to next 6 cells
6 ce
lls in
ser
ies
CB0
Charge balancing
capacitor
16 © Ricardo Inc 2015 March 2015
BMS Primary CPU
Car model BCM CPU
Ford Fusion Hybrid Freescale
MPC5534MVZ80 32 Bit
Chevrolet Malibu ECO Renesas
R5F714264FPV 32 Bit
Chevrolet Volt Freescale
S9S08DZ32 32 Bit
Mitsubishi I-MiEV NEC
F3612M2 32 Bit
Nissan Leaf Renesas
PD70F3236BM 32 Bit
Toyota Prius Renesas -
PD79F0121A 32 Bit
Ricardo McLaren P1 Freescale
MC9512XPE 32 Bit
17 © Ricardo Inc 2015 March 2015
Car
model
Vo
lta
ge
Se
nse
Tem
pera
ture
Sen
se
Pa
ss
ive B
ala
nc
e
Ac
tive
Ba
lan
ce
Iso
lati
on
Leakag
e D
ete
cti
on
Co
nta
cto
r D
riv
ers
Ac
ces
so
ry D
riv
es
Ch
arg
er
inte
rfac
e
CA
N
Co
mm
un
icati
on
Vo
lta
ge
< 2
50
V
CP
U 3
2 B
it
Dis
trib
ute
d
Ce
ntr
alize
d
Ford Fusion
Hybrid X X X X X X X X X X X X
Chevrolet
Malibu ECO X X X X X X X X X X X
Chevrolet
Volt X X X X X X X X X X X X
Mitsubishi I-
MiEV X X X X X X X X X X X X
Nissan Leaf X X X X X X X X X X X X
Toyota Prius X X X X X X X X X X X
Ricardo
McLaren P1 X X X X X X X X X X X X X
Architecture Functions Summary
18 © Ricardo Inc 2015 March 2015
• A global standard for BMS will further reduce high cost energy storage systems
– BMS electronic cost in the range of $500 to $1500 per battery Target $50
• BMS systems in use today have similar architectures and functions
• A BMS standard will improve safety and image of HEV/EV market
• A single specification can cover the two approaches to the architectures:
– Distributed
– Centralized
• BMS’s system hardware have 90% of common functions
– ASICs have common functions of voltage & temperature monitoring and
passive balancing
– Balancing is predominantly Passive with a couple of applications using Active
– Isolation is predominantly Opto-isolators
– Microcontroller are 32 bit
• Most of the IP is software related
• A common standard would further enable second life use of battery systems
Rational for a Standardized BMS System
19 © Ricardo Inc 2015 March 2015
top related