PHEV Vehicle Level Control StrategySSummary
November 2008November 2008
Aymeric RousseauyArgonne National Laboratory
Sponsored by Lee Slezak
This presentation does not contain any proprietary, confidential, or otherwise restricted information
Objective: Assess Fuel Displacement Potential of Various PHEVs Control StrategiesVarious PHEVs Control Strategies
Higher Electric Power
Higher Electric Energy Higher Control Freedom
Fuel Savings Potential
Depending on various driven distance, several possible modes:– Below AER (All‐Electric Range) : Electric‐only (EV)
– Above AER: EV? CS(Charge Sustaining)? Blended?
Common Strategy= EV+CS Blended Strategykm/h
SOC
% %km/h
lowICE
highICE
offoff
Veh. speed
SOChighICE
off
highICE
off off
s
EV MODE
off
CS MODE
s
BLENDED BLENDED
2
Control Optimization/Design Investigates Strategies Potential
Innovative 3-way Approach to Control OptimizationGl b l O ti i ti
tSOCtP
tJtSOCtP '0,000 ,1',1 tU 0 Optimal Control
Teng Tmcωmc
Global Optimization
tSOCtP nm ,
tJtSOCtP MMMM ',' ,1,1
0
tU M
TX 0X.
.
.
.
.
.
Optimal Control, Minimal Fuel Cons.
Backward model (not PSAT) Bellman Principle
ωeng
Backward model (not PSAT) p
PSAT Various PSAT C t l St t iVarious
Control Design
Control StrategiesVarious Control Principles Rule-Based Control Design
Heuristic OptimizationOptimally tuned
ParametersPSAT
Heuristic Optimization
3
Existing Control Logic DIRECT Algorithm
Global Optimization Showed Minimal Fuel Consumption Achieved in Blended Mode
Global Optimization
Consumption Achieved in Blended Mode
0 9
1
UDDS 10 ( ti )Japan 10-15 x25 (optim)Japan 10-15 x25 (simu)
0 7
0.8
0.9NEDC x10(optim)UDDS x10 (optim)
3 cycles demonstrated
0.6
0.7
SO
C
Blended strategy is optimal
0.4
0.5
Example of EV+CS Mode for
0 10 20 30 40 50 60 70 80 90 1000.2
0.3Example of EV+CS Mode for comparison
4
% of total distance
Control Strategy Design Showed Significant ImprovementsImprovements
EV/C
S
0 0%
2.0%
Up to 9% less
ase
Ove
r
-2.0%
0.0%0 16 32 48 64 80 96 112
fuel consumed when driving distance is 20 migy
Incr
ea
-6.0%
-4.0%
Differential EnginePower distance is 20 mi
uel E
nerg
-8.0%
Full Engine Power
Optimal EnginePower
Trip distance (km)Fu -10.0%
Power Split 10 miles AER (Prius), run on several UDDS cycles
5
p ( ), y
Different Optimization Methods Showed Control Depends on DistanceDepends on Distance
Global Optimization Heuristic Optimization (UDDS)30O
N
UDDS 20ld
25
ch IC
E is
UDDSHWFETLA92 18
20
N th
resh
ol
15
20
abov
e w
hic
14
16
W) I
CE
ON
10
15
eels
(kW
) a
12
14
whe
els
(kW
0 10 20 30 40 505
Distance (miles)
Pwr w
he
10 20 30 40 50 6010
Distance (miles)
Pwr w
6
Small SUV, Parallel Pre-tx, 10 miles AER
“ Engine On” Is Linked to Wheel Power35
=.95
Cycle:
Global Optimization
25
30
h P(
ICE=
ON
)= Cycle:
HWFETUDDS
LA92
15
20
) abo
ve w
hich
10 miles20 miles
Driven Distance
0
5
10
wr w
heel
s (kW
) 20 miles40 miles
Power at wheels above which ICE is on 95% of the time, similar to wheel
0 50 100 150 200 2500P
Plug-to-wheel Electric Consumption (Wh/km)
power threshold used in rule‐based controls
Higher Electric Energy Use → ICE starts “later”
Wheel Power Threshold follows linear trend
7
Wheel Power Threshold follows linear trend
Little influence of driving cycle
Engine Power Depends on Cycle and Electricity Use
Global Optimization
Electricity Use
30
35
W)
20
25
age
Pow
er (k
W
Cycle:
HWFETUDDS
LA92
10
15
Engi
ne A
vera
10 miles
Driven Distance
0 50 100 150 200 2500
5E
Plug-to-wheel Electric Consumption (Wh/km)
20 miles40 miles
UDDS: ICE power increases with Wh/km
LA92: ICE power constant, because cycle is aggressive enough for the ICE to operate
g p ( )
8
efficiently
Charging from ICE Likely When ICE is Predominant and Wheel Power is Low
Global Optimization
Predominant and Wheel Power is Low
0.35
0.4
Cycle:UDDS
0 2
0.25
0.3 Engine Efficiency
HWFET
Driven
LA92
0.1
0.15
0.2
Share of Engine Output Energy Used to Charge the Battery
10 miles20 miles40 miles
Distance
0 50 100 150 200 2500
0.05
Pl t h l El t i C ti (Wh/k )
UDDS: when ICE is used often, it has to operate often above requested wheel power
LA92: ICE operates efficiently, even in CS mode because average ICE power is high
Plug-to-wheel Electric Consumption (Wh/km)
9
LA92: ICE operates efficiently, even in CS mode because average ICE power is high
5 Vehicles with Different Energy and AERGlobal Optimization
90 90
Power to meet EV-mode requirements on UDDS…
70
80
90
65.8046
y (A
h) …Different UDDS All-Electric Range
10 miles AER on UDDS
40
50
60
44.5589
y C
apac
ity
UDDS
10
20
30
11.9045
23.2123
Bat
tery
5 10 20 30 400
10
All-Electric Range (miles)
10
For a Given Electric Consumption, AER Has Little Influence on Control
Global Optimization
For a given electric consumption, ICE‐On behavior does not depend on the energy sizing…
b ibl l i i i li i d l i
40954095
…but possible electric energy consumption is limited : same electric consumption corresponds to different SOC
25
30
35
40
ch P
(ICE=
ON
)=.9
Maximal battery depletion
25
30
35
40
ch P
(ICE=
ON
)=.9
AER5AER10AER20AER30AER40
10
15
20
AER10
AER20
AER30
AER40
(kW
) abo
ve w
hic
10
15
20
s (kW
) abo
ve w
hi AER40
0 0.2 0.4 0.60
5
AER5
SOC0 - SOC
end
Pwr w
heel
s CS mode
0 50 100 150 200 2500
5
Electric Energy Consumption (Wh/km)
Pwr w
heel
s
11
UDDS, 20 mi
Minimal Fuel Consumption Distance Traveled < AER
Global Optimization
Distance Traveled AER
– Decrease in fuel consumption is proportional to increase in AER
– 1.5 L/100 km difference between UDDS and LA92
5
6
7
(L/1
00km
)
Driven Distance: 10 miles
250
300
350
(Wh/
k)
Driven Distance: 40 miles
Wh/km (UDDS)Wh/km (HWFET)Wh/km (LA92)
3
4
5
el C
onsu
mpt
ion
L/100km (UDDS)L/100km (HWFET)L/100k (LA92)
150
200
250
Cti
( )
0
1
2
Min
imal
Fue
5 10 20 30 40
L/100km (LA92)
5 10 20 30 405 10 20 30 400
50
100
Elt
iE
12
5 10 20 30 40AER (miles)
5 10 20 30 405 10 20 30 40AER (miles)
Fuel Consumption Follows a Linear Trend and Depends on Wh/km
Global Optimization
p
Small AER vehicles consume slightly less fuel due to lower weight, but difference in consumption is minimal
For a given electric consumption, fuel consumption is comparable
For 2 vehicles (e.g. AER 30 & AER 40) travelling less than their AER (e.g. 20 mi on UDDS), same EV mode electric consumption
4
5
6
100k
m)
AER5AER10AER20AER30
UDDS, 20 miles
2
3
4
onsu
mpt
ion
(L/1 AER30
AER40
0 50 100 150 200 2500
1
2
Fuel
Co
13
0 50 100 150 200 250
Plug-to-wheel Electric Consumption (Wh/km)
5 Vehicles With Different PowerGlobal Optimization
120
Same Battery Energy… … Different Electric Power
100
89.4366
104.3427
W)
60
80
59.6244
74.5305
10 miles AER ony
Pow
er (k
W
20
4044.7183
AER on UDDS
Bat
tery
0.6 0.8 1 1.2 1.40
20
Power Scaling Ratio
14
Power Scaling Ratio
Maximum Power Impacts Wheel Power Threshold for Engine ON
Global Optimization
Threshold for Engine ON The engine starts “earlier” when the electric system has lower
powerpower
30
35
E=O
N)=
.95
Ratio 0.6Ratio 0.8
20
25
e w
hich
P(IC
E
Ratio 1Ratio 1.2Ratio 1.4
5
10
15
ls (k
W) a
bove
20 miles UDDS
0 50 100 150 2000
5
Electric Energy Consumption (Wh/km)
Pwr w
heel
15
Less Electric Power Results in Higher Fuel Consumption
Global Optimization
Consumption
Especially true on aggressive cycle (LA92), and distance traveled close to AER
Hi h l t i d t i ifi tl d f l ti Higher electric power does not significantly reduce fuel consumption
UDDS seems to be a good choice for power requirements (In terms of energy use)
4
5
L/10
0km
)
200
250
n (W
h/km
)
2
3
Con
sum
ptio
n (L
100
150
gy C
onsu
mpt
ion
L/100km (UDDS)L/100km (HWFET)L/100km (LA92)
Wh/km (UDDS)Wh/km (HWFET)Wh/km (LA92)
0
1
Min
imal
Fue
l
0
50
Elec
tric
Ene
rg
16
0.6 0.8 1 1.2 1.4
Power Scaling Ratio
Control Strategy Assessment Provided Insights on PHEVs Optimal OperationsPHEVs Optimal Operations
When the trip distance is greater than the All Electric Range, i h i h h h i (bl d d l) iusing the engine throughout the trip (blended control) is
preferable to depleting the battery as fast as possible
Optimum control depends on the distanceOptimum control depends on the distance
Engine On/Off is linked to wheel power demand and available electrical energy
When used, engine should be operated at high efficiency
17