Jairo A Sandoval

Ph.D. Dissertation Proposal

Center for Alternative Fuels, Engines and Emissions
Mechanical and Aerospace Engineering
West Virginia University
March 5
th
, 2010
Outline
1. Motivation
2. IBIS
3. Objectives
4. Methodology and Preliminary Results
5. Contribution
6. Timeline
2
1. Motivation
Environmental effects of transportation
Vehicle duty cycle Fuel economy & Emissions
Hybrid buses will not perform the same on all service
routes
Emissions and fuel economy correlate very well with
cycle metrics, but transit agencies do not know the
metrics corresponding to their operation (e.g.
acceleration or % idle)
3
Drive Cycle Metrics
Metric Units
Percentage of Idle
%
Stops per mile
mi
-1

Maximum Speed,
Acceleration (+/-),
Grade (+/-)
-
Average Speed
mph
Average Speed w/o idle
mph
Aerodynamic Speed
mph
Standard Deviation of
Speed
mph
Standard Deviation of
Speed w/o Idle
mph
Metric Units
Average Acceleration
ft/s
2
Average Deceleration
ft/s
2
Characteristic Acceleration
ft/s
2
Kinetic Intensity
mi
-1

Average Time of a Stop
s
Fractions in Accel./Decel-
Speed Bins
%
Fractions in Grade Bins %
Mean Grade Ascent/Descent
%
4
( ) 100 / % 1
1
Idle
V V
idle no

=
2
~
aero
V
a
ki =
Time, Distance,
Energy,
2. Integrated Bus Information System - IBIS
Some features are:
Searchable databases of transit vehicle emissions studies
Predictive emissions modeling tools
Construction of virtual fleets
F.E., CO2, CO,
NO
x
, HC, PM,
NMHC
Diesel, CNG,
Diesel-Electric
Hybrid

Source: IBIS Presentation, July 2009
5
IBIS (Continued)
Emissions data are available
on no more than 16
standard cycles + Idle cycle
Second-by-second data
micro-trips new entries
into the dataset
Chassis-dynamometer test
data Correlations for fuel
economy and emissions
Correction factors
Current inputs:
(Average speed, Stops per Mile)
(Average speed, Percentage Idle)
(Average speed, Std. Dev. Speed)
(Average speed, Kinetic Intensity)
SOC=0?
6
3. Objectives
1. To characterize transit bus operation
Correlations between the information available to
transit agencies and the cycle metrics
Evaluation of in-use routes and their properties
2. To evaluate the effects of driving cycle on the F.E.,
CO2 and NO
x
emissions of diesel-electric hybrids
Variety of driving conditions
Regression based model
Allow transit agencies to assess the best routes for the
hybrids
7
4. Methodology and Preliminary Results
Characterization of Transit Duty Cycles:
GPS logs of transit activity
Collect information from transit agency
Analysis
Develop correlations
Hybrid Bus Emissions Modeling
MY 2007-2009 Allison Hybrid Bus
ANN model for fuel, CO2, and NO
x
rates
Develop hybrid controller / Calibration
Simulations / Analysis
Develop correlations
8
4.1. Characterization of Transit Duty Cycles
GPS logs of in-use routes.
2,900 mi / 230 hr
Classify routes:
Stop-and-go
Urban
Suburban
Commuter
Express
9
Trip distance / Trip time
in light and heavy traffic
Option: Process user-
provided GPS logs and
extract the metrics

4.1. Characterization of Transit Duty Cycles
10
y= 55.1395 - 1.4102*x+ 0.0991(x-13.2385)
2

R
2
= 0.6937

NI
=13.8

NI
=17.4

NI
=35.7

NI
=42.8
4.2. Vehicle Dynamic Model
The Road Load Equation:
11
( ) ( ) ( )
dt
dV
V m mgV mgV c cV bV V V A c P
e r r f d road
+ + + + + = o o sin cos
2
1
2 2
Aerodynamic Loads due Rolling Elevation Inertial
Drag to Driveline Resistance Load Load
Friction
Vehicle Dynamic Model (Cont.)
ANLs PSAT:
12
Vehicle Dynamic Model (Cont.)
ANLs PSAT:
13
Vehicle Dynamic Model (Cont.)
2-Mode Transmission:
14
Engine /
Input
Ring 1
Sun 1
Motor A /
Motor 2
Ring 2
Sun 2
Motor B /
Motor 1
Ring 3
Sun 3
Output
C1
C2
E1
E2 E3
C
a
r
r
i
e
r

1
C
a
r
r
i
e
r

2
C
a
r
r
i
e
r

3
4 s / 4 s. 4 Eqns. 4 Degrees of Freedom (2 s / 2 s)+ Mode
Remove output and
4.3. ANN Engine Model
Emissions tests of MY 2006 Allison Hybrids (17 cycles)
and BAE Series Hybrids (6 cycles)
Cummins ISL280, ISB 260H
15
Speed (rpm)
T
o
r
q
u
e

(
N
m
)


800 1000 1200 1400 1600 1800 2000 2200
200
300
400
500
600
700
800
900
1000
1100
1200
0
0.2
0.4
0.6
0.8
1
1.2
1.4
In-use data ISL 280 Chassis Tests ISB 260H
ANN Engine Model (Cont.)
Trust ECU torque and speed
Inputs: torque, speed, delayed first derivatives (1, 5, 10
seconds)
60% training, 20% validation, 20% testing
Correct for engine NO
x
certification levels (2.5/1.5
g/brp-hr)
16
Input Output
radbas
Layer
logsig
Layer
tansig
Layer
Output Layer
(linear)
ANN Engine Model (Cont.)
Generalization (Partial Training):
17
Unseen Cycles
Cycles Seen in Training
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1 2 3 1 2 3 4 5 6 7 1 2 3 4 5 6 7 8
N
O
x
(
g
/
b
h
p
-
h
r
)
Measured Predicted
Arterial Braunschweig WMATA
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1 2 1 2 3 4 5 6 0 1 2 3 4 5 6 7 8 9
N
O
x
(
g
/
b
h
p
-
h
r
)
Measured Predicted
CBD New York Bus OCTA
Training Set
ANN Engine Model (Cont.)
Performance (OCTA cycle):
18
200 220 240 260 280 300 320 340 360 380 400
0
5
10
F
u
e
l

R
a
t
e

(
g
/
s
)


Target
Output
200 220 240 260 280 300 320 340 360 380 400
5
10
15
20
25
C
O
2

(
g
/
s
)


200 220 240 260 280 300 320 340 360 380 400
0
0.05
0.1
Time (s)
N
O
x

(
g
/
s
)


ANN Engine Model (Cont.)
Overall Performance
Fuel: 50% of results within 0.5% accuracy
80% 1.1%
95% 2.5%
CO2: 50% 1.3%
80% 2.8%
95% 5.8%
NO
x
: 50% 3%
80% 5%
95% 10%
19
ANN Engine Model (Cont.)
Performance of Fuel Model:
20
%

E
r
r
o
r

F
u
e
l
-4
-3
-2
-1
0
1
2
3
4
A
r
t
e
r
i
a
l
B
e
e
l
i
n
e
B
r
a
u
n
C
B
D
C
o
m
m
C
S
H
V
C
E
T
C
U
r
b
a
n
I
d
l
e
K
C
M
M
a
n
h
a
t
t
a
n
N
Y
B
u
s
N
Y
-
C
o
m
p
O
C
T
A
P
a
r
i
s
T
r
a
n
s
U
D
D
S
W
M
A
T
A
Cycle
S
.
E
.

F
u
e
l
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
A
r
t
e
r
i
a
l
B
e
e
l
i
n
e
B
r
a
u
n
C
B
D
C
o
m
m
C
S
H
V
C
E
T
C
U
r
b
a
n
I
d
l
e
K
C
M
M
a
n
h
a
t
t
a
n
N
Y
B
u
s
N
Y
-
C
o
m
p
O
C
T
A
P
a
r
i
s
T
r
a
n
s
U
D
D
S
W
M
A
T
A
Cycle
% Error Standard Error, g/s
( )
2

2
2

= = =

n
y y
DFE
SSE
MSE SE
i i
ANN Engine Model (Cont.)
Performance of CO2 Model:
21
% Error Standard Error, g/s
%

E
r
r
o
r

C
O
2
-6
-4
-2
0
2
4
6
8
10
A
r
t
e
r
i
a
l
B
e
e
l
i
n
e
B
r
a
u
n
C
B
D
C
o
m
m
C
S
H
V
C
E
T
C
U
r
b
a
n
I
d
l
e
K
C
M
M
a
n
h
a
t
t
a
n
N
Y
B
u
s
N
Y
-
C
o
m
p
O
C
T
A
P
a
r
i
s
T
r
a
n
s
U
D
D
S
W
M
A
T
A
Cycle
S
.
E
.

C
O
2
0.0
0.5
1.0
1.5
2.0
2.5
A
r
t
e
r
i
a
l
B
e
e
l
i
n
e
B
r
a
u
n
C
B
D
C
o
m
m
C
S
H
V
C
E
T
C
U
r
b
a
n
I
d
l
e
K
C
M
M
a
n
h
a
t
t
a
n
N
Y
B
u
s
N
Y
-
C
o
m
p
O
C
T
A
P
a
r
i
s
T
r
a
n
s
U
D
D
S
W
M
A
T
A
Cycle
ANN Engine Model (Cont.)
Performance of NO
x
Model:
22
% Error Standard Error, g/s
%

E
r
r
o
r

N
O
x
-15
-10
-5
0
5
10
15
20
A
r
t
e
r
i
a
l
B
e
e
l
i
n
e
B
r
a
u
n
C
B
D
C
o
m
m
C
S
H
V
C
E
T
C
U
r
b
a
n
I
d
l
e
K
C
M
M
a
n
h
a
t
t
a
n
N
Y
B
u
s
N
Y
-
C
o
m
p
O
C
T
A
P
a
r
i
s
T
r
a
n
s
U
D
D
S
W
M
A
T
A
Cycle
S
.
E
.

N
O
x
0.000
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.010
0.011
A
r
t
e
r
i
a
l
B
e
e
l
i
n
e
B
r
a
u
n
C
B
D
C
o
m
m
C
S
H
V
C
E
T
C
U
r
b
a
n
I
d
l
e
K
C
M
M
a
n
h
a
t
t
a
n
N
Y
B
u
s
N
Y
-
C
o
m
p
O
C
T
A
P
a
r
i
s
T
r
a
n
s
U
D
D
S
W
M
A
T
A
Cycle
5. Contribution
Correlate transit bus activity with information
available to transit agencies and cycle metrics
IBIS Module
Effect of driving cycle on NO
x
, CO2, and F.C.
Model to help transit agencies assess the best routes for
the hybrids
IBIS Module
23
6. Timeline
24
ACTIVITY Done
2010 2011
2 3 4 5 6 7 8 9 10 11 12 1 2 3
Collection of driving data
Sizing of hybrid bus components
Analysis of drive cycle data
Development and tuning of a controller for
the 2-mode hybrid bus
Development of engine fuel consumption,
CO
2
, and NO
x
model and inclusion in PSAT
Model validation
Computer simulations varying driving
conditions
Development of regression based emissions
model
Implementation in IBIS
Compilation of the document
Oral defense
Corrections and submission
Questions???
25