Sei sulla pagina 1di 14

SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 4, No. 2, November 2007, 119-132

Dynamic Modeling and Simulation of an Induction Motor with Adaptive Backstepping Design of an Input-Output Feedback Linearization Controller in Series Hybrid Electric Vehicle

Mehran Jalalifar 1 , Amir Farrokh Payam 2 , Seyed Morteza Saghaeian Nezhad 3 , Hassan Moghbeli 4

Abstract: In this paper using Adaptive backstepping approach an adaptive rotor flux observer which provides stator and rotor resistances estimation simulta- neously for induction motor used in series hybrid electric vehicle is proposed. The controller of induction motor (IM) is designed based on input-output feedback linearization technique. Combining this controller with adaptive backstepping observer the system is robust against rotor and stator resistances uncertainties. In additional, mechanical components of a hybrid electric vehicle are called from the Advanced Vehicle Simulator Software Library and then linked with the electric motor. Finally, a typical series hybrid electric vehicle is modeled and investigated. Various tests, such as acceleration traversing ramp, and fuel consumption and emission are performed on the proposed model of a series hybrid vehicle. Computer simulation results obtained, confirm the validity and performance of the proposed IM control approach using for series hybrid electric vehicle.

Keywords: Adaptive backstepping observer, Series electric vehicle.

1

Introduction

Nowadays the air pollution and economical issues are the major driving forces in developing electric vehicles (EVs). In recent years EVs and hybrid electric vehicles (HEVs) are the only alternatives for a clean, efficient and environmentally friendly urban transporta- tion system [1]. HEVs meet both consumer needs as well as car manufacturer needs. They give the consumer the ability to use the car for long periods of time

1 Islamic Azad University, Fereydan Branch, Esfahan, Iran; Email: mehran_j1356@yahoo.com 2 Dept. of Electrical & Computer Engineering, University of Tehran, Tehran 11365/4563, Iran; 3 Dept. of Electrical & Computer Engineering, Isfahan University of Technology, Isfahan, Iran. 4 Dept. of Electrical & Computer Engineering, Isfahan University of Technology, Isfahan, Iran; Email :hamoghbeli@yahoo.com

119

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

without recharging. HEVs also take a giant step forward in meeting low emission standards set by the Partnership for a New Generation of Vehicles. Because of simple and rugged construction, low cast and maintenance, high performance and sufficient starting torque and good ability of acceleration, squirrel cage induction motor is a good candidate for EVs [2]. In this paper by using an input-output feedback linearization technique combined with an adaptive backstepping observer in stator reference frame the induction motor [3] using in series hybrid electric vehicle is controlled. One of the best advantages of this control method is eliminating the flux sensor and decreases the cost of controller in addition the control system is robust respect to resistances variations and external load torque. Advanced vehicle simulator (ADVISOR) provides the vehicle engineering community with an easy-to-use, flexible, yet robust and supported analysis package for advanced vehicle modeling. It is primarily used to quantify fuel economy, the performance, and the emissions of vehicles that use alternative technologies including fuel cells, batteries, electric motors, and ICE in hybrid configurations. But the components in ADVISOR have been modeled simply and only with static model to decrease the simulation time [4]. In this paper using MATLAB/SIMULINK software, dynamic modeling of an induction motor that is used in series hybrid electric vehicle and controlled by input-output feedback linearization method combined with adaptive backstepping observer is investigated and then simulated separately by linking the mechanical components for a series hybrid electric vehicle from the ADVISOR software library. At the end, a typical HEV is modeled and investiga- ted. Simulation results obtained show the IM and other components perfor- mances for a typical city drive cycle.

2 The Performance of an Electric Vehicle

The first step in vehicle performance modeling is to write an equation for the electric force. This is the force transmitted to the ground through the drive wheels, and propelling the vehicle forward. This force must overcome the road load and accelerate the vehicle as shown in Fig. 1 [5]. The rolling resistance is primarily due to the friction of the vehicle tires on the road and can be written as:

(1)

where M is the vehicle mass, gravity acceleration.

The aerodynamic drag is due to the friction of the body of vehicle moving through the air. The formula for this component is as in the following:

f is the rolling resistance coefficient and g is

f

roll

r

= f Mg ,

r

120

Dynamic Modeling and Simulation of an Induction Motor with…

f

AD

1

= ξC AV ,

2

D

2

(2)

where ξ is the air mass density, and

coefficient, and the frontal area of the vehicle, respectively.

V , C

D and A are the speed, the aerodynamic

V , C D and A are the speed, the aerodynamic Fig. 1 – A summary

Fig. 1 A summary of forces on a vehicle.

The gravity force due to the slope of the road can be expressed by:

f

grade

= Mg sin α ,

where α is the grade angle.

(3)

In addition to the forces shown in Fig. 3, another one is needed to provide

the linear acceleration of the vehicle given by:

f

acc

=

M

α =

M

d

V

d

t

.

(4)

The propulsion system must now overcome the road loads and accelerate

, as follows:

(5)

A typical road load characteristic as a function of the speed and mass of a

the vehicle by the tractive force,

F tot

F

tot

= f

roll

+ f

AD

+ f

+ f

acc

.

grade

vehicle is shown in Fig. 2.

+ f AD + f + f acc . grade vehicle is shown in Fig. 2.

Fig. 2 The road profile as a function of speed and mass of a vehicle ( α = 0 ° ).

121

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

Wheels and axels convert

F tot

and the speed of vehicle to torque and

angular speed requirements for the differential as follow [4]:

T

wheel

=

F r

tot wheel

,

ω

wheel

=

V

,

r wheel

(6)

where

are the tractive torque, the radius, and the angular

velocity at the wheels, respectively. The angular velocity and torque of the wheels are converted to motor rpm and motor torque requirements using the gears ratio at differential and gearbox as follows:

(7)

T

wheel

, r

, and

wheel

ω

wheel

T wheel

G G

fd

gb

,

ω =

m

G G

fd

gb

ω

T

m

=

wheel

,

where

G

fd

and

G

gb

are respectively differential and gear box gears ratios.

A series HEV consists of two major group components as shown in Fig. 3:

- Mechanical components (engine, wheels, axels and transmission box) and

- Electrical components (batteries and electric motor).

and - Electrical components (batteries and electric motor). Fig. 3 – A series hybrid electric vehicle

Fig. 3 A series hybrid electric vehicle.

3 Electric Motor

Due to the simple and rugged construction, low cast and maintenance, high performance and sufficient starting torque and good ability of acceleration, squirrel cage induction motor is one of the well suited motors for the electric propulsion systems [6]. In this section first modeling of the induction motor in

122

Dynamic Modeling and Simulation of an Induction Motor with…

stator fixed reference frame and then designing an input-output controller combined with adaptive backstepping observer for IM [3] is investigated.

A. Input-Output Feedback linearization Controller Design The square of rotor flux amplitude is taken as the first output of the controlled system

y

e

1

1

= ψ

=

y

1

2

ra

+ ψ

2

rb

2

= ψ ,

r

y

1 d

= ψ − ψ ,

r

r

2

r 2

where

r

ψ r

is the reference signal for rotor flux amplitude.

Differentiating

ɺ

1

2

= ψ

ra

y

1

so much inputs (

r

L

r

ψ +

ra

R

r

L

r

sa

R

y Mi

u

sa

,

u

sb

+ 2 ψ

)

rb

appear in our equations,

R

r

L

r

ψ +

rb

R

r

L

r

Mi

sb

.

Then

+ 2

R

4

r

L

MR

r

r

ψ

ra

R

r

ψ −

R

ɺɺ y

= −

1

ψ

L

r

ψ +

ra

2

M R

r

+

L

L R

r

2

r

s

sa

i

+

r

Mi

+

u

sa

R

L

r

4

ra

σ

L L

s

2

r

ra

R R

r r

L

r

ψ

rb

L

r

ψ +

rb

R

r

L

r

σ

Mi

sb

2

r

L L

s

+

2

R

r

L

r

sa

M

ψ

σ

L

rb

.

s

r

M

If

e

1

we have

dynamic force to be

k

ɺɺe

11 1

+ k eɺ + k e = 0 ,

12 1

13 1

+

2

=

ψ

k (

11

ɺɺy ɺɺy

1

1

d

ɺɺ y

1

=

k

12

k

11

)

+ k ( yɺ

12

1

(

y ɺ

1

d

y ɺ

1

)

+

yɺ

k

d

1

13

k

11

)

(

+ k ( y y

13

1

1

y

1

d

y

1

)

+

ɺɺ y

d

)

1

d

=

=

0

,

k

13

k

11

rb

(

ψ − ψ )

r

r

r 2

2

k

12

k

11



R

r

L

r

ψ

rb

+

R

r

L

r

Mi

sb

2

ψ

     

ra

R

r

L

r

ψ

ra

− ψ ψ ɺ +

2

r

r

r

r

+

2

(

R

r

L

r

Mi

sa

+

ψ ɺ + ψ ψ ɺɺ

r

r

r

2

r

r

r

)

.

(8)

(9)

(10)

(11)

Now the motor developed electromagnetic torque is considered as the second output

123

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

y

2

=

te

=

n M

p

2

L

r

(

ψ

i

ra sb

− ψ

i

rb sa

)

where

r

y

te is the reference signal for developed electromagnetic torque.

e

2

=

y

2

2 d

=

te te

r

Time derivative of

y ɺ

2

=

n M

p

2 L

r



y

2

ψ

ra

+

i sb

R

r

L

r

ψ −

ra

n

p

ωψ

is:

rb

n M

p

σ

L L

s

r

ωψ −

ra

2

M R

r

+

2

L R

r

s

σ

2

r

L L

s

i

sb

+

u

sb

σ

L

s

− ψ

rb

n M

2

M R

r

+

2

L R

r

s

p

L L

ωψ −

rb

2

r

L L

σ σ

s

r

s

i

sa

R

L

− −

i

sa

r

r

ψ +

rb

n

p

ωψ

ra

 

.

+

+

u

sa

σ

L

s

By setting

e dynamic as

2

we have

eɺ +k e =0 ,

k

21 2

22 2

ɺ

y

2

=

k

21

k

22

k

21

( yɺ

2

(

y

2

d

yɺ +k y y =

2

d

)

22

(

2

2

d

)

y

2

)

+

y ɺ

2

d

=

k

22

k

21

(

te

r

0

te

,

)

+

ɺ

te

r

.

(12)

(13)

(14)

Setting right side of (13) equal to right side of (14), another equation for inputs is achieved

+

n M

p

R

r

2 L

r



ψ

ra

n M

p

2

M R

r

+

L R

r

s

i

2

+

+

u

sb

σ

L

s

2

L R

r

s

i

ωψ −

L L

σ

rb

s

r

− ψ

rb

ra

n M

p

L L

σ

2

r

L L

s

ωψ −

rb

sb

2

M R

r

+

i n

sb

L

r

ψ −

ra

p

ωψ

2

r

L L

sa

σ σ

s

r

s

i

sa

R

r

L

r

ψ +

rb

n

p

ωψ

ra

 

   

 

=

k

22

k

21

(

te

r

te

)

+

ɺ

te

r

.

+

u

sa

σ

L

s

(15)

Now, from system of two equations, (6 and 10), inputs (

To proof the stability of proposed controller, note that according to

e (0) = 0, then

u

sa

also

,

u

sb

2

)

otherwise,

are found.

e

1

( t ), e ( t )

2

dynamics

of

e

1

, t 0,

, e

if

initially

perfect

e (0) = eɺ (0) = 0

1

1

tracking

is

and

2

e

1

( t ), e ( t ) 0,

2

achieved;

converges to zero exponentially [7].

124

Dynamic Modeling and Simulation of an Induction Motor with…

B. Adaptive Backstepping Observer Design Stator current is measurable and is taken as output:

y

a

= i

sa

,

y = i

b

sb

.

(16)

The prediction model for the backstepping observer is chosen to be

p

p

ˆ

R

r

L

r

ˆ ˆ

ψ

= −

ra

ψ −

ra

ˆ

R

r

L

r

ˆ ˆ

ψ

= −

rb

ψ +

rb

n

n

p

p

ωψ ˆ +

rb

ωψ ˆ +

ra

R ˆ

r

L

R ˆ

r

r

L

r

My ,

a

My ,

b

ˆ

MR

r

σ

L

s

2

r

L

ˆ ˆ

pi

sa

=

ψ +

ra

n

p M

σ

L L

s

r

ωψ ˆ

rb

ˆ

M R

2

r

+

L R ˆ

2

r

s

σ

2

r

L L

s

y

a

+

ˆ

pi

sb

ˆ

MR

r

n

p M

ˆ

= ψ −

2

rb

σ σ

s

r

L L

L L

s

r

ωψ ˆ

ra

ˆ

M R

2

r

+

L R ˆ

2

r

s

σ

2

r

L L

s

y

b

+

1

σ

L

1

s

σ

L

s

u

u

sa

sb

+

+

v ,

a

v ,

b

(17)

where p is the differential operator and

designed by the backstepping method. The dynamical equations for the prediction errors are

v

a

, v

b

are the control input to

be

where

ɶ

pi

sa

=

ɶ

pi =

sb

2

r

L

ɶ R

r

r

ɶ

ψ −

ra

ɶ

ψ +

rb

n M

p

n

p

ɶ

ωψ +

rb

r

My ,

a

n

p

ɶ

ωψ +

ra

r

L

r

My ,

b

ɶ

ωψ −

rb

ɶ

M R

2

r

+

ɶ

L R

2

r

s

σ

2

r

L L

s

ɶ

R

r

ˆ

L

ˆ ɶ

R

ɶ

R

r

ψ ψ −

p

p

ɶ

= −

ra

L

r L

ɶ

R

r

ra

r

ɶ R

r

r

ψ ψ −

rb

= −

L

rb

r L

ψ +

ra

ˆ

MR

r

σ

2

r

L L

s

ɶ

ψ +

ra

MR

σ

L

s

σ

L L

s

r

ɶ

MR

r

σ

L

s

2

r

L

ψ +

rb

ˆ

MR

r

σ

2

r

L L

s

n

M

ɶ p

ψ −

rb

s

r

σ

L L

ɶ M

ωψ −

ra

2

ɶ

R

r

+

ɶ

L R

2

r

s

σ

2

r

L L

s

ɶ

y

ɶ

y

a

b

=

=

ɶ

i

ɶ

i

sa

sb

,

,

ψ ɶ

ra

ɶ

i

sb

ˆ

= ψ − ψ

ra

=

ˆ

i

sb

i

sb

ra

,

, ψ ɶ = ψ − ψ

R ˆ

ˆ

rb

rb

R ɶ

s

=

rb

,

s

R R ɶ

s

,

r

ɶ

i

sa

=

=

R ˆ

r

ˆ

i

sa

R

i

sa

r

.

,

y

a

y b

+

+

v ,

a

v ,

b

(18)

(19)

The first step in the backstepping strategy is to design a stable controller for

as virtual control

the integral of the prediction errors

ɶ

y

a

,

ɶ y

b

ɶ

ɶ

using

i sa , i sb

125

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

variables with stabilizing functions

variables. The integral of the prediction errors

φ

a

,

φ

b

which

x ɶ

a

px ɶ = i ɶ

b

,

sb

x ɶ

.

b

px ɶ = ɶ i

a

sa

,

is

are

Adding and subtracting

px ɶ

a

z

, φ

φ

=

a

z

b

a

φ

=

a

a

ɶ i

sa

= −

to above equations

px ɶ

z

=

φ = −

,

c x ɶ

1

a

− φ

a

,

,

c x ɶ

1

a

,

b

b

b

z

=

ɶ i

sb

b

− φ

c x ɶ

1

b

c x ɶ

1

b

.

b

,

reference

for

virtual

(20)

(21)

The second step in the backstepping strategy is the control of

z

a

Taking the derivative of

ɺ

z

ɺ

z

a

b

ɶ

MR

r

= ψ +

σ

2

L L

s r

ra

ˆ

MR

r

σ

2

r

L L

s

ɶ

MR

r

ˆ

MR

= ψ +

2

rb

r

σ L L

L

s

L

r

σ

s

2

r

= ɶ i + c x ɶ

z

and

a

sa

1

a

z ,

b

,

z = ɶ i + c x ɶ

b

sb

1

b

.

r

ωψ ɶ

rb

M

M

2 ɶ 2 R + L R ɶ r r 2 σ L L s
2
ɶ
2
R
+
L R ɶ
r
r
2
σ
L L
s
r
2
ɶ
2
R
+
L R ɶ
r
r

s

s

n

p M

L L

σ

n

s

M

ωψ ɶ

ra

σ

2

r

L L

s

ɶ

ψ +

ra

ɶ p

ψ −

rb

σ

L L

s

r

y

a

y

b

+

+

v

a

v

b

and selecting following control inputs

yields

where

c 1 and c

2

v

a

v

b

= −

= −

ɺ

z

ɺ

z

a

b

=

=

ˆ

MR

r

σ L

2

r

s L

ˆ

MR

r

σ

L

s

2

r

L

ɶ

MR

r

σ

L

s

2

r

L

ɶ

MR

r

σ

L

s

2

r

L

ɶ

ψ −

ra

ψ ɶ +

rb

ψ −

ra

ψ −

rb

n

p M

σ

L L

s

r

n

p

M

σ

L L

s

r

ɶ

M R

2

r

ωψ ɶ

rb

ωψ ɶ

ra

+

L R ɶ

2

r

s

σ

ɶ

M R

2

L L

s

r

+

L R ɶ

2

r

s

2

r

σ

2

r

L L

s

ɶ

c i

1

sa

ɶ

c i

1

sb

y

a

y

b

c z

2

a

c z

2

b

c z

2

a

c z

2

b

x ɶ ,

a

x ɶ ,

b

x ɶ ,

a

x ɶ ,

b

are positive constant design parameters.

+

+

ɶ

c i

1 sa

ɶ

c i

1 sb

,

(22)

(23)

(24)

(25)

Stability analysis of observer is done by the following Lyapunov candidate:

V

=

1

2



x ɶ

2

a

+

x ɶ

2

b

+

z

2

a

+

z

2

b

+ ψ ɶ + ψ ɶ +

ra

rb

2

2

1

γ

s

R ɶ

2

s

+

1

γ

r

R ɶ

2

r

.

(26)

Derivating V along the dynamics of (18, 21 and 25) yields

126

Dynamic Modeling and Simulation of an Induction Motor with…

+

ɶ

R

r

ɺ

V

= −

c x ɶ

1

2

a

2

b

c x ɶ

1

c z

2

2

a

+

ɶ

R

z L

2

s

ψ −

ra

σ

y

a

r

2

y

a

a

L L

s

2

r

M z

a

σ

2

r

L L

s

z M

a

σ

2

r

L L

s

+

c z

2

2

r

z L

b

2

b

σ

2

r

L L

s

z M

b

σ

2

r

L L

s

ˆ

R r

R ˆ

y

b

L

r

+

ψ ɶ

r

ra

L

r

s

ɶ

1 dR

γ

s

dt

2

ψ −

rb

2

M z

b

σ

2

r

L L

s

y

b

ψ ɶ

2

rb

ψ ɶ ψ L

ra

ra

r

+

M

L

r

y

a

ψ ɶ

ra

ψ ɶ ψ L

rb

r

rb

+

M

L

By selecting following adaptation laws

r

ɶ

dR

s

dt

= γ

s

2

r

z L

a

σ

2

r

L L

s

y

a

y

+

b

ψ ɶ

rb

1 dR

ɶ

r

+ γ

r

dt

2

r

z L

b

σ

2

r

L L

s

y

b

.

ɶ

dR

r

dt

= γ

+

z M

a

M z

a

y

2

z M

b

r

σ

ψ ɶ ψ L

ra

2

r

L L

s

ra

M

r

L

r

ψ +

ra

y

a

σ

2

r

L L

s

ψ ɶ +

ra

ψ ɶ ψ L

rb

a

r

σ

2

r

L L

s

rb

M

y

rb

ψ ɶ

rb

ψ +

L

r

b

2

M z

b

σ

,

2

r

L L

s

y

b

ɺ

we have V < 0

outside the equilibrium point

,

(

xɶ

a

,

xɶ , z

b

a

z ψɶ

,

ra

,

ψɶ

rb

)

= (0, 0, 0, 0, 0, 0).

b

(27)

(28)

Based on the Barbalat’s Lemma, we can obtain

xɶ

xɶ z

b

will

, converge to zero as t → ∞ . Therefore, the proposed observer is stable, even if parametric uncertainties exist [8]. The block diagram of the proposed controller is given in Fig. 4.

a

,

a

,

ra

,

, z

b

ψɶ

ψɶ

rb

r Ψ r u Input-Output sa u Feedbeack sb Park r Lineariztion Transformation te ω
r
Ψ
r
u
Input-Output
sa
u
Feedbeack
sb
Park
r
Lineariztion
Transformation
te
ω r PI
Controller
Controller
+
-
ɶ
ɶ
Ψ ,
Ψ
ra
rb
i
i
i
sA
sB
sC
Adaptive Backstepping
ɶ
ɶ
R R
Obsrever
u
s
r
sA
u
u
sB
sC
ω
IM

Fig. 4 Block diagram of proposed controller.

127

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

4 Battery Modeling

The battery considered in this paper is of the NiMH type for which a simple model is assumed. Therefore, a simplified version of the complex battery model reported in [9] is used.

5 Simulation Results

Simulation results presented in this section, focus on the dynamic behavior of IM and the battery of the vehicle. First the system is simulated for ECE+EUDC test cycle. This cycle is used for emission certification of light duty vehicles in Europe. Due to the electric motor has been modeled dynamically in SIMULINK. The data for IM and EV is in the appendix. Fig. 5 shows the simulation block diagram.

Controller Wheel Baterries Power Electric Transmission System Set Converter Motor and Differentional Power
Controller
Wheel
Baterries
Power
Electric
Transmission System
Set
Converter
Motor
and Differentional
Power
Diesel
Generator
Wheel
Converter
Engine
Electric Signal
Electric Energy
Electric Energy
Flow
Flow
Flow

Fig. 5 Simulation block diagram.

The drive cycle gives the required vehicle speed then the torque and speed requested from the electric motor. The current drawn from IM power supply shows the battery performance. The dynamic behavior of the IM in the ECE+EUDC drive cycle is shown in Figs. 6(a) and 6(d). Fig. 6(a) shows the ECE drive cycle. Figs. 6(c) and 6(d) show the IM torque and average torque. Finally, the system is simulated for fuel consumption and emissions test. The results obtained are shown in Fig. 7 and Fig. 8. Fig. 7 shows the fuel consu- mption and emissions of diesel engine when the motor is hot at the start of test, and Fig. 8 shows the results when the engine is cool at the start of test. It is

128

Dynamic Modeling and Simulation of an Induction Motor with…

obvious that when the engine is cool, fuel consumption and emissions is increases respect to the engine is hot.

and emissions is increases respect to the engine is hot. (a) (b) (c) (d) Fig. 6

(a)

and emissions is increases respect to the engine is hot. (a) (b) (c) (d) Fig. 6

(b)

emissions is increases respect to the engine is hot. (a) (b) (c) (d) Fig. 6 –

(c)

(d)

Fig. 6 – (a) ECE drive cycle, (b) Vehicle speed, (c) Average torque, (d) IM torque.

cycle , (b) Vehicle speed, (c) Average torque, (d) IM torque. (a) Fig. 7 – (a)

(a) Fig. 7 – (a) HC, (b) CO consumption.

129

(b)

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli (c) (d) (e) Fig. 7 – (c) NOX,

(c)

(d)

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli (c) (d) (e) Fig. 7 – (c) NOX,

(e)

Fig. 7 – (c) NOX, (d) PM, (e) Fuel consumption.

(c) (d) (e) Fig. 7 – (c) NOX, (d) PM, (e) Fuel consumption . (a) Fig.

(a) Fig. 8 – (a) HC, (b) CO consumption.

130

(b)

Dynamic Modeling and Simulation of an Induction Motor with…

Dynamic Modeling and Simulation of an Induction Motor with… (c) (d) (e) Fig. 8 – (a)

(c)

(d)

and Simulation of an Induction Motor with… (c) (d) (e) Fig. 8 – (a) HC, (b)

(e)

Fig. 8 – (a) HC, (b) CO, (c) NOX, (d) PM, (e) Fuel consumption.

6

Conclusion

Steady-state simulation tools such as ADVISOR have been developed in recent years for the design and analysis of electric and hybrid electric vehicles. In the past, dynamic simulation models have focused mainly on the analysis of control strategies. In this paper, the dynamic behavior of the electric motors with input-output state feedback controller combined with adaptive backstepping observer and batteries of a typical series hybrid EV is investigated and simulated by Matlab/Simulink, has been presented and the performance and ability of control strategy is investigated. Simulation results have also been shown the IM dynamic behavior and batteries for various tests such as ECE+EUDC drive cycle, maximum speed, traversing ramp and fuel consumption and emissions.

131

M. Jalalifar, A.F. Payam, S.M.S. Nezhad, H. Moghbeli

7 Appendix

A. EV Data:

Vehicle total mass of 8200 kg; Air drag coefficient of 0.79; Rolling resistance coefficient of 0.008; Wheel radius of 0.41 m; Level ground; Zero head wind.

B. The IM Parameters:

P

n

75 [kW]

X

m

1.95

[ ]

P

max

189

[kW]

X

ls

0.06

[ ]

f

60 [Hz]

X

lr

0.06

[ ]

T

n

209

[Nm]

n

s

3600 [r.p.m]

T

max

520

[Nm]

J

1.2

2

[kgm ]

R

s

0.02

[ ]

P

 

2

R

r

0.01

[ ]

   

8

References

[1]

M. Ehsani, K.M. Rahman, H.A. Toliyat: Propulsion System Design of Electric and Hybrid

[2]

Vehicle, IEEE Tran. On industrial Electronics, Vol. 44, No.1, February 1997. http://www.ott.doe.gov/hev/

[3]

R. Yazdanpanah, A. Farrokh Payam: Direct Torque Control of An Induction Motor Drive

[4]

Based on Input-Output Feedback Linearization Using Adaptive Backstepping Flux Observer, Proc. 2006 AIESP Conf., Madeira, Portugal. T. Markel, A. Brooker, T. Hendricks, V. Johnson, K. Kelly, B. Kramer, M. O' Keefe, S.

[5]

Sprik, K. Wipke: ADVISOR: A Systems Analysis Tool for Advanced Vehicle Modeling, ELSEVIER Journal of Power Sources 110, (2002), pp. 255-266. Electric Vehicle Technology Explained , Editted by James Larminie, and John Lowry, John

[6]

Wiley, England, 2003. A. Rahide: Vector Control of Induction Motor Using Neural Network in Electric Vehicle,

[7]

Scientific Report, IUST, 2000. R. Marino, S. Peresada, P. Valigi: Adaptive Input-Output Linearization Control of Induction

[8]

Motors, IEEE Trans. Automatic Cont., Vol. 38, No. 2, pp. 208-220, Feb. 1993. Applied Nonlinear Control , Editted by J. E. Slotine, Prentice-Hall International Inc.

[9]

S. Sadeghi, J. Milimonfared, M. Mirsalim, M. Jalalifar: Dynamic Modeling and Simulation of a Switched Reluctance Motor in Electric Vehicle, in Proc., 2006 ICIEA Conf.

132