ISSN 20088345

4 2 1389 1-10

1 2 3

1 rkardehi_moghaddam@yahoo.com
2 n-pariz@ferdowsi.um.ac.ir
3 shanechi@iit.edu
4 kamyad@math.um.ac.ir
) 1388/11/30 (1389/4/30

: .
.
.

.
- .

Using Directional Enlargement of Domain of Attraction to Increase Critical

Clearing Time of Nonlinear Systems
Reihaneh Kardehi Moghaddam, Naser Pariz, Hasan Modir Shanechi, Ali Vahidian Kamyad
Abstract: In this work we use directional enlargement of domain of attraction to increase
critical clearing time of nonlinear systems. To find the optimal control parameters, we use measure
theory which converts the enlargement problem to a linear programming problem. At first step we
find the critical directions of system, along them the system instability happens. After that we find
optimal controlling parameters to extend domain of attraction along directions of interest. The
efficiency of the proposed method is verified in simulation part for increasing critical clearing time
of a power system with four machines.
Keywords: DA enlargement measure theory optimization increasing critical clearing time
power systems

-1


.

.
.
] 1 [4
.

f ( 0, , t ) = 0
x (t = 0 ) = x 0

x0

x x 0
. ) x ( x0 ,0 , t

x 0 0.

:1

- > 0

( ) > 0
x0 < x( x0 , 0 , t ) < t 0

2 .

- .

]5 [6

c > 0

x0 < c lim x( x0 , 0 , t ) = 0

]12 .[7

] .[8

.

:2 ][12


.

:
-1
3 [9] 4
.[10] 5
-2
6 ]7. [9
.



4
5- .

)(2

}DA = {x 0 x0 D , lim x( x0 , 0 , t ) = 0
t

:3


.
:4
) - ( 7
8 ].[13
:5
h : X R 9
} {x X : h( x ) 0 ].[13
:6
10

X
x, y X U

-2

x V y

:
)(1

x& = f ( x, , t ) f : D P R + R n

U I V = ]13.[17

x D R n , x P Rl

- Post Fault System

- Critical Clearing Time
- Fast Voltage Regulators
4
- Break Resistor
5
- Line Reclosing
6
- Domain of Attraction
2
3

4 2 1389

- Heine Borel
- Compact
9
- Compact support
10
- Hausdorff
8

Journal of Control, Vol. 4, No. 2, Summer 2010

-3

) (3

e R n )(1

-1-4

] [11 ) (3

= [ 1 ... l ]T . e

) [0,

:1 e
:
Maximize
P

) J = [0,1 = t
t +1
:

Maximize

s.t. : x& = f ( x, , t ) f : D P R + R n

)(3

f (0, , t ) = 0

x0 = e

)s .t .: y& = g ( y , , ); J = [0,1
y 0 = y ( = 0) = e

)(4

lim x( x0 , , t ) = 0

lim y ( ) = 0

D P f .

g , y :

(y ( ) = x(t ) = x
)
1

) ( x0 = e
x0 =

)( 5

1
(1 ) 2

x0

e
.

e
.

-4


) (3
.
.

.

) g ( y, , ) = f ( y ( ),

g : D P J R n ; y D; J

) (3

.
:7
] w = [ y (.),
:

y(.) - ) y& = g ( y , ,
.
- P.
- lim y = 0 .

y 0 = y ( = 0) = e - 0.
) (3 ) w = ( 0,

W ) (3
.

4 2 1389

Journal of Control, Vol. 4, No. 2, Summer 2010

) (4

J
0
= ) F ( y , , ) C c (), F ( y1 , ,
h1 ( ) > 0 J
J
0
= ) F ( y 2 , ,
<
h
(

)
0
J
2

-2-4 ) I ( w w

, , )d > 0

) (4

] w = [ y(.),

, , )d < 0

= w 2 ( F ) = F (w 2 , )d
J

w w.
1
2

.
.

F (y

F (y

= w 1 (F ) = F (w 1 , )d

w .

= D P J

) Cc (

4-4

. ) I ( w :

I (w ) : F Cc (), [ y(.), ] W
w (F ) = F( y,, )d
J

-3-4

)(6
) F ( y, , ) C c (
) w (F w .
:1 ) I ( w :
F , G CC (), , R :
) w (F + G) = w (F ) + w (G

) (4

.
)
(
.

:2 ) I ( w ] [13 :

.
) y = f (x

( y , , ) : F ( y , , ) 0 w ( F ) 0

x
:1 I (w ) :W w
.

)) y = f ( x(t

) x(t .

:
w1 w2
.w w
1
2

) (7 ) (8
.

] w1 = [ y1 (), ] . w2 = [ y 2 (),

) y1 (. ) y 2 (. J
. y1 ( ) y 2 ( ) J
F :

B D P J B
R n +l +1 ) CD (B

1 B .

) CD (B g

- continually differentiable
4 2 1389

Journal of Control, Vol. 4, No. 2, Summer 2010

] .g +
= yT g +
y

)(7

[= g

w W :
(Ty g + )d

)( 8

-4-4

d = lim

] [0,m

m 1

= ) w ( g

[(

.
) 2 (:

)[0,1

dy
] .g + )d = lim ([ ]T . + )d
m 1
y

y d
] [0,m

] [0,m

= lim

m 1

d
( )d = lim ( y (m ),m ) ( y (0),0) =
m 1
d
] [0, m

) ( y0 = e, lim y ( ) = 0 ) (4
1

) C c (

) (F ) = Fd = (F ). F C c (

:] [11 ][16 .

w ( ) = g d = (0, m ) (e,0) =

)( 9

)[ 0,1

= y j
( y , = 0) = ( y , = 1) = 0
) S ( ) S (
.

m :
lim m = 1

) S () (10

+ g j

j = 1,..., n

[ y (.), ] W:

= + g j ) d

(y j

]m

[0,

( y , ) d = lim

m 1

{y& j g j } d

D P J m m

g j y j j g y
j

;m < 1

;] m = D P J m ; J m = [0,m

] [0,m

m 1

] [0,m

] .[13 m
R q q = 1, 2,...

j = yj

] [17
n+l +1

= ) w ( j

)[0,1

lim

} lim{y j

m 1

)(11

m = D P J m R .
m

.
I ( m w ) : F Cc ( m ), [ y, ] W
mw = F ( y, , )d
Jm

] [ y (.), ) (4

y& j = g j

( y , = 0) = ( y , = 1) = 0 .

) Cc ( m m
mw ) Cc (m
) m ( F .

( y , ) d = 0

)(12

)(14

= ) w ( j

)[0,1

]: [15

) (4

) m (F ) = F( y( ),, )d = F dm = m (F ) F Cc (m
m

Jm

)(15

.
max imize
)(13
) w ( g ) = (0, m ) (e,0) = CD( B
)j = 1,..., n S (

w ( j ) = 0

- Riesz Representation Theorem

4 2 1389

Journal of Control, Vol. 4, No. 2, Summer 2010

:2

)opt (g ) = lim( m (g ))i = (0,1) (e,0

m ) M (m

opt ( j ) = lim ( m ( j )) k = 0
k

:
Maximize
) m M + ( m

) CD( B

)s.t. : m ( g ) = (0,1) (e,0

m ( j ) = 0

S ( m ), j = 1,..., n

Am opt Am
. Am )
].([13

-5-4

) 0, m M + ( m

)(16

: 13 15 .

) (16

) ) ( M + ( m

) (16 Am

) (16 .

:3
opt Am

) (16 Am
) M + (m

) Am ( N1 , N 2

.
] [13

. Am

Am . ) M + ( m

) M + ( m

. )(16

Am ] [13 Am

N 1

i = 1,..., N 1

:
] [14

} Gm = {m : m M + ( m ), m (1) = m
. Am Gm

) =

)(7 = 1

) i CD (B
) CD ( B . i
n y .

)(8

( = ( g ) = (1) = m
] [13 Am
. } {( m ) i

N 2

j ,h

) S ( m

h
+ g j h j = 1,...., n, h = 1,..., N 2

) h {sin( 2r ),1 cos( 2r

}r = 1,2,..

j ,h = y j

Am ) Am M + ( m

} {( m ) i ) M + ( m .

) (16

M + (m ) m

) opt

Am

opt

.( ( m ) i
: opt

.
:3

) Am ( N1 , N 2

4 2 1389

Journal of Control, Vol. 4, No. 2, Summer 2010

) m (ig ) = i (0, m ) i (e,0

m ( j , h ) = 0

i = 1,..., N1
h = 1,..., N 2 , j = 1,..., n

lim

) N1 , N 2 A ( N , N
m
1
2

Am

i = 1,..., N1

j = 1,..., n, h = 1,..., N 2

) Sup ( ) = Sup(
: III.1 11 .

< ) j ,h ( z k* ) j ,h ( z k

) opt (1) = m 3 (

= m

*
k

k =0

) ( 1

) ) ( ( z ) (zk
*
k

g
i

)(16 4 ].[15

g
i

k =1

= (opt )

:4
) Am ( N1 , N 2

k =1

i ,k

) opt = ( z

0, z m , N = N1 + N 2
*
k

4 ) ( y , , z
*
k

*
k

k =1

) ) (opt ) j,h = ( j,h (z ) j ,h (zk

:
*
k

*
k

N
k =1

k =1

max ig (zk* ) ig (zk ) <

*
k

*
k

*
k

g
i

) k* ( ig (zk* ) ig (zk ) ) k* max ig (zk* ) ig (zk

i ,k

opt

< ) ig ( z k* ) ig ( z k

*
k

k =1

) k* ( j,h (zk* ) j,h (zk ) ) k* max j,h (zk* ) j,h (zk

h

) ( z ) M + ( m .

k =1

k =1

k* max j,h (zk* ) j,h (zk ) <

F :

k =1

( z) F = F ( z) F Cc ( m ), z m
4
*

* zk k

*zk

4 4 ) (16
:
Maximize

m
N

k ig (zk ) i (0,1) + i (e,0) i = 1,...,N1

s.t. :

k =1

:4
Z m
m

.
) M + ( m :

i = 1,..., N1
j = 1,..., n, h = 1,..., N 2

) ( z k

*
k

k =1

i j,h

j , h ( z k )

k =1

)(17
*k

( opt )ig <

( opt ) j ,h <

) ( z k ) j ,h ( zk .
g
i

) ( zk N N m

= . zk Z
m

zk .

:
g

j = 1,...,n, h = 1,...,N2

z k Z m

z k

-5

4 5 .
4 5 :
- Rosen Bloom
4 2 1389

Journal of Control, Vol. 4, No. 2, Summer 2010

2


.
3
) (
:1 4 5

]kA = [200 200 200 200

. 5 6 6-5

:1

)( .

]e1 = [0.39 -0.85 -0.35

]k1 =[92 90 92 90

]e2 = [.3 .6 .74

]k2 =[87 87 87 87

5

6

]k3 = [102 123 101 123] e3 = [-0.24 -0.85 -0.46

5-6

- 6
:2
-6

x 10

kA

0.0012

k1
0.0033

k2
0.002

k3
0.00225

) (5

41

) (6

0
8

2
-6

x 10

-6

x 10

1
0

21

0.21

0.23

0.208

0.343

) (5-6

6
4

0.0012

0.0028

0.0036

0.00225

31

- 5

-6

.



.


.

- 6-5

:2

4 2 1389

Journal of Control, Vol. 4, No. 2, Summer 2010

) (18 ) (19

:
1
1
)) E&qi _ = ( (EFDi_ Eqi _ + (xdi xdi )Idi
di
(1 )2

-7 :

1
1
)) E&di _ = ( (Edi _ + (xqi xqi )Iqi
qi
(1 )2

1
(Pmi (IdiEdi _ + IqiEqi _ ) + (xqi xqi )IdiIqi
Mi

-1-7

1
(1 )2

IEEE ].[18

1
) E& di = (Edi + ( xqi xqi )I qi
qi
1
)) (Pmi (I di Edi + I qi Eqi ) + (xqi xqi )I di I qi D(i s
Mi

= & i

) & j1_ = b ( j 1

j = 2,3,4; i = 1,..,4

)(18

))) D(i _ s

1
(1 )2
1
1

) EFDi _
(1 ) 2
TAi

:
1
) E& qi = (EFDi Eqi + ( xdi xdi )I di
di

) ) &j1_ = (b (j _ 1_

K K
E& FDi _ = ( Ai E 2i
TAi
1
1
1
X& E1i _ = (
) X E1i _ + VTi
TRi
TRi
(1 ) 2
T
1 T
1
X& E 2 i _ = ( ( Ci 1) X E1i _
X E 2i _ Ci VTi
TBi TRi
TBi
TBiTRi
1
1
) + Vrefi
TBi
(1 ) 2

-3-7

100 20

3-4

:
xdi = 1.8 pu, xqi = 1.7 pu
xdi = 0.3 pu, xqi = 0.55 pu

di = 8 sec, qi = .4 sec
] [18 .

= { j21 , i2 , j1i } j = 2,3,4; i = 1,...,4

) (7 .

= 1 :
2

1
( Pm1 ( I d 1 Ed 1 _ + I q1 Eq1 _ ) + ( xq1 xq 1 ) I d 1 I q1
Mi

( = &i _

1
(1 ) 2

( g = 21

) )) D(1 _ s

h :

}r = 1,2,..

:3

) h {sin( 2r ),1 cos(2r

:
K K
1
E& FDi = Ai E 2i
E FDi
TAi
TAi

)(19

1
1
X E1i +
VTi
=
TRi
TRi

h
+ g j h j = 1,...., 28, h = 1,..., 2

j ,h = y j

) (16
X& E1i

) (16 3

T
1 TCi
1
1
X& E 2i =
X E 2i Ci VTi +
( 1) X E1i
Vrefi
TBi TRi
TBi
TBiTRi
TBi

1 MATLAB .
1 3

TA = TR = 0.01, TB = 10, TC = 1

5 6 ) 6-5 (

-2-7
) (18 ) (19
:

[ 21 , 31 , 41 , 1 ,.., 4 , Eq1 ,..., Eq 4 , Ed 1 ,..., Ed 4 ,

1 ,.., E FD
4 , X E1 1 ,..., X E14 , X E 2 1 ,..., X E 2 4 ]T
E FD
4 2 1389

e1 = [0.39 0.85 0.35 0... 0 ]1T27

e2 = [0.3 0.6 0.74 0... 0 ]1T27
e3 = [ 0.24 0.85 0.46 0... 0 ]1T27
Journal of Control, Vol. 4, No. 2, Summer 2010

10

[7] A. S. Bazanella, P. Kototovic, A. S. Silva, 2004, A
dynamic extension for

Lg V controllers", IEEE trans

on automatic control, pp.588-592.

[8] M. A. Pai, 1981, "Power system stability", NorthHolland publication company.
[9] M. Galaz, R. Ortega, A. S. Bazanella, A. M.
Stankovic, 2003, An energy shaping approach to
[10] A. G. Bahbah, A. A. Girgis, 1999, An investigation
on the effect of line reclosing on transient stability
assessment for multi-mashine power systems, IEEE,
Power Engineering Society Summer Meeting, pp.
1082-1087.
[11] J. E. Rubio, 1986, Control and Optimization: The
Linear Treatment of Nonlinear Problems, Manchester
University Press, Manchester, and John Wiely, New
York and London.
[12] H. K. Khalil, 1996, Nonlinear systems Prentice
Hall 2nd Edition.
[13] Walter Rudin, 1986, Real and complex analysis,
McGraw-Hill publisher.
[14] G.Choque, 1996, Lectureson Analysis. NewYork:
Benjamin.
[15] A. V. Kamyad, M.Keyanpour, M. H. Farahi, 2007, A
New Approach for Solving of Optimal Nonlinear
Control Problems. , Appl. Math And Computations.
[16] Barra G, 1981, "Measure theory and integration", John
Willy and sons.
[17] J.Munkres, 2000, "Topology", published by Prentice
Hall; 2nd edition.

Journal of Control, Vol. 4, No. 2, Summer 2010

[1] R. W. Brockett, Hongyi L, December 2003, A light

weight rotary double pendulum: maximizing the
domain of attraction, Proceeding of 42nd IEEE
conference on decision and contro, pp 3299-4404.
[2] Y. Cao, Z. Li, D.G. Ward, January 2002, An
antiwindup approach to enlarging domain of attraction
for linear systems subject to actuator saturation, IEEE
transactions on automatic contro, vol. 47, no.1, pp
140-145.
[3] G. Chesi, 2004, Computing output feedback
controllers to enlarge the domain of attraction in
polynomial system", IEEE Trans. On automatic
contro, vol. 49, no. 1, pp. 1846-1850.
[4] F. Hashemzade, M. J. Yazdanpanah, 2006, Semiglobal enlargement of domain of attraction for a class
of affine nonlinear systems", IEEE, International
conference on control applications, pp. 2257-2262.
[5] Trong B.Ngye, M. A. Pai, I. A. Hisken, 2000, Direct
computation of critical clearing time using trajectory
sensitivities", IEEE, Power engineering society
summer mitting, vol 1, pp. 604-608.
[6] S. Jiriwibhakorn, A.H. Coonick , 2000, Fast critical
clearing time estimation of a large power systemusing
neural networks and Sobol sequences , IEEE Power
Engineering Society Summer Meeting , pp 522-527
vol. 1.

1389 2 4