TELKOM
NIKA
, Vol.9, No.1, April 2011,
pp. 65~7
2
ISSN: 1693-6
930
accredited by D
G
HE (DIKTI
), Decree No: 51/Dikti/Kep/2010
¢
65
Re
cei
v
ed Jan
uary
5
th
, 2011
; Revi
sed Ma
rch 5
th
, 201
1; Acce
pted April 7
th
, 2011
Power Oscillation Damping Control using Robust
Coordinated Smart Devices
Tumiran
1
, Cuk Supriy
adi Ali Nandar*
2
, Sarji
y
a
3
1,3
Department of Electrical En
gin
eeri
ng a
nd Informa
tio
n
T
e
chno
log
y
, Gad
j
a
h
Mada U
n
iver
sit
y
(UGM)
Jl. Grafika no.2
Kampus UGM
,
Yog
y
akarta 5
528
1
2
Badan Pe
ngk
ajia
n da
n Pen
e
r
apa
n T
e
knolo
g
i (BPPT
)
Jl. MH
T
hamrin No.8 Jakarta
Pusat
e-mail: tumiran@te.ugm.ac.id
1
, cuksupri
y
ad
i
@
gmai
l.com*
2
Abs
t
rak
Sistem i
n
terko
neksi tena
ga li
strik denga
n d
a
ya
da
mp
in
g yang ren
d
a
h
da
pat me
nja
d
i pe
nyeb
a
b
terjad
inya
mas
a
la
h osil
asi fre
k
uens
i rend
ah
pad
a sist
e
m
tenag
a listrik. Pa
da kon
d
isi k
e
rj
a yang ekstri
m,
stabilit
as siste
m
day
a (PSS)
bisa
m
e
n
g
a
l
a
m
i ke
ga
gal
an
dal
a
m
mered
a
m
osi
l
asi ters
e
but. Makala
h i
n
i
me
nya
jika
n
d
e
s
ain k
end
al
i ro
bust pa
da PSS
dan k
a
p
a
sitor
seri terke
nda
li t
h
yristor (T
CSC
)
secara s
i
mult
an
untuk mered
a
m
osil
asi pa
da
sistem interko
neksi t
ena
ga li
strik. Untuk meng
gara
n
si ke
nda
li yang kok
oh,
teknik pertu
ba
si aditif inv
e
rs
digu
nak
an u
n
tuk merepr
es
entasik
an keti
dakp
a
stia
n sis
t
em p
ada sist
e
m
tenag
a listrik s
eperti p
e
rub
a
h
an par
a
m
eter
sistem,
pe
mba
ngkita
n
siste
m
dan kon
d
isi p
e
mbe
ban
an ya
n
g
sulit
d
i
pre
d
iksi
secara pasti.
P
ada
stu
d
i ini, al
gorit
ma ge
neti
k
a
di
gu
naka
n
untuk
me
na
la para
m
eter
ke
n
dali
pad
a PSS dan
TCSC secara s
i
multa
n
. Studi simulas
i
te
la
h dilak
u
ka
n pa
da
sistem b
u
s tak berhi
ngg
a
me
s
i
n
tungg
al (SMIB
)
untuk
memb
uktikan
ba
hw
a kend
al
i
yan
g
telah
di
des
ai
n
me
mp
uny
ai
unj
uk ker
j
a
da
n
kekoko
ha
n yan
g
leb
i
h b
agus d
i
ba
ndi
ngk
an d
e
nga
n ken
dal
i konve
n
sio
n
a
l
.
Ka
ta
k
unc
i:
al
gorit
ma g
e
n
e
tika, kenda
li kok
oh, pertub
a
si a
d
itif invers, PS
S, T
C
SC
A
b
st
r
a
ct
T
he lack
of da
mp
in
g of the
el
ectr
omech
anic
a
l osc
ill
ation
modes
us
u
a
lly c
auses s
e
ver
e
prob
le
ms
of low
freque
n
cy oscill
ations
in interc
on
nect
ed pow
er syst
ems. In the
ex
trem
e
oper
atin
g cond
itio
ns, PSS
may fail to da
mp p
o
w
e
r oscillatio
n
. This pa
per pr
ese
n
ts a robust coordi
nated d
e
sig
n
of pow
er syst
em
stabili
z
e
r (PSS) and thyristor controll
ed
series ca
p
a
cit
o
r (TCSC) to damp pow
er
oscillati
on in
an
interco
nnect
e
d
pow
er system. The i
n
vers
e ad
ditive
per
turbatio
n is a
ppli
ed to r
epr
esent u
n
structur
e
d
uncerta
inties i
n
the pow
er system suc
h
as variatio
ns of syste
m
para
m
eters,
system ge
nera
t
ing and lo
ad
in
g
cond
itions. In a
dditi
on, ge
netic
algor
ith
m
is e
m
p
l
oy
ed to
sea
r
ch a robust tu
nin
g
to the con
t
roller p
a
ra
met
e
rs
of both PSS and TCSC simult
ane
ously. Si
mulati
on studies
have be
en do
n
e
in a single machi
ne infin
i
te bus
system to conf
irm that the
pe
rforma
nce a
n
d
robustness
of
the prop
ose
d
control
l
er
are s
uper
ior to that o
f
the conve
n
tio
n
a
l contro
ller.
Ke
y
w
ords
: ge
netic al
gorith
m
,
inverse a
dditiv
e
per
turb
atio
n, PSS, robust control, TCSC
1. Introducti
on
Powe
r sy
ste
m
bla
ck
out
due to l
o
w f
r
eque
ncy o
s
ci
llation be
com
e
s
se
riou
s p
r
oblem i
n
the powe
r
sy
stem. To pre
v
ent this pro
b
lem, t
he applicatio
n of smart techn
o
l
ogy such as
PSS
and flexible alternating current tran
smission sy
stems (FACTS) de
vice to provide an addition
al
dampin
g
of powe
r
system
is highly nee
ded. At prese
n
t, power sy
stem stabilizer
(PSS) has been
selected as a cost effective dev
ice to damp power oscillation
vi
a the excitation system
[1].
Several ap
proache
s ba
se
d on mod
e
rn cont
rol the
o
rie
s
have b
een succe
s
sfully applied
to
desi
gn PSSs, such as eigenvalue assi
gnment [2,
3], and linear quadratic
regulator [4]. These
works
have
c
onfirmed the s
i
gnific
ant perform
ance of PSS. However, P
SS may s
u
ff
er a
dra
w
ba
ck of
being li
able t
o
ca
use g
r
ea
t variati
ons i
n
the voltage
profile
and t
hey may eve
n
result in le
adi
ng po
we
r fa
ctor op
eratio
n
and lo
si
n
g
sy
stem
stability unde
r
severe distu
r
b
a
n
c
e
s
[5].
The appli
c
ati
on of FACTS devices
su
ch
as th
yristor controlle
d se
ri
es ca
pa
citor (TCSC),
unified power flow controll
er (UPF
C), a
nd static
var comp
en
sato
r (SVC), which using relia
ble
and hig
h
-sp
e
ed ele
c
tro
n
ic
device
s
to da
mp ele
c
trom
e
c
ha
nical oscil
l
ations in
po
wer sy
stem
s h
a
s
Evaluation Warning : The document was created with Spire.PDF for Python.
¢
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 9, No. 1, April 2011 : 65 – 72
66
been ob
se
rve
d
[6].
In this study, TCSC is installe
d
in the power sy
stem to tackle the limitation o
f
PSS. Several previous works has pai
d
attent
ions
to tuning convent
ional lead/lag PSS and
TCSC pa
ram
e
ters
simulta
neou
sly by heuri
s
tic meth
ods such as
simulate
d si
mulated an
ne
aling
[7] and geneti
c
algo
rithm [8
]. In these stu
d
ies, ho
weve
r, the uncerta
inty model is
not embed
de
d
in the mathematical model of the power system.
Therefore, the robust
stabilit
y margin of the
system in the
s
e works m
a
y not be guara
n
teed
in the face of severa
l unce
r
taintie
s
.
To get the robu
st controll
er,
H
∞
cont
ro
l has bee
n a
pplied to de
sign of a robu
st PSS
config
uratio
n
[9]. In
this work, the de
sign
ed
H
∞
PSS via mi
xed sensitivity
approach have
confirmed th
e signifi
cant
perfo
rman
ce
and hig
h
ro
b
u
stne
ss. In this ap
pro
a
ch, howeve
r
, du
e to
the
trade
-off relation between sen
s
itivity
function
an
d com
p
leme
ntary se
nsitiv
ity function, the
weig
hting fun
c
tion
s in
H
∞
control de
sig
n
can
not be
sele
cted e
a
si
ly. Moreover,
the stru
cture
of
conve
n
tional
H
∞
co
ntrolle
r is high order
and co
mplex whi
c
h
is different from the conve
n
tional
PI
or lead/la
g controlle
r. De
spite
the signif
i
cant potentia
l of cont
rol techni
que
s me
ntioned ab
ove,
power syste
m
utilities still
prefe
r
the
co
nventional l
o
w o
r
de
r PI or
lead/lag
co
ntroller. Thi
s
i
s
due
to the ease of
implementati
on, t
he long-t
e
rm reli
ability, etc.
This paper propose
s the
robust coor
di
nated PSS and TCS
C
to damp low frequency
oscillation in an interconnected power
system. To ta
ke system uncertaintie
s into account in the
control desi
g
n, an inverse additive perturb
ation
[10
]
is applied to represent all unstru
c
tu
red
uncertaintie
s
in the syste
m
modelin
g. More
ov
er, the perfo
rman
ce con
d
ition
s
in the dampi
ng
ratio an
d the
real p
a
rt of th
e domin
ant m
ode i
s
ap
plie
d to formul
ate the optimi
z
ation proble
m
. In
this work, the structur
e of the both proposed
controller of
PSS and TCS
C
are the first
-
order
lead/lag
com
pen
sator. It is ea
sy to implement in
th
e real
syste
m
. To achi
eve the co
ntroll
e
r
para
m
eters, the geneti
c
algorithm (GA) is used to
solve the optimization p
r
o
b
lem. Simula
tion
studies expli
c
itly show that
the proposed robust PSS and TCSC
are very robust to various
system u
n
ce
rtainties in co
mpari
s
o
n
to
that of conven
tional co
ntroll
er [3,7].
This pa
pe
r is orga
nized a
s
follows. First,
pow
e
r
syste
m
modeling i
s
explaine
d in se
ction
2. Section 3
pre
s
ent
s the
prop
osed de
sign meth
od
for optimization of coo
r
di
n
a
ted rob
u
st P
SS
and TCS
C
param
eters using GA. Subseq
uently,
se
ction 4 sho
w
s the result
s and analysi
s
.
Finally, the concl
u
si
on is g
i
ven.
2. A Robus
t Tuning to th
e Con
t
roller Parameters
2.1 Po
w
e
r S
y
stem Modeling
A single machine infinite bus (SMIB) system
sho
w
n in Figure 1 is used to explain the
desi
gn of propo
sed ro
bu
st coordinate
d
PSS and
TCSC.
The lineari
z
e
d
forth-o
r
de
r mod
e
l of
SMIB system [1]
for robust
PSS and TCSC desi
gn is
depi
cted in Figure 2. System param
eters
and notatio
ns are given in [7]. The initial con
d
it
ion u
s
e
d
as the de
si
gn co
ndition
of the propo
sed
PSS is
e
P
=
1.0
p.u.,
e
Q
= 0.015 p.u.
Figure 1. Single line of SMIB system.
Figure 2. Linearized of SMIB
system wit
h
PSS and TCSC
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
¢
Power Oscillation Dam
p
ing Control using Robu
st Coordinated Sm
art Devices (T
um
iran)
67
The linea
ri
ze
d system in F
i
gure 2
can b
e
writen a
s
⎥
⎥
⎥
⎥
⎥
⎦
⎤
⎢
⎢
⎢
⎢
⎢
⎣
⎡
Δ
Δ
Δ
Δ
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎦
⎤
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎣
⎡
−
−
−
′
′
−
′
−
−
−
−
=
⎥
⎥
⎥
⎥
⎥
⎦
⎤
⎢
⎢
⎢
⎢
⎢
⎣
⎡
′
Δ
′
Δ
′
Δ
′
Δ
fd
q
A
A
A
A
A
d
d
d
fd
q
E
E
x
T
T
K
K
T
K
K
T
T
K
T
K
M
K
M
D
M
K
E
E
'
6
5
0
0
3
0
4
2
1
'
1
0
1
0
0
0
0
377
0
ω
δ
ω
δ
⎥
⎦
⎤
⎢
⎣
⎡
Δ
Δ
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎦
⎤
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎣
⎡
−
′
−
−
+
TCSC
PSS
A
V
A
A
A
d
q
P
u
u
T
K
K
T
K
T
K
M
K
0
0
0
0
0
(1)
In short, the state equation
of
system ca
n be expre
ssed as
u
B
X
A
X
Δ
+
Δ
=
Δ
•
(
2
)
u
D
X
C
Y
Δ
+
Δ
=
Δ
(
3
)
ω
Δ
=
Δ
)
(
s
K
u
(
4
)
Whe
r
e the state vector
[
]
T
fd
q
E
e
X
Δ
Δ
Δ
Δ
=
Δ
'
ω
δ
, the outp
u
t vector
[
]
Y
ω
Δ=
Δ
,
[]
TCSC
PSS
u
u
u
Δ
Δ
=
Δ
are the cont
rol output sign
al of t
he
PSS and TCSC, whi
c
h uses o
n
ly th
e
angul
ar velo
city deviation (
ω
Δ
) as a fee
dba
ck in
put sig
n
a
l
.
2.2 Configur
ation of
PSS
and TCSC
Controller
As
shown in
Figure 3, the PSS c
ontroller (
K
PSS
) i
s
repre
s
e
n
ted b
y
a simple
1
st
orde
r
lead/lag cont
rolle
r and wa
sh out whi
c
h
uses
system
frequen
cy deviation (
∆ω
) as a feedba
ck
input si
gnal.
The power
system stabilizer (PSS) is
used to provide the a
dditional damping
vi
a
the excitation
system. Moreover, the TCSC blo
ck di
agra
m
is dep
icted in Figure 4. The TCSC
diagram con
s
ists of two transfe
r fun
c
tio
n
s, i.
e. the TCSC mo
del a
nd the lead/l
ag ba
sed p
o
w
er
oscillation
co
ntrolle
r. Base
d on [7], the
TCSC
ca
n
be
modele
d
by the first
-
orde
r
transfe
r fun
c
ti
on
with time c
ons
tant
T
C
= 0.
05 se
c. In this wo
rk, TCS
C
co
ntro
ll
er i
s
presented
by pra
c
tically
a 1
st
orde
r le
ad/la
g co
ntrolle
r
and
wa
sh o
u
t
with sin
g
le
feedba
ck in
put sig
nal, system freq
ue
ncy
deviation (
∆ω
). Note tha
t
the system
in equation
(2)
i
s
a mu
lti-input sin
g
l
e
-outp
u
t (MI
S
O)
system. He
re
, the proposed desi
gn ap
proa
ch is
ap
plied to desi
gn a robu
st coo
r
din
a
ted PSS
controlle
r (
K
PSS
) and TCS
C
co
ntrolle
r (
K
TCSC
) simultaneo
usly.
Figure 3. Bloc
k
diagram
of
PSS c
ontroller
(
K
PSS
)
Figure 4. Block di
agram of
TCSC contro
ller (
K
TCSC
)
Evaluation Warning : The document was created with Spire.PDF for Python.
¢
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 9, No. 1, April 2011 : 65 – 72
68
3. Resea
ch Metho
d
In this sectio
n, GA is applied to search t
he controll
er paramete
r
s with off lin
e tuning.
The flo
w
cha
r
t of prop
osed
cont
rol d
e
si
g
n
is
sh
own
in
Figure 5. E
a
ch
step
prop
ose
d
meth
od
is
explained a
s
follows.
Figure 5. Flow ch
art of the
propo
se
d de
sign
Figure 6. Fee
dba
ck
system
with inverse
additive pertu
rbation
Figure 7. D-shape regi
o
n
in the s-plane
Step 1
Gene
rate the obje
c
tive function for GA optimi
z
ation.
In this study, the perform
ance and ro
bust st
ability
condition
s in the inverse
additive
perturbation desi
gn ap
proach is adopted to desi
gn
a both
robust PSS and T
C
SC.
The
conventional
PSS and TCSC cont
roller with a 1
st
-ord
er lead/la
g co
ntrolle
r are
re
pre
s
ente
d
by
ω
Δ
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
+
+
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
+
=
Δ
1
1
1
2
1
sT
sT
sT
sT
K
u
W
W
P
PSS
(
5
)
ω
Δ
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
+
+
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
+
=
Δ
1
1
1
2
1
sT
sT
sT
sT
K
u
W
W
T
TCSC
(
6
)
whe
r
e,
PSS
u
Δ
,
TCSC
u
Δ
and
ω
Δ
are the co
ntrol output
sign
al and the roto
r spe
e
d
deviation at
both of PSS and TCSC,
res
p
ec
tively;
P
K
and
T
K
are
a
controller gain of PSS
and
TCSC, respec
tively;
W
T
is a wa
sh-out time con
s
tant (s): and
1
T
and
2
T
are time c
o
ns
tants
(s).
In this paper,
the control para
m
eters
K
,
1
T
and
2
T
are optimized by GA based on th
e
followin
g
con
c
ept. As
sh
o
w
n in
Figu
re
6, the sy
ste
m
with inve
rse additive p
e
rturb
a
tion [10
]
is
applie
d to ta
ke the robu
st stability against uncerta
inties into accou
n
t. For a stable additive
uncertainty
A
Δ
, the cl
ose
d
loo
p
system i
s
robu
st if the controlle
r
K
st
abilizes the
nominal pl
ant
G
. Based o
n
the small g
a
in
theorem, the
system is
sta
b
le if
()
1
1
/
<
−
Δ
∞
GK
G
A
(
7
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
¢
Power Oscillation Dam
p
ing Control using Robu
st Coordinated Sm
art Devices (T
um
iran)
69
Then
()
1
1
/
<
−
Δ
∞
∞
GK
G
A
(
8
)
This
yield
(
)
∞
∞
−
<
Δ
GK
G
A
1
/
/
1
(
9
)
The ri
ght ha
n
d
sid
e
of eq
u
a
tion (9
) impli
e
s the
maxim
u
m ro
bu
st sta
b
ility margin
again
s
t inverse
additive pert
u
rbatio
n. Then, the robu
st stabilit
y margin of the
close
d
loop
system can
be
guarantee
d in
terms of the additive stabil
i
ty margin (A
SM) as,
∞
−
<
))
(
)
(
1
/(
)
(
1
s
K
s
G
s
G
ASM
(
1
0
)
By minimizing
()
1
GG
K
∞
−
, the robust
stability margin of the close
d
-lo
op sy
stem is a
near optim
u
m
.
In
this study, the problem con
s
trai
nts are the co
n
t
roller pa
ram
e
ters b
ound
s. In
addition to e
nhan
ce the
robu
st stabilit
y, another ob
jective is to i
n
crea
se the
dampin
g
rati
o and
place the clo
s
ed
-loo
p eige
nvalue
s of hybrid wi
nd
-die
sel po
we
r system in a D-sh
ape re
gion [1
1].
The conditio
n
s
will pla
c
e the sy
ste
m
clo
s
ed
-lo
op eige
nval
ues i
n
the
D-sha
pe region
c
h
ar
ac
te
r
i
z
ed b
y
spec
ζ
ζ
≥
and
spec
σ
σ
≤
as sh
own in Fig
u
re
7.
Therefore, th
e desi
gn prob
lem can b
e
formulate
d
as t
he followi
ng o
p
timization p
r
oblem.
Minimize
()
1
GG
K
∞
−
(
1
1
)
Subject to
,
s
pe
c
s
p
e
c
ζζ
σ
σ
≥≥
max
,
,
,
min
,
,
i
c
i
c
i
c
K
K
K
≤
≤
,
max
,
min
,
i
i
i
T
T
T
≤
≤
,
2
,
1
=
i
Whe
r
e
ζ
and
spec
ζ
are a
c
tual
and
desi
r
e
d
dam
ping
ratio, re
spe
c
tively,
σ
and
s
pe
c
σ
are
actu
a
l
and de
sire
d
real part of the electro
m
ech
ani
cal mode,
,m
i
n
c
K
and
,m
a
x
c
K
are minimu
m
and
maximum gains of both
PSS and TCSC,
,m
i
n
i
T
and
,m
a
x
i
T
are minimum
and maximu
m time
constant
s of PSS and TCSC. The opti
m
izat
ion problem is solved by GA.
Step 2
Initialize the sea
r
ch para
m
eters for GA
. Define genetic parame
t
ers su
ch a
s
popul
ation
size, crossov
e
r, mutation rate, and maximum gen
erati
on.
Step 3
Ran
domly ge
nerate the init
ial solutio
n
.
Step 4
Evaluate obje
c
tive function
of each in
dividual in eq
uati
on (11
)
.
Step 5
Select the be
st individual i
n
the cu
rre
nt gene
ration. Check the max
i
mum gen
erat
ion.
Step 6
Incre
a
se the gene
ration.
Step 7
While th
e current gen
eration is le
ss t
han the
maximum ge
neratio
n, cre
a
te ne
w
popul
ation u
s
ing gen
etic o
perato
r
s and
go to step
4. If the current
gene
ration i
s
the maximu
m
gene
ration, th
en stop
4. Simulation Resul
t
s an
d Analy
s
is
In this sectio
n, simulation
studies in a single mach
ine con
nect
e
d to infinite
bus are
carrie
d out.
System para
m
eters are g
i
ven in [7].
In the optimization, the range
s of search
para
m
eters a
r
e set as foll
ows:
s
pe
c
ζ
is desi
r
ed dam
ping ratio is set as 0.5,
s
pec
σ
is desired re
al
part is
s
e
t as -0.2,
min
K
and
max
K
are minimum and maximum gain
s
of both PSS and TCSC
are set as 1 and 60,
min
T
and
max
T
are minimu
m and maximum time constant
s of both PSS
and T
C
SC
co
ntrolle
rs a
r
e
set as 0.01 a
n
d
1. More
ove
r
, the ran
g
e
s
of GA param
eters
are
set
as
follows: cro
s
sover p
r
o
bab
ility is 0.9, m
u
tation
pro
b
a
b
ility is 0.05, population
size is 10
0 a
n
d
Evaluation Warning : The document was created with Spire.PDF for Python.
¢
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 9, No. 1, April 2011 : 65 – 72
70
maximum ge
neratio
n is
10
0. The o
p
timization
pro
b
le
m is
solved b
y
genetic
alg
o
rithm [12]. A
s
a
result, the
desi
gned controllers
whi
c
h ar
e referred as “RPSS
and RT
CSC”
are
given
simultan
eou
sl
y as sho
w
n a
s
follows
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
1
0.4478
1
0.7437
1
5
5
58
.
38
+
s
+
s
+
s
s
=
RPSS
(12)
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
1
0.1336
1
0.4576
1
5
5
28
.
44
+
s
+
s
+
s
s
=
RTCSC
(13)
In simulation
studie
s
, the perfo
rman
ce
and ro
bu
stne
ss of the pro
posed co
ntrol
l
ers a
r
e
compared wit
h
PSS [3] an
d PSS and TCSC [7], that is
PSS[3] :
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
1
0.1000
1
0.685
1
5
5
7.091
+
s
+
s
+
s
s
=
PSS
(
1
4
)
PSS and TCSC [7]:
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
1
0.1000
1
0.2156
1
0.1000
1
0.1212
1
5
5
22.523
+
s
+
s
+
s
+
s
+
s
s
=
PSS
(
1
5
)
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
1
0.1000
1
0.1249
1
0.1000
1
0.0115
1
5
5
74.726
+
s
+
s
+
s
+
s
+
s
s
=
TCSC
(
1
6
)
Next, the perf
o
rma
n
ce and
robu
stne
ss o
f
the
RPSS and RT
CSC
is c
o
mp
ar
ed
w
i
th
PS
S
and T
C
SC [7
].
To evaluate the robustness of cont
rol
l
ers, the value of ASM of
PSS and TCSC
[7] and RPSS and RTCS
C are shown in Table 1.As shown in Table 1, the val
ue of ASM in
the
case of RPSS and RTCS
C is greater than that in
th
e case of PSS and RTCS
C [7].
It is indicate
that the better ro
bu
st sta
b
ility margin
of
the closed loop
syst
em
ca
n be
achi
eved by
th
e
prop
osed met
hod
In addition, t
he eig
envalu
e
s
corre
s
p
o
n
d
i
ng to the
electrome
c
h
a
n
ical m
ode
without
controllers, PSS and RTCSC [7] and RPSS and RTCS
C are listed in Table
2. Clearly, the
dampin
g
rati
o and
real p
a
rt of the o
s
cillation m
o
d
e
are
gre
a
tly enhan
ce
d with the pro
p
o
s
ed
controlle
r. On other hand, system witho
u
t controlle
r has neg
ative dampin
g
or un
stable
.
The limit
on ea
ch PSS
output (
PSS
u
Δ
)
and
T
C
SC
ou
tpu
t
(
TCSC
u
Δ
) is
±0.05
p.u. The
system re
spo
n
ses
with co
ntrolle
r are exami
n
e
d
unde
r thre
e
case studi
es
as in Tabl
e 3.
Table 1. Co
m
pari
s
on of ASM
Controller
ASM
PSS and TCSC [
7
]
0.6445
RPSS and RTCS
C
8.0319
Table 2. Com
pari
s
on of oscillation mode
S
y
stem
Eigen value and Damping ratio
Without Controlle
r
+0.30 ±
j
4.96,
ζ
= -0.06
-10.39 ±
j
3.28
,
ζ
= 0.954
PSS and TCSC
[7]
-5.074 ±
j
7.30
8,
ζ
= 0.57
-0.729 ±
j
3.17
9,
ζ
= 0.224
-14.35 ±
j
1.01,
ζ
= 0.98
RPSS and
RTCSC
-8.87 ±
j
8.62,
ζ
= 0.717
-9.61 ±
j
7.50 ,
ζ
= 0.788
-1.54 ±
j
1.31 ,
ζ
= 0.762
Table 3. Ope
r
ating co
nditio
n
Sy
s
t
e
m
paramete
r
Case 1
Normal
condition
Case 2
Light
Loading
Case 3
Heav
y
Loading
P (pu)
1.000
0.800
1.20
Q (pu
)
0.015
0.15
0.15
Xcsc (
pu)
0.000
0.000
-
0
.2
Cas
e
1:
Normal condition
First, a disturban
ce of 10 % (0
.1 p.u.) step resp
on
se
of
re
f
V
Δ
is applied
to
the system at
t
= 1 s.
Figure 8
sho
w
s th
e re
sp
o
n
se
s of
syste
m
frequ
en
cy deviation. System
without
cont
rolle
r ca
n’t
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
¢
Power Oscillation Dam
p
ing Control using Robu
st Coordinated Sm
art Devices (T
um
iran)
71
stabili
ze the power oscill
ation,
the oscill
ation become higher an
d unstabl
e. On other hand, PSS
[3], PSS and TCSC [7] and propose
d RPSSandRT
CSC are abl
e
to
damp power oscill
ations.
Ho
wever, the
overshoot a
nd setting ti
me of pow
er
oscillation in
ca
se of RPS
S
and RT
CS
C is
much lower
than those of
both PSS [3] and PSS and T
C
SC [7]. Next, to
evaluate power
capacities of both PSS and TCSC controller require
d for power oscillation stabilization. Figure
9
shows the controller outpu
t power deviation in case
1. Both
control
l
er power out
put of PSS and
TCSC [7] and RPSS and
RTCS
C
can
properly rem
a
in
withi
n
the allowable lim
its. However,
the
stabili
zing
effect of freq
uen
cy oscill
ation
by RP
SS and RT
CSC i
s
sup
e
rio
r
to that
of
PSSandTCS
C [7].
0
1
2
3
4
5
6
-3
-2
-1
0
1
2
3
x 10
-3
Ti
m
e
(
s
e
c
)
S
y
s
t
e
m
f
r
e
q
ue
n
c
y de
v
i
a
t
i
on (
p
u
)
W
i
th
o
u
t c
o
n
t
ro
l
l
er
PS
S [
3
]
PS
S &
T
C
SC
[
7
]
RPS
S
&
RTCS
C
Figure 8. Simulation re
sult
of case 1
0
1
2
3
4
5
6
-0
.
1
-0
.
0
8
-0
.
0
6
-0
.
0
4
-0
.
0
2
0
0.
02
0.
04
Ti
m
e
(
s
e
c
)
C
ont
rol
l
e
r
po
w
e
r out
put
de
vi
a
t
i
on (pu)
T
C
S
C
c
o
n
t
r
o
lle
r
o
u
tp
u
t
[
7
]
Pr
op
os
ed
T
C
S
C
c
o
n
t
r
o
l
l
e
r
ou
t
p
u
t
Pr
op
os
ed
P
S
S
ou
t
p
u
t
P
SS o
u
t
p
u
t
[
7
]
Figure 9. Con
t
roller p
o
wer
output deviati
on
0
5
10
15
20
25
30
35
40
45
50
0
1
2
3
4
5
6
7
x 1
0
-3
Ti
m
e
(
s
e
c
)
R
a
ndom
l
o
a
d
pow
e
r
de
vi
a
t
i
on (
pu kW
)
Figure 10. Ra
ndom loa
d
po
wer d
e
viation
0
5
10
15
20
25
30
35
40
45
50
-3
-2
-1
0
1
2
3
x 1
0
-5
T
i
m
e
(
s
ec)
S
y
s
t
em
freq
u
en
cy
d
e
v
i
at
i
o
n
(p
u
)
PS
S&
T
C
S
C
[
7
]
RP
S
S
&
R
T
C
S
C
Figure 11. Simulation resu
lt of case 2
Cas
e
2 :
Ligh
t loading con
d
ition
In ca
se 2, th
e ran
dom
po
wer input
(
m
P
Δ
) a
s
sho
w
n in
Fi
gure
10 i
s
inje
cted to the
system. The
respon
se of the system fre
quen
cy devia
tion in ca
se 2
is sho
w
n in
Figure11, the
damping eff
e
ct
of PSS and TCSC [7] i
s
deteriorated. On t
he
other hand, t
he fre
quency
oscillations
are
effectively stabilized by RPSS and RT
CSC. RPSS and
RT
CS
C is rarely sens
i
t
ive to the weak
line co
ndition.
Cas
e
3 :
He
a
vy loading co
ndition
In case 3, the electri
c
al power
output is
incr
eased. Fi
gure
12 shows
that the PSS and TCS
C
[7
]
fails to dam
p power
system. T
he frequency oscillation gradually
increases and diverges. In
Evaluation Warning : The document was created with Spire.PDF for Python.
¢
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 9, No. 1, April 2011 : 65 – 72
72
contrast, the prop
osed RP
SS and RTCSC can tole
rate
this situat
ion. The freq
uen
cy oscill
ation
is sig
n
ificantl
y
damped. These simul
a
tion re
sults
co
nfirm that the prop
os
ed controlle
r is very
robu
st agai
nst various op
erating co
nditio
n
s.
0
5
10
15
20
25
30
35
40
45
50
-3
-2
-1
0
1
2
3
x 1
0
-5
Ti
m
e
(
s
e
c
)
S
y
s
t
e
m
f
r
e
q
u
e
n
c
y
de
vi
a
t
i
o
n (
p
u)
P
SS
& T
C
S
C
[
7
]
RP
S
S
&
RT
C
S
C
Figure 12. Simulation resu
lt of case 3
5. Conclusio
n
In this
s
t
udy, a robus
t
design of coor
dinated PSS and TCS
C
has
been propos
ed. The
inverse additive perturbati
on is
used to take the robust stability
of the cont
roll
ed power
system
again
s
t syste
m
uncertainti
es. The d
e
si
gned r
obu
st
controlle
rs a
r
e the conve
n
t
ional 1
st
ord
e
r
lead/lag com
pen
sator. Mo
reove
r
, the controlle
rs use
only the speed deviation of generato
r
as
the feedba
ck signal input.
Therefor
e, the controlle
rs are ea
sy to realize in pra
c
tical po
wer
system. The
control effect
s and
robu
st
ness of the propo
se
d co
nt
rolle
r have b
een evaluate
d
by
variou
s ca
se
studie
s
. Simulation re
sults
confirm
that the pro
p
o
s
ed
controlle
r is superi
o
r to the
conve
n
tional
controlle
r in term
s of the
robu
stne
ss a
g
a
inst vario
u
s
uncertaintie
s
.
Referen
ces
[1]
Larse
n E, S
w
a
rm D. Appl
yi
ng
po
w
e
r s
y
stem
stabil
i
zers.
IEEE Transactio
n
s on P
o
w
e
r A
ppar
atus a
n
d
System
s
. 19
81
; 100(6): 30
34-
304
6
[2]
Gurrala G, Sen I. Po
w
e
r S
y
stem Stabi
li
ze
rs Desig
n
for Interconnec
ted Po
w
e
r S
y
stems.
IEEE
T
r
ansactio
n
s o
n
Pow
e
r Systems
. 20
10; 25(
2
)
: 1042-1
0
5
1
.
[3]
Abid
o MA. A nove
l
ap
pro
a
c
h
to conv
entio
nal
p
o
w
e
r s
y
s
t
em stabiliz
er
desi
gn us
ing t
abu s
earch
.
Internatio
na
l Journ
a
l of Electr
ical Pow
e
r an
d
Energy Syste
m
. 19
99; 21(
6): 443–
45
4.
[4
]
Gu
p
t
a
R
,
Bandy
o
p
a
dhy
ay
B, Ku
l
k
a
r
n
i
AM.
Po
w
e
r sy
ste
m
sta
b
i
li
se
r fo
r
mu
l
t
i
m
a
c
hi
ne
p
o
w
e
r sy
ste
m
usin
g robust d
e
centra
lise
d
pe
riodic
output fe
edb
ack.
IEE Procee
din
g
s - Co
ntrol T
heory &
Appl
icatio
ns
.
200
5; 152(
1): 3-8.
[5]
Rahim A, Nas
s
imi S. S
y
nchr
ono
us ge
nerat
or
damp
i
ng e
n
hanc
ement thr
oug
h coor
din
a
ted contro
l o
f
exc
i
ter an
d SVC.
IEE Proceedin
g
on Gen
e
r
a
tion, T
r
ans
mi
ssion & Distri
b
ution
. 19
96; 14
3(2): 211
–2
18.
[6]
Z
a
rgham
i M,
Cro
w
ML, Ja
gan
nath
an S. Nonli
n
e
a
r Co
ntrol of F
A
CT
S Controll
ers for Dampin
g
Interarea Oscillations in Po
w
e
r Sy
stems.
IEEE Trans. on Power Delivery
. 2
010; 25(
4): 311
3-31
21.
[7]
Abid
o M.A. Pole pl
acem
ent techn
i
qu
e for P
SS and T
C
SC-base
d
stabi
liz
er desi
gn us
in
g simul
a
te
d
ann
eal
in
g.
Transactio
n
s on El
ectrical p
o
w
e
r and e
ner
gy systems
. 20
00; 22
: 543-55
4.
[8]
Abde
l-mag
i
d YL, Abido MA. R
obust coord
i
na
ted
desig
n of excitati
on an
d
TCSC-bas
ed sta
b
iliz
er usin
g
gen
etic al
gorith
m
s.
Transactio
n
s on Electric
al
pow
er and e
n
e
rgy systems
.
200
4; 69(2): 12
9-14
1.
[9]
Yan T
C
. Appl
ying Optim
i
sa
tion Metho
d
to Po
w
e
r S
y
s
t
em Stabilis
er
Desig
n
-Parts
1 and 2.
T
r
ansactio
n
s o
n
Electrica
l
Po
w
e
r and Ener
g
y
System
. (199
7); 19: 29-4
3
.
[10]
Gu P, Petkov
Hr, Konstantin
ov MM. Robust
control des
ig
n w
i
t
h
MAT
L
AB.
Lon
do
n: Spring
er. 2005.
[11]
Rashi
d
i M, Rashidi F, Mona
var H.
T
uning
of pow
er system stabil
i
z
e
rs
via genetic a
l
gorit
hm for
stabili
z
a
tion of
power system
.
IEEE International Confer
enc
e on S
y
stems, Man and Cy
bernetics. Iran.
200
3; 5: 4649-
465
4.
[12]
GAOT
. A Genetic Algorit
hm fo
r F
unction Opti
mizatio
n
: A Matlab Implem
ent
ation. 20
05.
Evaluation Warning : The document was created with Spire.PDF for Python.