Indonesian J
ournal of Ele
c
trical Engin
eering and
Computer Sci
e
nce
Vol. 1, No. 3,
March 20
16, pp. 446 ~ 4
5
5
DOI: 10.115
9
1
/ijeecs.v1.i3.pp44
6-4
5
5
446
Re
cei
v
ed
No
vem
ber 1
2
, 2015; Re
vi
sed
Februar
y 8, 2
016; Accepte
d
February 1
7
, 2016
Power System Design using Firefly Algorithm for
Dynamic Stability Enhancement
Herlamba
ng
Setiadi*, Kar
l
O Jones
Dep
a
rtment of Electron
ics an
d Electrica
l
En
gin
eeri
ng,
Liver
poo
l Joh
n
Moores U
n
iver
sit
y
, B
y
r
o
m Street, Liverp
oo
l L3 3AE, Unit
ed
Kingd
om
e-mail: h
e
rlam
ban
g2
345
@g
mail.com
A
b
st
r
a
ct
Utilising additi
onal dev
ic
es i
n
pow
er systems have been devel
oped by i
n
dustry. Devices such as
a Pow
e
r System Stab
ili
z
e
r (PSS) and a Su
p
e
rcon
ductin
g
Magn
etic Ener
gy Storage (S
MES) are commo
n
l
y
empl
oyed
in
in
dustry. This w
o
rk inv
e
stigat
e
d
t
he c
oord
i
n
a
t
ion of
a PSS
and SMES
ap
plie
d to
a p
o
w
e
r
system
to enhance
dynamic
stability.
To obtain
optim
al coordinati
on, the param
e
ters
of the PSS and
SMES are tuned
using the F
i
refly
Algor
ithm (FA). The si
mulation
of the
power system
,
PSS, and SM
ES
has b
een
perf
o
rmed
usin
g M
A
T
L
AB and S
i
mu
link, a
nd th
e
F
A
run i
n
Matlab. F
o
r testi
ng the s
m
all s
i
gna
l
stability, th
e
ei
genv
alu
e
of th
e syste
m
w
i
l
l
be
invest
i
gate
d
, w
h
ile
for
dy
na
mic
stabi
lity
the syste
m
w
i
l
l
b
e
given an
exter
nal distur
bance. T
he rot
o
r angle
and frequency dev
iation
of
the power syst
em ar
e compar
ed
w
i
thout a controller, w
i
th a PSS and SME
S
inclu
ded, a
n
d
w
i
th the PSS and SMES tune
d by FA. The
simulati
on res
u
lts show
that the pr
o
pose
d
system can i
m
p
r
ove not only s
m
a
ll sig
nal sta
b
ility (steady st
ate
stability) b
u
t al
so dyna
mic stability.
Ke
y
w
ords
: P
o
w
e
r System
Stabil
i
z
e
r, Su
perco
nducti
ng
Ma
g
netic E
n
ergy Stor
age,
F
i
refly Al
gori
t
hm,
Dyna
mic Sta
b
il
ity
1. Introduc
tion
Electri
c
ity de
mand i
n
re
cent time
s h
a
s
b
e
come
a
liability and
i
s
m
o
re
than
likely to
increa
se rapi
dly in the future. Thi
s
gro
w
ing d
e
man
d
for powe
r
re
quire
s ele
c
tri
c
ity provide
r
s to
increa
se thei
r gen
eratin
g cap
a
city and
to ex
pand their di
strib
u
tion network, thus m
a
ki
ng th
e
entire
sy
ste
m
large. O
n
large
sy
stem
s the
r
e
are comm
on pro
b
lems
a
s
soci
ated with system
stability. Dist
urbances that affe
ct the
stability of the system
can be transi
ent
disturbance
or
dynamic di
sturba
nce. La
rge tra
n
si
ent
disturban
ce
i
s
o
ne that
o
c
curs a
s
a result
of a b
r
oke
n
transmissio
n line, whil
st dynamical distu
r
ban
ce
s a
r
e
minor o
n
e
s
, such a
s
loa
d
chang
es [1].
In the case of
dynami
c
sta
b
ility, the syst
em
is
often i
m
paired o
w
in
g to the effe
ct of load
changes, whi
c
h cause
th
e system
to have frequency oscillati
on.
Accordi
ng to Goshal
[2] the
oscillation
fre
quen
cy that
occurs i
s
rel
a
tively
low,
and
within
ra
nge
of 0.2 t
o
0.3
Hz. If this
oscillation n
o
t well da
mpe
d
,
magnitude
of these
os
cil
l
ation may ke
ep growi
ng u
n
til system lo
ss
synchro
n
ism. To overcom
e
these p
r
o
b
l
e
ms, ad
di
tion
al cont
rol eq
u
i
pment shoul
d be install
e
d
on
the sy
stem,
where the P
o
wer System
Stabilizer
(
PSS) [3] i
s
one
approach
that is often used.
PSS is an additional
cont
rol devi
c
e which
serv
es t
o
dampen the oscillation
frequency and
voltage on the generator.
Since t
he sy
stem is larger, the PSS o
n
its own is
not enough to
overcome th
e pro
b
lem, h
ence othe
r d
e
vice
s su
ch
as en
ergy storag
e sy
stem
s are re
quire
d. In
this e
r
a, th
ere a
r
e
a lot
of
energy
stora
ge a
pproa
ch
es th
at have
been
empl
oyed, sy
stem
s
su
ch
as Battery Energy Storage
(BES)
[4, 5, 6], Redox Flow Batteries
(RFB) [7, 8], Capasitive Energy
Storage
(CE
S
) [6, 9] and
Superco
ndu
cting Magn
et
ic Energy Storage (SMES
)
[6, 10-1
3
]. SMES
is on
e en
ergy
storage
app
roach that is
commonly u
s
e
d
in a p
o
wer
plant that giv
e
s a
c
tive po
wer
temporary in
the load. In Ansari a
nd
Velusa
mi [13
]
SMES has been u
s
ed
to improve the
perfo
rman
ce
of a wind-die
s
el hybri
d
po
wer
system.
The use of additional equi
pment on the elec
tri
c
power system such as PSS and SMES
that is n
o
t a
ppro
p
ri
ate in
stabili
zing th
e sy
stem
ca
n lead
to issues. T
he
pro
b
lems often
arise
owin
g to erro
rs in taki
ng the reference signal, or
the para
m
eters o
f
the
devices
are not optim
al.
Adding a con
t
roller that in
cre
a
ses
syst
em perfo
rma
n
ce
can ma
ke the system
more un
stab
le.
Owin
g to these issue
s
, an optimizatio
n met
hod shoul
d be used to overcome the
proble
m
s.
Nature al
way
s
give u
s
in
spiration, a
nd
ther
e a
r
e
a n
u
mbe
r
of co
mputation o
r
iteration
method
s in
sp
ired from the
appa
re
nt intelligen
ce
sh
o
w
n in
natu
r
e
[14]. Smart computation
can
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 446 – 455
447
be catego
rized into thre
e gro
u
p
s
: p
h
ysically inspired
su
ch
as the Sim
u
lated Anne
al
ing
Algorithm; so
cial in
spi
r
ed l
i
ke the T
abu
Search Al
go
ri
thm, or Impe
rialist Co
mpeti
t
ive Algorithm,
and la
stly bi
ologi
cally inspired
app
ro
a
c
he
s
su
ch
as
G
eneti
c
Algorithm
s, Particle
S
w
a
r
m
Optimizatio
n
, Artificial Im
mune Sy
ste
m
, Ant Colo
n
y
Optimizatio
n
, Bat Algorit
hm an
d the
Firefly
Algorithm [14
]. The Firefly Algorithm i
s
one of the
newer n
a
ture
inspi
r
ed al
g
o
rithm
s
, and
is
inspi
r
ed
by the beh
aviour
of firef
lies. A
pplication of t
he Firefly Al
gorithm
(FA) i
n
optimizatio
n of
power syste
m
s
have be
en
devel
ope
d. There are
a numb
e
r
o
f
application
s
su
ch a
s
in
the
resea
r
ch of
Niknam
[15]. In
that resea
r
ch a
FA
i
s
u
s
e
d
to fin
d
the
optimal
gene
ration level
for a
power sy
ste
m
, achiving the lowest pri
c
e with a re
serve co
nst
r
ai
nt. Another a
pplication of the FA
is a
s
a
tool t
o
tune th
e p
a
ram
e
ters of
a PI co
ntrolle
r ap
plied to
speed
co
ntrol
of a DC
se
ri
es
motor [16]. T
he Firefly Alg
o
rithm ha
s b
e
en used to a
ddre
s
s the e
c
onomi
c
emi
s
sion
s in a p
o
wer
system [17].
The FA ha
s n
o
t only been
use
d
in po
we
r syste
m
s b
u
t also in
cu
ltivation co
ntrol,
for
example
Roe
v
a O, Slavov [18] use
d
a
FA to
tune a
PID cont
roll
er for
glu
c
o
s
e co
ncentrati
on
control. All of
the resea
r
ch proven that F
A
is one
of the metaheu
ri
stic algo
rithm that very good
to
solve optimi
z
ation pro
b
lem
.
Some rese
arch h
a
ve be
e
n
pro
p
o
s
ed t
o
solve
dyna
mic
stability enha
ncement
. Like in
Parimi and friend
s re
sea
r
ch [19], they try to
enhance po
we
r system stability by using on
e
o
f
Flexible Alternating
Cu
rre
nt Tra
n
smi
ssions Sy
stem
(FACTS)
de
vices called
Interline P
o
wer
Flow
Controll
er (IPFC). An
other research to
enhan
ce
dynamic
sta
b
ility
of powe
r
system
have
prop
osed by Mohamm
ad
Kash
ki [20], they try to
im
prove dyna
mi
c stability in power sy
ste
m
by
combi
n
ing
P
SS and
one
FACT
devi
c
e
s
calle
d S
t
atic Pha
s
e
Shifter (SPS
). Kashki
u
s
ed
Combi
nato
r
ia
l Discrete a
n
d
Contin
uo
s Action Rei
n
fo
rce
m
ent Le
arning Automat
a
to tune PSS
and SPS parameter.
Usm
an re
se
a
r
ch [21], try to enhan
ce t
he st
ability by controllin
g excitation sy
stem of
power
system usi
ng Aut
o
matic V
o
ltage Regul
ator
(AVR
)
combine by PSS,
this time Particle
Swarm
Optim
i
zation
have
use
d
by them
to optimiz
e
d
para
m
eter
of AVR and PS
S. Different from
others, Haki
m and friend
s [22], try to improve dynamic
st
ability by designing PSS based of
Fuzzy PID. They combi
ne
2 Artificial Intel
ege
nt whi
c
h
were F
u
zzy and Ge
netic
Algorithm.
This
research proposed in
vo
lves the coordination of
a Po
wer Syst
em Stabilizer (PSS)
and an
ene
rg
y
storage
de
vice calle
d s Superco
ndu
cting Ma
gneti
c
Energy Storage
(SMES),
to
improve
the dynamic stability
of
a power
system. T
o
get improv
ed perform
a
nce of the
devi
c
es,
the parameters of the PSS and SMES will be tuned us
ing one of the nature inspi
r
ed algorithm
s
,
namely the Fi
refly Algorithm (FA). To find the best
solution the objective
function utilised
within
the Firefly Algorithm
is th
e Compe
r
h
e
n
sive
Dam
p
i
ng Ind
e
x (CDI). It is
expe
cted that
by u
s
ing
this method t
he dynami
c
stability of
the
power sy
ste
m
can be e
n
h
anced.
2. Rese
arch
Metho
d
2.1. Po
w
e
r S
y
stem Model
Gene
rally the
powe
r
plant system can b
e
des
cri
bed a
s
sho
w
n in Fi
gure 1. The g
o
verno
r
is a
pa
rt of a
gene
rating
u
n
it that serve
s
to
reg
u
la
te
the provisio
n
of fuel (stea
m
or
wate
r) in t
he
gene
rating
sy
stem in
orde
r
to obtain
a
stable
rotor
spe
ed. If there
is
a chang
e at t
he ou
put of th
e
gene
rato
r owing to a load
chan
ge, the
n
there
will be a feedb
a
ck fun
c
tion t
hat is set by
the
govern
o
r to readju
s
t the ro
tation of the rotor.
Tu
r
b
i
n
e
Ge
n
e
r
a
t
o
r
E
x
c
i
ta
ti
o
n
S
y
s
t
e
m
G
o
ve
r
nor
F
r
om
t
h
e
bo
i
l
e
r
/
r
a
pi
d
pi
pe
T
o
r
que
-
+
Re
f
e
r
e
n
c
e
ω
Ref
e
r
e
n
c
e
V
+
-
P
o
w
e
r
,
C
u
rre
n
t
&
V
o
lta
g
e
Figure 1. Gen
e
ral Po
wer Pl
ant Model
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Power System
Design usi
ng Firefl
y
Algorithm
for Dynam
ic Stability ...
(He
r
lam
bang Setiadi
)
448
The ex
citatio
n
syste
m
is
a co
ntrol
system
for the
gene
rato
r ou
put, su
ch
as voltage,
curre
n
t or p
o
w
er fa
cto
r
. If there i
s
a
ch
ange in
th
e g
enerator
oup
ut, then the e
x
citation sy
stem
serve
s
to co
n
t
rol the gene
rator in ord
e
r t
o
adju
s
t and find a new b
a
l
ance point. Both parts of the
control
syste
m
, the g
o
vernor and
the
excitati
on sy
stem
s,
have
different re
sp
onse
time
s. The
govern
o
r h
a
s a slow
re
sp
onse, while t
he excitati
on
system ha
s a fast repo
n
s
e. In the work
pre
s
ente
d
he
re, the powe
r
system mod
e
l that will be
use
d
is one that is con
n
e
c
ted to the infi
nite
bus. The p
o
w
er
system
will be model
led in Lapla
c
e domain a
s
sho
w
n in Fig
u
re 2. This m
odel
first int
r
odu
ce
in [2
3] jou
r
n
a
l, and
n
o
w this
mod
e
l a
r
e commo
nly
use
d
to
solve
the
small
sig
nal
stability and
dynamic
stabi
lity becau
se i
n
this
model
power sy
ste
m
have been
modele
d
to linear
model. All of
the value in t
h
is mo
del i
s
asum i
n
pe
r
unit value. So this mo
del l
i
mited in pe
r
unit
value, there are no var, watt, ampere, or
voltage value.
ga
ga
sT
1
K
1
1
Tst
R
1
Y
m
T
+
-
E
E
sT
K
1
A
A
sT
1
K
F
F
sT
1
sK
A
V
2
U
F
V
D
Ms
1
s
f
2
+
-
+
-
K
4
fd
E
q
E
+
-
K
2
K
1
K
6
t
V
+
+
K
5
1
U
-
-
-
w
(
t
)
3
d0
3
K
T
s
1
K
Figure 2. Power Syste
m
Dynami
c
Mod
e
l (Ta
k
en fro
m
[24])
2.2. Po
w
e
r S
y
stem Stabilizer
The bl
ock
diagram
of the P
o
we
r System
Stabilizer (PS
S
) is
shown i
n
Figure
3. T
he PSS
block
diag
ra
m co
nsi
s
ts of
a g
a
in bl
ock,
wa
sh
out
blo
ck, t
w
o
blo
c
ks of l
ead
-lag
and
a limiter.
The
block g
a
in fu
nction
is to
a
d
just th
e am
o
unt of reinforceme
n
t that i
s
o
bataine
d i
n
a
c
cord
an
ce
with
the desi
r
ed torque. The washout bl
ock serves to provi
de a steady
state bias ouput PSS
that will
modify the terminal voltage
of the generator. T
he PS
S is expecte
d
to only resp
o
nd the tran
sie
n
t
variation of the rotor
spe
e
d
signal to the
sign
al gene
ra
tor and DC of
fset [25].
s
T
s
T
w
w
1
ma
x
S
V
mi
n
S
V
s
T
s
T
B
A
1
1
s
T
s
T
D
C
1
1
s
V
PSS
K
Figure 3. Bloc
k
Diagram of PSS (Tak
en from [25])
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 446 – 455
449
The lea
d
-la
g
functio
n
s as p
h
a
s
e-l
ead p
r
od
uci
ng characte
ristics a
p
p
r
o
p
riate t
o
comp
en
sate
for the
pha
se
lag b
e
twe
e
n
the inp
u
t excitation a
nd t
o
rqu
e
g
ene
ra
tor. The li
miter
function
is re
stri
cts th
e PS
S actio
n
o
n
t
he AVRS.
Fo
r exam
ple, in
the
event of
load
shed
di
ng,
the AVR acts to reduce the terminal vol
t
age of
the generator
when the PSS generates a control
sign
al to rai
s
e the voltage
(be
c
au
se th
e gen
erato
r
rotor spee
d i
n
crea
se in
si
ze d
u
rin
g
loa
d
shedding). In this condition it i
s
necessary to
disable the PSS. It should be not
ed that, the high
negative limit values
can di
sru
p
t the
stab
ility of
the first swin
g [25].
2.3. Superco
nducting M
a
gnetic Ener
g
y
Storage
A Supe
rcond
ucting
Ma
gn
etic En
ery St
orag
e
(SME
S) sy
stem
wi
thin an
ele
c
tric p
o
wer
system is u
s
ed to control the balan
ce o
f
power
in th
e synchro
nou
s gene
rato
r d
u
ring p
e
rio
d
s of
dynamic
cha
nge. The SM
ES is installe
d in terminal bus ge
ne
rato
r on the mod
e
l of the power
system,
as shown in
the
b
l
ock di
agr
am
of Figu
re
4. A
ll of the
calcu
l
ation in
he
re
asum
a
nd lim
it
and La
pla
c
e
domain.
0
K
DC
sT
1
1
I
d
K
SM
E
S
P
1
s
L
d
I
0
d
I
0
dd
I
I
d
E
d
E
1
Figure 4. Block
Diag
ram o
f
SMES (Taken from [8] and [13])
The
state v
a
riabl
e e
quat
ion of th
e S
M
ES
unit m
a
y be
expre
s
sed
as foll
ows by
insp
ectio
n
of the block dia
g
ram
sho
w
n i
n
Figure 4 [12]:
∆
1
1
∆
∆
(1)
∆
1
∆
(2)
Whe
r
e
∆
is DC voltage ap
plied to the indu
ctor,
∆
is the rotor devi
a
tion,
∆
is the
curre
n
t
flowing throu
gh the indu
ct
or,
is the gai
n for feedba
ck
∆
,
is the con
v
erter time d
e
lay,
is a gain
co
n
s
tant and
L is the indu
ctan
ce of the co
il. The deviatio
n
in the indu
ctor real
powe
r
of
the SMES unit is expre
sse
d in the time domain a
s
fol
l
ows [12]:
∆
∆
∆
∆
∆
(3)
∆
is the
real
po
wer that will
be given
to the
g
r
id. Th
e
energy sto
r
e
d
in the SME
S
at any
instant in time,
,
is given b
y
[12]:
2
(4)
2.4. Firefly
Algorithm
Firefly Algorit
hm is
a natu
r
e inspire
d
al
gorithm
th
at
derive
s
fro
m
the ch
aracte
ristics of
fireflies. T
h
is
algorith
m
wa
s d
e
fined
by
Dr Xin-
Sh
e Y
ang from th
e
Univeri
s
ty of
Camb
rid
ge,
UK
in 20
08. In
th
is al
go
rithm t
here
a
r
e
thre
e ba
seli
ne
co
nsid
eratio
ns.
All of the
firef
lies
are u
n
isex
so that a firefly will be attracted to any other
firefly regadl
ess of gende
r [26]. The attractiven
e
ss
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Power System
Design usi
ng Firefl
y
Algorithm
for Dynam
ic Stability ...
(He
r
lam
bang Setiadi
)
450
is propo
rtion
a
l to the brig
htness of a firefly’s f
lickeri
ng light. The
r
efore, for every two firefl
y’s
blinki
ng, one
of them woul
d move to an
other li
ght. T
heir b
r
ightn
e
ss d
e
cre
a
ses as the di
sta
n
ce
betwe
en th
e
fireflies in
creases. If the
r
e i
s
n
o
b
r
ig
htest firefly
within th
e po
pulation, th
e
n
all
fireflies will move
randomly
[27].
The brigness
of a firefly is
inf
l
uenced
or determined by
the
obje
c
tive function [27].
The
objec
tive func
tion us
ed to tes
t
the
s
y
tem is
the Comprehen
s
i
ve D
a
mping
Index
(CDI) [28, 29]. Futhermo
re the dynami
c
model of power sy
stem
will be conv
erted into
state
matrix in equ
ation (5
) and
(6). By usin
g the st
ate matrix is easi
e
r to find the eigenvalue
s of
a
highe
r o
r
d
e
r syste
m
, the
eige
nvalue
itself c
an
be
prese
n
ted
a
s
e
quatio
n (7). To
find t
he
dampin
g
valu
e, then
equ
ation
(8) is u
s
e
d
. Finally th
e
CDI
obje
c
tive
functio
n
can
be d
e
termi
n
e
d
by equation
(9). The variable t
hat
will be optimiz
ed by the FA
is the PSS parameter
and SMES
para
m
eter, th
ere will be ra
nge of value of each pa
ra
meter. The reason why th
is paramete
r
is
chosen is because it repr
esents the PSS and SMES value.
∆
∆
∆
(5)
∆
∆
∆
(6)
Whe
r
e
∆
is st
ate matrix,
∆
is the ou
put variable m
a
trix,
is
the input matrix, A i
s
the
system
matri
x
, B is th
e in
p
u
t matrix,
C i
s
the
me
asurement m
a
trix, and
D i
s
the
input to
the
output matrix.
(7)
(8)
1
(9)
Whe
r
e
is th
e eigenval
ue,
is the re
al
comp
one
nt o
f
the eigenva
l
ue,
is the i
m
agina
r
y
comp
one
nt of the eigenval
ue, and
is the dampin
g
rat
i
o of the system.
3. Results a
nd Analy
s
is
Simulation
was p
e
rfom
ed
usin
g MAT
L
AB 2010,
whe
r
e th
e p
o
we
r
system
dynami
c
model, Power System Stabilizer (PSS)
and Super
conducting M
a
gnetic En
ergy Storage (SM
ES)
were create
d
within Simul
i
nk. The Fi
re
fly Algorithm (FA) wa
s i
m
pleme
n
ted
usin
g an M-f
ile
.
Re
sults and analysi
s
a
r
e pre
s
ente
d
in two
sectio
ns
,
first sectio
n
sho
w
s the ei
genvalu
e
an
a
l
ysis
and re
sult
s, while the second section
provide
s
information abo
ut the dynamic perform
an
ce
of
the system
by sho
w
ing
the frequ
en
cy and ro
to
r angle resp
onse. Table
1 provide
s
the
parameters of the PSS and SMES that
were optimi
z
ed usi
ng the
FA, while Figure 5
shows the
conve
r
ge
nce
grap
h for th
e Firefly Alg
o
rithm (FA)
as it find
s th
e optimal val
ue, or m
eets the
crite
r
iation of
the obje
c
tive function.
Table 1. Opti
mized Pa
ram
e
ters
Variable Value
Kpss 9.9803
T
w
10.5
T1
0.465
T2
0.2965
T3
0.3845
T4
0.2257
Ko 60
Kid 5.9143
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 446 – 455
451
Figure 5. Con
v
ergen
ce of the Firefly Algortihm
3.1. Small S
i
gnal Stability
Anal
y
s
is
Small signal
stability usual
l
y ca
lled
steady state stabi
lity [1],
relates to the stability of th
e
s
y
s
t
em
where there are
no exter
nal
dis
t
urbanc
e
,
or the s
t
ability of the s
ystem its
e
lft. This
stability is very important
is because
by finding thi
s
st
ability, we
can know whet
er thi
s
system is
stable o
r
not, if we kno
w
th
e system i
s
st
able than
we
can
start to ru
n the system.
The
stability
of the
system
ca
n b
e
dete
r
mined
by
find
ing the
eige
n
v
alue of the
system. If
the real pa
rts
of the eigenv
alue are ne
g
a
t
ive then the system i
s
sta
b
le.
Figure 6. Eigenvalue Plot
for Un
co
ntroll
ed System
Figure 7. Eigenvalue Plot for System wit
h
PSS and SMES (Conventional)
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Power System
Design usi
ng Firefl
y
Algorithm
for Dynam
ic Stability ...
(He
r
lam
bang Setiadi
)
452
Figure 8. Eigenvalue Plot for System wit
h
FA Tuned
PSS and SMES
Figures 6 to 8 sho
w
the ei
genvalu
e
plot for
the syste
m
without a c
ontrolle
r, the system
with
conventional PSS and SMES, and the
sy
stem
with PSS
and SMES tuned
by the
FA
respe
c
tively. Based
on
[1]
the sy
stem
can b
e
calle
d
stable
if the
real p
a
rt
s of th
e eig
envalu
e
are
negative, wh
ere the l
a
rg
er the m
agnitu
de then the f
a
ster th
e sta
b
ilit
y. From F
i
gure
6, it ca
n be
see
n
that the system
wi
thout
a cont
rolle
r is in i
t
self stable.
Ho
wever, for an improve
d
perfo
rman
ce,
the system
need
s to hav
e more
negat
ive magnitud
e
. Figure
s
7
and 8
sho
w
that
the stability
of the system increases when
the
system has the PSS and SMES installed.
Futhermore, enhanced results
are obt
ained when t
he PSS and SMES have their param
e
ters
tuned by the FA.
Another im
portant paramet
er fo
r
stability is the dam
pi
ng ratio. If the dampi
ng ratio has a
negative valu
e then the
r
e
will be a
po
sitive eigenv
alue that will
make th
e system un
sta
b
le.
Acco
rdi
ng to
Rog
e
rs [30],
a po
wer sy
stem with
dam
ping ratio g
r
e
a
ter tha
n
0.0
5
is
satisfa
c
t
o
ry.
For this work all of the damping ratio val
ues a
r
e
great
er than 0.05, as sh
ow
n in figure 6 to 8. By
installin
g the PID and SMES in the powe
r
system, t
he dampin
g
ratio
of the
system is increa
se
d:
the highe
r the
damping
rati
o then the oscillation level
decays.
3.2. D
y
namic
Stabilit
y
Anal
y
s
is
Dynami
c
stability is the
abi
lity of
the sy
stem to mai
n
tain a
synchronised co
ndition after
the first swin
g happ
en
s u
n
til the syste
m
rea
c
he
s
a
new e
quilib
ri
um stea
dy
state con
d
ition
[31].
This
stability is wh
ere the
r
e
is an externa
l
disturb
a
n
c
e
su
ch a
s
load
cha
nge.
To analy
s
e t
he dynami
c
behavio
ur of
the
power
system, the
system n
eed
s to be
modelle
d into
a li
nea
r m
o
d
e
l (th
a
t is through
La
pla
c
e
tran
sformati
on) an
d give
n a
load
cha
nge
0.01 pu: the rea
s
on
why 0
.
01 is ch
oo
se
n becau
se
0.
01-0.0
5
pu is the normal
way to simul
a
te
the load ch
an
ging in po
wer system. In this simu
l
a
tion
the load ch
an
ge is re
pre
s
e
n
ted as a ste
p
sign
al in Sim
u
link,
whe
r
e
the final valu
e of the
step
sign
al i
s
0.0
1
pu. An
alysi
s
i
s
cond
uct
e
d
throug
h
com
parin
g the
freque
ncy
re
spon
se
and
rotor a
ngle
re
spo
n
se of th
e po
we
r
syst
em
without controller, with the inclusion of
a PSS and
SMES, and also
with the PSS and SMES
tuned by th
e
FA. Figures
9 and
10
sh
o
w
the f
r
eq
u
e
n
cy respon
se
and
roto
r an
gle of the
po
wer
system for al
l three config
uration
s
, whil
e T
able
s
5 and 6 illustrat
e
the oversh
oot and settli
n
g
time for both frequ
en
cy and
rotor an
gle for the thre
e system arrang
ements.
Table 5. Co
m
pari
s
ion
Re
sp
onse Fre
que
ncy Value
Parameter
Uncontrolled
PSS SMES
PSS SMES
(FA tuned
)
Overshoot (pu)
-8.622e5
-6.46e5
-5.743e5
Settling Time (Se
c
)
11.57
7.53
5.33
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 446 – 455
453
Figure 9. Fre
quen
cy Re
sp
onse Com
p
a
r
ision
Table 6. Roto
r Angle Value
Compa
r
i
s
on
Re
spo
n
se
Parameter
Uncontrolled
PSS SMES
PSS SMES
(FA tuned
)
Overshoot (pu)
-0.0055
-0.0036
-0.0036
Settling Time (Se
c
)
12.21
14.33
13.53
Figure 10. Ro
tor Angle Respon
se Comp
ariso
Tables 5 and 6
show that
PSS and SM
ES can
decrease the
overs
hoot and
reduce the
settling time of the freque
ncy and the
rotor a
ngl
e in
a powe
r
system. The best result for th
e
freque
ncy
re
spo
n
se is tha
t
when the P
SS and SM
ES is tuned
by the FA. Whe
n
the sy
stem
is
installed with
the PSS and SMES then the system
overshoot i
s
reduced
to 0.00002979 pu and
the settling time is 6.24
seco
nd
s, which is fast
e
r
th
en the
sy
ste
m
witho
u
t a
controlle
r. Fo
r the
rotor a
ngle, the syste
m
in
cludi
ng the P
SS
and SMES can dam
pe
n the oversh
oot, howeve
r
the
time settling is
larger than
that for the
sys
tem
wi
thou
t a controller.
As
sh
own in
the g
r
a
ph, t
he
system
without a controll
er has quite an oscillato
ry response which is not
appropri
a
te because i
f
this oscillatio
n
rise
s ag
ain
and again, then the sy
ste
m
will be out
of synchroni
zation a
nd it will
c
o
llapse. That is
why us
ing an ad
ditional c
o
ntroller is
required. The PSS is
used to c
o
ntrol
the
excitation of the power sy
stem and the SMES wo
rks as energy storag
e to help
the governo
r.
Owin
g to th
at re
sp
on
se fre
quen
cy a
nd
rotor
angle
i
s
better th
an
system
withou
t cont
rolle
r. T
he
FA itself helps an engi
nee
r to find the best para
m
eter
of the system
: that is
why the system
with
a PSS and SMES tuned by the FA have the best parameters.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Power System
Design usi
ng Firefl
y
Algorithm
for Dynam
ic Stability ...
(He
r
lam
bang Setiadi
)
454
4. Conclusio
n
Includi
ng a
Power System Stabilizer
(PSS)
and Superconducti
ng Magnetic Energy
Storage
(SM
ES) can i
m
prove the eige
n
v
alue of a
po
wer
syste
m
, thus m
a
ki
ng t
he overall sy
stem
more
stable.
As well a
s
th
e eigenval
ue,
the dy
namic stability of the po
we
r sy
stem imp
r
ov
es
when the
system has the
PSS and SMES installed,
as
shown i
n
the respon
se
of the frequency
and
rotor angle. The
Firefly Algorithm
(FA)
can be empl
oyed
to tune the
PSS and S
M
ES
para
m
eters
so that it is e
a
s
ier to find th
e more
a
ppro
p
riate val
ues,
althougth
th
e FA itself mi
ght
never fin
d
th
e be
st pa
ram
e
ter a
s
th
e F
A
is a
sto
c
h
a
s
tic m
e
thod.
The
simulatio
n
re
sult
s
sho
w
that the syst
em witho
u
t a
cont
rolle
r ha
ve high o
s
ci
a
llation levels
that contin
ue
for quite a
long
time. This is not good for
the rotor of the
gene
rato
r and also not
advantageo
us for the pri
m
e
mover: the g
enerator
coul
d go out
of
synchro
n
ization, while t
h
e
prime move
r (turbine
) co
uld
brea
k. That is why the inclu
s
ion of an
additional
co
ntrolle
r is ne
eded. In this work that fact is
clea
rly
sho
w
n
.
Ackn
o
w
l
e
d
m
ent
The first a
u
th
or i
s
ve
ry gra
t
eful to the In
done
sia
n
Go
verment, e
s
p
e
cially the
Di
rectorate
Gene
ral of Hi
gher Edu
c
ati
on for the Co
mpetitiv
e Scholarship a
w
a
r
ded to him during hi
s stud
ies
for an MSc i
n
Electri
c
al P
o
we
r an
d Co
ntrol Engi
n
e
e
r
ing, at Liverpool John M
oore
s
University,
UK. Spe
c
ial t
han
ks to the
se
con
d
a
u
tho
r
, Dr Ka
rl O
Jone
s, who
was th
e Sup
e
rvisor th
at hel
ped
the first autho
r treme
ndou
sl
y.
Referen
ces
[1]
Kund
ur P. Po
w
e
r S
y
stem Sta
b
ilit
y a
nd Co
ntrol. Ne
w
York:
McGra
w
-
Hil
l. 1994.
[2]
Goshal SP,
C
hatterje A
an
d
Mukher
jee
V. Bio-i
n
spir
ed f
u
zz
y
lo
gic b
a
s
ed tun
i
n
g
of p
o
w
e
r s
y
stem
stabiliz
er.
Expe
rt System w
i
th
Appl
icatio
ns
. 2
009; 36; 9
281-
929
2.
[3]
Soni PB, Sa
xe
na A and Gupt
a V. A Minimax Pol
y
n
o
m
i
al A
ppro
x
imatio
n Objective F
unc
tion Appr
oac
h
for Optimal
D
e
sig
n
of
Po
w
e
r S
y
stem St
abil
i
ze
r by
Emb
e
d
d
i
ng
Pa
rti
c
l
e
S
w
arm
Op
ti
mi
za
ti
on
.
T
E
LKOMNIKA Indon
esi
an Jou
r
nal of Electric
al Eng
i
ne
eri
ng.
2015; 1
4
(2); 1
91-1
98.
[4]
Lu F
C
, Li
u C
C
an
d W
u
JC. Effect of Batter
y
E
ner
g
y
Storag
e S
y
stem
on L
oad F
r
e
q
uenc
y
Contro
l
Consi
der
ing G
o
vern
or De
ad
ban
d an
d Ge
nerati
on R
a
te
Constrai
nt.
IEEE Transaction on Energy
Conv
ersio
n
. 19
95; 10(3); 5
55-
561
[5]
Kal
y
a
n
i S, Na
gal
akshmi S a
nd Maris
ha R.
Loa
d F
r
equ
e
n
cy Contro
l us
ing Battery E
n
ergy Stora
g
e
System in Inter
c
onn
ected Po
w
e
r System
. Procee
din
g
s of ICCCNT
’12.
Co
imbator
e India.
2012.
[6]
Mahto T
and
Mukherj
ee V.
Energ
y
Stor
ag
e S
y
st
ems
for Mitigati
ng the
Varia
b
il
it
y
of Isolate
d
H
y
br
i
d
Po
w
e
r S
y
stem.
Elsevier R
ene
w
able an
d Sus
t
aina
ble E
nerg
y
Review
s
. 201
5; 51; 156
4-15
77.
[7]
Sasaki, T
., Kado
ya, T
.
and
Enomoto, K.,
Stud
y o
n
Loa
d F
r
eq
uenc
y
Contro
l Us
ing
Re
do
x F
l
o
w
Batteries.
IEEE Transactions
on Power System
. 2
004; 1
9
(1
); 660-66
7.
[8]
Chid
amb
a
ram
AI and Par
a
ma
sivam B. Optimized L
o
a
d
-F
requ
enc
y Sim
u
l
a
tion i
n
Restr
u
ctured Po
w
e
r
S
y
stem
w
i
th R
edo
x F
l
o
w
B
a
tteries
an
d Interl
ine P
o
w
e
r
F
l
o
w
C
o
ntroll
er.
E
l
ectrical
Pow
e
r
and
Ener
gy
System
s
. 20
13
; 50; 9-24.
[9]
Chatterj
ee A,
Goshal PS
an
d Mukh
erje
e V
.
T
r
ansient Per
f
ormance Impr
oveme
n
t of Grid Co
nnecte
d
H
y
dro S
y
stem
usin
g Distir
but
ed Gen
e
ratio
n
and C
a
p
a
citiv
e
Ener
g
y
Stor
age
Unit.
Else
vier Electric
al
Pow
e
r and En
ergy Syste
m
s
. 201
2; 43; 210-
221.
[10]
Ali HM, W
u
B
and
Do
uga
l A
R
.
An Overvi
e
w
of SMES Applic
atio
ns i
n
Pow
e
r an
d En
ergy Syste
m
s
.
IEEE Transactions on Sust
ainaible Ener
gy
. 2
010; 1(1); 3
8
-4
7.
[11]
Arso
y
BA
an
d
W
ang Z
.
T
r
ansient Mo
del
lin
g
and S
i
mul
a
tio
n
of a SMES Co
il an
d the P
o
w
e
r Electro
n
ics
Interface.
IEEE Transactions
on Applied Superconductivity
. 199
9; 9(4); 471
5-47
24.
[12]
Demiror
en
A a
nd Y
e
sil
E. Aut
o
matic Ge
nera
t
ion C
ontro
l
w
i
t
h
F
u
zz
y L
o
g
i
c
Contro
ller
in
P
o
w
e
r S
y
st
em
Includ
ing SME
S
Unit Contro
ll
ers.
Elsevier El
ectrical Pow
e
r
and En
ergy Sy
stems
. 20
04; 2
6
; 291-3
05.
[13]
Ansari
T
Md M
an
d V
e
lus
a
mi
S. Du
al
Mod
e
Li
ngu
istic
Hed
ge F
u
zz
y
L
ogi
c Co
ntroll
er for
an
Isol
ate
d
W
i
nd-Di
esel
H
y
br
id Po
w
e
r S
y
stem
w
i
t
h
Su
perco
nducti
ng
Magn
etic Ener
g
y
Stor
ag
e U
n
it.
Elsevier
Energy C
onver
sion a
nd Ma
na
ge
me
nt
. 2010;
51; 169-
18
1.
[14]
Puji
antara M
a
nd Ab
dil
l
a
h
M.
(In Engl
ish) A
daptiv
e PI Co
ntroll
er Base
d
on Bat Al
gorit
hm (BA) an
d
Extreme
Le
arn
i
ng M
a
chi
ne (
E
LM) for C
ont
rolli
ng S
p
e
ed
i
n
Perma
ne
nt
Magn
et S
y
nch
r
ono
us Motor.
Java Jour
nal E
l
ectrical
and El
ectronic En
gin
eeri
n
g
. 20
11; 9
(
2); 113-1
22.
[15]
Niknam
T
,
Abargho
oe A
R
a
n
d
Ro
osta A. R
e
serve
Co
nstrain
ed D
y
n
a
m
i
c
Econom
ic Dis
pacth; A N
e
w
F
a
st Self-Adap
tive Modifi
ed F
i
refl
y
Al
gor
ithm.
IEEE System
Journal
. 20
12; 6(4); 635-
64
6.
[16]
Ali ES. F
i
refl
y
Algorit
hm Bas
ed Spe
ed C
ont
rol of DC Seri
e
s
Motor.
W
S
EAS T
r
ansactio
n
s on Syste
m
and C
ontrol
. 2
015; 10.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 446 – 455
455
[17]
Apostol
o
p
oul
o
s
T
and Vlac
hos
A. Appl
ic
ation of the F
i
refl
y
Alg
o
rith
m for solvin
g
the Econom
i
c
Emissions
Loa
d Disp
a
tch Pro
b
lem.
Internati
ona
l Journ
a
l of
Combi
natorics
. 2011.
[18]
Roev
a O
and
Slavov
T
.
Firefly
Algorithm
of PID
Controller for
Glucos
e Co
ncentrati
on
Co
ntrol duri
n
g
E.
coli F
ed-
batch
Cultivati
on Pr
ocess
. Proce
e
d
in
gs of the F
eder
ated
C
onf
erenc
e on
Co
mputer Sci
e
ce
and Informati
o
n
S
y
stem. 20
1
2
; 455-4
62.
[19]
Parimi MA, E
l
amvazuthi I, Saad N.
D
a
mpi
ng
of Inter Ar
ea Osci
llati
ons
Usin
g Inter
lin
e Pow
e
r F
l
ow
Contro
ller
Bas
ed
Da
mp
in
g C
ontrol
l
ers
. Proc
eedings
2nd IEEE International
Confer
enc
e
on P
o
w
e
r
a
n
d
Energ
y
(PEC
o
n
08). Johor B
ahar
u, Mala
ysi
a
. Decemb
er 1
-
3, 2008.
[20]
Kashki M, Ab
id
o AM, Magi
d-A
bde
l LY. Po
w
e
r
S
y
stem D
y
n
a
m
ic Stabil
i
t
y
E
nha
nceme
n
t U
s
ing Optim
u
m
Desig
n
of PS
S and Static
Phase Sh
ifter Based Stab
ilizer.
Ara
b
ia
n
Journa
l for Scienc
e an
d
Engi
neer
in
g
. 2010; 38; 6
37-6
50.
[21]
Usman J, M
u
stafa W
M
, Ali
y
u
G. Desi
gn
of
AVR
and
P
SS for Po
w
e
r
S
y
stem
Stab
i
lit
y bas
ed
on
Iteration Particle s
w
arm Optimization.
Interna
t
iona
l Jour
nal
of Engi
neer
ing
and In
novativ
e
T
e
chnol
ogy
(IJEIT)
. 2012; 2; 307-3
14;
[22]
Hakim AE, Soeprij
anto
A, Purnom
o HM. Fuzz
y
PID bas
ed
PSS
Design Using Genet
ic Algorithm.
Internatio
na
l Journ
a
l of Electr
ical Pow
e
r an
d
Energy Syste
m
En
gin
eeri
n
g
.
2010; 3; 37-
40
.
[23]
De Me
ll
o F
P
a
nd
Corco
d
ia
C
.
Conc
epts of
s
y
nc
hro
nous
machi
ne sta
b
il
i
t
y
as
affected
b
y
e
x
cit
a
tio
n
control.
IEEE Transacti
on o
n
Pow
e
r Apparat
us and Syste
m
s
. 1959; 8; 316
-328.
[24]
Soepr
ija
nto A.
(In Eng
lish)
D
e
sign
Co
ntroll
er
for Dy
na
mic
S
t
ability
Electric
Pow
e
r Syste
m
. Surabay
a
:
Itsprees. 201
2.
[25]
Rob
and
i I. (In
Englis
h)
Desi
gn Mod
e
rn P
o
w
e
r System: Optimi
z
a
ti
on,
F
u
zz
y
Log
ic
and Geneti
c
Algorit
h
m
. Yogy
akarta: ANDI. 2005.
[26]
Yang
XS.
En
gin
eeri
ng Opti
mi
z
a
t
i
o
n
: An Introducti
on w
i
th Metahe
uristi
c Applic
ation
. Ne
w
Jersey
:
W
I
LEY. 2010.
[27]
Yang
XS.
Natu
re-Inspir
ed Met
ahe
uristic Alg
o
r
ithms. Sec
ond
Edition
. Un
ited
Kingd
om: Lun
i
v
er. 2010
[28]
Had
d
in
M, So
eba
gio,
Soe
p
ri
janto
A
an
d
Pur
nom
o HM.
Mod
e
li
ng
a
n
d
C
ontrol
of
Excitati
on
a
n
d
Governor B
a
s
ed o
n
Particl
e
S
w
arm Opti
mization for Micro Hy
dr
o Po
w
e
r Plant.
TEL
K
OMNIKA
Indon
esi
an Jou
r
nal of Electric
al Eng
i
ne
eri
n
g
.
2013; 1
1
(2); 2
97-3
04.
[29]
El-Z
onko
l
y MA
, Khali
l
AA a
n
d
Ahme
d MN.
Optimal T
uni
ng of L
e
a
d
-L
a
g
an
d F
u
zz
y
Log
ic Po
w
e
r
S
y
stem Sta
b
il
i
z
er usi
ng P
a
rticle S
w
a
rm Opti
mizatio
n
.
Expe
rt System w
i
th
Appl
icatio
n
. 20
09; 36;
209
7-
2
1
06
[30]
Rogers G.
Power System
Oscillation
. Ne
w
Yo
rk: Springer. 2
000.
[31]
Milles R
H
an
d Malin
o
w
ski JH.
Pow
e
r System Operation
. New
York: McGraw
-Hill. 1994.
Evaluation Warning : The document was created with Spire.PDF for Python.