1

Optimal Location of FACTS Devices for

Improving Performance of the Power
Systems
Rajive Tiwari, Member, IEEE, K.R.Niazi, Senior Member, IEEE, Vikas Gupta

controlled series capacitors (TCSCs) in order to eliminate line

Abstract- This paper presents a Genetic Algorithms (GA) based overloads under a single contingency. They introduced an
approach to seek the optimal location of FACTS devices to improve index called the single contingency sensitivity (SCS) index
voltage stability margin and minimize reactive power loss of the which is used to rank the system branches according to their
power systems. The location of FACTS devices, type, cost, and suitability for installing a TCSC. In [16], a methodology is
parameter values are optimized simultaneously. Static models of
carried out using a genetic algorithm to identify the optimal
three FACTS devices consisting of Unified Power Flow Controller
(UPFC), Thyristor Controlled Series Capacitor (TCSC) and Thyristor number and location of thyristor-controlled phase shifters for
Controlled Phase Shifting Transformer (TCPST) have been selected maximizing system capabilities. Rahimzadeh et al. [17]
for the steady-state analysis. Case studies with, IEEE 30-bus and 118- proposed a novel method based on genetic algorithm to find
bus systems are presented to illustrate the applicability of the the location, the operating point as well as the number of
algorithm. multi-type FACTS devices in the restructured environments
Index Terms-- Genetic algorithm, FACTS devices, voltage simultaneously and optimally. Ref. [18] described a
stability margin, optimal location. methodology for optimal placement of static var compensators
(SVCs) in a transmission network in such a manner that its
I. INTRODUCTION loading margin is maximized. Differential evolution (DE)

N OW a days, the electrical power system is experiencing a

new challenge of deregulation and liberalization. It has
created additional requirements for the operation of power
technique is applied to the optimal placement and parameter
setting of UPFC for enhancing power system security under
single line contingencies in [19]. Application of PSO to
systems. However, Issues such as limited investment in determine the optimal location of the shunt FACTS devices
generation, economic and environmental limitation, for transient stability improvement has been reported in the
competitive business environment etc. have motivated the literature [20]. Ref. [21] proposes a solution algorithm for the
modern electrical utilities to install flexible AC transmission preventive control problem through finding the optimum
systems (FACTS) devices to optimize the existing location, type and size of the static shunt compensation
transmission facilities while, at the same time, maintaining devices by using PSO. Wibowo et al. [22] presented an
acceptable levels of network reliability and stability. FACTS optimal allocation method for flexible ac transmission system
devices can improve the performance of the power systems by (FACTS) devices for market-based power systems considering
controlling the power flow in the network [1]. congestion relief and voltage stability. Techno-economic
Various FACTS controllers, their modeling and their contribution of FACTS devices to system operation are
impact on power systems have been reported in the literature assessed in [23]-[24].
[2]-[8]. Most effective use of the FACTS devices depend The review of the past research identified that most of the
largely on the fact, how these devices are placed in the power research is towards the enhancement of available transmission
systems, i.e. on type, location and size [9]. An optimal capacity while, the contributions of FACTS devices to
location of FACTS devices allows controlling its power flows minimize reactive power loss along with improvement in
and thus enhances the reliability of the power systems [10]. voltage stability margin are not comprehensively examined.
Optimal location can be determined using some well Moreover, these studies have not explored effectiveness of
established search algorithms such as; genetic algorithm [11], TCPST to improve voltage stability along with, UPFC, and
particle swarm optimization [12], tabu search [13], etc. Gerbex TCSC during contingency. This article, therefore, proposes a
et al. [14] provided an idea regarding the optimal locations of genetic algorithm (GA) search technique to determine the
FACTS devices using genetic algorithm. In their study, the optimal location of FACTS devices in transmission network
system loadability is applied as measure of power system to enhance the voltage stability margin and minimize the
performance. Lu et al. [15] described an approach for reactive power loss of the power systems. The proposed
determining the most suitable location for installing thyristor- method has been demonstrated on IEEE 30-bus and 118-bus
systems [25] for base load and contingency conditions.
Rajive Tiwari (e-mail: rajive.tiwari@yahoo.co.in), K. R. Niazi (e-mail:
krn152001@yahoo.co.in) and Vikas Gupta (e-mail:vikas_artist@yahoo.co.in)
are with the Department of Electrical Engineering, Malaviya National Institute
of Technology, Jaipur, Rajasthan, India.

978-1-4673-2729-9/12/$31.00 ©2012 IEEE

 2

II. MATHEMATICA
AL MODEL OF FACTS DEVICES
S
In a powerr system, FA ACTS technolo ogy has a widew
spectrum of im mpacts on the way the tran nsmission systtem
opperates. It maakes the transmission of power flow more m
relliable, more controllable, an nd more efficiient. The FAC CTS
deevices are ablee to manage the t interrelated d parameters thatt
coonstrain today’’s systems, in ncluding voltaage, phase ang gle,
serries impedance, shunt impedance, and reaactive power. By
addding to flexibility, the FACT TS devices enab ble a transmisssion
linne to operate close
c to its thhermal rating. Their high-speed Fig.1 Schematic diagram
m of TCSC
coontroller actionn help to maaintain stability y, avoid cascade
trippping, control power flow, support
s dynammic voltage, daamp B. Thyrristor-Controllled Phase Shiftting Transform
mer (TCPST)
oscillations and increase
i synchronizing torque.
On the basiss of type of compensation, FA ACTS controlllers TCP PST consists oof series and shhunt converterrs. The shunt
cann be categorizzed into three types;
t series controllers (SSS SC, converteer is coupled tto the transmisssion line throough a shunt
TC CSC, IPFC), shunt controlllers (SVC, STATCOM), S and
a transform
rmer and the series converrter is coupleed in series
coombined series--shunt controlllers (TCPST, UPFC).
U The serries throughh series transfoormer as shownn in Fig.2. Thhe Thyristor-
coontrollers can be
b used to allev viate line overlloads, to decreease Controllled Phase Shhifting Transfformer (TCPS ST) adds a
linne loss, and ennhance power transfer capab bility. The Sh hunt quadratuure componentt to the prevailling bus voltagge in order to
coontrollers can improve
i voltagge profile by injecting reacttive regulatee its angle so as to allow thhe control of power flow
poower at the lo ow voltage bu us. The comb bined series-sh hunt betweenn the transmisssion lines. TCPPST may be reppresented by
coontrollers can be
b employed in n the system too improve volttage power innjection model as shown in F Fig.3 where, Pi, Qi and Pj ,
staability margin,, to avoid the power flow co ongestion, and d to Qj are thhe injection poowers at the bbuses i and j rrespectively.
deecrease line losss, since it un nites the featurres of both serries The actiive and reactivve power invollved in the serries injection
coontrollers and shhunt controllerrs. is takenn from the linne and is circculated througgh the shunt
Three differrent types of controllers
c (TCCSC, TCPST, and a transform
rmer. Injected series voltagge is represennted by Vs,
UP PFC) have beeen considered to t be optimally y located in orrder whose magnitude annd phase anglle (θs) are coontrolled by
to enhance voltaage profile and d to reduce thhe reactive pow wer TCPST..
losss of the poweer systems. The effectivenesss of these deviices
is tested by developing
d theeir mathematiical models and a
inccorporating theem in load flow w program. TC CSC is modeled d as
vaariable reactan nce which caan modify th he reactance of
traansmission linee while TCPST T and UPFC are a represented d by
poower injection models
m [2]-[3]. The detailed modeling
m of th
hese
FA ACTS devices is described in n following secttion.
A. Thyristor-Con ntrolled Seriess Capacitor (TCCSC)
Thyristor- Controlled
C Series Capacitor (TTCSC) is inserrted
in series with the line. TCSC controllerrs use Thyristtor-
Coontrolled Reactor (TCR) in parallel
p with segments of serries
T combination of TCR and capacitor offfers
cappacitor bank. The
sm
mooth and flex xible control of the line im mpedance with h a
ressponse better than the conv ventional conttrol devices. The
T Fig.2 S
Schematic diagram
m of TCPST
TC CSC may havee only two possible characteeristics; inducttive
or capacitive. TCSC
T can be modeled as a variable serries
reaactance as shoown in Fig.1. TheT reactance XTCSC offered by
TC CSC is a function of the reacctance X12of thhe uncompensaated
linne in which it is located. Thee reactance of transmission line
l
is modified by TCSC in the folllowing way:

XLINE = X12 + XTCSC (1)

XTCSC = k. X12 (2) Fig.3 Power innjection model forr TCPST and UPF
FC

Where, XLINE
L is the reactance of o compensaated The expressions fo for injected poowers as suggeested by [3]
traansmission linee and k is the compensationn factor of TCS
SC. are giveen by
Thhe working ran nge of TCSC iss selected betw
ween −0.8 X12 and
a Pi = mbbseViVj sin (δi - δi + θs) (3)
0.22 X12 [24]. Qi= m2 bse Vi2 – mbseViVj cos(δi - δi + θs) (4)
Pj = – m
mbseViVj sin (δi - δi + θs) (5)
Q j= – m
mbseViVj cos(δi - δi + θs) (6)
 3

Where, bse iss susceptance of o series transfformer and Vi , Vj Qj = Vj2 bse+ VjVsebse ccos(δj - θse) – ViVjbse cos(δj - δi) (10)
annd δi, δj are magnitudes
m annd phase anglees of voltagess of
buuses i and j resspectively . Th he phase shift angle
a of TCPST T is Thee shunt voltagee source injectss active and reaactive
reppresented as θs and it can be controlled to /2 .We defi fine, powers at bus i which are given by (11)-(12).
m= = |Vs|/|Vi|, and it lies in the raange 0 to mmax.
Psh = V iVsh bsh sin(δi - θsh) (11)
C. Unified Poweer Flow Contro oller (UPFC) Qsh = V i2 bsh – ViVshbsh cos(δi - θsh)
UPFC con nsists of two o voltage so ource convertters + ViVj[ggse cos(δj – δi) + bse sin(δj–δi)]] (12)
emmploying Gatee Turn-off Th hyristors (GTOOs) as shown n in
Figg.4. The shu unt converter is coupled in i shunt to the Whhere, bse and bsh are suscepttances of seriees and shunt
traansmission linee through a sh hunt transform
mer and the serries rmers respectivvely and Vse∠ θse is the compplex value of
transform
coonverter is cou upled in seriees through seeries transformmer. series innjected voltagee while compllex value of shhunt voltage
Seeries converter controls poweer flows in thee line by injectting source iis represented bby Vsh∠θsh .
ann ac voltage in series witth transmission line which h is Moodels of aforrementioned F FACTS devicces can be
coontrollable in magnitude
m andd phase angle.. Shunt converter incorporrated in Newtton-Raphson looad flow (NR RLF) method
alsso controls thee reactive power flow in the transmission line
l by addiing FACTS eqquations (1)-(112) in the stanndard power
annd it can be useed to regulate th
he voltage mag
gnitude of the bus
b flow equuations.
to which it is connnected.
III. PROBLEM F
FORMULATIO
ON
Thee aim of optim mization algorrithm is to pllace FACTS
devices in order to ennhance voltage stability marggin of power
systemss, minimize reactive power loss of the syystem and to
reduce overall cost oof FACTS devvices. The muulti objective
optimizaation problem is formulated wwith non-lineaar constraints
as follow
ws:
A. Objjective Function
Opttimal placemeent of FACTS S devices is a composite-
objectivve optimizationn problem. T This optimizatiion problem
has threee objective fuunctions whichh must be simmultaneously
minimizzed. Mathemaatically, multti objective ooptimization
problemms can be writteen as per (13).
Fig.4 Schematic diagram of UPFC
C
Minn f (x) = [ f1 (x)), f2 (x), f3 (x)] (13)

It iis subjected tto a set of eequality and/oor inequality

constraiints. Where, x iis the vector off control variabbles.
Diffferent componnents of objecctive function f (x) are as
follows::

Nline
f1 (xx) = ∑ Q i,loss (14)
i=1

Funnction f1 (x) reepresents the tootal reactive poower loss of

the systtem. Where, Qi,loss is the rreactive powerr loss of ith
branch aand Nline repressents total num
mber of transmiission lines.
Fig.5 Steady statte model of UPFC Funnction f2 (x) reepresents the vvoltage stabilitty margin of
the systeem using L-inddex and given bby.
The UPFC canc be represen nted by a poweer injection mo
odel
simmilar to TCPS ST which is sh hown in Fig.3.. The total powwer f2 (xx) = max(Lj) (15)
injjection at the buses
b will be the sum of po owers injected by
serries and shunt voltage sourcces. The expresssions for pow wers Wheere, L-index forr the jth node iss computed as:
injjected by seriees voltage sourrce at both endds of transmisssion
linne as suggestedd in [5], are giv
ven by (7)-(10).
1 ∑ (16)
Pi = ViVse bse sin (δi - θse) + ViVj bse sin (δi - δj) (7)
Qi = Vi2 bse – ViVse bse cos(δi - θse) – ViVjbse coss(δi - δj) (8) Wheere, g is the nuumber of geneerator buses, Vi and Vj are
Pj= –VjVse bse sin (δj - θse) + ViVj bse sin (δj - δi) (9) the com
mplex values off bus voltages i and j respectivvely. Lower
value off L-index show
ws better marginn of voltage staability.
 4

The overall cost of FACTS controller depends upon size proposed method five numbers of locations have
and type of devices used which is represented by function f3(x) been chosen on the basis of [9].
in (17). (b) Types of devices to be located considered as the
second variable to be optimized. In the present work,
f3 (x) = ( Cupfc + Ctcsc+ Ctcpst ) (17) TCSC, TCPST and UPFC have been considered for
optimal placement.
Where, Cupfc, Ctcsc, and Ctcpst are the cost functions of (c) Size of FACTS devices is considered as third
UPFC, TCSC, and TCPST in (US$/ kVAr) respectively. Cost optimization variable.
functions for UPFC, TCSC, and TCPST are selected on the These variables are simultaneously optimized using
basis of [23]-[24] and given by (18)-(20). genetic algorithm search technique to obtain minimum
reactive power loss, improved voltage stability margin at
Cupfc = 0.0003s2 − 0.2691s+188.22 (18) minimum cost of FACTS devices.

Ctcsc = 0.0015s2 − 0.713s+153.75 (19) IV. GENETIC ALGORITHM FOR OPTIMAL PLACEMENT
OF FACTS DEVICES
Ctcpst = 100 (20)
The optimal locations of the FACTS devices can be
identified using the genetic algorithm (GA) search technique.
Where, s is the size of the FACTS devices in kVAr.
A brief overview GA is presented below:
B. Constraints
A. Overview of Genetic Algorithm
The optimal placement of FACTS devices is a constrained
Genetic algorithm is a global search and optimization
optimization problem which deals with equality and inequality
technique which is based on the mechanisms of natural
constraints. Equality constraints are power flow equations,
selection and genetics. It can search several possible solutions
which are given in general form as follows:
simultaneously and do not require any prior knowledge of the
objective function. It produces good quality solutions because
Nbus
it is independent of the choice of the initial parameters.
∑ PG − PD −PL = 0 (21)
Genetic algorithm is initiated with random creation of initial
i=1
population which represents possible solutions of the
optimization problem. Then the fitness of each individual is
Nbus
evaluated by the value of the objective function, which is
∑ QG − QD −QL = 0 (22)
called as fitness function. The new population is formed by
i=1
selecting the more fit individuals using genetic operators
(selection, crossover and mutation) until the assigned
Where, PG and QG denote total active and reactive power
maximum number of generations are reached or some form of
generations, PD and QD are the total active and reactive loads
convergence criterion has been met. Finally, the population
while PL and QL represents active and reactive line losses of
stabilizes and most of the individuals in the population are
the system.
found to be almost identical.
The Inequality constraints are given by (23)-(27).
B. Implementation of GA
1) Initialization
≤ ≤ (23) Each individual in the population represents a candidate
solution. A problem that involves more than one decision
≤ ≤ (24) variables, each variable is represented by a string. In this work
an individual is represented with three strings, where, the first
≤ (25) (24)
string of the individual is related to the locations of the
devices, second string corresponds to the types of the devices,
≤ ≤ (26) and the last string represents the values of the devices.
Three FACTS devices have been used in the present work
≤ ≤ (27) where TCSC is represented by 1while TCPST and UPFC are
represented by 2 and 3 respectively in second string. Only one
device is placed in a line. If more than one FACTS devices
Where, , , , and are the minimum
and maximum active and reactive power limits of appear in the same line, barring one remaining FACTS
devices are removed from that line and are placed in some
generator; , , ,and are the minimum and
other lines where FACTS devices are not located. Each
maximum limits of voltage and phase angle of ith bus,
FACTS device may appear at more than one location. The
while represents MVA limit of transmission system.
discrete values of the devices lie between 0 and 1; value 0
In this paper, the following variables are considered as the indicates minimum value while value 1 is corresponding to the
optimization variables: maximum value.
(a) Location for optimal placement of FACTS devices is
considered as the first variable to be optimized. In the
 5

2) Selection strategy
s locationns and sizes oof different typpes of FACTS S devices to
Selection iss a process in n which indiviiduals are cop pied improvee voltage stabbility margin, minimize reaactive power
according to theeir fitness funcction which is similar to thatt of loss andd cost of FACT TS devices. Thhe voltage stabbility margin
naatural selectio on found in biological systems. Poo orer is measuured by L-index. The lowerr value of L-inndex denotes
peerforming indiv viduals are weeeded out and betterb performming higher vvoltage stabilitty margin of thhe system. Thee problem is
haave greater th han average chance
c to be copied in next n formulaated as a multii-objective opttimization connsidering the
geeneration. Theere are a nu umber of sellection operattors minimizzation of reactiive power losss of the systemm, the cost of
prooposed in [11]]. Among them m, roulette wh heel selection anda FACTS devices to be installed, and L L-index of the system.
touurnament seleection are mo ost common. Roulette wh heel Maat lab programmming codes arre developed ffor GA, and
sellection is used in this work. Newtonn–Raphson looad flow alggorithm is m modified to
3) Crossover operation incorporrate FACTS ddevices. The reesults of simulaation studies
Crossover is a mixing op perator that sw waps informattion are sum
mmarized below w:
froom selected paarents. Crossov ver can occur att a single posittion
A. IEE EE 30-Bus Testt System
(siingle point croossover), or att number of different
d positions
(m
multiple point crossover).
c In th
his work two point
p crossoverrs is IEEEE 30-bus test system [25] ccontains 6 geneerator buses,
used in which tw wo crossover sites
s are choseen at random and a 24 loadd buses and 41 transmissionn lines. Genetiic algorithm
off
ffspring are produced
p by exchanging the informattion techniquue is used to find the optiimal size and location of
beetween the selected crossover sites. FACTS devices. In the proposedd method five numbers of
4) Mutation locationns have been chosen on the bbasis of ref. [99]. To obtain
Mutation accts as a backg ground operato or and is used d to the optimmal values of the variables tthe GA-based algorithm is
huunt the unexplo ored search spaace by randomlly altering the bits b implemeented with ddifferent contrrol settings. T The optimal
at one or more positions of the t chosen strring with a sm mall settings of the GA conntrol parameterrs are given bellow:
muutation probab bility. It main ntains the gen netic diversity by Populatiion size: 20
avvoiding premaature convergeence to locall optimum. The T Maximuum number of generations: 1000
proobability of mu utation is seleccted as 0.01, i.ee. one bit in ev
very Mutatioon probability: 0.01
huundred will be muted.
m Thee results of sstudies for baase load and contingency
The flow chart
c for the proposed genetic algorithm m is conditioons are presented below:
shown in Fig. 6. 1) B
Base load conddition
Thiis section preseents and discusses a case stuudy based on
the IEEEE 30-bus test system underr base load conndition. The
simulatiion results are shown in Tabble I. Out of fiive locations
exploredd, UPFCs of ssizes 1 and 9 M MVArs are to bbe placed in
lines 100-21 and 28-227 respectivelyy. The total coost of using
UPFCs are US$ 18600400/-. Lines 110-22 and 10-220 are found
to be ssuitable for pplacing TCSCss of values 00.118p.u.and
0.14p.u.. respectively, and it costs US$ 20345000/-. The line
6-10 is found to bee the optimum m locations foor installing
TCPST.. The size, cosst, and parameeters values off TCPST are
presenteed in the Tablee I.
TABLE I
SIMULATIONN RESULTS FORR IEEE 30-BUS SY
YSTEM
(BASE CA
ASE)
Type Parame Sizee Cost Qloss L- index
/Locati ter MVAAr (US $) (MVAr)
on in values w
with Wit wiitho With
line (p.u.) out h ut FACTS
(From F
FAC FAC FACT
TS TS
− to )
UPFC Vse = 1 188000 222 17 0.1136 0.072
10 -21 0.0051
θse
=61.320
UPFC Vse = 9 1672400
28- 27 0.0055
Fig. 6 Flo
ow chart θse =
32.400
TCSC Xtcsc = 1 153000
V. SIMU
ULATION RESU
ULTS 10- 22 -0.118
TCSC Xtcsc = 13 1881500
The effectivveness of pro
oposed approaach is illustraated 10- 20 -0.14
using the IEEE 30-bus and IE EEE 118-bus test
t systems. The
T TCPST VS = 18 5300000
tesst is carried out
o for base looad as well ass for contingenncy 6-10 0.0060
θS = -
coonditions. The objective off the test is to
t find optim mum 101.590
 6

Fig.88. Convergence chharacteristics of L--index for outage oof line 16-17

TABLE III
SIIMULATION RE ESULTS OF CONT TINGENCY ANA ALYSIS FOR
IEEE 300-BUS SYSTEM (BASE CASE)
Type Paramet Sizee Cost Qloss (MVAr) L
L- index
/Locati er values MVA Ar (US $)
on in (p.u.) WWith With W Witho With
Figg.7. Convergence characteristics
c of fitness
f function with GA technique for
f line oout FACT utt FACTS
IEEEE 30-bus system FFACT S FA
ACTS
(From S
− to )
Placement of these FAC CTS devices at the optim mum Simulationn results for continngency in line 3-4
loccations with op
ptimized parammeter values decreases
d the to
otal UPFC Vse = 1 187950 551 20 0.141 0.067
reaactive power lo
oss of the systtem from 22.24 4 MVAr to 177.72 28 - 27 0.0051
θse =
MV VAr. The L-index decreasses from 0.13 36 to 0.072 that
t 3.60
deenotes an imprroved voltage stability margin of the systeem. TCSC Xtcsc =- 2 304660
Thhe convergencee characteristiccs of fitness fu
unction using GA
G 10 - 20 0.1630
tecchnique is show
wn in Fig. 7. TCPST VS = 9 900000
18 - 19 0.0057
p.u.
2) Contingen ncy analysis θS =
In order to examine
e the efffectiveness of proposed meth hod -106.30
duuring line ou utage, conting gency analysis is perform med. TCPST VS = 3 300000
Coontingency anaalysis is condu ucted by remov ving one line ata a 12 - 13 0.0054.
θS =
tim
me consecutiveely. There aree 41 transmisssion lines, outt of -107.160
whhich only 38 siingle line outag ge contingenciies are consideered TCPST VS = 48 4800000
sinnce three tran nsmission linees (9–11, 12– –13, 25–26) are 19 - 20 0.0077
coonnected to iso olated buses. Although
A 38 single
s line outtage θS =
73.530
coontingencies an nalysis are cond ducted for IEE
EE 30-bus systeem, Simulationn results for continngency in line 16-17
buut due to lack of space, only y a small set of results for most
m UPFC Vse = 17 3123444 222 4.90 0.137 0.082
criitical lines [15]]are presented in Table II. 4 - 6 0.0059
The optimu um locations, optimized parameters
p vaalue θse =
settting, and the cost
c of FACTS S devices obtaained by propoosed 61.20
TCSC Xtcsc =- 7 1041800
GA A technique fo or different con ntingencies aree shown in Taable
6 - 10 0.4059
II. Comparison between reacttive power losss and L-index x of TCPST VS = 42 4200000
thee system beforre and after in nstallation of FACTS
F devices is 22 - 0.0073
alsso summarized d in this Table. It is observed d from this Taable 24 θS =
thaat reactive pow wer loss of thee system was 51.388
5 MVAr for -101.280
TCPST VS 9 900000
thee outage of linee 3-4. It was reeduced to 20.88 89MVAr after the 10 - =0.0056
FAACTS devicess were placed d at optimum m locations with
w 22 θS =
opptimized param meter values. L-- Index of the system decreaased - 101.030
froom 0.141 to 0.067which ind dicates improveed voltage pro ofile TCPST VS 8 800000
8 -28 =0.0055
T convergence characteristtic of L-index for
off the system. The θS =
ouutage of line 16 6-17 is shown in Fig. 8. Thee above resultss of - 99.580
stuudies confirm that
t proposed GA G method eff ffectively searcches
opptimum locatio on and size of different types t of FAC CTS B. IEEE E 118-Bus Testt System
deevices to impro ove the perform mance of poweer systems durring Thee proposed meethod is also applied on IE EEE 118-bus
coontingency. test systtem to investiggate its potentiial on large poower system.
 7

IE
EEE 118-bus test system [25] contains 54 geenerator buses,, 64
loaad buses and 186 transmisssion lines. Genetic
G algoritthm
tecchnique is useed to find the optimal sizes and locationss of
FAACTS devices.. In the prop posed method five numberss of
loccations have been chosen on the basis of ref. [9]. The T
meethod has beeen tested undeer heavy load and contingen ncy
coondition. The proposed GA G is implem mented with the
poopulation size of 20 and maximum generrations of 30 for
IE
EEE 118-bus teest system. Thee results of stud
dies are presen
nted
beelow:

1) Heavy loaad condition

The check the
t applicabilitty of proposed GA under heaavy
loaading, the load d on the systeem is increaseed by 50% of the
baase case. The optimal
o locationn, size, and paarameter valuess of Fig.9. Coonvergence characcteristics of reactivve power loss funcction for IEEE
118-buus system using GA technique
diffferent FACTS S devices are shown in Tab ble III along with
w
theeir optimum cost. Voltage stability marrgin and reacttive 2) C
Contingency annalysis for heaavily loaded sysstem
poower loss of thee original systeem is compared with the systtem To test the robbustness of thhe proposed ooptimization
inccorporated witth FACTS dev vices at optimu um locations. The
T techniquue, a larger size IEEE 118-bus system is chosen and
ressults of compaarison are presented in Tablee III. The L-ind dex contingeency is createdd by removing critical lines. The loading
off the system decreases
d from
m 0.107 to 0.0 082 when UPF FC, factor iss selected as 1..5 for all the looad buses to innvestigate its
TCCSC, and TCPST are instaalled at optim mum locations. It potentiaal under heavyy loading. Thee critical liness of IEEE
shows that voltage stability of o the system m is improved by 118-buss system are selected on the bbasis of [15].
prooper selection of FACTS devices
d at apprropriate locations TABLE IV
using proposed optimization
o method.
m A TCSC C of - 0.0307 p.u.
p SIMULAT TION RESULTS FOR IEEE 118-B BUS SYSTEM (CO ONTINGENCY
is installed in lin
ne 30-17. The negative
n valuee of Xtcsc deno otes ANA ALYSIS (1.5 BAS SE LOAD)
Type Paramet Sizee Cost Qloss (MVAr) LL- index
cappacitive naturee of series reaactance offered by TCSC. The T /Locati er values MVA Ar (US $)
prooper series co ompensation presented
p by TCSC
T at desiired on in (p.u.) W
With With W Witho With
loccation brings down reactivee power lossess of compensaated line out FACT
S
uut FACTS
F
FACT FACTS
system from 984 4.94 MVAr to 819.71MVAr. The convergence (From S
− to )
chharacteristic of reactive poweer loss function n is shown in the
Simulatioon results for contiingency in line 1144-115
Figg.9. UPFC Vse = 2 375366 985 838 00.107 0.105
Thus, it can be concluded from the analy ysis of the resu
ults 20 -21 4.86e-04
preesented in Taable III that performance
p of
o heavily load ded θse =
poower system can c be improv ved by propo osed optimizattion 7.200
TCSC Xtcsc =- 1 153040
strrategy. 11 - 13 0.0577
TCPST VS = 21 2100000
30 - 17 0.59e-03
TA
ABLE III
θS = -
SIMULAT TION RESULTS FORF IEEE 118-BU US SYSTEM
144.670
(1.5 TIMES BASE CASE)
TCPST VS= 12 1200000
Typpe Paramet Size Cost of
o Qloss (MVAAr) L- index
22 - 23 0.55e-03
/Loocati er values MVAr FACT TS with witth witho wiith θS =
on in (p.u.) devicee out FACCT ut FACTS
S 41.900
linee (US $)) FACT FACTS
S TCPST VS = 56 5600000
(Frrom
5 - 11 0.76e-03
− too )
θS = -
UPPFC Vse = 8 14887
700 984 819
9 0.107 0.0
082 151.920
20 -21 5.04e-04
Simulatioon results for contiingency in line 13--15
θse =
28.80 UPFC Vse = 4 748590 992 882 00.107 0.085
TC
CSC Xtcsc = - 1 15304
40 11 - 13 4.97e-04
30 -17 0.0307 θse =
14.40
TC
CPST VS = 20 000
20000 TCSC Xtcsc = - 8 1185136
5 - 11 0.59e-03 3 - 5 0.0778
θS = - TCSC Xtcsc = - 2 304660
149.930 5 - 11 0.0532
TC
CPST VS = 1 10000
00 TCPST VS = 17 1700000
16-- 17 0.50e-03 30 - 38 0.56e-03
θS = θS =
33.240 40.690
TC
CPST VS = 31 31000
000 TCPST VS= 42 4200000
44 -45 0.63e-03 43 - 44 0.70e-03
θS = θS
45.310 =38.330
 8

[3] M.Norroozian, L. Angquuist, M. ghandharri, and G.Anderssoon, “Improving

powwer system dynam mics by series-connnected FACTS ddevices,” IEEE
Tranns. Power Deliv., vvol. 12, no. 4, pp. 1635–1641, Oct. 11997.
[4] C. R R. Fuerte-Esquivell, E. Acha, and H. Ambriz-Perezz, “A thyristor
conttrolled series com mpensator model for the power floow solution of
pracctical power netwoorks” IEEE Trans. Power Syst., voll. 15, no. 1, pp.
58-664, Feb.2000.
[5] E.Achha, , C.R. Fuertee-Esquivel, H.Am mbriz-Perez, and C C. A.Camacho,
FAC CTS-Modelling annd simulation in power networks,, England:John
Wileey & Sons Ltd, 20004, pp.200–215.
[6] A.M. Vural and M. Tuumay, “Mathemattical modeling and analysis of a
unifi
fied power flow controller: A com mparison of two approaches in
powwer flow studies annd effect of UPFC Location”, Electr.. Power Energy
Syst.t., vol.29, pp. 6177-629, 2007.
[7] Bhow wmick, S., and D as, B., and Narenndra Kumar “An indirect UPFC
moddel to enhance rreusability of New wton power -flow w codes,” IEEE
Tranns. on Power Delivvery, vol. 23, no.4, pp.2079-2088, O Oct.2008.
[8] Rafaeel Mihalic,and U Uros Gabrijel, “Trransient stability assessment of
systeems comprising pphase-shifting FAC CTS devices by diirect methods” ,
Fiig.10. Convergencce characteristics of
o cost of FACTS devices
d for outagee of Elecct. Power and Enerrgy Syst., vol.26, ppp. 445-453, 2004.
line
l 114-115 for IE EEE 118-bus systeem [9] S. Gerrbex, R. Cherkaouui, and A. J. Germ mond, “Optimal loccation of multi-
The Table IV I provides a comparative study of the test typee FACTS devicees in a power system by meaans of genetic
ressults obtained for uncompen nsated 118-bus system and the algoorithms,”IEEE Traans. Power Syst.,, vol. 16, no. 3,, pp. 537–544,
system with optiimally located FACTS devices under differrent Augg.2001.
[10] D.J. Gotham and G.T T. Heydt, “Poweer flow control annd power flow
coontingencies. Out
O of five locaations explored d, UPFCs of sizze 2 studdies for systems w with FACTS deviices”, IEEE Transs. Power Syst.,
MV VAr is to be placed
p in lines 20-21 while lin ne 11-13 is fou
und vol.113, no.1, pp. 60–665, Feb.1998.
to be suitable for
fo placing TC CSCs of values 0.0577p.u. The T [11]D. E. Goldberg, Genetiic algorithms in seearch, optimizationn, and machine
linnes 30-17, 22--23, and 5-11 are found to be the optim mum learnning. New York: A Addison-Wesley, 11989.
[12]M. S Saravanan, S. M. R. Slochanal, P P.Venkatesh, P. S. Abraham. ,
loccations for insttalling TCPST T. The size, cosst, and parametters "Appplication of PSO technique for opptimal location of facts devices
vaalues of differeent FACTS dev vices are preseented in the Taable conssidering system loaadability and cost of installation" IE
EEE 2006.
IV
V. It is observeed from this Taable that reactive power losss of [13]H. Moori and Y. Goto, ““Parallel tabu searrch based method ffor determining
thee system is 985 5.544 MVAr for fo the outage ofo line 114-115 5. It optim mal allocation of FACTS in powerr systems.,” in Prooc. Power Con.
on PPower System Techhnology, pp. 1077––1082, 2000.
is reduced to 83 38.794 MVArr after the FA ACTS devices are [14] S. G Gerbex, R. Cherkkaoui, and A. J. Germond, “Optim mal location of
plaaced at optimu w optimized parameter valu
um locations with ues. FAC CTS devices to enhhance power systeem security,” in Prroc. 2003 IEEE
L- index of the system decreased from 0.1 107to 0.105wh hich Boloogna Power Tech C Conf., pp. 23–26.
inddicates improv ved voltage profile of the system. Fig. 10 [15] Y. LLu and A. Abur, “ Static security enhhancement via opttimal utilization
of thhyristor-controlledd series capacitors””, IEEE Trans. PPower Syst., vol.
shows the variaation of cost function with h respect to the 17, nno.2, pp, 324-3299, May 2002.
nuumber of generrations for outtage of line 11 14-115. It mayy be [16] L. Ipppolito and P. Siaano, “Selection off optimal number and location of
obbserved that the speed of convergence of the propo osed thyriistor-controlled phhase shifters usingg genetic based alggorithms,”Proc.
meethod is very high as the optimal
o solutio
on is achieved d in Inst.. Elect. Eng. Gen. Transm. Distrib, vvol. 151, no. 5, pp.630–637, Sep.
20044.
aroound 15 generaations which proves the excellent convergence [17]S. Raahimzadeh, M. T Tavakoli Bina, aand A.H. Viki, “Simultaneous
ratte of proposed method. appllication of mullti-type FACTS devices to thhe restructured
The results of studies fo or other contin ngencies are also
a enviironment: achievinng both optimal nuumber and locationn” , IET Gener.
shown in Table IV. Signifiicant improvement in volttage Trannsm. Distrib., vol. 4, no. 3, pp. 349––362, 2010.
[18]R. M Mínguez, F. Milanno, R. Z.Miñanoo, and A. J. Conejo, “Optimal
staability margin along with red duction in reactive power loss is netwwork placement off svc devices”, IE EEE Trans. Poweer Syst., vol.22,
obbserved from th his Table for otther critical con
ntingencies. no.44, pp, 1851-1860., Nov.2007.
[19]H.I. S Shaheen, G.I.Rashhed ,and S.J.Chengg , “Application aand comparison
of ccomputational inttelligence techniqques for optimall location and
VI. CONCLUSIONSS
paraameter setting of U UPFC”, Eng. Appp. Art. Int., vol. 223, pp.203-216,
In this pap
per, multiple FACTS
F devices allocation for 20100.
reaactive loss min
nimization andd voltage stabiility improvem
ment [20]S. Pananda, and N. P. P Padhy, “Optimal loocation and contrroller design of
STA ATCOM for pow wer system stabillity improvementt using PSO”,
off power systtems has been presented.. Besides otther Jourrnal of the Frank. Inst. , vol.345, pp. 166–181, 2008.
obbjectives, propposed GA teechnique also o considers cost
c [21] A. E E.D. Amgad, K.M M. Y. Hosam, M.M. E.Metwally, aand Z. Osman,
obbjective functio
on to find op ptimum locatiion and sizes of “O
Optimum VAR sizing and alloocation using pparticle swarm
diffferent FACTS S devices. The proposed metthod has excelllent opttimization” , Electt. Power Syst. Reseearch, vol.77, pp.9965-972, 2007.
[22] R. S.. Wibowo, N. Yorrino,, M. Eghbal, Y Y.Zoka, and Y. Saasaki ,“ FACTS
coonvergence ratee. The results of comparative studies conffirm deviices allocation w with control coorddination considerring congestion
thaat performancee of power sysstems are improoved by optimmum relieef and voltage stabbility” IEEE Tranns. Power Syst., vool. 26, no. 4, pp.
alllocation of FAC
CTS devices. 23022-2310, Nov. 20111.
[23] H.R R. Baghaee, M. Jaannati, B. Vahidi, S.H. Hosseinian, H. Rastegar, “
Imprrovement of voltaage stability and reduce power syystem losses by
VII. REFERENCES optim mal ga-based allo cation of multi-typpe facts devices” proc. IEEE Int.
[1] N.G. Hingoranii and L.Gyugyi, Understanding FACTS-concepts
F and Confnf. Optm. of Elect. EElectr. Equp. Rommania, pp.1–6 ,Mayy 2008.
technology off flexible AC trransmission systeems, India: Stand dard [24] A.A A. Alabduljabbar and J.V. Milannovic, “Assessmeent of techno-
Publishers Disttributors, 2001, pp. 2–5. econnomic contributionn of FACTS devices to power sysstem operation”
[2] M.Noroozian, L. Angquist, M. gh handhari, and G.A
Andersson, “ Use of Elecctric Power System ms Research, vol 80 , pp.1247–1255 ,2010.
UPFC for optim mal power flow coontrol”, IEEE Tra
ans. Power Deliv., vol. [25]Poweer System Test Archive-UWEE E (University off Washington)
12, no. 4, pp. 1629–1634, Oct. 19997. [Onlline].available: htttp://www.ee. washhington.edu/researcch/pstca.