Electrical Power and Energy Systems 78 (2016) 470–479

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

Loadability enhancement with FACTS devices using gravitational search

algorithm
Biplab Bhattacharyya ⇑, Sanjay Kumar
Dept. of Electrical Engineering, Indian School of Mines, Dhanbad, Jharkhand, India

a r t i c l e i n f o a b s t r a c t

Article history: In the present work, GSA (gravitational search algorithm) based optimization algorithm is applied for the
Received 2 September 2015 optimal allocation of FACTS devices in transmission system. IEEE 30 & IEEE 57 test bus systems are taken
Received in revised form 15 November 2015 as standards. Both active and reactive loading of the power system is considered and the effect of FACTS
Accepted 26 November 2015
devices on the power transfer capacity of the individual generator is investigated. The proposed approach
of planning of reactive power sources with the FACTS devices is compared with other globally accepted
techniques like GA (Genetic Algorithm), Differential Evolution (DE), and PSO (Particle Swarm
Keywords:
Optimization). From the results obtained, it is observed that incorporating FACTS devices, loadability
FACTS devices
Active power loss
of the power system increases considerably and each generator present in the system is being able to dis-
Operating cost patch significant amount of active power under different increasing loading conditions where the steam
Loadability flow rate is maintained corresponding to the base active loading condition. The active power loss & oper-
Gravitational search algorithm ating cost also reduces by significant margin with FACTS devices at each loading condition and GSA based
planning approach of reactive power sources with FACTS devices found to be the best among all the
methods discussed in terms of reducing active power loss and total operating cost of the system under
all active and reactive loading situations.
Ó 2015 Elsevier Ltd. All rights reserved.

Introduction voltage profile in the system as well as to minimise the system

transmission loss.
In power system, engineers and researchers are in process to The elementary idea of FACTS devices was first came into exis-
reduce reactive power and transmission loss to boost system effi- tence in 1988 [1]. Lagrangian decomposition approach is applied in
ciency. Reactive power has a deep effect on the security of power [2] for active power congestion management. Optimal placement
networks as it influence voltages throughout the entire network. of capacitor in a radial distribution system is presented in [3]. An
To increase the amount of active power that can be transferred elaborative discussion for the optimum placement of series capac-
across a congested transmission network, reactive power flows itor and phase shifter is presented in [4]. Sensitivity analysis and
must be minimized. Likewise, increase in reactive power genera- linear programming technique is presented for the optimal loca-
tion of a particular generator has impact on it’s active power gen- tion and size of Static Var Compensator (SVC) in a connected power
eration capacity. In addition, reactive power is essential (i) for the system in [5]. In [6], Authors have used TCSC device based on the
flow of active power through the transmission and distribution use of LMP (Locational Marginal Pricing) difference and congestion
system and (ii) to maintain the voltage to deliver active power rent. Real power performance index is used in [7], as an indicator
through transmission lines. Flexible AC Transmission System for the determination of location of Thyristor Controlled Series
(FACTS) devices can be effective for static as well as for dynamic Capacitor (TCSC) positions in a connected power network. In [8],
state of voltage control in power transmission and distribution. solution of transmission system congestion management problem
It’s principal function is to inject reactive power into the system is addressed by the authors using TCSC where the system loadabil-
which helps to support the system voltage profile. FACTS devices ity is increased keeping in view of the voltage stability of the sys-
regulate desired power flow in a power network provide the best tem. Optimal placement of TCSC for increasing loadability and
minimizing transmission loss by Genetic Algorithm (GA) is pre-
sented in [9]. Use of static phase shifters and series power flow
⇑ Corresponding author. controller (SPFC) and Unified Power Flow Controller (UPFC) to
E-mail addresses: [email protected] (B. Bhattacharyya), increase power transfer capacity in transmission lines is described
[email protected] (S. Kumar). in [10]. Solution technique for the power flow problem with TCSC

http://dx.doi.org/10.1016/j.ijepes.2015.11.114
0142-0615/Ó 2015 Elsevier Ltd. All rights reserved.
 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479 471

and UPFC devices is presented in [11]. Enhancement of available

transfer capacity with UPFC devices is described in [12]. Power
flow control approach in consideration with available transfer
capacity using static synchronous series compensator (SSSC), UPFC
and STATCOM devices is discussed in [13]. Authors have used
STATCOM, SSSC and UPFC devices in [14] for congestion manage-
ment considering voltage stability as loadability limit. Authors Fig. 1. TCSC model.
have proposed a technique in [15], for the optimal coordination
of SVC, TCSC and thyristor controlled phase angle regulator
(TCPAR) on the basis of demand responses in the restructured Cost function and problem formulation
power market. Optimal reactive power dispatch along with switch-
able TCSC and SVC devices is presented in [16]. Improvement in The objective of the proposed work is to minimize the transmis-
power flow control with TCSC, SVC and UPFC devices is presented sion loss of the system using FACTS devices under different loading
in [17]. This paper shows how the system loadability improves conditions. Increase in transmission loss as well as problem of volt-
with simultaneous use of multi type FACTS devices. An hybrid age stability is the main concern with the increased load. So, when
Genetic Algorithmic approach with TCSC and thyristor controlled the system loading is increased gradually, it requires reactive
phase shifter (TCPS) devices for optimal power flow is described power support to maintain voltage stability. Hence the main aim
in [18]. The placement of TCSC, TCPST, TCVAR and SVC devices in of the present work is to reduce the real power loss which is
a power system using GA is discussed in [19]. Utility of different expressed by Eq. (1) and to minimize voltage deviation at weak
types of UPFC, TCSC, PCPST and SVC devices in deregulated electric- buses under different loading conditions.
ity market is explained in [20]. In [21], congested areas of an inter- X
n
connected power networks are determined and then TCSC’s and PL ¼ g x ðv 2i þ v 2j  2v i v j Cos hij Þ ð1Þ
SVC are allocated using GA based optimization technique to solve x¼1
k¼ði;jÞ
transmission congestion. In [22], authors have developed a model
to solve congestion management problem by proper placement where g x is the conductance of line x, v i , v j are the voltages of ith
of unified power flow controller’s (UPFC) in suitable locations of and jth node respectively, and hi;j is the phase angle difference
the power system. PSO based solution methodology is applied in between ith and jth node.
determining proper size of UPFC to reduce the generation cost as Hence the objective of the present work is transmission loss
well as congestion cost in a restructured power market in [23]. minimization problem subject to the satisfaction of equality and
In order to minimize active power loss, improvement of voltage inequality constraints. Cost functions for TCSC’s and SVC’s are
profile and enhancement of voltage stability, GSA is proposed in given below:
[24]. In [25], simulation results indicate that GSA can provide effec- TCSC:
tive and robust high-quality solution for the OPF problem. Applica- C TCSC ¼ 0:0015S2  0:7130S þ 153:75 ðUS $=kVarÞ ð2Þ
bility of different computational algorithms for load ability
enhancement with TCSC, SVC, TCPST devices is presented in [26]. SVC:
In [27], authors have proposed the application of FACTS devices C SVC ¼ 0:0003S2  0:3051S þ 127:38 ðUS $=kVarÞ ð3Þ
in a deregulated environment for the solution of combined active
and reactive congestion management. In [28], authors suggested Here, S is the operating value of the FACTS devices. Energy cost is
model of three FACTS devices i.e. SVC, TCSC and TCPAR and unified taken as 0.06$/kW h and cost functions are obtained from [20].
into new FDLF (n-FDLF) program. By using above said program, a The main objective is to find the optimal location of FACTS
model of the Hellenic power system is developed. devices along with network constraints so as to minimize the total
In the present work, authors have implemented GSA based opti- operational cost and relieve transmission congestion at different
mization algorithm for the optimal planning of FACTS devices for loading conditions. Installation costs of various FACTS devices
the minimization of active power loss and operating cost of the
system under different loading conditions. Moreover, ability of
each generators to transfer active power in under different loading
conditions are investigated where the steam input to each genera-
tors are kept corresponding to the base demand.

Modeling of facts devices

Fig. 2. TCSC injection model.

For an interconnected congested power network FACTS devices

can be modeled as power injection model. The injection model
describes the FACTS as a device that injects a certain amount of real
and reactive power to a node. Both TCSC and SVC devices control
the power flow and voltages by adjusting the reactance of the sys-
tem. There are two possible characteristic for TCSCs; capacitive and
inductive, to increase or decrease the transmission line reactance.
These devices can cause increase in the transmission power capac-
ity of lines, static voltage security margin enhancement, voltage
profile improvement, and decrease in active power loss. SVCs have
also capacitive and inductive characteristics and are predominantly
utilized to improve and amend voltage in static and dynamic condi-
tions, reduce reactive power loss, and enhances static voltage secu-
rity margin. The injection power model and variable susceptance
model shown in Figs. 1–3. Fig. 3. Variable susceptance model of SVC.
472 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479

and the cost of system operation, namely, energy loss cost are com- TCSC SVC Transformer Reactive
bined to form the objective function to be minimized. Without tap setting Generation
of
FACTS devices, transmission loss can also be minimized by
Generators
optimization of reactive power, which is possible by controlling 4 nos. 4 nos. 4 nos. 5 nos.
reactive generations of the generator’s, controlling transformer
tap settings, and by the addition of shunt capacitors at weak buses. Fig. 4. String variables for IEEE-30 bus system.
As Var generations of the generators and controlling transformer
tap settings within their defined limits do not contribute any cost
to the operating cost of the system, here in the proposed approach TCSC SVC Transformer Reactive
setting of transformer tap positions and reactive generations of tap setting Generation
of
generators are included as controlling parameters along with the Generators
FACTS devices. 4 nos. 3 nos. 17 nos. 6 nos.
The objective function is combinatorial. It consists of two parts;
first is the cost due to the energy loss and second is the cost due to Fig. 5. String variables for IEEE-57 bus system.

investment of FACTS devices. So the optimization of the objective

function requires not only minimization of cost of energy loss
bus matrix for execution of load flow in evaluating the objective
which is related with the minimization of transmission loss by
function for each individual population of generation of
FACTS devices but also a optimization problem of investment cost
Gravitational Search and other evolutionary algorithms as DE, GA
of FACTS devices.
and PSO.
The optimal allocation of FACTS devices can be formulated as:
The objective function is calculated for all the individual of the
C TOTAL ¼ C 1 ðEÞ þ C 2 ðFÞ ð4Þ new generation and the procedure is repeated till the final goal is
reached. It is to be mentioned that the modified power flow equa-
where C1(E) is the cost due to energy loss and C2(F) is the total
tions represented by Eqs. (8)-(12) are also to be incorporated in the
investment cost of the FACTS devices.
load flow program in evaluating objective function. The number of
Objective function is minimized considering following active
each variables belonging to a particular type of parameter are
and reactive power balances and voltage magnitude constraints.
shown in Fig. 4 and 5 for both IEEE-30 and IEEE-57 bus systems
Pmin max respectively.
ni 6 P ni 6 P ni ð5Þ

Gravitational search algorithm

Q min max
ni 6 Q ni 6 Q ni ð6Þ
Gravitational search algorithm is a metaheuristic optimization
V min
i 6 V i 6 V max
i ð7Þ technique based on Newton’s law of gravity and motion. This algo-
Superscripts min, max are the minimum and maximum limits of the rithm was first developed by Rashedi et al. [23] in 2009. The work-
variables. ing of this algorithm is greatly influenced by the motion and the
Then, the power flow equations between the nodes i–j after mass of agents. Each agent experiences gravitational force of attrac-
incorporating FACTS devices would appear as tion with other agents present in the search space. Fitness of agents
in the search space is characterized by their masses. Hence, GSA can
TCSC:
be considered as collection of different masses. Heavier mass has
X
N1 bigger attraction force and attract other masses with a force propor-
PGi  PDi þ Pi Vi Vj ðGij cos hij þ Bij sin hij Þ ¼ 0 ð8Þ tional to the product of their masses and inversely proportional to
j¼1
the distance (not the square of distance) between masses.
The position of a mass corresponds to a solution in the search
X
N1
space. In due course of time all the masses or the agents will be
Q Gi  Q Di þ Q iðinjÞ  V i V j ðGij sin hij  Bij cos hij Þ ¼ 0 ð9Þ
attracted by the heaviest mass and the heaviest mass will repre-
j¼1
sent the optimum solution in the search space.Initially, N number
of agent are created and their positions are defined by
X
N1
PGi  PDi þ Pi  V j V j ðGjj cos hjj þ Bjj sin hjj Þ ¼ 0 ð10Þ X i ¼ ½x1i    xni  ð13Þ
j¼1
For i = 1, 2, . . . , n. Force of attraction between an active gravita-
X
N1 tional mass and a passive gravitational mass at any specific time t
Q Gj  Q Dj þ Q jðinjÞ  V j V j ðGjj sin hjj  Bjj cos hjj Þ ¼ 0 ð11Þ is given by
j¼1
M pi ðtÞ  Maj ðtÞ
SVC: F ij ðtÞ ¼ GðtÞ  ½X j ðtÞ  X i ðtÞ ð14Þ
Rij ðtÞþ 2
X
N1
where M pi is the passive gravitational mass and which is ith agent
Q Gi  Q Di þ Q iLðinjÞ  V i V j ðGij sin hij  Bij cos hij Þ ¼ 0 ð12Þ
j¼1
and Maj is the active gravitational mass which is the jth agent. Rij
is the Euclidian distance between the ith and jth mass. Є is a small
where PGi and Q Gi are the active and reactive generation and PDi and constant. The passive gravitational mass is attracted towards active
Q Di are the active and reactive demand at the ith node. G & B are the gravitational mass.
component of bus admittance matrix. Total force on ith agent is given by
P i and Q iðinjÞ are the real and reactive power flow change takes X
place at the nodes due to TCSC connected to a particular line F i ðtÞ ¼ ðrandÞ  F ij ðtÞ ð15Þ
i¼1;j–1
between the nodes i and j. Q iLðinjÞ is the reactive power injection
due to SVC. These changes in the power flow equations are taken where rand is a random number in the interval [0, 1].
into consideration by appropriately modifying the admittance The acceleration of the ith agent at time t is given by
 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479 473

Table 1 where G0 is the initial value of the Gravitational constant, chosen

Locations of different FACTS devices in the transmission network in IEEE-30 bus randomly, a is a constant, t is the current generation and T is the
system.
total number of generations.
TCSC in lines SVC in buses The velocity and the position of the ith agent are updated by the
25, 41, 28, 5 21, 7, 17, 15 following equation:

V i ðt þ 1Þ ¼ randi  V i ðtÞ þ ai ðtÞ ð18Þ

Table 2 X i ðt þ 1Þ ¼ X i ðtÞ þ V i ðt þ 1Þ ð19Þ

Locations of different FACTS devices in the transmission network in IEEE-57 bus
system. Mass of each agent is determined by it’s fitness. The masses of
TCSC in lines SVC in buses agents are updated by the following equation.
37, 13, 61, 57 49, 25, 38 mi ðtÞ
Mi ðtÞ ¼ Pn ð20Þ
j¼1 mij ðtÞ

F i ðtÞ fit i ðtÞworstðtÞ

ai ðtÞ ¼ ð16Þ where mi ðtÞ ¼ bestðtÞworstðtÞ
Mi ðtÞ
In the present problem reactive power sources along with the
Gravitational constant G(t), at time t is computed as follows: amount of FACTS devices are represented by a string and each
  string is nothing but the solution agent in the search space. Each
t
GðtÞ ¼ G0 exp a ð17Þ agent will have mass according to their fitness values evaluated
T from the objective function.

Table 3
Comparative study of total reactive power flow in line with GA, DE, PSO and GSA in IEEE-30 bus system.

Loading (%) Pnl Pnl Pnl Pnl Pnl

l Q L Without FACTS l Q L With FACTS using GA l Q L With FACTS using DE l Q L With FACTS using PSO l Q L With FACTS using GSA

100 0.2994 0.4197 3.5822 0.3614 0.6104

110 0.0185 0.4842 1.1510 1.1864 0.2222
120 0.2862 1.3166 2.5170 1.8457 2.1400

Table 4
Operating cost and active power loss analysis without and with FACTS devices in IEEE-30 bus system.

Loading Pd Active power loss Operating cost due to Active power loss with loss Evolutionary Operating Cost of FACTS Net
and Qd (%) without FACTS energy loss in ($) (A) with FACTS devices in (pu) methods with cost in ($) (B) devices in ($) saving
devices in (pu) FACTS devices (A–B)
100 0.0711 3,737,016 0.0406 GA 2.1786  106 44,664 1,558,416
0.0406 DE 2.1770  106 43,064 1,560,016
0.0445 PSO 2.4052  106 66,280 1,331,816
0.0390 GSA 2.1481  106 98,260 1,588,916
110 0.0974 5,120,900 0.0585 GA 3.1222  106 47,440 1,998,700
0.0584 DE 3.1222  106 52,700 1,998,900
0.0639 PSO 3.4361  106 77,516 1,684,800
0.0581 GSA 3.1224  106 68,700 1,998,500
120 0.1294 6,800,100 0.0839 GA 4.4915  106 81,716 2,308,600
0.0839 DE 4.4915  106 81,716 2,308,600
0.0891 PSO 4.7774  106 94,304 2,022,700
0.0824 GSA 4.4230  106 92,056 2,377,100

Table 5
Operating cost and active power loss analysis without and with FACTS devices in IEEE-57 bus system.

Loading Pd Active power loss Operating cost due to Active power loss with loss Evolutionary Operating Cost of FACTS Net saving
and Qd (%) without FACTS devices energy loss in ($) (A) with FACTS devices in (pu) methods with FACTS cost in ($) devices in ($) (A–B)
in (pu) devices (B)
100 0.2799 14,712,000 0.2165 GA 1.1440  107 60,760 3,272,000
0.2168 DE 1.1465  107 69,992 3,247,000
0.2276 PSO 1.2059  107 96,344 2,653,000
0.2145 GSA 1.1429  107 154,880 3,283,000
110 0.4168 21,907,000 0.2997 GA 1.5840  107 87,768 6,067,000
0.2997 DE 1.5846  107 93,768 6,061,000
0.3155 PSO 1.6674  107 91,320 5,233,000
0.2989 GSA 1.5830  107 119,816 6,077,000
120 0.6091 32,015,000 0.3075 GA 1.6318  107 156,000 15,697,000
0.3081 DE 1.6364  107 170,264 15,651,000
0.3221 PSO 1.7081  107 151,424 14,934,000
0.3012 GSA 1.5984  107 152,928 16,031,000
474 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479

Table 6 Results and discussion

Percentage of active power loss reduction at different loading using GA, DE, PSO and
GSA techniques in IEEE 30 bus system.
In this paper two types of FACTS devices namely TCSC (Thyris-
Evolutionary % Loss reduction % Loss reduction for % Loss reduction for tor Controlled Series Capacitor) and SVC (Static Var Compensator)
techniques for base loading 110% of base 120% of base
are considered. First the location of FACTS devices are defined by
loading loading
calculating the power flow in the transmission lines. SVC’s are con-
GA 75.12 66.50 54.23
nected at the buses 21, 17, 15 and 7 the finishing ends of the lines
DE 75.12 66.78 54.23
PSO 59.78 52.43 45.23 27, 26, 18 and 9 respectively, since these are the four lines carrying
GSA 82.31 63.97 52.77 highest, second highest, third highest and fourth highest reactive
powers respectively. Then TCSC’s positions are selected by
choosing the lines carrying large reactive power. In IEEE 30 bus test
system, lines 25, 41, 28 & 5 carry significant amount of reactive
Table 7 powers and are selected as candidate lines for the placement of
Percentage of active power loss reduction at different loading using GA, DE, PSO and Thyristor Controlled Series Capacitor. In IEEE-57 bus test system,
GSA techniques in IEEE 57 bus system. four locations are selected for TCSC’s and three locations are
Evolutionary % Loss reduction % Loss reduction for % Loss reduction for selected for SVC’s. Lines 37th, 13th, 61st and 57th are found as
techniques for base loading 110% of base 120% of base the lines for TCSC placement and simultaneously series reactance
loading loading of these lines are controlled. Bus number 49st, 25th and 38th are
GA 29.28 39.07 98.08 found as the candidate buses for the placement of SVC’s. Here
DE 29.11 39.07 97.70 GA, DE, PSO and GSA based optimization techniques are applied
PSO 22.98 32.11 89.10
on IEEE 30 and IEEE 57 bus system. GSA and other optimization
GSA 23.52 40.72 75.94
techniques are used to determine optimum magnitudes of FACTS

Table 8
Active power flow and reactive power flow with FACTS devices under different loading in IEEE 30 bus system.

Location of Generator Line in between Methods Active power flow between Reactive power flow between Phase angle between buses
generator connected generator and the buses buses
with bus bus (Eb)
Base 1.10⁄ 1.20⁄ Base 1.10⁄ 1.20⁄ Base 1.10⁄ 1.20⁄
(Eb)
(Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd)
2 4 3 GA 0.2916 0.3470 0.3888 0.0495 0.0165 0.0266 0.0458 0.0553 0.0636
DE 0.2921 0.3535 0.3981 0.0728 0.0159 0.1401 0.0448 0.0551 0.689
PSO 0.2149 0.2587 0.2885 0.0146 0.0931 0.0431 0.0443 0.0517 0.0537
GSA 0.2538 0.3123 0.3516 0.1196 0.1059 0.2599 0.0414 0.0491 0.0567
5 5 GA 0.5852 0.6656 0.7520 0.0384 0.2069 0.1315 0.1101 0.1207 0.1364
DE 0.5865 0.6582 0.7357 0.0383 0.2662 0.1760 0.1104 0.1146 0.1307
PSO 0.8647 0.9753 1.1152 0.0384 0.0492 0.1202 0.1112 0.0722 0.0752
GSA 0.7489 0.7744 0.8740 0.0502 0.1980 0.0162 0.0792 0.0955 0.1094
6 6 GA 0.3827 0.4535 0.5170 0.0033 0.0137 0.0393 0.0645 0.0744 0.0863
DE 0.3835 0.4520 0.5239 0.0084 0.0061 0.1821 0.0642 0.0732 0.0918
PSO 0.2860 0.3442 0.3965 0.0071 0.1388 0.0872 0.0632 0.0705 0.0705
GSA 0.3319 0.4168 0.4747 0.1961 0.1244 0.2555 0.0559 0.0667 0.0667

5 2 5 GA 0.5852 0.6656 0.7520 0.0384 0.2069 0.1315 0.1101 0.1207 0.1364

DE 0.5865 0.6582 0.7357 0.0383 0.2662 0.1760 0.1104 0.1146 0.1307
PSO 0.8647 0.9753 1.1152 0.0384 0.0492 0.1202 0.1112 0.0722 0.0752
GSA 0.7489 0.7744 0.8740 0.0502 0.1980 0.0162 0.0792 0.0955 0.1094
7 8 GA 0.1264 0.1591 0.1711 0.0499 0.0719 0.0710 0.0168 0.0163 0.0166
DE 0.1252 0.1635 0.1763 0.0607 0.1277 0.2219 0.0172 0.0279 0.0101
PSO 0.1034 0.0988 0.1211 0.0806 0.1969 0.1816 0.0190 0.0095 0.0117
GSA 0.0047 0.0640 0.2056 0.2039 0.0322 0.1526 0.0044 0.0061 0.0078

8 6 10 GA 0.0081 0.0081 0.0019 0.1918 0.0379 0.0004 0.0019 0.0003 0.0020

DE 0.0076 0.0121 0.0250 0.2414 0.1581 0.4152 0.0011 0.0013 0.0071
PSO 0.1266 0.0473 0.0453 0.2323 0.3944 0.3851 0.0041 0.0022 0.0020
GSA 0.0257 0.0210 0.0193 0.3057 0.1819 0.0616 0.0031 0.0015 0.0021
28 40 GA 0.0415 0.0135 0.0129 0.0378 0.0055 0.0153 0.0057 0.0075 0.0069
DE 0.0418 0.0127 0.0043 0.0463 0.0203 0.1084 0.0088 0.0085 0.0022
PSO 0.0721 0.0420 0.0736 0.0502 0.1668 0.1563 0.0039 0.0192 0.0248
GSA 0.0122 0.0038 0.0065 0.2016 0.0238 0.0229 0.0007 0.0008 0.0005

11 9 13 GA 0.1793 0.1709 0.1672 0.1433 0.2953 0.2835 0.0322 0.0288 0.0289

DE 0.1793 0.1711 0.1669 0.1140 0.2150 0.0632 0.0320 0.0283 0.0292
PSO 0.1750 0.1717 0.1683 0.0634 0.4542 0.4605 0.0315 0.0246 0.0249
GSA 0.1730 0.1717 0.0898 0.2316 0.4520 0.2502 0.0319 0.0323 0.0268

13 12 16 GA 0.1432 0.1590 0.1554 0.3766 0.2007 0.2204 0.0194 0.0185 0.0181

DE 0.1390 0.1581 0.1544 0.4045 0.6201 0.4264 0.0192 0.0154 0.0162
PSO 0.1642 0.1607 0.1579 0.0259 0.0511 0.0589 0.0193 0.0180 0.0195
GSA 0.1624 0.1606 0.1563 0.0457 0.1132 0.4026 0.0226 0.0226 0.0198
*
Sign is used to indicate multiplication operation.
 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479 475

Table 9
Active power flow and reactive power flow with FACTS devices under different loading in IEEE 57 bus system.

Location of Generator Line in Methods Active power flow between Reactive power flow between Phase angle between buses
generator connected between buses buses
with bus generator and
Base 1.10⁄ 1.20⁄ Base 1.10⁄ 1.20⁄ Base 1.10⁄ 1.20⁄
(Eb) the bus (Eb)
(Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd)
2 3 2 GA 1.0153 1.4324 1.7292 0.0044 0.0093 0.1103 0.0837 0.1153 0.1491
DE 1.0014 1.3489 1.7488 0.0045 0.1889 0.4796 0.0837 0.01195 0.1450
PSO 1.0161 1.3530 1.7650 0.0048 0.1891 0.0196 0.0839 0.1212 0.1549
GSA 1.0213 1.3754 1.8632 0.1282 0.1879 0.1601 0.1279 0.1223 0.1529

3 4 3 GA 0.6250 0.8774 1.1273 0.0857 0.0997 0.2522 0.0240 0.0322 0.0430

DE 0.6234 0.8759 1.1470 0.0839 0.02463 0.1468 0.0420 0.0296 0.0492
PSO 0.6130 0.8599 1.1307 0.0881 0.2028 0.3078 0.0237 0.0300 0.0440
GSA 0.6266 0.9270 1.1260 0.2032 0.2858 0.7901 0.0340 0.0326 0.0426
2 2 GA 1.0153 1.4324 1.7292 0.0044 0.0093 0.1103 0.0837 0.1153 0.1491
DE 1.0014 1.3489 1.7488 0.0045 0.1889 0.4796 0.0837 0.1195 0.1450
PSO 1.0161 1.3530 1.7650 0.0048 0.1891 0.0196 0.0839 0.1212 0.1549
GSA 1.0213 1.3754 1.8632 0.1282 0.1879 0.1601 0.1279 0.1223 0.1529

6 4 5 GA 0.1498 0.2666 0.3834 0.0292 0.0210 0.0936 0.0232 0.0465 0.0681

DE 0.1515 0.2593 0.3985 0.0296 0.1007 0.0378 0.0232 0.0412 0.0735
PSO 0.1453 0.2641 0.3787 0.0274 0.0475 0.0951 0.0227 0.0433 0.0706
GSA 0.1418 0.2856 0.3820 0.1733 0.0841 0.3182 0.0619 0.0516 0.0711
7 6 GA 0.1678 0.1115 0.0426 0.0127 0.0168 0.0370 0.0177 0.0120 0.0040
DE 0.1691 0.1099 0.0344 0.0145 0.0289 0.0915 0.0179 0.0114 0.0024
PSO 0.1742 0.1100 0.0441 0.0021 0.0340 0.0027 0.0185 0.0126 0.0033
GSA 0.1942 0.0591 0.0059 0.0241 0.1317 0.4013 0.0077 0.0039 0.0016
8 7 GA 0.4366 0.3860 0.3179 0.0417 0.1508 0.2225 0.0742 0.0651 0.0545
DE 0.4363 0.3838 0.3090 0.0417 0.1581 0.2963 0.0742 0.0641 0.0533
PSO 0.4427 0.3885 0.3243 0.0417 0.1052 0.1966 0.0742 0.0674 0.0539
GSA 0.4347 0.3987 0.3312 0.0293 0.1531 0.4221 0.0113 0.0504 0.0461

8 6 7 GA 0.4366 0.3860 0.3179 0.0417 0.1508 0.2225 0.0742 0.0651 0.0545

DE 0.4363 0.3838 0.3090 0.0417 0.1581 0.2963 0.0742 0.0641 0.0533
PSO 0.4427 0.3885 0.3243 0.0417 0.1052 0.1966 0.0742 0.0674 0.0539
GSA 0.4347 0.3987 0.3312 0.0293 0.1531 0.4221 0.0113 0.0504 0.0461
9 8 GA 1.7773 1.6595 0.3179 0.2288 0.6910 1.2047 0.0876 0.1117 0.0784
DE 1.7742 1.6666 0.3090 0.2288 0.6522 1.4197 0.0877 0.0849 0.0764
PSO 1.7731 1.6578 0.3243 0.2283 0.6231 1.2365 0.0880 0.0834 0.0764
GSA 1.8216 1.5499 0.3312 0.8540 0.6409 0.4433 0.0113 0.0588 0.0597

9 8 8 GA 1.7773 1.6595 1.5065 0.2288 0.6910 1.2047 0.0876 0.1117 0.0784

DE 1.7742 1.6666 1.4805 0.2288 0.6522 1.4197 0.0877 0.0849 0.0764
PSO 1.7731 1.6578 1.4902 0.2283 0.6231 1.2365 0.0880 0.0834 0.0764
GSA 1.8216 1.5499 1.3922 0.8540 0.6409 0.4433 0.0113 0.0588 0.0597
10 9 GA 0.1695 0.1379 0.1001 0.0879 0.0169 0.0677 0.0318 0.0273 0.0194
DE 0.1699 0.1380 0.0963 0.0860 0.0184 0.1290 0.0319 0.0270 0.0164
PSO 0.1667 0.1428 0.1009 0.0816 0.0237 0.1090 0.0333 0.0211 0.0210
GSA 0.1641 0.1274 0.1170 0.0130 0.0736 0.2100 0.0148 0.0201 0.0185
11 10 GA 0.1365 0.0575 0.0118 0.0329 0.0170 0.0129 0.0127 0.0050 0.0028
DE 0.1377 0.0596 0.0189 0.0279 0.0088 0.0703 0.0127 0.0054 0.0048
PSO 0.1374 0.0419 0.0341 0.0613 0.0077 0.0196 0.0133 0.0040 0.0068
GSA 0.1268 0.0465 0.0455 0.2477 0.0792 0.6476 0.0209 0.0037 0.0113
12 11 GA 0.0286 0.0049 0.0353 0.1213 0.0808 0.0274 0.0152 0.0053 0.0119
DE 0.0272 0.0034 0.0425 0.1213 0.0759 0.0223 0.0153 0.0053 0.0195
PSO 0.0284 0.0026 0.0355 0.1218 0.0712 0.0236 0.0189 0.0044 0.0186
GSA 0.0188 0.0101 0.0409 0.0054 0.1489 0.0867 0.0479 0.0046 0.0117
13 12 GA 0.0232 0.0600 0.1347 0.0722 0.0548 0.0772 0.0070 0.0085 0.0253
DE 0.0213 0.0583 0.1428 0.0707 0.0649 0.0255 0.0068 0.0076 0.0292
PSO 0.0228 0.0575 0.1386 0.0809 0.0317 0.0350 0.0078 0.0094 0.0331
GSA 0.0049 0.0908 0.1830 0.2079 0.0359 0.5832 0.0497 0.0235 0.0417
55 80 GA 0.1476 0.1511 0.1448 0.7389 0.6444 0.5190 0.0172 0.0203 0.0225
DE 0.1476 0.1520 0.1431 0.7377 0.6477 0.4858 0.0172 0.0203 0.0314
PSO 0.1479 0.1522 0.1454 0.2803 0.6510 0.5174 0.0184 0.0207 0.0222
GSA 0.2250 0.1290 0.1274 0.1133 0.4008 0.3069 0.0053 0.0153 0.0198

12 9 11 GA 0.0286 0.0049 0.0353 0.1213 0.0808 0.0274 0.0152 0.0053 0.0119

DE 0.0272 0.0034 0.0425 0.1213 0.0759 0.0223 0.0153 0.0053 0.0195
PSO 0.0284 0.0026 0.0355 0.1218 0.0712 0.0236 0.0159 0.0044 0.0186
GSA 0.0188 0.0101 0.0409 0.0054 0.1489 0.0867 0.0479 0.0046 0.0117

(continued on next page)

476 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479

Table 9 (continued)

Location of Generator Line in Methods Active power flow between Reactive power flow between Phase angle between buses
generator connected between buses buses
with bus generator and
Base 1.10⁄ 1.20⁄ Base 1.10⁄ 1.20⁄ Base 1.10⁄ 1.20⁄
(Eb) the bus (Eb)
(Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd) (Pd, Qd)

10 23 GA 0.1642 0.1999 0.2156 0.1632 0.2092 0.1523 0.0166 0.0220 0.0313

DE 0.1688 0.1959 0.2225 0.1655 0.1988 0.1169 0.0166 0.0217 0.0359
PSO 0.1598 0.2004 0.2136 0.1720 0.1957 0.0887 0.0174 0.0167 0.0396
GSA 0.1814 0.2028 0.2520 0.0035 0.2519 0.0858 0.0331 0.0247 0.0302
13 25 GA 0.0443 0.1084 0.1813 0.4305 0.2568 0.0885 0.0082 0.0138 0.0134
DE 0.0373 0.1136 0.1811 0.4347 0.2039 0.1953 0.0085 0.0129 0.0097
PSO 0.0511 0.1167 0.2048 0.4094 0.2723 0.2248 0.0081 0.0138 0.0145
GSA 0.0624 0.1401 0.2785 0.5771 0.6687 1.1523 0.0018 0.0189 0.0270
16 26 GA 0.3334 0.5073 0.6460 0.0961 0.0660 0.2505 0.0292 0.0447 0.0615
DE 0.3423 0.5004 0.6424 0.0962 0.0634 0.3341 0.0275 0.0441 0.0619
PSO 0.3256 0.5024 0.6253 0.0964 0.0641 0.2971 0.0275 0.0438 0.0633
GSA 0.3431 0.5126 0.6155 0.6780 0.0086 1.2483 0.0399 0.0436 0.0600
17 27 GA 0.4632 0.9892 0.8560 0.1130 0.0925 0.2855 0.0888 0.1322 0.1740
DE 0.4717 0.6819 0.8528 0.1131 0.0898 0.3714 0.0871 0.1308 0.1693
PSO 0.4557 0.6840 0.8337 0.1136 0.0906 0.3342 0.0873 0.1303 0.1784
GSA 0.4725 0.6946 0.8249 0.5317 0.0163 0.9740 0.1173 0.1293 0.1705
*
Sign is used to indicate multiplication operation.

Fig. 6. Variation of operating cost with generation for base loading using GA, DE, Fig. 7. Variation of operating cost with generation for 110% of base loading using
PSO and GSA for IEEE 30 bus system. GA, DE, PSO and GSA for IEEE 30 bus system.

devices placed at different locations already defined and system per- The loss reduction is calculated on the basis of the loss occurred
formance is observed without and with FACTS devices. Tables 1 and without and with FACTS devices for 100%, 110% and 120% of base
2 shows the locations of different FACTS devices in the transmission loading conditions. Hence it is clear from the Tables 6 and 7 that
lines in IEEE 30 and IEEE 57 bus systems. Table 3 shows the total optimal placement of FACTS devices in the systems can effectively
reactive power loss of the system without and with FACTS devices. reduce the transmission loss of the system. GSA method can be an
A comparative study of the operating cost of the system without and effective method for the planning of FACTS devices in reducing
with FACTS devices using GA, DE, PSO and GSA technique under dif- active power loss and reactive power flow in all the congested
ferent loading conditions are shown in Table 4 for IEEE 30 bus sys- lines. Effect of FACTS devices on the active power flow, reactive
tem and Table 5 for IEEE 57 bus system. Tables 6 and 7 shows power flow and the phase angle in lines connected between
comparative study of the percentage loss reduction without and generator bus and load bus is shown in Tables 8 and 9 for both
with FACTS devices using GA, DE, PSO and GSA technique under dif- the systems with proposed approach. From the results, as observed
ferent loading conditions for both the systems respectively. from Tables 8 and 9, it is clear that each generator in both IEEE 30
From Table 3, it is observed that after installation of FACTS and IEEE 57 bus test systems are being able to transfer significant
devices in the pre-defined location, the sum of reactive power flow amount of active powers through the lines connected to other load
of all lines reduces considerably. The decrease is more in case of buses. Similarly, there is considerable reduction of reactive power
GSA based optimization algorithm. flow in these lines. This phenomenon is found to be true for all
It has been observed that operating cost and active power loss cases of increased loading condition.
using GA, DE, PSO and GSA algorithms is reduced significantly in Variations of operating cost with generation using GA, DE, PSO
all cases of loading with FACTS devices as well as significant eco- and GSA techniques under different loading conditions for IEEE-30
nomic gain is obtained. The economic gain obtained is much higher bus test system are shown in Figs. 6–8. Similarly, Figs. 9–11 shows
than the installation cost of FACTS devices in all cases of loading. variation of operating cost with generation using GA, DE, PSO and
 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479 477

Fig. 11. Variation of operating cost with generation for 120% of base loading using
Fig. 8. Variation of operating cost with generation for 120% of base loading using
GA, DE, PSO and GSA for IEEE 57 bus system.
GA, DE, PSO and GSA for IEEE 30 bus system.

Fig. 9. Variation of operating cost with generation for base loading using GA, DE, Fig. 12. Active power flow with active and reactive loading for generator 2
PSO and GSA for IEEE 57 bus system. connected to bus 4.

Fig. 10. Variation of operating cost with generation for 110% of base loading using Fig. 13. Active power flow with active and reactive loading for generator 11
GA, DE, PSO and GSA for IEEE 57 bus system. connected to bus 9.
478 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479

GSA techniques for IEEE-57 bus test system under different cases of
loadings. Optimization algorithms are run for 100 generations and
number of populations are taken as 80 for IEEE-30 bus system. For
IEEE-57 bus system, optimization algorithms are run for 200 gen-
erations as number of string variables are more in 57 bus system
compared to 30 bus system. The number of populations are taken
as 80 as in 30 bus system.
Effect of FACTS devices on some selected generator buses are
shown by, Figs. 12 and 13 for IEEE 30 bus test system. Similarly
effect of FACT devices on some generator buses of IEEE 57 bus test
system is shown in Figs. 14–16.

Conclusion

In this paper the usefulness of GA, DE, PSO and GSA based opti-
mal placement of FACTS devices in a transmission network is
tested for the increased loadability of the power system as well
as to minimize the total operating cost and total active power loss.
Fig. 14. Active power flow with active and reactive loading for generator 2 Results showed that the proposed GSA algorithm is efficient for
connected to bus 3. reduction of power losses, improvement of the voltage profile as
well as reduction of total operating cost maintaining all the con-
straints. From the convergence characteristics of operating cost
with iteration for different optimization techniques, it is found that
GSA is the best among all the methods. It gives minimum operating
cost and active power loss among all the methods. Furthermore,
the effect of FACTS devices on transfer capability of each generator
present in the transmission system is investigated and it is found
that even at increased loading condition, generators are being able
to dispatch significant amount of active power without increasing
steam flow rate corresponding to the base loading situation. This is
the significant contribution of FACTS (Flexible AC Transmission
System) devices when both active & reactive loading is considered
though it is obvious that FACTS devices is able to inject large
amount of reactive power into the system under only reactive
loading cases. Hence, key issue is the optimum co-ordination of
FACTS (Flexible AC Transmission System) devices with the existing
reactive power sources for achieving all the mentioned benefits.
Here, in the proposed work, the result shows that GSA can be a
proper optimization algorithm for optimal planning of FACTS
devices for the enhancement of loadability of the power system.
Fig. 15. Active power flow with active and reactive loading for generator 12
connected to bus 10. References

[1] Hingorani N. Flexible AC transmission. IEEE Spect 1993;30(4):40–5.

[2] Wang X, Song YH, Lu Q. Lagrangian decomposition approach to active power
congestion management across interconnected regions. IEE Proc Gen, Trans Dis
2001;148(5):497–503.
[3] Mohamed Shuaib Y, Surya Kalavathi M, Christober Asir Rajan C. Optimal
reconfiguration in radial distribution system using gravitational search
algorithm. Electr Power Comp Syst 2014;42(7):703–15.
[4] Lie TT, Deng W. Optimal flexible AC Transmission systems (FACTS) devices
allocation. Int J Electr Power Energy Syst 1997;19(2):125–34.
[5] Ibrahim E, Abusorrah A. Optimal allocations of SVCs for improvement of power
system performance. Electr Power Comp Syst 2003;31(1):27–46.
[6] Acharya Naresh, Mithulananthan N. Locating series FACTS devices for
congestion management in deregulated electricity markets. Electr Power
Syst Res 2007;77(3–4):352–60.
[7] Besharat Hadi, Taher Seyed Abbas. Congestion management by determining
optimal location of TCSC in deregulated power systems. Int J Electr Power
Energy Syst 2008;30(10):563–8.
[8] Colvara LD, Araujo SCB, Festraits EB. Stability analysis of power system
including facts (TCSC) effects by direct method approach. Int J Electr Power
Energy Syst 2005;27:264–74.
[9] Abdelaziz AY, EL-Sharkawy MA, Attia MA. Optimal location of thyristor-
controlled series compensators in power systems for increasing loadability by
genetic algorithm. Electr Power Comp Syst 2011;39(13):1373–87.
[10] Nelson R, Bian J, Williams S. Transmission series power flow control. IEEE
Trans Power Deliv 1995;10(1):504–10.
[11] Padhy Narayana Prasad, Abdel Moamen MA. Power flow control and solutions
with multiple and multi-type FACTS devices. Electr Power Syst Res
2005;74:341–51.
Fig. 16. Active power flow with active and reactive loading for generator 12 [12] Xiao Y, Song YH, Liu Chen-Ching, Sun YZ. Available transfer capability
connected to bus 13. enhancement using FACTS devices. IEEE Trans Power Syst 2003;18(1):305–12.
 B. Bhattacharyya, S. Kumar / Electrical Power and Energy Systems 78 (2016) 470–479 479

[13] Xiao Y, Song YH, Sun YZ. Power flow control approach to power systems with [22] Verma KS, Singh SN, Gupta HO. Location of unified power flow controller for
embedded FACTS devices. IEEE Trans Power Syst 2002;17(4):943–50. congestion management. Electr Power Syst Res 2001;58(2):89–96.
[14] Kumar Ashwani, Sekhar Charan. Congestion management with FACTS devices [23] Hajforoosh S, Nabavi SMH, Masoum MAS. Coordinated aggregated-based
in deregulated electricity markets ensuring loadability limit. Int J Electr Power particle swarm optimisation algorithm for congestion management in
Energy Syst 2013;46:258–73. restructured power market by placement and sizing of unified power flow
[15] Yousefi A, Nguyen TT, Zareipour H, Malik OP. Congestion management using controller. IET Sci Meas Technol 2012;6(4):267–78.
demand response and FACTS devices. Int J Electr Power Energy Syst 2012;37 [24] Duman S, Sonmez Y, Guvenc U, Yorukeren N. Optimal reactive power dispatch
(1):78–85. using a gravitational search algorithm. IET Gen Trans Distrib 2012;6
[16] Preedavichit P, Srivastava SC. Optimal reactive power dispatch considering (6):563–76.
FACTS devices. Int J Electr Power Syst Res 1998;46(3):251–7. [25] Duman Serhat, Guvenc Ugur, Sonmez Yusuf, Yorukeren Nuran. Optimal power
[17] Gotham DJ, Heydt GT. Power flow control and power flow studies for system flow using gravitational search algorithm. Energy Conver Manage
with FACTS devices. IEEE Trans Power Syst 1998;13:60–5. 2012;59:86–95.
[18] Chung TS, Li YZ. A hybrid GA approach for OPF with consideration of FACTS [26] Nagalakshmi S, Kamaraj N. Comparison of computational intelligence
devices. IEEE Power Eng Rev 2000;August:54–7. algorithms for loadability enhancement of restructured power system with
[19] Gerbex S, Cherkaoui R, Germond AJ. Optimal location of multi-type FACTS FACTS devices. Swarm Evol Comput 2012;5(August):17–27.
devices in a power system by means of genetic algorithms. IEEE Trans Power [27] Phichaisawat S, Song YH, Wang XL, Wang XF. Combined active and reactive
Syst 2001;16(3):537–44. congestion management with FACTS devices. Electr Power Comp Syst 2002;30
[20] Cai LJ. Optimal choice and allocation of FACTS devices in deregulated (12):1195–205.
electricity market using genetic algorithms. In: IEEE PES, 2004. p. 56. [28] Vlachogiannis John G. FACTS applications in load flow studies effect on the
[21] Alhasawi FB, Milanovic JV. Techno-economic contribution of FACTS devices to steady state analysis of the Hellenic transmission system. Electr Power Syst
the operation of power systems with high level of wind power integration. Res 2000;55(September):179–89.
IEEE Trans Power Syst 2012;27(3):1414–21.