Electrical Power and Energy Systems 80 (2016) 240–251

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

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

Optimal location of STATCOM in multimachine power system

for increasing loadability by Cuckoo Search algorithm
S.M. Abd-Elazim, E.S. Ali ⇑
Electric Power and Machine Department, Faculty of Engineering, Zagazig University, Zagazig, Egypt

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

Article history: A new metaheuristic method, the Cuckoo Search (CS) algorithm, based on the life of a bird family is
Received 8 February 2014 proposed in this paper for optimal design of static synchronous compensator (STATCOM) in a multima-
Received in revised form 15 December 2015 chine environment. PV curves are illustrated to determine the best location of STATCOM. The STATCOM
Accepted 12 January 2016
parameter tuning problem is converted to an optimization problem which is solved by CS Algorithm. The
performance of the proposed CS based STATCOM (CSSTATCOM) is compared with Genetic Algorithm (GA)
based STATCOM (GASTATCOM) and open loop STATCOM under various operating conditions and distur-
Keywords:
bances. The superiority of the proposed technique in damping oscillations and enhancing voltage profiles
Cuckoo Search algorithm
Genetic Algorithm
is confirmed through eigenvalues and time domain simulation results over the GA and open loop one.
STATCOM Ó 2016 Elsevier Ltd. All rights reserved.
Multimachine system
Power system stability
Loadability

Introduction
The recent development of the high-power electronics presents
the use of Flexible AC Transmission Systems (FACTS) controllers in
power systems [1]. Subsequently, it has been demonstrated that
variable shunt compensation is highly effective in both controlling
power flow in the lines and hence the system voltage profiles and
stability [2,3]. Static synchronous Compensator (STATCOM) is a
member of FACTS family that is connected in shunt with the power
system [4]. By controlling the magnitude of the STATCOM voltage,
the reactive power exchanges between the STATCOM and the
transmission line and hence the amount of shunt compensation
in the power system can be controlled. In addition to reactive
power exchange, a properly controlled STATCOM can also provide
great damping to the power system oscillations [3,4].
Recently, Artificial Intelligence (AI) techniques have been dis-
cussed in literatures to solve problems related to STATCOM design.
Artificial Neural Network (ANN) for designing STATCOM is
addressed in [5–8]. The ANN approach has its own merits and
demerits. The performance of the system is improved by ANN
based controller, but the main problem of this controller is the long
training time, the selecting number of layers and the number of
⇑ Corresponding author at: Faculty of Engineering, Zagazig University, Zagazig
44519, Egypt. Tel.: +20 11 2669781, +20 55 2321407.
E-mail addresses: sahareldeep@yahoo.com (S.M. Abd-Elazim), ehabsalimalisalama@
yahoo.com (E.S. Ali).
neurons in each layer. Another AI approach likes Fuzzy Logic Con-
trol (FLC) has received much attention in control applications. In
contrast with the conventional techniques, FLC formulates the con-
trol action of a plant in terms of linguistic rules drawn from the
behavior of a human operator rather than in terms of an algorithm
synthesized from a model of the plant [9–11]. However, it can be
designed on the basis of linguistic information obtained from the
previous knowledge of the control system and gives better perfor-
mance results than the conventional controllers; a hard work is
inevitable to get the effective signals when designing FLC. Robust
techniques [12–14] have been also used for STATCOM design, but
these methods are iterative, sophisticated and the system uncer-
tainties should be carried out in a special format. On the other
hand, the order of the controllers is as high as that of the plant. This
gives rise to complex structure of such controllers and reduces
their applicability.
Global optimization techniques have been applied to STATCOM
design problem. Optimal design of STATCOM via Genetic Algorithm
(GA) is developed in [15–19] but, it requires a very long run time
depending on the size of the system under study. Also, it gives rise
to repeat revisiting of the same suboptimal solutions. STATCOM
parameters tuning using Particle Swarm Optimization (PSO) are
illustrated in [20–24], but it pains from the partial optimism.
Moreover, the algorithm cannot work out the problems of scatter-
ing and optimization. Furthermore, the algorithm suffers from
slow convergence in refined search stage, weak local search ability
and algorithm may lead to possible entrapment in local minimum

http://dx.doi.org/10.1016/j.ijepes.2016.01.023
0142-0615/Ó 2016 Elsevier Ltd. All rights reserved.
 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251 241

Nomenclature

E0q ; Efd the internal and the field voltages respectively Po ; Q o the base initial loading conditions
d; x the rotor angle and speed respectively k loading factor
V ref ; V t the reference and terminal voltage respectively Dwi the change in speed for generator (i)
Tm; Te the mechanical input and electrical torques respectively Pa the probability to abandon a nest
KA; TA the gain and time constant of excitation system respec- n number of nests
tively Ts the settling time
sj ; D the interia coefficient and damping coefficient respec- J objective function
tively
X d ; X 0d the d axis reactance of the generator and transient reac- List of abbreviations
tance respectively CS Cuckoo Search
Itd ; s0do the d axis generator current and time constant of exci- GA Genetic Algorithm
tation system respectively STATCOM static synchronous compensator
V DC ; IDC the STATCOM DC voltage and current respectively CSSTATCOM cuckoo based STATCOM
ILoq ; ILod the STATCOM output current in quadrature and direct GASTATCOM GA based STATCOM
axis respectively FACTS Flexible AC Transmission Systems
C DC the capacitance of STATCOM DC capacitor link ANN Artificial Neural Network
Kpdc ; Kidc the gains of PI controller for DC voltage regulator FLC Fuzzy Logic Control
respectively PSO Particle Swarm Optimization
Tc the washout time constant BF Bacteria Foraging
Kpac ; Kiac the gains of PI controller for AC voltage regulator ITAE The Integral of Time multiply Absolute Error
respectively ISE The Integral of Square Error
T c1 ; T w the washout time constants ITSE The Integral of Time multiply Squared Error
T 1  T 4 the time constants of lead lag circuits IAE The Integral of Absolute Error
K the gain of lead lag circuits
Ui ; V L the supplementary signal and voltage of the bus where
STATCOM is connected respectively

solutions. Artificial Bee Colony (ABC) is developed in [25] to design sj x_ ¼ T m  T e  Dx ð2Þ

STATCOM controller, but it is slow to converge and the processes of
the exploration and exploitation contradict with each other, so the 1
 ðX  X 0 Þ
E_ 0q ¼ 0 Efd  E0q þ
d
two abilities should be well balanced for achieving good optimiza- 0
d
Itd ð3Þ
sdo sdo
tion performance. A relatively newer evolutionary computation
algorithm, called Bacteria Foraging (BF) scheme has been estab-
1 KA  
lished recently by [26–32]. The BF algorithm depends on random E_ fd ¼  Efd þ V ref  V t ð4Þ
search directions which may lead to delay in reaching the global TA TA
solution.
In order to overcome these drawbacks, a Cuckoo Search (CS)
optimization algorithm is proposed in this paper for optimal design STATCOM dynamic model
of STATCOM parameters. The problem of a robust STATCOM design
is formulated as an objective optimization problem and CS algo- The power circuit of STATCOM is composed of a boosting trans-
rithm is used to handle it. The effectiveness of the proposed former, three phase GTO-based VSCs and a DC capacitor link
CSSTATCOM is tested on a multimachine power system under var- [34,35]. c, w are the amplitude modulation ratio and phase angle
ious operating conditions in comparison with GASTATCOM and of the control signal of each VSC respectively, which are the input
open loop STATCOM (without supplementary signal) through
eigenvalue and time domain analysis. Results evaluation show that
the proposed algorithm attains good robust performance for sup- 8 9
load C
pressing the low frequency oscillations under various operating 2
3
conditions and disturbances, and increasing the loadability of
system. ~ ~
6
Problem formulation
4 5
Power system model
load A
load B
Fig. 1 shows the single line diagram of the tested system [33].
7
The system and generator loading levels are given in Table 1. Each
generator is represented by the third order model and equipped
with static exciter (IEEE type STI). The electromechanical swing 1
equations, the generator internal voltage equation and the exciter
equation for one machine are given below: ~

d_ ¼ x  1 ð1Þ Fig. 1. The study system.

242 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251

Table 1
Loading conditions for the system (in p.u).

Light Normal case Heavy

P Q P Q P Q
Generator
G1 0.566 0.0752 0.9176 0.2082 2.1796 1.2427
G2 1.00 0.2503 1.630 0.0268 2.20 0.7397
G3 0.45 0.2949 0.85 0.1289 1.35 0.4456
Load
A 0.7 0.35 1.25 0.5 2.0 0.9
B 0.5 0.3 0.9 0.3 1.8 0.6
C 0.6 0.2 1.0 0.35 1.6 0.65

Fig. 4. Block diagram of AC voltage regulator with supplementary signal.

the AC voltage control signal of the STATCOM as shown in Fig. 4,

where the feedback signal for the supplementary controller is the
locally speed deviation [34–36]. The function of supplementary
signal is to counteract the negative damping effect brought by
the interaction of both AC and DC regulator. The AC voltage regu-
lator dynamic equations are given below:
 
Kiac ST c1  
c¼ Kpac þ V Lref  V L þ U i ð7Þ
S ð1 þ T c1 SÞ
  
1 þ ST 1 1 þ ST 3 ST w
VS ¼ K ð xÞ ð8Þ
1 þ ST 2 1 þ ST 4 ð1 þ T w SÞ

Hopf bifurcation and location of STATCOM

Fig. 2. Active and reactive power control of STATCOM.
Hopf bifurcation is characterized by periodic orbits emerging
around an equilibrium point and can be determined with the help
control signals to the STATCOM as shown in Fig. 2. The parameters of linearized analysis, at these bifurcation points the system jaco-
of STATCOM are given in Appendix. bian has a pair of purely imaginary eigenvalues as shown in
The DC voltage dynamic equation is given below: Fig. 5 [39,40].
PV curves of the system will be obtained for a specific load and
c  
V_ DC ¼ ILoq cos w þ ILod sin w ð5Þ generation direction by increasing the active and reactive power in
C DC
load as follows [41,42].
P ¼ Po ð1 þ kÞ; Q ¼ Q o ð1 þ kÞ ð9Þ
DC voltage regulator
The DC voltage regulator controls the DC voltage across the DC where Po and Q o correspond to the base initial loading conditions
capacitor. Fig. 3 shows the dynamic model of the DC voltage regu- and the parameter k, which corresponds to the loading factor, rep-
lator, which adopts PI controller. The DC voltage regulator resents the slow varying parameter used in bifurcation studies. The
exchanges the active power between the STATCOM and the power load is considered as constant impedance load.
system [36–38]. As w increasing, more active power is sent to the In order to determine the most suitable placement of the STAT-
power system from the STATCOM. The DC voltage regulator COM in the system, PV curves are studied to find the most critical
dynamic equation is given below: bus for the greatest favorable impact on transmission system volt-
  ages. This placement technique results in optimal STATCOM con-
Kidc ST c   troller location to enhance voltage profiles, and to improve at the
w ¼ Kpdc þ V DCref  V DC ð6Þ
S ð1 þ T c SÞ
Imaginary

AC voltage regulator with supplementary signal

The AC voltage regulator controls the reactive power exchange λ •
with the power system. A supplementary signal can be imposed on

Hopf
Re
Bifurcation

λ •

Fig. 3. Dynamic model of DC voltage regulator. Fig. 5. Locus of the eigenvalues at Hopf bifurcation.
 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251 243

Voltage magnitude in p.u.

Fig. 6. PV curves for the load buses.

same time stability margins by increasing the distance to a voltage The aim of the optimization is to search for the optimal set of STAT-
collapse point. Fig. 6 shows the PV curves of buses 4, 5 and 6 when COM parameters setting via CS that reflect the settling time and
loading these buses individually to their maximum loadability overshoots of the system. Furthermore, the aims are enhancing
points until divergence. Bus 4 is the most sensitive bus due to the damping characteristics, acquiring a good performance under
the lowest maximum loadability point. Moreover, it is the lower different operating conditions and also improving the voltage
voltage profile. Finally, this location agrees with that obtained in profiles of system.
[29,30].
Overview of Cuckoo Search algorithm
Objective function
The breeding behavior of cuckoo species and the basic items of
A performance index can be defined by the Integral of Time the proposed algorithm are discussed below.
multiply Absolute Error (ITAE) of the speed deviation of each gen-
erator and DC voltage link. The merit of this chosen performance Cuckoo breeding behavior
index is that minimal dynamic plant information is required. Other
indices, the Integral of Square Error (ISE) and Integral of Time mul- CS is a metaheuristic search algorithm which has been intro-
tiply Squared Error (ITSE) are very offensive criterions because of duced recently by Yang [45]. The algorithm is inspired by obligate
squaring the error creates unrealistic evaluation. Also, the Integral brood parasitism of some cuckoo species by laying their eggs in the
of Absolute Error (IAE) is unqualified compared with the ITAE nests of other host birds, which may be of different species. At the
which illustrates more realistic error index [43,44]. appropriate moment, the hen cuckoo flies down to the host’s nest,
Accordingly, the objective function J is set to be pushes one egg out of the nest, lays an egg and flies off. The whole
Z t sim
process takes about 10 s. A female may visit up to 50 nests during a
J¼ t fjðDx1 þ Dx2 þ Dx3 Þj þ jDV DC jgdt ð10Þ breeding season. The host birds may detect that the eggs are not
0 their own and either throw them away or abandon the nest and
The values of the washout time constants T W , T c and T c1 are kept at build new one elsewhere. This has resulted in the evolution of
10, 8 and 8 s respectively. The values of time constants T 2 and T 4 are cuckoo eggs which mimic the eggs of local host birds [46]. More-
fixed at a reasonable value of 0.05 s. Typical ranges of the optimized over, the timing of egg lying of some species is also amazing. Par-
parameters are [1–100] for K; Kpdc ; Kidc ; Kpac ; Kiac , and [0.06–1.0] asitic cuckoos often select a nest where the host bird just laid its
for T 1 and T 3 . Optimization problem based on the objective function own eggs. In general, the cuckoo eggs hatch slightly earlier than
J can be stated as: minimize J subjected to: their host eggs. It methodically dislodges all host progeny from
host nests. It is a much larger bird than its hosts, and needs to
K min 6 K 6 K max monopolize the food supplied by the parents. The chick will roll
the other eggs out of the nest by pushing them with its back over
T min
1 6 T 1 6 T max
1
the edge. If the host’s eggs hatch before the cuckoos, the cuckoo
T min
3 6 T 3 6 T max
3 chick will push the other chicks out of the nest in a similar way.
Kpmin max
ac 6 Kpac 6 Kpac
min max Lévy flights
Kiac 6 Kiac 6 Kiac
Kpmin max
dc 6 Kpdc 6 Kpdc The use of Lévy flights for both local and global searching is an
min max important component of the CS [47]. The Lévy flight process is a
Kidc 6 Kidc 6 Kidc ð11Þ
random walk that is characterized by a series of instantaneous
244 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251

Table 3
Start Parameters of STATCOM for different algorithms.

CS GA
Initialization of CS parameters
DC regulator Kpdc 18.0959 12.0959
Kidc 23.4575 49.4575
Initial generation of host nest population AC regulator Kpac 55.3694 65.3694
Kiac 25.9627 35.9627
Supplementary signal K 49.4526 42.3005
Evaluation of fitness function for the generated
T1 0.6875 0.2475
host nest population
T3 0.4168 0.6801
Move all cuckoos
Modification of nest position using Levy
flight equation towards better
environment jumps selected from a probability density function which has a
Evaluate the new solution power law tail. This process represents the optimum random
search pattern and is frequently found in nature [48]. When gen-
erating a new egg, a Lévy flight is done starting at the location of
No a randomly chosen egg, if the objective function value at these
Check the condition is
satisfied? new coordinates is better than another randomly chosen egg then
that egg is shifted to this new location. One of the advantages of
Yes CS over other optimization algorithm is that only one parameter,
Print the result the fraction of nests to abandon P a , needs to be adjusted. The use
of Lévy flights as the search method means that the CS can simul-
taneously find all optima in a design space and the method has
Stop been shown to perform well in comparison with other algorithms
[49].
Fig. 7. Flow chart of cuckoos optimization algorithm.

-4
x 10
1.8
CS
GA
1.6

1.4
Change of objective function

1.2

0.8

0.6

0.4

0.2

0
0 10 20 30 40 50 60 70 80 90 100
Generations

Fig. 8. Change of objective function.

Table 2
Mechanical modes and f under different loading conditions and controllers.

Open loop GASTATCOM CSSTATCOM

Light load 1.15 ± 6.88j, 0.16 1.43 ± 6.53j, 0.21 1.13 ± 6.92j, 0.16
0.034 ± 6.49j, 0.005 1.09 ± 6.92j, 0.155 2.16 ± 6.05j, 0.34
0.067 ± 5.72j, 0.011 1.4 ± 5.6j, 0.243 1.51 ± 5.68j, 0.26
Normal load 1.3 ± 9.13j, 0.14 1.35 ± 9.07j, 0.147 1.43 ± 9.1j, 0.155
0.026 ± 7.9j, 0.003 1.28 ± 7.9j, 0.16 1.69 ± 7.7j, 0.21
0.019 ± 5.79j, 0.003 0.91 ± 5.77j, 0.156 1.7 ± 5.7j, 0.29
Heavy load 0.59 ± 10.88j, 0.05 1.06 ± 10.81j, 0.1 1.16 ± 10.56j, 0.11
0.024 ± 8.4j, 0.003 0.92 ± 8.56j, 0.11 0.99 ± 8.25j, 0.12
0.015 ± 5.76j, 0.003 1.14 ± 5.6j, 0.2 1.37 ± 5.27j, 0.25
 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251 245

Cuckoo Search implementation
Each egg in a nest performs a solution and a cuckoo egg intro-
duces a new solution. The target is to use the new and potentially
better solutions to replace a not so good solution in the nests. Each
nest has one egg for simple case. The algorithm can be developed
to more sophisticated cases in which each nest has multiple eggs
representing a set of solutions. To apply this search as an optimiza-
tion algorithm, three approximation rules are presented below
[49,50]:
(1) Cuckoos select random nest for laying their eggs. Artificial
cuckoo can lay only one egg at the time.
(2) Elitist selection process is applied, so only the eggs with
highest quality are passed to the next generation.
(3) The number of available hosts nests is specified and a host
can detect a foreigner egg with a probability Pa 2 ½0; 1. If
cuckoo egg is discovered by the host, it may be thrown away,
or the host may abandon its own nest and commit it to the
cuckoo intruder.
The last assumption can be approximation by a factor P a of the n
nests being exchanged by new nests. The quality of a solution is
relative to the objective function for a maximization problem.
Other forms of fitness can be determined in an identical way to
the fitness function in GA. Based on these three rules, the flow
chart of CS is shown in Fig. 7. The parameters of CS are shown in
Appendix.
When generating a new solution xtþ1
i for the ith cuckoo, a Lévy
flights is carried out as follows:

x 10-4
CSSTATCOM
GASTATCOM
3 Open loop

2
Change in w12 (rad/second)

-1

-2

-3

-4

-5
0 2 4 6 8 10 12
Time in second

Fig. 9. Change of Dx12 under normal load condition.

x 10-4
3
CSSTATCOM
GASTATCOM
Open loop
2
Change in w13 (rad/second)

-1

-2

-3
0 2 4 6 8 10 12
Time in second

Fig. 10. Change of Dx13 under normal load condition.

246 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251

x 10-4
4
CSSTATCOM
GASTATCOM
Open loop
3

Change in w23 (rad/second)

2

-1

-2

-3
0 1 2 3 4 5 6 7 8 9 10
Time in second

Fig. 11. Change of Dx23 under normal load condition.

x 10-4
4
CSSTATCOM
GASTATCOM
3
Open loop

1
Change in w12 (rad/second)

-1

-2

-3

-4

-5

-6
0 2 4 6 8 10 12
Time in second

Fig. 12. Change of Dx12 under light load condition.

0
xitþ1 ¼ xti þ a  Levy ðkÞ
where a > 0 is the step size which should be proportional to the
scales of the optimization problem. The product  means entrywise
walk during multiplications. Lévy flights essentially provide a ran-
dom walk while their random steps are drawn from a Lévy distribu-
tion for large steps.
0
k
L e vy  u ¼ t ; ð1 < k 6 3Þ
Which has an infinite variance with an infinite mean. Here the steps
essentially form a random walk process with a power law step-
length distribution with a heavy tail. Some of the new solutions
should be generated by Lévy walk around the best solution obtained
ð12Þ so far, this will speed up the local search. However, a substantial
fraction of the new solutions should be generated by far field ran-
domization and whose locations should be far enough from the cur-
rent best solution, this will make sure the system will not be
trapped in a local optimum.
Results and simulations
ð13Þ
In this section, the superiority of the proposed CS algorithm in
designing STATCOM compared with optimized STATCOM with
GA and open loop STATCOM (DC and AC regulators) is illustrated.
Fig. 8 shows the change of objective functions with two optimiza-
tion algorithms. The objective functions decrease over iterations of
 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251 247

-4
x 10
2
CSSTATCOM
GASTATCOM
1.5 Open loop

Change in w13 (rad /second) 0.5

-0.5

-1

-1.5

-2
0 2 4 6 8 10 12
Time in second

Fig. 13. Change of Dx13 under light load condition.

-4
x 10
5
CSSTATCOM
GASTATCOM
4 Open loop

3
Change in w23 (rad/second)

-1

-2

-3

-4
0 2 4 6 8 10 12
Time in second

Fig. 14. Change of Dx23 under light load condition.

CS and GA. Moreover, CS converges at a faster rate (53 generations) GASTATCOM and open loop, CSSTATCOM provides good robust
compared with that for GA (72 generations). Furthermore, compu- performance and achieves superior damping characteristics of
tational time (CPU) of both algorithms is compared based on the electromechanical modes. Results of CSSTATCOM parameters set
average CPU time taken to converge the solution. The average values based on the proposed objective function using CS and GA
CPU for CS is 36.4 s while it is 48.6 s for GA. The proposed CS are given in Table 3.
methodology and GA are programmed in MATLAB 7.1. The men-
tioned CPU time is the average of 10 executions of the computer Response under normal load condition
code.
Table 2 shows the system eigenvalues and damping ratios of The validation of the system performance due to 20% increase of
mechanical modes with three different loading conditions. It is mechanical torque for generator 1 as a small disturbance is veri-
clear that, the CSSTATCOM shift the electromechanical modes to fied. Figs. 9–11 show the response of Dx12 ; Dx13 and Dx23 due
the left of the S-plane and the values of the damping factors with to this disturbance under normal loading condition. It can be seen
the proposed CSSTATCOM are significantly enhanced for different that the system with the proposed CSSTATCOM is more stabilized
loading conditions. Also, the damping ratios corresponding to than GASTATCOM and open loop. In addition, the required mean
CSSTATCOM controllers are larger than those corresponding to settling time to mitigate system oscillations is approximately
GASTATCOM and open loop STATCOM. Hence, compared with 1.05 s with CSSTATCOM, 1.6 s for GASTATCOM. On the other hand,
248 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251

-4
x 10
5
CSSTATCOM
GASTATCOM
4
Open loop

Change in w12 (rad/second) 2

-1

-2

-3

-4

-5
0 2 4 6 8 10 12
Time in second

Fig. 15. Change of Dx12 under heavy load condition.

x 10-4
5
CSSTATCOM
GASATACOM
4
Open loop

3
Change in w13 (rad/second)

-1

-2

-3

-4

-5
0 2 4 6 8 10 12 14 16 18 20
Time in second

Fig. 16. Change of Dx13 under heavy load condition.

the system with open loop STATCOM still has oscillations due to
the interaction between AC and DC regulators so the designed con-
troller is qualified for supplying adequate damping to the low fre-
quency oscillations.
Response under light load condition
Figs. 12–14 show the system response under light loading con-
dition with fixing the controller parameters. It is clear from these
figures, that the proposed CSSTATCOM has good damping charac-
teristics to system oscillatory modes and stabilizes the system
rapidly. Also, the mean settling time of oscillations is T s = 1.46
and 2.6 s for CSSTATCOM and GASTATCOM respectively. In addi-
tion, the system with open loop STATCOM cannot reach steady
state value till 12 s. Hence, the proposed CSSTATCOM outlasts
GASTATCOM and open loop controller in attenuating oscillations
effectively and minifying settling time. Consequently, the proposed
CSSTATCOM extend the power system stability limit.
Response under heavy load condition
Figs. 15–17 show the system response under heavy loading con-
dition. These figures indicate the superiority of the CSSTATCOM in
reducing the settling time and suppressing the power system oscil-
lations. Moreover, the mean settling time of these oscillation is
T s = 1.0 and 2.43 s for CSSTATCOM and GASTATCOM respectively.
Also, the system without supplementary signal is suffered from sus-
tained oscillations. Hence, CSSTATCOM controller greatly improves
the system stability and enhances the damping characteristics of
power system. Furthermore, the settling time of the proposed con-
troller is smaller than that in [35,41].
 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251 249

-4
x 10
4
CSSTATCOM
GASTATCOM
3 Open loop

Change in w23 (rad/second)

1

-1

-2

-3

-4
0 2 4 6 8 10 12 14 16 18 20

Time in second

Fig. 17. Change of Dx23 under heavy load condition.

-3
x 10
4

CSSTATCOM

3 GASTATCOM
Open loop

2
Change in w12 (rad/second)

-1

-2

-3

-4
0 2 4 6 8 10 12 14 16 18 20
Time in second

Fig. 18. Change of Dx12 under severe disturbance.

Response under severe disturbance imately 0.2373 p.u., a Hopf bifurcation condition is detected. Based
on eigenvalue analysis, one can load the system until loading factor
The validation of the performance under severe disturbance is of 0.2373 p.u. which represents the stable equilibrium points. The
confirmed by applying a three phase fault of 6 cycle duration at system is unstable after this value of loading until the nose point.
1.0 s near bus 8. Fig. 18 shows the response of Dx12 under normal The maximum loading point corresponds to the saddle node bifur-
loading condition. From this figure, it is can be seen that the CS cation point, where a real eigenvalue becomes zero is at loading
based STATCOM using the proposed objective function introduces factor of approximately 2.2542 p.u. On the other hand, Fig. 20
superior damping and attains better robust performance in com- shows the PV curves of load buses when loading bus 4 only based
parison with the other methods. on its initial values with insertion of the proposed CSSTATCOM on
the same bus. These results show that the Hopf bifurcation has
Loadability enhancement been removed to the maximum loadability point in spite of
increasing the maximum loadability point by 210%. The voltage
Fig. 19 shows the corresponding PV curves of load buses for profiles are enhanced since one can load the system near the nose
loading bus 4 based on its initial value. At loading factor of approx- point while it is still stable.
250 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251

1.05

0.95

0.9
Bifurcation Point

0.85
Voltage (p.u)

0.8

0.75

V4
0.7
V5

0.65 V6
V7

0.6 V8

V9
0.55
0 0.5 1 1.5 2
Loading (p.u)

Fig. 19. PV curves without STATCOM.

1.05

0.95
Voltage (p.u)

0.9

0.85
V4
V5
V6 Bifuraction point
0.8 V7
V8
V9
0.75
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Loading (p.u)

Fig. 20. Effect of STATCOM on PV curves.

Conclusions Appendix

A new optimization algorithm known as CS algorithm, for opti-
mal setting of STATCOM parameters is proposed in this paper. PV
curves are investigated to find the optimal location of controller.
The STATCOM parameters tuning problem are formulated as an
optimization problem and CS algorithm is employed to seek for
optimal parameters. A time domain objective function involving
the change of synchronous speed of generator and DC voltage is
proposed to alleviate power system oscillations and enhance sys-
tem performance in terms of the settling time and overshoots. Sim-
ulation results confirm the robustness and superiority of the
proposed CSSTATCOM in providing good damping characteristic
to system oscillations over a wide range of loading conditions com-
pared with GASTATCOM and open loop one. Finally, proper selec-
tion of STATCOM location can effectively improve the overall
voltage profiles and system loadability.
(a) DC link parameters (p.u): V DC = 1.25; C DC = 1.0.
(b) STATCOM w and c parameters have been recalculated, at
each loading conditions.
(c) The parameters of CS algorithm are as follows: Max genera-
tion = 100; Number of nests = 50; Pa = 0.25.
(d) The parameters of GA are as follows: Max generation = 100;
Population size = 50; Crossover probabilities = 0.75; Muta-
tion probabilities = 0.1.
References
[1] Song YH, Johns AT. Flexible AC transmission systems (FACTS). United Kingdom,
London: IEE; 1999.
[2] Arrillaga J, Watson NR. FACTS and HVDC transmission, in computer modelling
of electrical power systems. 2nd ed. West Sussex, England: John Wiley & Sons,
Ltd; 2001.
 S.M. Abd-Elazim, E.S. Ali / Electrical Power and Energy Systems 80 (2016) 240–251
[3] Sood VK. HVDC and FACTS controllers: applications of static converters in
power system. Springer; 2004. edition.
[4] Zhang XP, Rehtanz C, Pal B. Flexible AC transmission systems: modeling and
control. Berlin Heidelberg: Springer-Verlag; 2012.
[5] Sumathi S, Bansilal S. Artificial Neural Network application for control of
STATCOM in power systems for both voltage control mode and reactive power
mode. In: Int. conference on power systems. ICPS ’09, Kharagpur; 2009. p. 1–4.
[6] Balavar M. Using neural network to control STATCOM for improving transient
stability. J Artif Intell Electr Eng 2012;1:26–31.
[7] Dhal PK, Ranjan CCA. Analysis of transient stability based STATCOM for neural
network controller in cascaded multilevel inverter. In: 4th International
conference, SEMCCO 2013, Chennai, India; December 19–21, 2013. p. 342–53.
[8] Sumathi S, Bansilal S. Artificial Neural Network for coordinated control of
STATCOM, generator excitation and tap changing transformer. Electr Power
Energy Syst 2015;64:536–41.
[9] Mak LO, Ni YX, Shen CM. STATCOM with fuzzy controllers for interconnected
power systems. Electr Power Syst Res 2000;55:87–95.
[10] Kumkratug P. STATCOM control strategy based on lyapunov energy function
and fuzzy logic control for improving transient stability of multimachine
power system. WSEAS Trans Circuits Syst 2012;11(5):159–68.
[11] Awasthi A, Gupta SK, Panda MK. Design of a fuzzy logic controller based
STATCOM for IEEE9 bus system. Eur J Adv Eng Technol 2015;2(4):62–7.
[12] Latha S, Vikhram GYR. Power system damping improvement by robust
STATCOM voltage controller design using loop-shaping technique. In: Int.
conference on process automation, control and computing (PACC).
Coimbatore; 2011. p. 1–6.
[13] Gupta S, Tripathi RK. An LQR and pole placement controller for CSC based
STATCOM. In: Proceedings of the Int. Conference on Power, Energy and Control
(ICPEC ’13); February 2013. p. 115–19.
[14] Safari A, Shayeghi H, Shayanfar HA. Robust state feedback controller design of
STATCOM using chaotic optimization algorithm. Serb J Electr Eng 2010;7
(2):253–68.
[15] Eshtehardiha S, Poodeh MB, Kiyoumarsi A. Optimized performance of
STATCOM with PID controller based on genetic algorithm. In: Int. conference
on control, automation and systems. ICCAS ’07, Seoul; 2007. p. 1639–44.
[16] Panda S, Patidar NP, Singh R. Robust coordinated design of excitation and
STATCOM-based controller using genetic algorithm. Int J Innovative Comput
Appl 2008;1(4):244–51.
[17] Mandour ME, Abd-Elazeem SM. Robust design of STATCOM controller for
improving power system damping performance via GA. In: Al_AZHAR
engineering tenth international conference, vol. 3, No. 12; December 2008.
p. 159–69.
[18] Golkar MA, Zarringhalami M. Coordinated design of PSS and STATCOM
parameters for power system stability improvement using genetic
algorithm. Iran J Electr Comput Eng 2009;8(2):80–8.
[19] To’aima FM, Al-Aani YN, Salbi HAA. Optimal location of static synchronous
compensator (STATCOM) for IEEE 5-bus standard system using genetic
algorithm. J Eng 2015;21(July):72–84.
[20] Azadani EN, Hosseinian SH, Hasanpor P. Optimal placement of multiple
STATCOM for voltage stability margin enhancement using particle swarm
optimization. Electr Eng 2008;90(7):503–10.
[21] Kumarasamy K, Raghavan R. Particle swarm optimization algorithm for
voltage stability improvement using multiple STATCOM. In: Int. conf. on
emerging trends in electrical engineering and energy management
(ICETEEEM), Chennai; 2012. p. 235–42.
[22] Sagarika A, Jyothsna TR. Tuning of PSO algorithm for single-machine and
multi-machine power system using STATCOM controller. Int Res J Eng Technol
(IRJET) 2015;2(4):175–82.
[23] Ravi K, Rajaram M. Optimal location of FACTS devices using improved particle
swarm optimization. Int J Electr Power Energy Syst 2013;49:333–8.
[24] Rocha L, Castro R, de Jesus JMF. An improved particle swarm optimization
algorithm for optimal placement and sizing of STATCOM. Int Trans Electr
Energy Syst 2015. http://dx.doi.org/10.1002/etep.2110.
[25] Safari A, Ahmadian A, Golkar MAA. Controller design of STATCOM for power
system stability improvement using honey bee mating optimization. J Appl
Res Technol 2013;11(1):144–55.
251
[26] Ali ES, Abd-Elazim SM. Coordinated design of PSSs and TCSC via bacterial
swarm optimization algorithm in a multimachine power system. Int J Electr
Power Energy Syst 2012;36(1):84–92.
[27] Ali ES, Abd-Elazim SM. TCSC damping controller design based on bacteria
foraging optimization algorithm for a multimachine power system. Int J Electr
Power Energy Syst 2012;37(1):23–30.
[28] Abd-Elazim SM, Ali ES. Coordinated design of PSSs and SVC via bacteria
foraging optimization algorithm in a multimachine power system. Int J Electr
Power Energy Syst 2012;41(1):44–53.
[29] Abd-Elazim SM, Ali ES. Bacteria foraging optimization algorithm based SVC
damping controller design for power system stability enhancement. Int J Electr
Power Energy Syst 2012;43(1):933–40.
[30] Abd-Elazim SM, Ali ES. A hybrid particle swarm optimization and bacterial
foraging for power system stability enhancement. J Complex 2015;21
(2):245–55.
[31] Ali ES, Abd-Elazim SM. Optimal SSSC design for power systems via hybrid
approach. J Electr Eng (JEE) 2014;14(1):138–47.
[32] Abd-Elazim SM, Ali ES. New coordinated design of PSSs and SVC to improve
power system stability via BSO. J Electr Eng (JEE) 2015;15(1):21–30.
[33] Anderson PM, Fouad AA. Power system control and stability. Iowa: Iowa State
University Press; 1977.
[34] Wang HF. Philips-Heffron model of power systems installed with STATCOM
and applications. IEE Proc Gener, Transm Distrib 1999;146(5):521–7.
[35] Abido MA, Bamasak SM. Robust coordinated design of PSS & STATCOM
controllers for damping power system oscillation. In: 15th PSCC, Liege; August
2005. p. 22–26.
[36] Al-Jowder F. Improvement of synchronizing power and damping power by
means of SSSC and STATCOM, a comparative study. Int J Electr Power Syst Res
2007;77(8):1112–7.
[37] Rao P, Crow ML, Young Z. STATCOM control for power system voltage control
applications. IEEE Trans Power Delivery 2000;15(4):1311–7.
[38] Sensarma PS, Padiyar KR, Ram-narayanan V. Analysis and performance
evaluation of a distribution STATCOM for compensating voltage fluctuation.
IEEE Trans Power Delivery 2001;16(2):259–64.
[39] Laufenberg MJ, Pai MA, Padiyar KR. Hopf bifurcation control in power systems
with static VAR compensators. Int J Electr Power Eng Syst 1997;19(5):339–47.
[40] Mithulananthan N, Canizares CA, Reeve J. Indices to detect Hopf Bifurcations in
power systems. In: North American power system symposium, Waterloo;
October 2000. p. 1–7.
[41] Mandour ME, Abd-Elazeem SM. Hopf Bifurcation improvement by STATCOM
controller employing genetic algorithm approach. In: 44th int. universities
power engineering conference UPEC, 1-4 September 2009, University of
Strathclyde, Glasgow, Scotland.
[42] Cañizares CA, Hranilovic S. Transcritical and Hopf Bifurcations in AC/DC
systems. In: Proceedings of bulk power system voltage phenomena III- voltage
stability and security. Fairfax, VA; August 1994. p. 105–14.
[43] Ali ES, Abd-Elazim SM. Power system stability enhancement via bacteria
foraging optimization algorithm. Int Arab J Sci Eng (AJSE) 2013;38
(3):599–611.
[44] Ali ES. Firefly algorithm based speed control of DC series motor. Int J WSEAS
Trans Syst Control 2015;10:137–47.
[45] Yang XS. Engineering optimization: an introduction with metaheuristic
applications. John Wiley and Sons; 2010.
[46] Yang XS, Deb S. Engineering optimization by Cuckoo Search. Int J Math Model
Numer Optim 2010;1(4):330–43.
[47] Yang XS, Deb S. Cuckoo Search via Lévy Flights. In: Proceedings of World
congress on nature & biologically inspired computing (NaBIC 2009, India), IEEE
Publications, USA; 2009. p. 210–14.
[48] Abdelaziz AY, Ali ES. Load frequency controller design via artificial Cuckoo
Search algorithm. Electr Power Compon Syst 2015. http://dx.doi.org/10.1080/
15325008.2015.1090502.
[49] Abd-Elazim SM, Ali ES. Optimal power system stabilizers design via Cuckoo
Search algorithm. Int J Electr Power Energy Syst 2016;75 C(February):99–107.
[50] Yang XS, Deb S. Multiobjective Cuckoo Search for design optimization. Comput
Oper Res 2013;40:1616–24.