Electrical Power and Energy Systems 78 (2016) 801815

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

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

A novel method based on adaptive cuckoo search for optimal network

reconfiguration and distributed generation allocation in distribution
network
Thuan Thanh Nguyen a,b,, Anh Viet Truong a, Tuan Anh Phung c
a
b
c

Faculty of Electrical and Electronics Engineering, HCMC University of Technology and Education, 1 Vo Van Ngan Str., Thu Duc Dist., Ho Chi Minh City, Viet Nam
Dong An Polytechnic, 30/4 Str., Di An Dist., Binh Duong Province, Viet Nam
Ha Noi University of Technology, 1 Dai Co Viet Str., Ha Noi, Viet Nam

a r t i c l e

i n f o

Article history:
Received 4 September 2015
Received in revised form 3 December 2015
Accepted 17 December 2015

Keywords:
Distribution network reconfiguration
Distributed generator
Adaptive cuckoo search algorithm
Power loss reduction
Voltage stability enhancement

a b s t r a c t
This paper proposes a new methodology to optimize network topology and placement of distributed generation (DG) in distribution network with an objective of reduction real power loss and voltage stability
enhancement. A meta-heuristic cuckoo search algorithm (CSA) inspired from the obligate brood parasitism of some cuckoo species which lay their eggs in the nests of other birds of other species for solving
optimization problems is adapted to simultaneously reconfigure and identify the optimal location and
size of DG units in a distribution network. The graph theory is used to determine the search space which
reduces infeasible network configurations of reconfiguration process and check the radial constraint of
each configuration of distribution network. The effectiveness of the proposed method has been validated
on three different distribution network systems at seven different scenarios. The obtained results show
well the effectiveness and performance of the proposed method in distribution network reconfiguration
with optimal location and size of DG problems.
2015 Elsevier Ltd. All rights reserved.

Introduction
A distribution network is the last stage in delivery of electric
power. It carries electricity from the transmission system to consumers. Distribution systems usually have high system losses
and poor voltage regulation because of the high current and low
voltage level in distribution systems [1,2]. In addition, due to the
rapid expansion of distribution networks, the voltage stability of
distribution systems has become an important issue. Therefore,
many efforts have been made to decrease the losses and improve
the voltage stability in distribution systems. Network reconfiguration and distributed generator placement are among those efforts
to mitigate this problem [3].
Distribution network reconfiguration (DNR) is the process of
varying the topology of distribution network by changing the
closed/open status of sectionalizing and tie switches while respecting system constraints upon satisfying the operators objectives
[4]. The first publication about the DNR problem was presented
by Merlin and Back [5]. They solved DNR problem through a
Corresponding author at: Dong An Polytechnic, 30/4 Str., Di An Dist., Binh
Duong Province, Viet Nam. Tel.: +84 0916664414.
E-mail address: thuan.dap@gmail.com (T.T. Nguyen).
http://dx.doi.org/10.1016/j.ijepes.2015.12.030
0142-0615/ 2015 Elsevier Ltd. All rights reserved.

discrete branch-and-bound type heuristic technique. Civanlar

et al. [6] proposed a switch exchange method to estimate the loss
reduction based on particular switching option. Since the method
is based on heuristics technique, it is difficult to take a systematic
way to evaluate an optimal solution. In recent years, new metaheuristic methods have been proposed for solving optimization
problems to obtain an optimal solution of global minimum in the
literature with good results. In [7], a method based on an enhanced
genetic algorithm was developed for DNR problem to minimize the
power loss and maximize the system reliability. Souza et al. [8]
proposed two new approaches for solving the DNR problem using
the Opt-aiNet (artificial immune network for optimization) and
Copt-aiNet (artificial immune network for combinatorial optimization) algorithms to minimize power loss. In [9], the network reconfiguration and capacitor placement are simultaneously employed
to enhance the system efficiency in a fuzzy multi-objective optimization problem by using a binary gravitational search algorithm
(BGSA).
Distributed generations (DGs), which are connected to the grid
at distributed level voltages are generating plant serving a customer on-site. Because of the reasons of energy security and economical benefit, the presence of DGs into distribution networks
has been increasing rapidly [10,11]. Impact of DG units on power

802

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Nomenclature
Prec
loss
P0loss
Nbr
Nbus
NDG
Pi
Qi
Vi

total power loss of the system after reconfiguration

total power loss of the system before reconfiguration
total number of branches
total number of buses
number of distributed generators
real power load at bus i
reactive power load at bus i
voltage magnitude at bus i

system has attracted the interest of some recent research efforts.

In [12], authors proposed comparison of novel power loss sensitivity (NPLS), power stability index (PSI), and voltage stability
index (VSI) methods for optimal allocation and size of DG in
radial distribution network. In [13], a method based on bacterial
foraging optimization algorithm (BFOA) is proposed to find the
optimal location and size of DG with an objective of power losses
reduction, operational costs and improving voltage stability. In
[14], authors proposed a method based on the artificial neural
network to find optimal DG size and locations due to complexity
of multiple DG concepts. Kayal and Chanda [15] proposed a new
constrained multi-objective particle swarm optimization (PSO)
based wind turbine generation unit and photovoltaic array placement approach to reduce power loss and improve voltage stability of radial distribution system. In [16], a novel application of
multi-objective particle swarm optimization was developed for
determining the place and size of DGs, and the contract price of
their generated power.
Recently, some researches have integrated both the DNR and
DG placement problems to improve the effectively of distribution
network [1719]. In [18], the DNR problem in the presence of DG
with an objective of minimizing real power loss and enhancing
voltage profile in distribution network is solved based on the harmony search algorithm (HSA). In [19], a method based on fireworks
optimization algorithm (FWA) is proposed for solving DNR
together with DG placement to minimize power loss and improve
voltage stability. Both researches [18,19] had used the different
techniques to pre-identify the candidate bus locations for DG
installation such as loss sensitivity factor (LSF) [18], and voltage
stability index (VSI) [19].
Due to pre-identifying of location of DGs based on LSF or VSI in
initial network configuration, the above methods focused only on
sizing of DG units. However, these parameters may change during
network reconfiguration process and DGs installation. In addition,
in distribution systems with multi-DG units, these parameters may
change more noticeable because of the interaction between DGs. In
this paper, a method based on cuckoo search algorithm (CSA) [20]
which is a recent meta-heuristic is proposed for solving the DNR
problem in the presence of distributed generation. Compared to
other algorithms, CSA has fewer control parameters and is more
effective [4]. Recently, CSA has been applied to solve many power
system problems and other fields such as optimal power flow
(OPF) [21], power system stabilizers (PSSs) [22], load frequency
control (LFC) [23], and automatic generation control (AGC) [24].
The results obtained from the above problems have proven the
effectiveness of CSA compared to other optimization algorithms.
In this study, the proposed method based on CSA uses power
loss and VSI index as objective functions to find the optimum configuration of distribution network, and the optimum bus location
and size of DGs. The algorithm is tested on 33-bus, 69-bus and
119-bus systems and results obtained are compared with other
techniques available in the literature.

PDgmax,i
Vmin
Vmax
Ii
Imax,i
Rk
Xk

maximum size of distributed generator ith

minimum acceptable bus voltage
maximum acceptable bus voltage
current at branch ith
upper limit of line current as defined by the manufacturer
resistance of the line section between buses k and k + 1
reactance of the line section between buses k and k + 1

Problem formulation
Objective functions
One of the main advantages of the optimal network reconfiguration and DG installation is the reduction in power loss. The net
power loss reduced (DP Rloss ) is taken as the ratio of total power loss
before and after the reconfiguration considering DGs of the system:

DPRloss

Prec:
loss

P0loss

The total power loss of the system is determined by the summation of losses in all line sections:

Ploss

Nbr
X
P2i Q 2i

Ri 
V 2i
i1

On the other hand, as DGs are installed in distribution network,

the bus voltages will increase and voltage security will enhance.
Therefore, to obtain maximum benefit from the DG, suitable location and sizing have to be determined before its installation based
on voltage stability index (VSI). VSI is a parameter that identifies
the near collapse nodes. The node with small VSI is more sensitive
to voltage collapse. According to Fig. 1, VSI of node 2th to node Nth
is calculated as follows [25,16]:

VSIk1 jV k j4  4Pk1 X k  Q k1 Rk 2
 4Pk1 Rk  Q k1 X k jV k j2

where Pi+1, Qi+1 are total real power and reactive power load fed
through node (k + 1), respectively.
If the VSI for each bus is higher, the stability of that relevant
node shall be better. In distribution network reconfiguration considering DGs, the voltage stability deviation index (DVSI) can be
defined as follows:

DVSI max



1  VSIi
8i 2; . . . Nbus
1

Vk

Vk+1
Ik

kth branch
Rk+jXk

Sending node

Receiving node
Pk+1 + jQ k+1

Fig. 1. A branch of a radial distribution system.

803

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

The proposed objective function (F) of the problem is formulated to minimize the total power loss and voltage stability
deviation index. The objective function can be described as
follows:

minimize F DPRloss DVSI

Constraints
Each proposed configuration in DNR considering DGs process,
the power flow analysis should be carried out to calculate the voltage stability index, power loss of system and current of each
branch. The constraints of objective function are as follows:
(1) The computed voltages and currents should be in their
premising range for the proposed configuration.

V min 6 V i 6 V max ;
0 6 Ii 6 Imax;i ;

I 1; 2; . . . Nbus

i 1; 2; . . . Nbr

6
7

(2) The radial nature of distribution network must be maintained and all loads must be served.
(3) Distributed generation capacity limits:

0 6 PDGi 6 PDGmax;i ;

i 1; 2; . . . NDG

Adaptive CSA for DNR considering DGs

CSA is a meta-heuristic search algorithm which has been proposed recently by Yang and Deb [20]. The algorithm is inspired
based on the reproduction strategy of cuckoos. At the most basic
level, cuckoos lay their eggs in the nests of other host birds. The
host bird may discover that the eggs are not its own and either
destroy the egg or abandon the nest all together. To apply this as
an optimization tool, Yang and Deb [20] used three idealized rules:
At a time, each cuckoo lays one egg and puts its egg in a randomly chosen nest among the available host nests which is fixed.
The best nests with high quality of egg will be carried over to
the next generation.
A host bird can detect an alien egg. In this case, it can either
throw the egg away or abandon the nest so as to build a new nest
in a new location.
When the CSA algorithm is employed to solve the DNR considering DGs. The radial constraint imposes the main hurdle since a
large number of infeasible individuals appear during initial stage
and intermediate stages of the evolutionary process because a set
of definite number of tie lines will act as a host nests, such that
when they open, a feasible radial configuration is formed or not.
Therefore, the CSA needs modification using some engineering
knowledge base to make it compatible with the DNR problem. In
the proposed adaptive CSA (ACSA), the host nests generating are
adapted using graph theory to reduce the number of infeasible
individuals at each stage of the optimization process.
Variable expression for DNR

is equal 1, and if any branch i is directed toward node j, element is

equal 1, otherwise its element is zero [26].
Fundamental loops (FLs) based on a graph theory are determined for the meshed network by closing all tie-switches. It is
found that, the number of fundamental loops of the system is
equal to the number of normal open branches (NO) [27,28]. For
finding FLs of the network after determining connection matrix
A, in each level, a normal open branch is added to system to form
a new loop [28]. Based on the method proposed in [28], branches
connected to the nodes, which have sum of absolute each element of a corresponding column in the connection matrix A is
1, are removed. This process is repeated until this node type no
longer remains in the system. Finally the numbers of remaining
branches can be saved in a vector that this vector represents a
FL [28]. Fig. 2 shows the topology of a 16-node system and the
first FL is determined in Fig. 3. As presented in Fig. 3, after determining connection matrix A, to identify the first FL, the following
steps are done.
Step 1: Add branch 14 to the connection matrix A.
Step 2: Calculate sum of absolute each element of a corresponding column in matrix A, the nodes which have the result is 1, are
node 3, 8, 9, 10, and 12. Therefore, branch 2, 6, 12, 13, and 8 are
removed. Because these branches connected to node 3, 8, 9, 10,
and 12 respectively.
Step 3: Similarly to step 2, calculate sum of absolute each element of a corresponding column for rest branches in matrix A
and the result is the branch 5 removed.
Step 4: Similarly to step 2, calculate sum of absolute each element of a corresponding column for rest branches in matrix A
and the result is the branch 4 removed.
Step 5: There are not any nodes that have sum of absolute each
element of a corresponding column in matrix A are 1. Therefore,
the first FL consists of branches {1, 3, 7, 9, 10, 11, and 14}. Similarly to branch 14, by adding branch 15 the second FL which
consists of branches {1, 2, 4, 5, 12, and 15} will be obtained.
Continually, the third FL with branches {4, 6, 7, 8, and 16} will
be obtained by adding branch to matrix A. Pseudo code of the
finding fundamental loop algorithms is given in Fig. 4.
Each radial configuration which involves a set of open branches
are randomly chosen from corresponding FLs. This helps to reduce
the generation of infeasible configuration during each stage of the
optimization algorithm. However, some of the branches are common in some of FLs [28]. Therefore, radial condition of network
must be checked. In each configuration, the connection matrix A
is determined. Then, the first column corresponding to the reference node in the network will be removed to form a square matrix
if the configuration of the network is radial. It is found that, if the

1
s4

s1

s3
5

s13

s14

s8

s16
s9

10

s15
s11

12

s12

s2

11

s6

s7

6
s5

In the conventional CSA, the initial population is randomly created, which consists of a large number of infeasible individuals violating the radial constraint, particularly in medium/large
distribution networks. In the proposed ACSA, these infeasible individuals are reduced by using graph theory. A distribution network
can be represented with a graph that contains a set of branches B
and a set of nodes N. They are presented by connection matrix A,
which has one row for each branch and one column for each node.
In this matrix if any branch i is directed away from node j, element

14

s10

Fig. 2. The topology of a 16-node system.

13

804

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Fig. 3. Determination FL when closing branch 14.

Input: Connected matrix A for the initial configuration of distribution network,

initial open branches.
Output: Fundamental loops
For (k: =1 to the number of normal open branches) do
Add the normal open branch k to the matrix A
Sum_column:= Sum of absolute each element of each column in the
matrix A
While (Sum_column jth =1, j=1 N) do
For (i=1 to the number of branches) do
If element (i, j) =1 or -1 then
Remove the branch ith from matrix A
End if
End for i
Update sum of absolute elements of each column in the matrix A
End while
Save the remaining branches of matrix A to the FL kth
End for k
Fig. 4. Pseudo code of the finding fundamental loop algorithms.

determinant of square matrix A is equal to 1 or 1 then the system

is radial [26]. Pseudo code of the checking system radially algorithm is given in Fig. 5.
Implementation of adaptive CSA for DNR considering DGs
Based on the three rules of CSA in Section Adaptive CSA for DNR
considering DGs, the ACSA method is implemented for DNR considering DGs as follows:
Step 1: Determine a fundamental loops.
Step 2: Determine the upper bound and lower bound of each
tie-line based on the size of the corresponding fundamental
loop.
Step 3: Initialization.
In DNR process using ACSA, each radial structure of the network
is represented as a host nest. A population of N host nests is repreh
i
sented by X i X i1 ; . . . ; X id1 ; X id with i = 1, 2 . . . N. To solve DNR

problem with simultaneous DG allocation, the first part of the solution vector is taken as the number of open branches in the distribution network, and the second part is the number of buses chosen for
DG installation and the third part is the sizes of DGs. Thus the solution vector for simultaneous reconfiguration and DG installation
considering their location and size is formed as follows:

h
i
i
i
X i Tie1 ; . . . ; TieNO ; Lo:DGi1 ; . . . ; Lo:DGim ; Size:DGi1 ; . . . ; Size:DGim
9
where Tie1, Tie2, . . . , TieNO are open branches in the fundamental
loops FL1 to FLNO formed corresponding to the tie lines; Lo.DG1, Lo.
DG2, . . . , Lo.DGm are the buses for DG installation; and Size.DG1,
Size.DG2, . . . , Size.DGm are sizes of DG units (MW) to be installed
at buses respectively. NO and m are the number of tie lines and
the number of DGs, respectively.
In the ACSA, each egg can be regarded as a solution which is
randomly generated in the initialization. Therefore, each nest i of
the population is randomly initialized as follows:

805

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Input: a candidate configuration with set of open branches

Output: a candidate configuration is a radial configuration or not
Determine connection matrix A for the network, which involves initial open
branches.
Remove the first column of matrix A
Remove the rows of matrix A corresponding to open branches in candidate
configuration
If (matrix A is a square matrix)
Calculate determinant of square matrix A
If (determinant of square matrix A = 1 or -1)
Output: = a candidate configuration is a radial configuration
Else
Output: = a candidate configuration is not a radial configuration
End if
Else
Output: = a candidate configuration is not a radial configuration
End if
Fig. 5. Pseudo code of the checking system radially algorithm.

h
i
i
i
i
Tiei round Tielower;d1 rand  Tieupper;d1  Tielower;d1
h
i
Lo:DGi round Loilower;d2 rand  Loiupper;d2  Loilower;d2
h
i
i
i
i
Size:DGi Sizelower;d3 rand  Sizeupper;d3  Sizelower;d3

11

Some nests of population may violate the boundary condition of

the optimization problem. Therefore, they are redefined at the end
of the generation of new solution process, as follows:

12

new
new
X new
roundX new
d
d1 ; roundX d2 ; X d3 

10

where d1 = 1, 2 . . . NO, d2 = 1, 2 . . . m and d3 = 1, 2 . . . m. Tielower,d1

and Tieupper,d1 are minimum tie-line and maximum tie-line which
are encoded in fundamental loop d1. DGs are located from any
nodes excluding the reference node. Therefore, the lower bound
(Lolower,d2) and upper bound (Loupper,d2) of each location of DG is
from node 2 to the maximum node of the network and the size of
each DG is from zero to maximum power of DG.
Based on the initialized population of the nests, the radial topology checking algorithm is run to check the nests and the objective
function of each nest is calculated by the power flows using NewtonRaphson method. Based on the values of objective function,
the nest with the best fitness function is saved to the best nest
Gbest.

Except for the best nest, all the other nests are replaced based
on the quality of new cuckoo eggs which are generated by Lvy
flights from their position as follows:

X new
Xbest d a  rand  DX new
d
d

randx
jrandy j

1=b

rx b
 Xbesti  Gbesti
ry b

14

where randx and randy are two normally distributed stochastic variables with standard deviation rx b and ry b given by:

31=b

C1 b  sinp2b5

rx b 4
ry b 1

b1

C1b
 b  2
2

lim
Tied1

8
Tielower;d1
>
>
>
<
Tieupper;d1
>
>
>
:
Tied1

Lo:DGlim
d2

if Tied1 < Tielower;d1

if Tied1 > Tieupper;d1

18

otherwise

8
2
>
>
>
<
Loupper;d2
>
>
>
:
Lo:DGd2

if Lo:DGd1 < 2
if Lod2 > Loupper;d2

19

otherwise

8
>
< Sizelower;d3
Sizeupper;d3
>
:
Size:DGd3

if Sized3 < Sizelower;d3

if Sized3 > Sizeupper;d3

20

otherwise

Based on the new population of the nests, the radial topology

checking algorithm is run to check the nests. Then, the fitness values are calculated to find the best value of each nest Xbestd.

13

where a > 0 is the step size parameter; rand is a random number in

is determined by:
range [0, 1] and the increased value DX new
i

DX new

The limits for each tie-line, location of each DG and power are
determined as follows:

Size:DGlim
d3

Step 4: Generation of new solution via Lvy flight

17

15

Step 5: Alien eggs discovery

A fraction pa of the worst nests can be abandoned so that new
nests can be built at new locations by random walks. Existing eggs
will be replaced by a good quality of new generated ones from
their current positions through random walks with step size as
follows:

X new
Xbesti K  DX new
i
i

where K is the updated coefficient determined based on the probability of a host bird to discover an alien egg in its nest:

K
16

where b is the distribution factor 0:3 6 b 6 1:99 and C is the

gamma distribution function.

21

if rand < Pa

0; otherwise

22

And the increased value DX new

is determined by:
i

DX new
rand  randp1 Xbesti  randp2 Xbesti 
i

23

806

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Input: line and load data of the distribution network

Output: optimal configuration and optimal the location and size of DGs
Step 1: Determine a fundamental loops.
Step 2: Determine the upper bound and lower bound of each tie-line based on size
of the corresponding fundamental loops.
Step 3: Generate initial population of N host nests
with i =
1, 2 N.
Check radially condition of each host nests by checking system radially
algorithm.
If hosts nest Xi is radial configuration then
Calculate the fitness function of Xi to find the best nest Gbest.
Else
Fitness function of Xi = inf
End if
Step 6: While (Iter < Itermax) do
Step 4: Get new solution by Lvy flight
Redefine the new solution depending on the boundary condition
Check radially condition of each host nests by checking system radially
algorithm
If hosts nest Xi is radial configuration then
Evaluate fitness function to choose new Xbestd
Else
Fitness function of Xi = inf
End if
Step 5: Get new solution by Alien eggs discovery
Redefine the new solution depending on the boundary condition
Check radially condition of each host nests by checking system radially
algorithm
If hosts nest Xi is radial configuration then
Evaluate fitness function to choose new Xbestd and Gbest
Else
Fitness function of Xi = inf
End if
If fitness (Gbest) < Fmin then
Fmin = fitness (Gbest)
Best nest = Gbest
End if
End While
Post process result: best fitness value Fmin and the Best nest
Fig. 6. Pseudo code of the ACSA for DNR considering DGs algorithm.

where rand is the random numbers in [0, 1] and

randp1 Xbesti ; randp2 Xbesti are the random perturbation for positions of the nests in Xbesti : Some nests of population may violate
the boundary condition of the optimization problem. Hence, they
are redefined at the end of the generation of new solution process,
as Eq. (17):

new
new
X new
roundX new
d
d1 ; roundX d2 ; X d3 

17

For the new solution, its lower and upper limits should be satisfied according to their limits by using Eqs. (18)(20).
Based on the new population of the nests, the radial topology
checking algorithm is run to check the nests. Then, the fitness values are calculated by the power flows using NewtonRaphson
method to find the best value of each nest Xbestd and the nest corresponding to the best fitness value is set to the best nest Gbest.

Step 6: Termination criterion

The generating new cuckoos by Lvy flight and discovering alien
eggs steps are alternatively performed until the number of iterations (Iter) reaches the maximum number of iterations (Itermax).
The Pseudo code of the proposed ACSA for DNR considering DGs
is given in Fig. 6.

Application results and analysis

In order to demonstrate and examine the applicability of the
proposed technique in solving the network reconfiguration and
installation of DG units simultaneously from small scale to large
scale distribution networks using ACSA, it is applied to three test
systems consisting of 33 buses, 69 buses and 119 buses. The maximum number of DGs installed for the given test systems is limited

807

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

to three. The limits of DG unit sizes chosen for installation are 0 to

2 MW for 33-bus and 69-bus system and 0 to 5 MW for 119-bus
system. In the simulation of network, seven scenarios are considered to analyze the superiority of the proposed method.

Table 1
Fundamental loops of the 33-bus system.

Scenario 1: Base case (without reconfiguration and distributed

generators).
Scenario 2: The system is only reconfigured.
Scenario 3: Allocation and size of DGs are optimized on base
case.
Scenario 4: Allocation and size of DGs are optimized after
reconfiguration of the network.
Scenario 5: The system is reconfigured after DGs installed based
on scenario 3.
Scenario 6: The System is simultaneous reconfigured and optimized size of DGs (VSI for all nodes of the system is computed
from power flow to identify the candidate bus location of DGs).
Scenario 7: The System is simultaneous reconfigured and optimized allocation and size of DGs.

24

FL1
FL2
FL3
FL4

2, 3, 4, 5, 6, 7, 18, 19, 20, 33

9, 10, 11, 12, 13, 14, 34
2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 18, 19, 20, 21, 35
6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 25, 26, 27, 28, 29, 30,
31, 32, 36
3, 4, 5, 22, 23, 24, 25, 26, 27, 28, 37

using scenarios 2, 3, 4, 5, 6 and 7, respectively. It is observed that

the power loss reduced using scenario 7 is the highest, which
demonstrates that the bus location of DGs need to optimize simultaneously with the reconfiguration and optimum size of DGs process. The voltage profiles (which are shown in node voltages and
VSI of nodes) of all seven scenarios are compared and shown in
Figs. 8 and 9. From the figures, it is observed that the voltage profile at all buses has been improved significantly after using reconfiguration and optimization of location and size of DGs.
In order to illustrate the performance of the proposed method,
the performance of ACSA is compared with the results of fireworks
algorithm (FWA) [19] and harmony search algorithm (HSA) [18]
available in the literature and is presented in Table 2. From the
table, it is perceived that at all scenarios, the performance of the
ACSA is better than FWA and HSA in terms of power loss and minimum voltage. The convergence results of system indices from scenario 2 to scenario 6 are shown in Fig. 10. It can be seen from
Fig. 10 that the fitness value of scenario 7 is the lowest compared
to other scenarios.

The 33-bus distribution system, which is a small-scale distribution networks, includes 37 branches, 32 sectionalizing switches
and 5 tie switches. The line and load data of this system are taken
from [29]. The total real and reactive power loads of the system are
3.72 MW and 2.3 MVAr, respectively. Fig. 7 shows the single line
diagram of this network. The parameters of ACSA algorithm used
in the simulation of network are number of nets Np = 30, probability of an alien egg to be discovered Pa = 0.2, number of iterations
Itermax = 2000.
Fundamental loops of the network are obtained by the finding
fundamental loop algorithms as given in Table 1. The performance
of the proposed method is presented in Table 2. From Table 2, it
can be seen clearly that in the initial case, power loss (kW) in the
system is 202.68, which is reduced to 139.98, 74.26, 58.79, 62.98,
63.69, and 53.21 using scenarios 2, 3, 4, 5, 6, and 7, respectively.
The percentage power loss reduction for scenario 27 is 30.93,
63.26, 71.0, 68.93, 68.58, and 73.75, respectively. It can be also
seen from Table 2 that, the minimum voltage magnitude of the system is improved remarkably in all the scenarios. In the base case,
the minimum voltage magnitude is improved from 0.9108 p.u. to
0.9413, 0.9778, 0.9802, 0.9826, 0.9786, and 0.9806 p.u. for using
case 2 to case 7. In addition, the VSI is also improved from
0.6978 to 0.7878, 0.9118, 0.9264, 0.9354, 0.9202, and 0.9318 by

23

Tie-line

FL5

33-bus test system

23

Fundamental
loop

69-bus test system

To demonstrate the applicability of the proposed method using
ACSA in medium-scale system. It is simulated on 69-bus system.
The 69-bus distribution system includes 69 nodes, 73 branches.
There are 5 tie switches and total loads are 3.802 MW and
2.696 MVAr [30]. The schematic diagram of the test system is
shown in Fig. 11. In a normal operation, switches {69, 70, 71, 72,
and 73} are opened. The parameters of ACSA algorithm used in
the simulation of network are number of nets Np = 30, probability
of an alien egg to be discovered Pa = 0.2, number of iterations
Itermax = 2000.
Similar to 33-bus test system, fundamental loops of the network are also obtained by the finding fundamental loop algorithms

25
24

37
26

27
26

22

29 30

28
27

28

29

31
30

32
31

25
1

2
1

3
2

18

8
7

11

10

9
9

10

12
11

13
12

33
19 20
19

21
20

22
21

36

32

34

64 7
5

33

35
Fig. 7. IEEE 33-bus test system.

15 16

14
13

14

15

17
16

18
17

808

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Table 2
Performance analysis of proposed method on the 33-bus system.
Scenario

Item

Proposed ACSA

FWA [19]

HSA [18]

Base case (Scenario 1)

Switches opened
Power loss (kW)
Minimum voltage (p.u.)
Minimum VSI

33, 34, 35, 36, 37

202.68
0.9108
0.6978

Only reconfiguration (Scenario 2)

Switches opened
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

07, 14, 9, 32, 28

139.98
30.93
0.9413
0.7878

7, 14, 9, 32, 28
139.98
30.93
0.9413

7, 14, 9, 32, 37
138.06
31.88
0.9342

Only DG installation (Scenario 3)

Switches opened
Size of DG in MW (Bus number)

33, 34,
0.7798
1.1251
1.3496
74.26
63.26
0.9778
0.9118

33, 34,
0.5897
0.1895
1.0146
88.68
56.24
0.9680

33, 34,
0.1070
0.5724
1.0462
96.76
52.26
0.9670

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
DG installation after reconfiguration
(Scenario 4)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
Reconfiguration after DG installation
(Scenario 5)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
Simultaneous Reconfiguration and DG
installation (Scenario 6)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
Switches opened
Size of DG in MW (Bus number)

Simultaneous Reconfiguration, DG
installation and location of DG
(Scenario 7)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

35, 36, 37
(14)
(24)
(30)

35, 36, 37
(14)
(18)
(32)

7, 14, 9, 32, 28
1.7536 (29)
0.5397 (12)
0.5045 (16)
58.79
71.00
0.9802
0.9264

7, 14, 9, 32, 28
0.5996 (32)
0.3141 (33)
0.1591 (18)
83.91
58.59
0.9612

7, 14, 9, 32, 37
0.2686 (32)
0.1611 (31)
0.6612 (30)
97.13
52.07
0.9479

33, 9, 8, 36, 27
0.7798 (14)
1.1251 (24)
1.3496 (30)
62.98
68.93
0.9826
0.9354

7, 34, 9, 32, 28
0.5897 (14)
0.1895 (18)
1.0146 (32)
68.28
66.31
0.9712

7, 10, 13, 32, 27

0.4263 (32)
1.2024 (29)
0.7127 (18)
63.69
68.58
0.9786
0.9202

7, 14, 11, 32, 28

0.5367 (32)
0.6158 (29)
0.5315 (18)
67.11
66.89
0.9713

7, 14, 10, 32, 28

0.5258 (32)
0.5586 (31)
0.5840 (33)
73.05
63.95
0.9700

33, 34, 11, 31, 28

0.8968 (18)
1.4381 (25)
0.9646 (7)
53.21
73.75
0.9806
0.9318

1
0.99

Voltage (p.u.)

0.98
0.97
0.96
0.95
0.94
0.93
0.92
0.91

Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Case 7

10

35, 36, 37
(18)
(17)
(33)

15

20

25

Node No.
Fig. 8. Comparison of node voltages of 33-bus system.

30

33

809

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

1
0.95

VSI

0.9
0.85

Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Case 7

0.8
0.75
0.7
0

10

15

20

25

30

33

Node No.
Fig. 9. Comparison of VSI-nodes of 33-bus system.

1
0.9

Case 2: Only Rec.

Case 3: Only DG ins.
Case 4: DG ins. after Rec.
Case 5: Rec. after DG ins.
Case 6: Rec. and size DG ins.
Case 7: Rec., loc. and siz. DG

Fitness value

0.8
0.7
0.6
0.5
0.4
0.3

200

400

600

800

1000

1200

1400

1600

1800

2000

Iterations
Fig. 10. Comparison of 33-bus system indices for scenarios in fitness function.

29

28

29
48

28
47

31

30

47

32

30

32

31

49

48

33

35

33

34

50
72
56

49
53

54

55
54

53
27

34

55

66

46

58

57
56

57

59

4
2

5
4

7
7
50

52
35

9 10
8

36

37

38

39
38

40

11

69

41

42
41

43
42

65

64
63

64

70

67

51

40
39

63
62

73

11 12

10

51
36 37

62
61

66
65

61
60

67

52
1

60

59

58

43

13 14

12

13

68

69

15 16
14 15 16

18 19

17 18 19

71

68
44
45
44

17

46
45

Fig. 11. IEEE 69-bus test system.

20

21 22
20

21

23

24 25

26

22 23 24 25 26

27

810

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Table 3
Fundamental loops of the 69-bus system.
Fundamental
loop

Tie-line

FL1
FL2
FL3

3, 4, 5, 6, 7, 8, 9, 10, 35, 36, 37, 38, 39, 40, 41, 42, 69

13, 14, 15, 16, 17, 18, 19, 20, 70
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 35, 36, 37, 38, 39, 40, 41,
42, 43, 44, 45, 71
4, 5, 6, 7, 8, 46, 47, 48, 49, 52, 53, 54, 55, 56, 57, 58, 72
9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,
26, 52, 53, 54, 55, 56, 57 58, 59, 60, 61, 62, 63, 64, 73

FL4
FL5

as given in Table 3. This test system is also simulated for seven scenarios and the results are presented in Table 4. It is observed from
Table 4, base case power loss (in kW) in the system is 224.89 which
is reduced to 98.59, 72.44, 37.23, 41.13, 40.49, and 37.02 using scenarios 2, 3, 4, 5, 6, and 7, respectively. The percentage loss reduc-

tion for scenario 27 is 56.16, 67.79, 83.45, 81.71, 82.0, and

83.54, respectively. In six scenarios, power loss reduction using
scenario 7, which is the proposed method, is the highest, which
elicits the superiority of the proposed method over the others.
From Table 4, it is also seen that improvement in power loss reduction and voltage profile for scenario 7 are higher when compared to
other scenarios. This illustrates that reconfiguration presence DGs
need to concern simultaneously the location and size of DG.
By using scenario 2 to scenario 7, the minimum voltage magnitude is improved from 0.9092 p.u. to 0.9495, 0.9890, 0.9870,
0.9828, 0.9873, and 0.9869 p.u. The VSI is enhanced significantly
from 0.6859 to 0.8414, 0.9546, 0.9390, 0.9260, 0.9403, and
0.9433. The voltage profiles of the system for seven scenarios are
compared and shown in Figs. 12 and 13. From the figures, it is
observed that the voltage profile at all buses has been improved
significantly after using proposed method. Fig. 14 shows the convergence characteristics of the algorithms for the best solution in

Table 4
Performance analysis of proposed method on the 69-bus system.
Scenario

Item

Proposed ACSA

FWA [19]

HSA [18]

Base case (Scenario 1)

Switches opened
Power loss (kW)
Minimum voltage (p.u.)
Minimum VSI

69, 70, 71, 72, 73

224.89
0.9092
0.6859

Only reconfiguration (Scenario

2)

Switches opened
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

69, 70, 14, 57, 61

98.59
56.16
0.9495
0.8414

69, 70, 14, 56, 61

98.59
56.16
0.9495

69, 18, 13, 56, 61

99.35
55.85
0.9428

Only DG installation (Scenario 3)

Switches opened
Size of DG in MW (Bus number)

69, 70, 71, 72, 73

0.6022 (11)
0.3804 (18)
2 (61)
72.44
67.79
0.9890
0.9546

69, 70, 71, 72, 73

0.4085 (65)
1.1986 (61)
0.2258 (27)
77.85
65.39
0.9740

0.1018 (65)
0.3690 (64)
1.3024 (63)
86.77
61.43
0.9677

69, 70, 14, 57, 61

1.7254 (61)
0.4666 (64)
0.3686 (12)
37.23
83.45
0.9870
0.9390

69, 70, 14, 56, 61

1.0014 (61)
0.2145 (62)
0.1425 (64)
43.88
80.49
0.9720

69, 18,
1.0666
0.3525
0.4257
51.3
77.2
0.9619

69, 70, 14, 58, 64

0.6022 (11)
0.3804 (18)
2 (61)
41.13
81.71
0.9828
0.9260

69, 70, 12, 58, 61

0.4085 (65)
1.1986 (61)
0.2258 (27)
39.69
82.35
0.9763

69, 70, 12, 58, 61

1.7496 (61)
0.1566 (62)
0.4090 (65)
40.49
82.0
0.9873
0.9403

69, 70, 13, 55, 63

1.1272 (61)
0.2750 (62)
0.4159 (65)
39.25
82.55
0.9796

69, 17,
1.0666
0.3525
0.4257
40.3
82.08
0.9736

69, 70, 14, 58, 61

0.5413 (11)
0.5536 (65)
1.7240 (61)
37.02
83.54
0.9869
0.9433

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
DG installation after
reconfiguration (Scenario 4)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
Reconfiguration after DG
installation (Scenario 5)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
Simultaneous Reconfiguration
and DG installation (Scenario
6)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
Simultaneous Reconfiguration,
DG installation and location
of DG (Scenario 7)

Switches opened
Size of DG in MW (Bus number)

Power loss (kW)

% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

13, 56, 61
(61)
(60)
(58)

13, 58, 61
(61)
(60)
(62)

811

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

1
0.99

Voltage (p.u.)

0.98
0.97
0.96
0.95
Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Case 7

0.94
0.93
0.92
0.91

10

20

30

40

50

60

69

Node No.
Fig. 12. Comparison of node voltages of 69-bus system.

1
0.95

VSI

0.9
0.85

Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Case 7

0.8
0.75
0.7
0

10

20

30

40

50

60

69

Node No.
Fig. 13. Comparison of VSI-nodes of 69-bus system.

Case 2: Only Rec.

Case 3: Only DG ins.
Case 4: DG ins. after Rec.
Case 5: Rec. after DG ins.
Case 6: Rec. and size DG ins.
Case 7: Rec., loc. and siz. DG

0.9

Fitness value

0.8
0.7
0.6
0.5
0.4
0.3
0.2
0

200

400

600

800

1000

1200

1400

1600

1800

2000

Iterations
Fig. 14. Comparison of 69-bus system indices for scenarios in fitness function.

six scenarios. From the figure, the fitness function in scenario 7 is

the most minimum in six scenarios.
Similar to 33-bus test system, the performance of ACSA on
69-bus system is also compared with the results of FWA and
HSA, and the results are presented in Table 4. From the table, it
is observed that the performance of the ACSA is better compared
to HSA in terms of the quality of solutions in most scenarios. Also

from Table 4, the best results are identical to results obtained by

FWA in the scenario 2. It is also shown from Table 4 that the ACSA
has outperformed FWA in terms of power loss minimization and
voltage stability enhancement in the scenario 3 and 4. In the scenario 5 and 6, although the proposed algorithm finds the optimal
configuration with the power loss reduction in percent are 81.71
and 82.0, respectively. This values are 0.64 and 0.55 lower than

812

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

s12
s11
s10

s18

s17

s9

s19
19

18
s3

s4

s5

s6

s7

s27

s20

38

s29 s30

29

s21

17

s119

s22
22

s39

s31

23

s41

s33

36

s34

43
s46

35
s123

34

s55

s56
56

s57

s89
89

s88
s66

67

s78

s79

78

s90

s48

s121

46
s49

49

s50

s51
51

50

s52
52

s53
53

54

s122

79
s85

s59

s68

61 s61 62

s97

s93

s70

s71

68

69

70

71

s80

s81

s82

s83

80
82 83
81
s87 s87

99
s128

s94
95

94

93

s69

s98
98

97

96

92

s60
60

s91 s92
91

90
s67

65 s65 66

s77

s45
45

48

59
s125
s96

s95

s62

s99

s47
47

s58

57 58

64

s44
44

37

55

63

27

26 s26

s118

s42 s43
42

41

s32

s54

s63 s64

s25
25

s132

s40
40

39

s0

s24
24

s124

31 32
33
s35
s36

30

s23

s120
9

s38

s28

s16
16

21

20

s37

28

s15
15

s8

s2

s1

s14
14

11

10

s13
13

12

s72
72

s126
s73

73

s74
74

s75
75

76 s76

77

s84
84

85

s127

86

88
87
s129
s130
s100 s101 s102 s103 s104 s105 s106 s107 s108 s109 s111 s112
100
s113

101

102

103

104 105

106

107

108

s114 s115 s116 s117

114

115 116

117

118

s131

109 110 112

s110

113

111

Fig. 15. 119-bus test system.

Table 5
Fundamental loops of the 119-bus system.
Fundamental loop

Tie-line

FL1
FL2
FL3
FL4
FL5
FL6
FL7
FL8
FL9
FL10
FL11
FL12
FL13
FL14
FL15

3, 9, 10, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 37, 38, 39, 40, 41, 42, 43, 44, 45, 118
11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 119
3, 4, 5, 6, 7, 9, 10, 17, 18, 19, 20, 21, 22, 23, 120
29, 30, 31, 32, 33, 34, 37, 38, 39, 40, 41, 42, 46, 47, 48, 49, 50, 51, 52, 53, 121
29, 30, 31, 32, 33, 34, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 122
29, 35, 36, 54, 55, 56, 57, 58, 59, 60, 61, 123
4, 5, 6, 7, 8, 27, 28, 37, 38, 39, 124
1, 3, 27, 28, 54, 55, 56, 57, 62, 63, 64, 88, 89, 90, 95, 125
65, 66, 67, 68, 69, 70, 71, 72, 88, 89, 90, 126
64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 77, 78, 85, 86, 87, 127
65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 88, 89, 90, 95, 96, 97, 98, 128
62, 63, 77, 78, 79, 80, 81, 82, 99, 100, 101, 102, 103, 104, 105, 106, 107, 129
62, 63, 77, 78, 85, 99, 100, 101, 102, 103, 104, 130
100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 113, 114, 115, 116, 117, 131
3, 9, 10, 17, 18, 19, 20, 21, 22, 23, 24, 27, 28, 29, 30, 31, 32, 33, 34, 132

the optimum solution obtained from FWA in the scenario 5 and 6,

respectively. However, it is very clear that ACSA has outperformed
FWA in term of voltage stability enhancement.
119-bus test system
The 119-node distribution system, which is a large-scale distribution networks, consists 118 sectionalizing switches, 119 nodes

and 15 tie switches as shown in Fig. 15. The total power loads
are 22709.7 kW and 17041.1 kVAr [31]. The parameters of ACSA
algorithm used in the simulation of network are number of nets
Np = 30, probability of an alien egg to be discovered Pa = 0.2.
Because the number of variables need to optimize in scenario 6
and scenario 7 are 18 and 21, respectively compared with 15, 6,
6, and 15 in scenarios 2, 3, 4 and 5, respectively. Thus, to ensure
the obtained solutions are minimum, the number of iterations

813

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815
Table 6
Performance analysis of proposed method on the 119-bus system.
Scenario

Item

Proposed ACSA

Base case (Scenario 1)

Switches opened
Power loss (kW)
Minimum voltage (p.u.)
Minimum VSI

118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130,
131, 132
1273.45
0.8678
0.5676

Only reconfiguration (Scenario 2)

Switches opened
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

42, 25, 23, 121, 50, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34
855.04
32.86
0.9298
0.7535

Only DG installation (Scenario 3)

Switches opened

118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130,
131, 132
3.2664 (71)
3.1203 (109)
2.86267 (50)
648.10
49.11
0.9515
0.8199

Size of DG in MW (Bus
number)
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI
DG installation after reconfiguration (Scenario 4)

Switches opened
Size of DG in MW (Bus
number)
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

Reconfiguration after DG installation (Scenario 5)

Switches opened
Size of DG in MW (Bus
number)
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

Simultaneous Reconfiguration and DG installation (Scenario 6)

Switches opened
Size of DG in MW (Bus
number)
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

Simultaneous Reconfiguration, DG installation and location of DG

(Scenario 7)

Switches opened
Size of DG in MW (Bus
number)
Power loss (kW)
% Loss reduction
Minimum voltage (p.u.)
Minimum VSI

42, 25, 23, 121, 50, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34
1.7145 (111)
1.7565 (96)
5 (65)
631.19
50.43
0.9526
0.8208
42, 25, 21, 121, 122, 58, 39, 125, 70, 127, 128, 129, 85, 131, 33
3.2664 (71)
3.1203 (109)
2.86267 (50)
613.79
51.80
0.9608
0.8523
42, 25, 23, 121, 50, 61, 39, 125, 126, 70, 75, 129, 130, 109, 34
2.9585 (75)
0.1924 (76)
1.3397 (77)
682.09
46.44
0.9298
0.7535
42, 25, 22, 121, 122, 58, 39, 125, 70, 127, 128, 81, 130, 131, 33
2.5331 (50)
3.6819 (109)
3.7043 (73)
586.24
53.96
0.9644
0.8700

Table 7
Comparison of simulation results for 119-node network in Scenario 2.
Methods

Open Switches

Delta P (kW)

Vmin (p.u.)

Initial
Proposed ACSA
ITS [31]
MTS [26]
CSA [4]
FWA [32]

119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133
42, 25, 23, 121, 50, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34
42, 26, 23, 51, 122, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34
42, 26, 23, 51, 122, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34
42, 25, 23, 121, 50, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34
42, 25, 23, 121, 50, 58, 39, 95, 71, 74, 97, 129, 130, 109, 34

1273.45
855.04
867.4
867.4
855.04
855.04

0.8678
0.9298
0.9323
0.9323
0.9298
0.9298

selected for scenario 6 and 7 are Itermax = 5000 and Itermax = 2000
for the rest of scenarios.
The fundamental loops of the network are obtained by the
finding fundamental loop algorithms as given in Table 5. Table 6
show the results obtained from proposed method for seven

scenarios. It can be seen from Table 6, base case power loss (in
kW) in the system is 1273.45 which is reduced to 855.04,
648.10, 631.19, 613.79, 682.09, and 586.24 using scenarios 2, 3,
4, 5, 6, and 7, respectively. The percentage loss reduction for
scenario 27 is 32.86, 49.11, 50.43, 51.80, 46.44, and 53.96,

814

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

Voltage (p.u.)

0.98
0.96
0.94
0.92

Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Case 7

0.9
0.88
0

20

40

60

80

100

118

Node No.
Fig. 16. Comparison of node voltages of 119-bus system.

1
0.95
0.9

VSI

0.85
0.8
0.75

Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Case 7

0.7
0.65
0.6
0.55

20

40

60

80

100

118

Node No.
Fig. 17. Comparison of VSI-nodes of 119-bus system.

Case 2: Only Rec.

Case 3: Only DG ins.
Case 4: DG ins. after Rec.
Case 5: Rec. after DG ins.
Case 6: Rec. and size DG ins.
Case 7: Rec., loc. and siz. DG

1.8

Fitness Value

1.6
1.4
1.2
1
0.8
0.6
0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Iteration
Fig. 18. Comparison of performance of ACSA in six scenarios for minimization of the 119-bus system.

respectively. In six scenarios, power loss reduction using scenario

7, which is the proposed method, is the highest, which elicits the
superiority of the proposed method over the others. By using case
2 to case 7, the minimum voltage magnitude is improved from
0.8678 p.u. to 0.9298, 0.9515, 0.9526, 0.9608, 0.9298, and
0.9644 p.u. Similar to the minimum voltage magnitude, the VSI

is also improved from 0.5676 to 0.7535, 0.8199, 0.8208, 0.8523,

0.7535, and 0.87 by using scenarios 2, 3, 4, 5, 6 and 7, respectively. From Table 6, it is seen that enhancement in power loss
reduction and voltage profile for scenario 7 are higher when compared to scenario 6. This implies that simultaneous reconfigured
and optimized size of DGs, which is the scenario 6, does not yield

T.T. Nguyen et al. / Electrical Power and Energy Systems 78 (2016) 801815

desired results of minimizing power loss and maximizing voltage

profile.
Table 7 presents the comparison is made with previous study in
scenario 2. These results are identical to the results obtained by the
methods proposed in Refs. [4,32] and better than the results
obtained by the modified tabu search (MTS) algorithm [26] and
the improved tabu search (ITS) algorithm [31]. The voltage profiles
of the system for seven scenarios are compared and shown in
Figs. 16 and 17. From the figures, it is observed that the voltage
profile has been improved drastically after using proposed method.
Fig. 18 shows the convergence characteristics of the algorithms for
the best solution in six scenarios. As illustrated in this figure, the
fitness function in scenario 7 is the most minimum in six scenarios.
This results elicit the superiority of the proposed method.
Conclusion
In this paper, the ACSA method has been successfully applied
for distribution network reconfiguration and simultaneous location and size of DG problem. The objective is to minimize the active
power loss and enhance voltage stability index of power distribution systems. In addition, different loss reduction and enhancement voltage stability methods such as only network
reconfiguration, only DG installation, DG installation after reconfiguration, reconfiguration after DG installation, simultaneous
reconfiguration and DG installation are also simulated to establish
the superiority of the proposed method. The proposed method
based on graph theory is used to determine the search space of
each tie-line, which helps the cuckoo search algorithm reduces
infeasible network configurations at each stage of the optimization
process and it is also is adapted to check the radial constraint of
each generated configuration. The proposed method is tested on
33-bus, 69-bus, and 119-bus test systems. The results demonstrate
that network reconfiguration with simultaneous location and size
of DG is more effective in reducing power loss and improving the
voltage profile compared to other scenarios. The simulated results
are also compared with the results of FWA and HSA available in the
literature. The computational results have demonstrated that the
performance of the ACSA is better than FWA and HSA in most of
scenarios.
References
[1] Gopiya Naik S, Khatod DK, Sharma MP. Optimal allocation of combined DG and
capacitor for real power loss minimization in distribution networks. Int J Electr
Power Energy Syst 2013;53:96773.
[2] Sathish Kumar K, Jayabarathi T. Power system reconfiguration and loss
minimization for an distribution systems using bacterial foraging
optimization algorithm. Int J Electr Power Energy Syst 2012;36(1):137.
[3] Kalambe S, Agnihotri G. Loss minimization techniques used in distribution
network: bibliographical survey. Renew Sustain Energy Rev 2014;29:184200.
[4] Nguyen TT, Truong AV. Distribution network reconfiguration for power loss
minimization and voltage profile improvement using cuckoo search algorithm.
Int J Electr Power Energy Syst 2015;68:23342.
[5] Merlin A, Back H. Search for a minimal loss operating spanning tree
configuration in an urban power distribution system. In: Proceeding 5th
power syst comput conf (PSCC), Cambridge, UK, vol. 118; 1975.
[6] Civanlar S, Grainger JJ, Yin H, Lee SSH. Distribution feeder reconfiguration for
loss reduction. IEEE Trans Power Deliv 1988;3(3):121723.
[7] Duan D-L, Ling X-D, Wu X-Y, Zhong B. Reconfiguration of distribution network
for loss reduction and reliability improvement based on an enhanced genetic
algorithm. Int J Electr Power Energy Syst 2015;64:8895.

815

[8] Souza SSF, Romero R, Franco JF. Artificial immune networks Copt-aiNet and
Opt-aiNet applied to the reconfiguration problem of radial electrical
distribution systems. Electr Power Syst Res 2015;119:30412.
[9] Esmaeilian HR, Fadaeinedjad R. Distribution system efficiency improvement
using network reconfiguration and capacitor allocation. Int J Electr Power
Energy Syst 2015;64:45768.
[10] Hung DQ, Mithulananthan N, Bansal RC. An optimal investment planning
framework for multiple distributed generation units in industrial distribution
systems. Appl Energy 2014;124:6272.
[11] Doagou-Mojarrad H, Gharehpetian GB, Rastegar H, Olamaei J. Optimal
placement and sizing of DG (distributed generation) units in distribution
networks by novel hybrid evolutionary algorithm. Energy 2013;54:12938.
[12] Murty VVSN, Kumar A. Optimal placement of DG in radial distribution systems
based on new voltage stability index under load growth. Int J Electr Power
Energy Syst 2015;69:24656.
[13] Mohamed IA, Kowsalya M. Optimal size and siting of multiple distributed
generators in distribution system using bacterial foraging optimization.
Swarm Evol Comput 2014;15:5865.
[14] Ugranli F, Karatepe E. Multiple-distributed generation planning under load
uncertainty and different penetration levels. Int J Electr Power Energy Syst
2013;46(1):13244.
[15] Kayal P, Chanda CK. Placement of wind and solar based DGs in distribution
system for power loss minimization and voltage stability improvement. Int J
Electr Power Energy Syst 2013;53:795809.
[16] Ameli A, Shahab B, Farid K, Mahmood-Reza H. A multiobjective particle swarm
optimization for sizing and placement of DGs from DG Owners and
distribution companys viewpoints. IEEE Trans Power Deliv 2014;29
(4):183140.
[17] Tan S, Xu JX, Panda SK. Optimization of distribution network incorporating
distributed generators: an integrated approach. IEEE Trans Power Syst
2013;28(3):242132.
[18] Rao RS, Ravindra K, Satish K, Narasimham SVL. Power loss minimization in
distribution system using network reconfiguration in the presence of
distributed generation. IEEE Trans Power Syst 2013;28(1):19.
[19] Mohamed Imran A, Kowsalya M, Kothari DP. A novel integration technique for
optimal network reconfiguration and distributed generation placement in
power distribution networks. Int J Electr Power Energy Syst 2014;63:46172.
[20] Yang XS, Deb S. Cuckoo search via Lvy flights. In: 2009 world congress on
nature and biologically inspired computing, NABIC 2009 proceedings; 2009.
p. 2104.
[21] Balasubbareddy M, Sivanagaraju S, Suresh CV. Multi-objective optimization in
the presence of practical constraints using non-dominated sorting hybrid
cuckoo search algorithm. Eng Sci Technol Int J 2015;18(4):60315.
[22] Abd Elazim SM, Ali ES. Optimal power system stabilizers design via cuckoo
search algorithm. Int J Electr Power Energy Syst 2016;75:99107.
[23] Abdelaziz AY, Ali ES. Cuckoo search algorithm based load frequency controller
design for nonlinear interconnected power system. Int J Electr Power Energy
Syst 2015;73:63243.
[24] Dash P, Saikia LC, Sinha N. Comparison of performances of several FACTS
devices using Cuckoo search algorithm optimized 2DOF controllers in multiarea AGC. Int J Electr Power Energy Syst 2015;65:31624.
[25] Chakravorty M, Das D. Voltage stability analysis of radial distribution
networks. Int J Electr Power Energy Syst 2001;23(2):12935.
[26] Abdelaziz AY, Mohamed FM, Mekhamer SF, Badr MAL. Distribution system
reconfiguration using a modified Tabu search algorithm. Electr Power Syst Res
2010;80(8):94353.
[27] Gupta N, Swarnkar A, Niazi KR, Bansal RC. Multi-objective reconfiguration of
distribution systems using adaptive genetic algorithm in fuzzy framework. IET
Gener Transm Distrib 2010;4(12):1288.
[28] Dolatdar E, Soleymani S, Mozafari B. A new distribution network
reconfiguration approach using a tree model. World Acad Sci Eng Technol
2009;3(10).
[29] Baran ME, Wu FF. Network reconfiguration in distribution systems for loss
reduction and load balancing. IEEE Trans Power Deliv 1989;4(2):14017.
[30] Chiang H-D, Jean-Jumeau R. Optimal network reconfigurations in distribution
systems: Part 2: solution algorithms and numerical results. IEEE Trans Power
Deliv 1990;5(3):156874.
[31] Zhang D, Fu Z, Zhang L. An improved TS algorithm for loss-minimum
reconfiguration in large-scale distribution systems. Electr Power Syst Res
2007;77(56):68594.
[32] Mohamed Imran A, Kowsalya M. A new power system reconfiguration scheme
for power loss minimization and voltage profile enhancement using Fireworks
Algorithm. Int J Electr Power Energy Syst 2014;62:31222.