Received: November 5, 2018 290

Optimal Allocation and Sizing of Multiple DG in Radial Distribution System

Using Binary Particle Swarm Optimization

Banka Jyothsna Rani1* Ankireddipalli Srinivasula Reddy2

1
Government Model Residential Polytechnic, India
2
CMR Engineering College, India
* Corresponding author’s Email: jyothsna.jyothsna3@gmail.com

Abstract: In modern days, utility grids are facing major issues in the power system due to an uninterrupted load
development and shortage of capital investment in transmission systems. Due to this, network is not capable to meet
the estimated demand and transfer the produced power from the integrated power generation into the distribution
network. This may generate high power losses in the distribution side that also disturbs the overall efficiency of the
network. By employing an optimal placement of Distributed Generation (DG) in power systems is the only way to
minimize this kind of problems in economical and practical way. DG has several benefits over centralized power
generations that comprising system steadiness, improvement and dismissing transmission/distribution system
bottleneck to meet the power demand. So in this paper, Binary Particle Swarm Optimization (BPSO) is used for
optimal DG reconfiguration. BPSO iteratively moves the DG in each load bus of the test system that provides the
smallest value of the objective function, which is considered as the finest candidate for DG connection. This method
is simulated on IEEE 69-bus system that showed enhancement in voltage profile and decrease in power loss
compared to other optimization methods.
Keywords: Binary particle swarm optimization (BPSO), Distributed generation (DG), Optimal placement, Power
loss, Voltage profile.

complicated optimization issue [5]. Currently,

1. Introduction several researchers improved the existing
procedures by altering the development and
In recent years, there has been a developing
hybridization of different algorithms [6]. Numerous
concern in Renewable Energy Sources (RES) due to
methods used to recognize both fundamental and
the growing demand of the electricity and
progressive methods to solve DG location and sizing
substantial reduction of fossil fuel. For that reason,
problems [7].
research on the incorporation of DG into the
The optimal location of DG is essential to attain
distribution system has turned out to be very popular.
the maximum profits with less investment. After
The incorporation of DG is anticipated to play an
deregulation of power system, non-utility
important part in the electric power system [1, 2].
corporations are investing in distribution network to
DG is one of the small-scale power generations that
meet the active power demand (MW) and get the
is frequently associated with distribution system.
maximum profit [8]. Growing machineries play an
DG is playing a major role in minimizing the losses
important role in the development of power
and enhancing voltage stability in power systems [3],
generation; DGs could rapidly counter the demand
[4]. Along with these assistances, use of DG
variants to carry out the demand side management
components of the non-optimal location can lead to
[9]. Therefore, it is extremely important to allocate
disadvantages. As a result, the choice of best
each DG at the correct position to evade the poor
placement for the connection and the desirable size
performance of the electrical network [10, 11]. More
of DG units in the distribution system is a
International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 291

than a few sets of rules were used for DG important contributions to the existing literatures is
employment like Artificial Bee Colony Algorithm represented below.
(ABC), Particle Swarm Optimization (PSO), S. Kaur, G. Kumbhar, and J. Sharma [17]
Modified Teaching–Learning Based Optimization presented Mixed Integer Non-Linear
(MTLBO), Genetic Algorithm (GA), and Harmony Programming (MINLP) method for finding the best
Search Algorithm (HSA) [12, 13]. The main location of multiple DG units in the distribution
disadvantages of all the above approaches are network to improve the stability of the system. The
insufficient reactive power support in distributed established MINLP shortens the difficulties by
network that leads to higher power losses, bus distributing the system into two models, namely
voltage reduction, low convergence speed and Siting Planning Model (SPM) and Capacity
obtain nearby optimal solutions only. The DG units Planning Model (CPM) that provide reduction in
have the ability to supply the reactive power when computational time and search space. In case of
connected to any of the bus in the network [14]. multiple DGs, optimal positions are same for
Additionally, the non-optimal location of DG can MINLP techniques. However, this method gives
result in increase of the system losses that makes the more loss, which is slightly higher than the PSO
voltage profile lower than the permissible range. algorithm.
Therefore, there are numerous problems concerning D. B. Prakash, and C. Lakshminarayana [18]
the incorporation of DGs with the existing power proposed multiple DG employments in the
system [15]. From that, it is essential to establish a distribution network for decreasing the power loss
new procedure that can be robust, efficient, and and enhancing the voltage stability using Particle
comparatively more consistent with the other Swarm Optimization (PSO) Algorithm. In this
constraints [16]. Here, network reconfiguration concept, PSO was used to find out the optimal
based on Binary Particle Swarm Optimization location and sizing of DG. However, this method
(BPSO), which is recommended in this research only focuses on power loss issues, does not consider
work. The velocity update formulation of BPSO is load restoration and cost-related problems.
same as PSO (Particle Swarm Optimization) Additionally, this method easily falls into local
algorithm, but the position update formula is optima in high dimensional space and it has a very
different. During the initialization process of BPSO low convergence rate.
particle population, the result of the issues is P. Chiradeja, S. Yoomak, and A. Ngaopitakkul
transformed into binary code and particle’s position [19] presented an optimal placement of multiple
vectors can take the value as 0 or 1. Then these DGs on distribution network to verify its reliability
vectors drawn from the n-dimensional binary space using Differential Evolution (DE) algorithm. In this,
(B) into the real numbers (F) that means a particle’s both DG size and its placement were connected
location must belong to B that is considered by F to nearby to the maximum load, so it minimized the
discover the ideal distribution system structure for active and reactive power losses. In the DE
power loss minimization. The main objective of this algorithm, modification of parameter is
paper is to minimize the power loss and improve the indispensable, so identical parameters may not
voltage profile in the system by employing DG. This assure the global optimal resolution.
algorithm is applied on IEEE 69-bus Radial . Das, V. Mukherjee, and D. Das [20] developed
Distribution Network (RDN), and then it has been symbiotic organism search algorithm for the best
verified by BPSO. This proposed work gives an placement of DG in RDN. Power loss minimization
efficient performance over the distributed system can be attained with the help of this proposed
when compared with other traditional techniques. algorithm. In this work, location of DG was
The paper is organized as follows. Section 2 depended upon generation and deterministic load
provides a brief description of the related works. demand. The outcome also showed better
Section 3 focuses on BPSO methodology. In section convergence characteristics and better
4, comparative study of the results and discussion of computational efficiency of the proposed method.
proposed and existing system is presented. Section 5 However, the voltage profile of the system is not
gives a summary of this paper. considered while optimizing the DG location and
size.
2. Literature review P. V. Babu, and S. P. Singh [21] presented NLP
& PLS approaches for the optimal allocation of DG
Several researchers recommended many
in distribution system. By integrating General
techniques in the field of optimal allocation of DG.
Algebraic Modelling System (GAMS), optimal
In this scenario, a brief evaluation of some
sizing and allocation of DG in RDN were efficiently
International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 292

done. For that reason, power loss in the system was ω = ω𝑚𝑎𝑥 – {(ω𝑚𝑎𝑥 − ω𝑚𝑖𝑛 ) − k 𝑚𝑎𝑥 )} × 𝑘
slightly decreased. However, Percentage loss (2)
reduction is same and optimum power factor is not
attained in this method. Where k 𝑚𝑎𝑥 is the number of the maximum
To overcome the above mentioned problems in iteration. At the end of each iteration, a new position
the previous works, an effective optimization for each particle is obtained by summing of old
algorithm is proposed, named as Binary Particle position and new velocity.
Swarm Optimization (BPSO) which is described
briefly in the below section. 𝑋 𝑖 𝑘+1 = 𝑋 𝑖 𝑘 + 𝑉 𝑖 𝑘+1 (3)

3. BPSO algorithm The particle swarm formula remained

Particle Swarm Optimization (PSO) is presented unchanged. A logistic transformation 𝑆(𝑉 𝑖 𝑘 ) is used
as one of the most computation-efficient to accomplish this modification that is written in Eq.
experimental method. Compared to Genetic (4) and (5).
Algorithm (GA) and other traditional methods, PSO
1
delivers improved performance in DG placement 𝑆(𝑉 𝑖 𝑘+1 ) = 𝑠𝑖𝑔 𝑚𝑜𝑑 𝑒(𝑉 𝑖 𝑘+1 ) = 1+exp(𝑉 𝑖 𝑘+1 )
complications. A new PSO method is suggested for
(4)
selecting the variables from a selective space. This
process becomes convenient when available DGs
have their regular sizes in which the optimizer If rand ∝ 𝑆(𝑉 𝑖 𝑘+1 ) then: 𝑋 𝑖 𝑘+1 = 1; (5)
should choose the ideal ones. A typical IEEE 69-bus Else: 𝑋 𝑖 𝑘+1 = 0 ;
RDN is employed to point out the practicality of the
proposed method. However, there are various The function 𝑆(𝑉 𝑖 𝑘 ) is a sigmoid limiting
optimization issues, which are discrete in nature and transformation and rand is a quasi-random number
categorized as Combinatorial Optimization selected from a uniform distribution in [0,1]. Eq. (6),
Difficulties. A “Continuous PSO” method would not (7) and (8) describe the limits of the particle’s
suit here, so a “Binary PSO” algorithm is developed. dimension.
Binary versions of the PSO has also been
established that have higher benefits for resolving 1 ∝ 𝐵𝑖 ∝ 𝐵𝑚𝑎𝑥 (6)
optimization issues. Therefore, the alterations make
the PSO algorithm to work for these different types 0 ∝ 𝑃𝑖 ∝ 𝑃𝑚𝑎𝑥 (7)
of problems. This solution creates the particle’s bit
string and constrain the value of velocity in the 𝑇𝑖 = {1,2, … . , 𝑇𝑓 } (8)
interval of [0 1]. The velocity is stated as the
possibility of a bit to achieve a value of 1. BPSO 3.1 Optimal DG allocation
appears like to be more appropriate to improve the
efficiency of the distribution system. So, in this The incorporation of DG into the distributed
study, a novel BPSO algorithm is established for system may have an effect on the frequency,
enlightening the steadiness of the system by protection discrimination, voltage regulation and
reducing the power losses. stability of the system. Improper choice of size and
In PSO, the position and the velocity of every location of DG may produce more system losses. By
particle at the iteration 𝑘 in the search space are optimal placement of DGs, distributed system take
described by 𝑋 𝑖 𝑘 and 𝑉 𝑖 𝑘 . The velocity of the benefits of improving the reliability of supply. DG
particle 𝐼 in the iteration 𝑘 + 1 𝑃𝑖 𝑙𝑏𝑒𝑠𝑡 is obtained could be deliberated as one of the sustainable
from the following Eq. (1). choices to alleviate some of the issues such as, low
reliability, high power loss and poor power quality
𝑉 𝑖 𝑘+1 = ω. 𝑉 𝑖 𝑘 + 𝐶1. 𝑅1 (𝑃𝑖 𝑙𝑏𝑒𝑠𝑡 - 𝑋 𝑖 𝑘 ) which helps in meets the energy requirement of
ever-increasing loads. For a specific bus, the high
+𝐶2. 𝑅2 (𝑃𝑖𝑔𝑙𝑜𝑏𝑎𝑙 − 𝑋 𝑖 𝑘 ) (1)
value of DG produces less power loss and if the size
of DG is greater than before, the losses will
Where 𝑅1 and 𝑅2 are the random functions and gradually increase. The output of active and reactive
𝐶1, 𝐶2 are the training coefficients. ω is the inertia power of each DG unit must be selected in-between
weight factor. ω can be obtained from the following its minimum and maximum values which are
Eq. (2). expressed in Eq. (9) and (10).

International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 293

Figure. 1 Block diagram for allocation of DG

𝑃𝑚𝑖𝑛 𝐷𝐺,𝑖 ≤ 𝑃𝐷𝑔,𝑖 ≤ 𝑃𝑚𝑎𝑥 𝐷𝐺,𝑖 (9)

3.2 BPSO reconfiguration
𝑄𝐺,𝑚𝑖𝑛 ≤ 𝑄𝐺 ≤ 𝑄𝐺,𝑚𝑎𝑥 (10) The initial progress is completely based on
continuous valued search spaces. BPSO is used for
Where 𝑃𝐷𝑔,𝑖 and 𝑄𝐺 is nominal active and real value optimization issues once the real-binary
reactive power of DG, 𝑃𝑚𝑖𝑛 𝐷𝐺,𝑖 and 𝑃𝑚𝑎𝑥 𝐷𝐺,𝑖 are transformation process is completed. While, in the
minimum and maximum range of DG’s active Binary version of PSO, the following state of every
power, while 𝑄𝐺,𝑚𝑖𝑛 and 𝑄𝐺,𝑚𝑎𝑥 are minimum and particle is decided to conferring the agent’s affinity
maximum range of DG’s reactive power. So, the to be zero or one. The location of particles changed
size of distribution network in terms of the load will by selecting the appropriate mutation bits. Then the
plays an essential part in choosing the size of DG. particle may perceive to transfer near and far corners
Likewise, the position of DG also plays a key role in of the hypercube by flipping bits. One of the main
reducing the losses. The block diagram for the issues to be addressed in the establishment of BPSO
optimal placement in IEEE 69-bus system is shown is how to interpret the velocity of a binary vector.
in below Fig. 1. BPSO methodology is used for solving the DG
Step 1: Initialize the process with common allocation problem. In this context, the procedure
control parameters that present inside the algorithm. stated above is not used because of the level of
Step 2: Read the line data and bus data for IEEE randomness. Due to this randomness level, the
69 bus radial distribution system. constraint equation becomes very problematic, while
Step 3: Initially, the random particle generation running the program. To deal with the difficulty,
will occur in that data. For that, load flow analysis is BPSO selects the binary values 0 and 1 as a range of
checked. particles. Then the velocity is considered in the
Step 4: Then, run the fitness function (Power normal process; BPSO is a velocity limited process
loss, Voltage profile and Cost) of this particular data. and it makes sure the resolutions which do not fall
Step 5: From the data, find the best fitness into the local optima. BPSO is used to replicate the
values that are given to system data, which will be mapping from particle velocity, probability should
processed again for the next iteration. be selected, and it should confirm the outcome to be
Step 6: While considering the algorithm (with global optimal one.
reconfiguration only), same load flow analysis is
checked with a proposed methodology to find the 3.3 Flowchart for reconfiguration with BPSO
best fitness values.
BPSO used in the reconfiguration of the
Step 7: The fitness values are to be calculated
distribution system and it decides the operation of
for random location for placing DG, and it controls
the switches. The main problems of this system is to
the real power and reactive power values.
select the sectionalizing switches. Moreover, this
Step 8: From the best values, DG will be
problem lies in two states, either open or close the
optimally located with the help of the proposed
sectionalizing switches. The binary values of 0 and
methodology, and multi-objectives are evaluated
1 signifies the opening and closing of switches
with proper placement of DG.

International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 294

𝑟2 Set the maximum iteration to check each node in

the network for establishing the incidence matrix.
Fitness function calculated with the help of velocity
updation in randomized value updates particle
swarm coordinates. Like this, the BPSO method is
developed to simulate the IEEE 69 Bus distribution
network.

4. Result and discussion

The optimization techniques implemented in this
study have been analysed on standard radial
networks consist of IEEE 69 bus systems. Ever
since DGs are normally installed at only few number
of locations, the set of exploratory solutions are
obtained from each candidate solution which is not
very large. However, some of the exploratory
solutions may violate problem constraints and thus
become infeasible. These issues are corrected by
using BPSO algorithm. BPSO search algorithm used
in searching within the group to improve search
efficiency and avoid early maturing. In the proposed
method, two major parameters of each DG
determined using BPSO method. These major
parameters are: size and location of each DG. BPSO
identifies the test system through load flow, optimal
DG placement and DG rating. The simulation results
show that BPSO for the distribution network with
DG, which can not only reduce the network loss and
support system voltage but also improve the stability
of the system. BPSO is suitable to search for the
best switch combination of distribution network
with DG. The simultaneous optimal placement of
DG and optimal DG rating improve the power
quality in terms of loss reduction and enhanced
voltage profile have been compared and discussed.
The effectiveness of the proposed BPSO method is
investigated on 69-bus test distribution system. This
Figure. 2 Flowchart for the reconfiguration of power system has 68 sectionalizing lines, 5 tie-lines and 8
distribution system feeders. The normally open switches are 69-73. The
proposed method is programmed in MATLAB
respectively. The states of switch designated by software. In the simulation of network, four
binary vector. BPSO explored the binary vector that scenarios are considered to analyse the superiority
gives a solution for solving the reconfiguration of the proposed technique. The simulated results are
issues when compared to other process. The compared with the results of other techniques to
flowchart for reconfiguration of power distribution assess the performance and effectiveness of the
system is given in Fig. 2. proposed technique.
Several swarm particles are exploited to get the Scenario I: The system with reconfiguration
resolution of the given issue. In the proposed only;
method, BPSO procedure primarily creates the Scenario II: The system with DG units only;
randomized velocity by loading the swarm values Scenario III: The system with reconfiguration
that have the dimensional matrix and it’s treated as and installation of single DG unit;
swarms. In the next, zero padding for the variable Scenario IV: The system with reconfiguration
sig with the number of particles and dimensions, and installation of multi DG units;
which created the randomized matrix called 𝑟1 and

International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 295

Figure. 3 Reconfiguration of power distribution system for first scenario

For considering scenario 1, the test bus system is Table 1. Results for first scenario
carried out with only reconfiguration without the Simulation Results of 69 Bus Distribution Network
inclusion of DG units. The results for first scenario Items (scenario BEFORE AFTER
is shown in Fig. 3 and the values are tabulated in 1) Reconfiguration Reconfiguration
Table 1. Tie switches 69 70 71 72 73 14 56 61 69 70
For considering scenario 2, the test bus system is Power loss 224.9804 kW 98.5952 kW
Power loss ------------- 56.1761 %
carried out with only the inclusion of DG unit. The
reduction
results for second scenario is shown in Table 2.
Minimum 0.90919 pu 0.94947 pu
The results of second scenario are shown in Fig. voltage:
4 and Table 2 represents the optimal DG size
without the reconfiguration for 69-bus distribution Table 2. Results for second scenario
system. It can be concluded that optimal DG Simulation Results of 69 Bus Distribution Network
allocation with a size of 0.4 MW harnessing a Items (scenario BEFORE DG AFTER DG
reduction in the total real power loss from 224.6 to 2)
82.1119 kW that shows the 63.4886% of overall Tie switches 69 70 71 72 72 69 70 71 72 73
reduction. The percentage improvements in power Power loss 224.9804 kW 82.1119 kW
losses from the base case show better results. Power loss ------------- 63.4886 %
For considering scenario 3, the test bus system is reduction
carried out with reconfiguration and the inclusion of Size of DG 0.1018 MW 4 KW
DG unit. The results for third scenario is shown in Minimum 0.9677 0.94693
Table 3. voltage:
Table 3 represents the optimal DG size with
reconfiguration for the IEEE 69-bus distribution Table 3. Results for third scenario
Simulation Results of 69 Bus Distribution Network
system. It can be concluded that bus 39 is the best
Items (scenario BEFORE AFTER
bus for optimal DG allocation with a size of 0.4 3) Reconfiguration Reconfiguration
MW that produces reduction in the total real power with DG with DG
loss from 224.6 to 46.9193 kW that shows the Tie switches 69 70 71 72 73 40 60 5 30 6
79.1372% of overall reduction. The percentage Power loss 224.9804 kW 46.9193 kW
improvements in power losses from the base case Power loss ------------- 79.1372 %
show better results. However, the total real power reduction
loss is reduced more due to the availability of Minimum 0.90919 pu 0.94693
reactive power generation source at load locations. voltage:
Size (location of 4 KW 4 KW (39)
DG)

International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 296

Figure. 4 Reconfiguration of power distribution system for second scenario

Figure. 5 Reconfiguration of power distribution system for fourth scenario

Table 4. Results for fourth scenario

multi DG units. The results for fourth scenario is
Simulation Results of 69 Bus Distribution Network shown in Table 4.
Items (scenario BEFORE AFTER Table 4 represents the optimal DG size with
4) Reconfiguration Reconfiguration reconfiguration for the IEEE 69-bus distribution
with DGs with DGs system. Fig. 5 shows the voltage profile for the
Tie switches 69 70 71 72 73 17 14 48 13 36 fourth scenario with reconfiguration technique. It
Power loss 224.9804 kW 35.9239 kW can be concluded that bus 21, 32, 63 is the best bus
Power loss ------------- 84.0263 % for optimal DG allocation with a size of 0.4 MW
reduction which harnessing a reduction in the total real power
Minimum 0.90919 pu 0.95907 pu loss from 224.6 to 35.9239 kW which shows
voltage: 84.0263% of overall reduction. The percentage
Size (location of 4 KW 4 KW (21 32 63) improvements in power losses from the base case
DG)
show better results that are shown in comparison
Table 5 and 6.
For considering scenario 4, the test bus system is From the above results, better voltage regulation
carried out with reconfiguration and the inclusion of

International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 297

Table 5. Comparison table for all scenarios

Base case Existing Methods [22] BPSO Algorithm
Scenario 1 (Only Tie switch = 69 18 13 56 61 Tie switch = 14 56 61 69 70
Reconfiguration) Power loss = 99.35 Power loss = 98.5952
Power loss reduction = 55.85 % Power loss reduction = 56.1761 %
Min voltage = 0.9428 Min voltage = 0.94947
Scenario 2 (Only DG) Tie switch = 69 70 71 72 72 Tie switch = 69 70 71 72 73
Power loss = 86.77 Power loss = 82.1119
Power loss reduction = 61.43 % Power loss reduction = 63.488 %
Min voltage = 0.9697 Min voltage = 0.9494
Scenario 3 Tie switch = 69 18 13 56 61 Tie switch = 40 60 5 30 6
(Reconfiguration with Power loss = 51.30 Power loss = 46.9193 kW
single DG) Power loss reduction = 77.2 % Power loss reduction = 79.137 %
Min voltage = 0.9619 Min voltage = 0.94693
Scenario 4 Tie switch = 69 17 13 58 61 Tie switch = 17 14 48 13 36
(Reconfiguration with multi Power loss = 40.30 Power loss = 35.9239 kW
DGs) Power loss reduction = 82.08 % Power loss reduction = 84.026 %
Min voltage = 0.9736 Min voltage = 0.95907
DG size (location) = 1.0666(61), 0.3525 DG size (location) = 0.4(21), 0.4(32),
(60), 0.4527 (62) 0.4(63)

Table 6. Comparison table for final scenario

Base case Genetic Algorithm with Stud Krill Herd Algorithm BPSO Algorithm
capacitor placement [23] (SKHA) [24]
Scenario 4 Power loss = 99.742 Power loss = 99.742 Power loss = 35.9239 kW
(Reconfiguration Min voltage = 0.95814 Power loss reduction = Power loss reduction = 84.026 %
with multi DGs) DG size (location) = 0.4(60), 69.0967 % Min voltage = 0.95907
0.4 (62), 0.4 (64) Min voltage = 0.95814 DG size (location) = 0.4(21),
DG size (location) = 0.4(32), 0.4(63)
1719.0677 [61], 370.8802
[17] 527.1736 [11]

and much reduction in power losses can be obtained achieve. The results showed that, in case of voltage
without encountering adverse problems in power problem, algorithm could improve voltage to a
system operation. In addition, BPSO is more fast better limit, while keeping the losses reduced.
and robust in solving radial distribution systems
especially large ones. The advantage of this method 5. Conclusion
over other methods is that the sub-problems can be
In this research work, BPSO algorithm is applied
solved using algorithms that can deal with a very
for resolving optimal distribution network
large number of variables. Moreover, another
reconfiguration and optimal DG employment with
benefit of this algorithm is that it updates the
the objective of decreasing power loss and
multipliers adaptively whose outcomes have fast
enhancing voltage profile. At first, the algorithm is
solution and more precise. These advantages have
exercised for resolving optimal DG placement that
been proven through application to a large test
assesses optimal allocation for connecting DG and
system with a large number of variables. The
their significance for solving reconfiguration is a
applied methodology in this research has attained
transient manner. The obtained outcomes clearly
much reduction in real power losses and better
indicate that scenario IV (Multi DG installation with
improvement in voltage profile. The Table 5 shows
network reconfiguration) is found to be more
the comparison tables for all cases along with
effective in reducing the power loss from 224.9804
existing systems. Table 6 shows the comparison for
to 35.9239 KW with the reduction of 84.026%
final scenario. The performance and efficiency of
compared to the other scenarios. Moreover, the
distribution network reconfiguration are largely
voltage profile of the system improved from
dependent on an efficient search algorithm. BPSO is
0.95814 to 0.95907 while inserting multi DG units
a swarm intelligence optimization algorithm. The
with 0.4 MW size. To demonstrate the efficiency,
principle of BPSO is simple, robust and easy to
International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 298

BPSO algorithm is compared with genetic algorithm [8] M. M. Aman, G. B. Jasmon, A. H. A. Bakar,
(GA) and harmony search algorithm (HAS). In view and H. Mokhlis, “A new approach for optimum
of the uncomplicatedness and rapid presentation of simultaneous multi-DG distributed generation
this BPSO algorithm, one can use it in the intelligent Units placement and sizing based on
systems with a number of buses in real time maximization of system loadability using
applications. In addition, this algorithm is valuable HPSO (hybrid particle swarm optimization)
for investigating remaining systems that support in algorithm”, Energy, Vol. 66, pp.202-215, 2014.
the development of the upcoming system, which is [9] B. Singh and B. J. Gyanish, “Impact
specifically appropriate for a large-scale real-world assessment of DG in distribution systems from
system. minimization of total real power loss viewpoint
by using optimal power flow
References algorithms”, Energy Reports, Vol. 4, pp.407-
417, 2018.
[1] I. B. Hamida, S. B. Salah, F. Msahli, and M. F.
[10] C. Yammani, S. Maheswarapu, and S. Matam,
Mimouni, “Optimal network reconfiguration and
“Multiobjective optimization for optimal
renewable DG integration considering time
placement and size of dg using shuffled frog
sequence variation in load and DGs”, Renewable
leaping algorithm”, Energy Procedia, Vol.14,
Energy, Vol.121, pp.66-80, 2018.
pp.990-995, 2012.
[2] P. Mehta, P. Bhatt, and V. Pandya, “Optimal
[11] A. M. Abd-Rabou, A. M. Soliman, and A. S.
selection of distributed generating units and its
Mokhtar, “Impact of DG different types on the
placement for voltage stability enhancement and
grid performance”, Journal of Electrical
energy loss minimization”, Ain Shams
Systems and Information Technology, Vol.2,
Engineering Journal, Vol.9, No.2, pp.187-201,
No.2, pp.149-160, 2015.
2015.
[12] D. Q. Hung, N. Mithulananthan, and Y. L.
[3] O. A. Zongo and A. Oonsivilai, “Optimal
Kwang, “Optimal placement of dispatchable
placement of distributed generator for power
and nondispatchable renewable DG units in
loss minimization and voltage stability
distribution networks for minimizing energy
improvement”, Energy Procedia, Vol.138,
loss”, International Journal of Electrical Power
pp.134-139, 2017.
& Energy Systems, Vol.55, pp.179-186, 2014.
[4] U. Sultana, A. B. Khairuddin, M. M. Aman, A.
[13] B. Mohanty and S. Tripathy, “A teaching
S. Mokhtar, and N. Zareen, “A review of
learning based optimization technique for
optimum DG placement based on minimization
optimal location and size of DG in distribution
of power losses and voltage stability
network”, Journal of Electrical Systems and
enhancement of distribution system”, Renewable
Information Technology, Vol.3, No.1, pp.33-44,
and Sustainable Energy Reviews, Vol. 63, pp.
2016.
363-378, 2016.
[14] S. A. ChithraDevi, L. Lakshminarasimman, and
[5] H. Doagou-Mojarrad, G. B. Gharehpetian, H.
R. Balamurugan, “Stud Krill herd Algorithm
Rastegar, and J. Olamaei, “Optimal placement
for multiple DG placement and sizing in a
and sizing of DG (distributed generation) units
radial distribution system”, Engineering
in distribution networks by novel hybrid
Science and Technology, an International
evolutionary algorithm”, Energy, Vol.54, pp.
Journal, Vol.20, No.2, pp.748-759, 2017.
129-138, 2013.
[15] M. M. Aman, G. B. Jasmon, H. Mokhlis, and A.
[6] A. K. Singh and S. K. Parida, “Novel sensitivity
H. A. Bakar, “Optimal placement and sizing of
factors for DG placement based on loss
a DG based on a new power stability index and
reduction and voltage
line losses”, International Journal of Electrical
improvement”, International Journal of
Power & Energy Systems, Vol.43, No.1,
Electrical Power & Energy Systems, Vol.74
pp.1296-1304, 2012.
pp.453-456, 2016.
[16] S. Sahu, A. K. Barisal, and A. Kaudi, “Multi-
[7] R. Ishak, A. Mohamed, A. N. Abdalla, and M. Z.
objective optimal power flow with DG
C. Wanik, “Optimal placement and sizing of
placement using TLBO and MIPSO: A
distributed generators based on a novel MPSI
comparative study”, Energy Procedia, Vol.117,
index”, International Journal of Electrical
pp.236-243, 2017.
Power & Energy Systems, Vol.60, pp.389-398,
[17] S. Kaur, G. Kumbhar, and J. Sharma, “A
2014.
MINLP technique for optimal placement of
multiple DG units in distribution systems”,
International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28
Received: November 5, 2018 299

International Journal of Electrical Power &

Energy Systems, Vol.63, pp.609-617, 2014.
[18] D. B. Prakash and C. Lakshminarayana,
“Multiple DG placements in distribution
system for power loss reduction using PSO
Algorithm”, Procedia Technology, Vol.25,
pp.785-792, 2016.
[19] P. Chiradeja, S. Yoomak, and A. Ngaopitakkul,
“Optimal Allocation of Multi-DG on
Distribution System Reliability and Power
Losses Using Differential Evolution
Algorithm”, Energy Procedia, Vol.141, pp.512
-516, 2017.
[20] B. Das, V. Mukherjee, and D. Das, “DG
placement in radial distribution network by
symbiotic organisms search algorithm for real
power loss minimization”, Applied Soft
Computing, Vol.49, pp.920-936, 2016.
[21] P. V. Babu and S. P. Singh, “Optimal
Placement of DG in Distribution network for
Power loss minimization using NLP & PLS
Technique”, Energy Procedia, Vol.90, pp.441-
454, 2016.
[22] R. S. Rao, K. Ravindra, K. Satish, and S. V. L.
Narasimham, “Power loss minimization in
distribution system using network
reconfiguration in the presence of distributed
generation”, IEEE Transactions on Power
Systems, Vol.28, No.1, pp.317-325, 2013.
[23] B. J. Rani and A. S. Reddy, “Optimal
Distribution Planning with Finite Value Dg
Placement: A GA Based
Approach”, International Journal of Applied
Engineering Research, Vol.10, No.44, pp.
30932-30937, 2015.
[24] S.S. ChithraDevi, L. Lakshminarasimman, and
R. Balamurugan, “Stud Krill herd Algorithm
for multiple DG placement and sizing in a
radial distribution system”, Engineering
Science and Technology, an International
Journal, Vol.20, No.2, pp.748-759, 2017.

International Journal of Intelligent Engineering and Systems, Vol.12, No.1, 2019 DOI: 10.22266/ijies2019.0228.28