2016 International Conference on Electrical Power and Energy Systems (ICEPES)

Maulana Azad National Institute of Technology, Bhopal, India. Dec 14-16, 2016

Real Power Loss Reduction in Distribution Network

through Distributed Generation Integration by
Implementing SPSO
Sunaina Saini1 Gagandeep Kaur2
Department of Electrical Engineering Department of Electrical Engineering
I. K. Gujral Punjab Technical University I. K. Gujral Punjab Technical University
Jalandhar, India Jalandhar, India
[email protected] [email protected]

Abstract— Renewable sources based Distributed Generation researchers have applied different optimization techniques to
(DG) are placed in the vicinity of either customers or the load analyze and demonstrate the DG placement and rating
centers with the intention of minimization of the network power problem.
losses, improve voltage profile and recover substation capacity
delivery. Thus, while placing DG units caution has to be taken so The incomes of the power companies as well as the power
as to maximize the benefits. The current work presented in this transfer expenses are significantly affected by the real power
paper describes several techniques to get optimal siting of DG losses. A possible way to reduce these losses is the integration
units and their preliminary sizes to reduce power loss that of DG units in the distribution networks [5].
enhance the reliability of power distribution system. Optimal
sizing of DG is carried out using PSO and SPSO with arbitrarily The deregulation and restructuring market in some nations
generation of DG sizes. Here, objective function is taken to conveys novel discernments for DG with utilization of
reduce the real losses of distribution network. An analysis of renewable energy resources with little evaluations. To analyze
system operation has been done on IEEE-12 and 33 test systems the multi-objective optimization problem for the reduction of
and results obtained express the efficiency of proposed voltage error at DG and pilot buses, a technique called Optimal
approaches. The distinct approaches have been implemented on Coordinated Voltage Control (OCVC) was suggested by J. R.
standard networks in MATLAB®, in which SPSO diminishes Castro et al. [6]. N. Mohandas et al. suggested a multi
real power losses up to 49.587 for 12-bus network while 48.09% objective performance index (MOPI) for the betterment of
for 33-bus network. voltage stability of the distribution network under various
operating limits through Chaotic Artificial Bee Colony
Keywords— optimal placement; distributed generation (DG); (CABC) [7]. Other methods were also applied by P. S. Babu et
distribution network; real power loss; particle swarm optimization al. like fuzzy approach and Harmonic Search Algorithm (HSA)
(PSO); selective particle swarm optimization (SPSO) for DG placements in distribution network for the minimization
of power loss and voltage profile improvement [8]. E. K.
I. INTRODUCTION Bindumol et al. used a fast load flow method to obtain real
The power system is conventionally composed of power losses and voltage profile [9]. For this, backward
centralized power plant that foster a complex interconnected forward sweep algorithm (BFSA) was carried out by setting up
high voltage transmission and related distribution networks. different sizes of DG units at various sites in distribution
This network is utilized to supply power to the customer sites network. R. S. Al Abri et al. illustrated a scheme to make
present at large distances [1]. With the recent increase in the available a solution for DG location and rating, by choosing
power demand there is a need to deregulate and restructure the candidate buses where DG units are installed. In this method,
electric generating units. This can be viewed as a prospect for candidate buses sensitive to voltage profile were prioritized and
the emergence of discrete sources like distributed generators consequently, voltage stability margin was improved [10].
(DG) and sophisticated operating bases to provide an ample Kalman filter algorithm and power loss sensitivity calculation
and consistent supply of power. The betterment in the power was used to find a more authentic method for the optimal
systems is promoting the implementation of electricity solution of DG placement and size based on renewable
generation and storage at the site of distribution. Application resources. Optimal locator index (OLI) was developed to place
of DG and load management, both are important factors that DG units [11]. Other important techniques to prioritize the
contribute to consistent supply of power [2]. placement of DG units for minimization of power consumption
during power shortage are Continuous load flow and Modal
The level of incursion of DG units is the determining factor analysis [12]. Reconfiguration of the network in the presence
which minimizes the challenges in power system. The location
of DG units is mostly at the vicinity of load centers in the
power distribution systems. A suitable incursion of DG units
will influence the network in the desired way [3]. Several

978-1-5090-2476-6/16/$31.00 ©2016 IEEE 35

of DG units may also be considered to check the system The optimization of DG location and sizing problem can be
performance on power reliability. Sensitivity analysis has also defined by various objective functions involving power quality
been done to identify optimal positioning of DG units [13]. An improvement, techno-economic profits, power loss reduction,
analytical algebraic approach to establish optimal solution for increment of power supply reliability, voltage profile
consistently distributed load demand was demonstrated so as to improvement, etc. [12]. For DG placement, the objective
enhance network voltage profile [14]. function has been taken which minimizes network losses
II. POWER LOSS A. Power Loss Reduction Index (PLRI)
The real power loss in the system varies due to the flow of PRLI objective function has been used to evaluate the
time varying line current and line resistance. The capacitance reduction level of network power loss by integrating DG in
of the line and inductance of the line are notable as a result of distribution network for minimization of system losses. PLRI
production of magnetic and electric field by network current. can be defined by the following equation:
Reactive power is generated as the current flows through these
ܲ‫ ܹ ܮ‬Ȁ‫ܩܦ‬
elements which do not produce heat as much by resistance. ܲ‫ ܫܴܮ‬ൌ (4)
ܲ‫ ܱܹ ܮ‬Ȁ‫ܩܦ‬
Reactive power flows in the network aids to further power
losses in summation to real power losses discussed earlier.
Where, ܲ‫ܹܮ‬Ȁ‫ ܩܦ‬is power loss with DG integration.
The real power loss can be defined by:
ܲ‫ ܱܹܮ‬Ȁ‫ ܩܦ‬is power loss without DG.
݈ܲ‫ ݏݏ݋‬ǡ݅ ൌ σܰ‫ܤ‬ ʹ
݅ൌͳ ͵ ൈ ‫ ݅ܫ‬ൈ ܴ݅ (1) B. Objective Function
ܵ There are number of objective functions for DG unit
Where,‫ ݅ܫ‬ൌ ሺ ݅ ሻ‫כ‬ (2) integration in power distribution network that alternatively
ܸ݅
ܵ݅ ൌ  ܲ݅ ൅ ݆ܳ݅ (3) improve power reliability of system.In this section, objective
function contains PLRI which has to be minimized. The
NBare the number of network branches objective function can be represented as:
ܵ݅ is the total apparent power of the network for i-th branch
ܲ݅ is the total real power of the network for i-thbranch ‫ ܨ‬ൌ ݉݅݊݅݉݅‫݊݋݅ݐܽݖ‬ሺܹͳ ݂ͳ ሻ (5)
ܳ݅ is the total reactive power of the network for i-thbranch
‫ ݅ܫ‬is the line current of network for branch i Where, ݂ͳ ൌ ݉ܽ‫݉ݑ݉݅ݔ‬ሺܲ‫ܫܴܮ‬ሻ
ܴ݅ is the line resistance of the network for branch i
݈ܲ‫ ݏݏ݋‬ǡ݅ is the real power loss for i-thbranch The ܹͳ is the weight having range 0-1, here 0.5 weight has
been taken for the ease of calculation. And, ݂ͳ is objective
In addition of DG unit in distribution network,real power function, where ݂ͳ is the maximum value of PLRI which has
to be minimized. Therefore, the overall objective function ‘F’
losses are diminished at a certain level because power loss at
has to be optimized to determine optimal solution. The optimal
upstream portion has already reduced due to DG integration.
DG siting & sizing has been determined on that candidate bus
III. PROBLEM FORMULATION where overall distribution network power loss is found to be
minimized. Consequently, for this DG size and location, PLRI
The existence of DG is considerable challenge in is calculated for each branch.
stipulations of practical, technical and network protection
issues in distribution system. Therefore, techno-economic IV. COMPUTATIONAL METHODOLOGY
impact of DG is difficult to examine in power distribution
network. Thus, DG should be connected in distribution A. Particle Swarm Optimization (PSO)
networks in such a way that it evades deprivation of power The basic Particle Swarm Optimization (PSO) is a
reliability, stability and quality. The estimation of techno- populace based stochastic search evolutional implementation
economic impacts of DG in distribution network is a serious methodology, grouping of evolutionary programming and
issue. Inappropriate placement of DG in terms of its position genetic algorithm that gives optimized solution. Kennedy and
and capability may bring about increase in voltage variations, Eberhart primarily discovered this method by observing
fault currents, intervenes the voltage control activities, behavior of group of fishes and flock of birds. It was
maximize or minimize system losses, raise the capital and introduced by analyzing a simplified communal system, and
operating cost of equipment’s used for power protection. The has been originated to be vigorous for non-linear optimization
geographical topology, environmental aspects, political problems. In PSO algorithm, each element called particle
influences, solar radiation etc. affect the allocation of
present in n-dimensional search space and transfer from their
renewable sources based DG units which can conceivably lead
places in multi-dimensional search spaces with a certain
to possible risks in optimizing the DGunit allocating and sizing
in power distribution network planning. The time-varying velocity on the basis of their knowledge with respect to time.
characteristics of electricity demand are the chief motive for Each particle has sufficient knowledge and updates its
DG unit penetrations in electrical distribution networks. direction towards the best position corresponding to its own
Therefore, adequate placement and capacity of DG units ability, called as pbest, and overall best position equivalent to
should be determined for stable operation of power system. the transfer of location of neighboring particles called as

36
gbest. The particles modify their location according to best ‫ݐ݆݅ݒ‬൅ͳ ൌ ‫ ݐ݆݅ݒ‬൅ ܿͳ ‫ ݐ݆ͳݎ‬൫ܾܲ݁‫ݐݏ‬
‫ݐ‬ ‫ݐ‬ ‫ݐ‬
ǡ݅ െ ‫ ݆݅ݔ‬൯ ൅ ܿʹ ‫ ݆ʹݎ‬ሺ‫ݐݏܾ݁ܩ‬
position confronted by them and their neighbors [16].
The particle position in n-dimensional search space can be െ ‫ ݐ݆݅ݔ‬ሻ
written as: Step-6: If new value is improved than pbest value
ܼ݉  ൌ ሺ‫ ͳ݉ݖ‬ǡ ‫ ʹ݉ݖ‬ǡ ǥ ǥ Ǥ Ǥ ǡ ‫ ݊݉ݖ‬ሻ (6) than pbest value is modified as new value and pbest
location is modified as new location.
The current position of particle can be calculated as: Step-7: Compare the estimated best new fitness with
݇൅ͳ ݇ ݇൅ͳ
the gbest of population. If new value is improved
ܵ݅‫ܦ‬ ൌ  ܵ݅‫ܦ‬ ൅ ܵ݅‫ܦ‬ ݅ ൌ ͳǡʹǡ ǥ ǡ ݊ǡ than gbest. Then, gbest value is modified as best new
‫ ܦ‬ൌ ͳǡ ʹǡ ǥ ǥ Ǥ ǡ ݉ (7) fitness value.
Step-8: Check the maximum number of evaluations
Where ܵ ݇ is the present particle position and ܵ ݇൅ͳ the new is met if not go to the step 5.
position. Step-9: Gbest value is the optimal value of DG size.
Calculate PLRI for corresponding DG location according to
The particle velocity in n-dimensional search space is: DG size.

ܸ݉  ൌ ሺ‫ ͳ݉ݒ‬ǡ ‫ ʹ݉ݒ‬ǡ ǥ ǥ Ǥ Ǥ ǡ ‫ ݊݉ݒ‬ሻ (8) The DG unit constraints are fixed (0, 4) when PSO technique
is applied in MATLAB. The DG values are selected as
The updated velocity of particle is shown by the following particles to be optimized to minimize the network power loss.
equation:
B. Selective Particle Swarm Optimization (SPSO)
݇ ൅ͳ
‫ܦ݅ݒ‬ ݇
ൌ  ‫ ݅ݓ‬ൈ ‫ܦ݅ݒ‬ ݇
൅ ܿͳ ൈ ‫ ݀݊ܽݎ‬ൈ ൫‫ ܦ݅ݐݏܾ݁݌‬െ ܵ݅‫ܦ‬ ൯ ൅ ܿʹ ൈ Currently, a variation came in PSO technique which
‫ ݀݊ܽݎ‬ൈ ሺܾ݃݁‫ ܦ݅ݐݏ‬െ ܵ݅‫ ܦ‬ሻ݇
(9) recommends advance versions of improved PSO technique
that have many advantages. The convergence speed increment,
Where ‫ ݇ ݒ‬൅ͳ is the updated velocity; ‫ ݇ ݒ‬is the present best optimal solution provident, applications widespread etc.
velocity; pbest shows the individual best velocity; gbest are the benefits of advanced PSO techniques. The binary PSO
represents the best velocity over all participated particles; n is (BPSO) is also a novel version of PSO by modulating search
the number of particles in a populace; m is the number of space to be binary, which is a discrete version of PSO, has
participants in a particle; rand is the random values originated; been proposed by Kennedy and Eberhart [18]. A sigmoid
ܿ݅ acceleration coefficient of particle; ‫ ݅ݓ‬is the inertia weight transformation is operated to restraint the velocity components
for every participated particle. in a prescribed limits [0, 1] and particles constraint (0, 1) in
BPSO.
݇ ൅ͳ ͳ
The inertia weight is given by the following equation: ‫݃݅ݏ‬൫‫ܦ݅ݒ‬ ൯ൌ ݇ ൅ͳ (11)
ͳ൅‡š’ ൫െ‫ ܦ݅ݒ‬൯

‫ ݔܽ݉ ݓ‬െ‫݊݅݉ ݓ‬
‫ ݓ‬ൌ ൈ ݇ (10) ݇ ൅ͳ
ͳǡ ݂݅ߪ ൏ ‫݃݅ݏ‬൫‫ܦ݅ݒ‬ ൯

݇ ݉ܽ‫ݔ‬ ݇൅ͳ
Where ‫ ݔܽ݉ݓ‬is the maximum and ‫ ݊݅݉ݓ‬is the minimum inertia ‫ܦ݅ݔ‬ ൌቊ ݇ ൅ͳ
(12)
Ͳǡ ݂݅ߪ ൒ ‫݃݅ݏ‬൫‫ܦ݅ݒ‬ ൯
weights; k is the current inertia weight and ݇݉ܽ‫ ݔ‬is
themaximum iterations. The appropriate value of ܿͳ and ܿʹ is An advanced version of BPSO is SPSO was developed by
2 and the range randomize between 1 & 2. The values of ‫ݔܽ݉ݓ‬ Khalil and Gorpinich [19], where search space is undertaken
and ‫ ݊݅݉ݓ‬have been chosen by hit and trial method [17].The as selected space by approving sigmoid transformation of
steps given below demonstrate DG optimal placement and velocity function and upsurge the superiority of solution. In
sizing in distribution system through PSO. SPSO technique, a set of DN positions for search space is
ܵ‫ ܦ‬ൌ ሾܵ‫ ͳܦ‬ǡ ܵ‫ ʹܦ‬ǡ ǥ ǥ ǡ ܵ‫ ܰܦ‬ሿ for each D dimension, where DN
Step-1: Select the parameters which have to be are the selected positions in dimension D.
optimized through PSO. ݇ ൅ͳ
‫݃݅ݏ‬൫‫ܦ݅ݒ‬ ൯ ൌ ‫ܰܦ‬
ͳ
(13)
݇ ൅ͳ
ͳ൅‡š’ ൫െ‫ ܦ݅ݒ‬൯
Step-2: Select swarm size.
Step-3: Randomly generate the DG size between (0, ݇ ൅ͳ
ܵ ݂݅‫݃݅ݏ‬൫‫ܦ݅ݒ‬ ൯൏ͳ
4). ‫ͳܦ ۓ‬
݇ ൅ͳ
Step-4: Assign pbest value to the initial particle. ۖ ܵ‫݃݅ݏ݂݅ ʹܦ‬൫‫ ܦ݅ݒ‬൯ ൏ ʹ
ۖ Ǥ
Step-5: Change the particle position and velocity ۖ
݇൅ͳ
‫ܦ݅ݔ‬ ൌ Ǥ
(14)
according to equations Ǥ
‫۔‬
‫ݐ݅ݔ‬൅ͳ ൌ ‫ ݐ݅ݔ‬൅ ‫ ݐ݅ݒ‬൅ͳ ۖ Ǥ
ۖ Ǥ
ۖ
݇ ൅ͳ
‫݃݅ݏ݂݅ ܰܦܵە‬൫‫ ܦ݅ݒ‬൯ ൏ ܰ

37
Where ܵ‫ ͳܦ‬, ܵ‫ ʹܦ‬,…..,ܵ‫ ܰܦ‬are the selected values for each dimension for IEEE-12, 33 bus systems. For a particular bus,
particle in dimension D. the DG sizes are arbitrarily created which should perform only
Velocity constraints ሾܸ݉݅݊ ǡ ܸ݉ܽ‫ ݔ‬ሿ can be calculated using in one dimension at a time. The SPSO can be introduced to
the following equation: determine optimal solution for improvement in electric power
݇ ൅ͳ
ܸ݉ܽ‫ ݔ‬ǡ ݂݅‫ܦ݅ݒ‬ ൐ ܸ݉ܽ‫ݔ‬ quality after detection of search space for every dimension.
‫ ܦ݅ݒ‬ൌ  ቐ‫ ܦ݅ݒ‬ǡ ݂݅ห‫ ܦ݅ݒ‬ห ൑ ܸ݉ܽ‫
ݔ‬
݇ ൅ͳ ݇ ൅ͳ ݇ ൅ͳ
(15) V. TEST RESULTS AND SIMULATIONS
݇ ൅ͳ
ܸ݉݅݊ ǡ ݂݅‫ܦ݅ݒ‬ ൏ ܸ݉݅݊ The proposed approaches have been widely studied on two
The consistency of i-th particle velocity value in a D- standard IEEE-networks consisting of 12 & 33 radial
dimension at the lower and upper limits can be avoided using distribution network. The DG size has been optimized through
the equation given below and drive the particles moving load flow by applying all approaches (presented here) in order
through the selected space. to diminish the objective function. Then, Power Loss
Reduction Index (PLII) is computed for optimal solution
‫ ݀݊ܽݎ‬ൈ ‫ܦ݅ݒ‬ ݇ ൅ͳ
ǡ ݂݅ห‫ܦ݅ݒ‬ ݇ ൅ͳ ݇
ห ൌ ห‫ܦ݅ݒ‬ ห
composing two cases which are without DG& with DG
݇ ൅ͳ
‫ܦ݅ݒ‬ ൌ  ቊ ݇ ൅ͳ (16) integration in distribution network. Here, comparisons of how
‫݁ݏ݅ݓݎ݄݁ݐ݋ ܦ݅ݒ‬ integration of optimal DG placement effect network real
The following steps show the SPSO algorithm for power losses are presented through operating proposed
obtaining DG optimal location and size: approaches for two test systems.
Step-1: Select swarm size and the parameters which A. 12-Bus System
has to be minimized using SPSO. The analysis has been performed on IEEE-12 bus distribution
Step-2: Initialize the values of particles of system to diminish network real power losses. The radial
acceleration constants ‘ܿͳ ’, ‘ܿʹ ’; iteration number ‘k’, network of 12-bus system is shown in figure 1. The system
and inertia weight ‘w’. voltage considered here is 12.66 kV, total real and reactive
Step-3: Initialize the values of particles (DG) in power losses before DG integration are 18.564 kW and
each D-dimension with the initial values of position 7.2149kVAR. The total apparent power loss is 19.916 kVA
and velocity. for the base case and minimum voltage profile is 0.99268 at
Step-4: Randomize the search space in each D- bus-12. The network for 12-bus distribution system is shown
dimension. by figure 1.
Step-5: Compute power flow for distribution
network.
Step-6: Checkradially constraints exit or not, if not,
then go to step 8.
Step-7: Compute the network power loss ܲ‫ ݏݏ݋ܮ‬ሺ݇ሻ. Fig.1. IEEE-12 Bus Distribution System
Step-8: Check ܲ‫ ݏݏ݋ܮ‬ሺ݇ ൅ ͳሻ is less than ܲ‫ ݏݏ݋ܮ‬ሺ݇ሻ.
Then Through implementing various approaches, optimal DG
ܲ‫ ݏݏ݋ܮ‬ሺ݇ ൅ ͳሻ ൌ ܲ‫ ݏݏ݋ܮ‬ሺ݇ሻ size has been obtained on that bus where overall real power
And, identify the equivalent bus number. loss is found to be minimized. For this corresponding size of
Step-9: Modify the values of velocity and position DG, power losses for every branch of distributed network are
of ݅ ‫ ݄ݐ‬particle for each D-dimension using equations calculated. Comparisons of power loss for different
(13), (14), (15), and (16). approaches are shown in figure 2.
Step-10: Increment the value of ‘k’ as
݇ ൌ݇൅ͳ
Step-11: Check the maximum number of iterations is
met. If not go to step 7.
Step-12: Print optimal DG size and bus location
Step-13: Obtain PLRI for equivalent DG location and
capacity.

There would signify the number of dimensions, categorize

search space for each and every dimension and selection of
optimal solution for current problem from search spaces by
applying SPSO to solve optimal placement and capacity of
DG problem in power distribution network. The network
candidate buses represent dimensions for a particular search
space. Here, standard IEEE-12, 33 radial networks are studied
Fig.2. Comparisons of power loss for different techniques with or without DG
for load flow analysis. Thus, numbers of candidate buses are integration
12 and 33thosecharacterize search space for a specific

38
The results after DG integration are demonstrated by table-I. power losses before DG integration are 210 kW and 143.0
kVAr. The total apparent power loss is 254.065 for the base
TABLE-I case minimum voltage profile is 0.90377 at bus-18.
RESULTSAFTER DG INTEGRATION (33-RADIAL DISTRIBUTION SYSTEM)
12-RADIAL BUS PSO SPSO
SYSTEM
Real Power Loss 10.5 kW 9.3585 kW
Optimum DG size & 3.143 MW 1.58MW
location Bus-6 Bus-6
Minimum Voltage 0.99767 0.99800
profile Bus 12 Bus 12
% Reduction in Loss 43.438% 49.587%

After computing power loss, PLRI also obtained which Fig.4. IEEE-33 Bus Distribution Network
shows how much real power losses are affected for the
presence of DG in distribution network by applying various By implementing various approaches, optimal DG size has
approaches. Figure 3 illustrates about PLRI with DG been obtained on that bus where overall real power loss is
integration for PSO, and SPSO. found to be minimized. For this corresponding size of DG,
power loss in distribution network has been calculated.
Comparisons of real power loss for different approaches are
displayed in figure 5.

Fig.3. PLRI for PSO, and SPSO

Fig.5. Comparisons of power loss for different techniques with base case
Then, function f is the minimum value of power loss
The results after DG integration are displayed in table-III.
evaluated for all branches of 12 bus system. The objective
function ‘F’ has to be minimized which tells about the
TABLE-III
minimum value of power loss on that particular branch. The RESULTSAFTER DG INTEGRATION (33-RADIAL DISTRIBUTION SYSTEM)
table-II demonstrates the values of objective functionf&‘F’and 33-RADIAL BUS PSO SPSO
function is minimized up to 0.43934through SPSO which SYSTEM
clearly shown by table given below: Real Power Loss 111.07 kW 109.79 kW
Optimum DG size & 2.6509 MW 2.5173 MW
TABLE-II location Bus-6 Bus-6
VALUESOF OBJECTIVE FUNCTIONFOR IEEE 12-TEST SYSTEM Minimum Voltage 0.94324 0.94863
OBJECTIVE PSO SPSO profile Bus 18 Bus 18
FUNCTION
VALUES % Reduction in Loss 47.10% 48.09%
f 0.99004 0.87869
‘F’ 0.49502 0.43934 Afterward power loss description, PLRI also computed
which expresses how much real power losses are affected for
It can be stated that SPSO gives best result. DG integration by applying distinct approaches. The PLRI has
B. 33-Bus System been evaluated using MATLAB is shown by the figure 6.
Then, function f is the minimum value of power loss evaluated
The analysis has been performed on IEEE-33 bus radial
for all branches of 33 bus system. The multi-objective
distribution network to diminish power loss. The radial
function ‘F’gives minimum value of real power loss. The
structure of 33-bus system is shown in figure 3. The system
table-6 demonstrates the values of objective function f&‘F’and
voltage considered here is 12.66 kV, total real and reactive

39
function is minimized up to 0.15499which clearly shown by REFERENCES
table IV. [1] A. S. Safigianni, G. N. Koutroumpezis, and V. C. Poulios, “Mixed
Distributed Generation Technologies in a medium Voltage Network”,
Electric power Systems Research 96, pp. 75-80, 2013.
[2] P. Chiradeja, “Benefit of Distributed Generation: A Line Loss Reduction
Analysis”, IEEE/PES Transmission and Distribution Conference &
Exhibition: Asia and Pacific Dalian, China, pp. 1-5, 2005.
[3] B. Poornazaryan, P. Karimyan, G. B. Gharepetian, and M. Abedi,
“Optimal Allocation and Sizing of DG Units Considering Voltage
Stability, Losses and Load Variations”, Electrical Power and Energy
Systems 79, pp. 42-52, 2016.
[4] R. B. Prada, and L. J. Souza, “Voltage Stability and Thermal Limit:
Constraints on the Maximum Loading of Electrical Energy Distribution
Feeders”, Proc. Inst. Electr. Eng-Gen. Transm. Distrib. Vol. 145, pp.
573-577, 1998.
[5] D. Q. Hung, N. Mithulananthan, R. C. Bansal, “Analytical Expressions
for DG Allocation in Primary Distribution Networks”, IEEE Trans
Energy Conv., Vol. 25, No. 3, pp. 814-820, 2010.
[6] J. R. Castro, M. Saad, S. Lefebvre, D. Asber, and L. Lenoir, “Optimal
Fig.8. PLRI for PSO and SPSO Voltage Control in Distribution Network in the Presence of DGs”,
Electrical Power and Energy Systems 78, pp. 239-247, 2016.
TABLE-IV [7] N. Mohandas, R. Balamurugan, and L. Lakshminarasimman, “Optimal
VALUESOF OBJECTIVE FUNCTION FOR IEEE 33-TEST SYSTEM Location and Sizing of Real Power DG Units to Improve the Voltage
Stability in the Distribution System using ABC Algorithm United with
OBJECTIVE PSO SPSO
Choas”, Electrical Power and Energy Systems 66, pp. 41-52, 2015.
FUNCTION
VALUES [8] P. S. Babu, and R. M. Mohan, “Optimal Performance Enhancement of
f 0.3299 0.30997 DG for Loss Reduction Using Fuzzy and Harmonic Search Algorithm”,
‘F’ 0.16252 0.15499 IEEE Conf., 2015.
[9] E. K. Bindumol, and C. A. Babu, “A Simple and Fast Load Flow
Algorithm for Sizing and Placement of DG in Radial Distribution
The above table shows SPSO approach gives best results which System”, IEEE Conf., 2015.
reduce ‘F’ up to 0.15499. [10] R. S. Al Abri, E. F. El-Saadany, Y. M. Atwa, “Optimal Placement and
Sizing Method to Improve the Voltage Stability Margin in a Distribution
C. Network Comparisons System Using Distributed Generation”, IEEE Trans on Power System,
The value of function ffor 12-bus test system for SPSO is Vol. 28, No. 1, pp. 326-334, Feb 2013.
0.87869; on the other hand 0.30997for 33-bus test system and [11] S. H. Lee, and J. W. Park, “Optimal Placement and Sizing of Multiple
for PSO 0.99004 for 12-bus system and conversely 0.3299 for DGs in a Practical Distribution System by Considering Power Loss”,
IEEE Trans. on Industry Applications, Vol. 49, No. 5, pp. 2262-2270,
33-bus system which shows optimization techniques give Sep/Oct 2013.
better result for large networks as compare to small one.For [12] M. Ettehadi, H. Ghasemi, and S. Vaez-Zadeh, “Voltage Stability-Based
33-bus system, value of ‘F’ for PSO is 0.16252 which is DG Placement in Distribution Networks”, IEEE Trans. on Power
minimum than for 12-bus system i.e., 0.49502.And for Delivery, Vol. 28, No. 1, pp. 171-178, Jan 2013.
SPSO‘F’ is 0.43934for 12-bus system has maximum value [13] R. S. Rao, K. Ravindra, K. Satish, and S. V. L. Narasimham, “Power
Loss Minimization in Distribution System Using Network
than 33-bus system i.e., 0.15499. Therefore applied techniques Reconfiguration in the Presence of distributed Generation”, IEEE Trans.
give best results for 33-bus network in terms of objective Power Systems, Vol. 28, No. 1, Feb 2013.
function contrariwise on the basis of system real power loss [14] K. M. Muttaqi, An D. T. Le, M. Negnevitsky, and G. Ledwich, “An
reduction, results are in favor of 12-bus network. Algebraic Approach for Determination of DG Parameters to Support
Voltage Profiles in Radial Distribution Networks”, IEEE Trans. on
VI. CONCLUSION Smart Grid, Vol. 5, No. 3, pp. 1351-1360, May 2014.
[15] N. Acharya, P. Mahat, and N. Mithulananthan, “An Analytical Approach
In this paper, usage of objective function accomplishes result for DG Allocation in Primary Distribution Network”, Electrical Power
of PLRI. PLRI is obtained for two approaches utilized here by and Energy Systems 28, pp. 669-678, 2006.
computing network real power losses for different cases and [16] J. Kennedy, and R. Eberhart, “Particle Swarm Optimization”,
used for 12-bus and 33-bus IEEE test networks. For 12-bus Proceedings of IEEE International Conference on Neural Networks
(ICNN’95), Vol. IV, pp.-1942-1948, Perth, Australia, November 29-
radial system, PSO diminished real power loss of distribution December 1, 1995.
network up to 43.438% and SPSO reduced up to 49.587% [17] R. C. Eberhart , and Y. Shi, “Comparing inertial weights and
whereas for 33-bus system, PSO minimized up to 47.10% and constriction factor in Particle Swarm Optimization”, Proceedings of
by implementing SPSO network losses are decreased up to International Congress on Evaluating Computation, pp. 84-88, San
48.09%.And, it can be concluded that if power loss is reduced Diego California , Piscataway, NJ: IEEE Service Center, 2000.
at a certain level then network voltage profile also improves [18] J. Kennedy, and R. Eberhart, “A Discrete Binary Version of the Particle
Swarm Algorithm”, IEEE International Conference on Systems, Man,
through inserting real DG unit in distribution network. It can and Cybernetics (SMC 97), Vol. 5, pp. 4104-4108, 1997.
also be observed that SPSO provides best result as relate to [19] T. M. Khalil, and A. V. Gorpinich, “Optimal Conductor Selection and
other approaches. Capacitor Placement for Loss Reduction of Radial Distribution Systems
by Selective Particle Swarm Optimization”, IEEE Conference, pp. 215-
220, 2012.

40