Adaptive PSO Technique for Optimal Placement

and Sizing of DG in 3-Phase Unbalanced Radial

Secondary Distribution System
Ponnam Venkata K Babu Dr.K. Swarnasri
Research Scholar Professor, EEE Department
Acharya Nagarjuna University RVR & JC College of Engineering
Nagarjuna Nagar, Guntur, A.P,India Chowdavaram, Guntur, A.P,India
kishore.ponnam@gmail.com swarnasrik@gmail.com

Abstract— Actual operating conditions of Secondary Radial into several segments like spider network. The social
Distribution System are unbalanced in nature and for stable Information sharing between the swarms are made strong and
operation of the system, the distribution system engineer should adaptive. Authors [5] proposed two methods, first one uses
account for operational imbalances within the system. With the the index vector method and the second one uses loss
continuing increase in load demand, future expansion of the
network is to be done basing on the load flow study of the
sensitivity factors for optimal size and sitting of capacitors in
distribution system network and thus, it is one of the most secondary distribution networks. In both methods, loss
important research fields for electrical engineering. Optimally reduction is same but the second method requires less
placed Distributed Generation (DG) units reduce system losses reactive power compensation. S. Sudhakar Reddy et.al [7]
and lead an improvement in the voltage profile, system proposes the optimal size and location of DG and kVAr
reliability, load ability, voltage stability, voltage security and compensation by using the PSO method on the 37 bus system.
power quality. There are several techniques available for the This will be increasing the loading capacity of the secondary
optimum size and placement of DGs in the unbalanced radial distribution system within the thermal limits and energy
secondary distribution system (URSDS). In this paper, in saving. The article [8] presents the optimal placement and
URSDS, optimal placement of DGs is determined by using
modified sensitivity factor method and optimal size is
size of multiple DGs by applying hybrid PSO & Honey Bee
determined by using Adaptive Particle Swarm Optimization Mating Optimization (HBMO) methods. This method
(APSO) technique. The proposed technique is tested on 25 & 37 precisely improves voltage fluctuations and minimizes power
bus URSDS. Results show that network power losses are losses. M.H. Moradi, M. Abedini [9] proposed novel
reduced and voltage profiles are improved in each phase with combined Genetic Algorithm (GA) for optimal location of
optimal allocation. DG and PSO for optimal size of DG to minimize the power
losses, embellish the voltage stability in the system. In [10]
Keywords— Secondary radial distribution system, H. Manafi et.al proposed both PSO and Differential
Unbalanced, Distributed Generation, PSO.
Evolution (DE) techniques for optimal place and size of the
I. INTRODUCTION DG. DE technique has good convergence rate and provided
the optimum results compared to PSO. In [11] Mohamed
Secondary Distribution system is a vital part of the Imran.A and Kowsalya.M used Loss sensitivity factor
electric power system since the reliability of electric power method to find the best locations for DG units. Bacterial
supply to the consumers is necessary. About 70% of losses Foraging Optimization Algorithm (BFOA) is used to find the
are occurred in the distribution system, while 30% losses optimal size of DG in order to minimize the operational costs,
occurred in the transmission system [14]. DG is a small scale network power losses and improving voltage stability. Banaja
power generation that is typically embedded within the Mohanty, Sasmita Tripathy [12] used Teaching Learning
secondary distribution system. By introducing DG units at Based Optimization (TLBO) method in order to find the
proper positions with the appropriate size distribution system optimum site and size of DG in the secondary distribution
losses can be reduced considerably. Main aim of the network. This method is tested against 33 and 69 bus systems
Distribution Generation is to minimize the losses, cost, and the results showed the good performance and
improve voltage profile, stability and to reduce greenhouse effectiveness of this method over Genetic algorithm and
gas emissions. Thus, DG placement and sizing are important Evolutionary methods. P.Vijay Babu and S.P. Singh [13]
factors in distribution systems. presented a two-stage procedure based on Loss Sensitivity
A simple but efficient 3-ϕ load flow algorithm for factor and analytical methods in order to find the location and
secondary radial distribution networks is proposed by size of the capacitors for optimizing the energy savings. This
Thukaram.D et.al [1]. This method uses forward-backward technique is applied to 15, 34 and 69 bus systems. Apart from
propagation to find line currents and voltages. In [2], PSO the energy savings, operation and installation costs are
technique is used for determining the optimal size for included in this paper. In [14], Ant Lion optimization method
capacitive compensation and loss sensitivity factors for is used to find the optimal DG in order to reduce the losses
determining the optimal location to improve the voltage and voltage profile improvement. This method is based on the
profile as well as to minimize the active power loss in the hunting nature of ant lions. The effects of DG are investigated
secondary distribution system. Md. Sakhawat Hossen et.al on the 37-bus unbalanced system in [15]. In this paper, the
[4] proposed a new evolutionary optimization based on PSO size of the DGs obtained based on the ambiguity methods and
technique called adaptive particle swarm optimization the location is found by using weighted multi-objective
technique. In this technique, the given search space is split Improved Particle Swarm Optimization technique.

978-1-7281-1423-1/XX/$XX.00 ©2019 IEEE

 In the present work, modified sensitivity factor iii. Form pathline vectors.
method is proposed for 3-ϕ unbalanced radial distribution iv. Calculate Zline matrix. The diagonal elements of Zline
networks in order to decide the optimal placement of DG and matrix can be calculated by adding primitive
Adaptive PSO technique is applied for obtaining DG capacity impedances of the lines connecting the reference bus
to optimize the total real power losses and enhance the to any preferred bus. The off-diagonal elements are
voltage profile of the secondary radial distribution system. calculated by adding the primitive impedances of the
The paper is organized as follows: Suggested Power lines which appears common to the lines of m & n
Flow Algorithm is presented in Section II. Problem buses from the reference bus.
formulation is given in Section III. Modified sensitivity 𝑎 𝑏
analysis and loss sensitivity factors method are shown in v. Initialize the Bus Voltages, 𝑣𝑚 = 1 + 𝑗0 , 𝑣𝑚 =
𝑐
Section IV. Adaptive PSO technique is presented in Section −0.5 − 𝑗0.866, 𝑣𝑚 = −0.5 + 𝑗0.866 pu.
V. The proposed Algorithm for DG placement and sizing is vi. Calculate the current injections at node ‘m’ by using
given in Section VI. Results on the test systems are illustrated (𝑃 + 𝑗𝑄 )∗
𝐼𝑚 = 𝑚 𝑚 .
in Section VII. 𝑉𝑚

vii. Set iteration count p =1.

II. PROBLEM FORMULATION
𝑜𝑙𝑑
The primary goal of optimal DG size and placement viii. Assign 𝐼𝑚 = 𝐼𝑚 for all the buses.
problem in URSDS is to optimize the power loss while ix. Calculate the voltage drop in each branch by
satisfying all network constraints. Mathematical expressions multiplying the Zline matrix with the current injection
for real power loss and operational constraints are given matrix.
below.
x. Update the bus voltages for all the 3-phases by
A. Real Power Loss subtracting the voltage drop from the previous bus
The main advantage of optimal placing of DG in URSDS voltage.
is to decrease the power losses and it is the main objective xi. Compute current injections Im with the updated bus
function. It can be written Mathematically as: voltages.
2 2
[𝑖]+𝑄𝑒𝑓𝑓𝑎 [𝑖])∗𝑅𝑎𝑎𝑖
(𝑃𝑒𝑓𝑓𝑎 𝑝+1 𝑝
P𝑙𝑜𝑠𝑠 (i) = ∑( 2 + xii. If max ( | 𝐼𝑚 | − |𝐼𝑚 | ) > Prescribed value, then
𝑉𝑖𝑎
2 2 2 2 increase the iteration count by one and go to step viii.
(𝑃𝑒𝑓𝑓𝑏 [𝑖]+𝑄𝑒𝑓𝑓𝑏 [𝑖])∗𝑅𝑎𝑏𝑖 (𝑃𝑒𝑓𝑓𝑐 [𝑖]+𝑄𝑒𝑓𝑓𝑐 [𝑖])∗𝑅𝑐𝑎𝑖
2 + 2 ) ----(1) xiii. Calculate real and reactive power losses by using eqns.
𝑉𝑖𝑏 𝑉𝑖𝑐
(1) & (2) with converged voltages and currents and
2 2
stop.
(𝑃𝑒𝑓𝑓𝑎 [𝑖]+𝑄𝑒𝑓𝑓𝑎 [𝑖])∗𝑋𝑎𝑎𝑖
𝑄𝑙𝑜𝑠𝑠 (𝑖) = ∑( 2 + IV. MODIFIED SENSITIVITY ANALYSIS AND LOSS SENSITIVITY
𝑉𝑖𝑎
2 2 2 2 FACTORS FOR IDENTIFYING THE LOCATION OF DG
(𝑃𝑒𝑓𝑓𝑏 [𝑖]+𝑄𝑒𝑓𝑓𝑏 [𝑖])∗𝑋𝑎𝑏𝑖 (𝑃𝑒𝑓𝑓𝑐 [𝑖]+𝑄𝑒𝑓𝑓𝑐 [𝑖])∗𝑋𝑐𝑎𝑖
2 + 2 ) -----(2)
𝑉𝑖𝑏 𝑉𝑖𝑐 A new technique is employed to find a suitable bus for the
placement of DG using Modified Loss Sensitivity Factors.
The assessment of these candidate buses will reduce the search
Where, Peff[i] = Active power delivered beyond the bus ‘i’. area for Adaptive PSO optimization.
Qeff[i] = Reactive power delivered beyond the bus ‘i’. The Active and reactive power losses can be obtained at ith
bus by using eqns. (1) & (2).
B. Network Constraints
The constraints are The modified loss sensitivity factors can be obtained by
using the following equations.
• Voltage of each phase, should be between 0.95 to 1.05
(2𝑃𝑒𝑓𝑓𝑎 [𝑖])∗𝑅𝑎𝑎𝑖 (2𝑃𝑒𝑓𝑓𝑏 [𝑖])∗𝑅𝑎𝑏𝑖 (2𝑃𝑒𝑓𝑓𝑐 [𝑖])∗𝑅𝑐𝑎𝑖
(per unit). + +
∂P𝑙𝑜𝑠𝑠 𝑉2
𝑖𝑎 𝑉2
𝑖𝑏 𝑉2
𝑖𝑐
• The sum of the generating power and distributed (i) =( )
∂P𝑒𝑓𝑓 3
generation power equal to the sum of the power ------(3)
demand and power loss in the URSDS. (2𝑃𝑒𝑓𝑓𝑎 [𝑖])∗𝑋𝑎𝑎𝑖 (2𝑃𝑒𝑓𝑓𝑏 [𝑖])∗𝑋𝑎𝑏𝑖 (2𝑃𝑒𝑓𝑓𝑐 [𝑖])∗𝑋𝑐𝑎𝑖
+ +
∂Q𝑙𝑜𝑠𝑠 𝑉2 𝑉2 𝑉2
• The URSDS losses after placing the DG should be less (i) =( 𝑖𝑎 𝑖𝑏 𝑖𝑐
)
∂P𝑒𝑓𝑓 3
than the losses without having DG.
------(4)
III. POWER FLOW ALGORITHM (2𝑄𝑒𝑓𝑓𝑎[𝑖])∗𝑅𝑎𝑎𝑖 (2𝑄𝑒𝑓𝑓𝑏 [𝑖])∗𝑅𝑎𝑏𝑖 (2𝑄𝑒𝑓𝑓𝑐 [𝑖])∗𝑅𝑐𝑎𝑖
+ +
∂P𝑙𝑜𝑠𝑠 𝑉2
𝑖𝑎 𝑉2
𝑖𝑏 𝑉2
𝑖𝑐
Power flow in a URSDS is determined by using the (i) =( )
∂Q𝑒𝑓𝑓 3
following algorithm [3] and it uses only primitive impedances
of the branches. The proposed technique is efficient as there is ------(5)
(2𝑄𝑒𝑓𝑓𝑎 [𝑖])∗𝑋𝑎𝑎𝑖 (2𝑄𝑒𝑓𝑓𝑏 [𝑖])∗𝑋𝑎𝑏𝑖 (2𝑄𝑒𝑓𝑓𝑐[𝑖])∗𝑋𝑐𝑎𝑖
no need to form the BIBC, BCBV and DLF matrices as + +
∂Q𝑙𝑜𝑠𝑠 𝑉2 𝑉2 𝑉2
suggested in the literature. (i) =( 𝑖𝑎 𝑖𝑏 𝑖𝑐
)
∂Q𝑒𝑓𝑓 3
i. Read the URSDS line and load data. ------(6)
ii. Calculate the per unit values of line and load data. The bus that has the most change in real power loss with
respect to reactive power is considered as the proper place to
install the DG. In the next step, the size of the DG will be the global position with the previous global position. If the
estimated by using the APSO technique. present global position > the previous one, then set the global
position to the present global position.
V. ADAPTIVE PARTICLE SWARM OPTIMIZATION FOR FINDING
THE OPTIMAL SIZE OF THE DG Step8:Change the velocity vector by using
v[i]=(w[i]*v[i]+1/rank(i)*rand()*(pbestparticle[i]-particle[i])
In PSO algorithm,each particle produces a solution within + Social_Information[i].
the search space. Every particle has a velocity, a position in
search space and its past performance. In the given search Step9: If the velocity v[i][j] violates its limits (-vmax, vmax),
space, the initial population is distributed randomly. In the set it at its proper limits
course of a flight, each individual particle explores for the Step10: move each particle to is updated position by
optimum value of a function by changing its position through adding the velocity vector (v[i][j]) to position vector.
a variety of generations. Every particle follows an elementary
equation to change its position and velocity. The normal PSO Step11: Increase the iteration count by 1.
does not use crossover and mutation operator.
Step12: Go to step 5 and repeat until the convergence
In Adaptive PSO algorithm, there are three modifications criteria are achieved.
to the PSO algorithm [4]. The first Modification is, total
search space is split into number of segments like a spider net. VII. RESULTS
By using the standard methods, the initial population was The APSO algorithm for reducing the losses by placing the
created indiscriminately where they depend much on the suitable size of DG is tested on 25 and 37-bus test systems.
mutation operator. It gives better searching capability by The modified sensitivity factor method is used to identify the
having individuals in each and every segment. The Second appropriate bus in order to install DG. The several constants
Modification is , in the given model, the data sharing part of used in the APSO technique are no. of particles=30,
the swarms are made very strong. Every swarm takes data K=0.7259, w=1.2, C1=C2=2.05.
from all other swarms that have higher fitness value than its.
The swarms that have higher fitness value will guide other A. 25-Bus Test System [16]
swarms to enhance their fitness value. The third Modification The 25-bus system has 24 branches. Total active power
is, this technique uses the fitness value to adjust the learning load in each phase is 1073.3,1083.3 & 1083.3 kW. The total
factors C1 and C2 of the swarms. In the traditional PSO, reactive power load each phase is 792, 801 & 800 kVAr. From
fitness is never used. All swarms are ranked depending on the power flow, modified sensitivity factors of the busses are
their fitness value. In the case of minimization problem, the calculated and they are ranked. The bus which is more
swarm that has the smallest fitness value is ranked no.1, and sensitive is preferred as candidate bus for installing the DG
the others will be ranked in this way. unit and in this case, it is 13. The optimal size of DG installed
at this bus is determined by using APSO technique. It is
VI. PROPOSED ALGORITHM FOR DG SIZE AND PLACEMENT observed that the size of the DG is 205 kW at 0.8 p.f. The bus
Step1: Run the URSDS load flow algorithm as given in voltages with and without placing DG are tabulated in Table
section II without placing the DG and compute the real power I. % loss reduction is tabulated in Table II. From the results,
loss (PLwithout DG). the losses and voltage profile of the system has been enhanced
significantly.
Step2: Identify the Suitable buses for locating DG using
modified Loss Sensitivity Factors. The simulated results of the APSO technique are
compared with the results obtained by variational algorithm
Step3: Set the Iteration count t =0, the search space is [6] and are tabulated in Table III. It is noticed that the optimal
divided into ‘m’ number of segments like spider net and size of DG obtained by APSO approach is significantly small
randomly initialize 'S' number of particles in each segment in as compared to the variational algorithm approach with great
the given search space. increase in the percentage of power loss and also the
Step4: Generate the velocity vector (v[i]) between -Vmax & minimum voltage magnitude improved significantly. This
Vmax. indicates that the proposed technique accurately predicts the
optimal location and size of DG with the maximum
Where, Vmax = (DGmax-DGmin)/N percentage power loss reduction and improved voltage
DGmax = maximum DG rating, stability.

DGmin = minimum DG rating & TABLE I. BUS VOLTAGES WITHOUT & WITH DG ON 25- BUS
SYSTEM.
N= number of steps to move the particle from one
position to the other. Without DG With DG
Bus No.
Step5: Run the URSDS load flow algorithm as given in |Va| pu |Vb| pu |Vc| pu |Va| pu |Vb| pu |Vc| pu
section II by placing a particle 'i' at the candidate bus and store 1 1 1 1 1 1 1
the real power loss (PLwithDG).
2 0.9720 0.9726 0.9767 0.9771 0.9776 0.9809
Step6: Evaluate the fitness function value (PLwithoutDG -
3 0.9644 0.9652 0.9705 0.9696 0.9702 0.9746
PLwithDG) of the particle ‘i’ and check with previous particle
best (pbest) value. If the current fitness function value > pbest 4 0.9613 0.9620 0.9682 0.9664 0.9670 0.9724
value, then assign the pbest value to the current value. 5 0.9600 0.9605 0.9671 0.9651 0.9655 0.9713
Step7: Find the current global best (gbest) maximum value
6 0.9596 0.9603 0.9654 0.9699 0.9704 0.9742
among the particles individual best (pbest) Values and check
 TABLE III. COMPARISON OF 25- BUS SYSTEM WITH DG.
7 0.9498 0.9503 0.9562 0.9651 0.9654 0.9695
Variational Algorithm
8 0.9566 0.9578 0.9632 0.9668 0.9679 0.9720 APSO
Method [8]
9 0.9450 0.9458 0.9518 0.9655 0.9659 0.9698 Ph-A Ph-B Ph-C Ph-A Ph-B Ph-C
10 0.9403 0.9412 0.9475 0.9658 0.9664 0.9700 DG
Size(kW) 205 205 205 215 215 215
11 0.9380 0.9391 0.9454 0.9667 0.9673 0.9707 DG
12 0.9371 0.9381 0.9442 0.9657 0.9663 0.9696 Location
13 13
(Bus
13 0.9373 0.9383 0.9446 0.9697 0.9704 0.9738 Number)
Real Power 28.6 33.3
14 0.9469 0.9473 0.9533 0.9623 0.9624 0.9667 32.01 25.04
Loss (kW) 27.23 7 21.21 9
15 0.9463 0.9468 0.9527 0.9616 0.9619 0.9661 Reactive
33.6
Power Loss 32.7 30 31.89 37.08 35.4
16 0.9485 0.9488 0.9550 0.9639 0.9639 0.9684 6
(kVAr)
17 0.9460 0.9461 0.9522 0.9613 0.9612 0.9655 Minimum 0.95 0.95 0.955
0.9557 0.9613 0.9504
Voltage 78 11 7
18 0.9567 0.9580 0.9631 0.9618 0.9630 0.9673 % Voltage
4.43 4.22 3.87 4.96 4.89 4.43
Regulation
19 0.9532 0.9544 0.9589 0.9583 0.9594 0.9630
20 0.9519 0.9529 0.9572 0.9570 0.9579 0.9613
The performance of the 37-bus system with DG is also
21 0.9529 0.9548 0.9594 0.9580 0.9598 0.9636 compared with the results of variational algorithm similar to
22 0.9506 0.9530 0.9575 0.9557 0.9580 0.9616 the 25-bus system and the results are tabulated in Table VI.
23 0.9584 0.9590 0.9657 0.9635 0.9640 0.9698 TABLE IV. BUS VOLTAGES WITHOUT & WITH DG ON 37- BUS SYSTEM.
24 0.9564 0.9572 0.9641 0.9616 0.9622 0.9682
Without DG With DG
25 0.9541 0.9555 0.9622 0.9593 0.9605 0.9663 Bus No.
|Va| pu |Vb| pu |Vc| pu |Va| pu |Vb| pu |Vc| pu

TABLE II. PERFORMANCE OF THE PROPOSED METHOD ON 25- BUS 701 1 1 1 1 1 1

SYSTEM.
702 0.9931 0.9944 0.9918 0.9949 0.9961 0.9938
Without DG With DG 703 0.9862 0.9906 0.9852 0.9880 0.9924 0.9873
Description
Ph-A Ph-B Ph-C Ph-A Ph-B Ph-C 704 0.9901 0.9900 0.9882 0.9956 0.9952 0.9940
Real Power 705 0.9925 0.9931 0.9905 0.9943 0.9948 0.9925
43.59 45.64 34.32 27.23 28.67 21.21
Loss (kW)
% Real 706 0.9888 0.9862 0.9856 0.9977 0.9946 0.9950
Power Loss - - - 37.53 37.18 38.2 707 0.9871 0.9815 0.9841 1.0029 0.9964 1.0004
Reduction
Reactive 708 0.9771 0.9858 0.9764 0.9789 0.9876 0.9784
Power Loss 49.99 45.58 48.58 32.7 30 31.89
(kVAr) 709 0.9795 0.9868 0.9786 0.9813 0.9886 0.9806
% Reactive 710 0.9708 0.9824 0.9698 0.9726 0.9841 0.9718
Power Loss - - - 34.59 34.18 34.36
Reduction 711 0.9666 0.9820 0.9672 0.9684 0.9837 0.9693
Minimum 0.938 0.944 0.955 0.961 712 0.9924 0.9929 0.9899 0.9941 0.9946 0.9919
0.9371 0.9578
Voltage 1 2 7 3
% Voltage 713 0.9918 0.9925 0.9900 0.9951 0.9957 0.9936
6.29 6.19 5.58 4.43 4.22 3.87
Regulation
% 714 0.9898 0.9899 0.9881 0.9954 0.9951 0.9940
Improvement 718 0.9885 0.9897 0.9877 0.9940 0.9949 0.9936
- - - 27.8 31.82 30.65
in Voltage
Regulation 720 0.9889 0.9867 0.9857 0.9978 0.9951 0.9951
722 0.9869 0.9810 0.9839 1.0036 0.9967 1.0011
B. 37-Bus Test System [16]
The 37-bus system has 36 branches. Total active power 724 0.9868 0.9805 0.9839 1.0026 0.9954 1.0002
load in each phase is 727,639 & 1091 kW. Total reactive 725 0.9887 0.9859 0.9855 0.9976 0.9943 0.9949
power load in each phase is 357, 314 & 530 kVAr. From the
727 0.9851 0.9900 0.9843 0.9869 0.9917 0.9863
power flow, modified sensitivity factors of the busses are
calculated and they are ranked. The bus which is more 728 0.9840 0.9893 0.9835 0.9858 0.9910 0.9855
sensitive is preferred as candidate bus for installing the DG 729 0.9840 0.9896 0.9837 0.9858 0.9913 0.9858
unit and in this case, it is 722. The optimal size of DG installed
730 0.9811 0.9877 0.9800 0.9829 0.9895 0.9820
at this bus is determined by using APSO technique. It is
observed that the size of the DG is 150 kW at 0.8 p.f. The bus 731 0.9793 0.9860 0.9783 0.9811 0.9877 0.9803
voltages with and without placing DG are tabulated in table 732 0.9770 0.9857 0.9760 0.9788 0.9874 0.9780
IV. The % loss reduction is tabulated in table V. From the
results, the losses and voltage profile of the system have been 733 0.9747 0.9849 0.9745 0.9765 0.9866 0.9765
enhanced significantly. 734 0.9713 0.9834 0.9713 0.9731 0.9852 0.9733
 735 0.9707 0.9822 0.9693 0.9725 0.9840 0.9713 Detailed results are presented and the algorithm can be applied
to any practical unbalanced distribution network.
736 0.9703 0.9808 0.9695 0.9721 0.9825 0.9715
737 0.9680 0.9825 0.9692 0.9698 0.9842 0.9713 REFERENCES
738 0.9669 0.9821 0.9681 0.9687 0.9839 0.9702 [1] Thukaram D, Wijekoon Banda HM, Jovitha J,” A robust three phase
power flow algorithm for radial distribution systems,” Electric Power
740 0.9665 0.9818 0.9667 0.9683 0.9835 0.9688 Systems Research, vol. 50, no. 3, pp.227-236, June 1999.
741 0.9666 0.9819 0.9669 0.9684 0.9836 0.9690 [2] K.Prakash, M.Sydulu, "Particle Swarm Optimization Based Capacitor
Placement on Radial Distribution Systems," 2007 IEEE Power
742 0.9921 0.9923 0.9903 0.9939 0.9940 0.9923 Engineering Society General Meeting, Tampa, FL, pp. 1-5, June 2007.
744 0.9844 0.9896 0.9839 0.9862 0.9914 0.9859 [3] K.Prakash, M.Sydulu, "An effective topological and primitive
impedance based three phase load flow method for radial distribution
775 0.9795 0.9868 0.9786 0.9813 0.9886 0.9806 systems," TENCON 2008 - 2008 IEEE Region 10 Conference,
Hyderabad, pp. 1-6, Nov. 2008.
TABLE V. PERFORMANCE OF THE PROPOSED METHOD ON 37- BUS [4] Md. Sakhawat Hossen, Fazle Rabbi, Md.Mainur Rahman, “Adaptive
SYSTEM. Particle Swarm Optimization (APSO) for multimodal function
optimization,” International Journal of Engineering and Technology,
Without DG With DG vol. 1, no. 3, pp.98-103, Dec. 2009.
Description [5] K.V.S.Ramachandra Murthy, M.Ramalinga Raju, G.Govinda Rao,
Ph-A Ph-B Ph-C Ph-A Ph-B Ph-C K.Narasimha Rao, “Comparison of Loss Sensitivity Factor & Index
Real Power Vector methods in Determining Optimal Capacitor Locations in
24.08 19.29 22.54 18.79 2.59 14.28 Agricultural Distribution,” 16th National Power Systems Conference,
Loss (kW)
% Real Power pp.26-30, Dec.2010.
- - - 21.97 86.57 36.65 [6] T.Ramana, V.Ganesh, S.Sivanagaraju, “Distributed Generator
Loss Reduction
Reactive Power Placement and Sizing in Unbalanced Radial Distribution System,”
19.67 17.89 23.4 14.32 2.85 16.86 Cogeneration & Distributed Generation Journal, vol. 25, no. 1, pp.52-
Loss (kVAr)
% Reactive 71, Jan. 2010.
Power Loss - - - 27.19 84.06 27.95 [7] S.Sudhakar Reddy, S.Halder Nee Dey, Subrata Paul, “Optimal Size
Reduction and Location of Distributed Generation and KVAR Support in
Minimum 0.966 0.980 0.966 0.968 0.968 Unbalanced 3-ϕ Distribution System using PSO,” International
0.9825 Conference on Emerging Trends in Electrical Engineering and Energy
Voltage 5 5 7 3 8
% Voltage Management (ICETEEEM-2012), pp.77-83, Dec.2012.
3.35 1.95 3.33 3.17 1.75 3.12
Regulation [8] M. Afzalan, M. A.Taghikhani, “DG Placement and Sizing in Radial
% Distribution Network Using PSO & HBMO Algorithms,” Energy and
Improvement Power 2012, vol. 2, no. 4, pp.61-66, 2012.
- - - 5.37 10.25 6.30
in Voltage [9] M.H. Moradi, M.Abedini, “A combination of genetic and particle
Regulation swarm optimization for optimal DG location and sizing in distribution
system,” International Journal of Electrical Power & Energy Systems,
TABLE VI. COMPARISON OF 37- BUS SYSTEM WITH DG. vol. 34, no. 1, pp.66-74, Jan.2012.
[10] H.Manafi, N.Ghadimi, M.Ojaroudi, P.Farhadi, “Optimal Placement of
Variational Algorithm Distributed Generations in Radial Distribution Systems Using
Method APSO
[8] Various PSO and DE Algorithms.” Elektronika ir Elektrotechnika,
Ph-A Ph-B Ph-C Ph-A Ph-B Ph-C vol. 19, no. 10, pp.53-57, 2013.
DG [11] Mohamed Imran.A, Kowsalya.M, “Optimal size and siting of multiple
Size(kW) 150 150 150 160 160 160 distributed generators in distribution system using bacterial foraging
DG optimization,” Swarm and Evolutionary Computation, vol. 15, pp.58-
Location 65, Apr. 2014.
722 722
(Bus [12] Banaja Mohanty, Sasmita Tripathy, “A teaching learning based
Number) optimization technique for optimal location and size of DG in
Real Power distribution network,” Journal of Electrical Systems and Information
18.79 2.59 14.28 25.13 2.91 23.16 Technology, vol. 3, no. 1, pp.33-44, May. 2016.
Loss (kW)
Reactive [13] P.Vijay Babu, S. P. Singh, “Capacitor allocation in radial distribution
Power Loss 14.32 2.85 16.86 22.76 2.91 23.16 system for maximal energy savings,” 2016 National Power Systems
(kVAr) Conference (NPSC), pp.1-6, Dec. 2016.
Minimum 0.98 0.98 0.953 [14] P.Dinakara Prasad Reddy, V.C.Veera Reddy, T.Gowri Manohar, “Ant
0.9683 0.9688 0.9814 Lion optimization algorithm for optimal sizing of renewable energy
Voltage 25 63 4
% Voltage resources for loss reduction in distribution systems,” Journal of
3.17 1.75 3.12 1.87 1.37 4.70 Electrical Systems and Information Technology, vol. 5, no. 3, pp.663-
Regulation
680, Dec. 2018.
VIII. CONCLUSIONS [15] Fariba Amini, Rasool Kazemzadeh, “Distributed Generations Optimal
Placement and Sizing in Unbalanced Distribution Systems with
In this paper, to reduce the system real power losses and Respect to Uncertainties,” International Journal of Renewable Energy
improve the voltage regulation, an Adaptive PSO Technique Research, vol. 7, no. 2, pp.915-925, June. 2017.
is proposed to identify the optimal size of DG unit and [16] Radial distribution test feeders web site at http://sites.ieee.org/pes-
modified sensitivity factor method is used to locate DGS testfeeders/resources/
optimally in URSDS. Moreover, a simplified way for load
flow analysis in URSDS is addressed. This methodology has
been tested on IEEE 25 & 37 bus URSDS and the significant
result has been achieved with the proposed technique. The
results showed that the voltage profile is improved and real
power losses are decreased because of the DG placement.