ISSN 2278-3091

Abdelkader boukaroura et al., International Journal of Advanced

Volume TrendsNovember
11, No.6, in Computer Science and Engineering,
- December 2022 11(6), November - December 2022, 227 - 235
International Journal of Advanced Trends in Computer Science and Engineering
Available Online at http://www.warse.org/IJATCSE/static/pdf/file/ijatcse021162022.pdf
https://doi.org/10.30534/ijatcse/2022/021162022

Optimal Placement and Sizing of DGs in Distribution

Networks Using Dandelion Optimization Algorithm: Case
Study of an Algerian Distribution Network
Abdelkader boukaroura1, Mohammed Amroune2
1
Department of Electrical Engineering, University of Kasdi Merbah, Ouargla 30000, Algeria,
[email protected]
2
Department of Electrical Engineering, University of Ferhat Abbas Setif 1, Setif 19000, Algeria,
[email protected]

Received Date: October 17, 2022 Accepted Date: November 22, 2022 Published Date: December 06, 2022

appropriate locations and sizes. A non-optimal location with

ABSTRACT an optimal size or a non-optimal size with an optimal location
The optimal placement of distributed generation (DG) in can result in an increase in system losses and costs, and
power distribution systems involves identifying the best degradation in voltage profile, protection and stability. Thus,
locations for the generators to be installed and sizing them the simultaneous optimization of location and sizing of DGs in
appropriately, to optimize the performance of the system. In distribution systems can be very useful for the distribution
this paper, the recently proposed nature-inspired optimization power system [4].
algorithm namely: Dandelion Optimizer (DO) has been used In recent years, several meta-heuristic optimization techniques
for the optimal placement and sizing of DG in the radial have been employed for solving the optimal placement and
distribution network. The objectives are to minimize active size of DG sources connected to the distribution network.
power loss and voltage deviation and to enhance the voltage These methods apply an iterative procedure to find the optimal
stability of the distribution network. The efficiency of the
solution or sub-optimal solutions to an optimization problem.
proposed method has been verified over the IEEE 33-bus and
Some of the methods that adopt meta-heuristics notions
Algerian 112-bus distribution systems. The result comparisons
include Genetic Algorithm (GA), Simulated Annealing (SA),
indicated that the proposed method can obtain higher quality
solutions than many other methods for the considered Tabu Search (TS), Particle Swarm Optimization (PSO), Ant
scenarios from the test systems. Therefore, the DO algorithm Colony Optimization (ACO), Non-dominated Sorting GA-II
can be a very effective method for solving the optimal (NSGA-II), Plant Growth Simulation Algorithm (PGSA),
allocation of the DG problem. Artificial Bee Colony Algorithm (ABC), Bacterial Foraging
Algorithm (BFA), Cat Swarm Optimization (CSO), Grey Wolf
Key words: Distribution network, Distributed generation, Optimization (GWO), Krill Herd Algorithm (KHA) and
Dandelion optimizer, Power losses, Voltage deviation, Invasive Weed Optimization (IWO) [6, 7, 8].
Voltage stability. In this paper, the recently proposed met-heuristic method
named Dandelion Optimization (DO) algorithm [9] has been
1. INTRODUCTION adopted for multiple DG allocation and sizing to minimize
active power loss, voltage deviation, and voltage stability
The electric power system can be divided into four major improvement in the distribution network. Case studies with
components; generation, transmission, distribution and standard IEEE 33-bus and Algerian 112-bus distribution
utilities. Amongst these four components, the distribution networks are performed. The obtained results have indicated
network is the final and most critical link in the power system that the proposed algorithm provides higher-quality solutions
[1]. It is a more complex network and has a higher power loss than several other algorithms in the literature for the
as compared to a transmission network due to the high R/X considered scenarios.
ratio. Previous literature studies show that losses in the The rest of this paper is organized as follows: Section 2 deals
distribution network are high and can exceed 13% [2]. with the formulation of the optimal placement and sizing of the
The integration of distributed generation (DG) in the DG problem. The general framework of the employed
distribution networks has several benefits such as reducing Dandelion optimization (DO) algorithm is presented in
power losses, improving voltage profile along feeders and Section 3, whereas Section 4 and 5 illustrates the numerical
increasing the maximum transmitted power in cables and simulations to investigate the performance of the solutions
transformers [3]. However, the installation of DG in the obtained by the optimization algorithm. Finally, conclusions
distribution systems requires consideration of their are drawn in Section 6.

227
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

2. PROBLEM FORMULATION problem.

The power flow equations are defined as equality constraints
2.1 Objective function
in the optimal allocation of DGs problem. The mathematical
Objective functions can be classified as single-objective or model is given by [12]:
multi-objective. The single objective functions are such as Nbus
minimizing system power losses, cost, or enhancing voltage PG ,i  Pl ,i  Vi Y ij V j cos  i   j  ij  (6)
profile or system stability, etc. and multi-objective functions j 1
Nbus
V j sin  i   j  ij 
would be the combination of two or more single objective
functions by considering suitable parameters and constituting QG ,i  Ql ,i  Vi Y
j 1
ij (7)
the objective function [10].
where PG,i is the active power output of the generator at bus i;
A. Single objective function PL,i is the active power of load at bus i; QG,i is the reactive
The main objective of DG siting and sizing in the distribution power output of the generator at bus i; QL,i is the reactive
network is to minimize network active power losses while power of load at bus i; and Yij and θij are the modulus and angle
satisfying some operating constraints. The objective function of ith element in the admittance matrix of the system related to
for the minimization of active power loss is described as: bus i and bus j, respectively.
F1  min  Ploss 
The inequality constraints subjected to DG setting and sizing
(1)
problems include [12]:
where Ploss is the total active power loss of the system
Vmin  Vi  Vmax i  1,2,..., Nbus (8)
expressed as follows:
Nbr where Vmin and Vmax are taken as 0.95 and 1.05 (p.u),
Ploss   R I i
2
(2) respectively.
i 1 I i  I i max i  1, 2,..., Nbr (9)
where Ii and Ri are the current magnitude and the resistance of
the ith branch, respectively, Nbr is the number of branches.
min
PDG  PDGi  PDG
max
(10)

B. Multi-objective function 2  DGbus  Nbus (11)

To represent all the objectives in a combined mathematical where DGbus is the bus number of the DG installation, Vi is the
expression, we divide every single objective function by its bus voltage, Ii is the current of the DG at branch i, PDG is the
base value and link them together by coefficients. The use of total power of DG, Nbr is the total number of branches.
weighting coefficients helps transform three single objective
functions into one objective function, and the whole fitness 3. DANDELION OPTIMIZATION ALGORITHM
function is given by:
 Dandelion optimization (DO) is a new swarm intelligence
P VD VSI 1 
F2  min  w1 loss  w2  w3 1  (3) bioinspired optimization algorithm was proposed by Zhao et al
 P VDbase VSI base 
 loss _ base  in 2022 [9]. DO simulates the process of dandelion seed
long-distance flight relying on wind. The mathematical
VD is the total voltage deviation given by: modelling of the DO can be summarized as follows:
Nbus
VD   Vi  Vrated 
2
(4) 3.1 Initialization
i 1

where Vi is the voltage magnitude at bus i, Vrated is the rated Similar to other natured-inspired metaheuristic algorithms,
voltage (1.0 p.u.) and Nbus is the number of buses in the DO fulfils population evolution and iterative optimization on
distribution network. the basis of population initialization. In the proposed DO
VSI is the voltage stability index given by the following algorithm, it is assumed that each dandelion seed represents a
equation [11]: candidate solution, whose population is expressed as:
VSI j  Vi  4  Pj xij  Q j rij   4  Pj rij  Q j xij  Vi
4 2 2
(5)  x11 ... x1Dim 
 
where VSIj is the voltage stability index of bus j, xij is the population    (12)
reactance of the line connected between buses i and j.  x1pop Dim 
... x pop
w1, w2 and w3 are penalty factors. These factors are attuned  
based on the significance of the objective function. In this Where pop denotes the population size and Dim is the
paper, w1, w2 and w3 are taken as 0.6, 0.3 and 0.1, respectively. dimension of the variable. Each candidate solution is
randomly generated between the upper bound (UB) and the
2.2 Constraints lower bound (LB) of the given problem, and the expression of
the 𝑖th individual 𝑋𝑖 is:
Two types of constraints, which include equality and X i  rand  (UB  LB)  LB (13)
inequality constraints, are considered in the optimization

228
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

where 𝑖 is an integer between 1 and 𝑝𝑜𝑝 and 𝑟𝑎𝑛𝑑 denotes a 𝛼 is a random perturbation between [0, 1] in the process of a
random number between 0 and 1. 𝐿𝐵 and U𝐵 are expressed as: nonlinear decrease that approaches 0. Such fluctuations make
LB  lb1 ,..., lbDim  (14) the algorithm pay much attention to the global search in the
early stage and turn to a local search in the later stage, which is
UB  ub1 ,..., ubDim  (15) beneficial to ensure accurate convergence after a full global
search. 𝑣𝑥 and 𝑣𝑦 represent the lift component coefficients of a
During initialization, DO regards the individual with the dandelion due to the separated eddy action. Eq. (22) is utilized
optimal fitness value as the initial elite, which is approximately to calculate the force on the variable dimension.
considered the most suitable position for the dandelion seed to 1
flourish. Taking the minimum value as an example, the r
e
mathematical expression of the initial elite Xelite is:
vx  r  cos 
f best  min( f ( X i )) (16) (22)
v y  r  sin 
X elite  X ( find ( fbest  f ( X i ))) (17)
where 𝜃 is a random number between [−𝜋, 𝜋].
Where find () denotes two indexes with equal values. Case 2: On a rainy day, dandelion seeds cannot rise
appropriately with the wind because of air resistance, humidity
3.2 Rising stage and other factors. In this case, dandelion seeds are exploited in
In the rising stage, dandelion seeds need to reach a certain their local neighbourhoods, and the corresponding
height before they can float away from their parent. Under the mathematical expression is:
influence of wind speed, air humidity, etc., dandelion seeds X t 1  X t  k (23)
rise to different heights. Here, the weather is divided into the
following two situations. where 𝑘 is used to regulate the local search domain of a
Case 1: On a clear day, wind speeds can be regarded to have a dandelion, and Eq. (24) is used to calculate the domain.
lognormal distribution ln 𝑌∼𝑁(𝜇, 𝜎2). Under this distribution, 1 2 1
q 2 t2  2 t 1 2 (24)
random numbers are more distributed along the 𝑌 -axis, which T  2T  1 T  2T  1 T  2T  1
increases the chance for dandelion seeds to travel to far
Where k=1-rand()×q.
regions. Therefore, DO emphasizes exploration in this case. In
In conclusion, the mathematical expression of dandelion seeds
the search space, dandelion
in the rising stage is
seeds are blown randomly to various locations by the wind.
The rising height of a dandelion seed is determined by wind  X t    vx  vy  ln Y  ( X s  X t ) randn<1.5
X t 1   (25)
speed. The stronger the wind is, the higher the dandelion flies  X t  k else
and the farther the seeds scatter. Affected by wind speed, the
vortexes above the dandelion seeds are constantly adjusted to Where randn() is the random number that follows the standard
make them rise in a spiral form. The corresponding normal distribution.
mathematical expression in this case is: Figure 1 shows the behaviour of dandelion seeds flying under
X t 1  X t    vx  v y  ln Y  ( X s  X t ) (18) different weather conditions. The approximate regeneration
locations of dandelion seeds are given in the figure. First,
Where Xt represents the position of the dandelion seed during when the weather is clear, dandelion seeds are updated based
iteration t. Xs represents the randomly selected position in the on randomly selected location information to emphasize the
search space during iteration t. Eq. (19) provides the exploration process. The eddy above the seed acts on the
expression for the randomly generated position. moving vector by multiplying the 𝑥 and 𝑦 components to
correct the direction of the dandelion’s movement in a spiral.
X s  rand (1, Dim)  (UB  LB)  LB (19)
In the second case, dandelion seeds are exploited in all
ln 𝑌 denotes a lognormal distribution subject to 𝜇 = 0 and 𝜎2 = directions in the local community. The normal distribution of
1, and its mathematical formula is: random numbers is used to dynamically control exploitation
 1  1  and exploration. To make the algorithm more global
 exp   2 (ln y ) 2  y  0 search-oriented, the cut-off point is set to 1.5. This setting
ln Y   y 2  2  (20)
 makes dandelion seeds traverse the entire search space as
 0 y<0 much as possible in the first stage to provide the correct
direction for the next stage of iterative optimization.
In Eq. (20), 𝑦 denotes the standard normal distribution N (0,
1). 𝛼 is an adaptive parameter used to adjust the search step
length, and the mathematical expression is:
1 2
  rand ()  ( 2 t 2   1) (21)
T T

229
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

Figure 3 shows the process of the search agent gradually

updating to the global optimal solution in the final phase. To
accurately converge to the global optimum, the linear
increasing function is applied to individuals to avoid excessive
exploitation. In this stage, the Levy flight coefficient is used to
simulate the individual movement step size. The reason is that
the Levy flight coefficient can be used by agents to stride to
other positions with a large probability under a Gaussian
distribution, which develops more local search domains with a
limited number of iterations.

Figure 1: Schematic diagram of the rising stage of dandelion

seeds.
3.3 Descending stage
In this stage, the proposed DO algorithm also emphasizes
exploration. Dandelion seeds descend steadily after rising to a
certain distance. In DO, Brown motion is used to simulate the
moving trajectory of dandelions. It is easy for individuals to
traverse more search communities in the process of iterative
updating because Brownian motion obeys a normal
distribution at each change. To reflect the stability of
dandelion descent, the average position information after the
Figure 3: Schematic diagram of the dandelion seed landing stage
rising stage is employed. This facilitates the development of
the population as a whole towards promising communities. 3.4 Landing stage
The corresponding mathematical expression is
X t 1  X t     t  ( X mean _ t    t  X t ) (26) In this part, the DO algorithm focuses on exploitation. Based
on the first two stages, the dandelion seed randomly chooses
where 𝛽𝑡 denotes Brownian motion and is a random number where to land. As the iterations gradually progress, the
from the standard normal distribution. algorithm will hopefully converge to the global optimal
Xmean_t denotes the average position of the population in the solution. Therefore, the obtained optimal solution is the
𝑖th iteration, and its mathematical expression is approximate position where dandelion seeds will most easily
1 pop survive. To accurately converge to the global optimum, search
X mean _ t   Xi
pop i 1
(27) agents borrow the eminent information of the current elite to
exploit in their local neighbourhoods. With the evolution of
Figure 2 shows the regeneration process of dandelion seeds the population, the global optimal solution can eventually be
during descent. According to this figure, the average position found. This behaviour is expressed in Eq. (28).
information of the population is essential for the iterative X t 1  X elite  levy ( )    ( X elite  X t   ) (28)
updating of individuals, which directly determines the
evolution direction of individuals. The trajectory of Brownian where Xelite represents the optimal position of the dandelion
motion, which is based on a global search, is also presented in seed in the 𝑖th iteration. Levy (𝜆) represents the function of
the figure. The irregular movement causes the search agents to Levy flight and is calculated using Eq. (29) [13].
escape the local extremum with a high probability during the  
levy ( )  s  1 (29)
iterative update and then pushes the population to seek the
region near the global optimum. t
In Eq. (29), 𝛽 is a random number between [0, 2] (𝛽 = 1.5 in
this paper). 𝑠 is a fixed constant of 0.01. 𝑤 and 𝑡 are random
numbers between [0, 1]. The mathematical expression of 𝜎 is:
1
    
  1     sin   
   2  
 1    1   (30)
     
    2  2  
Figure 2: Schematic diagram of the descending stage of dandelion   2  
seeds

230
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

where 𝛽 is fixed at 1.5. 𝛿 is a linearly increasing function

between [0, 2] and is calculated by Eq. (31).
2t
  (31)
T

4. IMPLEMENTATION OF DO FOR OPTIMAL

ALLOCATION OF DG

The flowchart of the optimal allocation and sizing of DG for

power loss reduction, voltage deviation minimization and Figure 4: Single-line diagram of the IEEE 33-bus system
voltage stability improvement, using the DO algorithm, can be
summarized in the following steps. A. Case 1: Active power loss minimization
The performance of the DO algorithm is tested, firstly, for
Step 1: Read the bus and branch data of the considered active power loss minimization.
network. Figure 5 shows the convergence track of the DO algorithm.
Step 2: Read DG data (DG power limits). The convergence of the DO algorithm occurs on the first
Step 3: Set the parameters of the DO algorithm and the limits iterations, showing the algorithm's ability to explore the
of decision variables (locations and sizing of DG). search space quickly.
Step 4: Generate the initial population for the decision Figure 6 shows how the voltage profile was improved in the
variables (DG locations and sizes). IEEE 33-bus system after DG units integration at buses 14, 24
Step 5: Run backwards and forward to sweep power flow, and 30. Table 1 shows the best solutions obtained by DO,
incorporating DG. algorithm and the comparison with other optimization
Step 6: Compute the active power loss (Ploss), voltage techniques in the literature. From these tables, it can be seen
deviation (VD), and voltage stability index (VSI). that the proposed algorithm has greatly reduced the real power
Step 7: Compute the objective functions represented by Eq. loss and improved the voltage deviation and voltage stability
(1) and Eq. (3). of the distribution network. Moreover, this algorithm has
Step 8: Update the fitness of the objective function. shown high performance compared to competitive
Step 9: Repeat steps 5–8 until the maximum number of optimization algorithms in the literature.
iterations is reached.
Step 10: Print the optimal solution (optimal location and
sizing of DG).

5. RESULTS AND DISCUSSIONS

The proposed algorithm is tested on IEEE 33-bus and Algerian

112-bus radial distribution networks taking into account the
following cases.
 Case 1: in this case, a single objective function of
minimizing active power loss (Ploss) is considered.
 Case 2: in this case, a multi-objectives function is
considered. This function includes the minimization Figure 5: DO convergence characteristic for 33-bus system (Case 1)
of active power loss (Ploss), the minimization of
voltage deviation (VD) and the enhancement of
voltage stability (VSI).
The setting parameters of DO have been taken as follows:
Population size=50, the maximum number of iterations=200.
While the generated power from the DG units is in the range [0
3] MW.

5.1 Test system 1: IEEE 33-bus system

The IEEE 33-bus test system consists of 33 buses, and 32

branches along with a total load of 3.72 MW and 2.30 MVAr.
The substation voltage is 12.66 kV. The single-line diagram of Figure 6: Voltage profile of 33-bus system (Case 1)
the IEEE 33-bus system is shown in Figure 4 and the overall
data of this system is available in [14].

231
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

Table 1: Results for installing DGs in the 33-bus system (Case 1) As can be seen from Table 2, the active power losses and the
Algorithm PL With DGs voltage deviation were severely reduced by proper DG
without Bus DG size PL (kW) PL allocation. It can be seen also that the voltage stability was
DG (kW) no. (kW) reduction improved. The results shown in this case reveal that the DO
(%) algorithm is effective in site DG units in distribution networks,
DO 201.8925 14 759.08 69.3833 65.6335
24 1071.1
finding better solutions and presenting lower losses, lower
30 1099.9 voltage deviation as well as higher voltage stability index
QOGWO 210.98 14 801.81 72.784 65.5019 when compared to the other metaheuristics.
[15] 24 1091.29
30 1053.01
QOCSOS 210.99 13 801.7 72.7869 65.5000
[16] 24 1091.3
30 1053.7
210.98 13 801.8 72.785 65.5000
OTCDE [17] 24 1091.31
30 1053.6
210.98 13 802 72.785 65.50
CMSFS [18] 24 1091
30 1054
MRFO [19] 210.98 13 788.27 72.876 65.4583
24 1017.1
30 1035.3
CBGA-VSA 210.98 13 801.8 72.785 65.50
[20] 24 1091.3 Figure 7: DO convergence characteristic for 33-bus system (Case 2)
30 1053.6
HHO [21] 210.98 14 745.69 72.98 65.40
24 1022.69
30 1135.78
IHHO [21] 210.98 14 757.54 72.79 65.50
24 1080.83
30 1066.69
DA [22] 201.89 14 760 69.3833 65.6335
24 1070
30 1100
CSCA [23] 202.68 13 871.00 71.94 64.5056
24 1091.47
30 954.08
CTLBO [24] 210.99 13 801.7 72.79 65.5007
24 1091.3
30 1053.6
SFSA [12] 210.988 13 802.0 72.785 65.50
Figure 8: Voltage profile of 33-bus system (Case 2)
24 1092.0
30 1053.7 5.2 Test system 1: Algerian 112-bus system

The Algerian 112-bus radial distribution system, shown in

B. Case 2: Multi-objective function Figure 9, has a total active load of 3.36 MW and, a total
In this sub-section, the proposed DO algorithm was employed reactive load of 3.72 MVAr. The rest of the system data is
to site three DG units in the IEEE 33-bus system, aiming to available in [25].
minimize the active power loss and voltage deviation and to A. Case 1: Active power loss minimization
improve voltage stability.
In this sub-section, the effectiveness of the proposed DO for
The convergence characteristic of the DO algorithm for the
solving optimal allocation of DG problem considering active
best solution found is given in Figure 7.
power loss minimization is demonstrated.
The voltage profiles of the IEEE 33-bus test system with and
The convergence characteristic for the best fitness obtained
without DG are given in Figure 8. In comparison to the base
after applying DO is shown in Figure 10.
case, the voltage profile with DG units has significantly
improved.
Table 2: Results for installing DGs in the 33-bus system (Case 2)
Algorithm Without DG With DG
PL (kW) VD (p.u.) VSI Bus no. Size(kW) PL (kW) PL reduction (%) VD (p.u.) VSI
DO 201.89 0.1164 0.69 14 879.4796 71.7686 64.4521 0.0047 0.93
24 1096.7
30 1281.0
QOCSOS [16] 210.99 0.1338 0.67 13 956.4 77.0414 63.4857 0.0065 0.91
24 1030.9
30 1293.5
SCA [23] 202.68 0.1337 0.67 13 1247.61 89.92 55.6344 0.0023 0.95

232
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

25 1061.68
32 1023.50
CSCA [23] 202.68 0.1373 0.67 13 1098.02 88.43 56.3796 0.0016 0.96
24 986.57
30 1584.90
SFSA [12] 210.98 0.1338 0.67 13 964.7 77.41 63.31 0.0062 0.92
24 1133.7
30 1301.8
CTLBO [24] 210.99 - 0.67 13 1192.6 96.17 54.41 0.0009 0.96
25 870.6
30 1629.6

An enhanced voltage profile is obtained with DG integration

in the Algerian network, as shown in Figure 11.
The final results of real power losses and the optimal
allocation of DGs obtained by the DO algorithm are shown in
Table 3. The best objective function value of 43.6717 kW is
obtained via the integration of DGs at buses 15, 24, and 94
with sizes in kW, 1533.9, 1129.4 and 488.7845.

Figure 11: Voltage profile of 112-bus system (Case 1)

Table 3: The best solution obtained by the DO algorithm

for the 112-bus system (Case 1)
PL without With DGs
DG (kW) Bus no. DG size (kW) PL (kW) PL reduction (%)
77.9423 15 1533.9 43.6717 43.9708
24 1129.4
94 488.7845

B. Case 2: Multi-objective function

In this case, the DO algorithm is applied to optimize the
Figure 9: Single-line diagram of the Algerian 112-bus system multi-objective function of active power loss, voltage
deviation plus voltage stability index via optimally
simultaneous allocation of three DGs in the system.
Figure 12 shows the convergence curve for the objective
function over the iterations of the DO algorithm.
The voltage profiles of the test system in the base case and
with DGs are given in Figure 13. It is clear that the voltage
profile of the Algerian network is improved when DGs are
installed.
The power loss reduction, the percentage power loss
reduction, the voltage deviation and the minimum voltage
stability index are presented in Table 4.
As shown in this table, the best locations for DGs units
installation are buses 15, 75 and 94, and the best sizes in kW
Figure 10: DO convergence characteristic for 112-bus system (Case
1) are 1829.9, 1420.0 and 624.5237, respectively.

233
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

Figure 12: DO convergence characteristic for 112-bus system (Case Figure 13: Voltage profile of 112-bus system (Case 2)
2)
Table 3: The best solution obtained by the DO algorithm for the 112-bus system (Case 2)
Without DG With DG
PL (kW) VD (p.u.) VSI Bus no. Size PL (kW) PL reduction (%) VD (p.u.) VSI
(kW)
77.9423 0.1164 0.69 15 1829.9 45.4798 41.6511 0.0065 0.96
75 1420.0
94 624.5237
5. A. Aida Fazliana, M. Azah, H. Sharee, A. Asrul Ibrahim,
6. CONCLUSION T. Khatib and W. Elmenreich. An improved
In this paper, optimal allocation and sizing of DGs are gravitational search algorithm for optimal placement
determined in a radial distribution network through Dandelion and sizing of renewable distributed generation units
in a distribution system for power quality
Optimizer (DO) to achieve the benefits of power loss
Enhancement, Journal of renewable and sustainable
reduction along with improvement of bus voltage profile and
energy, vol. 6, no. 3, pp. 1-17. 2014
voltage stability. The DO algorithm is tested with the optimal
6. V. Rajkumar and D. K. Khatod. Optimal planning of
location and sizing of three DGs in IEEE 33-bus and Algerian distribute d generation system s in distribution
112-bus systems. A comparative study has been carried out to system: A review, Renewable and Sustainable Energy
evaluate the performance of the DO algorithm among other Reviews, vol. 16, pp. 5146 -5165. 2012.
algorithms in the literature. The obtained results reveal that 7. P. Prakash and D. K. Khatod. Optimal sizing and siting
the DO algorithm outperforms other existing methods such as techniques for distributed generation in distribution
SCA, HHO, DA, QOGWO, CTLBO and SFSA. systems: A review, Renewable and Sustainable Energy
Reviews, vol. 57, pp. 111-130. 2016.
REFERENCES 8. W. L. Theo, J. S. Lim, W. S. Ho, H. Hashim and C.T. Lee.
1. T. Yuvaraj and K. Ravi. Multi-objective simultaneous Review of distributed generation (DG) system
DG and D-STATCOM allocation in radial planning and optimisation techniques: Comparison of
distribution networks using cuckoo searching numerical and mathematical modelling methods,
algorithm, Alexandria Engineering Journal, vol. 57, no. Renewable and Sustainable Energy Reviews, vol. 67, pp.
4, pp. 2729-2742, Jan. 2018. 531-573. 2017.
2. A. A. EI-Fergany. Optimal capacitor allocations using 9. S. Zhao, T. Zhang, S. Ma and M. Chen. Dandelion
evolutionary algorithms, IET Generation, Transmission Optimizer: A nature-inspired metaheuristic
& Distribution, vol. 7, no. 6, pp. 593-601. Feb. 2013. algorithm for engineering applications, Engineering
3. R. Sirjani, A. Mohamed and H. Shareef. Heuristic Applications of Artificial Intelligence, vol. 144, 105075.
optimization techniques to determine optimal 2022.
capacitor placement and sizing in radial distribution 10. P. Prakash and K. K. Dheeraj. Optimal sizing and siting
networks: A Comprehensive Review, Przegląd techniques for distributed generation in distribution
Elektrotechniczny, 2012, vol. 7, no. 12, pp. 1-7, 2012. systems: A review, Renewable and Sustainable Energy
4. O. Amanifar and M. E. Hamedani Golshan. Optimal Reviews, vol. 57, pp. 111–130. 2016.
distributed generation placement and sizing for loss 11. M. Chakravorty and D. Das. Voltage stability analysis
and THD reduction and voltage profile improvement of radial distribution networks, International Journal
in distribution systems using particle swarm of Electric Power and Energy Systems, vol. 23, pp.
optimization and sensitivity analysis, International 129–135. 2001
Journal on Technical and Physical Problems of 12. T. P. Nguyen and D. N. Vo. A Novel Stochastic Fractal
Engineering, vol. 3, no. 2, pp. 47-53, Jun. 2011. Search Algorithm for Optimal Allocation of
Distributed Generators in Radial Distribution

234
Abdelkader boukaroura et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(6), November - December 2022, 227 - 235

Systems, Applied Soft Computing Journal, vol. 70, pp. 24. A. Quadri, S. Bhowmick and D. Joshi. A comprehensive
773-796. 2018. technique for optimal allocation of distributed energy
13. R. Klages, G. Randos and I. Sokolov. Anomalous resources in radial distribution systems. Applied
Transport Foundations and Applications. Wiley-VCH Energy, vol. 211, pp.1245–1260. 2018.
Verlag GmbH & Co. KGaA, Weinheim, 2008, pp. 25. A. Boukaroura. Contribution à la modélisation et à
129-159. l’optimisation des réseaux de distribution sous
14. A. Khodabakhshian and M. H. Andishgar. Simultaneous incertitudes, Thèse de doctorat sciences, Université
Placement and Sizing of DGs and Shunt Capacitors in Larbi Ben M’hidi-Oum El Bouaghi, Algérie, 2021.
Distribution Systems by Using IMDE algorithm,
Electrical Power and Energy Systems, vol. 82, pp.
599-607. 2016.
15. S. Kumar, K. Mandal and N. Chakraborty. Optimal
placement of different types of DG units considering
various load models using a novel multiobjective
quasi-oppositional grey wolf optimizer, Soft
Computing, vol. 25, pp. 4845–4864. 2021.
16. H. Truong, P. Nallagownden, I. Elamvazuthi and N. Vo.
A Quasi-Oppositional-Chaotic Symbiotic Organisms
Search algorithm for optimal allocation of DG in
radial distribution networks, Applied Soft Computing
Journal, vol. 88, 106067.2020.
17. S. Kumar, K. Kamal and N. Chakraborty. A novel
opposition-based tuned-chaotic differential evolution
technique for techno-economic analysis by optimal
placement of distributed generation, Engineering
Optimization, vol. 52, no. 2, pp. 303–324.2020.
18. T. Duong, P. Nguyen, N. Vo and M. Le. A newly
effective method to maximize power loss reduction in
distribution networks with highly penetrated
distributed generations, Ain Shams Engineering
Journal, vol. 12, no. 2, pp.1787−1808. 2021.
19. G. Hemeida, A. Ibrahim, A. Mohamed, S. Alkhalaf and
M. Bahaa El-Dine. Optimal allocation of distributed
generators DG based Manta Ray Foraging
Optimization algorithm (MRFO), Ain Shams
Engineering Journal, vol. 12, pp. 609−619. 2021.
20. O. Montoya, W. Gil-González and C. Orozco-Henao.
Vortex search and Chu-Beasley genetic algorithms
for optimal location and sizing of distributed
generators in distribution networks: A novel hybrid
approach, Engineering Science and Technology, an
International Journal, vol. 23, pp.1351–1363. 2020.
21. A. Selim, S. Kamel, S. Alghamdi and F. Jurado. Optimal
Placement of DGs in Distribution System Using an
Improved Harris Hawks Optimizer Based on Single-
and Multi-Objective Approaches, IEEE Access, vol. 8,
pp. 52815−52829. 2020.
22. A. Boukaroura, L. Slimani and T. Bouktir. Optimal
Placement and Sizing of Multiple Renewable
Distributed Generation Units Considering Load
Variations Via Dragonfly Optimization Algorithm,
Iranian Journal of Electrical and Electronic
Engineering, vol. 16, pp. 353–362. 2020.
23. A. Selim S, Kamel and F. Jurado. Efficient optimization
technique for multiple DG allocation in distribution
networks, Applied Soft Computing Journal, vol. 86,
105938. 2019.

235