Ain Shams Engineering Journal 13 (2022) 101786

Contents lists available at ScienceDirect

Ain Shams Engineering Journal

journal homepage: www.sciencedirect.com

Improvement of distribution networks performance using renewable

energy sources based hybrid optimization techniques
Omima M Bakry a, Abdullah Alhabeeb b, Mahrous Ahmed c, Salem Alkhalaf d, Tomonobu Senjyu e,
Paras Mandal f, Mostafa Dardeer a
a
Department of Electrical Engineering, Faculty of Engineering, Aswan University, Aswan 81542, Egypt
b
Department of Information Science, College of Arts, King Saud University, Riyadh, Saudi Arabia
c
Department of Electrical Engineering, College of Engineering, Taif University, P.O. Box 11099, Taif 21944, Saudi Arabia
d
Department of Computer Science, Alrass College of Science and Arts, Qassim University, Qassim, Saudi Arabia
e
Department of Electrical and Electronics Engineering, Faculty of Engineering, University of the Ryukyus, Japan
f
Department of Electrical and Computer Engineering, University of Texas at El Paso, El Paso, TX 79968, USA

a r t i c l e i n f o a b s t r a c t

Article history: A very critical and integral part of the power system is the distribution networks. The final component of
Received 17 July 2021 the power system, including transmission systems or consumers, is the distribution system (DS). This
Revised 9 February 2022 leads to the highest energy loss that happens. So improving the performance of distribution networks
Accepted 17 March 2022
is required not only to provide the reliability of power supply but also to achieve the most economic cost.
Available online 6 April 2022
By optimizing the power flow and simultaneously minimizing the total emission cost and generation cost
and taking into account the power losses, these objectives can be achieved. In recent years, heuristic
Keywords:
methods are widely employed for solving such complex problems, and The main modern optimization
Distribution networks
Genetic Algorithms
techniques are genetic algorithm(GA). The most important issue in Evolutionary Algorithms is explo-
Equilibrium optimizer ration vs. exploitation. Maybe GA is restricted for exploration features, what causes slow convergence
Renewable Energy Sources and poor robustness Therefore, using the hybridization strategy, the main reason behind this is that such
Power loss minimization a hybrid approach is expected to create swape between the exploration and exploitation. This work pre-
Pollutant emissions sents performance improvement of a radial distribution networks using a new hybrid optimization tech-
nique of Genetic Algorithms (GA) with Equilibrium optimizer (EO) algorithm called Hybrid Genetic
Algorithm Equilibrium optimizer (GAEO). It is used for optimum location and size of Renewable
Energy Sources (wind energy, photovoltaic, fuel cell) on distribution systems. DG source locations and
capacity have strongly influenced the improvement of the distribution network performance by reducing
the entire system’s power loss, enhancing the voltage profile, reducing fuel costs and emissions of
contaminants.
Ó 2022 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Ain Shams University
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-
nd/4.0/).

1. Introduction with a high R/X ratio are typical in these distribution networks,
resulting in an unacceptable voltage profile and excessive power
‘‘Due to the simplicity of control and security operations, deliv- losses [1-5]. Facts devices were introduced and applied for power
ery systems are naturally managed in a centralized and vertical systems stabilization based many control systems [6-15]. Many
manner. Radial topologies or weakly interconnected structures research studies have been conducted to improve the efficiency
of DS power transmission and guarantee supply permanency for
various load demands in order to reduce such unfavorable high
Peer review under responsibility of Ain Shams University. power loss. Optimization techniques were introduced to deal with
many engineering problems, with accurate solution [16-22]. DG
integration is the most common method for reducing power loss
and increasing voltage in DS [23]. The generation of electricity
Production and hosting by Elsevier from fossil fuels has resulted in a variety of negative effects, includ-

https://doi.org/10.1016/j.asej.2022.101786
2090-4479/Ó 2022 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Ain Shams University
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

ing air pollution, water pollution, and temperature increase. As a sion that represents one or more quantities that must be
consequence, reducing the use of fossil fuels is important. Around minimized or maximized. When an optimization problem has
the world, renewable energy resources are being commonly used more than one objective function, the aim of a multi-objective
because of its economic, technical, and environmental benefits optimization algorithm is to find one or more optimal solutions
[24-32]. These sustainable renewable sources would ensure [63]. Since classical search and optimization algorithms use a
reduced power loss, improved voltage profile, reduced peak point-by-point method, the effect of using a classical approach is
demand, and eliminated line tension, resulting in a more stable a single optimized solution.
power supply and lower levels of air emissions around the world The present work proposes a novel optimization technique to
[33,34]. Distributed generation (DG), also known as on-site gener- improve the distribution system performance. This technique
ation, is usually conducted by a number of small-scale (typically Using combination between Genetic Algorithm and Equilibrium
1 kW – 50 MW) electric power generators placed near loads or optimizer (GAEO), aspire to satisfy the optimum locations and
attached to an electric distribution grid. Distributed generators capacities of DG units considering total active power loss reduc-
are used to complement or bypass existing electric power systems. tion, improving voltage stability, minimizing of fuel cost, and pol-
These include conventional power generators such as induction lutant emissions. GAEO has been tested on IEEE 33-bus, 69-bus and
generators, synchronous generators, micro turbines (combustion 118-bus radial distribution networks. To prove the effectiveness of
turbines that operate on high-energy fossil fuels such as oil, pro- new algorithm, eight cases have been studied. These cases are with
pane, natural gas, gasoline, or diesel), combustion gas turbines, penetration of different type and number of DG. The system simu-
and reciprocating engines [35]. Energy systems can be built from lated using MATLAB 2014a. The results were compared with GA,
a variety of various energy sources and storage devices. Wind GASBO, SBO, TLBO, PSO, GA/PSO,BFOA,SA,QOTLBO,IA, Hybrid and
and solar energy are the most important renewable energy sources EO methods, to show the strength and weakness of GAEO and to
used in the energy grid. WT and PV DGs are often run close to their prove the validation of novel algorithm over the other.
rated capacity, which ensures that producing electric power using
WT and PV is more cost efficient than generating power conven-
2. Distributed generation modeling
tionally and buying it from the grid. PV and WT are also green
energy sources for DG technologies [36]. However, the optimal
The small DG resources generally operates in constant power
position and scale of DGs guarantees a significant influence on
mode, i.e., the generation node is assumed to model as a constant
power system performance and regulation, as well as the avoid-
negative PQ load [64]. Several advantages have attributed to opti-
ance of any adverse effects such as increased power loss and volt-
mal DG placement (ODGP) in DN, such as system stability, reliabil-
age variations [37]. The optimization of DG position and size is a
ity, loss minimizations, voltage stability, load-ability enhancement,
difficult non-linear problem. The study studies conducted in this
capacity release from sub-stations (SS), improvement in voltage
domain are classified into many groups based on the goals, limits,
stability margin (VSM), improved system efficiency, release of
and solution algorithms investigated [38]. The researchers’ goal
thermal loading in feeders, etc. The total net load will get reduced
functions include power loss reduction [39-41], voltage profile
by the amount of power penetrated by the DG [65]. DGs are also
change [42-44], fuel cost reduction [45,46], efficiency enhance-
classified as Type 1, Type 2, Type 3, and Type 4 DGs (Hung et al.,
ment [47,48], and pollution reduction [49]. In the literature, a wide
2010).
variety of solution methodologies and strategies for optimal posi-
tion and scale of active and reactive power sources in radial distri-

Type 1 DGs only deliver active power and include photovoltaics,
bution systems have been presented. The primary classifications of
fuel cells, and microturbines.
solution techniques are mathematical programming algorithms,

Type 2 DGs deliver only reactive power. They include syn-
heuristics, meta-heuristic methods, and empirical approaches.
chronous compensators such as gas turbines.
The methods that 8have been used for searching the optimum

Type 3 DGs deliver both active and reactive power. These DGs
position and sizing of DG are thought to have two key goals: reduc-
are based on synchronous machines.
ing power loss and optimizing voltage profile. There are several

Type 4 DGs deliver active power while consuming reactive
approaches for solving the considered problem, the most com-
power. Induction Generators used in wind farms are in this cat-
monly one is the Genetic algorithm (GA) [50-61]. GA was applied
egory [66]
to find the suitable location and sizing of DGs for enhancing the
voltage profile and the reduction of power losses, without consid-
ering total harmonic distortion (THD) and individual harmonic dis- 3. Problem formulation
tortion (IHD). The reduction in power loss and voltage stability was
very significant after properly connecting and distributing DGs in The mathematical representation of objective function, F is
the distribution network. GA is very popular for optimizing prob- expressed as follows:
lem solving, but the greatest drawback is the lack of assurance of
Min:Fðx; uÞ ð1Þ
finding global solutions in solving the optimal problems due to
its premature convergence to local optimization zone. In [62]
authors used the loss sensitivity factor (LSF) for sizing optimization Subject to:h(x,u) = 0 ð2Þ
and simulated annealing (SA) for deciding the optimum position of
DGs. The optimal solution of DGs was successfully found by inte- And g(x,u)  0. ð3Þ
grating these two proposed optimization approaches with the
aim of minimizing branch losses. The proposed approach was eval- where x is vector of system state/dependent variables, u is vector of
uated on a number of machine power parameters, and the findings control/independent variables (like generated active and reactive
were compared to those obtained from other approaches. The sim- power, generation bus voltage magnitudes, transformer taps etc).
ulation results shows that this suggested approach outperformed F(x,u) is the objective function, h(x,u) is the equality constraints
other approaches in terms of different power variables, but the and g(x,u) is the inequality constraints. The optimal power flow
method’s apparent drawback is the complicated calculation. As a problem resides the essence of reducing the objective function
consequence, it isn’t commonly used to address other delivery net- and in the same time satisfying the load flow equations (equality
work concerns.”.An objective function is a mathematical expres- constraints) without violation the inequality constraints [67]. The
2
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

set of variables, which describe the state of the power system, can DG technical constraints.As DG capacity is inherently limited by
be defined as follows: the energy resources at any given location, it is necessary to con-

 strain capacity between the maximum and the minimum levels
x ¼ PG1 ; VL1 ;    ; VLNL ; Q G1 ;    ; Q GNG ; Sl1 ;    ; Slnl ð4Þ
[69].
where, PG1 , Q G , VL and Sl are the active power generation at
slack bus, reactive power outputs of the generators, the voltage
Pmin max
gni  Pgni  Pgni ð14Þ
magnitude at load bus and apparent power flow, respectively.NL,
NG and nl are the number of load buses (P-Q buses), generators
5. The objective functions
buses (P-V buses) and the transmission lines, respectively.

5.1. Total active power losses

4. Constraints
The objective functions to be minimized are given by the sum of
The system needs to meet both the requirements of inequality line losses.
and equality. The limitations on power balance are seen as a con-
straint on equality. Inequality restrictions are considered the oper- X
N1

ational limits of power system components. PL ¼ Plk ð15Þ

K¼1
1.1 Equality constraints:
The physics of the power system, as well as the desired voltage
set points throughout the system, are reflected in equality con- Plk ¼ gk ½V2i þ V2j  2Vi Vj cosðdi  dj Þ ð16Þ
straints. The power flow equations, which require that the net The objective function can now be written as.
injection of real and reactive power at each bus sum to zero,
enforce the physics of the power system. [68].  X
N1
minf 1 X ¼ PL ¼ gk ½V2i þ V2j  2Vi Vj cosðdi  dj Þ ð17Þ
X
nb
    i¼1
}PGi  PDi  Vi Vj ½Gij cos di  dj þ Bij sin di  dj  ¼ 0 8ienb
j¼1

ð5Þ 5.2. Voltage deviation index

X The reduction in voltage deviation is one of the primary benefits

nb
   
}Q Gi  Q Di  Vi Vj ½Gij sin di  dj  Bij cos di  dj  ¼ 0 8ienb} of optimal network reconfiguration and DG installation.
j¼1  
V1  Vk
ð6Þ minf 2 X ¼ DV ¼ max 8k ¼ 1; 2;    ::; n ð18Þ
D V1
where, Q G is the generator reactive power, nb is the total number of
If the state of the system has voltage limit violations during
buses, Q D is the reactive load demand, PD is the active load
DG installation, the proposed technique will try to minimize
demand,Gij and Bij are the transfer conductance and susceptance
DVD closer to 0, improving voltage stability and network execu-
between bus i and bus j, respectively.
tion [71].
1.2 Inequality constraints:
These outline the power system operating limits as elaborated
5.3. Minimizing the total emission produced by DGs and the grid
below:Generation constraints: For stable case, real and reactive
power of the generators and the voltages are restricted by the
Some DGs emit pollutant gases into the atmosphere. CO2 ,NOX ,
lower and upper limits as follows:
SO2 , CO and PM10 are the primary polluting gases. Table 1 displays
Q min max
Gi  Q Gi  Q Gi 8ieN ð7Þ the values of the grid and DG parameters. The total annual emis-
sions generated by a hybrid energy system and the grid are as
GPmin max
8ieNG follows:
Gi  PGi  PGi ð8Þ
 XN WT X
N FC X
NPV
Vmin max
Gi  VGi  VGi 8ieNG ð9Þ minf 3 X ¼ EWT i þ EFC i þ EPV i þ EGrid ð19Þ
i¼1 i¼1 i¼1
Constraints on security: The constraints on load bus voltage
magnitudes and transmission line loadings should be kept within
 
EWT i ¼ COWind
2 þ NOWind
x þ SOWind
2  PWT i ð20Þ
the following limits:

Vmin max
8ieNL  
Bi  VBi  VBi ð10Þ
EFC i ¼ COFC FC FC
2 þ NOx þ SO2  PFC i ð21Þ
SLi  Smax
Li 8ienl ð11Þ  
Shunt VAR compensator constraints: The shunt VAR compensators
EWT i ¼ COPV PV PV
2 þ NOx þ SO2  PPV i ð22Þ
are restricted by their limits as follows:  
Q min  Q ci  Q max 8ieNC ð12Þ EGrid ¼ COGrid
2 þ NOGrid
x þ PGrid  PGrid ð23Þ
ci ci

Transformer constraints: The tap settings of the transformers must ‘‘where E and P are design emissions produced and active power
be restricted by their upper and lower limits as follows: generation by the ith energy sources including wind turbine,fuel
cell, Photovoltaic and grid. Also,NWT ; NFC and NPV is the numbers
Tmin
i  Ti  Tmax
i 8ieNT ð13Þ of the WT,FC and PV units, respectively[68].”.

3
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Table 1
Specification of different DG technologies.

DG Rated capacity Fuel cost ($/ Emission factor (Ib/MWh) Life time
type (Mw) Kw h) (year)
NOx SO2 CO2
WT 3.00 0 00 0 20
FC 2.00 0.029 0.03 1078 10
PV 1.00 0 0.006 0 20
Grid - - 00 2031 0
5.06
11.6

5.4. Minimizing the generation cost or fuel cost (FC) based on Darwinian evolutionary theory [73,74], which employs
the cross-over principle to generate improved solutions, referred
A quadratic curve can be used to calculate the fuel cost of elec- to as offspring, from a collection of fitted solutions, referred to as
tricity per hour for a power plant or DG as follows: parents. Cross-over, which happens naturally in nature and aids
! in the preservation of habitat diversity; or, in this case, to dis-
X
NG
cover the domain. Mutations allow offspring to have different
f ¼ ai P2Gi þ bi PGi þ ci þ Penalty ð=hÞ ð24Þ
i¼1 characteristics than their parents. This operator in GA is for local
search and result manipulation. Some solutions and their dimen-
where, and ai , bi and ci are the cost coefficients of ith generators sions are subjected to mutation, which is determined by a func-
[72]. tion and a parameter such as mutation likelihood and
percentage. Evolution flow and how genetic algorithm working
5.4.1. Hybrid genetic algorithm equilibrium optimizer (GAEO) is shown in Fig. 1 and Fig. 2. Equilibrium optimizer (EO) is a
This work presents a hybrid optimization technique based on novel optimizer which mixes dynamic mass balance on a control
GA and EO algorithm. For optimization problems that are too volume firstly proposed by Farmarzi in 2020 [75]. It uses a
complex to be solved using deterministic techniques like linear chunk mass equation to assess the quantity of chunk that step
programming or gradient (Jacobian) approaches, evolutionary out, take in, and produce in a control volume over time and tries
algorithms such as Genetic Algorithms (GA) have emerged as to find the state that will bring the volume to equilibrium
the method of choice. Because of their universality, simplicity [76,77]. In EO, the search agents are made up of particles (solu-
of implementation, and parallel computing suitability. GA is tions) and their concentrations (positions). And then change

Fig. 1. Evolution flow of genetic algorithm [74].

4
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 2. How Genetic Algorithm Works.

their concentration at random in relation to the best-so-far solu- (1) Initialization is the first step in the Equilibrium Optimizer
tions, i.e. equilibrium candidates, in order to arrive at the equi- (EO): EO employs a large number of particles in this step,
librium state (optimal result). The term ‘‘generation pace” is each of which represents a vector of concentration contain-
used to inspire EO’s discovery, exploitation, and avoidance of ing the solution to the optimization problem. The initial con-
local minima (see Fig. 1). (3). Fig. 4 depicts the collaboration centration vector in the search space is generated at random.
of all equilibrium candidates on a sample particle in 2-D and (2) Candidates and the equilibrium pool: The aim of EO is to
how they impact concentration updating one by one [78]. The achieve the system’s equilibrium state. As EO enters equilib-
following steps illustrate how to use the EO algorithm model: rium, it will arrive at a near-optimal solution. EO does not
(see Fig. 3). know the amount of concentrations needed to achieve the

Fig. 3. EO concentrations updating aid in exploration and exploitation [78].

5
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

optimal conditions during the optimization process. As a Table 2

result, it chooses the perfect four particles from the equilib- The control parameters of GAEO.

rium candidate group, as well as another with the best four- a1 : controls the exploration quantity (magnitude) of the algorithm. 2
particle average. The first four particles cause EO to have a a2 : is a constant used to balance the exploitation ability
greater degree of diversification, while the norm aids in 1
GP: controls the participation probability of concentration updating by
extraction. The equilibrium pool, which contains these five
the generation rate. 0.5
candidates, is a vector. selection 0.5
(3) Updating the concentration: Since the turnover rate con- Max. No. of iterations 100
trasts after a little time in an effective control size, which No. of trails 30
is presumed to just be a random vector between 0 and 1
[79], this definition enables EO to have a realistic harmony
between concentration and enhancement

To address GA’s drawbacks and improve its performance, a new particle is sorted and the best solution is selected based on its cal-
hybrid optimization method is put forward, which combines GA culated fitness.
with conventional optimization procedures. GA stands for ‘‘global
quest.” When GA is put together with a problem-specific local 5.4.2.2. Main loop. apply GA (crossover and mutations):
search process, the overall output and solution quality can be Step 4: ‘‘The initialized population is passed to the GA algorithm
improved. A search algorithm with some well- with the first half of the defined population (selection = 0.5). The
adjusted exploitative capacity will increase the probability of gen- reason behind using the first half of population is to reduce the
erating ‘‘better” solutions from ‘‘weak” ones that have already been complexity of the proposed algorithm, as the performance of the
found. As a result, most metaheuristics techniques seek to strike a GA algorithm depends mainly on the size of the population
balance between discovery and exploitation by beginning with a (fitness).”.
high population diversity (high exploration) and gradually Step 5: GA algorithm performs pairing and mating using single
decreasing it during the quest process. Since GA’s rate of conver- point crossover, then Mutate the population.
gence isn’t great, it restricts the features that can be explored. Fur- Step 6: the GA algorithm evaluate the position and cost function
thermore, Because of the small population size, a genetic algorithm for each chromosome in the population.
can sample bad representatives of good search regions and good Step 7: select the best half population members (selection = 0.5)
representatives of bad search regions, necessitating a lengthy eval- and survive to EO and half worse population are updating their
uation process for generation and population. In this regard, location according to GA crossover.
the authors proposed the hybrid GAEO algorithm to produce and apply EO on best half solutions:
provide a better initial population than GA-based algorithms’ ran- Step 8: ‘‘In EO Particles(population members) are evaluated for
domly selected initial population. EO has the potential to be faster their fitness function and then are sorted to determine the equilib-
and provide a better solution than GA. The EO algorithm, on the rium candidates.”.
other hand, can become stuck in a local optimal solution. The solu- ! n! ! ! ! ! o
tions are referred to as chromosomes in the GA algorithm, and the C eq;pool ¼ C eqð1Þ ; C eqð2Þ ; C eqð3Þ ; C eqð4Þ ; C eqðaveÞ ð26Þ
chromosomes are gradually improved through the GA operators at
!
each iteration (i.e., selection, crossover, and mutation). The chro- where C eq;pool is the equilibrium state vec-
mosomes are referred to as particles in the EO algorithm, and the ! ! ! !
tor, C eqð1Þ ; C eqð2Þ ; C eqð3Þ and C eqð4Þ are the four best-so-far particles
particles are gradually improved with each iteration. The particle !
with the lowest fitness value is chosen to represent the optimiza- identified during the whole optimization process. C eqðaveÞ is the
tion problem’s initial solution. At the second half of the given iter- mean value of first four candidate solutions. These solutions are
ations, the resulting particles are transferred to the GA algorithm, assumed as approximated equilibrium states, because no informa-
which then applies EO to the best half solutions.‘‘. tion is available about the equilibrium state, when the optimization
!
Sample particles around an equilibrium candidate (.C 1 ; C 2 Þ process begins. The exponential term F tries to maintain a balance
An equilibrium candidate (C eq ). between exploitation and exploration during the search. The
Probable positions of particles with.k ¼ 0:5 !
expression of vector F :
Probable positions of particles with.k ¼ 0:05
Mathematical model of GAEO algorithm. ! !
F ¼ e k ðtt0 Þ ð27Þ

5.4.2. Appling GAEO algorithm can be divided into two stages as follow !
where k is the turnover rate, and t is the time defined as a function
5.4.2.1. Initialization. Step 1: ‘‘Read The power data of power sys- of iteration (Iter) and thus decreases with the number of iterations:
tem including characteristics of DG units, cost coefficients of Iter
a2 Max
DGs, the network configuration, line impedance, prices of substa- Iter iter
t ¼ ð1  Þ ð28Þ
tion bus and emission functions.”. Max iter
Step 2: ‘‘Assign equilibrium candidates’ fitness value and the
where Iter and Max iter are the current and the maximum number
control parameters of GA and EO as shown in Table 2.”.
of iterations. The value of the t0 is expressed by‘‘
Step 3: ‘‘Set the time counter t = 0 and initializing the popula-
   !
tion of candidate solutions using EO Optimizer as follows in search 1 !
space [xmin ; xmax  randomly.”.
t0 ¼ ! ln a1 sign r  0:5 1  e k t þt ð29Þ
i i k
 
C nitial ¼ randi ðUB  LBÞ þ LB i ¼ 1; 2; 3;    :: ð25Þ !
i where r is a random vector between 0 and 1 and sign r  0:5

where Cnitial is the initial concentration of i particle, UB and LB is the effects on the direction of exploration and exploitation.
i
upper and lower bounds respectively. The rand is uniform random By substitution of Eq. (27) into Eq. (25) The value of exponential
!
value 2[0,1]. After initialization, the location and fitness of each term F can be expressed by.
6
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

!   !
! Step 10: Sift the new population and old population as per cost
F ¼ a1 sign r  0:5 e k t  1 ð30Þ
function-obtained values, where the new population is developed
and others are eliminated.
One of the main factors in the EO is the generation rate Gi that
Step 11: ‘‘(time updating): update the time counter t = t + 1.”.
improving the exploitation phase.
Step 12: ‘‘If one of the termination conditions is satisfied then
! ! ! stop, else go to step 5.”.
G ¼ G 0 e k ðtt0 Þ ð31Þ Fig. 5 Shows the flow chart of GAEO optimization.
Test system and Simulation results.
!
where G 0 is initial value and can be found using Eq. (30).and k is a The new GAEO algorithm is implemented to solve problems
constant. for the standard IEEE 33-bus as shown in Fig. 6, 69-bus and
118-bus distribution systems, these systems are simulated using
! !! !!
G 0 ¼ GCPð C eq  k C Þ ð32Þ MatlabR2014 a environment software. GAEO is solved for single
objective function. The maximum number of iterations is set at
! 0:5r1 r2  GP 100.
GCP ¼ ð33Þ There are eight cases for each objective function based on type
0r2 < GP
and number of DG, as listed below (see Fig. 7).
!
GCP is the ‘‘possibility of contributing the generation term dur-
ing the position update process, it’s the parameter that control the
Case 1: with penetration of PV and FC for 33 bus system.
generation rate. So the concentration C can by defined in Eq. (32):”.
Case 2: with penetration of WT and FC for 33 bus system.

Case 3: with penetration of FC,WT, and PV for 33 bus system.
! ! ! !  ! ! G !
Case 4: with penetration of PV and FC for 69 bus system.
C ¼ C eq þ C  C eq : F þ ! ð1  F Þ ð34Þ
Case 5: with penetration of WT and FC for 69 bus system.
kV
Case 6: with penetration of FC, WT, and PV for 69 bus system.
is the control volume and considered as unit.
Case 7: with penetration of PV and FC for 118 bus system.
Step 9: Evaluation. The old population and the population
Case 8: with penetration of WT and FC for 118 bus system.
resulting from changes are measured at the end of each period.
Case 9: with penetration of FC, WT, and PV for 118 bus system

Fig. 4. Equilibrium candidate concentration in 2D dimensions.

7
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 5. GAEO Flowchart [78].

8
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 6. Single line diagram for IEEE 33 bus system [80].

Fig. 7. Single line diagram for the IEEE–69-bus radial system.

5.4.3. IEEE 33-bus system test 6. Result for IEEE 33-bus system
The system total active and reactive loads are 3,715KW and
2,300KVAr respectively substation voltage = 12.66 kV, base 6.1. Case 1: With penetration of PV and FC
MVA = 10.The voltage limit is 0.90p.u to 1.0 p.u and the initial
active power loss are 208 kW. The line diagram of this distribution In this case GA, EO and GAEO algorithms are employed to iden-
system is shown in Fig.(6). The total loads for this test system are tify the optimal PV size and location and FC units we can see the
3,715 kW and 2,300 kVAr. [80]. effect of penetration of PV and FC in our system, where the value
of power losses, voltage deviation, Emission and Fuel cost become
less than the initial value. It is observed that the GAEO algorithm
5.4.4. IEEE 69 bus radial distribution test system succeeded in minimizing all objective functions and leads to the
‘‘As shown in Fig. 1, the IEEE 69 bus radial distribution system most optimal execution, and EO finds the least optimal one. Table 3
consists of 69 buses and 68 branches (7). The cumulative active show the results obtained of case 1. The values of power losses,
and reactive loads are respectively 3,800KW and 2,690 KVar. The voltage deviation, emission and fuel cost over the iterations are
total real power loss at the substation was 225KW, with a voltage demonstrated in Fig. 9.
of 12.66 kV and a base MVA of 10. 0.95p.u and 1.05p.u are the
lower and upper voltage thresholds, respectively [81].”.

6.2. Case 2: With penetration of WT and FC.

5.4.5. IEEE 118 bus radial distribution system
An IEEE 118-bus system on a wide scale. 22709.72 kW and In this case a combination between WT and FC for the purpose
17041.07 kVAr are the active and reactive power demands for of improving overall system performance.it is clear that this case
the system, respectively. There are 54 generators, 9 transformers, gives better results than case 1 for all objective functions. GAEO
and 30 VAR compensators in this system. The lower and upper algorithm has a great effect in reducing power losses, voltage devi-
voltage thresholds are 0.95p.u.V1.1p.u. and 0.95p.u.V1.1p.u., ation, emission and fuel cost over the iterations. the power losses is
respectively. The limits of the tap setting transformer are 0.9p.u. reduced from 0.078985 to 0.0696 MW and voltage deviation from
Ti1.1p.u [82]. The single line diagram of IEEE118 is shown in Fig. 8. 0.0354 to 0.025654. Table 4 show the Results obtained of case 2.
9
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 8. Single line diagram for the IEEE-118-bus distribution system.

Table 3
‘‘Results obtained for case 1with FC and PV in the IEEE 33-bus distribution system.”.

Power losses (MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)
Method
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.) size
DG
type
PV 14 0.08493 13 0.10000 7 0.06803 8 0.09987
EO 8 0.10000 8 0.08367 24 0.02768 23 0.10007
FC 13 0.20000 14 0.19991 14 0.10811 25 0.18299
32 0.20000 32 0.19923 8 0.19999 32 0.19982
Function
value 0.083325 0.039415 4157.2727 34.0413

GA PV 30 0.09968 29 0.09811 30 0.09984 24 0.09802

14 0.09818 30 0.09888 25 0.09977 30 0.09756
FC 32 0.19593 18 0.19845 31 0.19770 25 0.19540
18 0.19906 14 0.19109 8 0.19932 14 0.19961
Function
value 0.081349 0.035726 3297.8028 32.9799
PV 32 0.09883 32 0.09614 32 0.09895 8 0.09716
GAEO 18 0.09611 15 0.09733 25 0.09657 7 0.09764
FC 8 0.19454 18 0.19801 7 0.19183 24 0.19927
30 0.19559 31 0.19589 30 0.18467 25 0.19760
Function
value 0.078985 0.0354 3086.9896 30.03668

Fig. 10 shows the comparison convergence characteristics of this that power generation by WT and PV is more cost-effective than
Simulation case. grid power. Therefore, PV units are added to the combination
between WT and FC. This case provides the greatest reduction
6.3. Case 3: With penetration of PV, WT and FC. in power loss voltage deviation, emission and fuel cost, as shown
in Table 5. The convergence plots of GA, EO and GAEO are shown
As mention in previous case, the combination between WT in Fig. 11. The superiority of GAEO in achieving a global optimal
and FC gives better overall system. Otherwise, the power pro- solution in a smaller number of convergence measures is
duced by WT and PV is close to their rated capacity, meaning apparent.
10
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 9. ‘‘Comparison of convergence characteristics of Simulation case 1 for 33-bus system.”.

Table 4
‘‘Results obtained for case 2with FC and WT in the IEEE 33-bus distribution system.”.

Power losses (MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)
Method
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.)
DG
size
type
WT 30 0.29937 32 0.29998 32 0.30000 24 0.29936
EO 14 0.29917 14 0.29983 18 0.29999 8 0.28755
FC 8 0.19970 18 0.20000 25 0.19952 25 0.19726
25 0.19966 8 0.20000 4 0.19918 32 0.19892
Function
value 0.074877 0.026032 2545.2584 27.806

GA WT 8 0.28223 18 0.29882 25 0.29437 30 0.28165

31 0.28937 14 0.27275 30 0.29300 24 0.29478
FC 25 0.19614 30 0.18290 24 0.19457 7 0.19061
14 0.19834 31 0.19726 7 0.19910 8 0.19834
Function
value 0.071567 0.025759 2400.8045 23.6523
WT 14 0.27342 30 0.29432 8 0.29605 7 0.28155
GAEO 30 0.29871 18 0.27897 30 0.29940 8 0.29871
FC 7 0.19671 14 0.19941 7 0.19991 32 0.19823
32 0.19294 32 0.19327 25 0.19980 24 0.19294
Function
value 0.0696 0.025654 2310.6233 22.7489

11
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 10. ‘‘Comparison of convergence characteristics of Simulation case 2 for 33-bus system”.

Table 5
‘‘Results obtained for case 3 with FC, PV and WT in the IEEE 33-bus distribution system.”.

Power losses (MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)
Method
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.)
DG
size
type

PV 18 0.10009 33 0.03734 24 0.09963 8 0.09919

EO 25 0.09606 30 0.09838 14 0.10003 25 0.10023
WT 13 0.29371 8 0.29766 32 0.29981 7 0.27980
8 0.30077 31 0.27833 25 0.29796 24 0.29917
FC 30 0.18114 18 0.19844 8 0.20098 14 0.17804
Function 32 0.13956 14 0.19484 7 0.19904 18 0.19901
value 2023.6185
0.072335 0.022558 23.9336

GA PV 18 0.09948 14 0.09966 30 0.09600 7 0.09902

32 0.09835 30 0.09421 25 0.09163 24 0.09780
WT 7 0.29275 32 0.29983 32 0.29916 32 0.29442
25 0.29572 18 0.29808 24 0.29289 30 0.28912
FC 8 0.19865 29 0.19494 8 0.19518 8 0.19614
Function 30 0.19982 58 0.19878 7 0.18997 25 0.19834
value
0.069686 0.023853 1959.426 19.544

PV 30 0.09923 32 0.09478 24 0.09375 32 0.09577

GAEO 32 0.09786 30 0.09208 7 0.09675 25 0.09929
WT 8 0.29448 31 0.29385 30 0.25969 7 0.28488
18 0.29815 14 0.27640 8 0.27326 8 0.28702
FC 24 0.17946 8 0.19086 32 0.19934 24 0.19823
Function 29 0.19906 18 0.19901 25 0.19298 30 0.19294
value
0.021168 1907.9218 18.8473
0.065752

12
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 11. ‘‘Comparison of convergence characteristics of Simulation case 3 for 33-bus system”.

Table 6
‘‘Results obtained for case 4with FC and PV in the IEEE 69-bus distribution system”.

Power losses (MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)
Method
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.) size
DG
type
PV 11 0.03944 64 0.10000 49 0.09952 64 0.07908
EO 60 0.10001 21 0.04201 3 0.06901 12 0.08327
FC 12 0.19891 69 0.19845 60 0.08175 21 0.18645
21 0.19806 60 0.19960 63 0.16118 50 0.2000
Function
value 0.069072 0.018599 1973.1846 21.3901

GA PV 64 0.09399 21 0.09902 11 0.09785 50 0.09573

11 0.09809 61 0.09780 64 0.09784 11 0.09495
FC 21 0.19976 64 0.19594 12 0.19862 61 0.19481
61 0.18314 62 0.19834 50 0.19786 49 0.19537
Function
value 0.068081 0.016114 1233.0523 15.406
PV 50 0.09921 59 0.08934 49 0.09481 11 0.09563
GAEO 21 0.09109 21 0.09690 61 0.09719 50 0.09415
FC 64 0.19400 12 0.19823 50 0.19884 49 0.19961
12 0.19389 64 0.19294 8 0.17166 61 0.17186
Function
value 0.065156 0.014992 1334.1524 15.0157

13
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 12. Comparison of convergence characteristics of Simulation case 4 for 69-bus system.

Table 7
‘‘Results obtained for case 5 with FC and WT in the IEEE 69-bus distribution system.”.

Power losses(MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)

Method
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.)
DG
size
type
WT 61 0.29683 64 0.29998 49 0.29453 61 0.29998
EO 69 0.23528 29 0.06000 28 0.29996 50 0.27058
FC 11 0.19832 61 0.19448 1 0.19964 65 0.20001
21 0.17492 21 0.20000 61 0.19913 49 0.20000
Function
value 0.06772 0.013607 1023.7396 10.8553

GA WT 61 0.21718 17 0.29350 61 0.29527 50 0.29882

21 0.21492 61 0.28612 64 0.29576 61 0.27647
FC 12 0.19457 65 0.18214 49 0.19922 64 0.18102
61 0.05107 12 0.19726 50 0.19982 49 0.19576
Function
value 0.067681 0.0085316 589.5122 9.4199
WT 61 0.29282 59 0.29432 61 0.29448 64 0.29432
GAEO 21 0.27260 61 0.29349 64 0.29815 61 0.29349
FC 59 0.14255 24 0.19941 49 0.19739 50 0.19941
10 0.19544 16 0.19327 21 0.19906 59 0.19327
Function
value 0.065056 0.0065907 675.6921 8.0702

14
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 13. Comparison of convergence characteristics of Simulation case 5for 69-bus system.

Fig. 14. ‘‘Comparison of convergence characteristics of Simulation case 6 for 69-bus system”.

15
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Table 8
Results obtained for case 6 with FC,PV and WT in the IEEE 69-bus distribution system.

Power losses(MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)

Method DG
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.)
type
size

PV 3 0.00166 61 0.09958 61 0.06809 51 0.09219

EO 11 0.10000 64 0.07804 12 0.10021 61 0.07366
WT 61 0.30000 65 0.28848 64 0.29934 11 0.29733
17 0.30000 14 0.23032 8 0.29146 49 0.29505
FC 3 0.00301 68 0.10021 11 0.19374 63 0.19413
Function 21 0.00176 21 0.12826 49 0.19582 58 0.16306
value 23.2618
0.071335 0.010458 5.4897

GA PV 11 0.09806 64 0.09940 61 0.09902 49 0.09966

65 0.09996 11 0.09842 49 0.09780 21 0.09627
WT 22 0.29026 65 0.29613 64 0.28109 64 0.29983
61 0.29029 19 0.29441 21 0.29537 12 0.29808
FC 12 0.17007 21 0.19951 18 0.19594 61 0.19999
Function 49 0.19803 61 0.19498 50 0.19834 50 0.19878
value
0.067686 0.0058853 477.5252 3.6667

PV 61 0.09486 12 0.09629 49 0.09577 59 0.09478

GAEO 17 0.08282 18 0.09742 11 0.09690 61 0.09680
WT 64 0.17061 61 0.28795 12 0.28038 49 0.29385
12 0.27786 64 0.27084 61 0.28702 12 0.29327
FC 24 0.16573 21 0.19943 64 0.19671 64 0.19086
Function 50 0.19980 59 0.19241 50 0.19294 11 0.19292
value
0.0055168 302.8319 8.9068
0.064752

Table 9
‘‘Results obtained for case 7 with FC, PV and WT in the IEEE 118-bus distribution system.”.

Power losses (MW) Voltage deviation Emission (Ib/h) Fuel Cost ($/h)
Method
Location (bus no.) size Location (bus no.) size Location (bus no.) size Location (bus no.) size
DG
type

PV 20 0.99402 116 0.00117 59 0.79422 71 0.87916

EO 97 0.35910 9 0.07286 74 0.96650 31 1.00067
WT 110 2.00986 74 2.66637 20 2.96534 74 2.99536
73 2.07071 50 0.51125 102 2.95886 28 2.99286
FC 42 0.61895 111 1.96001 50 1.88356 107 1.98677
Function 50 1.74548 51 1.47401 51 2.01571 111 1.98962
value
0.58799 0.043661 17605.7362 156.8406

GA PV 41 0.95393 54 0.97967 74 0.95733 50 0.99504

50 0.98086 48 0.96835 111 0.92257 111 0.93336
WT 92 1.43082 74 2.71375 96 2.89888 20 2.99203
109 2.59178 110 2.88323 31 2.95532 96 2.99564
FC 79 1.99313 92 1.62155 102 1.94810 102 1.98174
Function 74 1.48963 50 1.96936 28 1.93817 74 1.94545
value
0.03941 16019.1366 145.2252
0.56197

PV 42 0.78122 111 0.91666 50 0.94775 71 0.91391

GAEO 81 0.97100 66 0.97193 111 0.83756 74 0.96934
WT 50 1.66859 74 2.88784 74 2.98336 107 2.99908
73 2.58341 54 1.78710 28 2.85036 32 2.99292
FC 112 1.92032 50 1.98843 20 1.97147 50 1.88557
Function 99 1.32227 103 1.60486 102 1.72556 28 1.99305
value
0.5139 0.039744 15799.7087 138.6509

16
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 15. ‘‘Comparison of convergence characteristics of Simulation case 7 for 118-bus system”.

Table 10
‘‘comparison result of the GAEO with EO, GA, GASBO and SBO for the IEEE 118-bus system”.

Objective function GAEO EO GA GASBO SBO

Power losses(MW) 0.5139 0.58799 0.56197 0.58750 0.59513

Voltage deviation 0.036074 0.042702 0.03843 0.045895 0.040895
Emission (Ib/h) 13863.47 17419.68 14989.216 14385.62 17514.530
Fuel Cost ($/h) 148.0478 153.7663 152.7624 149.914 155.1474

7. Result for IEEE 69-bus system of case 4. Fig. 12 shows the comparison convergence characteris-
tics of this Simulation case.
7.1. Case 4: With penetration of PV and FC.
7.2. Case5: With penetration of WT and FC.
In this case IEEE 69-bus system used as test system.as mention
before total power losses in this system is 225 KW, GAEO produces As previous case GAEO, EO and GA algorithms deployed to out-
a better solutions compared to GA and EO methods and the power line the optimal size and location of FC and WT in the context of
losses is reduced to 651.56 KW. Table 6 show the results obtained the 69 bus system. GAEO provides optimal results, such that they
17
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 16. ‘‘Comparison of convergence characteristics of Simulation case 8 for 118-bus system.”.

Table 11
Comparison and performance of 33-bus system, 69-bus system and 118-bus system with DG.

33-bus system 69-bus system 118-bus system

Method
PLOSS(MW) Voltage deviation PLOSS(MW) Voltage deviation PLOSS(MW) Voltage deviation
PSO[70] 0.10535 0.0194 0.0832 0.0099 0.72648 0.08066
GA/PSO[70] 0.10340 0.0192 0.0811 0.0075 NON NON
SA[65] 0.08203 0.0324 0.0771 0.0189 0.8588133 0.08095
BFOA[83] 0.08990 0.0295 0.07523 0.0192 NON NON
GA 0.06823 0.01343 0.067545 0.00587 0.5632 0.0394
EO 0.07561 0.01953 0.071335 0.010456 0.5819 0.0421
GAEO 0.06509 0.01089 0.06475 0.005516 0.5139 0.039744
TLBO[41] 0.07554 NON 0.07554 NON 0.5906 NON
QOTLBO[41] 0.0741 NON 0.0741 NON 0.576 NON
IA [84] 0.08105 NON 0.08105 NON NON NON
SBO[85] 0.07182 0.02045 0.06997 0.00863 0.5919 0.045895
GASBO 0.06886 0.01158 0.06751 0.00609 0.5845 0.040895
Hybrid[86,87] 0.07289 NON 0.07289 NON 0.677 0.0526

Fig. 17. Comparison of performances for 33-bus system.

18
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

Fig. 18. ‘‘Comparison of performances for 69-bus system.

Fig. 19. ‘‘Comparison of performances for 118-bus system.

are an improvement over EO and GA for all objective functions. The 8.1. Case 8: Applying GASBO and SBO.
values of power losses, voltage deviation, emission and fuel cost
are shown in Table 7and sketched in Fig. 13. Same as previous case, but with applying SBO and GASBO algo-
rithms also. Table 10 show the Results obtained of case 8. The val-
7.3. Case 6: With penetration of PV, WT and FC. ues of power losses, voltage deviation, emission and fuel cost over
the iterations are sketched in Fig. 16.
The performance of three optimization methods in this case is The proposed Algorithm was compared in Table 11 with GA, EO,
similar to the performance obtained in case 5, the minimum loss PSO, GA/PSO, SA, SBO, TLBO, QOTLBO, IA, Hybrid, GASBO and BFOA.
obtained by the GAEO method is lower than other methods The It is evident that, GAEO gives a better solution than other algo-
convergence plots of GA, EO and GAEO are shown in Fig. 14. It is rithms. Figs. 17–19 show a Comparison of performances for 33-
obvious that GAEO has the advantage of being able to reach a glo- bus system, 69-bus system and 118-bus system.
bal optimal solution with less convergence measures. GAEO
achieves the best results, while EO achieves the worst. Table 8 dis-
9. Conclusion
plays the results.

A novel GAEO algorithm is introduced and implemented in this

8. Result for IEEE 118-bus system work, to Improve the Distribution Networks Using Renewable
Energy Sources by solve several optimization problem. The GA
Case 7: with penetration of PV, WT and FC. and EO algorithms are implemented and compared, as well as
Same as case 3 and 6, a combination between PV, WT and FC is GAEO, which is a combination of the two. For a single objective
used to improve the performance of IEEE 118 system. Table 9 function, GAEO is solved. Based on the form and number of DG,
shows the effect of adding DG on the value of power losses, Emis- the objective function has eight different instances. As compared
sion, fuel Cost and voltage deviation. The convergence plots of GA, to other metaheuristic algorithms, GAEO is a superior optimization
EO and GAEO are shown in Fig. 15. It is clear that, Superiority of method for handling global optimization tasks, making it ideal for
GAEO to achieve best solution in less number of convergence steps. solving complex problems. Furthermore, it supports the proposed
19
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

algorithm’s primacy and its ability to find true and correct [21] Gamil MM, Sugimura M, Nakadomari A, Senjyu T, Howlader HOR, Takahashi H,
et al. Optimal Sizing of a Real Remote Japanese Microgrid with Sea Water
solutions.
Electrolysis Plant Under Time-Based Demand Response Programs. Energies
2020;13(14):3666.
Declaration of Competing Interest [22] Matayoshi H, Kinjo M, Rangarajan SS, Ramanathan GG, Hemeida AM, Senjyu T.
Islanding operation scheme for DC microgrid utilizing pseudo Droop control of
photovoltaic system. Energy Sustain Develop 2020;55:95–104.
The authors declare that they have no known competing finan- [23] Naggar H. Saad, Ahmed A. El-Sattar, Abd El-Aziz M. Mansour, ‘‘A novel control
cial interests or personal relationships that could have appeared strategy for grid connected hybrid renewable energy systems using improved
particle swarm optimization”, Ain Shams Engineering Journal 9 (2018) 2195–
to influence the work reported in this paper.
2214.
[24] Abaza A, Fawzy A, El-Sehiemy RA, Alghamdi AS, Kamel S. Sensitive reactive
Acknowledgment power dispatch solution accomplished with renewable energy allocation using
an enhanced coyote optimization algorithm. Ain Shams Eng J 2021;12
(2):1723–39.
The authors would like to acknowledge the financial support [25] Shaheen MAM, Hasanien HM, Alkuhayli A. A novel hybrid GWO-PSO
received from Taif University Researchers Supporting Project Num- optimization technique for optimal reactive power dispatch problem
solution. Ain Shams Eng J 2021;12(1):621–30.
ber (TURSP-2020/146), Taif University, Taif, Saudi Arabia.
[26] Hemeida MG, Ibrahim AA, Mohamed A-A, Alkhalaf S, Ayman M. Bahaa El-Dine,
‘‘Optimal allocation of distributed generators DG based Manta Ray Foraging
References Optimization algorithm (MRFO) ”. Ain Shams Eng J 2021;12:609–19.
[27] Hamour H, Kamel S, Mohamed EA. Harris Hawks Optimization Technique for
Optimal Reconfiguration of Radial Distribution Networks. Aswan University J.
[1] Wan, Shuo, et al. ‘‘A Communication-efficient Federated learning assisted by
of Sciences and Technlogy 2021;1(1):41–54.
Central data: Implementation of vertical training into Horizontal Federated
[28] Ali ZM, Diaaeldin IM, El-Rafei A, Hasanien HM, Abdel Aleem SHE, Abdelaziz AY.
learning.” arXiv preprint arXiv:2112.01039 (2021).
A novel distributed generation planning algorithm via graphically-based
[2] Senjyu T, Alkhalaf S, Mohamed A-A. Nature-Inspired Algorithms Applications
network reconfiguration and soft open points placement using Archimedes
to Power System Optimization. Aswan University Journal of Science and
optimization algorithm. Ain Shams Eng J 2021;12(2):1923–41.
Technology 2021;1(1):74–109.
[29] Elkazaz MH, Hoballah AA, Azmy AM. Operation optimization of distributed
[3] de Graaff R, Adrianus A. Flexible distribution systems through the application
generation using artificial intelligent techniques. Ain Shams Eng J 2016;7
of multi back-to-back converters: Concept, implementation and experimental
(2):855–66.
verification. Eindhoven Technische Universiteit Eindhoven 2010:210.
[30] Mohammad Jafar Hadidian-Moghaddam, Saber Arabi-Nowdeh, Mehdi Bigdeli,
[4] Islam QNU, Ahmed A, Abdullah SM. Optimized controller design for islanded
Davood Azizian, ‘‘A multi-objective optimal sizing and siting of distributed
microgrid using non-dominated sorting whale optimization algorithm
generation using ant lion optimization technique”, Ain Shams Engineering
(NSWOA). Ain Shams Engineering Journal 2021;12:3677–89.
Journal 9 (2018) 2101–2109.
[5] Mokhtar M, Marei MI, Attia MA. ‘Hybrid SCA and adaptive controller to
[31] Swief RA, El-Amary NH. Optimal probabilistic reliable hybrid allocation for
enhance the performance of grid-connected PV system. Ain Shams Engineering
system reconfiguration applying WT/PV and reclosures. Ain Shams Eng J
Journal 2021;12:3775–81.
2020;11(1):109–18.
[6] Hemeida AM, Hamada MM, Mobarak YA, El-Bahnasawy A, Ashmawy MG,
[32] AL Ahmad A, Sirjani R. Reza Sirjani, ‘‘ Optimal placement and sizing of multi-
Senjyu T. TCSC with auxiliary controls based voltage and reactive power
type FACTS devices in power systems using metaheuristic optimisation
controls on grid power system. Ain Shams Eng J 2020;11(3):587–609.
techniques: An updated review ”. Ain Shams Eng J 2020;11(3):611–28.
[7] Mohamed A-A, El-Gaafary AAM, Mohamed YS, Hemeida AM. Multi-objective
[33] Mahmoud G Hemeida, Salem Alkhalaf, Tomonobu Senjyu, Abdalla Ibrahim,
states of matter search algorithm for TCSC-based smart controller design.
Mahrous Ahmed , Ayman M. Bahaa-Eldin, ‘‘ Optimal probabilistic location of
Electr Power Syst Res 2016;140:874–85.
DGs using Monte Carlo simulation based different bio-inspired algorithms ”,
[8] Hemeida AM, Alkhalaf S, Alfarraj O. Control quality assessment of fuzzy logic
Ain Shams Engineering Journal.
controller based static VAR compensator (SVC). In: In 2015 SAI Intelligent
[34] Bakry OM, Dardeer M, Senjyu T, Alkhalaf S. Multi-Objective Hybrid Genetic
Systems Conference (IntelliSys). p. 507–17.
Algorithms and Equilibrium Optimizer GAEO to Integrate Renewable Energy
[9] Hemeida AM, Mobarak YA, Mousa MR. Power systems stabilization using svc
Sources with Distribution Networks. Aswan University J. of Sciences and
and statcom. Int Rev Model Simulat 2010;3(5):900–10.
Technlogy 2021;1(2):35–69.
[10] Mobarak YA, Hemeida AM, Nithiyananthan K. Voltage and frequency based
[35] Mohamed A-A, Al-Jaafary A, Mohamed Y. Optimal Number, Location and Sizing
load dependent analysis model for Egyptian power system network. J Adv Res
Of FACTS Devices For Optimal Power Flow Using Genetic Algorithm. Aswan
Dyn Control Syst 2019;11(6):971–8.
University Journal of Science and Technology 2021;1(1):55–73.
[11] Hemeida, Ashraf M., and Gaber El-Saady. ‘‘Damping power systems
[36] Ravi Kumar Soni, Deepesh Agarwal,and Parmeshwar kumawat. ‘‘Optimal
oscillations using FACTS combinations.” In 39th International Universities
Allocation of DG to Radial Distribution Using GA. Overview of Different
Power Engineering Conference, 2004. UPEC 2004, vol. 1, pp. 333-337. IEEE,
Approaches.” International Journal of Latest Trends in Engineering and
2004.
Technology (IJLTET): ISSN: 2278-621X, Vol. 5 Issue 4 July 2015.
[12] Hemeida AM. Improvement of voltage stability and critical clearing time for
[37] Kefayat M, Lashkar Ara A, Nabavi SA, Niaki. A hybrid of ant colony optimization
multi-machine power systems using static var compensator. ICGST-ACSE
and artificial bee colony algorithm for probabilistic optimal placement and
2009;9(2):41–7.
sizing of distributed energy resources. Energy Convers Manage
[13] Shair J, Xie X, Liu W, Li X, Li H. Modeling and stability analysis methods for
2015;92:149–61.
investigating subsynchronous control interaction in large-scale wind power
[38] Mohd Dahalan W, Mokhlis H, Ahmad R, Abu Bakar AH, Musirin I. Int. Trans.
systems. Renew Sustain Energy Rev 2021;135:110420.
Electr. Energy Syst. 2015;25(11):2577–94.
[14] Mohamed A-A, El-Gaafary AAM, Mohamed YS, Hemeida AM. Design static VAR
[39] Pesaran HAM, Huy PD, Ramach, Aramurthy VK. A review of the optimal
compensator controller using artificial neural network optimized by modify
allocation of distributed generation: Objectives, constraints, methods, and
Grey Wolf Optimization. In: In 2015 International Joint Conference on Neural
algorithms. Renew Sustain Energy Rev 2017;75:293–311.
Networks (IJCNN). p. 1–7.
[40] Abu-Mouti FS, El-Hawary ME. Optimal distributed generation allocation and
[15] Molina-Cabrera A, Ríos MA, Besanger Y, Hadjsaid N, Montoya OD. Latencies in
sizing in distribution systems via artificial bee colony algorithm. IEEE Trans
power systems: A database-based time-delay compensation for memory
Power Deliv 2011;26(4):2090–101.
controllers. Electronics 2021;10(2):208.
[41] Li Y, Feng B, Li G, Qi J, Zhao D, Mu Y. Optimal distributed generation planning in
[16] Mohamed A-A, Hassan SA, Hemeida AM, Alkhalaf S, Mahmoud MMM, Baha
active distribution networks considering integration of energy storage. Appl
Eldin AM. Parasitism–Predation algorithm (PPA): A novel approach for feature
Energy 2018;210:1073–81.
selection. Ain Shams Eng J 2020;11(2):293–308.
[42] Sultana S, Roy PK. Multi-objective quasi-oppositional teaching learning based
[17] Susowake, Yuta, Hasan Masrur, Tetsuya Yabiku, Tomonobu Senjyu, Abdul
optimization for optimal location of distributed generator in radial
Motin Howlader, Mamdouh Abdel-Akher, and Ashraf M Hemeida. ‘‘A multi-
distribution systems. Electr Power Energy Syst 2014;63:534–45.
objective optimization approach towards a proposed smart apartment with
[43] Muttaqi KM, Le ADT, Negnevitsky M, Ledwich G. An algebraic approach for
demand-response in Japan.” Energies 13, no. 1 (2020): 127.
determination of DG parameters to support voltage profiles in radial
[18] Howlader, Harun Or Rashid, Oludamilare Bode Adewuyi, Ying-Yi Hong, Paras
distribution networks. IEEE Trans Smart Grid 2014;5(3):1351–60.
Mandal, Ashraf Mohamed Hemeida, and Tomonobu Senjyu. ‘‘Energy storage
[44] Al-Abri RS, El-Saadany EF, Atwa YM. Optimal placement and sizing method to
system analysis review for optimal unit commitment.” Energies 13, no. 1
improve the voltage stability margin in a distribution system using distributed
(2020): 158.
generation. IEEE Trans Power Syst 2014;28(1):326–34.
[19] Mohamed A-A, El-Gaafary AAM, Mohamed YS, Hemeida AM. Multi-objective
[45] Quadri I, Bhowmick S, Joshi D. A comprehensive technique for optimal
modified grey wolf optimizer for optimal power flow. In: In 2016 eighteenth
allocation of distributed energy resources in radial distribution systems. Appl
international Middle East power systems conference (MEPCON). p. 982–90.
Energy 2018;211:1245–60.
[20] Sugimura M, Senjyu T, Krishna N, Mandal P, Abdel-Akher M, Hemeida AM.
[46] Tanzil AH, Brandt K, Wolcott M, Zhang X, Garcia-Perez M. Strategic assessment
Sizing and Operation Optimization for Renewable Energy facilities with
of sustainable aviation fuel production technologies: Yield improvement and
Demand Response in Micro-grid. In: In 2019 20th International Conference
cost reduction opportunities. Biomass Bioenergy 2021;145:105942.
on Intelligent System Application to Power Systems (ISAP). p. 1–5.

20
Omima M Bakry, A. Alhabeeb, M. Ahmed et al. Ain Shams Engineering Journal 13 (2022) 101786

[47] El-Khattam W, Bhattacharya K, Hegazy Y, Salama MMA. Optimal investment [66] Kazmi SAA, Shin DR. DG Placement in Loop Distribution Network with New
planning for distributed generation in a competitive electricity market. IEEE Voltage Stability Index and Loss Minimization Condition Based Planning
Trans Power Syst 2004;19(3):1674–84. Approach under Load Growth. Energies 2017;10:1203. doi: https://doi.org/
[48] Qiu Yu, Xu Y, Li Q, Wang J, Wang Q, Liu B. Efficiency enhancement of a solar 10.3390/en10081203.
trough collector by combining solar and hot mirrors. Appl Energy [67] Njuguna CS. Optimization of Placement and Sizing of Distributed Generation
2021;299:117290. and Capacitors for Voltage Stability Improvement. Diss. JKUAT-COHRED 2021.
[49] Wang DTC, Ochoa LF, Harrison GP. DG impact on investment deferral: network [68] Yang, Xin-She. ‘‘Nature-inspired metaheuristic algorithms”. Luniver press,
planning and security of supply. IEEE Trans Power Syst 2010;25(2):1134–41. 2010.
[50] Hung DQ, Mithulananthan N, Bansal RC. An optimal investment planning [69] Hemeida AM, El-Ahmar MH, El-Sayed AM, Hasanien HM, Alkhalaf S, Esmail
framework for multiple distributed generation units in industrial distribution MFC, et al. Optimum design of hybrid wind/PV energy system for remote area.
systems. Appl Energy 2014;124:62–72. Ain Shams Eng J 2020;11(1):11–23.
[51] Kashyap M, Mittal A, Kansal S. Optimal placement of distributed generation [70] Hantash N, Khatib T, Khammash M. An improved particle swarm optimization
using genetic algorithm approach. Computing & Communication algorithm forOptimal allocation of distributed generation units in radial power
Systems. India: Jharkhand; August 2017. p. 587–97. systems. Applied Computational Intelligence and Soft Computing
[52] R. Sulistyowati, C. D. Riawan, and M. Ashari, ‘‘PV farm placement and sizing 2020;2020:1–8.
using GA for area development plan of distribution network,” in Proceedings of [71] Moradi MH, Abedini M. A combination of genetic algorithm and particle
the International Seminar on Intelligent Technology and Its Application, pp. swarm optimization for optimal DG location and sizing in distribution
509–514, Lombok, Indonesia, July 2016. systems. Int J Electr Power Energy Syst 2012;34(1):66–74.
[53] Ali II, Saeed YS. Optimal reconfiguration and distributed generation placement [72] Imran A, Mohamed MK, Kothari DP. A novel integration technique for optimal
in Baghdad distribution sector. Eng Technol J 2018;36(4):333–43. network reconfiguration and distributed generation placement in power
[54] Saleh AA, Mohamed A-A, Hemeida AM. Optimal Allocation of Distributed distribution networks. Int J Electr Power Energy Syst 2014;63:461–72.
Generations and Capacitor Using Multi-Objective Different Optimization [73] Singh RP, Mukherjee V, Ghoshal SP. Particle swarm optimization with an aging
Techniques. In: In 2019 International Conference on Innovative Trends in leader and challengers algorithm for the solution of optimal power flow
Computer Engineering (ITCE). p. 377–83. problem. Appl Soft Comput 2016;40:161–77.
[55] Saleh AA, Mohamed A-A, Hemeida AM, Ibrahim AA. Multi-Objective Whale [74] Abdelaziz M. Distribution network reconfiguration using a genetic algorithm
Optimization Algorithm for Optimal Allocation of Distributed Generation and with varying population size. Electr. Power Syst. Res. 2017;142:9–11.
Capacitor Bank. In: In 2019 International Conference on Innovative Trends in [75] Noraini Mohd and Geraghty, ‘‘Biologically inspired genetic algorithm to
Computer Engineering (ITCE). p. 459–65. minimize idle time of the assembly line balancing” 3rd World Congress on
[56] Alkhalaf S, Senjyu T, Saleh AA, Hemeida AM, Mohamed A-A. A MODA and Nature and Biologically Inspired Computing (NaBIC2011), Oct 2011, Spain.
MODE comparison for optimal allocation of distributed generations with ISBN 978-1-4577-1122-0.
different load levels. Sustainability 2019;11(19):5323. [76] Abdul-hamied DT, Shaheen AM, Salem WA, Gabr WI, El-sehiemy RA.
[57] Ali SM, Mohamed AA, Hemeida AM. A Pareto Strategy based on Multi- Equilibrium optimizer based multi dimensions operation of hybrid AC/DC
Objective for Optimal Placement of Distributed Generation Considering grids. Alexandria Engineering Journal 2020;59(6):4787–803.
Voltage Stability. Int Conf Innovative Trends Computer Eng (ITCE) [77] M. H. El-Ahmar, Abou-Hashema M. El-Sayed, A. M. Hemeida, ‘‘ Mathematical
2019;2019:498–504. doi: https://doi.org/10.1109/ITCE.2019.8646383. Modeling of Photovoltaic Module and Evalute The Effect of Varoius
[58] Ali SM, Mohamed A-A, Hemeida AM. Impact of Optimal Allocation of Paramenters on its Performance ”, Accepted for oral presentation at 2016
Renewable Distributed Generation in Radial Distribution Systems Based on Eighteenth International Middle East Power Systems Conference (MEPCON),
Different Optimization Algorithms. In: In 2019 International Conference on Faculty of Engineering, Helwan University, Egypt, December, 27-29, 2016, pp.
Innovative Trends in Computer Engineering (ITCE). p. 472–8. 741-746.
[59] Aly N, Hemeida AM, Ibrahim AA. Moth-flame algorithm and loss sensitivity [78] Hemeida, Ashraf M., and Gaber Shabeeb. ‘‘Assessment of photovoltaic power
factor for optimal allocation of shunt capacitor banks in radial distribution system in Aswan city.” In 39th International Universities Power Engineering
systems. In: In 2017 Nineteenth International Middle East Power Systems Conference, 2004. UPEC 2004., vol. 2, pp. 610–613. IEEE, 2004.
Conference (MEPCON). p. 851–6. [79] Faramarzi A, Heidarinejad M, Stephens B, Mirjalili S. Equilibrium optimizer: A
[60] Mohamed A-A, Ali S, Alkhalaf S, Senjyu T, Hemeida AM. Optimal allocation of novel optimization algorithm. Knowledge-Based Syst. 2020;191:105190.
hybrid renewable energy system by multi-objective water cycle algorithm. [80] Rabehi A, Nail B, Helal H, Douara A, Ziane A, Amrani M, et al. Optimal
Sustainability 2019;11(23):6550. estimation of Schottky diode parameters using a novel optimization
[61] Mohamed A-A, El-Gaafary AA, Mohamed YS, Hemeida AM. Energy algorithm: Equilibrium optimizer. Superlattices Microstruct
management with capacitor placement for economics low carbon emissions 2020;146:106665.
using modified multi-objective grey wolf optimizer. In: In 3rd International [81] Essallah S, Bouallegue A, Khedher A. Int. J. Renew. Energy Res. 2018;8
Conference on Energy Systems and Technologies. p. 261–70. (1):166–77.
[62] Saleh AA, Mohamed A-A, Hemeida AM, Ibrahim AA. Comparison of different [82] Ali A, Raisz D, Mahmoud K. Int. J. Electr. Power Energy Syst. 2019;105:500–13.
optimization techniques for optimal allocation of multiple distribution [83] The IEEE 118–Bus Test System [online]. Available from: http://www.ee.
generation. In: In 2018 International Conference on Innovative Trends in washington.edu/research/pstca/pf118/pgtca118bus.htm.
Computer Engineering (ITCE). p. 317–23. [84] Kansal S, Kumar V, Tyagi B. Hybrid approach for optimal placement of multiple
[63] Injeti SK, Prema Kumar N. A novel approach to identify optimal access point DGs of multiple types in distribution networks. Electr Power Energy Syst
and capacity of multiple DGs in a small, medium and large scale radial 2016;75:226–35.
distribution systems. Int J Electr Power Energy Syst 2013;45(1):142–51. [85] Kowsalya M. Optimal size and siting of multiple distributed generators in
[64] Gopiya Naik, S., D. K. Khatod, and M. P. Sharma. ‘‘Planning and operation of distribution system using bacterial foraging optimization. Swarm Evol Comput
distributed generation in distribution networks.” ISSN 2250-2459, Volume 2, 2014;15:58–65. doi: https://doi.org/10.1016/j.swevo.2013.12.001.
Issue 9, September 2012 . [86] Hung DQ, Mithulananthan N. Multiple distributed generator placement in
[65] Parihar SS, Malik N. Optimal Integration of Multi-type DG in Radial primary distribution networks for loss reduction. IEEE Trans Ind Electron
Distribution Network. In: Dubey HM, Pandit M, Srivastava L, Panigrahi BK, 2013;60(4):1700–8.
editors. Artificial Intelligence and Sustainable Computing. Algorithms for [87] Samareh Moosavi SH, Khatibi Bardsiri V. Satin bowerbird optimizer: A new
Intelligent Systems. Singapore: Springer; 2022. doi: https://doi.org/10.1007/ optimization algorithm to optimize ANFIS for software development effort
978-981-16-1220-6_9. estimation. Eng Appl Artif Intell 2017;60:1–15.

21