Ain Shams Engineering Journal 12 (2021) 697–708

Contents lists available at ScienceDirect

Ain Shams Engineering Journal

journal homepage: www.sciencedirect.com

Electrical Engineering

ABC algorithm based optimal sizing and placement of DGs in distribution

networks considering multiple objectives
Essam A. Al-Ammar a, Kiran Farzana b, Asad Waqar b,⇑, Muhammad Aamir c, Saifullah d, Azhar Ul Haq e,
Muhammad Zahid f, Memoona Batool b
a
Department of Electrical Engineering, College of Engineering, King Saud University, 11421 Riyadh, Saudi Arabia
b
Department of Electrical Engineering, Bahria University, 44000 Islamabad, Pakistan
c
Department of Electrical & Computer Engineering, PAF-IAST, 22620 Haripur, Pakistan
d
Department of Industrial Engineering, University of Engineering and Technology, 47050 Taxila, Pakistan
e
College of Electrical & Mechanical Engineering, National University of Science and Technology, 44000 Rawalpindi, Pakistan
f
Department of Electrical Engineering, Riphah International University, 55150 Lahore, Pakistan

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

Article history: In recent years, due to increase in load demand, distribution networks have got substantial attention to
Received 3 December 2019 get optimized. It is significant to optimally place the distributed generators (DGs) in distribution net-
Revised 4 April 2020 works to minimize power losses and voltage drops. However, the DGs incur certain investment and oper-
Accepted 6 May 2020
ational costs, and their placement is only viable if these costs overcome energy losses. Therefore, current
Available online 27 June 2020
paper investigates the optimal sizing and placement of DGs in distribution networks with a novel concept
to simultaneously minimize total energy cost along with total power loss and average voltage drop.
Keywords:
Artificial bee colony (ABC) algorithm is proposed to solve the considered multi-objective problem. The
Distributed generation (DG)
Newton Raphson Load flow (NRLF) analysis
performance of the proposed ABC algorithm is tested with standard algorithms. Newton Raphson load
Optimal sizing and placement of DGs flow (NRLF) analysis is conducted on IEEE 33 and 69-bus radial networks and on CIGRE medium voltage
Line loss (MV) benchmark grid. Two test cases have been formed and investigated. The results prove that proposed
Voltage drop ABC algorithm mostly outperforms other algorithms.
Artificial bee colony algorithm Ó 2020 The Authors. Published by Elsevier B.V. on behalf of Faculty of Engineering, Ain Shams University.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction is utilized in serving peak load for 5% of overall operating time [3].
In recent years, the technology of distributed generators (DGs) has
The electrical load demand has been seen to have an extensive gained enough popularity due to its limited commissioning time
rise in the last decades, which in turn has increased overall line loss and fast response [4]. These DGs could minimize line loss and volt-
and voltage drop. As distribution systems have low X/R ratio when age drop in electricity grids, if they are optimally sized and placed.
matched with transmission systems, therefore line loss and voltage A DG can provide both active and reactive powers, and therefore
drop is more significant in distribution systems [1]. Convention- can decrease the peak load demand. This eventually results in
ally, this loss was overcome by expanding the infrastructures [2]. low power loss and voltage drop, and improves the power factor
However, it was observed that only 20% of expanded infrastructure and stability of power system [5].
The published literature shows numerous methods which have
been applied to optimally allocate DGs in different distribution
⇑ Corresponding author.
grids. In [6], power loss minimization was achieved by optimally
E-mail addresses: [email protected] (E.A. Al-Ammar), 01-244172-012@student.
bahria.edu.pk (K. Farzana), [email protected] (A. Waqar), muhamma- sizing and allocating DGs in IEEE 16, 37 and 75-bus systems using
[email protected] (M. Aamir), [email protected] (Saifullah), genetic algorithm (GA). In [7], total power loss in IEEE 33 and 69-
[email protected] (A. Ul Haq), [email protected] (M. bus networks was minimized by optimally placing and sizing DGs
Zahid), [email protected] (M. Batool). using sensitivity test. The authors in [8], have introduced an
Peer review under responsibility of Ain Shams University. improved re-initialized socially structured particle swarm opti-
mization (IRS-PSO) algorithm to optimally locate multiple DGs in
a microgrid. The objectives were to minimize active power loss
as well as to improve voltage profile in IEEE 69-bus radial system.
Production and hosting by Elsevier In [9], nature intelligent water drops (IWD) optimization has been

https://doi.org/10.1016/j.asej.2020.05.002
2090-4479/Ó 2020 The Authors. Published by Elsevier B.V. on behalf of Faculty of Engineering, Ain Shams University.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
698 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708

Nomenclature

b net present factor Qq magnitude of reactive power at qth bus

CT total energy cost Rnew updated solution (food source)
CDGw rating of wth DG Roldm old solution at random food source position m
fitnessR fitness of a food source R (solution) Roldn old solution at random food source position n
ipq current phasor between bus p and q RR probability of fitness of solution R
ICw installation cost of wth DG Sp phasor of apparent power at pth bus
iNL line current of NLth bus Sq phasor of apparent power at qth bus
iNLN maximum thermal limit of line current TG grid tariff
p
MCw maintenance cost of wth DG j complex number i.e 1
NB total number of buses u random number between [-2-2]
Ne total number of best solutions vk voltage at phasor at kth bus
Ploss total active power loss vi voltage at phasor at ith bus
Ploss1(NL-1) active power loss between bus 1 and NL-1 vD average voltage drop
Ploss2(NL-2) active power loss between bus 2 and NL-2 vavg magnitude average voltage
Plosspj active power loss between bus pth and jth buses vrated rated voltage with magnitude of 1 (p.u)
Pin total input power from the utility grid vlow minimum value of voltage
PDGz
P active power injected by zth DG vup maximum value of voltage
Pload total load Ykk phasor of admittance at bus k
Pt_loss % loss reduction Yki phasor of admittance between bus k and i
P0_loss total loss before adding DGs Zpj phasor of line impedance between bus p and j
PDG_loss total loss after adding DGs DT time period in hours
Pp magnitude of active power at pth bus hki angle of Yki
Pq magnitude of active power at qth bus di angle of voltage vi
Pk magnitude of active power at kth bus dk angle of voltage vk
Qk magnitude of reactive power at kth bus Obj objective function
Qp magnitude of reactive power at pth bus

presented to optimally place DGs in IEEE 10, 33 and 69-bus radial (IWO) has been evaluated while optimally allotting a single DG
grids to reduce total line loss. In [10], power loss minimization was in IEEE 33 and 69-bus radial distribution systems to minimize total
achieved by optimally placing multiple DGs in IEEE 33, 69 and power loss, operating cost and voltage violation. The authors in
Indian 85-bus radial networks by using hybrid grey wolf optimizer [24], have proposed a hybrid method (HA) to optimally size and
(HGWO). The authors in [11], have proposed bat algorithm (BA) for site DGs in standard IEEE networks to minimize overall power loss
optimal assignment of PV power in IEEE 33-bus distribution model and to enhance voltage profile. The size of DGs has been optimized
to reduce power loss. PSO algorithm has been used in [12], to by analytical method and location has been found by PSO. The
locate capacitors and DGs in IEEE 33 and 69-bus networks. The authors in [25], have proposed backtracking search optimization
objective was to maximize the profit. In [13], line loss was reduced (BSO) to search optimal capacity and site of DGs in IEEE 33-bus
by optimally engaging DGs in IEEE 33, 69-bus and Portuguese 94- and Portuguese 94-bus radial networks. The objectives were to
bus radial distribution grid by using strud krill herd algorithm reduce power loss and voltage drop. In [26], bacterial foraging opti-
(SKHA). The authors in [14], have optimally placed DGs to decrease mization (BFOA) has been proposed to place and size a single DG in
active power loss in IEEE 34-bus network, and therefore have max- IEEE distribution systems to minimize active power loss, operating
imized profit by proposing an improved version of PSO algorithm. cost and voltage drop. The authors in [27], have presented a com-
The authors in [15], have implemented PSO to evaluate the effect of bined GA-PSO algorithm to minimize power loss and to improve
sizing and siting of a single DG on active power loss in IEEE 33 and the voltage stability in IEEE 33 and 69-bus distribution systems.
69-bus systems. The authors in [16], have applied the bat-inspired In [28], the authors have proposed Harmony Search Algorithm
algorithm (BA) to optimally locate and size a single DG in IEEE 33 (HSA) for optimal sizing of DGs in IEEE 33 and 69-bus distribution
and 69-bus network. The authors in [17], have introduced a bacte- systems. The authors in [29], have introduced a hybrid ant colony
rial foraging algorithm (BFA) to optimally allocate DGs in IEEE 12, optimization (ACO) with ABC algorithm called hybrid ACO–ABC
34 and 69-bus distribution models to improve bus voltages and to algorithm to optimally allocate DGs in IEEE 33 and 69-bus distribu-
reduce power losses. In [18], an efficient analytical (EA) approach tion networks. The authors in [30], have proposed water cycle algo-
was adopted for optimal distribution of various DGs to minimize rithm (WCA) for optimal placement and sizing of DGs and CBs in
power loss in IEEE standard distribution systems. In [19], GA was IEEE 33 and 69-bus distribution systems. In [31], the authors have
proposed to optimally place, size and select capacitor banks (CBs) proposed salp swarm algorithm (SSA) for the optimal location of
and DGs in IEEE radial systems to minimize power loss and to DGs and CBs in IEEE 33 and 69-bus distribution networks. The
improve reliability. The authors in [20], have proposed authors in [32], have proposed chaotic sine cosine algorithm
biogeography-based optimization (BBO) to place PV arrays in IEEE (CSCA) to reduce losses and to improve voltage profile in distribu-
33 and 69-bus radial systems to reduce power loss. The cuckoo tion networks. The authors in [33], have proposed a Taguchi
search algorithm (CSA) has been proposed in [21], to search opti- method (TM) for optimal integration of DGs in distribution net-
mal site and capacity of DGs in IEEE 69-bus radial distribution grid. works. In [34], the authors have introduced exchange market algo-
The objectives were to lower line losses and voltage drops. In [22], rithm (EMA) for optimal sizing and placement of DGs in
authors have presented a reconfiguration of IEEE 33, 69 and 119- distribution networks. In [35], the authors have used quasi-
bus distribution models by using CSA to minimize overall power oppositional chaotic symbiotic organisms search (QOCSOS) algo-
loss and total voltage drop. In [23], invasive weed optimization rithm to find optimal location of DGs in radial distribution
 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708 699

networks. The authors in [36], have utilized strength pareto evolu- Additionally, investigation of minimization of power losses and
tionary algorithm 2 (SPEA2) for optimal integration of DGs in dis- voltage drops has not been carried out on CIGRE MV benchmark
tribution networks. The authors in [37], have proposed a hybrid grid. Hence, this contribution is novel is as well. Key contributions
gravitational search algorithm and general algebraic modelling of this paper include:
system (GSA-GAMS) for optimal location and sizing of DGs in dis-
tribution networks. The authors in [38], have introduced an adap- d Simultaneous minimization of multiple objectives including
tive shuffled frog leaping algorithm (ASFLA) for optimal DG total active power loss, average voltage drop and total energy
allocation in reconfiguration in distribution networks. An artificial cost has been considered.
bee colony (ABC) based optimal allocation of a single DG in IEEE 33 d ABC algorithm has been proposed to solve the multi-objective
and 69-bus systems is proposed to minimize power loss [39]. optimization problem. ABC algorithm is swarm intelligence-
based algorithm which searches solution based on foraging
1.1. Research gap behaviors of bees. The limit cycle ability of the ABC algorithm
reduces the chance of local optimization, and therefore
From above literature survey, following deductions are increases its diversity.
presumed. d Inclusion of CIGRE MV benchmark grid as a test network. A
NRLF analysis has been conducted and simultaneous minimiza-
d Majority of authors optimally sized or placed DGs to minimize tion of multiple objectives including total active power loss,
power losses mostly in IEEE 33 and 69-bus radial networks. average voltage drop and total energy cost has been carried out.
However, none of the authors have taken CIGRE (International d Two test cases have been formed and investigated. In the 1st
Council on Large Electric Systems) medium voltage (MV) bench case, two objectives, including total active power loss and aver-
mark grid as a test network. Therefore, this element has been age voltage drop, have been simultaneously minimized in IEEE
found to be missing in existing literature. 33 and 69-bus radial networks with proposed ABC algorithm,
d In terms of objectives, some of the authors have improved the and the results have been compared with previously applied
voltage profile as well by optimally sizing or placing DGs. Some standard algorithms. In the 2nd case, three objectives, including
authors have simultaneously considered both power losses and total active power loss, average voltage drop and total energy
voltage profiles as objective functions in their optimal sizing or cost have been simultaneously minimized in IEEE 33 and 69-
placement of DGs problem. bus radial networks and in CIGRE MV benchmark grid with pro-
d As far as solution algorithms are concerned GA, PSO, IRS-PSO, posed ABC algorithm and the results have been compared with
HGWO, BA, SKHA, BFA, EA, BBO, CSA, IWO, HA, BSO, BFO, IWD GA.
and ABC have been applied to solve the optimization problem.
It has been observed that application of ABC algorithm in the Rest of paper is organized as follows. In Section 2 proposed
context of optimal sizing and placement of DGs in distribution solution is drafted with explanation of NRLF and ABC algorithm.
networks was limited to a single objective of power loss mini- In Section 3 optimization problem is formulated. In Section 4
mization. None of the authors have considered ABC algorithm results and discussions are made. Section 5 concludes the paper.
to solve a multi-objective DG placement and sizing problem
using ABC algorithm. This element is found to be missing in
2. Proposed solution
existing literature as well.
d Most importantly, none of the authors have considered the
2.1. NRLF analysis
economic aspect of DGs while solving their optimal sizing
and placement problems. Whenever a DG is optimally placed
In order to find out the total active power loss and average volt-
in an electrical network, there are certain investment and
age drop in in IEEE 33 and 69-bus radial networks [16], and in
operational costs attached to it. The impact of these costs
CIGRE MV benchmark grid [41], an NRLF analysis has been per-
couldn’t be ignored, as addition of DGs must be economically
formed. NRLF analysis results in determination of magnitudes of
viable. Unfortunately, this element is found to be missing in
active and reactive powers, voltages and angles at all buses in
existing literature in terms of both investment and operational
any electrical network. These parameters then could be used to
costs of DGs.
find out total active power loss and average voltage drop [42].
For any electrical network, the total active power loss is the sum
1.2. Key contributions
of active power losses in all lines. For IEEE 33-bus radial network
the total lines count to 32. Similarly, for IEEE 69-bus radial network
In current paper while solving the optimal sizing and placement
the total lines count to 68. For CIGRE MV benchmark grid the total
problem of DGs, the authors have introduced a novel concept of
lines count to 15. If NL represents total lines count in any generic
simultaneous minimization of total energy cost along with total
electrical network, the total active power loss could be represented
active power loss and average voltage drop in distribution net-
as in Eq. (1).
works. The total energy cost is the cost of power flow from grid
towards load demand. It includes the cost of total input power Ploss ¼ Ploss1ðNL1Þ þ Ploss2ðNL2Þ þ :::::::::: ð1Þ
and loss power. The proposed solution uses Newton Raphson Load
where Ploss is the total active power loss in a network of NL lines.
Flow (NRLF) analysis to minimize the considered objectives in IEEE
Similarly, Ploss1(NL-1) and Ploss2(NL-2) are active power losses between
33 and 69-bus radial networks and in CIGRE MV benchmark grid.
buses 1 and NL-1, and buses 2 and NL-2 respectively. If p and q rep-
Beforehand, ABC algorithm has been applied to different problems
resent two random buses in any electrical network, then apparent
in power systems including optimal power dispatch (PD), loss min-
powers at these buses could be represented as shown in Eq. (2)
imization in distribution networks, economic dispatch (ED) of
and Eq. (3).
power sources and tuning of power system stabilizers [40]. Accord-
Sp ¼ v p  ipq

ing to authors’ best of knowledge, the simultaneous minimization ð2Þ
of considered objectives has not been investigated before in litera-
Sq ¼ v q  ipq
ture and ABC algorithm has not been applied to same kind of prob- 
ð3Þ
lem. Therefore, the authors consider this contribution as novel.
700 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708

where Sp, vp, i*pq, Sq, and vq are phasors. Sp and Sq represent apparent If there are total NB buses in an electrical network, the voltage
powers at pth and qth buses. Similarly, vp and vq represent voltage phasor at kth bus could be calculated by using Eq. (8) [43].
phasors at pth and qth buses. i*pq is the conjugate of current phasor 1 ðPk  jQ k Þ X
NB
vk ¼ ½  Y ki  v i 
th th
between p and q bus. The current phasor ipq could be calculated Y kk vk

ð8Þ
as shown in Eq. (4). i¼1
vp  vq i–k
ipq ¼ ð4Þ
Z pq where, vk is the phasor of voltage at bus k, Ykk is the phasor of
admittance at bus k, v*k is the conjugate of phasor of voltage at
where vp and vq are phasors of bus voltages at buses p and q respec-
bus k, Pk is the magnitude of active power at bus k, Qk is the magni-
tively. Zpq is the phasor of line impedance between buses p and q.
tude of reactive power at bus k, Yki is the phasor of admittance
The apparent powers Sp and Sq at buses p and q could be repre-
between buses k and i, and vi is the phasor of voltage at bus i. NB
sented in rectangular form as shown in Eq. (5) and Eq. (6).
is total number of buses. Pk and Qk could be found out by using
Sp ¼ P p þ jQ p ð5Þ Eq. (9) and Eq. (10).
X
NB
Sq ¼ P q þ jQ q ð6Þ Pk ¼ Y ki v i v k cosðhki þ di  dk Þ ð9Þ
i¼1
th th
where, Pp and Pq are magnitutes of active powers at p and q
buses. Similiarly, Qp and Qq are magnitudes of reactive powers at X
NB
pth and qth buses. j is the complex number representation and its Qk ¼  Y ki v i v k sinðhki þ di  dk Þ ð10Þ
p
value is 1. The active power loss Plosspq between pth and qth buses i¼1

could be calculated as shown in Eq. (7). In the same way active where Yki is the magnitude of admittance between buses k and i. vi
power loss could be found out in any line between two buses and and vk are magnitudes of voltages at buses i and k respectively. hki is
total active power loss could be calculated by using Eq. (1) in any the angle of Yki. di and dk are the angles of voltages vi and vk respec-
electrical network. tively. The magnitude of average voltage could be calculated by
Plosspq ¼ Pp  Pq ð7Þ using Eq. (11).
P
NB
vk
v av g ¼ k¼1 ð11Þ
NB
where vavg is the magnitude average voltage in a network of NB
buses. Similarly, the average voltage drop could be calculated by
using Eq. (12).
v D ¼ v rated  v av g ð12Þ
where vD is the average voltage drop and vrated is the rated voltage
with a magnitude of 1 (p.u).
While performing NRLF analysis, the total input power from the
utility grid could be calculated by using Eq. (13).
X
Pin ¼ Ploss þ Pload ð13Þ

where Pin is the total input power from the utility grid. Similarly,
P
Pload is the total load and Ploss is the total active power loss of
the electrical network. When a DG is placed in the network, Eq.
(11) could be modified as shown in Eq. (14).
X X
Z
Pin ¼ Ploss þ Pload  PDGz ð14Þ
z¼1

where PDGz is the active power injected by zth DG.

2.2. ABC algorithm

ABC algorithm was introduced by Karobaga [44] in 2005, and

after that it has been adapted in different research problems
[45–47]. It is an optimization algorithm which is based on the for-
aging behaviors of honey bees in a hive in search of food sources. A
food source represents a solution and its fitness is measured in
terms of nectar amount (quality). ABC algorithm is centered
around the behaviors of three types of bees, including employee,
onlooker and scout bees. The number of employee bees is equal
to the number of onlooker bees in a hive. The basic algorithm is
structured in three phases namely employee bees’ phase, onlooker
bees’ phase and scout bees’ phase. In employee bees’ phase, the
Fig. 1. Flowchart of ABC algorithm. employee bees search for food sources, check the nectar amount
 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708 701

and pass this information to the onlooker bees. In onlooker bees’ Step II: Define parameters of ABC algorithm including number
phase, the onlooker bees evaluate the food sources’ information of employee bees, number of onlooker bees, number of limit
(including nectar amount) brought by the employee bees and cycles, initial food sources (random solutions) and number of
decide to whether accept or reject it. Based on decision, the iterations (MI).
employee bees are guided for the next iteration of search. Those Step III: Read the line and load data and perform NRLF analysis
employee bees which come up with same results for a specified with incorporation of defined DG sizes and locations.
number of cycles (called as limit cycles) and don’t give improved
results are designated as scout bees. In scout bees’ phase, the scout Again, find and store the values of objectives (employee bees’
bees are directed to find new food sources in random directions. phase).
This search process reduces the chances of local optimization and Evaluate the values of objectives and select the best ones (on-
can help to increase the diversity of algorithm. looker bees’ phase).
In current paper, a food source represents a solution which Set iteration = 1.
includes size and location of a DG in a distribution network. The
nectar amount (quality) of a food source represents the objectives Step IV: Initiate more food sources (random solutions). Read the
which include total active power loss, average voltage drop and line and load data, and perform NRLF analysis with incorpora-
total energy cost. The step by step search process of ABC algorithm tion of new DG sizes and locations.
is as follows:
Again, find and store the values of objectives (employee bees’
Step 1: In this step, the algorithm parameters including the phase).
number of employee bees, number of onlooker bees, number Evaluate the values of objectives and select the best ones (on-
of limit cycles, initial food sources (population) and number looker bees’ phase).
of iterations are defined. These food sources represent a popu- Find scout bees (same random solutions as of first iteration).
lation of R random solutions, where R is a 3-dimensional vector. Set iteration = iteration + 1.
Step 2: In this step, the employee bees look for food sources.
Each employee bee searches a food source, checks the nectar Step V: Repeat step IV until iterations = MI.
amount, stores its position in memory and reports the food Step VI: Find the probability RR of fitness of solutions using Eq.
source information to an onlooker bee. (16). Store the final solution in memory.
Step 3: In this step, the onlooker bees evaluate the food sources’
information (including nectar amount) brought by the 3. Optimization problem
employee bees. As the number of employee bees is same as
the onlooker bees, when each employee bee comes up to the The ultimate target is to simultaneously minimize three objec-
hive with a food source information each onlooker bee evalu- tives, including total active power loss Ploss, average voltage drop vD
ates this information. All the gathered information from the and total energy cost CT in IEEE 33 and 69-bus radial networks [16],
employee bees is stored in memory by the onlooker bees and and in CIGRE MV benchmark grid [41]. These objectives are subject
the best food source is selected. The employee bees are then to constraints of voltage limitations of buses, generation capacity
redirected by the onlooker bees to search more food sources. limitations of DGs, limitations of branch currents and power
The employee bees continue looking for more food sources balance.
and come up with new food sources’ information. Once again,
each onlooker bee evaluates the information brought by each 3.1. Active power loss
employee bee and process of selecting the best food source is
repeated. In this way the iterative process starts and is repeated The first objective (Obj1) is minimization of total active power
till all iterations are completed. During the process of evalua- loss Ploss and is represented as Eq. (17) [42].
tion, if any onlooker bee finds constant repetition of food source
information by any employee bee for a specified number of Obj1 ¼ minPloss ð17Þ
cycles that employee bee is designated as a scout bee. This scout where Ploss is total active power loss in kW and can be calculated by
bee is directed to look for food sources in other random direc- using Eq. (1). Ploss may also be represented in percentage (%) while
tions. Each onlooker bee checks the food source information doing comparison, as shown in Eq. (18) [49].
with the neighboring food source information and updates this
information in memory by using Eq. (15). P0  PDG
loss loss
Pt loss ¼  100 ð18Þ
P0 loss
Rnew ¼ Roldm þ uðRoldm  Roldn Þjm–n ð15Þ
where Rnew is the updated solution, Roldm is the old solution at a ran-
dom food source position m and Roldn is the old solution at food Table 1
Input parameters for ABC algorithm.
source position n. u is the random number between [2, 2] [48].
The probability RR of the fitness of the food sources is calculated Parameters IEEE 33-Bus IEEE 69-Bus CIGRE MV
by using Eq. (16). Network Network Grid
Iterations 100 150 100
fitnessR
RR ¼ ð16Þ Initial Solutions 400 800 400
P
Ne
(Population)
fitnessR Size Range of DGs 100–1000 100–1000 100–1000
R¼1
(kW)
The flowchart of the proposed ABC algorithm for the DG sizing ICw ($/kW) 350 350 350
MCw ($/kW) 350 350 350
and placement problem is shown in Fig. 1. The steps are as follows:
TG ($/kWh) 0.15 0.15 0.15
IR (%) 12.5 12.5 12.5
Step I: Read the line and load data, perform NRLF analysis, and IF (%) 9 9 9
find the values of objectives without adding the DGs.
702 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708

19 20 21 22 be calculated by using Eq. (14). DT is the time period in hours for a

26 27 28 29 30 31 32 33
year (8760 h), TG is the grid tariff, and b is the net present factor and
could be calculated by using Eq. (22) [12]. Where, IF is the inflation
rate and IR is the interest rate.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Utility grid

X
W X
W
CT ¼ ðC DGw  IC w Þ þ ðC DGw  MC w Þ þ Pin DTT G b ð21Þ
w¼1 w¼1
23 24 25

1 þ IF
b¼ ð22Þ
1 þ IR
Fig. 2. IEEE 33-Bus radial network. As mentioned before, these objectives are subject to four con-
straints [49]. These constraints are listed below.

where Pt_loss represents % loss reduction after adding DGs. P0_loss is 3.4. Voltage limitation of buses
total active power loss before adding DGs and PDG_loss is total active
power loss after adding DGs. The bus voltages are constrained by upper and lower limits, as
shown in Eq. (23).
3.2. Voltage drop v low 6 v k 6 v up ð23Þ
where vlow is the minimum value of vk and vup is the maximum
The second objective (Obj2) is minimization of average voltage
value. All voltages are in (p.u).
drop at all busses and is represented as Eq. (19).

Obj2 ¼ minv D ð19Þ 3.5. Capacity limitations of DGs

where vD is the average voltage and can be calculated by using Eq.
The active power supplied by each DG is also constrained by its
(12).
upper and lower limits, as shown in Eq. (24).

3.3. Total energy cost X

Z X
06 PDGz 6 P load ð24Þ
z¼1
The third objective (Obj3) is minimization of total energy cost CT
and is represented as Eq. (20). where PDGz is active power in kW supplied by each zth DG and RPload
is total load in kW.
Obj3 ¼ minC T ð20Þ

where CT is the sum of installation, maintenance and operational 3.6. Limitations of branch currents
costs in $. It can be calculated by using Eq. (21) [12]. The first part
in Eq. (21) represents the installation cost, the second part repre- The placement of DGs might increase the thermal overloading
sents the maintenance cost and third part represents the opera- of the lines. Therefore, the line currents are also constrained by
tional cost. CDGw is rating of wth DG in MW and ICw is installation upper limits, as shown in Eq. (25).
cost in $/MW of wth DG. Similarly, MCw is maintenance cost in
iNL < iNLN ð25Þ
$/MW of wth DG. Pin is total input power from utility grid and could

Table 2
Comparison for IEEE 33-Bus Radial Network (Case-1)

Algorithm Year DGs’ Sizes (kW) DGs’ Location Pt_loss (%) vmin (p.u)
CSCA [32] 2020 871, 1091.47, 954.08 13, 24, 30 64.5 0.9690
GSA-GAMS [37] 2019 801.22, 1091.3, 1053.59 13, 24, 30 65.64 0.9686
ASFLA [38] 2019 545.7, 993.6, 1209.4 24, 29, 12 67 0.9781
SFLA [38] 2019 563.9, 318.2, 514.4 28, 30, 14 58.86 0.9770
FWA [38] 2019 589.7, 189.5, 1014.6 14, 18, 32 56.24 0.9680
QOCSOS [35] 2019 801.7, 1091.3, 1053.6 13, 24, 30 65.50 NA
SSA [31] 2019 753.6, 1100.4, 1070.6 13, 23, 29 64.8 0.9686
SPEA2 [36] 2018 691, 733.4, 742.9 18, 29, 8 71.1 0.9616
WCA [30] 2018 854.6, 1101.7, 1181 14, 24, 29 65 0.973
EMA [34] 2018 976.6, 1169.09, 943.54 30, 24, 12 64.32 0.9684
TM [33] 2017 719.9, 719.9, 1439.7 15, 26, 33 49.52 0.996
MOTA [33] 2017 980, 960, 1340 7, 14, 30 52.4 0.9986
HGWO [10] 2017 802, 1090, 1054 13, 24, 30 64.4 NA
SKHA [13] 2017 801.81, 1091, 1053.6 13, 24, 30 64.4 0.9687
BA [11] 2016 816.3, 952.35, 952.35 15, 25, 30 63 0.98
IWO [23] 2016 624.7, 104.9, 1056 14, 18, 32 57.7 0.9716
PSO & Analytical [24] 2016 790, 1070, 1010 13, 24, 30 64.1 NA
ACO-ABC [29] 2016 754.7, 1099.9, 1071.4 14, 24, 30 62.8 0.9735
BSOA [25] 2015 632, 487, 550 13, 28, 31 56.1 0.9554
BFOA [26] 2014 633, 90, 947 17, 18, 33 51.5 0.964
HSA [28] 2013 572.4, 107, 1046.2 17, 18, 33 52.3 0.967
GA-PSO [27] 2012 925, 863, 1200 11, 16, 32 49.2 0.967
IWD [9] 2011 600.3, 300, 1011.2 9, 16, 30 57.7 0.9696
Proposed ABC 2020 514.9, 948.9, 635.16 14, 28, 23 76.8 0.9770
 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708 703

Fig. 3. Convergence curve of Ploss in IEEE 33-Bus Network (Case-1). Fig. 5. Convergence curve of Ploss in IEEE 33-Bus Network (Case-2).

Fig. 6. Voltage profile and vavg in IEEE 33-Bus Network (Case-2).

Fig. 4. Voltage profile and vavg in IEEE 33-Bus Network (Case-1).
X
Z X
P DGz þ Pin ¼ Pload þ Ploss ð26Þ
z¼1
Table 3
Comparison for IEEE 33-Bus Radial Network (Case-2)

System Without With DGs (with With DGs (with 4. Results and discussions
Parameter DGs GA) ABC)
DGs’ Location – 19, 20, 18 32, 27, 14 The NRLF analysis have been done with the proposed ABC algo-
DGs’ Size (kW) – 655, 520, 532 500, 500, 500 rithm on IEEE 33 and 69-bus radial networks and, on CIGRE MV
Ploss (kW) 232.9 150.5 56.9
benchmark grid. The complete system has been coded and realized
Pt_loss (%) – 35.5 75.6
vmin (p.u) 0.8778 0.9364 0.9678 in MATLAB, and two test cases have been formed. The utility grid
vavg (p.u) 0.9297 0.9566 0.9855 has been selected as the swing bus for NRLF analysis. In the 1st
vD (p.u) 0.0703 0.0434 0.0145 case, simultaneous minimization of Ploss and vD has been done in
CT ($) 4,298,861 2,932,969 2,814,898 IEEE 33 and 69-bus radial networks by adding three DGs with
the proposed ABC algorithm. In 2nd case, simultaneous minimiza-
tion of Ploss, VD and CT has been done in IEEE 33 and 69-bus radial
networks and, in CIGRE MV benchmark grid by adding three DGs
where iNL is the line current NLth line and iNLN is its maximum ther- with the proposed ABC algorithm. Table 1 shows the details of
mal limit. input parameters of ABC algorithm and related costs.

3.7. Power balance 4.1. IEEE 33-Bus radial network

The total active power supplied by utility grid and DGs must be The IEEE 33-bus radial network with 33 buses and 37 lines is
enough to fulfill total load and losses, as shown in Eq. (26). shown in Fig. 2. The total active and reactive load demand of this
704 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708

Fig. 7. IEEE 69-Bus Radial Network.

Table 4
Comparison for IEEE 69-Bus Radial Network (Case-1)

Algorithm Year DGs’ Sizes (kW) DGs’ Location Pt_loss (%) vmin (p.u)
CSCA [32] 2020 365.9, 1675.85, 65.52 17, 61, 67 68.8 0.98
QOCSOS [35] 2019 526.8, 380.4, 1719.0 11, 18, 61 69.14 NA
ASFLA [38] 2019 1962.6. 627.4, 993.9 62, 16, 40 69.13 0.9913
SFLA [38] 2019 1088.7, 167.3, 980.9 57, 63, 26 65.43 0.9752
FWA [38] 2019 480.5, 1198.6, 225.8 65, 61, 27 65.39 0.9740
SSA [31] 2019 380,527, 1718 17, 10, 60 69.1 0.9789
WCA [30] 2018 775, 1105, 438 61, 62, 23 68.2 0.987
EMA [34] 2018 910.72, 1263.9, 689.06 69, 69, 50 69.30 0.9817
HGWO [10] 2017 527,380, 1718 11, 17, 61 69.14 0.98
BFOA [26] 2014 295.4, 447.6, 1345.1 27, 65, 61 66.56 0.9808
HSA [28] 2013 1302.4, 369, 101.8 63, 64, 65 61.4 0.967
GA [27] 2012 929.7,1075.2, 984.8 21, 62, 64 60.4 NA
PSO [27] 2012 1199.8, 795.6, 992.5 61,63,17 63.02 NA
GA-PSO [27] 2012 910.5, 1192.6, 884.9 21, 61, 63 63.9 NA
Proposed ABC 2020 1000, 200, 338.2 61, 51, 62 86.8 0.96

network is 3720 kW and 2300kVAR respectively. The NRLF analysis By performing the NRLF analysis with the proposed ABC algo-
of this system results in a Ploss of 232.9 kW and a vmin of 0.8778 (p. rithm Ploss has been minimized to 54 kW which shows a Pt_loss of
u) without DGs. The vavg of this system is 0.9297 (p.u) which 76.8%. Similarly, vD has been minimized to 0.011 (p.u) with a vmin
results in a vD of 0.0703 (p.u). CT without DGs is $4298861. of 0.9770 (p.u). Table 2 shows the comparison of results of the
NRLF analysis with the proposed ABC algorithm and other standard
1) Case-1 (Simultaneous minimization of Ploss and VD)

Fig. 8. Convergence curve of Ploss in IEEE 69-Bus Network (Case-1). Fig. 9. Voltage profile and vavg in IEEE 69-Bus Network (Case-1).
 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708 705

the proposed ABC algorithm and with GA to $2932969. Fig. 5 com-

pares the convergence of the proposed ABC algorithm with GA
while calculating Ploss. It can be seen that ABC algorithm converges
in less iterations than GA. Fig. 6 shows the voltage profile and vavg
for case-2 (with and without DGs).

4.2. IEEE 69-Bus radial network

The IEEE 69-bus radial network with 69 buses and 73 lines is

shown in Fig. 7. The total active and reactive load demand of this
system is 3802 kW and 2696kVAR respectively. The NRLF analysis
of this system results in a Ploss of 318.4.9 kW and a vmin of 0.8592 (p.
u) without DGs. The vavg of this system is 0.9592 (p.u) which
results in a vD of 0.0408 (p.u). CT without DGs is $5195114.

1) Case-1 (Simultaneous minimization of Ploss and VD)

By performing the NRLF analysis with the proposed ABC algo-

rithm Ploss has been minimized to 42 kW which shows a Pt_loss of
Fig. 10. Convergence curve of Ploss in IEEE 69-Bus Network (Case-2). 86.8%. Similarly, vD has been minimized to 0.0168 (p.u) with a vmin
of 0.9600 (p.u). Table 4 shows the comparison of results of the
NRLF analysis with the proposed ABC algorithm and other standard
algorithms (which were applied to minimize same objectives in algorithms (which were applied to minimize same objectives in
IEEE 33-bus radial network). It can be seen from Table 2 that for IEEE 69-bus radial network). It can be seen from Table 4 that for
same problem, highest Pt_loss and vmin (p.u) have been achieved same problem, highest Pt_loss has been achieved with the proposed
with the proposed ABC algorithm. Similarly, both of these objec- ABC algorithm. However, vmin has been minimized to 0.9600 (p.u).
tives have been minimized with fairly compact sizes of DGs as In this context, the other algorithms have given better results. Nev-
well. Fig. 3 shows the convergence curve of Ploss minimization with ertheless, it is acceptable that vmin is still in limits. Similarly, both
the proposed ABC algorithm for case-1. Fig. 4 shows the voltage objectives have been minimized with drastically compact sizes of
profile and vavg for case-1 (with and without DGs). DGs as well. Fig. 8 shows the convergence curve of Ploss with the
proposed ABC algorithm for case-1. Fig. 9 shows the voltage profile
2) Case-2 (Simultaneous minimization of Ploss, VD and ct) and vavg for case-1 (with and without DGs).

By performing the NRLF analysis with the proposed ABC algo- 2) Case-2 (Simultaneous minimization of Ploss, VD and ct)
rithm Ploss has been minimized to 56.9 kW which shows a Pt_loss
of 75.6%, vD has been minimized to 0.0145 (p.u) with a vmin of By performing the NRLF analysis with the proposed ABC algo-
0.9678 (p.u), and CT has been minimized to $2814898. Table 3 rithm Ploss has been minimized to 38.7 kW which shows a Pt_loss
shows the comparison of results of the NRLF analysis with the pro- of 87.8%, vD has been minimized to 0.0157 (p.u) with a vmin of
posed ABC algorithm and GA. It can be seen from Table 3, that Pt_loss 0.9644 (p.u), and CT has been minimized to $3706484. Table 5
has been significantly minimized with the proposed ABC algorithm shows the comparison of results of the NRLF analysis with the pro-
to 75.6% as compared to GA, which is 35.5%. Similarly, vD is mini- posed ABC algorithm and GA. It can be seen from Table 5, that Pt_loss
mized to 0.0145 (p.u) as compared to GA, which is 0.0434 (p.u). has been significantly minimized with proposed ABC algorithm to
CT without DGs is $4298861, and it is minimized to $2814898 with 87.8% as compared to GA which is 75.3%. Similarly, vD is minimized
to 0.0157 (p.u) by proposed ABC algorithm as compared to GA,
which is 0.0242 (p.u). CT without DGs is $5195114 and it is mini-
mized to $3706484 with the proposed ABC algorithm as compared
to GA, which is $4281574. Fig. 10 compares the convergence of
proposed the ABC algorithm with GA while calculating Ploss. It
can be seen that ABC algorithm converges in less iterations than
GA. Fig. 11 shows the voltage profile and vavg for case-2.

4.3. CIGRE MV benchmark grid

The CIGRE MV benchmark grid with 14 buses and 30 branches

is shown in Fig. 12. The total active and reactive load demand of
this system is 31310 kW and 7783kVAR respectively. The NRLF
analysis of this system results in a Ploss of 22170 kW and a vmin of
0.6363 (p.u) without DGs. The vavg of this system is 0.8761 (p.u)
which results in a vD of 0.01239 (p.u). CT without DGs is
$65816893.

1) Case-2 (Simultaneous minimization of Ploss, VD and ct)

By performing the NRLF analysis with the proposed ABC algo-

rithm Ploss has been minimized to 11400 kW which shows a Pt_loss
Fig. 11. Voltage profile and vavg in IEEE 69-Bus Network (Case-2). of 48.5%, vD has been minimized to 0.0764 (p.u) with a vmin of
706 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708

Fig. 13. Voltage profile and vavg in CIGRE MV Benchmark Grid (Case-2).

0.9236 (p.u), and CT has been minimized to $14374860. Table 6

shows a comparison of results of the NRLF analysis with the pro-
posed ABC algorithm and GA It can be seen from Table 6, that Pt_loss
has been significantly minimized with proposed ABC algorithm to
48.5% as compared to GA, which is 30.17%. However, vD is 0.0764
(p.u) which is approximately same when compared to GA, which
is 0.0765 (p.u). CT without DGs is $65816893, and it is minimized
to $14374860 with the proposed ABC algorithm, and with GA to
$19143103. Fig. 13 shows the voltage profile and vavg for case-2.

5. Conclusion

In current paper, the optimal sizing and placement of DGs has

been done with a novel concept to simultaneously minimize total
power loss, average voltage drop and total energy costs. The NRLF
analysis is conducted on IEEE 33 and 69-bus radial networks and
Fig. 12. CIGRE MV benchmark grid. on CIGRE MV benchmark grid. ABC algorithm is proposed to simul-
taneously minimize considered multiple objectives. Two test cases
have been solved by using proposed ABC algorithm. In the 1st case,
Table 5 two objectives including total active power loss and average volt-
Comparison for IEEE 69-Bus Radial Network (Case-2)
age drop were simultaneously minimized and results were com-
System Without With DGs (with With DGs (with pared with other standard algorithms. It has been seen that
Parameter DGs GA) ABC) proposed ABC algorithm almost gives better results in both IEEE
DGs’ Location – 64, 62, 61 62, 17, 64 33 and 69-bus radial networks. In the 2nd case, three objectives,
DGs’ Sizes (kW) – 233.6, 210, 540.9 1000, 200, 200 including total active power loss, average voltage drop and total
Ploss (kW) 318.4 78.6 38.7
energy cost were simultaneously minimized, and results were
Pt_loss (%) – 75.3 87.8
vmin (p.u) 0.8592 0.9434 0.9644 compared with GA. It has been seen that proposed ABC algorithm
vavg (p.u) 0.9592 0.9758 0.9843 gives minimized results of total active power loss, average voltage
VD (p.u) 0.0408 0.0242 0.0157 drop and total energy cost than GA in both IEEE 33 and 69-bus
CT ($) 5,195,114 4,281,574 3,706,484 radial networks. In CIGRE MV distribution grid, ABC algorithm
gives better results than GA for total active power loss and total
energy cost, however, average voltage drop seems to be at par with
Table 6 both algorithms. Moreover, proposed ABC algorithm converges in
Comparison for CIGRE MV Benchmark Grid (Case-2) less iterations than GA in both cases.
System Without With DGs (with With DGs (with
Parameter DGs GA) ABC)
DGs’ Location – 8, 11, 12 12, 12, 12 References
DGs’ Size (kW) – 1200, 1100, 1000 1000, 1000, 1000
Ploss (kW) 22,170 15,480 11,400 [1] Nath Vijay, Mandal Jyotsna Kumar. Optimal placement of distributed
Pt_loss (%) – 30.17 48.5 generators using genetic algorithm approach. In: Preceding of 2nd
vmin (p.u) 0.6363 0.7151 0.7708 International Conference on Microelectronics, Computing & Communication
vavg (p.u) 0.8761 0.9235 0.9236 System, vol 476, 2017, pp. 587.
[2] Saha Subhodip, Mukherjee Vivekananda. Optimal placement & sizing of DGs in
vD (p.u) 0.1239 0.0765 0.0764
RDS using Chaos embedded SOS Algorithm. In: IET Gener Transm Distrib, 2016,
CT ($) 65,816,893 19,143,103 14,374,860
Vol. 10, Iss. 14, pp. 3671–3680.
 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708 707

[3] Hyoung Soo. Optimal placemet & sizing of multiple DGs in a practical [30] El-Ela AAA, El-Sehiemy RA, Abbas AS. Optimal placement and sizing of
distribution system by considering power loss. IEEE Transactions on Industry distributed generation and capacitor banks in distribution systems using
Applications, vol. 49, NO. 5, September/October 2013. water cycle algorithm. IEEE Syst J Dec. 2018;12(4):3629–36.
[4] Choton K. Das, Octavian Bass, Ganesh Kothapalli, Thair S. Mahmoud, Daryoush [31] Sambaiah KS, Jayabarathi T. Optimal allocation of renewable distributed
Habibi. Optimal placement of distributed energy storage systems in generation and capacitor banks in distribution systems using Salp Swarm
distribution networks using artificial bee colony. In: Applied Energy, vol 232, algorithm. Int J Renew Energy Res 2019; 9(1).
2018, pp. 212–228. [32] Selim Ali, Kamel Salah, Jurado Francisco. Efficient optimization technique for
[5] Mehta Pankita, Bhatt Praghnesh, Pandya Vivek. Optimal selection of multiple DG allocation in distribution networks. Appl Soft Comput 2020;86.
distributed generating units and its placement for voltage stability [33] Meena NK, Swarnkar A, Gupta N, Niazi KR. Multi-objective taguchi approach
enhancement and energy loss minimization. Ain Shams Eng J 2018;9: for optimal DG integration in distribution systems. IET Gener Transm Distrib
187–201. 2017;11(9):2418–28.
[6] Singh D, Verma K. GA based energy loss minimization approach for optimal [34] Daneshver Mohamamd Reza, Babaei Ebrahim. Exhange market algorithm for
sizing & placement of distributed generation. Int J Knowl Intell Eng Syst multiple DG placement and sizing in a radial distribution system. J Energy
2008;12:147–56. Manage Technol (JEMT) 2018; 2(1).
[7] Abu-Mouti FS, El-Hawary ME. Heuristic curve-fitted technique for distributed [35] Truong Khoa H, Nallagownden Perumal, Baharudin Zuhairi, Vo Dieu N. A
generation optimization in radial distribution feeder systems. Proc IET Gener Quasi-Oppositional-Chaotic Symbiotic Organisms Search algorithm for global
Trans Distrib 2011;5(2):172–80. optimization problems. Appl Soft Comput J 2019;77:567–83.
[8] Prommee WO, Ongsakul W. Optimal multiple distributed generation [36] Hamida Imen Ben, Salah Saoussen Brini, Msahli Faouzi, Mimouni Mohamed
placement in microgrid system by improved reinitialized social structures Faouzi. Optimal network reconfiguration and renewable DG integration
particle swarm optimization. Eur Trans Electr Power 2011;21(1):489–504. considering time sequence variation in load and DGs. Renew Energy 2018;
[9] Rama Prabha D, Jayabarathi T, Umamageswari R, Saranya S. Optimal location 121: pp 66–80.
and sizing of distributed generation unit using intelligent water drop [37] Veera Vallem, Satya Venkata, Murty Narayana, Kumar Ashwani. Optimal DG
algorithm. Sustain Energy Technol Assess 2015;11:106–13. integration and network reconfiguration in microgrid system with realistic
[10] Sanjay R, Jayabarathi T, Raghunathan T, Ramesh V, Mithulananthan N. Optimal time varying load model using hybrid optimization. IET smart grid; 2019.
allocation of distributed generation using hybrid grey wolf optimizer. IEEE [38] Onlam Arun, Daranpob Yodphet, Rongrit Chatthaworn, Chayada Surawanitkm,
Access 2017;5:14807–18. Apirat Siritaratiwat, Pirat Khunkitti. Power loss minimization and voltage
[11] Sudabattula Suresh Kumar, Kowsalya M. Optimal allocation of solar based stability improvement in electrical distribution system via network
distributed generators in distribution system using Bat algorithm. Perspect Sci reconfiguration and distributed generation placement using novel adaptive
2016;8:270–2. shuffled frogs leaping algorithm. Energies 2019; 12: p553.
[12] Kansal Satish, Tyagi Barjeev, Kumar Vishal. Cost-benefit analysis for optimal [39] Abu-Mouti FS, El-Hawary ME. Optimal distributed generation allocation and
DG placement in distribution system. Int J Ambient Eng, 2015 sizing in distribution systems via Artificial Bee Colony algorithm. IEEE Trans
[13] ChithraDevi SA, Lakshminarasimman L, Balamurugan R. Stud Krill herd Power Deliv Oct. 2011;26(4):2090–101.
Algorithm for multiple DG placement and sizing in a radial distribution [40] Karaboga Dervis, Gorkemli Beyza, Ozturk Celal, Karaboga Nurhan. A
system. Eng Sci Technol, Int J 2017;20(2):748–59. comprehensive survey: Artifical Bee Colony(ABC) algorithm and applications.
[14] Akbari K, Rahmani E, Abbasi A, Askari M. Optimal placement of distributed In: Artif Intell Rev vol 42; 2014. pp. 21–57.
generation in radial networks considering reliability and cost indices. J Intell [41] Rudion K, Orths A, Styczynski ZA, Strunz K. Design of benchmark of medium
Fuzzy Syst 2016;30(2):1077–86. voltage distribution network for investigation of DG integration. In: IEEE;
[15] Kansal S, Kumar V, Tyagi B. Optimal placement of different types of DGs in 2006.
distribution networks. Elect Power Energy Syst 2013;53:752–60. [42] Zahid M, Chen J, Li Y, Duan X, Lei Q, Bo W, et al. New approach for optimal
[16] Candelo-Becerra John Edwin, Hernández-Riaño Helman Enrique. Distributed location and parameters setting of UPFC for enhancing power systems stability
generation placement in radial distribution networks using a bat-inspired under contingency analysis. Energies 2017;10:1738.
algorithm. DYNA 2015;82(192):60–7. [43] Teng Jen Hao. A modified Gauss-Siedel algorithm of three phase power flow
[17] Devi S, Geethanjali M. Application of modified bacterial foraging optimization analysis in distribution network. Electr Power Energy Syst 2002;24:97–102.
algorithm for optimal placement and sizing of distributed generation. Expert [44] Karaboga D. An idea based on honey bee swarm for numerical optimization.
Syst Appl 2014;41:2772–81. Technical Report-TR06, Erciyes University, Engineering Faculty, Computer
[18] Mahmoud K, Yorino N, Ahmed A. Optimal distributed generation allocation in Engineering Department; 2005.
distribution systems for loss minimization. IEEE Trans Power Syst March [45] Karaboga Dervis, Basturk Bahriye. A powerful and efficient algorithm for
2016;31(2):960–9. numerical function optimization: Artificial bee colony (abc) algorithm. J Global
[19] Ayodele TR, Ogunjuyigbe ASO, Akinola OO. Optimal location, sizing, and Optim 2007;39:459–71.
appropriate technology selection of distributed generators for minimizing [46] Karaboga Dervis, Basturk Bahriye. An artificial bee colony (abc) algorithm for
power loss using genetic algorithm. J Renew Energy 2015. numeric function optimization. In: IEEE Swarm Intelligence Symposium,
[20] Duong MQ, Pham TD, Nguyen TT, Doan AT, Tran HV. Determination of optimal Indianapolis, Indiana, USA; 2006.
location and sizing of solar photovoltaic distribution generation units in radial [47] Karaboga Dervis, Basturk Bahriye. On the performance of artificial bee colony
distribution systems. Energies 2019;12:174. (ABC) algorithm. Appl Soft Comput 2008;8:687–97.
[21] Mohd Zaid N et al. Multi-objective optimization for sizing of distributed [48] Eajal AA, Al Hajri MF, El-Saadany EF. Distributed generation planning in smart
generation using cuckoo search algorithm. Appl Mech Mater 2015; distribution grids via a meta-heuristic approach. Electrical and Computer
785:34–7. Engineering (CCECE), 2014 IEEE 27th Canadian Conference on; 2014. pp. 1–6.
[22] Nguyen Thuan Thanh, Truong Anh Viet. Distribution network reconfiguration [49] Dawoud Samir M, Lin Xiangning, Okba Merfat I. Optimal placement of
for power loss minimization and voltage profile improvement using cuckoo different types of RDGs based on maximization of microgrid. J Clean Prod
search algorithm. Int J Electr Power Energy Syst 2015;68:233–42. 2017.
[23] Rama Prabha D, Jayabarathi T. Optimal placement and sizing of multiple
distributed generating units in distribution networks by invasive weed
optimization algorithm. Ain Shams Eng J 2016;7:683–94.
[24] Kansal Satish, Kumar Vishal, Tyagi Barjeev. Hybrid approach for optimal Dr. Essam A. Al-Ammar (SM) worked as a Power/Soft-
placement of multiple DGs of multiple types in distribution networks. Int J ware Engineer at Lucent Technologies in Riyadh for two
Electr Power Energy Syst 2016;75:226–35. years. In 2003, he received his MS degree from Univer-
[25] El-Fergany Attia. Optimal allocation of multi-type distributed generators using sity of Alabama, Tuscaloosa, AL, and Ph.D. degree from
backtracking search optimization algorithm. Int J Electr Power Energy Syst Arizona State University in 2007. He is an Associate
2015;64:1197–205. Professor in Electrical Engineering Department, King
[26] Mohamed Imran A, Kowsalya M. Optimal size and siting of multiple Saud University, Riyadh, Saudi Arabia. He was Gover-
distributed generators in distribution system using bacterial foraging nor’s Advisor in Electricity Co-Generation Regulatory
optimization. Swarm Evol Comput 2014;15:58–65. Authority (ECRA). Also, he was an advisor at Ministry of
[27] Moradi MH, Abedini M. A combination of genetic algorithm and particle Water and Electricity (MOWE) and a former Energy
swarm optimization for optimal DG location and sizing in distribution Consultant for Riyadh Techno Valley (RTV). He involved
systems. Int J Electr Power Energy Syst 2012;34(1):66–74.
in many local and international committees in different
[28] Rao RS, Ravindra K, Satish K, Narasimham SVL. Power loss minimization in
aspects in the area of power and electrical energy. His current interests include high
distribution system using network reconfiguration in the presence of
distributed generation. IEEE Trans Power Syst Feb. 2013;28(1):317–25. voltage engineering, power system transmission and distribution as well as
[29] Kefayat M, Lashkar Ara A, Nabavi Niaki SA. A hybrid of ant colony optimization renewable energy and smart grid. He is a Senior Member of IEEE since 2007 and
and artificial bee colony algorithm for probabilistic optimal placement and sizing Saudi Engineering Committee since 1997. He published nearly 110 papers and 20
of distributed energy resources. Energy Convers Manage 2016; 92: 149–161. patents in Energy and Water.
708 E.A. Al-Ammar et al. / Ain Shams Engineering Journal 12 (2021) 697–708

Ms. Kiran Farzana is an Electrical Engineer by profes- Dr. Azhar Ul Haq has been working as an Assistant
sion. She did her master’s degree in Electrical Engi- Professor of Electrical Engineering at National Univer-
neering from Bahria University Islamabad Pakistan. Her sity of Sciences and Technology, Islamabad since Oct
research interests include optimization of distribution 2016. He worked as a as a Research Assistant with ECE
systems and microgrids Department of University of Waterloo, Canada for year
2015. He holds Ph.D. in Electrical Engineering obtained
under the joint Research Doctoral Program of University
of L’Aquila, Italy and University of Waterloo, Canada. His
research interests include grid integration of renewable
energies, PV powered smart charging of electric vehi-
cles; and direct load control strategies for peak-shaving
in power system.

Dr. Asad Waqar received his Electrical Engineering Dr. Muhammad Zahid is working as Assistant Professor
degree from UET Taxila in 2002. He did his master’s in department of Electrical Engineering at Riphah
degree in Electrical Power Engineering from RWTH International University Lahore, Pakistan. He did his
Aachen Germany in 2011 and PhD degree in Electrical PhD degree in Electrical Engineering from Huazhong
Engineering from Huazhong University of Science and University of Science and Technology China in 2018. His
Technology China in 2016. Currently he is working as research interests include microgrids operation and
Associate Professor in department of Electrical Engi- control, FACTS and planning of distribution systems.
neering at Bahria University Islamabad, Pakistan. His
research interests include smart grids, microgrids
operation and control, power quality, power electronics
and network reinforcement planning.

Dr. Muhammad Aamir is working as Assistant Profes- Ms. Memoona Batool is an Electrical Engineer by pro-
sor in department of Electrical & Computer Engineering fession. She did her master’s degree in Electrical Engi-
at PAF-IAST Haripur Pakistan. He did his PhD degree in neering from Bahria University Islamabad Pakistan. Her
Electrical Engineering from University of Malaya research interests include optimization of distribution
Malaysia in 2017. His research interests include power systems, microgrids and DSM.
electronics, power systems and smart grids.

Dr. Saifullah is working as Assistant Professor in

department of Industrial Engineering at University of
Engineering and Technology Taxila, Pakistan. He did his
PhD degree in Electrical Engineering from Huazhong
University of Science and Technology China in 2015. His
research interests include application of heuristic
methods to real time problems.