This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE SYSTEMS JOURNAL 1

Optimal Placement and Sizing of Distributed

Generation and Capacitor Banks in Distribution
Systems Using Water Cycle Algorithm
Adel A. Abou El-Ela, Ragab A. El-Sehiemy, and Ahmed Samir Abbas

Abstract— Integration of distributed generation units (DGs) and Ri Resistance of ith branch in the network.
capacitor banks (CBs) in distribution systems aim to enhance the LB and UB Lower and upper boundaries, respectively.
system performance. This paper proposes water cycle algorithm PF DG units’ power factor.
(WCA) for optimal placement and sizing of DGs and CBs. The
proposed method aims to achieve technical, economic, and environ- PL an d QL Active and reactive power losses,
mental benefits. Different objective functions: minimizing power respectively.
losses, voltage deviation, total electrical energy cost, total emissions Pd and Qd Load demand active and reactive powers.
produced by generation sources and improving the voltage stabil- PrGrid Generated power cost at substations.
ity index are considered. WCA emulates the water flow cycle from C Generation cost for each source.
streams to rivers and from rivers to sea. Five different operational
cases are considered to assess the performance of the proposed a and b Fixed and variable generation cost
methodology. Simulations are carried out on three distribution coefficients.
systems, namely IEEE 33-bus, 69-bus test systems, and East Delta Gr Annual rate of benefit ($/h).
network, as a real part of Egyptian system. The simulated results LF Load factor of DGs.
demonstrate the effectiveness of the proposed method compared ei , Cci CBs installation and purchase cost.
with other optimization algorithms. Also, the results demonstrate
that the proposed WCA gives superior performance for the system QCi Reactive power of the CBs.
and give distinguished improvements in both economic and envi- EGrid and EDGi Emissions generated by the grid and ith
ronmental benefits. Moreover, the results give the flexible operation DG unit, respectively
with controllable power factor DGs that is better than those using Rand Random number between 0 and 1.
DGs at fixed power factor. Rij and Xij Resistance and reactance of branch ij, re-
Index Terms—Capacitor bank (CB), distributed generation spectively.
(DG), power loss, voltage deviation (VD), voltage stability index m1, m2 Slack bus, bus number (2, 3, . . . , nbus).
(VSI), water cycle algorithm (WCA). ffi Fitness functions of a raindrop.
RP Raindrops population matrix.
NOMENCLATURE Npop, Nvar Number of raindrops and design variables,
respectively.
O& M cost Operation and maintenance cost. i i
Xstream and Xriver Position of the stream and river for ith
CO2, SO2, NOx Carbon dioxide, sulfur dioxide, and nitro- iteration, respectively.
gen oxides, respectively. N Sn Number of streams which flow into rivers
nL Number of branches. or sea.
PV Photovoltaic power plant. Nsr Number of rivers and sea.
WT and GT Wind and gas turbine power plants, dmax Very small number near zero.
respectively. nbus Total number of system buses.
NDG Numbers of DG units. Vi and vi Voltage at ith bus and its magnitude.
P Gi and QGi Active and reactive power at ith DG unit,
respectively.
Manuscript received February 6, 2017; revised April 22, 2017 and July 3, Pg Grid Real power produced by substation.
2017; accepted August 22, 2017. (Corresponding author: Ahmed Samir Abbas.)
A. A. A. El-Ela is with the Faculty of Engineering, Electrical Engineering QCBtotal Reactive power of CBs.
Department, Menoufiya University, Shebeen El-Kom 32511, Egypt (e-mail: Ii Current magnitude of ith branch.
draaa50@hotmail.com). visp ec Specified voltage magnitude (1.0 p.u.).
R. A. El-Sehiemy is with the Faculty of Engineering, Electrical Engineering
Department, and the Intelligent Systems Reseach Group, Kafrelsheikh Univer- k1, k2 , and k3 Priorities factors which are as 0.5, 0.25,
sity, Kafrelsheikh 33516, Egypt (e-mail: elsehiemy@eng.kfs.edu.eg). and 0.25, respectively.
A. S. Abbas is with the Mechanical and Electrical Research Institute, NWRC, Life time Investment years of DGs/CBs.
Cairo 13621, Egypt (e-mail: Ahmd_sameer@yahoo.com).
Digital Object Identifier 10.1109/JSYST.2018.2796847 min, max Minimum and maximum limits.

1937-9234 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications standards/publications/rights/index.html for more information.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE SYSTEMS JOURNAL

I. INTRODUCTION C. Paper Contribution

A. Motivation This study proposes the water cycle algorithm (WCA) to
obtain the optimal DGs/CBs placement and sizing in distribution
NSTALLATION and integration of DGs in distribution sys-
I tems can provide several technical, economic, and environ-
mental benefits. The technical benefits are power loss reduction,
systems. The proposed method aims to realize the following
benefits.
1) Studying the penetration of DGs and CBs to enhance the
voltage profile improvement, and power quality enhancement
technical, environmental, and economic issues of distri-
[1]. The environmental benefits are the reduction of pollution
bution systems.
and emission levels in the system. Economic benefits aim to
2) Three technical objectives are satisfied that are: power
reduce the operational costs as much as possible and to increase
loss reduction, voltage profile improvement, and stability
the profits of all system participants.
index enhancement.
On the other hand, CBs installation in distribution system can
3) Two economic issues are considered as minimizing the
produce reactive power that able to improve load-bus voltage
costs of generated power and CBs.
and to reduce the distribution power losses. Thus, the required
4) Reducing the emission is considered as environmental
reactive power demand from the main grid is reduced [2]. Added
benefit for achieving clean operation.
to that, the voltage fluctuations caused by some types of DGs can
5) Providing a controllable power factor strategy for flexible
be reduced by using (fixed-switched) capacitor banks (CBs) [3].
operation of distribution systems.
Therefore, it is expected that a hybrid penetration of DGs and
6) Three operational cases of DGs/CBs are considered with
CBs reduce distribution power losses, improve voltage profile
single objective optimization to find the effectiveness of
and therefore enhance the overall distribution system perfor-
the proposed WCA compared with other techniques.
mance. But to achieve the earlier mentioned benefits and to de-
7) Two additional multiobjective cases are considered to esti-
crease voltage deviation (VD) problems, placement and sizing
mate the technical, economic, and environmental impacts
of DGs and CBs need to be investigate using proper optimization
of the optimal placement and sizing of DGs/CBs.
tools [4].
8) Appling the proposed method to standard and real radial
distribution system.
B. Literature Survey 9) Increasing the awareness of the importance of combined
penetration of DGs and CBs for enhancing the operation
1) DG Units Placement: DG placement has been studied of electrical systems.
using several techniques. El-Sehiemy and Ahmed [5] studied
the insertion of the DGs into distribution systems and the ef-
fect of its placement on protective devices setting. In [6], a cat D. Paper Organization
swarm optimization method was proposed for finding the opti- This paper is organized as follows. Section I presents the
mal placement and sizing of DGs in the network to reduce total introduction, Section II presents the problem formulation,
power loss, total generation cost and emissions. Injeti and Kumar Section III presents the structure of water cycle optimization
[7] proposed a particle swarm optimization (PSO) algorithm to algorithm. Applications of the proposed placement and siz-
find the optimal placement and sizing of the DGs in distribution ing method are introduced and discussed in Section IV, while
systems to minimize power loss. Saravanamutthukumaran and Section V presents the conclusion of this paper.
Kumarappan [8] considered different load levels while finding
the optimal placement of the DG using multiobjective optimiza-
tion technique. Amini et al. [9] utilized a two stages optimization II. PROBLEM FORMULATION
method for getting the optimal placement of renewable DG and
The objective functions (OFs), equality and inequality con-
electric vehicles parking lots.
strains are introduced for optimal placement and sizing of DGs
2) CBs Placement: Optimal placement and sizing of CBs
and CBs in distribution systems as follows.
can offer several benefits, such as minimizing power loss cost
and harmonic distortion level [10]. Optimization techniques for
CBs placement and sizing problem can be classified into many A. Objective Functions
categories [11], such as numerical methods, heuristic, and other The proposed method aims to achieve three types of OFs:
optimization techniques. A review of CBs placement and sizing technical, economic, and environmental OFs.
techniques has been presented in [12]. In [13], the iterative allo- 1) Technical OF: Three technical OFs are considered in this
cation method of CBs is introduced for reactive power planning section. The first one aims to minimize the distribution
problem. power losses (f1 ) that can be expressed as [6]
3) Hybrid DGs/CBs Placement: Saonerkar and Bagde [14]
proposed genetic algorithm for getting the optimal placement 
nL

and sizing of combined DGs and CBs in distribution systems. f1 (x) = min Ri ∗ |Ii |2 . (1)
i=1
Jannat and Savic [15] proposed a method for solving the place-
ment and sizing problem considering the effect of renewable The second technical OF aims to improve the voltage pro-
energy uncertainty. file and preserve better voltage profile. This function can be
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

EL-ELA et al.: OPTIMAL PLACEMENT AND SIZING OF DISTRIBUTED GENERATION AND CBs IN DISTRIBUTION SYSTEMS 3

described as [16] 1) Generation operating limits

N 
 2
vi − visp ec PGmi
in
≤ PGi ≤ PGm i
ax
, QGm
i
in
≤ QGi ≤ QGm
i
ax
.
f2 (x) = min . (2) (15)
i=0
vim ax − vim in
2) Installed capacitors limits
Voltage stability is one of the most significant indices. The
third OF (f3 ) for voltage stability index (VSI) can be described QCB
total < Qd . (16)
as follows [16]: 3) Bus voltage limits
 
1 0.95 ≤ Vi ≤ 1.05, i = 1, 2, . . . , nbus. (17)
f3 (x) = min (3)
VSI(m2 )
4) DG Power factor limit
where
0.8 ≤ PF ≤ 1. (18)
VSI(m2) = ABS(|V (m1)|4 − 4 ∗ [P (m2) ∗ Xij − Q(m2) ∗ Rij]2
− 4 ∗ [P (m2 ) ∗ Rij + Q(m2 ) ∗ Xij ] ∗ |V (m1 )|2 ) . III. PROPOSED WCA
(4) A. WCA Procedure

2) Economic OF: The economical OF (f4 ) aims to minimize WCA [18] begins with an initial population similar to other
the power generation costs that can be calculated from [2] metaheuristic algorithms, this initial population called the rain-
and [17]. drops (RD). The values of the problem controlled variables xi
(PGi , QGi , QCB
total , and placement of DG and CB) can be formed

N DG
as an array called “RD” for single solution. This array can be
f4 (x) = min (CDG i + Csub + CCB ) (5)
defined as follows:
i=1

where RD = [x1 , x2 , x3 , . . . , xN ] . (19)

CDG i = a + b ∗ PGi (6) RP matrix contains random solutions in iteration #1 as


capital cost($/kW) ∗ capacity(kW) ∗ Gr RP = xjk : j = 1 : Np op and k = 1 : Nvar . (20)
a= ($) (7)
life time( year) ∗ 8760 ∗ LF Then, (ffi ) is obtained from
b = O&M cost ($/kW·h) + fuel cost ($/kW·h) (8) ffi = f (xi1 , xi2 , xi3 , . . . , xiN var ), i = 1, 2, 3, . . . , Np op . (21)
Csub = Pg Grid ∗ Pr (9) After that, the best RD is selected to be the sea, number of
Grid
N C good RD is chosen to be the rivers and the remainder RD are
(ei + Cci |Qci |)
i=1 assumed to be streams that flow to the sea or the rivers. Equation
CCB = . (10)
lifetime ∗ 8760 (22) calculates the streams that flow to a sea or a river depending
3) Environmental OF: (Minimization of Generation Units’ on the flow intensity as
Emissions (f 5)): CO2 , SO2 , and NOx are considered as
  
 ff 
most effective pollutants in power generation sources. The  n 
NSn = round  N s r  × Np op , n = 1, 2, 3, . . . , Nsr .
mathematical formulation of this OF can be described as  i=1 ffi 
follows [9]: (22)
A stream flows to the river along the path between them

N DG

f5 (x) = EDG i + EGrid (11) using a random distance (x) [18]. The same concept is applied
i=1 for flowing rivers to the sea, so the new position for the streams
  and rivers can be given as
EDG i = CODG
2 + NODG
x + SODG
2 × PGi (12)
 
i+1
Xstream i
= Xstream + rand ∗ U ∗ ( Xriver
i
− Xstream
i
) (23)
EGrid = COGrid
2 + NOGrid
x + SOGrid
2 × Pg Grid .
(13)
i+1
Xriver i
= Xriver + rand ∗ U ∗ (Xsea
i
− Xriver
i
) (24)
where U is a constant and it equals to 2.To avoid getting
B. Constraints trapped in local optimal it is assumed that an evaporation
1) Equality Constraints: The constraints for power balance process will happen and the clouds will be formed and then
requirements it will begin to rain (new random solutions). The condition
in (25) is checked, if it is satisfied then evaporation will be

NG 
NG
PGi − PL = Pd QGi − QL = Qd . (14) started.
 i 
i=1 i=1 Xsea − Xriver
i  < dm ax , i = 1, 2, 3, . . . , Nsr − 1. (25)
2) Inequality Constraints: Maximum admissible generated
After each evaporation process, the value of dm ax is as
power from DGs/CBs should not exceed to permissible limita-
m ax = dm ax − (dm ax / max iteration).
di+1 i i
tions of the distribution systems. (26)
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4 IEEE SYSTEMS JOURNAL
Fig. 1. Flow chart of the proposed method.
The raining process starts after evaporation process. The new
streams are formed due to newly RD (random solution).
The streams that directly flow to the sea [see (27)] aim to
encourage streams production to improve the optimal solution
for constrained problems.
√ objectives (f1 , f4 , and f5 ) are optimized. Therefore, the OF can
new
Xstream = Xsea + µ × rand(1, Nvar ) (27)
where µ is the coefficient that shows the range of searching near
the sea and its value is usually equals to 0.1.
B. Investigation WCA for Placement and Sizing Problem
Fig. 1 shows the proposed steps of solving placement and
sizing problem using WCA as follows.
Step 1: Loading distribution system data and defining the
power limits of DGs and CBs in the system.
Step 2: Identifying the WCA parameters.
Step 3: Randomly initialize the set of RD (solution).
Step 4: Checking the system constrains for each solution.
Step 5: Determining the fitness function of each RD.
Step 6: Determining the best solution in the rain drops.
Step 7: Generating the new set of solutions.
Step 8: Repeating steps from 4–7 until stopping criteria or
maximum iteration is satisfied.
IV. APPLICATIONS
A. Test Distribution Systems
The proposed WCA is applied to three distribution sys-
tems. These systems are IEEE 33-bus distribution system
[17], IEEE 69-bus distribution system [18], and the East
Delta network (EDN) as a real part of the Egyptian distri-
bution network [20]. The total real and reactive loads of the
three systems are 3.715 + j2.3 MVA, 3.802 + j2.694 MVA, and
22.441 + j14.162 MVA, respectively.
The power flow calculations are carried out using backward
forward sweeping [20], [21]. The total real power losses for the
three distributions systems in the initial case are 202.67, 225,
and 805.73 kW, respectively.
B. Cases Studied
Five operational cases are introduced to show the effective-
ness of the proposed WCA and to study the impact of DGs and
CBs installation on the system performance.
Case 1: Single OF, power loss minimization, is considered
for placement and sizing of CBs alone.
Case 2: Single OF, power loss minimization, is considered
for placement and sizing of DGs that operate at unity PF.
Case 3: Single OF power loss minimization is considered
for optimal placement and sizing combination of CBs/DGs.
Case 4: Multiobjective optimal placement and sizing of
CBs/DGs. The DGs are operated with controllable PF. Three
technical objectives (f1 , f2 , f3 ) are considered. The multiOF is
implemented by using weighting factors approach. The OF can
be expressed as follows:
OF = min(k1 f1 + k2 f2 + k3 f3 ). (28)
Case 5: Multiobjective optimal placement and sizing of
multi CBs/DGs. DGs are controllable units to supply active
and reactive power. Technical, economic, and environmental
be formulated as follow:
OF = min(k1 f1 + k2 f4 + k3 f5 ). (29)
The economic and environmental properties of DGs are de-
pendent on their type. Table I shows the characteristics of the
considered DGs which are customized from [9] and [17].
In order to increase the DGs penetration level, three types of
DGs are considered (PV, WT, and GT). Generated power costs
at substations are considered to be 0.044 $/kW·h [17]. ei and
Cci are taken equal to 1000 and 30 000 $/Mvar, respectively [2].
The generated power cost in EDN grid is (0.0625 + 0.72 Prsub )
$/h according to [22].
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

EL-ELA et al.: OPTIMAL PLACEMENT AND SIZING OF DISTRIBUTED GENERATION AND CBs IN DISTRIBUTION SYSTEMS 5

TABLE I TABLE IV
CHARACTERISTICS OF DGS PLACEMENT AND SIZING OF DGS AND CBS FOR 33-BUS SYSTEM (CASE 3)

DG Rated Life Capital O&M Fuel cost Emission factors Method Power loss DG size (MW) Capacitor size (MVAR) Min. voltage
type capacity time cost costs ($/kWċh) (lb/MWh) (kW) and location and location (p.u.)
(MW) (year) ($/kW) ($/kWċh)
NO X SO 2 CO 2 GA [14] 71.25 0.25 (16), 0.25 0.30 (15), 0.30 (18) 0.971∗
(22), 0.50 (30) 0.30 (29), 0.60 (30),
Grid 25 25 – – 0.044 5.06 11.6 2031 0.30 (31)
PV 1 20 3985 0.01207 – – – – BFOA [23] 41.41 0.542 (17), 0.163 (18), 0.338 (33), 0.978∗
WT 5 20 1822 0.009 52 – – – – 0.160 (18), 0.541 (30),
GT 3 12 1224 0.06481 0.0667 0.279 0.93 1239.2 0.895 (33)
WCA 24.688 0.973 (25), 0.465 (23), 0.565 (30) 0.980 (33)
1.04 (29), 0.535 (14)
0.563 (11)
TABLE II
OPTIMAL PLACEMENT AND SIZING OF CBS FOR 33-BUS SYSTEM (CASE 1) ∗
Placement of minimum bus voltage has not been mentioned.

Method Power loss CBs size (MVAR) and Min. voltage TABLE V
(kW) location (p.u.) MULTIOBJECTIVE PLACEMENT AND SIZING FOR 33-BUS SYSTEM (CASE 4)

BFOA 144.04 0.349 (18), 0.821 (30), 0.936∗

[23] 0.277 (33) Method DG Capacitor Min.
CSA 131.5 0.6 (11), 0.3 (33), 0.943∗ size voltage
[24] 0.45 (24), 0.6 (30) Power Size (MW) (MVAR) (p.u.)
PSO 132.48 0.9 (2), 0.45 (7), 0.45 (31), 0.945∗ factor and location and location
[24] 0.3 (15), 0.45 (29)
WCA 130.912 0.3973 (14), 0.4511 (24), 0.951 (18) WCA 0.905 0.9917 (11) 0.325 (19) 0.989 (18)
1.0 (30) 0.985 0.9823 (31) 0.3116 (23)
0.959 1.652 (24) 0.5432 (30)
∗ OF f1 = 19.848 f2 = 0.041 f3 = 0.015
Placement of minimum bus voltage has not been mentioned.

TABLE III
OPTIMAL PLACEMENT AND SIZING OF DGS FOR 33-BUS SYSTEM (CASE 2)
The proposed WCA gives a significant reduction in total
active power loss to be 71.052 kW with reduction of 64.9%
Method Power DG size (MW) and location Min.
loss voltage
referred to initial case. Three DGs are installed at buses 14,
(kW) (p.u.) 24, and 29 with penetration 0.8546, 1.017, and 1.181 MW, re-
spectively. The minimum voltage level (0.973) is obtained at
FWA [16] 88.68 0.5897 (14), 0.189 (18), 1.0146 (32) 0.968∗
BFOA [23] 98.3 0.633 (17), 0.09 (18), 0.9470 (33) 0.964∗
bus 33.
HSA [25] 96.76 0.5724 (17), 0.107 (18), 1.0462 (33) 0.967 (29) Table IV represents the results of Case 3 that shows that the
TM [26] 91.305 0.5876 (15), 0.1959 (25), 0.783 (33) 0.958 (30) WCA gives superior power loss reduction compared with Cases
GA/PSO [27] 103.4 0.9250 (11), 0.8630 (16), 1.2 (32) 0.980 (25)
PSO [27] 105.35 1.1768 (8), 0.9816 (13), 0.8297 (32) 0.980 (30)
1 and 2. Case 3 suggests installing three DGs at buses 11, 25,
GA [27] 106.3 1.50 (11), 0.4228 (29), 1.0714 (30) 0.981 (25) and 29 and three CBs at buses 14, 23, and 30. A significant
WCA 71.052 0.8546 (14), 1.1017 (24), 1.181 (29) 0.973 (33) power loss reduction (24.688 kW) is achieved compared to GA
∗
(71.25 kW) and BFOA (41.41 kW). The minimum voltage level
Placement of minimum bus voltage has not been mentioned.
of 0.98 p.u. is obtained at bus 33.
From the results of Case 4 that is shown in Table V, it is clear
C. Results and Comments that the optimal placement and sizing of DGs (with controllable
1) Results of 33-Bus Network: Results of case 1 are shown PF (and CBs will be very effective to minimize system power
in Table II where the optimal placement and sizing of CBs loss, improve VSI and voltage deviation. A comparison between
is determined using the proposed WCA. The results of this bus voltage profiles of different cases can be seen in Fig. 2.
case are compared with other algorithms [bacterial foraging Sample of convergence curves for IEEE 33-bus distribution sys-
optimization algorithm (BFOA) [23], crow search algorithm tem is presented in Fig. 3. The evaporation process in WCA leads
(CSA) [24], and PSO [31]]. Three capacitors are installed at to optimal solutions with little number of iterations.
buses 14, 24, and 30. The results show that the proposed algo- Table VI presents the technical, economic, and environmental
rithm is efficient to find the optimal solution with lowest real benefits of simultaneous placement of DGs and CBs in 33-
power losses of 130.91 kW and the total installed CBs capacity bus distribution system. Results show that the total emission
equals 1.8484 MVAR. The lowest voltage (0.951 p.u.) is found at is reduced by 58% due to the penetration of Renewable DGs
bus 18. (PV with 0.7149 and 0.6397; WT with 0.6476; and GT with
Table III presents the optimal solution of Case 2. It shows 0.2008 MW). Also generated power cost is reduced by 18%,
the effectiveness of the proposed WCA for finding the optimal and distribution power loss is reduced to 28.962 kW.
placement and sizing of DGs only. The obtained results are 2) Results of 69-Bus System: Tables VII and VIII present
assessed comparing with that of BFOA in [23] and [25]–[27]. the results of Cases 1 and 2, respectively. It can be observed
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6 IEEE SYSTEMS JOURNAL

TABLE VIII
PLACEMENT AND SIZING OF DGS ONLY FOR 69-BUS SYSTEM (CASE 2)

Method Power loss DG size (MW) and location Min voltage

(kW) (p.u.)

FWA [16] 77.85 0.2258 (27), 1.1986 (61), 0.974 (62)

0.4085 (65)
HSA [32] 86.77 1.7732 0.967∗
GA [34] 88.5 1.9471 0.969∗
RGA [35] 87.65 1.7868 0.968∗
CVSI [36] 83.18 1.895 (61) 0.968 (27)
WCA 71.5 0.775 (61), 1.105 (62), 0.987 (65)
0.4380 (23)
Fig. 2. Voltage profiles of different cases for 33-bus system. ∗
Placement of minimum bus voltage has not been mentioned.

TABLE IX
PLACEMENT AND SIZING OF DGS AND CBS FOR 69-BUS SYSTEM
FOR CASES 3 AND 4 USING WCA

Case DGs CBs Min. voltage

(p.u.)
Location Size (MW) Power Location Size
factor (MVAR)

Case 3 17 0.5408 1.0 2 1.1879 0.994 (50)

(Single OF) 61 2 1.0 62 1.2373
69 1.1592 1.0 69 0.2697

OF: f1 = 33.339 kW
Fig. 3. Convergence curves of Case 2 for the first 5 runs 33-bus distribution
system. Case 4 61 1.8247 0.877 15 0.0188 0.994 (50)
(multi-OF) 36 1.0414 0.916 33 0.4578
19 0.1063 0.904 22 0.5586
TABLE VI
MULTIOBJECTIVE PLACEMENT AND SIZING FOR 33-BUS SYSTEM (CASE 5) OFs f1 = 18.7048 kW f2 = 0.0092 f3 = 0.0313

Location Active power (MW) Reactive power

(MVAR) that the power loss obtained by the proposed algorithm is better
Grid 1 1.5410 0.6572
than those obtained by other methods, where the power loss ob-
PV 32, 27 0.7149, 0.6397 0.4385, 0.2776 tained by WCA is decreased to 144.53 and 71.5 kW for Cases 1
WT 25 0.6476 0.1467 and 2, respectively. Add to that, the voltage profiles have the
GT 18 0.2008 0.0516
CB 15, 26, 19 0, 0, 0 0.3, 0, 0.45
lowest voltage levels at node 65 for Cases 1 and 2 (0.95 and
0.979 p.u.), respectively. Table IX shows the placement of DGs
OFs Power losses (kW) Cost ($/h) Emission (lb/h)
and CBs for Cases 3 and 4. In Case 3, the power loss is
Before 202.6 304.8966 8.0267e+006
After 28.9615 249.3429 3.4045e+006 33.339 kW that is the smallest value compared with Cases 1
and 2, with DGs at nodes 69, 61, and 17 using active power pen-
etration levels of 1.159, 2, and 0.541 MW, respectively. Also,
TABLE VII
OPTIMAL PLACEMENT AND SIZING OF CBS FOR 69-BUS SYSTEM (CASE 1) three CBs at nodes 2, 62, and 69 are used with reactive power
penetration levels of 1.188, 1.237, and 0.27 MVAR, respectively.
Method Power Capacitor size (MVAR) and Min. voltage
Therefore, the DGs/CBs placement enhances the power loss re-
loss (kW) location (p.u.) duction and maximizes the energy utilization of the distribution
system. Multiobjective optimal placement and sizing of CBs and
PSO [19] 156.14 1.015 (59), 0.241 (61), 0.934∗
0.365 (65) DGs with controllable PF (Case 4) gives better results than using
DE [28] 149.55 0.2 (16), 0.7 (60), 0.5 (61) 0.928∗ DG with unity PF as shown in Table IX. Also the voltage profile
(DE-PS) [29] 146.13 0.95 (61), 0.2 (64), 0.05 (65), 0.931∗ of the system in Case 4 is better than that of the other cases
0.15 (95), 0.3 (21) as shown in Fig. 4. Table X shows the application of Case 5
CSO [30] 147.95 1.2 (62), 0.25 (21) 0.930∗
TLBO [31] 146.35 0.6 (12), 1.050 (61), 0.150 (64) NR for 69-bus distribution system. A high reduction in power loss
GSA [32] 145.9 0.15 (26), 0.15 (13), 1.050 (15) 0.952∗ (22.36 kW) is obtained. Added to that, significant economic
DSA [33] 147 0.9 (61), 0.45 (15), 0.45 (60) NR and environmental benefits are satisfied for Case 5. However,
WCA 144.53 1.2882 (61), 0.2134 (69), 0.95 (65)
0.27 (18) the economic benefits is achieved by reducing production costs
from 309.7134 to 297.47 $/h, while the emissions pollution is
∗
Placement of minimum bus voltage has not been mentioned. reduced from 82.508e5 to 4.247e5 lb/h.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

EL-ELA et al.: OPTIMAL PLACEMENT AND SIZING OF DISTRIBUTED GENERATION AND CBs IN DISTRIBUTION SYSTEMS 7

TABLE XII
MULTIOBJECTIVE SOLUTION OF EDN FOR CASE 5

DG type Location Active power (MW) Reactive power (MVar)

Grid 1 12.663 4.251

PV 30, 7 1.0, 1.0 0. 006, 0.001
WT 19 5 3.750
GT 28 3 2.250
CB 3 0 4.050
OF Power losses (kW) Cost ($/h) ×103 Emission (lb/h)×106
Fig. 4. Voltage profiles of different cases for 69-bus distribution system. Before 805.7327 16.870 47.602
After 257.4912 9.7174 32.146

TABLE X
MULTIOBJECTIVE PLACEMENT AND SIZING SOLUTION
FOR 69-BUS SYSTEM (CASE 5)

Location Active power (MW) Reactive power (MVAR)

Grid 1 1.7467 0.2946

PV 58 66 0.1024 0.7314 0.0352 0.2913
WT 63 0.7030 0.2742
GT 64 0.5405 0.3130
CB 23, 62, 42 0, 0, 0 0.6, 0.6, 0.3
OFs Power losses (kW) Cost ($/h) Emission (lb/h)×106
Before 225 309.7134 8.2508
After 22.36 297.47 4.247

Fig. 6. Power losses for different cases of the test systems.

TABLE XI
PLACEMENT AND SIZING OF DGS AND CBS OF EDN FOR CASES 1–4
4 has the best voltage profile with the lowest VD of 0.0029 p.u.
Method Case1 Case2 Case3 Case4 compared with Cases 1–3. Cases 3 and 4 present acceptable
levels of VSI compared with Cases 1 and 2.
DG size (MW, – 0.59 (21) 3.5586 (24) 7.49, 2.74 (16), Simulation results obtained by multiobjective optimization
MVAR) and 0.59 (6) 7.5307 (5) 6.11, 1.99 (22),
location 0.726 (16) 6.537 (19) 5.06, 3.393 (7) of the technical, economic, and environmental objectives are
CB size (MVAR) 3.302 (22) – 5.119 (14) 3.61 (17), 2.35 shown in Table XII. The PV and WT have active roles in emis-
and location 3.736 (6) 5.825 (18) (2) sion reduction, and both DGs and CBs have active roles in power
5.047 (16) 2.472 (11)
Min. voltage (p.u.) 0.957 (30) 0.983 (30) 0.995 (15) 0.997 (30) loss and generation cost reduction. For Case 5, a large reduction
Power loss (kW) 581.368 250.18 53.836 47.1367 in power loss (257.49 MW) is obtained. Significant economic
VD 0.698 0.328 0.0071 0.0029 and environmental benefits are satisfied in Case 5. The eco-
VSI 0.039 0.036 0.035 0.035
nomic benefits are achieved by reducing production cost from
16870 to 9717.4 $/h while the emissions pollution is reduced
from 47.602 to 32.146 Mlb/h.
The above results for different test systems assure the WCA
capability to achieve significant technical, economic, and envi-
ronmental benefits. The power losses of the different cases for
the test systems are shown in Fig. 6.

Fig. 5. Voltage profiles of different cases for EDN system.
3) Results of EDN: The results of WCA Cases 1–4 applied
to EDN are shown in Table XI. Results show a significant im-
provement in system performance (VSI and VD). A comparison
between bus voltage profiles of each case is shown in Fig. 5 for
EDN. The highest reduction in power losses (47.1369 kW) is
obtained with placement of DGs and CBs in Case 4. Also, Case
V. CONCLUSION
Water cycle optimization algorithm has been proposed as
single and multiobjective frameworks for optimal placement
and sizing of combined DGs/CBs in distribution networks. It
aims at maximizing technical, economic, and environmental
benefits. Five operational cases of DGs and CBs have been
applied to three different distribution systems and compared
with other optimization algorithms. The salient findings of the
simulation results are summarized as follow.
1) More effectiveness of the proposed WCA have been
obtained, for solving the optimal placement and sizing
problem, compared with other optimization algorithms.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 IEEE SYSTEMS JOURNAL
2) The system power losses have been minimized using the
placement of CBs alone or DGs alone, but major reduction
in power losses have been obtained using the optimal
placement and sizing of combined DGs and CBs.
3) More improvement in distribution system performance
has been achieved using DGs with controllable power
factor.
4) Simultaneous DGs and CBs placement achieve substantial
technical, economic, and environmental benefits.
5) More emission reduction up to 58%, more power loss
reduction up to 86%, more energy cost reduction up to
18%, as in Case 5 for 33-bus system, have been achieved
using the proposed method.
6) Good convergence characteristics of the proposed WCA
have been noticed.
REFERENCES
[1] S. Saha and V. Mukherjee, “Optimal placement and sizing of DGs in RDS
using chaos embedded SOS algorithm,” IET Gener. Transmiss. Distrib.,
vol. 10, no. 14, pp. 3671–3680, Apr. 2016.
[2] A. M. Shaheen, R. A. El-Sehiemy, and S. M. Farrag, “Adequate planning
of shunt power capacitors involving transformer capacity release benefit,”
in IEEE Syst. J., to be published, doi: 10.1109/JSYST.2015.2491966.
[3] S. Kongtonpisan and S. Chaitusaney, “Loss reduction in distribution
system with photovoltaic system by considering fixed and automatic
switching capacitor banks using genetic algorithm,” in Proc. Asia–Pacific
Power Energy Eng. Conf., Shanghai, China, Mar. 2012, pp. 1–4, doi:
10.1109/APPEEC.2012.6307191.
[4] M. Doostan, S. Navaratnan, S. Mohajeryami, and V. Cecchi, “Concurrent
placement of distributed generation resources and capacitor banks in dis-
tribution systems,” in Proc. 2016 North Amer. Power Symp., Nov. 2016,
pp. 1–6.
[5] R. El-Sehiemy and E. Ahmed, “Integration of distributed generating units
in to distribution networks,” in Proc. 14th Int. Middle East Power Syst.
Conf., Giza, Egypt, Dec. 2010, Paper 288.
[6] A. A. A. El-Ela, R. A. El-Sehiemy, A.-M. Kinawy, and E. S. Ali, “Optimal
placement and sizing of distributed generation units using different cat
swarm optimization algorithms,” in Proc. IEEE 2016 18th Int. Middle
East Power Syst. Conf., 2016, pp. 975–981.
[7] S. K. Injeti and N. P. Kumar, “Optimal planning of distributed generation
for improved voltage stability and loss reduction,” Int. J. Comput. Appl.,
vol. 15, no. 1, pp. 40–46, 2011.
[8] S. Saravanamutthukumaran and N. Kumarappan, “Sizing and siting of
distribution generator for different loads using firefly algorithm,” in Proc.
2012 Int. Conf. Adv. Eng., Sci. Manage., Mar. 2012, pp. 800–803.
[9] M. H. Amini, M. P. Moghaddam, and O. Karabasoglu, “Simultaneous
allocation of electric vehicles’ parking lots and distributed renewable re-
sources in smart power distribution networks,” Sustain. Cities Soc., vol. 28,
pp. 332–342, 2017.
[10] M. A. Masoum, M. Ladjevardi, A. Jafarian, and E. F. Fuchs, “Optimal
placement, replacement and sizing of capacitor banks in distorted distri-
bution networks by genetic algorithms,” IEEE Trans. Power Del., vol. 19,
no. 4, pp. 1794–1801, Oct. 2004.
[11] H. N. Ng, M. M. A. Salama, and A. Y. Chikhani, “Classification of ca-
pacitor allocation techniques,” IEEE Trans. Power Del., vol. 15, no. 1,
pp. 387–392, Jan. 2000.
[12] G. A. Bortignon and M. E. El-Hawary, “A review of capacitor place-
ment techniques for loss reduction in primary feeders on distribution sys-
tems,” in Proc. 1995 Can. Conf. Electr. Comput. Eng., Sep. 1995, vol. 2,
pp. 684–687.
[13] A. M. Shaheen, R. A. El-Sehiemy, and S. M. Farrag, “A novel adequate
bi-level reactive power planning strategy,” Int. J. Elect. Power Energy
Syst., vol. 78, pp. 897–909, 2016.
[14] A. K. Saonerkar and B. Y. Bagde, “Optimized DG placement in radial
distribution system with reconfiguration and capacitor placement using
genetic algorithm,” in Proc. 2014 Int. Conf. Adv. Commun. Contr. Comput.
Technol., May 2014, pp. 1077–1083.
[15] M. Jannat and A. Savi´c, “Optimal capacitor placement in distribution
networks regarding uncertainty in active power load and DG units pro-
duction,” IET Gener. Transmiss. Distrib., vol. 10, pp. 3060–3067, 2016.
[16] A. M. Imran, M. Kowsalya, and D. P. Kothari, “A novel integration
technique for optimal network reconfiguration and distributed generation
placement in power distribution networks,” Int. J. Elect. Power Energy
Syst., vol. 63, pp. 461–472, 2014.
[17] M. Kefayat, A. L. Ara, and S. N. Niaki, “A hybrid of ant colony op-
timization and artificial bee colony algorithm for probabilistic optimal
placement and sizing of distributed energy resources,” Energy Convers.
Manage., vol. 92, pp. 149–161, 2015.
[18] A. Jabbar and S. Zainudin, “Water cycle algorithm for attribute reduction
problems in rough set theory,” J. Theor. Appl. Inf. Technol., vol. 61, no. 1,
pp. 107–117, 2014.
[19] K. Prakash and M. Sydulu, “Particle swarm optimization based capacitor
placement on radial distribution systems,” in Proc. 2007 IEEE Power Eng.
Soc. Gen. Meeting, Jun. 2007, pp. 1–5.
[20] A. A. A. El-Ela, R. A. El-Sehiemy, A. M. Kinawy, and M. T. Mouwafi,
“Optimal capacitor placement in distribution systems for power loss re-
duction and voltage profile improvement,” IET Gener. Transmiss. Distrib.,
vol. 10, no. 5, pp. 1209–1221, 2016.
[21] A. A. Abou El-Ela, A. M. Kinawy, M. T. Mouwafi, and R. A. El-Sehiemy,
“Optimal sitting and sizing of capacitors for voltage enhancement of dis-
tribution systems,” in Proc. 2015 50th Int. Univ. Power Eng. Conf., 2015,
pp. 1–6.
[22] [Online]. Available: http://egyptera.org/en/t3reefa.aspx
[23] M. Kowsalya, “Optimal distributed generation and capacitor placement in
power distribution networks for power loss minimization,” in Proc. 2014
Int. Conf. Adv. Electr. Eng., Jan. 2014, pp. 1–6.
[24] A. Askarzadeh, “Capacitor placement in distribution systems for power
loss reduction and voltage improvement: A new methodology,” IET Gener.
Transmiss. Distrib. vol. 10, no. 14, pp. 3631–3638, 2016.
[25] R. S. Rao, K. Ravindra, K. Satish, and S. V. L. Narasimham, “Power
loss minimization in distribution system using network reconfiguration in
the presence of distributed generation,” IEEE Trans. Power Syst., vol. 28,
no. 1, pp. 317–325, Feb. 2013.
[26] N. K. Meena, A. Swarnkar, N. Gupta, and K. R. Niazi, “A Taguchi-based
approach for optimal placement of distributed generations for power loss
minimization in distribution system,” in Proc. 2015 IEEE Power Energy
Soc. Gen. Meeting, Jul. 2015, pp. 1–5.
[27] M. H. Moradi and M. Abedini. “A combination of genetic algorithm
and particle swarm optimization for optimal DG location and sizing in
distribution systems,” Int. J. Elect. Power Energy Syst., vol. 34, no. 1,
pp. 66–74, 2012.
[28] S. Neelima and P. S. Subramanyam, “Efficient optimal sizing and alloca-
tion of capacitors in radial distribution systems using drdlf and differential
evolution,” ACEEE Int. J. Elect. Power Eng. vol. 2, no. 3, pp. 56–61, 2011.
[29] A. A. El-Fergany, “Optimal capacitor allocations using evolutionary al-
gorithms,” IET Gener. Transmiss. Distrib., vol. 7, no. 6, pp. 593–601,
2013.
[30] A. A. El-Fergany and A. Y. Abdelaziz, “Capacitor allocations in radial dis-
tribution networks using cuckoo search algorithm,” IET Gener. Transmiss.
Distrib., vol. 8, no. 2, pp. 223–232, 2014.
[31] S. Sultana and P. K. Roy, “Optimal capacitor placement in radial distri-
bution systems using teaching learning based optimization,” Int. J. Elect.
Power Energy Syst., vol. 54, pp. 387–398, 2014.
[32] Y. M. Shuaib, M. S. Kalavathi, and C. C. A. Rajan, “Optimal capacitor
placement in radial distribution system using gravitational search algo-
rithm,” Int. J. Elect. Power Energy Syst., vol. 64, pp. 384–397, 2015.
[33] M. R. Raju, K. R. Murthy, and K. Ravindra, “Direct search algorithm for
capacitive compensation in radial distribution systems,” Int. J. Elect.Power
Energy Syst., vol. 42, no. 1, pp. 24–30, 2012.
[34] K. Nara, A. Shiose, M. Kitagawa, and T. Ishihara, “Implementation of ge-
netic algorithm for distribution systems loss minimum re-configuration,”
IEEE Trans. Power Syst., vol. 7, no. 3, pp. 1044–1051, Aug. 1992.
[35] J. Z. Zhu, “Optimal reconfiguration of electrical distribution network using
the refined genetic algorithm,” Elect. Power Syst. Res., vol. 62, no. 1,
pp. 37–42, 2002.
[36] A. Gantayet and S. Mohanty, “An analytical approach for optimal place-
ment and sizing of distributed generation based on a combined voltage
stability index,” in Proc. 2015 IEEE Power, Commun. Inf. Technol. Conf.,
Oct. 2015, pp. 762–767.
Authors’ photographs and biographies not available at the time of publication.