Arab J Sci Eng (2017) 42:2851–2865
DOI 10.1007/s13369-017-2518-y
RESEARCH ARTICLE - ELECTRICAL ENGINEERING
Optimization of Energy Loss Cost of Distribution Networks
with the Optimal Placement and Sizing of DSTATCOM Using
Differential Evolution Algorithm
Joseph Sanam1 · Sanjib Ganguly2 · A. K. Panda1 · Chaduvula Hemanth3
Received: 24 August 2016 / Accepted: 5 April 2017 / Published online: 26 May 2017
© King Fahd University of Petroleum & Minerals 2017
Abstract This paper presents an optimization of energy loss
cost (ELC) of distribution networks (DN) with the opti-
mal allocation and sizing of DSTATCOM to maximizing
the total net profit (TNP)/cost savings by using the differen-
tial evolution algorithm per annum and planning horizon of
DSTATCOM installation scheme. In this approach, the opti-
mal reactive power compensation is the main vital role in
solving the objective function. The optimal reactive power
compensation with the optimal placement and sizing of
DSTATCOM and the improvement in voltage profile of the
DN are obtained based on certain objectives such as best
reduction in network power loss and the total ELC and the
maximization of TNP. The modeling on the size of DSTAT-
COM is used and was incorporated in DN through the
forward–backward sweep load flow technique to evaluate
the parameters of load flows in DN. Present worth factor is
instigated to evaluate the TNP of the DSTATCOM installa-
tion scheme. The proposed method is validated on the 30,
33, and 69-bus DN.
Keywords DSTATCOM · Reactive power compensation ·
Power loss · Voltage improvement · ELC · TNP · PWF
B Joseph Sanam
joseph.nitr@gmail.com
1 Department of Electrical Engineering, National Institute
of Technology, Rourkela, India
2 Department of Electronics and Electrical Engineering,
Indian Institute of Technology, Guwahati, India
3 Department of Electrical Engineering, Indian Institute
of Technology, Kharagpur, India
1 Introduction
Distribution networks are typically radial in structure because
of the simple design, operational simplicity, and low cost
[1,2]. The substation is the only one point at which DN are
fed by the integrated power generating stations through the
transmission system [2]. In conventional DN, the flow of
electricity is unidirectional, explicitly from the distribution
transformer through distribution feeders called passive net-
work to load point transformers [3]. The increase in electrical
demand, lack of reactive power and the high resistance to
reactance ratios in distribution lines results in significantly
high power losses compared to the power loss in transmission
systems, large voltage drops, low voltage stabilities, and load
imbalances at certain areas of DN, etc., [1,2,4]. These prob-
lems have an impact on the reliable and economical electric
power supply to the consumers. Hence, it is important that the
DISCOs should design DN properly to optimize their opera-
tion with an objective of the minimization of power loss and
energy loss cost, i.e., maximization of TNP/savings analysis
(minimization of an economic cost function) and improv-
ing the bus voltages [5,6]. Therefore, optimal development
of the DN saves expenditure for the utilities with the lowest
cost and highest benefit and reduces the cost of power supply
to the customers [7].
There have been some previous investigations that have
allocated certain devices such as DGs, shunt capacitors, and
DSTATCOM. These devices are optimally allocated at the
locations of the DN in anticipation to compensate the reac-
tive power, to optimize the power loss and total economic
cost and to improve the voltage profile of DN. DGs are used
for the DN to optimize the energy loss and benefit–cost anal-
ysis of DG installation by optimally sizing and allocating
it on DN [8–11]. However, DG sources are relatively high
costs, and intermittency [12–15]. Shunt capacitor banks are
123
2852
deployed to distributed transformers to reduce the power loss
and decrease the cost of capacitor installation scheme, and
this approach adopted the NPV principle for the evaluation
of cost–benefit of the capacitor installation scheme [16–19].
A capacitor is allocated on DN to minimize the investment
cost and energy loss at different load levels. Authors of [5]
proposed the minimization of power loss and cost of energy
loss with the optimal allocation of the capacitor in DN by
using TLBO technique. The optimal placement and sizing of
fixed capacitor banks in DN is done to minimize the kW loss
and maximize net savings [4]. But capacitors are not capa-
ble of providing smooth reactive power compensation and
suffering from inevitable oscillations [20], and the proposed
approach obtained higher total net savings compared to the
approach of [21] in which capacitor is located in of 33 and
69 bus DN.
The DSTATCOM can alleviate the difficulties that are dis-
cussed to minimize the energy loss and to maximize the
TNP/savings of DN. Some of the power quality issues of
electrical DN influenced by the allocation of DSTATCOM
with distribution generator are given in [22]. These same
type devices are optimally sized and allocated in DN by
using the particle swarm optimization algorithm to compen-
sate the reactive power for the reduction of power loss [23].
Reconfiguration and optimal size and allocation of PV arrays
and DSTATCOM are simultaneously done on DN by using
the combination of fuzzy multi-objective approach and ACO
metaheuristic algorithm to reduce the power loss and improve
voltage profile [24]. The optimal allocation of DSTATCOM
along with network reconfiguration by using DEA is pro-
vided to minimize the power loss in DN [25]. However, the
approaches with the combination of DSTATCOM and recon-
figuration are an expensive issue because the reconfiguration
problem is defined as the process of changing the topology
of the network using some sectionalizing switches (closed
switches), and tie switches (open switches), so as the num-
ber of lateral branches increases the circuit length increases
which results in increase in the cost of scheme [7].
The modeling and the optimal allocation of single and
multiple DSTATCOM in DN to reduce the power loss has
been investigated in [26]. The Authors of [27] proposed opti-
mal sizing and allocation of DSTATCOM in DN by using the
immune algorithm for the compensation of reactive power
to optimize the power loss and total cost savings. How-
ever, in this approach maintenance cost, the operational cost
of DSTATCOM, PWF, and PH of DSTATCOM installation
scheme for the evaluation of total ELC after DSTATCOM is
installed is not considered except for the initial capital invest-
ment cost of the DSTATCOM. NPV or PWF is taken into
account to evaluate an economic assessment of the hybrid
voltage control scheme with PV penetration and DSTAT-
COM in Taipower DN. However, in this work, the initial
capital investment cost, O&M cost of DSTATCOM are not
123
Arab J Sci Eng (2017) 42:2851–2865
taken into account to evaluate the benefit and cost of the
entire scheme, but these are considered for PV alone with a
proper selection of PV and DSTATCOM capacities [28]. In
most of these approaches in using DSTATCOM, the objec-
tive function is formulated for the minimization of power
loss and improvement of voltage profile; a very few works
are appraised minimization of total ELC of the DN to maxi-
mizing the TNP/savings of the entire scheme.
Out of these reviews of studies, this paper proposes an
interest of using DSTATCOM in DN appropriately. DEA
[29,30] is used as a solution strategy for optimal sizing and
placement of DSTATCOM. The appropriate modeling on the
size of DSTATCOM is used so as to incorporate it into the
FBS load flow algorithm [31] for the reactive power com-
pensation. To realistically appraise the TNP/savings of the
DSTATCOM installation project, PWF criterion is applied.
The main objectives of proposed work are improving the volt-
age profile, minimizing the power loss so that minimizing
the energy loss, and total ELC of the DN, thereby maximiz-
ing the TNP/savings over PH of DSTATCOM installation
scheme. IEEE 30, 33 and 69-bus DN are used for the case
study. Simulation results have demonstrated the effectiveness
of proposed work. The reduction in power loss and ELC and
maximization of TNP/savings are found to be better in the
proposed approach as compared to IA approach used in [27]
and GSA used for optimal capacitor placement [21]. The
abbreviations used in the proposed approach are given in
Table 1.
The paper is organized as follows: The modeling of
DSTATCOM and the incorporation of DSTATCOM model
in FBS load flow are shown in Sects. 2 and 3, respectively.
Objective function formulation is provided in Sect. 4. Imple-
mentation of proposed approach using DEA is provided in
Sect. 5. Simulation results are provided in Sect. 6, and the
paper concluded in Sect. 7.
2 Modeling of DSTATCOM
DSTATCOM is a shunt connected VSC device that absorbs or
injects both active and reactive current, respectively, through
PCC. In this work, the modeling of DSTATCOM is referred
from our previous work [20,31]. DSTATCOM is used only
for reactive power compensation in DN. Figure 1. is the single
line diagram of two successive buses n and n + 1 of DN,
there are real and reactive power demands connected to these
buses, and it is used to place the DSTATCOM in DN. The
Kirchhoff’s voltage law equation of Fig. 1. is given by Eq. (1)
V n+1  βn+1 = V n  αn − (Rb + j X b ) I b  θ. (1)
The phasor diagram corresponding to Eq. (1) is shown in
Fig. 2. The Kirchhoff’s voltage law equation of Fig. 3. is as
expressed in Eq. (2)
Arab J Sci Eng (2017) 42:2851–2865 2853

Vn Vn +1 V'
n
R jX
b b
jX I
I b b
b
I DSTAT R I
V b b
P + jQ P + jQ
n +1
n n n +1 n +1 jX b I DSTAT
Rb I DSTAT V'
α' n +1
Fig. 1 Two successive buses of DN drawn as a single line diagram n β
n +1
I β'
V θ b n +1
n
jX I Reference axis
b b
Fig. 4 Phasor diagram for the network shown in Fig. 3

R I
V b b
α n +1 Real part:
n
β I   
θ n + 1b IDSTAT cos βn+1 + 90◦
Reference axis     
V n Rb cos αn V n X b sin αn V n+1 Rb cos βn+1
= + −
Fig. 2 Phasor diagram for the network shown in Fig. 1 Rb2 + X b2 Rb2 + X b2 Rb2 + X b2
 

V n+1 X b sin βn+1
Vn' Vn' +1 − − Ib cos θ (4)
Rb2 + X b2
R jX
b b
PCC Imaginary part:
I I
b DSTAT    
j IDSTAT sin βn+1 + 90◦
    
P + jQ
V n Rb sin αn V n X b cos αn V n+1 Rb sin βn+1
n n P + jQ = + −
n +1 n +1 Rb2 + X b2 Rb2 + X b2 Rb2 + X b2
 

V n+1 X b cos βn+1
DSTATCOM
(VSC )
− − I b sin θ (5)
Rb2 + X b2

  
V R V n+1 X b V n Rb
Energy Storage Let, a = − Rn+1
2 +X 2 , b =
b
Rb2 +X b2
,c = Rb2 +X b2
,d =
b b

Fig. 3 Single line diagram with a DSTATCOM placed at bus i + 1 V n Xb 
, e = I DS , ψ = βn+1 + 90◦ , and φ = I b
Rb2 +X b2
By replacing these parameters in Eqs. (4) and (5), the mag-
  
V n+1  βn+1 = V n αn − (Rb + j X b ) nitude of the injected current of DSTATCOM has obtained
 π  as:

× I b θ + I DSTAT  + βn+1 . (2)
2  
K 1 + K 2 sin αn − θ − ϕ
I DS = (6)
The phasor diagram corresponding to Eq. (2) is shown in sin (ψ + θ )
Fig. 4. Due to the injection of IDSTAT at PCC, the voltage
 . The angle of the I      
at bus n + 1 is changed to Vn+1 DSTAT is where K 1 = a sin βn+1 + θ + b cos βn+1 − θ , K2 =
expressed as follows √ V

c2 + d 2 =  2 n 2 , “ϕ” is a unique angle satisfying the
Rb +X b
π 
I DSTAT = + βn+1 (3) following conditions:
2
The real and imaginary parts of Eq. (2) are separated and (i) −π < ϕ < π,
are computed as Eqs. (4) and (5). (ii) tan ϕ = dc = XRbb , Since d = 0, c = 0

123
2854 Arab J Sci Eng (2017) 42:2851–2865

Table 1 Abbreviation Table 2 Nomenclature

S. no. Notation Abbreviation S. no. Notation Description

1 DSTATCOM Distribution static 1 θ Angle of Ik

synchronous compensator 2 αn Angle of Vk
2 DEA Differential evolution 3 ϕ Unique angle
algorithm
4 βn+1 Angle of Vk+1
3 ELC Energy loss cost
5 n, n + 1 Bus numbers
4 PWF Present worth factor
6 Rn + j X n Line impedance
5 DN Distribution networks
7 Vn Voltage in kth bus
6 DISCO Distribution companies
8 IDSTAT DSTATCOM current
7 DG Distribution generators
9 K1, K2 Properties of trigonometry
8 NPV Net present value
10 In Current flowing from kth to
9 PV Photovoltaic k + 1th bus
10 ACO Ant colony optimization 11 Pn , Q n Real and reactive power demand at
11 O&M Operating and maintenance the kth bus
12 FBS Forward–backward sweep 12 Vn Voltage in kth bus after
13 VSC Voltage source converter DSTATCOM is placed
14 PCC Point of common coupling 13 αn Angle of Vk after the placement of
DSTATCOM
15 TNP Total net profit  
14 βn+1 Angle of Vk+1 after the placement
16 PH Planning horizon of DSTATCOM
15 Q DSTAT DSTATCOM size to compensate
the reactive power
Finally, Eq. (6) is the current, which must be injected at a
required n + 1th bus of DN to compensate the reactive power
to reduce the power loss. Hence, the reactive power that can where n = 1, . . . m, PLoad (n) and, Q Load (n) are active and
be provided by the DSTATCOM is expressed as follows: reactive power demand at the nth bus.
Now then, branch currents are calculated using the expres-
 ∗ sion
j Q DSTAT = V n+1 · IDSTAT (7)
m
I Loadn + I Loadn+1 = I bn,n+1 (10)
The symbol ‘∗ ’ in Eq. (7) designates the complex conjugate.
n=1
Eq. (7) is integrated into the load flow algorithm through the
DEA to compute the load flow parameters voltage magni- The reactive power, which is to be compensated, can be deter-
tude of each bus and the total power loss of the network. mined by Eq. (11)
The variables of Eq. (7), IDSTAT , depends on DSTATCOM
location, and the size of the DSTATCOM depends on βn+1  , Q DSTATn+1 = Q Loadn+1 − Q new
Loadn+1 (11)
which is a variable angle. Thus, the location of DSTATCOM

and βn+1 is considered to be as optimizing variables. The where Q DSTAT is evaluated by Eq. (7).
nomenclature of this modeling is given in Table 2.
Step 2-Forward sweep computation of voltage profile at each
bus. The node voltage can be computed using following Eq.
(12)
3 Incorporation of DSTATCOM into  
Forward–Backward Sweep Load Flow V n = V n+1 + I b Z b n,n+1 (12)
Technique
where n + 1 = the receiving end bus and, n = thesending
Step 1-Backward sweep computation of load and branch cur- endbus. Ibn,n+1 = he branch current between the buses n and
rents and the required reactive power to be compensated n + 1. Z bn,n+1 = the impedance between the buses n and
n + 1.
PLoadn + j Q Loadn = Vn × I ∗ Loadn (8) Step 3-Convergence criteria check convergence using the
1   following equation and steps 1 and 2 are repeated until con-
I Loadn = × PLoadn − j Q Loadn (9) vergence is satisfied
Vn

123
Arab J Sci Eng (2017) 42:2851–2865 2855

Power f 2 = DSTATTotalcost = f 21 + f 22 + f 23 (18)

(kW) nl  
Peak load level
f 21 = kck × Cin × Q DSTAT
k × PWF (19)
k=1
Medium load level ph nl  
f 22 = kck × Cop × Q DSTAT
k × PWF (20)
Light load level
y=1 k=1
ph nl  
f 23 = kck × Cin × Q DSTAT
k × Cma × PWF
y=1 k=1
Time
T1 T2 T3 (hour/year) (21)

Fig. 5 Time duration curve where f 21 and f 22 and f 23 are the total initial capital
investment cost and the total operational cost and the total
count
 count   count−1  maintenance costs [34] of the DSTATCOM, respectively, in
V n+1 = abs V n+1 − abs V n+1 (13)
the whole PH of DSTATCOM installation scheme, Ce is the
count energy cost per kWh; Tk is the duration of time in kth load
If V n+1 < accuracy. (14)
level; Cin is the initial capital investment cost (purchase cost)
Count = the iteration number, Accuracy = 0.0001 of DSTATCOM per kVAr; Cop is the operational cost of the
DSTATCOM per kWh; Cma is the DSTATCOM maintenance
cost which in terms of the % of initial cost of DSTATCOM
4 Mathematical Problem Formulation per a year; Q DSTAT
k is the size of the DSTATCOM placed
at optimal location during kth load level; kck is the propor-
4.1 Objective Function (F) tionality constant of kth load level time duration to the total
duration of the time formulated as following Eq. (22)
In the proposed method, the optimal location and the size
Tk
(kVAr rating) of DSTATCOM in DN in a steady-state condi- kck = (22)
nl
tion are used to optimize the objectives such as ELC of DN Tk
to achieve the maximization of TNP. Hence, minimizing the k=1
ELC ( f 1 ) and total installation cost of DSTATCOM ( f 2 ) in
DN is considered in the objective function (F). The total ELC 4.1.1 Real Power Loss
and TNP of the DN are evaluated per for a given period ‘T’,
i.e., per annum and PH of DSTATCOM installation scheme The real power loss encountered in Eq. (17) is expressed by
by taking the three load levels (light, medium and peak lev- Eq. (23) [31]
els) into account in the network as shown in Tables 4 and 5. ⎡ ⎤
Here the load duration curve is divided into discrete intervals nl n−1
DSTAT
PLoss = ⎣ Ib2 ( j) × Rb ( j)⎦ (23)
as shown in Fig. 5 [27,30]. k

The objective function is mathematically expressed given k=1 j=1

by Eq. (15)
DSTAT is the active power loss during kth load level
where Ploss k
F = min (( f 1 ) + ( f 2 )) (15) after DSTATCOM is installed, nl is the number of load levels
mentioned in Table 5, k is the load level, n is the total number
The first part of ‘F’ is the total ELC ( f 1 ). The second part is of buses in the DN, Ib ( j) and Rb ( j) are the line current and
the total installation cost of DSTATCOM ( f 2 ) that includes the resistance of jth branch, respectively.
initial capital investment cost, the O&M cost (running costs)
of the DSTATCOM placed in DN. Therefore, the objective of 4.1.2 Present Worth Factor (PWF) Analysis
optimal DSTATCOM placement problem is defined by Eq.
(16) [32,33] To evaluate the economic value of the DSTATCOM installa-
tion scheme, it is required to compare the expected revenue
Objective function = F = f 1 + f 2 (16) and investment costs over the whole PH of DSTATCOM
ph nl installation scheme. In the proposed objective function as
f 1 = E Losscost
DSTAT
= Ce DSTAT
PLoss × Tk ×PWF (17) given in Eq. (17), the PWF principle is adopted for cost–
k
y=1 k=1 benefit analysis of the scheme. The PWF offers a net worth of

123
2856 Arab J Sci Eng (2017) 42:2851–2865

the scheme in today’s dollars by discounting each year’s cash (1). Initialization, (2). Mutation, (3). Crossover and (4).
flow back to the present and then, deducing the initial invest- Selection
ment. The mathematical expression of PWF is expressed by
Step 1. Initialization this operation is based on following Eq.
Eq. (24): [3,35,36]
(26); in this operation, the target vector (TV) is generated
(1 + γ ) y × γ randomly
PWF = . (24)
(1 + γ ) y − 1    
j j j
TVi,gen = round TVmin + TVmax
j
− TVmin .rannd()
4.1.3 TNP/Savings
(26)
The TNP which is to be maximized is the difference between
where j = 1, 2, . . . , D; i = 1, 2, . . . , N p and gen =
the expenditure of the energy loss cost of a DN without
the generation number.
DSTATCOM and the total expenditure of the energy loss
cost of a DN with the scheme of DSTATCOM. It is given by Step 2. Mutation in this operation, the mutant vector MVi,Gen
Eq. (25) [27] j
is generated by Eq. (27) for each target vector TVi,Gen
 
ph nl
1 MVi,gen = TVbest,gen + F. TVr1 ,gen − TVr2 ,gen (27)
TNP = Ce w/o DSTAT
PLoss × Tk × − F.
k
(1 + γ ) y
y=1 k=1 r1 , r2 are the integers ranged between the ranges [1, Np]. ‘F’
(25) is the scaling factor ranged from 0 to 2.

TNP, in fact, yields an economic savings or benefit in the DN
with DSTATCOM for the total PH of DSTATCOM installa-
tion scheme, [1,30]. Therefore, the purchasing cost of power
from substation according to the customer’s demand can be
reduced.
4.2 Constraints
Proposed Eq. (17) is bounded by various active constraints
to meet the limitations on DSTATCOM operation and elec-
trical requirements for the DN. Penalty factor is considered
when the operating variables such as bus voltages, line cur-
rents and the capacity of the DSTATCOM violate the desired
safe limits. The minimum and maximum voltages are con-
sidered as 0.9 and 1.1 p.u., respectively, [27]. If the voltage
at bus i crosses these limits, the penalty factor is considered
in objective function Eq. (17). Maximum limit of the current
in the network is taken as 1.2 times the base current of the
branch j. The line will be melted if the maximum limit of
the current exceeds. The maximum capacity is considered as
10,000 kVAr beyond which the penalty factor is considered.
5 Implementation of Proposed Approach Using
DEA
5.1 The Operation of DEA
DEA is the population-based metaheuristic algorithm [29,
30]. It has strings, and each string consists of population Np
and the dimension (D) of optimization variables. The overall
algorithm has mainly four steps of operation; they are:
Step 3. Crossover the operation of crossover takes place after
operation of mutation; this operation is performed between
the target vector and its corresponding mutant vector and
produces the trial vector. This operation is performed based
on the following equation

MV ji,gen if rand () ≤ CR
CV ji,gen = (28)
TV ji,gen or else
where CR = C crossover rate ranged between [0, 1].
Step 4. Selection in this operation, the fitness of trial vector
and corresponding target vector in the present population is
compared. If trial vector is best compared to target vector
fitness, then the target vector is replaced in trail vector and
is entered into the population of the future generation. If it is
not so, the target vector has remained in the population in the
coming generation. The operation of selection is as follows:

CVi,gen if f (C V ) ≤ f (TVi,gen )
TV ji,gen = (29)
TVi,gen or else
where f = the fitness (objective) function; the above steps
are repeated until the population is converged to the best
value.
5.2 Solution Strategy Using DEA
The main purpose of the DEA in this approach is the min-
imization of an objective function given in Eq. (16) by
determining the capacity of the DSTATCOM at the candi-
date locations. A typical string structure is shown in Fig. 6.
The candidate locations for DSTATCOM are considered to

123
Arab J Sci Eng (2017) 42:2851–2865 2857

'
formance and run on the same parametrical basis. The DEA
n+1,N (which yields QDSTAT)
N parameters, initial population size (NP ), dimension of each
population (D), weighting factor (F), crossover rate (CR),
and maximum generations, are taken as mentioned in Table 3.
Initial strings that are produced randomly contain bus loca-
Location of Value of 'i+1 corresponding to 
DSTATCOM
tion for compensation as well as optimal variable angle βn+1
the location of DSTATCOM
of DSTATCOM for the respective location at all(three) load
Fig. 6 A typical string for DEA levels. The objective function is calculated for each string by
running load flow algorithm at all load levels.
24 25 26 27

20 21 22 23 18 19 6 Simulation Results

In this section, the impact of DSTATCOM on total ELC of the

14 15 16 17
DN per annum and PH of DSTATCOM installation scheme
under three load levels is analyzed. The usefulness of pro-
posed approach is demonstrated on three DN as mentioned in
Sect. 5.2 of Sect. 5. Three load levels are selected as referred
1 2 3 4 5 6 7 8 9
to in Sect. 4.1 of Sect. 4 to model the annual load profile. The
time duration, total load for each load level, and base power
loss in three load levels are shown in Table 4. The fitness
10 11 12 13
(objective) function parameters used in this work are shown
in Table 5 [27,38].

28 29 30
6.1 Impact of DSTATCOM Allocation

When DSTATCOM is allocated one at a time at each bus

Fig. 7 Typical IEEE 30-bus DN except substation bus, there is a significant reduction in power
loss and an improvement in minimum bus voltage of the DN
23 24 25 at some certain buses. Figures 12, 13 and 14 show the impact
of DSTATCOM on power loss and minimum bus voltage of
26 27 28 29 30 31 32 33 IEEE 30-bus DN, respectively, and it is observed that these
parameters have been affected much when DSTATCOM is
placed at buses 3–7, 10, 14, 20 and 21. The DSTATCOM
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 location corresponding to the minimum power loss is found
to be bus 5. Thus, the voltage profile of the network with
and without DSTATCOM allocation at bus 5 is shown in
19 20 21 22 Fig. 15. Similarly, the impact of DSTATCOM on power loss
in IEEE 33-bus DN is observed that the buses 3–18, 23 and
26–33 have been affected much in view of power loss and
minimum bus voltage and the DSTATCOM location corre-
Fig. 8 Typical IEEE 33-bus DN
sponding to the minimum power loss is found to be bus 30.
The impact of DSTATCOM on power loss and minimum bus
be all the busses for each load level except the substation bus, voltage in IEEE 69-bus is observed that these parameters are

and the range for the angle βn+1 should lie in between 0◦ and affected much when DSTATCOM is placed at the buses 6–27
◦ 
90 . The optimal value of βn+1 decides the optimal size of and 51–69. The DSTATCOM location corresponding to the
DSTATCOM with the help of Eqs. (6) and (7). To study the minimum power loss is found to be bus 61. Table 6 shows
validity of the proposed approach, three standard sample DN optimal power loss, voltage profile, and size and location of
such as 30, 33 and 69 bus networks are taken [27,37] and DSTATCOM in all three load levels. Figure 15 shows the plot
their typical single line diagrams are shown in Figs. 7, 8 of the optimal size and location of DSTATCOM in IEEE 30,
and 9, respectively. To assess the virtue of DEA, it has been bus DN.
compared with an approach based on IA [27] in view of per-

123
2858 Arab J Sci Eng (2017) 42:2851–2865

36 37 38 39 40 41 42 43 44 45 46

47 48 49 50

51 52 68 69

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

66 67

53 54 55 56 57 58 59 60 61 62 63 64 65

28 29 30 31 32 33 34 35

Fig. 9 Typical IEEE 69-bus DN

Table 3 Parameters of DEA NP D F CR Generations currents connected to that bus according to Eq. (10). If the
demand of the load at the bus is high, then the current drawn
50 2 1 0.8 100 by the load is high. This results in an increase in ‘Ib ’, which
in turn causes the increase in Ploss . Also, the bus voltage
will fall due to the increase in voltage drop in the branch.
To minimize the power loss, should either active power be
6.2 Analysis of Power Loss Reduction injected or reactive power be compensated, which in turn
causes the decrease in ‘Ib ’. When ‘Ib ’ is decreased, the volt-
Power loss depends on branch current (Ib ) and resistance age drop in the branch will be decreased, and thus there is
(Rb ) since Ploss = Ib2 Rb according to Eq. (23). If ‘Ib ’ an improvement in the bus voltage profile. The proposed
increases, Ploss will also increase. Basically, ‘Ib ’ depends approach aims mainly to compensate the reactive power in
on two currents. One is the load current at the sending end DN to minimize the power loss, ELC, and to improve the TNP
bus according to Eq. (9), and the other is the lateral branch and voltage profile using DSTATCOM. If DSTATCOM volt-

Table 4 Load duration time and

Load level Light load Medium load Peak load
load level
Time duration (h/year) 2000 5260 1500

Total load (kVA)

30 bus 1603.2 + j1196.4 2084.1 + j1196.4 2306.7 + j1196.4
33 bus 3715 + j2300 4829.5 + j2300 5944 + j2300
69 bus 3802 + j2694.6 4942.8 + j2694.6 6083.5 + j2694.6
Total power loss(kW)
30 bus 146.07 220.32 320.02
33 bus 202.66 305.81 442.39
69 bus 224.97 342.96 502.47

123
Arab J Sci Eng (2017) 42:2851–2865 2859

350
Table 5 Parameters of objective function Power loss at light load with DSTATCOM
Power loss at medium load with DSTATCOM
Power loss at peak load with DSTATCOM
Objective function parameter Value 300

Power loss (kW)

Number of load levels 3 250
Cost of energy loss (ke ) US($/kWh) 0.06
200
Cost of DSTATCOM (kin ) US($/kVAr) 50
Operational cost of DSTATCOM (kop ) US($/kWh) 0.02 150
Maintenance cost of DSTATCOM (kma ) US($/kWh) 0.05
100
Discount rate of interest (γ ) 0.1 0 5 10 15 20 25 30

PH (years) 30 Bus Number

Fig. 12 Power loss at different loads with DSTATCOM at each bus of

IEEE 30-bus distribution network
5
2 x 10 Total net savings per annum
Total planning cost with DSTATCOM per annum 0.94
Energy loss cost without DSTATCOM per annum

Minimum voltage (p.u.)

0.93
1.5
0.92
0.91
Cost($)

1 0.9
0.89
0.88
0.5
0.87
Minimum voltage at light load with DSTATCOM
0.86 Minimum voltage at medium load with DSTATCOM
0 Minimum voltage at peak load with DSTATCOM
IEEE 30 Bus System IEEE 33 Bus System IEEE 69 Bus System 0.85
0 5 10 15 20 25 30
Bus System Bus Number

Fig. 10 Cost analysis per annum Fig. 13 Minimum bus voltage at different loads with DSTATCOM at
each bus of IEEE 30-bus distribution network

5
13 x 10 1 Light load base voltage
Total planning Cost
Ligh load voltage with DSTATCOM
12 Total Eloss Cost With DSTATCOM Medium load base voltage
Total Eloss Cost Without DSTATCOM
0.98 Medium load voltage with DSTATCOM
Bus Voltage (p.u.)

Peak load base voltage

11 Peak load voltage with DSTATCOM
0.96
10
Cost ($)

0.94
9
0.92
8
0.9
7
6 0.88

5 0.86
0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 25 30
Generation Bus Number

Fig. 11 Total scheme mean cost of IEEE 30-bus distribution network Fig. 14 Voltage magnitude at various loads with DSTATCOM at bus
5 of IEEE 30-bus distribution network

4000
Size of DSTATCOM at light load
age is greater than the bus voltage while it is being located 3500 Size of DSTATCOM at medium load
Size of DSTATCOM at peak load
in the bus, then DSTATCOM injects current into the bus 3000
with a phase angle of 90◦ as shown in Eq. (3) and Fig. 4.
Size (kVAr)

2500
Thus, the reactive power demand shall be compensated which 2000
minimizes the power loss and improves the bus voltage even- 1500
tually. 1000
In the case of the 33-bus DN, the complex load power 500
demand (P + jQ) at “bus 30” is higher than the load demand 0
0 5 10 15 20 25 30 35
in remaining buses as shown in the network data given in
Bus Number
“Table 8” in “Appendix”. Mainly, the reactive power demand
in “bus 30” is higher than all loads connected in remain- Fig. 15 Size of DSTATCOM at each bus of IEEE 30-bus distribution
ing buses. Hence, it is quite natural that the compensation network at various loads

123
2860 Arab J Sci Eng (2017) 42:2851–2865

of reactive power highly occurs at “bus 30”. Thus, in pro-

F (kW) Min voltage

posed approach, the best location for DSTATCOM allocation

(p.u.)

0.8888

0.8832

0.8899
is found to be “bus 30” with the size of 1252.7 kVAr as

–
shown in “Table 6”. Similarly, there is certain impact on
network power loss when DSTATCOM is located at buses

261.2

370.4
407.7
410.5
472
3–18, 23 and 26–33 as mentioned in Sect. 6.1. In case of the

–
69-bus DN, the load power demand at “Bus 61” is higher

Optimal size
than those of the remaining buses as shown in the network

(kVAr)
data in “Table 9” in “Appendix”. Hence, the best location

1271.3

1314.0
1222.6
1404.0
2606.8
for DSTATCOM allocation is found to be at “bus 61” with

–
the size of 1312.1 kVAr as shown in “Table 6”. There is

Peak load level

the certain impact on network power loss when DSTAT-

location
COM is located at buses 6–27 and 51–69 as mentioned in

Optimal
Sect. 6.1.

30
12
61
61
5
–
F (kW) Min voltage
(p.u.)

0.9189

0.9058

0.9124
6.3 Analysis of ELC

–
The total ELC of the all three DN is described in this sec-

166.7

241.2

261.9
272.0

274.4
tion. Parameters of Table 5 are used to evaluate the objective

–
function F. Table 7 shows the comparative results of the
total annual ELC of the DN with DSTATCOM installation

Optimal size
scheme. In the approach of [27], to evaluate TNP the whole

(kVAr)

1204.9

1278.4

1360.8
1008.1

1911.2
PH, as shown in Table 5, is considered only for the evalua-
tion of the DSTATCOM initial capital investment cost ( f 21 )

–
Table 6 Comparative results of reactive power compensation with DSTATCOM for three load levels
Medium load level

but not for remaining costs such as energy (power) loss cost
( f 1 ), operational and maintenance cost ( f 22 and f 23 ) of the
location
Optimal

DSTATCOM, so to compare the effectiveness of proposed

approach, the same scenario is considered along with same 30

61
12

61
5
–

PWF which is used in existed approach [27]. As can be

seen, compared with IA, DEA offers an improved optimal
F(kW) Min voltage

solution with its lower F and higher TNP. It is noteworthy

(p.u.)

0.9358

0.9256

0.9338

that the values of F and TNP are minimized and maxi-

–

mized by 10.5% and 136.18%, respectively, in the scheme

100.8

143.5

152.0
171.8

157.5

of IEEE 33-bus DN and 8.49% and 61.25%, respectively,

–

in the scheme of IEEE 69-bus DN using DEA compared

with IA [27] method. Using the proposed approach, the
Optimal size

TNP per annum with respect to total ELC with DSTAT-

(kVAr)

COM is valued to be of the order of 17.91, 16.93, and

1159.6

1252.7

1312.1
1704.4
962.4

19.63%, in 30, 33 and 69 buses DN, respectively, which is

–

quite profitable and useful for “DISCOs” (distribution com-

Light load level

panies).
location

Figure 11 shows the mean economic evaluation of total

Optimal

ELC result of 30 bus DN including f 22 and f 23 considering

30

61
IA based [27] 12

IA based [27] 61
5
IA based [27] –

the PWF as given in Eq. (24) in the total PH of DSTATCOM

installation scheme. It is observed from Fig. 11 that the total
Test network Approach

Proposed

Proposed

Proposed

planning cost with DSTATCOM is $5,39,491 and the TNP

is $1,60,078 since the total energy loss cost without DSTAT-
COM is $6,99,570. Therefore, 22.88% of TNP is achieved
using proposed approach. As can be seen in both Figs. 10
30-bus

33-bus

69-bus

and 11, there is an obvious rise in TNP by using proposed

approach compared to IA approach [27].

123
Arab J Sci Eng (2017) 42:2851–2865 2861

Table 7 Comparative results of

Test network Total energy loss cost with- Approach F($) TNP($)
annual cost of DN with
out DSTATCOM($)
DSTATCOM installation
without considering operational f 1 ($) f 21 ($)
and maintenance cost of
DSTATCOM 30-bus 1,16,374 Proposed 89,127 6399 20,848
33-bus 1,60,670 Proposed 1,26,679 6780 27,211
IA [27] 1,43,160 5989 11,521
GSA [21] – – 12,389
69-bus 1,80,470 Proposed 1,37,841 7,198 35,431
IA [27] 1,47,980 10,518 21,972
GSA [21] – – 12,837

7 Conclusion TNP. The reduction in total ELC after the installation of

This paper presents optimization of total ELC of DN with
DSTATCOM. In this scheme, minimization of power loss,
improvement in voltage profile, minimization of energy loss
cost and installation cost of the DSTATCOM, and the max-
imization of TNP are obtained by sizing and allocating the
DSTATCOM optimally using DEA method with a prede-
termined current, voltage and reactive power constraints in
DN. The objective function is defined by using energy losses
and its associated cost after the installation of DSTATCOM.
Forward–backward sweep load flow algorithm is used for
the solution of the network. Simulation results show that
the objective function is optimized with DSTATCOM using
DEA. Compared with IA, DEA technique offers minimum
power loss, CPU time, objective function and maximum
DSTATCOM by the proposed approach for the whole PH
of DSTATCOM installation scheme leads to 23.86, 21.14
and 23.61% in 30, 33 and 69 buses DN, respectively. TNP,
because of this approach, is valued to be of the order of
22.88%, in 30 bus DN. The simulation results in Table 7
of proposed work are found better compared to [21,27]. The
voltage at each bus and current in each line are within the
permissible boundaries.
Appendix
Tables 8 and 9 show the line and bus data in 33 and 69 bus
distribution networks in light loading.

Table 8 Data of 33 bus DN

Bus Line data Bus data
Send Receive Line resistance Line reactance Active load Reactive load
(Ω) (Ω) power (kW) power (kVAr)

1 2 0.0922 0.0470 100 60

2 3 0.4930 0.2511 90 40
3 4 0.3660 0.1864 120 80
4 5 0.3811 0.1941 60 30
5 6 0.8190 0.7070 60 20
6 7 0.1872 0.6188 200 100
7 8 0.7114 0.2351 200 100
8 9 1.0300 0.7400 60 20
9 10 1.0440 0.7400 60 20
10 11 0.1966 0.0650 45 30
11 12 0.3744 0.1238 60 35
12 13 1.4680 1.1550 60 35
13 14 0.5416 0.7129 120 80
14 15 0.5910 0.5260 60 10
15 16 0.7463 0.5450 60 20
16 17 1.2890 1.7210 60 20
17 18 0.7320 0.5740 90 40

123
2862 Arab J Sci Eng (2017) 42:2851–2865

Table 8 continued
Bus Line data Bus data
Send Receive Line resistance Line reactance Active load Reactive load
(Ω) (Ω) power (kW) power (kVAr)

2 19 0.1640 0.1565 90 40
19 20 1.5042 1.3554 90 40
20 21 0.4095 0.4784 90 40
21 22 0.7089 0.9373 90 40
3 23 0.4512 0.3083 90 50
23 24 0.8980 0.7091 420 200
24 25 0.8960 0.7011 420 200
6 26 0.2030 0.1034 60 25
26 27 0.2842 0.1447 60 25
27 28 1.0590 0.9337 60 20
28 29 0.8042 0.7006 120 70
29 30 0.5075 0.2585 200 600
30 31 0.9744 0.9630 150 70
31 32 0.3105 0.3619 210 100
32 33 0.3410 0.5302 60 40

Table 9 Data of 69 bus DN

Bus Line data Bus data
Send Receive Line resistance Line reactance Active load Reactive load
(Ω) (Ω) power (kW) power (kVAr)

1 2 0.0005 0.0012 0 0
2 3 0.0005 0.0012 0 0
3 4 0.0015 0.0036 0 0
4 5 0.0251 0.0294 0 0
5 6 0.366 0.1864 2.6 2.2
6 7 0.3811 0.1941 40.4 30
7 8 0.0922 0.047 75 54
8 9 0.0493 0.0251 30 22
9 10 0.819 0.2707 28 19
10 11 0.1872 0.0691 145 104
11 12 0.7114 0.2351 145 104
12 13 1.03 0.34 8 5.5
13 14 1.044 0.345 8 5.5
14 15 1.058 0.3496 0 0
15 16 0.1966 0.065 45.5 30
16 17 0.3744 0.1238 60 35
17 18 0.0047 0.0016 60 35
18 19 0.3276 0.1083 0 0
19 20 0.2106 0.0696 1 0.6
20 21 0.3416 0.1129 114 81
21 22 0.014 0.0046 5.3 3.5
22 23 0.1591 0.0526 0 0
23 24 0.3463 0.1145 28 20
24 25 0.7488 0.2745 0 0
25 26 0.3089 0.1021 14 10
26 27 0.1732 0.0572 14 10

123
Arab J Sci Eng (2017) 42:2851–2865 2863

Table 9 continued
Bus Line data Bus data
Send Receive Line resistance Line reactance Active load Reactive load
(Ω) (Ω) power (kW) power (kVAr)

3 28 0.0044 0.0108 26 18.6

28 29 0.064 0.1565 26 18.6
29 30 0.3978 0.1315 0 0
30 31 0.0702 0.0232 0 0
31 32 0.351 0.116 0 0
32 33 0.839 0.2816 14 10
33 34 1.708 0.5646 19.5 14
34 35 1.474 0.4873 6 4
3 36 0.0044 0.0108 26 18.55
36 37 0.064 0.1565 26 18.55
37 38 0.1053 0.123 0 0
38 39 0.0304 0.0355 24 17
39 40 0.0018 0.0021 24 17
40 41 0.7283 0.8509 1.2 1
41 42 0.31 0.3623 0 0
42 43 0.041 0.0478 6 4.3
43 44 0.0092 0.0116 0 0
44 45 0.1089 0.1373 39.22 26.3
45 46 0.0009 0.0012 39.22 26.3
4 47 0.0034 0.0084 0 0
47 48 0.0851 0.2083 79 56.4
48 49 0.2898 0.7091 384.7 274.5
49 50 0.0822 0.2011 384.7 274.5
8 51 0.0928 0.0473 40.5 28.3
51 52 0.3319 0.1114 3.6 2.7
9 53 0.174 0.0886 4.35 3.5
53 54 0.203 0.1034 26.4 19
54 55 0.2842 0.1447 24 17.2
55 56 0.2813 0.1433 0 0
56 57 1.59 0.5337 0 0
57 58 0.7837 0.263 0 0
58 59 0.3042 0.1006 100 72
59 60 0.3861 0.1172 0 0
60 61 0.5075 0.2585 1244 888
61 62 0.0974 0.0496 32 23
62 63 0.145 0.0738 0 0
63 64 0.7105 0.3619 227 162
64 65 1.041 0.5302 59 42
11 66 0.2012 0.0611 18 13
66 67 0.0047 0.0014 18 13
12 68 0.7394 0.2444 28 20
68 69 0.0047 0.0016 28 20

123
2864
References
1. Moradi, M.H.; Zeinalzadeh, A.; Mohammadi, Y.; Abedini, M.: An
efficient hybrid method for solving the optimal sitting and sizing
problem of DG and shunt capacitor banks simultaneously based
on imperialist competitive algorithm and genetic algorithm. Int. J.
Electr. Power Energy Syst. 54, 101–111 (2014)
2. Moradi, M.H.; Abedini, M.: A combination of genetic algorithm
and particle swarm optimization for optimal DG location and sizing
in distribution systems. Int. J. Electr. Power Energy Syst. 34(1), 66–
74 (2012)
3. Naderi, E.; Seifi, H.; Sepasian, S.M.: A dynamic approach for dis-
tribution system planning considering distributed generation. IEEE
Trans. Power Deliv. 27(3), 1313–1322 (2012)
4. Shuaib, Y.N.; Kalavathi, M.S.; Rajan, C.C.A.: Optimal capacitor
placement in radial distribution system using gravitational search
algorithm. Int. J. Electr. Power Energy Syst. 64, 384–397 (2015)
5. Sultana, S.; Roy, P.K.: Optimal capacitor placement in radial dis-
tribution systems using teaching learning based optimization. Int.
J. Electr. Power Energy Syst. 54, 387–398 (2014)
6. Georgilakis, P.S.; Hatziargyriou, N.D.: A review of power distribu-
tion planning in the modern power systems era: models, methods
and future research. Electr. Power Syst. Res. 121, 89–100 (2015)
7. Ganguly, S.; Sahoo, N.C.; Das, D.: Multi-objective planning of
electrical distribution systems using dynamic programming. Int. J.
Electr. Power Energy Syst. 46, 65–78 (2013)
8. Hung, D.Q.; Mithulananthan, N.; Bansal, R.C.: An optimal invest-
ment planning framework for multiple distributed generation units
in industrial distribution systems. Appl. Energy 124, 62–72 (2014)
9. Niknam, T.; Taheri, S.I.; Aghaei, J.; Tabatabaei, S.; Nayeripour,
M.: A modified honey bee mating optimization algorithm for mul-
tiobjective placement of renewable energy resources. Appl. Energy
88(12), 4817–4830 (2011)
10. Soroudi, A.; Ehsan, M.; Zareipour, H.: A practical eco-
environmental distribution network planning model including fuel
cells and non-renewable distributed energy resources. Renew.
Energy 36(1), 179–188 (2011)
11. Kucuksari, S.; Khaleghi, A.M.; Hamidi, M.; Zhang, Y.; Szi-
darovszky, F.; Bayraksan, G.; Son, Y.J.: An integrated GIS,
optimization and simulation framework for optimal PV size and
location in campus area environments. Appl. Energy 113, 1601–
1613 (2014)
12. Hung, D.Q.; Mithulananthan, N.: Multiple distributed generator
placement in primary distribution networks for loss reduction.
IEEE Trans. Ind. Electron. 60(4), 1700–1708 (2013)
13. Georgilakis, P.S.; Hatziargyriou, N.D.: Optimal distributed gener-
ation placement in power distribution networks: models, methods,
and future research. IEEE Trans. Power Syst. 28(3), 3420–3428
(2013)
14. Li, W.; Joos, G.; Belanger, J.: Real-time simulation of a wind tur-
bine generator coupled with a battery supercapacitor energy storage
system. IEEE Trans. Ind. Electron. 57(4), 1137–1145 (2010)
15. Pigazo, A.; Liserre, M.; Mastromauro, R.A.; Moreno, V.M.;
Dell’Aquila, A.: Wavelet-based islanding detection in grid-
connected PV systems. IEEE Trans. Ind. Electron. 56(11), 4445–
4455 (2009)
16. Xu, Y.; Dong, Z.Y.; Wong, K.P.; Liu, E.; Yue, B.: Optimal capacitor
placement to distribution transformers for power loss reduction in
radial distribution systems. IEEE Trans. Power Syst. 28(4), 4072–
4079 (2013)
17. Chiou, J.P.; Chang, C.F.; Su, C.T.: Ant direction hybrid differential
evolution for solving large capacitor placement problems. IEEE
Trans. Power Syst. 19(4), 1794–1800 (2004)
18. Haghifam, M.R.; Malik, O.P.: Genetic algorithm-based approach
for fixed and switchable capacitors placement in distribution sys-
123
Arab J Sci Eng (2017) 42:2851–2865
tems with uncertainty and time varying loads. IET Gen. Trans. Dist.
1, 244–252 (2007)
19. Huang, S.J.: An immune-based optimization method to capacitor
placement in a radial distribution system. IEEE Trans. Power Deliv.
15(2), 744–749 (2000)
20. Sanam, J.; Ganguly, S.: Impact of distribution STATCOM allo-
cation on radial distribution networks. In: IEEE International
Conference on Energy, Power and Environment: Towards Sustain-
able Growth (ICEPE), pp. 1–6 (June 2015)
21. Shuaib, Y.M.; Kalavathi, M.S.; Rajan, C.C.A.: Optimal capacitor
placement in radial distribution system using gravitational search
algorithm. Int. J. Electr. Power Energy Syst. 64, 384–397 (2015)
22. Eldery, M.A.; El-Saadany, E.F.; Salama, M.M.A.: Effect of dis-
tributed generator on the allocation D-STATCOM in distribution
network. In: Power Engineering Society IEEE General Meeting,
pp. 2360–2364 (2005)
23. Devi, S.; Geethanjali, M.: Optimal location and sizing determi-
nation of distributed generation and DSTATCOM using particle
swarm optimization algorithm. Int. J. Electr. Power Energy Syst.
62, 562–570 (2014)
24. Tolabi, H.B.; Ali, M.H.; Rizwan, M.: Simultaneous reconfigura-
tion optimal placement of DSTATCOM and photovoltaic array in
a distribution system based on fuzzy-ACO approach. IEEE Trans.
Sustain. Energy 6(1), 210–218 (2015)
25. Jazebi, S.; Hosseinian, S.H.; Vahidi, B.: DSTATCOM alloca-
tion in distribution networks considering reconfiguration using
differential evolution algorithm. Energy Convers. Manag. 52(7),
2777–2783 (2011)
26. Sanam, J.; Panda, A.K.; Ganguly, S.: Optimal phase angle injection
for reactive power compensation of distribution systems with the
allocation of multiple distribution STATCOM. Arab. J. Sci. Eng.
1–9 (2016)
27. Taher, S.A.; Afsari, S.A.: Optimal location and sizing of DSTAT-
COM in distribution systems by immune algorithm. Int. J. Electr.
Power Energy Syst. 60, 34–44 (2014)
28. Hsieh, S.C.: Economic evaluation of the hybrid enhancing scheme
with DSTATCOM and active power curtailment for PV penetration
in taipower distribution systems. IEEE Trans. Ind. Appl. 51(3),
1953–1961 (2015)
29. Das, S.; Suganthan, P.N.: Differential evolution: a survey of the
state-of-the-art. IEEE Trans. Evolut. Comput. 15(1), 4–31 (2011)
30. Varadarajan, M.; Swarup, K.S.: Differential evolutionary algorithm
for optimal reactive power dispatch. Int. J. Electr. Power Energy
Syst. 30(8), 435–441 (2008)
31. Sanam, J.; Ganguly, S.; Panda, A.K.: Placement of DSTATCOM in
radial distribution systems for the compensation of reactive power.
In: IEEE PES Intl. Conf on Innovative Smart Grid Technologies-
Asia (ISGT ASIA), pp. 1–6 (2015)
32. Sajjadi, S.M.; Haghifam, M.R.; Salehi, J.: Simultaneous placement
of distributed generation and capacitors in distribution networks
considering voltage stability index. Int. J. Electr. Power Energy
Syst. 46, 366–375 (2013)
33. Ganguly, S.: Multi-objective planning for reactive power compen-
sation of radial distribution networks with unified power quality
conditioner allocation using particle swarm optimization. IEEE
Trans. Power Syst. 29(4), 1801–1810 (2014)
34. Gitizadeh, M.; Vahed, A.A.; Aghaei, J.: Multistage distribution-
system expansion planning considering distributed generation
usinghybrid evolutionary algorithms. Appl. Energy 101, 655–666
(2013)
35. Raju, M.R.; Murthy, K.V.S.R.; Ravindra, K.: Direct search algo-
rithm for capacitive compensation in radial distribution systems.
Int. J. Electr. Power Energy Syst. 42(1), 24–30 (2012)
36. Sundhararajan, S.; Pahwa, A.: Optimal selection of capacitors for
radial distribution systems using a genetic algorithm. IEEE Trans.
Power Syst. 9(3), 1499–1507 (1994)
Arab J Sci Eng (2017) 42:2851–2865 2865

37. Savier, J.S.; Das, D.: Energy loss allocation in radial distribution 38. Jeynes, P.H.: Present worth, or the time cost of money, as a factor in
systems: a comparison of practical algorithms. IEEE Trans. Power the economic comparison of alternative facilities. Trans. Am. Inst.
Deliv. 24(1), 260–267 (2009) Electr. Eng. 70(2), 1956–1962 (1951)

123