Electrical Power and Energy Systems 67 (2015) 191201

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

An analytical approach for sizing and siting of DGs in balanced radial

distribution networks for loss minimization
Rajkumar Viral , D.K. Khatod
Alternate Hydro Energy Centre, Indian Institute of Technology Roorkee, Roorkee, Uttarakhand 247667, India

a r t i c l e

i n f o

Article history:
Received 22 February 2014
Received in revised form 9 November 2014
Accepted 21 November 2014
Available online 10 December 2014
Keywords:
Analytical approach
Loss minimization
Loss saving
Optimal siting
Optimal sizing

a b s t r a c t
This paper presents a novel analytical approach to determine the optimal siting and sizing of distributed
generation (DG) units in balanced radial distribution network to minimize the power loss of the system.
The proposed analytical expressions are based on a minimizing the loss associated with the active and
reactive component of branch currents by placing the DG at various locations. This method rst identies
a sequence of nodes where DG units are to be placed. The optimal sizes of DG units at the identied nodes
are then evaluated by optimizing the loss saving equations and need only the results of base case load
ow. To nd out the best location for DG placement, a computational method is also developed. The proposed method has been tested and validated on two IEEE test distribution systems (DSs) consisting of 15
and 33-buses and it has been found that a signicant loss saving can be obtained by placing DG units in
the system using proposed analytical method.
2014 Elsevier Ltd. All rights reserved.

Introduction
R/X ratio in distribution networks is much higher rather than
transmission systems, and result of higher power losses and gradually loss of electrical energy along the distribution feeders [15].
Consequently, for many utilities all over the world loss minimization is one of the biggest question. Two extensive methods for loss
minimization in distribution networks are network reconguration
and capacitors placement well known and used frequently [2,36].
In recent past, DG has attained signicant interest and can be
defended, aspects such as environmental concerns, the restructuring of electricity market, the development in advance technologies
for small-scale power generation, power electronics, and energy
storage devices for transient backup into electric power DS [6,7].
However, this inclination has extended considerable opportunities
but devised several confrontations in planning and operations of
DSs. DGs are dened as electric power generations directly connected to loads or distribution networks; they range from a few
kWs to a few MWs [7,8]. Today, there are many DG technologies
in trend cover conventional (such as micro turbines, combustion
turbine, combined cycle, and internal combustion engines) to
non-conventional (such as ocean, photovoltaic solar, fuel cell,
wind, and geothermal) [1,7,8].
Corresponding author. Tel.: +91 9897 675190; fax: +91 1332 286125.
E-mail addresses: [email protected] (R. Viral), [email protected] (D.K.
Khatod).
http://dx.doi.org/10.1016/j.ijepes.2014.11.017
0142-0615/ 2014 Elsevier Ltd. All rights reserved.

The essential objective of DG units is energy injection; despite,

strategically placed and operated DG units can offer several other
benets (i.e. technical and economical) to utilities as well as to customers [9]. Typical cases of such benets are the application of DG
units for loss reduction, voltage and loadability improvement,
enhanced system reliability and security, improved power quality,
increased overall energy efciency, and relieved transmission and
distribution (T&D) [68]. While, economical benets cover saving
world fuel, saving T&D cost and reducing whole sale electricity
price. Deferred investments for upgrades of facilities, reduced
operational and maintenance (O&M) costs, enhanced productivity,
reduced fuel costs due to increased overall efciency, reduced
reserve requirements and the associated costs, lower operating
costs due to peak shaving are the additional economical benets
[8,10,11]. As far concern to the electricity market security todays
deregulation of power industry, DG units play an important role
in ancillary services such as reactive power support, spinning
reserve, loss compensation, frequency control and other fast
response services [9,12]. Moreover, in achieving the benets of
these ancillary services; DG units have been come into view as
an integral part of DS. Although, inadequately operated and poorly
planned DG units may also have some adverse effect on the functioning of the DS; based on the size, location, and inltration level
they can lead to reverse power ows, voltage rise, increased fault
levels, more power losses, harmonic distortion, stability problems
and consecutive feeder overloads [1,911,13].

192

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

It is evident that loss reduction is one of the most substantial

and benecial factor to be treated in DG planning and operation
apart from factors discussed above. The major challenges in planning of DG for loss minimization are suitable location, proper sizes,
and operating strategies. The DG optimal sizing and siting for minimizing losses has drawn increasing attention of the extensive
group of researchers in the recent years. There have been diverse
techniques/approaches employed to cover the DG siting and sizing
problem in DS for power loss minimization considering different
type of DG technologies with their relative advantages and disadvantages, in attaining this distinct objective along with their practicability as examined in [1,7,8,10,11]. Therefore, this paper reports
the development of some simple analytical expressions for sizing
and siting of DG units, which can be easily implemented in a
balanced radial DS.
Remaining of the paper is set out as below: Section Loss minimization techniques describe a concise literature review on earlier
loss minimization methods/techniques for DG planning. In Section
Proposed methodology, proposed analytical method for optimal
size and siting for single and multiple DG is discussed. The detailed
computational procedure is elaborated in Section Computational
procedure. Numerical results and simulation of developed analytical method applied in two IEEE test systems, interesting ndings
along with discussion addressed in Section Numerical results. Section Conclusion, summarizes the major contributions and
conclusions.

Loss minimization techniques

It has been realized that most of the existing work on DG siting
and sizing in the DS, discussed different issues such as minimization of system power loss [1,6,10,11,1325], abatement of harmonic pollution [19], enhancement of system voltage prole and
stability [12,13,1518,2225], investment minimization or prot
maximization [26,27], and loading margin [28] have been intended
by researchers in their single or multi-objective problem formulations. Different optimization techniques, such as analytical
approach [1,911,15,16,19,25], mixed integer non-linear programming (MINLP) [12,13,17,18], evolutionary algorithms (EA) technique [9], metaheuristic approaches: meta-heuristic harmony
search algorithm (HSA) [6], particle swarm optimization (PSO)
[20], heuristic approaches [27]; trade-off method [20], genetic
algorithm (GA) technique [26,28], Kalman lter algorithm [23],
multi-period AC optimal power ow (OPF) solver tool [20], and
multi-objective non-linear programming (NLP) [15] have been
used to solve the optimization problems for DG siting and sizing.
Except these, there have been many interesting studies on the
DGs siting and sizing of DS for loss minimization.
An analytical technique was noticed in [16] to nd out the allocation of a single DG in radial as well as mesh networks to minimize the losses, based on unity power factor. However, optimal
sizing is not taken into account. A more faster and precise analytical method than the classical methods [6,20,27] based on the
equivalent current injection technique and without the use of
impedance or Jacobian matrices for optimum size and location of
DG in radial systems has been implemented in [19]. Moreover, this
method was in near concurrence with the analytical method
inscribed in [15]; in which an exact loss formula based analytical
approach has been investigated to identify the optimal size and
location of single DG in two load ow solution.
Pursuing the aforesaid work, various analytical expressions
based on exact loss formula for optimal allocation of DGs were
addressed in [1,10,11]. An efcient solution based on improved
analytical (IA) expression to locate and size of four type of (renewable and non-renewable) DGs for loss minimization has been

examined in [1]. Although multiple DGs allocation was not considered. Contrary to this, the same authors in [10] applied the same
approach for multiple DG unit placement to get an utmost loss
minimization in large size primary DSs. Similar kind of work was
also noticed in [11] using three analytical expressions to obtain
the optimum sizes and locations of renewable DGs for power loss
reduction considering the combination of time-varying demand
and different DG output curves.
Moreover, in [12,17,18] technique based on probabilistic planning and formulated as MINLP problem have been acquainted to
the readers. In [17,18], this technique enforced to identifying the
best supply, unify of various classes of non-conventional DGs (i.e.
wind, biomass, and solar) to reduce the power losses yearly in
DS; although, DGs competent of bringing active power only is
taken into account in both the studies. Similarly, in [12] same
approach was implemented on renewable DGs for best location
and size so as to enhance the voltage stability margin (VSM).
In the line of above, in [23] the optimal size of DGs is determined using the Kalman lter algorithm so that total power losses
are minimized. A multi-objective index-based technique to determine optimal size and location of DG units in DS with non-unity
power factor considering different load models has been exposed
in [24].
A multi-period AC-OPF solver based method is discussed for to
determine optimal power of renewable DG sources and there size
to minimize the total energy losses during a period in [20]. Authors
in [22] considered an iterative DG placement technique to improve
the VSM. Though losses and optimal size of DGs not considered and
a xed value is assumed for all DGs. A multi-objective method is
examined in [28] for optimal placement of DGs with for loading
margin and prot to be maximized considering network constraints. Although, losses and xed reactive limits for unknown
DG sizes are not studied. Recently, a new multi-objective index
(IMO) based analytical expressions to accommodate a combination
of photovoltaic and battery energy storage DG units for reducing
energy loss and enhancing voltage stability suggested in [25] using
self-correction algorithm (SCA), while considering the time-varying demand and probabilistic generation.
Most of the studies reported above, DGs considered as only pure
active power source. However, it is more benecial to improve performance of DS, when the DG units supply reactive power. Depending on the type of DG used; they can able to inject or absorbs
reactive power within their capability limits [13]. Furthermore,
large number of the commonly used analytical techniques for DG
siting and sizing are depend on exact loss formula and expect the
evaluation of the Jacobian matrix and computationally demanding
more time. Therefore, the above said methods are not quite appropriate due to the intricacy, capacity and the distinct property of the
DS. Consequently, the optimal allocation of either type of DG using
optimal solution methodology draws added consideration.
To overcome the obstacles of earlier studies and motivated by
the work of [35,29], this paper proposes to apply a novel and
simple analytical approach which is based on the DG active
and reactive branch currents and the associated loss saving for
allocating the DG units for loss reduction in the radial DS. The
procedure rst determines the location of the DG in a consecutive way. Erstwhile the DG locations are obtained, the optimal
DG capacity at each chosen locations are nd out by optimizing
the loss saving equation.

Proposed methodology
This segment set forth on a detailed mathematical formulation
of the proposed analytical method. To develop the formulation following are the assumptions and constraints used in this paper:

193

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

i. Considered distribution system is a single source, radial, and

balanced.
ii. The lower and upper voltage limits are set at 0.90 and
1.05 pu.
iii. Load level is constant.
iv. Maximum DG capacity for different test systems is assumed
to be equal to the total load of the system.
v. Maximum loss saving by the DGs for different test systems
should not be less than zero (i.e. maximum loss saving <0)
The power ow solution will be used to check the limit violations of all above assumption and constraints.
Consider a N-bus radial DS having n number of branches as
shown in Fig. 1. Here, Ii is the current in branch i before DG
placement.
The total power loss of the system, i.e. PL can be given as

PL

N1
N1 

X
X
I2i  Ri
I2ai I2ri  Ri
i1

i1

PLa

i1
N1
X
I2ri  Ri
PLr


 

m
m
I new
I i  Di  I m
i
dg I ai  Di  I dga j Iri  Di  Idgr
Im
dg

Im
dga

m
jIdgr .

where I i Iai jIri and

m
Here, Im
and
I
are
the
real
and imaginary components of I m
dga
dgr
dg
and Di can be given as:


Di

1; if branch i is between bus 1 and m

0;

otherwise

Now the total compensated losses P new

after DG placement may
L
be expressed as

Pnew

N 1
X


Inew
i

2

 Ri

i1

Here, Iai and Iri are the real and imaginary components of Ii (= Iai + jIri)
respectively, the complex current in branch i and Ri is the resistance
of ith branch. The Ii can be obtained from the load ow solution of
the given DS [5]. The loss associated with the active and reactive
components of branch currents could be expressed as PLa and PLr
respectively and given as
N1
X

I2ai  Ri

When a single DG is placed at bus m; injecting a current I m

dg . This
current changes the currents in all the branches connected
between substation (bus no. 1) to bus m and the currents in the
remaining branch are unaffected by the DG. Therefore, modied
phasor current in ith branch, I new
can be written mathematically
i
as:

i1

The procedure of single and multiple DG placement for loss

minimization is addressed in the following section.

N 1 
2 
2 
X
m
 Ri
Iai  Di  Im

I

D

I
ri
i
dga
dgr

i1

Loss saving in single DG case

The associated loss saving, SL may be given using (1) and (5) as:

SL PL  Pnew
L
N1 
 
2 
2 
X
 Ri :

Iai 2 I2ri  Iai  Di  Im

Iri  Di  Im
dga
dgr

i1

Simplifying the above equation, we get

SL 2

N1 h
X

N1 
i
2  2 
X
m
 D2i  Ri
Iai  Im
Im
Im
dga I ri  I dgr  Di  Ri 
dga
dgr

i1

i1

Loss minimization by single DG placement

Fig. 2 demonstrates a N-bus radial distribution network with
DG at mth bus. Here, Inew
is the current in branch i after DG placei
ment in the system.

The maximum loss saving can be obtained corresponding to the

m
DG currents Im
dga and I dgr using (7) when the following conditions
are to be satised

Fig. 1. A N-bus radial distribution network.

Fig. 2. A N-bus radial distribution network after DG placement.

194

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

9
@SL
>
>

0
>
>
=
@Im
dga

>
@SL
>
>
>
m 0;
@Idgr

SL PL  Pnew
L
2
!2 3
N1
k
X
X
j
2
4I  Iai 

Dij  Idga 5  Ri
ai

Hence from (8) we get

N1
N1
X
X
@SL
2
Iai  Di  Ri  2 Im
m 2
dga  Di  Ri 0
@Idga
i1
i1

N1
N1
X
X
@SL
2
Iri  Di  Ri  2 Im
m 2
dgr  Di  Ri 0
@Idgr
i1
i1

j1

2
!2 3
N1
k
X
X
2
j
4I  Iri 

Dij  Idgr 5  Ri
ri
i1

j1

"
#
N1
k
k
X
X
X
2
Iai 
Dij  Ijdga Iri 
Dij  Ijdgr  Ri
i1

PN1

i1 Di  I ai  Ri
PN1 2
i1 Di  Ri

11

Similarly, the reactive current Im

dgr corresponding the maximum
loss saving can be obtained from (10) and given as

PN1
i1 Di  I ri  Ri
Im

PN1 2
dgr
i1 Di  Ri

12

j1

N1
X

k
X

i1

j1

Dij  Ijdga

j1

!2

k
X

Dij  Ijdgr

!2 3
5  Ri

16

j1

The maximum loss saving could be achieved by placing multiple DGs if following conditions are to be satised:

9
>
0>
>
>
>
>
>
>
>
>
@SL
>
0>
>
1
>
>
@Idgr
>
>
=
..
.
>
>
>
>
@SL
>
>

0
>
>
k
>
@Idga
>
>
>
>
>
@SL
>
>
>

0
;
k
@Idgr
@SL

@I1dga

Optimal size in single DG case

The corresponding single DG size may be obtained using (10)
and (12) as

Sm
dg

i1

10

Simplifying (9) active DG current Im

dga corresponding to the maximum loss saving can be given as

Im
dga

Loss saving in multi DG case

The associated loss saving in multi DG case can be obtained by
subtracting (15) from (1) gives the loss saving SL due to integration
of the DGs in the DS as:

 


m
V m  Im
V m  Im
dg
dga  I dgr

13

In (13) S m
dg is the capacity of mth DG in complex form and Vm is the
phasor voltage at bus to which mth DG is connected.

17

Loss minimization by multiple DG placement

m
The partial derivative of SL with respect to Im
dga and I dgr can be
given as:

This section extends the previously developed approach for

placement of multiple DG units simultaneously. Now, suppose in
an N-bus distribution network, k numbers of DGs are to be placed.
Integration of DGs at different buses alters the ow of branch current; therefore the modied current in branch i can be given mathematically as:

"
#
N1
k
X
X
@SL
j
2
Dim  Iai 
Dij  Idga  Ri 0
@Im
dga
i1
j1
"
#
N1
k
X
X
@SL
j
2
Dim  Iri 
Dij  Idgr  Ri 0
@Im
dgr
i1
j1

I new
Ii 
i

k
X
Dij  I jdg
j1

Iai 

k
X

!
Dij 

Ijdga

k
X
j Iri 
Dij  Ijdgr

j1

!
14

j1

8
>
< 1; if ith branch is between S=s bus and bus at
th
Dij
which j DG is placed
>
:
0; otherwise

2
Inew
i

i1

 Ri

k
X
Iai 
Dij  Ijdga
j1

20

Akk  I dgr k1 Ck1

21

The elements of A, B, and C can be calculated as:

Apq
App

N1
X
Dip  Diq  Ri

22

N1
X
D2ip  Ri

23

i1

Bp

N1
X

Dip  Iai  Ri

24

N1
X
Dip  Iri  Ri

25

i1

Cp

 Ri

i1
N 1
X

Akk  I dga k1 Bk1

i1

The detail evaluation of Dij for multi DG is given in Appendix.

When DGs are connected in the DS, the associated compensated
losses, Pnew
can be written as
L
N1
X


19

Corresponding to (17), there will be total 2k linear algebraic

equations, k out of which are similar to (18) and remaining are
similar to (19). These two sets of equations can be represented as:

where I new
is the complex current in ith branch after DGs placei
ment, I jdg is the current injected by jth DG with I jdga and Ijdgr being
its real and imaginary parts, respectively.

Pnew
L

18

i1

!2
 Ri

N 1
X
i1

k
X
Iri 
Dij  Ijdgr

!2

j1

15

where Apq and App are the off-diagonal and diagonal elements of
matrix A. Similarly, Bp and Cp are the respective elements of matrix
B and C. It is evident from the (22)(25), that only the branch resistance, active and reactive components of branch current of base
case are required to nd the elements A, B and C. From (20) and

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

(21), the active and reactive components of DG currents for maximum loss saving can be computed as:

I dga k1 A1

kk  Bk1

26

I dgr k1 A1

kk  Ck1

27

Optimal size in multi DG case

Once the active and reactive components of DG currents are
known from (26) and (27), the optimal size of DGs can be calculated as:

 


m
m
Sm
V m  Im
dga  jI dgr
dg V m  I dg

28

Computational procedure
To nd out the optimal DG size and location for minimization of
loss in radial DS, following computational steps are involved.

195

Step 1: Read the system data and run the base load ow program for the original uncompensated system; obtain the branch
currents, bus voltages, real power losses other necessary data.
The load ow program for the proposed methodology implementation is taken from [5]. The detail description of the base
load ow interpreted in ow chart in Fig. 3.
Step 2: Assume that every node is candidate node. Calculate the
loss saving and consequent DG size using (7) and (13) at each
bus except the source bus.
Step 3: Select the bus that yields the maximum loss saving and
it is corresponding DG size for compensation and is called a sensitive bus/node. This is case of single DG.
Step 4: Place the DG obtained from step (3) at the bus which has
maximum loss saving and repeat step (1)(3) again to get the
next DG bus. Find out the sequence of nodes which is to be
replaced/compensated until no such convincing loss saving
achieved or reached to zero value by further DG placement
otherwise stop the program.
Step 5: As from step (4), the sequence of nodes is known now.
Calculate the optimal DG sizes and loss saving using (28) and

Fig. 3. Flow chart of the proposed analytical method to allocate single and multiple DGs.

196

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

(16) respectively. This is the case of multi DG. Finally, obtain the
optimal number of DGs to be placed.
This process is repeated iteratively until the total loss saving
reach zero value or no further loss saving could be achieved. The
method of obtaining the optimal DG sizing and siting is outlined
in owchart in Fig. 3.
It is worth mentioning here that the proposed technique is a
SCA (i.e. after setting two or more DGs in the system). The next
decision could be to reduce the size of DG that is already set at a
certain node to obtain more loss saving. This implies that this
bus is overcompensated and we have to reduce the DG size placed
at this bus to obtain further reduction in loss. This is an advantage
in the proposed technique. Additionally, another advantage of SCA
implementation is that it requires less number of iteration to
achieve the convergence because the self-correction process is only
implemented at the selected buses obtained earlier. Another benet is that the total number of load ow used normally remains
unchanged for larger system such as 69 or 118 bus system [25,29].
Numerical results
Test systems
Two test systems have been employed to test the proposed analytical approach for optimal sizing and placement of DGs. The rst
system is an 11 kV; IEEE 15-bus radial DS is taken into consideration with the total load of (1.2264 + j1.2512) MV A and the total
I2R losses are 61.79 kW [30]. The second system is 12.66 kV, IEEE
33-bus test system with total load of (3.715 + j2.300) MV A radial
DS and the total I2R losses of the base system are 197.94 kW [2].
To implement the above algorithm an analytical software tool
has been developed in MATLAB environment to run load ow,
determine losses and optimal sizes of DG.
Simulation results
15-bus IEEE test system
The single-line diagram of this system is shown in Fig. A.1 in
Appendix (considering no DGs). The data of the system are taken

from [30]. The losses consociated with the active and reactive elements of branch currents prevailed with the power ow method
are 30.42 kW and 31.37 kW respectively.
First, the optimal loss saving and corresponding DG size is
determined. Figs. 4 and 5 depicts the DG size and loss saving
respectively for all the buses in the system excluding the source
bus. It can be observed from Fig. 5 that the highest loss saving of
24.55 kW can be realized by placing a DG of 1421.26 kV A (Fig. 4)
at bus 3 in the rst iteration. The detailed summary of results
obtained by the proposed method is illustrated in Table 1. When
the above procedure is applied again, after placing 1421.26 kV A
of DG at bus 3, it was evident that a second DG of 728.73 kV A at
bus 6, will give a further loss saving of 3.67 kW in the same
iteration.
Again, when 728.73 kV A of DG is placed at bus 6, we achieve a
loss saving of 1.77 kW with next DG size of 575.38 kV A at bus 7 in
the rst iteration. Once again, repeating the same process and
place 575.38 kV A of DG at bus 7, it was found that no further loss
saving could be obtain because the optimization problem was
solved to nd the size of a singly located DG (as described in technique of Section Loss minimization by single DG placement) and
loss saving till reaches zero value. Thus the sequence of buses to
be redeemed is 3, 6 and 7. The total power losses PL retrieved in
case of aforementioned single DG placement reduced from
61.79 kW to 36.69 kW with the percentage loss reduction of
40.62%. As said above, the optimization technique is a SCA, therefore important to mention here that in case of single DG, the MATLAB codes are set in such a way that it automatically disconnects
the previously connected DG and consider only the next single
DG available in the system except the DG size achieve in base case
and this operation takes only one iteration as illustrated in Table 1.
However, when multiple DG is to be placed in the system, technique reported in Section Loss minimization by multiple DG placement has to be implemented. The above techniques contribute a
total loss saving of 29.94 kW by placing the two DG 1129.69 kV A
and 544.64 kV A at buses 3 and 6 respectively in the second iteration. When the original base system is redeemed with above DGs,
the load ow result pointed out that the total power losses PL
reduced by 61.7936.60 kW with the percentage loss reduction
of 40.76%.

Fig. 4. Optimal size of singly located DG of 15-bus system.

197

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

Fig. 5. Estimated loss saving in 15-bus system for a singly located DG.

However, when all three buses (3, 6 and 7) are considered for
DG placement it was observed that a total 1673.86 kV A of DGs
(1129.69 kV A at bus 3, 344.18 kV A at bus 6, and 199.99 kV A at
bus 7) contribute a loss saving of 30.02 kW in third iteration. The
associated power losses PL by placement of all three DGs are
reduced by 61.7935.63 kW with the percentage loss reduction
of 42.33%. Consequently, the respective loss reduction associated
with the active and reactive component of branch currents are
30.4229.00 kW and 31.376.63 kW respectively. It is also realized
that the proposed technique able to improve the voltage prole of
the system under consideration as represented in Fig. 6, with and
without DG in single DG case. Similarly, Table 2 highlighted the
minimum and maximum voltages deviations before and after DG
by proposed method for 15-bus system. It could be examined that
the voltage at various buses maintain within the acceptable constraints limits. In the end, the results obtained by the proposed
method are compared based on size, location, percentage loss
reduction and computation time, with those methods addressed

in the literature for 15-bus test system and illustrated in Table 3

for validation purpose. The results indicate in this table are for siting and sizing of single DG to minimize the real power loss only. It
is evident from this table that better active power loss is possible
to reduce by the proposed method as distinguished in the other
methods revealed in the literature.
33-bus IEEE test system
Fig. 7 shows the single line diagram of the IEEE 33-bus system.
The data of the system acquired from [2]. Whereas the losses
linked with the active and reactive components of branch currents
abound with the power ow in base case are 131.41 kW and
66.52 kW, respectively. Enforced the methodology addressed in
Section Loss minimization by single DG placement for single DG
placement and nd the DG size and loss saving. Similar to the
15-bus system, rst the optimal loss saving and corresponding
DG size is calculated and delineate in Figs. 9 and 8. It can be recognized from Fig. 9 that the highest loss saving of 53.30 kW can be

Table 1
Summary of results of 15-bus system.
S. no.

System description

Base system

Single DG
placement

Base system

Multi DG
placement

Base system

Multi DG
placement

1421.26 kV A DG is placed at bus 3

728.73 kV A DG is placed at bus 6
575.38 kV A DG is placed at bus 7

Placed 1129.69 kV A DG at bus 3 and

544.64 kV A DG at bus 6

Placed 1129.69 kV A DG at bus 3,

344.18 kV A DG at bus 6 and
199.99 kV A DG at bus 7

System losses

DG size

PLa
(kW)

PLr
(kW)

30.42

31.37

61.79

29.12
29.01
28.89

7.98
7.67
7.58

37.11
36.69
36.47

30.42

31.37

61.79

29.04

7.56

30.42

29.00

PL
(kW)

Pdga
(kW)

DG
location
(bus no.)

Number
of
iteration

Total
loss
saving SL
(kW)

% Loss
reduction

Qdga
(kVAr)

S (kV A)

996.99

1012.92

1421.26

24.55

651.38
539.02
486.23

326.72
201.31
187.56

728.73
575.38
520.93

6
7
3

1
1
1

3.67
1.77
43.23

39.94
40.62
40.97

792.19
382.56

805.39
387.66

1129.69
544.64

3
6

36.60

1174.74

1193.04

1674.33

3 and 6

29.94

40.76

31.37

61.79

792.19
241.69
140.54

805.39
245.04
142.29

1129.69
344.18
199.99

3
6
7

6.63

35.63

1174.42

1192.71

1673.86

3, 6 and
7

30.02

42.33

198

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

Fig. 6. Voltage prole of 15-bus test system with and without DG.

Table 2
Voltage before and after DG at 15-bus system.
System

15-bus

Voltage at bus before DG

Voltage at bus after DG

Min

Max

Min

Max

0.9452 at 13

1.000 at 1

0.9721 at 13

0.9899 at 9

Table 3
Comparison of results for 15-bus test system.
Particulars

Gozel and
Hocaoglu [19]

Murthy and
Kumar [31]

Proposed
method

Optimal bus
DG size (MV A)
% Loss reduction
CPU time (s)

3
1.418
39.041
0.032

3
1.411
42.008
0.041

3
1.421
42.331
0.023

Fig. 7. Single line diagram of the IEEE 33-bus test system.

realized by placing a DG of 2968.53 kV A (Fig. 8) at bus 6. The

detailed summary of results accessed by the suggested method is
demonstrated in Table 4.

Fig. 8. Optimal size of DG in 33-bus system for a singly located.

199

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

Fig. 9. Estimated loss saving in 33-bus system for a singly located DG.
Table 4
Summary of results of 33-bus system.
S. no.

System description

Base system

Single DG
placement

Base system

Multi DG
placement

Base system

Multi DG
placement

DG
location
(bus no.)

Number
of
iteration

Total loss
saving SL
(kW)

2968.53

53.30

1133.41
1680.25

14
24

1
1

6.79
2.98

29.69
33.35

1656.11

33.88

1053.75
370.29

1537.97
858.89

6
14

1895.23

1424.04

2396.86

6 and 14

58.45

33.55

197.94

1023.31
725.22
1068.39

1006.51
346.05
520.56

1435.35
803.55
1188.46

6
14
24

127.86

2816.92

1873.12

3427.36

6, 14
and 24

62.94

35.40

System losses

2968.33 kV A DG is placed at bus 6

1133.41 kV A DG is placed at bus
14
1680.25 kV A DG is placed at bus
24

Placed 1537.97 kV A DG at bus 6

and 858.89 kV A DG at bus 14

Placed 1435.35 kV A DG at bus 6,

803.55 kV A DG at bus 14 and
1188.46 kV A DG at bus 24

DG size

PLa
(kW)

PLr
(kW)

PL
(kW)

Pdga
(kW)

Qdga
(kVAr)

131.41

66.52

197.94

2428.40

1707.23

126.13
125.67

13.03
6.26

139.16
131.92

1077.94
1609.24

350.22
483.33

124.68

6.18

130.86

1585.16

479.56

131.41

66.52

197.94

1120.26
774.97

125.58

5.95

131.53

131.41

66.52

125.28

2.58

S (kV A)

Fig. 10. Voltage prole of 33-bus test system with and without DG.

% Loss
reduction

200

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

It is evident from Table 4, that we get two more single DG at

buses 14 and 24 in rst iteration. The corresponding DG size and
loss saving are 1133.41 kV A, 1680.25 kV A and 6.79 kW, 2.98 kW
at buses 14 and 24 respectively. It is also seen from Table 2 that
after 2.98 kW no further loss saving could be obtain in rst iteration. Thus the sequence of buses to be redeemed for this system
is 6, 14 and 24. The total power losses PL claim in case of aforementioned single DG placement diminished from 197.94 kW to
131.92 kW with the percentage loss reduction of 33.35%. When
only rst two buses (6 and 14) are to be redeemed in the second
iteration using technique reported in Section Loss minimization
by multiple DG placement; it contribute a total loss saving of
58.45 kW. When the original base system is redeemed with above
DGs, the load ow result indicated that the total power losses PL
reduced by 197.94131.53 kW with the percentage loss reduction
of 33.55%.
Further, when all three buses (6, 14 and 24) are treated for DG
placement it was detected that a total 3427.36 kV A of DGs
(1435.35 kV A at bus 6, 803.55 kV A at bus 14, and 1188.46 kV A
at bus 24) add a loss saving of 62.94 kW in third iteration. The
associated power losses PL by placement of all three DGs are shortened by 197.94127.86 kW with the percentage loss reduction of
35.40%. The respective loss reduction associated with the active
and reactive component of branch currents can also be observed
from Table 4. The enhancement of voltage prole can also be

examine as depicted in Fig. 10, with and without DG in single DG

case. In the same sequence, the voltage deviation of various buses
for this test system can be examine in Table 5. Moreover, it is realized from this table that the voltage at various buses manage
within the acceptable constraints limits. Table 6 display the comparison of results achieve by dissimilar methods for allocation of
single DG to lessen the real power loss in 33-bus system. It could
be realized from this table that extra active power loss saving by
the proposed method is in close compromise with the other methods as bring out in the literature.
Conclusion
In this paper a simple and novel analytical method for minimizing the loss associated with the active and reactive components of
DG branch current is proposed. The proposed method is tested on
two IEEE 15 and 33 distribution networks. In both the test system
it was demonstrated that on which optimal bus, the optimal number of sizes of DG is to be placed so that the maximum loss saving
could achieve. In the 15-bus system it was found that by placing
the optimal DGs at buses 3, 6 and 7 the total power losses can be
reduced from 61.79 kW to 35.63 kW (whereas, PLa and PLr
29.00 kW and 6.63 kW respectively due to Idga and Idgr). In the
33-bus system it was found that by placing optimal DGs at buses
6, 14 and 24 the total power losses associated can be reduced from
197.94 kW to 127.86 kW (whereas, PLa and PLr 125.28 kW and
2.58 kW respectively due to Idga and Idgr). It may also conclude that
using proposed method, the voltage prole is also improved. Further, the proposed method compared with those other methods
reported in literature to show the credibility. Additionally, since
the proposed method uses the SCA, which requires less number
of iteration (as only 3 iteration used for each system) to achieve
the convergence and hence demands less computation time.
Another benet of the suggested method is that the total number
of load ow used normally remains unchanged for larger system
such as 69 or 118 bus system. Besides that, the proposed method
not applicable for unbalanced and meshed distribution system.
Acknowledgements
This work was supported by Grant of Ministry of Human
Resources and Development (MHRD), Government of India and
Indian Institute of Technology Roorkee. The authors would like to
thank the editor and anonymous reviewers whose comments and
suggestions have rened the quality of this paper.

Fig. A.1. Single-line diagram of 15-bus test system with DG.

Appendix

Table 5
Voltage before and after DG at 33-bus system.
System

Voltage at bus before DG

Min
Max

Voltage at bus after DG

Min
Max

33-bus

0.9141 at 18

0.9183 at 33

1.000 at 1

0.9971 at 19

Consider a 15-bus radial system as in Fig. A.1 [30], if four DGs

(i.e. k = 4) are to be placed at buses 7, 10, 12 and 15, then the DG
branch set aj and the transpose of matrix D can be written as DT.
For this set of DG buses the matrix DT can be evaluated as given
below:

k 4; number of DGs buses

a1at DG bus 7 1; 7; 8
Table 6
Comparison of results for 33-bus test system.

a2at DG bus 10 1; 5; 6

Particulars

Gozel and
Hocaoglu [19]

Murthy and
Kumar [31]

Proposed
method

Optimal bus
DG size (MV A)
% Loss reduction
CPU time (s)

6
2.988
31.599
0.043

6
2.536
34.868
0.053

6
2.968
35.405
0.038

a3at DG bus 12 1; 2; 10; 11

a4at DG bus 15 1; 2; 3; 14
For these set of branches the formulation of matrix DT is as
following

R. Viral, D.K. Khatod / Electrical Power and Energy Systems 67 (2015) 191201

T
D

0
0

0 0 0 0 1 1 0 0 0 0 0
0 0 1 1 0 0 0 0 0 0 0

1 0 0 0

0 0 1 1 0 0

1 1 0 0

0 0 0

0 0 0

kN1

In matrix D rows and columns represent the number of branches

and buses of DG respectively. In above case DT is the (k  N  1 or
4  14) matrix. Similarly, the Dij can be evaluated for n number of
DG.
References
[1] Hung DQ, Mithulananthan N, Bansal RC. Analytical expressions for DG
allocation in primary distribution networks. IEEE Trans Energy Convers
2010;25(3):81420.
[2] Baran ME, Wu FF. Network reconguration in distribution systems for loss
reduction and load balancing. IEEE Trans Power Del 1989;4(2):14017.
[3] Chis M, Salama MMA, Jayaram S. Capacitor placement in distribution systems
using heuristic search strategies. IEE Proc Gener Transm Distrib 1997;144(3):
22530.
[4] Haque MH. Capacitor placement in radial distribution systems for loss
reduction. IEE Proc Gener Transm Distrib 1999;146(5):5015.
[5] Haque MH. Efcient load ow method for distribution systems with radial or
mesh conguration. IEE Proc Gener Transm Distrib 1996;143(l):338.
[6] Rao RS, Ravindra K, Satish K, Narasimham SVL. Power loss minimization in
distribution system using network reconguration in the presence of
distributed generation. IEEE Trans Power Syst 2013;28(1):31725.
[7] Ackermann T, Andersson G, Soder L. Distributed generation: a denition. Electr
Power Syst Res 2001;57(3):195204.
[8] Viral R, Khatod DK. Optimal planning of distributed generation systems in
distribution system: a review. Renew Sustain Energy Rev 2012;16(7):514665.
[9] Khatod DK, Pant V, Sharma J. Evolutionary programming based optimal
placement of renewable distributed generators. IEEE Trans Power Syst
2013;28(2):68395.
[10] Hung DQ, Mithulananthan N. Multiple distributed generator placement in
primary distribution networks for loss reduction. IEEE Trans Ind Electron
2013;60(4):17008.
[11] Hung DQ, Mithulananthan N, Bansal RC. Analytical strategies for renewable
distributed generation integration considering energy loss minimization. Appl
Energy 2013;105(April):7585.
[12] Al Abri RS, El-Saadany EF, Atwa YM. Optimal placement and sizing method to
improve the voltage stability margin in a distribution system using distributed
generation. IEEE Trans Power Syst 2013;28(1):32634.
[13] Esmaili M. Placement of minimum distributed generation units observing
power losses and voltage stability with network constraints. IET Gener Transm
Distrib 2013;7(8):81321.
[14] Singh D, Misra RK, Singh D. Effect of load models in distributed generation
planning. IEEE Trans Power Syst 2007;22(4):220412.

201

[15] Acharya N, Mahat P, Mithulananthan N. An analytical approach for DG

allocation in primary distribution network. Int J Electr Power Energy Syst
2006;28(10):66978.
[16] Wang C, Nehrir MH. Analytical approaches for optimal placement of
distributed generation sources in power systems. IEEE Trans Power Syst
2004;19(4):206876.
[17] Atwa YM, El-Saadany EF, Salama MMA, Seethapathy R. Optimal renewable
resources mix for distribution system energy loss minimization. IEEE Trans
Power Syst 2010;25(1):36070.
[18] Atwa YM, El-Saadany EF. Probabilistic approach for optimal allocation of windbased distributed generation in distribution systems. IET Renew Power Gener
2011;5(January):7988.
[19] Gozel T, Hocaoglu MH. An analytical method for the sizing and siting of
distributed generators in radial systems. Electr Power Syst Res 2009;79(June):
9128.
[20] Ochoa LF, Harrison GP. Minimizing energy losses: optimal accommodation and
smart operation of renewable distributed generation. IEEE Trans Power Syst
2011;26(1):198205.
[21] Kim JO, Nam SW, Park SK, Singh C. Dispersed generation planning using
improved Hereford ranch algorithm. Electr Power Syst Res 1998;47(October):
4755.
[22] Hedayati H, Nabaviniaki SA, Akbarimajd A. A method for placement of DG
units in distribution networks. IEEE Trans Power Del 2008;3(July):16208.
[23] Lee SH, Park JW. Selection of optimal location and size of multiple distributed
generations by using Kalman lter algorithm. IEEE Trans Power Syst 2009;
24(3):1393400.
[24] El-Zonkoly AM. Optimal placement of multi-distributed generation units
including different load models using particle swarm optimisation. IET Gener
Transm Distrib 2011;5(7):76071.
[25] Hung DQ, Mithulananthan N, Bansal RC. Integration of PV and BES units in
commercial distribution systems considering energy loss and voltage stability.
Appl Energy 2014;113(January):116270.
[26] Celli G, Ghiani E, Mocci S, Pilo F. A multiobjective evolutionary algorithm for
the sizing and siting of distributed generation. IEEE Trans Power Syst 2005;
20(2):7507.
[27] El-Khattam W, Bhattacharya K, Hegazy Y, Salama MMA. Optimal investment
planning for distributed generation in a competitive electricity market. IEEE
Trans Power Syst 2004;19(3):167484.
[28] Akorede MF, Hizam H, Aris I, Ab Kadir MZA. Effective method for optimal
allocation of distributed generation units in meshed electric power systems.
IET Gener Transm Distrib 2011;5(2):27687.
[29] Mekhamer SF, EI-Hawary ME, Soliman SA, Moustafa MA, Mansour MM. New
heuristic strategies for reactive power compensation of radial distribution
feeders. IEEE Trans Power Del 2002;17(4):112835.
[30] Das D, Kothari DP, Kalam A. Simple and efcient method for load ow solution
of radial distribution systems. Electr Power Energy Syst 1995;17(5):33546.
[31] Murthy VVSN, Kumar Ashwani. Comparison of optimal DG allocation methods
in radial distribution systems based on sensitivity approaches. Int J Electr
Power Energy Syst 2013;53(10):45067.