Comparison of optimal DG allocation methods in radial distribution

systems based on sensitivity approaches

V.V.S.N. Murthy, Ashwani Kumar

Department of Electrical Engineering, National Institute of Technology, Kurukshetra, India

a r t i c l e i n f o
Article history:
Received 3 February 2013
Received in revised form 27 April 2013
Accepted 2 May 2013
Keywords:
Distributed generation
Radial distribution system
Load ow
Optimal locations
Optimal sizes
Novel method
a b s t r a c t
Integration of renewable energy based distributed generation (DG) units provides potential benets to
conventional distribution systems. The power injections from renewable DG units located close to the
load centers provide an opportunity for system voltage support, reduction in energy losses, and reliability
improvement. Therefore, the location of DG units should be carefully determined with the consideration
of different planning incentives. This paper presents a comparison of novel, combined loss sensitivity,
index vector, and voltage sensitivity index methods for optimal location and sizing of distributed gener-
ation (DG) in a distribution network. The main contribution of the paper is: (i) location of DGs based on
existing sensitivity methods, (ii) proposing combined power loss sensitivity based method for DG loca-
tion, (iii) modied Novel method for DG location, (iv) comparison of sensitivity methods for DG location
and their size calculations, and (v) cost of losses and determining cost of power obtained from DGs and
the comparison of methods at unity and lagging power factors. The results show the importance of
installing the suitable size of DG at the suitable location. The results are obtained with all sensitivity
based methods on the IEEE 33-bus and 69-bus systems.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
Distributed power generation (DG) is electricity production that
is on-site and refers to small generating units installed near local
loads or load centers to avoid the need of the network expansions
in order to cover new load areas or to support the increased energy
transfer which would be necessary for satisfying consumers
demand. DG can be an alternative for residential, commercial,
and industrial applications. However, distributed generation can
be dened in a variety of ways as reported in the literature. The
Electric Power Research Institute (EPRI) denes distributed gener-
ation as generation from a few kilowatts up to 50 MW [1]. Interna-
tional Energy Agency (IEA) denes distributed generation as
generating plant serving a customer on-site or providing support
to a distribution network, connected to the grid at distributed level
voltages. The International Conference on large High Voltage Elec-
tric Systems (CIGRE) denes DG as smaller than 50100 MW [1].
The share of DGs in power system is increasing worldwide and
their contribution in the future power system is expected to be
even more [2]. There are many reasons behind the increasingly
widespread use of DG deferring the Transmission and Distribution
(T&D) costs, good efciencies especially in cogeneration and in
combined cycles, creating opportunities for new utilities in the
power generation sector, and provides a exible way to choose a
wide range of combinations of cost and reliability. DG impacts dif-
ferent parameters of a power system, comprising voltage prole,
line losses, and short circuit current, amount of injected harmonic,
and system reliability and stability. The parameters have to be
appropriately investigated prior to installation of DG units. The
problem of allocating DG units to optimal places and also their siz-
ing is of higher priority amongst all issues. However, installation of
DG units in non-optimal places may results in an increase in sys-
tem losses and a bad effect on voltage prole and other parameters
which may lead to a growth of costs, and consequently an opposite
effect on what is expected. Therefore, DG should be allocated in an
optimal way to maximize the system efciency. Studies also show
that if the DG units are connected at non-optimal locations or have
non-optimal sizes, the system losses may increase. To analyze the
distributed energy resources (DER) impacts, different types of gen-
erator groups can be considered in [3]. Many authors have pro-
posed sensitivity based approaches and optimization based
methods for optimal location and sizing of DGs in distribution
systems.
The impact of DG on radial distribution network is explained
i.e., voltage support, loss reduction, and distribution capacity
release and power quality issues in [4]. In this referenced paper,
a new method based on sensitivity indices derived from voltage
stability improvement with respect to changes in injected active
and reactive power at a bus for determining the suitable location
0142-0615/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijepes.2013.05.018

Corresponding author. Tel.: +91 9416366091; fax: +91 1744 238050.

E-mail address: ashwa_ks@yahoo.co.in (A. Kumar).
Electrical Power and Energy Systems 53 (2013) 450467
Contents lists available at SciVerse ScienceDirect
Electrical Power and Energy Systems
j our nal homepage: www. el sevi er . com/ l ocat e/ i j epes
for embedded generator is proposed. Evolutionary programming
optimization technique has been developed in order to determine
the optimal size of the embedded generation [5]. Two new ap-
proaches based on sensitivity of real and reactive power losses
with respect to size and operating pint of DG has been proposed
to determine the most suitable DG size and location towards min-
imizing power losses in the distribution systems. The proposed
techniques have been developed considering the constant imped-
ance and constant current load characteristics into account. The
developed methods have been tested on a practical long radial sys-
tem [6]. In [7] a new method has been suggested based on nodal
pricing for optimally allocating DG in radial distribution system.
An analytical expression based on real power loss sensitivity to cal-
culate optimal DG size and optimal location of DG minimizing
power losses in a distribution network was proposed in [8]. A sim-
ple and effective cumulative performance index, utilizing voltage
prole improvement, loss reduction, and voltage stability index
improvement is considered. Loss sensitivity factor, based on equiv-
alent current injection method for sizing and sitting of DG in radial
distribution systemis given in [9]. Calculation of cost of DG is given
in [10] based on conventional, triangular, and complex power limit.
Authors in [11] described a technique for selection of buses in a sub
transmission system for location of distributed generation (DG)
and determination of their optimum capacities by minimizing
transmission losses using incremental voltage (dV/dP) sensitivities.
Reference [12] presented two new methodologies for optimal
placement of distributed generation sources using an optimal
power ow (OPF) based model in real time wholesale electricity
market. The problem of optimal placement, including size, is for-
mulated for two different objectives, namely, social welfare maxi-
mization and prot maximization. The candidate locations for DG
placement are identied on the basis of locational marginal price
(LMP). Optimal sizing and sitting decisions for DG capacity
planning using heuristic approach was proposed in [13]. In [14]
describes a Novel methodology to calculate optimal DG sizes based
on real power loss. This method gives optimal DG sizes at unity
power factor and it requires less computation when compared
with [8].
Authors presented a methodology for optimal distributed gen-
eration (DG) allocation and sizing in distribution systems consider-
ing the losses minimization, and to guarantee acceptable reliability
level and voltage prole. The GA based optimization technique has
been utilized to obtain the results. The results for voltage prole
and losses have been obtained based on the load ow [15]. Authors
in [16] presented a method for optimal sitting and sizing of multi-
ple distributed generators (DGs) using particle swarm optimization
(PSO) based approach. A new methodology using Fuzzy and Arti-
cial Immune System (AIS) for the placement of Distributed Gener-
ators (DGs) in a radial distribution system to reduce the real power
losses and to improve the voltage prole [17]. Paper [18] deals
with impact of voltage dependent load models on the predicted
energy losses in DG planning. A multi-objective optimization
approach considering losses reduction and voltage prole
improvement for DG allocation using GA was proposed in [19]. A
mixed-integer linear programming approach to determine optimal
size and allocation of distributed generators (DGs) in radial distri-
bution systems is presented in [20]. The proposed formulation
accounts for the steady-state operation of the radial distribution
system, considering different load levels, different types of DGs
with their capability curves, the short-circuit current capacity of
the circuits, and different topologies of the radial distribution
system.
A multi-objective optimization approach using evolutionary
algorithm with an objective of minimizing cost of energy losses,
network upgrading and service interruptions for sizing and sitting
of DG in distribution systems has been presented in [21]. A simple
method for optimal placement of DG in radial distribution system
to minimize real power loss, voltage prole improvement, substa-
tion capacity release and is based on voltage sensitivity index (VSI)
analysis is presented in [22]. Ref. [23] presented an optimal pro-
posed approach (OPA) to determine the optimal sitting and sizing
of DG with multi system constraints to achieve a single or multi-
objectives using genetic algorithm (GA). It deals with the benets
(voltage prole improvement, spinning reserve increasing, power
ow reduction and total line loss reduction) obtained with optimal
DG installation. Authors in [24] presented a simple method for
investigating the problem of optimal location and capacity of DG
in three-phase unbalanced radial distribution systems (URDS) for
power loss minimization and to improve the voltage prole of
the system using voltage index (VSI) analysis. Loss sensitivity
factors (LSFs) are used to select the candidate locations for the
multiple DG placements and Simulated Annealing (SA) is used to
estimate the optimal size of DGs at the optimal locations in [25].
A new method for optimal sizing and sitting of DG in radial
distribution systems was proposed in [26]. In this, optimal location
for DG obtained by power loss sensitivity and optimal size is given
by Harmony Search Algorithm (HSA). Power loss minimization in
radial distribution system with network reconguration and dis-
tributed generation is presented in [27]. In [28] optimal placement
of DG is given based on loss sensitivity and voltage stability index.
A simple conventional iterative search technique along with New-
ton Raphson method of load ow study is implemented for DG siz-
ing and location with an objective to lower down both cost and loss
very effectively. The paper also focuses on optimization of weight-
ing factor, which balances the cost and the loss factors [29]. The
study is carried out for time invariant as well as time invariant
loads incorporating single and multiple DG. A sensitivity analysis
to locate DG at the respective buses and GA to obtain optimal sizes
has been presented in [30]. A methodology for obtaining the opti-
mal sizes of DG adopting two nested calculation stages taking into
account time dependent generation and load is presented in [31]. A
novel load ow method is proposed in [34] for RDSs which in-
cludes identifying the nodes and branches beyond any node using
a sparse technique [32,33]. In [35] a new load ow method is pro-
posed which takes into account of voltage dependent load models,
and line charging capacitance. The method is based on the forward
and backward voltage updating by using polynomial voltage equa-
tion for each branch and backward ladder equation (Kirchhoffs
Laws). Many methods have been proposed by authors for DG loca-
tion based on different approaches viz. sensitivity and optimization
based. The sensitivity based methods are based on losses and volt-
age variation. It is essential to compare the sensitivity based meth-
ods for DG allocation and to observe the impact of DGs on cost
savings and other parameters viz. losses and voltage prole.
In this work, the comparison of existing sensitivity based meth-
ods for optimal allocation of DG in radial distribution system has
been presented. A method based on combined loss sensitivity
and index vector methods are proposed for optimal location of
DGs and comparison of the results with existing novel and voltage
sensitivity index methods are presented. The Novel method was
proposed for unity power factor only, however, the Novel method
has been modied at lagging power factor to obtain the location
and size of DGs as DGs can supply reactive power also for better
voltage prole meeting real power demand also. In this paper,
results have been obtained for DGs operating at unity power factor
injecting only real power into the system and other operating at
0.9 power factor (lag), providing reactive power injection into the
system. Optimal sizes of DG at unity and 0.9 power factor (lag)
are calculated. The cost of loss savings and cost of power supplied
from DGs are also calculated and comparison has been provided.
The results have been obtained on 33-bus [36] and 69-bus [37]
systems.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 451
2. Optimal location of DGs using Novel method
Novel method [17] was proposed for DG allocation at unity
power factor based on real power loss minimization. This method
does not require the calculation of Y
bus
and Z
bus
as in case of ana-
lytical method. The formulae were derived for unity power factor
and real loss minimization. However, DG can also supply reactive
power in a system and therefore it is essential to study the impact
of reactive loss along with real power loss in the system. In this
paper, we have included the reactive power loss also along with
real power loss for calculating optimal DG sizes by considering
both real and reactive power loss sensitivities into account. This
method where both the real and reactive power losses have been
included is called as Modied Novel Method. Optimal DG sizes
at unity and lagging power factors at each bus are obtained and
compared the performance of DGs allocation in the system. By
placing these DG sizes at each bus except source node and run
the load ow to plot the total real power loss variation with DG
size. Then select the node at which loss saving is maximum and
corresponding DG size is the optimal DG size.
TLP = total active power loss.
TLQ = total reactive power loss.
I
i
= branch current.
R
i
= branch resistance.
I
ai
= active component of branch current.
I
ri
= reactive component of branch current.
TLP
a
= loss associated with active component of branch current.
TLP
r
= loss associated with reactive component of branch
current.
TLP

br
i1
I
2
i
R
i
1
TLP

br
i1
I
2
a
R
i

br
i1
I
2
r
R
i
2
TLP TLP
a
TLP
r
3
2.1. DG at unity power factor placed at bus k (Novel method)
I
adgk
= active component of current supplied by DG at node k
TLP

k
i1
I
ai
I
adgk

2
R
i

br
ik1
I
2
ai
R
i

br
i1
I
2
r
R
i
4
Subtract Eq. (2) from Eq. (4)
DTLP

br
i1
I
2
a
R
i

br
i1
I
2
r
R
i

k
i1
I
ai
I
adgk

2
R
i

br
ik1
I
2
ai
R
i

br
i1
I
2
r
R
i

br
i1
I
2
a
R
i

k
i1
I
ai
I
adgk

2
R
i
5
@DTLP
@I
adgk
0 or 2I
adgk

k
i1
R
i
2

k
i1
I
ai
R
i
0 6
For maximum loss saving required current to be supplied by DG is
given by
I
adgk

k
i1
I
ai
R
i

k
i1
R
i
7
V
dgk
= is the voltage magnitude of DG at node k.
Optimal size of DG at unity power factor is given as
P
dgk
I
adgk
V
dgk
8
2.2. DG at lagging power factor placed at bus k (proposed modied
Novel method)
TLP

k
i1
I
ai
I
adgkp

2
R
i

br
ik1
I
2
ai
R
i

k
i1
I
ri
I
rdgkp

2
R
i

br
ik1
I
2
r
R
i
9
Subtract Eq. (2) from Eq. (6)
DTLP

br
i1
I
2
a
R
i

br
i1
I
2
r
R
i

k
i1
I
ai
I
adgkp

2
R
i

br
ik1
I
2
ai
R
i

k
i1
I
ri
I
rdgkp

2
R
i

br
ik1
I
2
r
R
i
10
@DTLP
@I
adgk
p
0 11
and
@DTLP
@I
rdgk
p
0 12
From Eq. (10)
I
adgkp

k
i1
I
ai
R
i

k
i1
R
i
13
I
rdgkp

k
i1
I
ri
R
i

k
i1
R
i
14
TLQ

br
i1
I
2
i
X
i
15
TLQ TLQ
a
TLQ
r
; TLQ

br
i1
I
2
a
X
i

br
i1
I
2
r
X
i
16
TLQ

k
i1
I
ai
I
adgkq

2
X
i

br
ik1
I
2
ai
X
i

k
i1
I
ri
I
rdgkq

2
X
i

br
ik1
I
2
r
X
i
17
Eq. (13) can be written as:
DTLQ

br
i1
I
2
a
X
i

br
i1
I
2
r
X
i

k
i1
I
ai
I
adgkq

2
X
i

br
ik1
I
2
ai
X
i

k
i1
I
ri
I
rdgkq

2
X
i

br
ik1
I
2
r
X
i
18
From Eq. (14), the equations can be derived for maximum loss sav-
ings as:
@DTLQ
@I
adgk
q
0 19
@DTLQ
@I
rdgkq
0 20
Solving Eqs. (15) and (16), we get components of currents as:
I
adgkq

k
i1
I
ai
X
i

k
i1
R
i
21
I
rdgkq

k
i1
I
ri
X
i

k
i1
R
i
22
I
adgk
= active component of the current to be supplied by DG for
maximum loss saving at node k.
452 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
I
rdgk
= reactive component of the current to be supplied by DG
for maximum loss saving at node k.
I
adgk

I
2
adgkp
I
2
adgkq
_
23
I
rdgk

I
2
rdgkp
I
2
rdgkq
_
24
P
dgk
= optimal real power supplied by DG at power factor cos/
at bus k.
Q
dgk
= optimal reactive power supplied by DG at power factor
cos/ at bus k.
S
dgk
= Optimal complex power supplied by DG at bus k.
P
dgk
I
adgk
V
dgk
cos / 25
Q
dgk
I
rdgk
V
dgk
sin/ 26
S
dgk
I
adgk
V
dgk
27
Using Eqs. (5), (21), and (22), optimal DG sizes can be obtained for
unity and lagging power factor respectively (Figs. 1, 4 and 7).
3. Optimal Location of DG based on proposed combined power
loss sensitivity method
Real power loss sensitivity factors [22,25] are calculated for
determining the candidate nodes for placement of DGs. Estimation
of these sensitive nodes helps in reducing the search space. Instal-
lation of DGs not only impacts the real power loss however, as DGs
can supply reactive power and have considerable impact on reac-
tive power losses. Therefore, combined loss sensitivity is calculated
based on both real and reactive power loss.
@Ploss
@Q
2

2 Q
2Rj
V
2
2
28
@Qloss
@Q
2

2 Q
2Xj
V
2
2
29
Combined loss sensitivity with respect to reactive power

@Sloss
@Q
2

@ploss
@Q
2
j
@Qloss
@Q
2
30
Combined loss sensitivity with respect to real power

@Sloss
@P
2

@ploss
@P
2
j
@Qloss
@P
2
31
Loss sensitivity matrix
@ploss
@P
2
@Qloss
@P
2
@ploss
@Q2
@Qloss
@Q
2

32
Loss sensitivity factors are calculated from load ow analysis and
values are arranged in descending order for all the lines. Normalized
voltage magnitudes are calculated for all the buses by using the fol-
lowing formula,
Normalized; Vi Vi=0:95 33
Buses, whose normalized values are less than 1.01 are considered as
candidate nodes for DG location. For 33 bus system based on Com-
bined Power Loss Sensitivity prole (Fig. 7) it is found that loss sen-
sitivity is maximum at buses 24 and 25 but at this bus normalized
voltage (Fig. 8) is greater than 1.01 so it is not selected as candidate
node for DG allocation. Then next most sensitive bus is 8 and at this
bus normalized voltage is less than 1.01 so it is selected as candi-
date node for DG location. For 69 bus system, based on Combined
Power Loss Sensitivity prole (Fig. 11), it is found that loss sensitiv-
ity is maximum at bus 61 (Fig. 12) and at this bus normalized volt-
age is also less than 1.01 so it is selected as candidate node for DG
location.
4. Optimal location of DG based on index vector Method
Index Vector method has been utilized for optimal allocation of
capacitor in radial distribution system [21]. Index Vector is formu-
lated by running the base case load ow on a given radial distribu-
tion network, and calculating reactive component of current in the
branches and reactive power load concentration at each node [21].
In this work, the index vector method has been used for optimal
DG allocation problem. Based on the elements of the Index Vector,
this method identies a sequence of nodes to be connected with
DG. The Index-Vector for bus n is given by:
Indexn
1
Vn
2

Iqk
Ipk

Qeff n
total Q
34
Index[n] = Index for nth bus.
V[n] = voltage at nth bus.
Iq[k] = imaginary component of current in kth branch.
Ip[k] = real component of current in kth branch.
Qeff[n] = effective load at nth bus.
Total Q = total reactive load of the given distribution system.
Thus, the potential locations of DG are obtained directly. Ar-
range the Index vector in descending order so that highest priority
bus will come rst and the lowest priority bus will come at the end.
Normalized voltage magnitudes are calculated for all the buses by
the following formula: Normalized, V(i) = V(i)/0.95. Buses, whose
normalized values are less than 1.01 are considered as candidate
nodes for DG location. Now place the DG at the rst potential loca-
tion and vary the DG size in steps and run the load ow to compute
the losses. The DG size at which losses is minimum is selected as
Fig. 1. Optimal DG sizes for 33 bus system using Novel method.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 453
the optimal size. From index vector prole (Figs. 15 and 16) and
normalized voltage plots (Figs. 19 and 20), it is found that the opti-
mal location of DG is at bus 30 and bus 61 for 33, 69-bus systems
respectively since at these buses, index vector value is maximum
and normalized voltage is below 1.01. Optimal DG sizes (Figs. 17
and 21) are found to be 1550, 1850 kVA at unity power factor
and 1950, 2200 kVA at 0.9 power factor lag for 33, 69 bus systems
respectively.
5. Optimal location based on voltage sensitivity index
The main objective of voltage sensitivity index (VSI) is to nd
the distance from current operating point to the marginally stable
point. The purpose of nding VSI is to nd most sensitive node of
the system from voltage sensitivity point of view. Voltage sensitiv-
ity index [18,20] is numerical solution which helps operator to
monitor how close the system is to collapse or to initiate automatic
remedial action schemes to prevent voltage collapse. Voltage col-
lapse starts at the most sensitive node and then spread out to other
sensitive nodes. The voltage sensitivity index (VSI) of all nodes of
distribution network is computed at unity power factor [20]. Nodes
having minimum voltage sensitivity index have been selected rst
and then the node having the minimum value of voltage stability
index is declared as most sensitive node. In the present work, we
have considered the impact of lagging power factor for computa-
tion of VSI and accordingly the location of DGs has been identied.
0
1000
2000
3000
4000
5000
6000
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
D
G

s
i
z
e

(
k
V
A
)
Bus Number
AT 0.9 PF LAG AT UNITY PF
Fig. 4. Optimal DG sizes for 69 bus system using Novel method.
Fig. 2. Total Real Power Loss variation with DG size using Novel method.
Fig. 3. Voltage prole variation with DG size using Novel method.
454 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
The voltage sensitive nodes are rst identied by penetrating
DG with 25% of the total feeder loading capacity at each node
at a time and then, calculating the voltage sensitivity index
(VSI). When DG is connected at bus i, VSI for bus i is dened
as:
VSI
i

n
k1
1 V
k

2
n

35
where V
k
is voltage at kth node and n is the number of nodes.
0
0.001
0.002
0.003
0.004
0.005
0.006
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 28 29 30 31 32 33
C
o
m
b
i
n
e
d

P
o
w
e
r

L
o
s
s

S
e
n
s
i
t
i
v
i
t
y
Bus Number
Fig. 7. Combined power loss sensitivity prole for 33 bus system.
0
50
100
150
200
250
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
T
o
t
a
l

R
e
a
l

P
o
w
e
r

l
o
s
s

(
k
W
)
Bus Number
DG AT 0.9 PF LAG DG AT UNITY PF
Fig. 5. Total Real Power Loss variation with DG size.
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1832.536 KVA AT UNITY PF
WITH DGOF 2236.704KVA AT 0.9 PF LAG
Fig. 6. Voltage prole for 69 bus system using Novel method.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 455
The node with least VSI will be picked as the best location for
the DG placement.
To determine the optimal size of DG, the following steps are
taken:
(1) First, the DG is placed at the node with least VSI.
(2) DG size is varied from a minimum value to a value equal to
feeder loading capacity in constant steps until the minimum
system losses is found.
(3) The DG size which results in minimum losses is taken as
optimal.
Bus-16, 65 are most sensitive buses for 33, 69 bus systems
respectively since at these buses VSI (Fig. 23 and 26) value ob-
tained is lowest. Optimal DG sizes are found to be 1000,
1450 kVA at unity power factor and 1200, 1750 kVA at 0.9 pf lag
(Figs. 24 and 27) since total loss obtained is minimum by placing
DG at these buses.
0.9
0.92
0.94
0.96
0.98
1
1.02
1.04
1.06
1.08
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 28 29 30 31 32 33
N
o
r
m
a
l
i
s
e
d

V
o
l
t
a
g
e


(
p
.
u
.
)
Bus Number
Fig. 8. Normalized voltage prole for 33 bus system.
0
50
100
150
200
250
T
o
t
a
l

R
e
a
l

P
o
w
e
r

L
o
s
s

(
k
W
)
DG Size (kVA)
WITH DG AT UNITY PF WITH DG AT 0.9 PF LAG
Fig. 9. Total real power loss variation with DG size using combine power loss sensitivity method.
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
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 28 29 30 31 32 33
V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1800KVA AT UNITY PF
WITH DG OF 2100KVA AT 0.9 PF LAG
Fig. 10. Voltage prole for 33 bus system using combined power loss sensitivity method.
456 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
6. Results and discussion
6.1. Results for 33 bus test system using Novel method and modied
Novel method
Results have been obtained for IEEE-33 and 69-bus test systems
with Novel, modied Novel, VSI based, index vector, and combined
power loss sensitivity based methods. The base MVA and base kV
have been taken as: MVA
Bass
100 MVA kV
base
12:66 kV and
results have been obtained at unity and 0.9 pf(lag).
6.2. Results for 33 bus test system using Novel method and modied
Novel method
For 33-bus system without installation of DG, real and reactive
power losses are 210.9761 kW and 143.0171 kVAR respectively.
Real and reactive power from the substation is 3925.9761 kW
and 2443.011kVAR respectively. It is found that total loss is mini-
mal at bus 6 as shown in Fig. 2. Therefore, it has been selected as a
candidate node for DG allocation. With installation of DG at unity
pf, real, reactive power losses are 111.1462 kW and 81.6507 kVAR
0
0.002
0.004
0.006
0.008
0.01
0.012
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
C
o
m
b
i
n
e
d

P
o
w
e
r

L
o
s
s

S
e
n
s
i
t
i
v
i
t
y
Bus Number
Fig. 11. Combined power loss sensitivity prole for 69 bus system.
0.9
0.92
0.94
0.96
0.98
1
1.02
1.04
1.06
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
N
o
r
m
a
l
i
s
e
d

V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
Fig. 12. Normalized voltage prole for 69 bus system.
Fig. 13. Total real power loss variation with DG size using combined power loss sensitivity method.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 457
respectively. The real and reactive power fromthe substation is ob-
tained as 1431.1761 kW and 2443.0171 kVAR. It is observed that
with DG, both real and reactive power drawn from the system re-
duces based on the DG output. With DG at 0.9 pf lag, real, reactive
power losses are 70.9072 kW and 56.7162 kVAR respectively. Real
and, reactive powers from the substation become 1215.7761 kW
and 1130.7171 kVAR at lagging power factor. Real power loss var-
iation with DG at unity and lagging power factor and voltage pro-
le is shown in Figs. 2 and 3. It is observed from the gures that the
DG size obtained is higher at lagging power factor compared to the
size obtained at unity power factor, however, the losses are found
lower with DGs at lagging power factor rather than DGs at unity
power factor. This is due the reason of reactive power available lo-
cally for the loads and thereby decrease in the reactive power
available from substation. The voltage prole also improves with
DGs at lagging power factor and it is observed from the Fig. 3.
The minimum voltage obtained for the system is better compared
to the voltage obtained with DGs at unity power factor. Thus, it is
essential to consider the reactive power available from DGs for its
size calculations and its impact on losses reduction and voltage
prole improvement. The results obtained are also given in Table
1 using Novel method and modied Novel method. The results ob-
0
0.5
1
1.5
2
2.5
3
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 28 29 30 31 32 33
I
n
d
e
x

V
e
c
t
o
r
Bus Number
Fig. 15. Index vector prole for 33 bus system.
0.9
0.92
0.94
0.96
0.98
1
1.02
1.04
1.06
1.08
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 28 29 30 31 32 33
N
o
r
m
a
l
i
s
e
d

V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
Fig. 16. Normalized voltage prole for 33 bus system.
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1850 KVA AT UNITY PF
WITH DG OF 2200 KVA AT 0.9 PF LAG
Fig. 14. Voltage prole for 69 bus system using combined power loss sensitivity method.
458 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
tained with consideration of reactive power are better than the re-
sults obtained with DGs at unity power factor.
6.3. Results for 69 bus test system using Novel method and modied
Novel method
The results have been obtained for 69 bus test system with
Novel method as well as modifying Novel method at 0.9 pf (lag)
considering the base MVA and base kV as MVA
Base
100 MVA.
kV
Base
12:66 kV.
For 69 bus system, candidate node is bus 61 as shown in Fig. 5,
since total real power loss is obtained minimum with installation
of DG at this bus. Without DG real, reactive power losses are
224.8799 kW and 102.1091 kVAR respectively. With the installa-
tion of DG at unity pf, the real and reactive power losses are
83.1951 kW and 40.5795 kVAR respectively. With DG at unity pf
real, reactive power from the substation becomes 2193.7439 kW
0
0.5
1
1.5
2
2.5
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
I
n
d
e
x

V
e
c
t
o
r
Bus Number
Fig. 19. Index vector prole for 69 bus system.
Fig. 17. Total real power loss variation with DG size using index vector method.
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
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 28 29 30 31 32 33
V
o
l
t
a
g
e


(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1550KVA AT UNITY PF
WITH DG OF 1950KVA AT 0.9 PF LAG
Fig. 18. Voltage prole for 33 bus system using Index Vector method.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 459
and 2795.7091 kVAR respectively. With DG at 0.9 pf lag real, reac-
tive power losses are 27.3795 kW and 16.1734 kVAR. Real and
reactive power from the substation is 2013.2463 kW and
1847.9534 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 5. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 6. The optimal DG sizes
obtained at optimal bus 61 and other buses are also calculated
and are shown in Fig. 4. It is observed from the gures that the
DG size obtained is higher at lagging power factor compared to
the size obtained at unity power factor, however, the losses are
found lower with DGs at lagging power factor rather than DGs at
unity power factor. This is due the reason of reactive power avail-
able locally for the loads and thereby decrease in the reactive
power available from substation. The voltage prole also improves
with DGs at lagging power factor and it is observed from the Fig. 6.
The minimum voltage obtained for the system is better compared
to the voltage obtained with DGs at unity power factor. Thus, it is
essential to consider the reactive power available from DGs for its
size calculations and its impact on losses reduction and voltage
0.9
0.92
0.94
0.96
0.98
1
1.02
1.04
1.06
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
N
o
r
m
a
l
i
s
e
d

V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
Fig. 20. Normalized voltage prole for 69 bus system.
0
50
100
150
200
250
0

1
0
0
2
0
0
3
0
0
4
0
0
5
0
0
6
0
0
7
0
0
8
0
0
9
0
0
1
0
0
0
1
1
0
0
1
2
0
0
1
3
0
0
1
4
0
0
1
5
0
0
1
6
0
0
1
7
0
0
1
8
0
0
1
9
0
0
2
0
0
0
2
1
0
0
2
2
0
0
2
3
0
0
2
4
0
0
2
5
0
0
T
o
t
a
l

R
e
a
l

P
o
w
e
r

L
o
s
s

(
k
W
)
DG Size (kVA)
DG AT UNITY PF DG AT 0.9 PF LAG
Fig. 21. Total real power loss variation with DG size using index vector method.
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1850 KVA AT UNITY PF
WITH DG OF 2200 KVA AT 0.9 PF LAG
Fig. 22. Voltage prole for 69 bus system using Index Vector method.
460 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
prole improvement. The results obtained are also given in Table 2
using Novel method and modied Novel method. The results ob-
tained with consideration of reactive power are better than the re-
sults obtained with DGs at unity power factor.
6.4. Results for 33 bus test system using proposed combined power loss
sensitivity method
For 33 bus system without installation of DG, the real and reac-
tive power losses are 210.9761 kW and 143.0171 kVAR respec-
tively. The real and the reactive power received from the
substation is 3925.9761 kW and 2443.011 kVAR respectively. Plac-
ing DG at bus 8 and varying the sizes of DG in steps, the variation of
total real power loss with DG size is obtained and is plotted in
Fig. 9. Total real power loss is obtained minimum with DG of
1800 kW and 2100 kVA at unity and 0.9 pf lag respectively. With
installation of DG at bus 8 at unity pf real, reactive power losses
are 118.1293 kW and 82.9355 kVAR respectively. Also real and
reactive power from the substation reduces to 2125.9761 kW and
2443.0171 kVAR. With DG at 0.9 pf lag, real and reactive power
losses are 84.472 kW and 62.085 kVAR respectively. Real and reac-
tive powers received from the substation obtained are
2035.9761 kW and 1527.8371 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 9. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 10. It is observed from
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
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 28 29 30 31 32 33
V
o
l
t
a
g
e

S
e
n
s
i
t
i
v
i
t
y

I
n
d
e
x
Bus Number
Fig. 23. VSI prole for 33 bus system.
0
50
100
150
200
250
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800
T
o
t
a
l

R
e
a
l

P
o
w
e
r

L
o
s
s

(
k
W
)
DG Size (kVA)
DG AT UNITY PF DG AT 0.9 PF LAG
Fig. 24. Total real power loss variation with DG size using VSI method for 33 bus system.
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
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 28 29 30 31 32 33
V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1200KVA AT 0.9 PF LAG
WITH DG OF 1000KVA AT UNITY PF
Fig. 25. voltage prole for 33 bus system using VSI method.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 461
the gures that the DG size obtained is higher at lagging power fac-
tor compared to the size obtained at unity power factor, however,
the losses are found lower with DGs at lagging power factor rather
than DGs at unity power factor. This is due the reason of reactive
power available locally for the loads and thereby decrease in the
reactive power available from substation. The voltage prole also
improves with DGs at lagging power factor and it is observed from
the gure. The minimum voltage obtained for the system is better
compared to the voltage obtained with DGs at unity power factor.
Thus, it is essential to consider the reactive power available from
DGs for its size calculations and its impact on losses reduction
and voltage prole improvement. The results obtained are also gi-
ven in Table 3 using combined power loss sensitivity based meth-
od. The results obtained with consideration of reactive power are
better than the results obtained with DGs at unity power factor.
6.5. Results for 69 bus test system using proposed combined power loss
sensitivity method
For 69 bus system, without DG real, reactive power losses are
224.8799 kW and 102.1091 kVAR. With DG optimal, location at
bus 61 and with different sizes taken in steps, the variation of total
real power loss with DG size is obtained and is shown in Fig. 13.
Minimum real power loss is obtained with DG of 1850 kW and
2200 kVA at unity and 0.9 pf (lag). With installation of DG at bus
61 at unity pf, real and reactive power losses are 83.1568 kW
and 40.5373 kVAR respectively. The real and reactive power re-
ceived from the substation is 2176.2799 kW and 2795.7091 kVAR
respectively. With DG at 0.9 pf lag, the real and reactive power
losses obtained are 27.9091 kW and 16.4639 kVAr. The real and
reactive power received from the substation received is
2046.2799 kW and 1836.7731 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 13. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 14. It is observed from
the gures that the DG size obtained is higher at lagging power fac-
tor compared to the size obtained at unity power factor, however,
the losses are found lower with DGs at lagging power factor rather
than DGs at unity power factor. This is due the reason of reactive
power available locally for the loads and thereby decrease in the
reactive power available from substation. The voltage prole also
improves with DGs at lagging power factor and it is observed from
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68
V
o
l
t
a
g
e

S
e
n
s
i
t
i
v
i
t
y

I
n
d
e
x
Bus Number
Fig. 26. VSI prole for 69 bus system.
0
50
100
150
200
250
0

1
0
0
2
0
0
3
0
0
4
0
0
5
0
0
6
0
0
7
0
0
8
0
0
9
0
0
1
0
0
0
1
1
0
0
1
2
0
0
1
3
0
0
1
4
0
0
1
5
0
0
1
6
0
0
1
7
0
0
1
8
0
0
1
9
0
0
2
0
0
0
2
1
0
0
2
2
0
0
2
3
0
0
2
4
0
0
2
5
0
0
T
o
t
a
l

R
e
a
l

p
o
w
e
r

L
o
s
s

(
k
W
)
Bus Number
DG AT UNITY PF DG AT 0.9 PF LAG
Fig. 27. Total real power loss variation with DG size using VSI method for 69 bus system.
Table 1
Results for 33 bus system using Novel method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 6 6
DG size (kVA) 2494.8 3011.3
Total real power loss (kW) 210.9761 111.1462 70.9072
Total reactive power loss (kVAR) 143.0171 81.6507 56.7162
Minimum bus voltage (p.u.) 0.9040 0.9412 0.9566
Real power from the substation (kW) 3925.9761 1431.1761 1215.7761
Reactive power from the substation (kVAR) 2443.0171 2443.0171 1130.7171
462 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
the gure. The minimum voltage obtained for the system is better
compared to the voltage obtained with DGs at unity power factor.
Thus, it is essential to consider the reactive power available from
DGs for its size calculations and its impact on losses reduction
and voltage prole improvement. The results obtained are also
given in Table 4 using combined power loss sensitivity based
method. The results obtained with consideration of reactive
power are better than the results obtained with DGs at unity power
factor.
6.6. Results for 33 bus test system using Index vector method at unity
and lagging power factor
For 33 bus system without installation of DG real, reactive
power losses are 210.9761 kW and 143.0171 kVAR respectively.
Real, reactive power from the substation is 3925.9761 kW and
2443.011 kVAR. With installation of DG at unity pf real, reactive
power losses are 125.1516 kW and 89.3088 kVAR respectively.
Also real and reactive power from the substation is decreased to
2375.9761 kW and 2443.0171 kVAR. With DG at 0.9 pf lag real,
reactive power losses are 78.4193 kW and 58.9994 kVAR respec-
tively. Real, reactive powers from the substation become
217.9761 kW and 1593.0511 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 17. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 18. It is observed from
the gures that the DG size obtained is higher at lagging power fac-
tor compared to the size obtained at unity power factor, however,
the losses are found lower with DGs at lagging power factor rather
than DGs at unity power factor. This is due the reason of reactive
power available locally for the loads and thereby decrease in the
reactive power available from substation. The voltage prole also
improves with DGs at lagging power factor and it is observed from
the gure. The minimum voltage obtained for the system is better
compared to the voltage obtained with DGs at unity power factor.
Thus, it is essential to consider the reactive power available from
DGs for its size calculations and its impact on losses reduction
and voltage prole improvement. The results obtained are also
given in Table 5 using index vector based method. The results
obtained with consideration of reactive power are better than the
results obtained with DGs at unity power factor.
6.7. Results for 69 bus test systems using Index vector method at unity
and lagging power factor
For 69 bus system, without DG real, reactive power losses are
224.8799 kW and 102.1091 kVAR, with installation of DG at unity
pf real, reactive power losses are 83.1568 kW and 40.5373 kVAR
respectively. Real, reactive power from the substation becomes
216.2799 kW and 795.7091 kVAR respectively. With DG at 0.9 pf
lag real, reactive power losses are 27.9091 kW and 16.4639 kVAR.
Real, reactive power from the substation becomes 2046.2799 kW
and 1836.7731 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 21. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 22. It is observed from
the gures that the DG size obtained is higher at lagging power fac-
tor compared to the size obtained at unity power factor, however,
the losses are found lower with DGs at lagging power factor rather
than DGs at unity power factor. This is due the reason of reactive
power available locally for the loads and thereby decrease in the
reactive power available from substation. The voltage prole also
improves with DGs at lagging power factor and it is observed from
the gure. The minimum voltage obtained for the system is better
compared to the voltage obtained with DGs at unity power factor.
Thus, it is essential to consider the reactive power available from
DGs for its size calculations and its impact on losses reduction
and voltage prole improvement. The results obtained are also
given in Table 6 using combined power loss sensitivity based
method. The results obtained with consideration of reactive power
are better than the results obtained with DGs at unity power factor.
6.8. Results for 33 bus test system using voltage sensitivity method
For 33 bus system without installation of DG, real and reactive
power losses are 210.9761 kW and 143.0171 kVAR respectively.
Real and reactive power received from the substation is
3925.9761 kWand 2443.011 kVAR. With installation of DG at unity
pf, real and reactive power losses obtained are 136.7553 kW and
92.6599 kVAR respectively. Also real and reactive power received
from the substation obtained is 2925.9761 kW and 2443.0171
kVAR respecively. With DG at 0.9 pf lag real, reactive power losses
are 112.7864 kW and 77.449 kVAR respectively. Real and reactive
Table 2
Results for 69 bus system using Novel method and modied Novel method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 61 61
DG size (kVA) 1832.536 2236.704
Total real power loss (kW) 224.8799 83.1951 27.37953
Total reactive power loss (kVAR) 102.1091 40.5795 16.17349
Minimum bus voltage (p.u.) 0.909253 0.9685 0.972844
Real power from the substation (kW) 4026.2799 2193.7439 2013.2463
Reactive power from the substation (kVAR) 2795.7091 2795.7091 1847.9534
Table 3
Results for 33 bus system using combined power loss sensitivity method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 8 8
DG size (kVA) 1800 2100
Total real power loss (kW) 210.9761 118.1293 84.472
Total reactive power loss (kVAR) 143.0171 82.9355 62.085
Minimum bus voltage (p.u.) 0.9040 0.9433 0.9534
Real power from the Substation (kW) 3925.9761 2125.9761 2035.9761
Reactive power from the Substation (kVAR) 2443.0171 2443.0171 1527.8371
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 463
powers received from from the substation become 2845.9761 kW
and 1919.9611 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 24. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 25. It is observed from
the gures that the DG size obtained is higher at lagging power fac-
tor compared to the size obtained at unity power factor, however,
the losses are found lower with DGs at lagging power factor rather
than DGs at unity power factor. This is due the reason of reactive
power available locally for the loads and thereby decrease in the
reactive power available from substation. The voltage prole also
improves with DGs at lagging power factor and it is observed from
the gure. The minimum voltage obtained for the system is better
compared to the voltage obtained with DGs at unity power factor.
Thus, it is essential to consider the reactive power available from
DGs for its size calculations and its impact on losses reduction
and voltage prole improvement. The results obtained are also gi-
ven in Table 7 using voltage sensitivity based method. The results
obtained with consideration of reactive power are better than the
results obtained with DGs at unity power factor.
6.9. Results for 69 bus test system using voltage sensitivity method
For 69 bus system, without DG, the real and reactive power
losses are 224.8799 kW and 102.1091 kVAR respectively. With
the installation of DG at unity pf, real and reactive power losses ob-
tained are 112.0217 kW and 5.1172 kVAR, respectively. The real
and reactive power received from the substation is 2576.2799
kW and 2795.7091 kVAR respectively. With DG at 0.9 pf lag, the
real and reactive power losses obtained are 65.4502 kW and
35.62506 kVAR. The real and reactive power received from the
substation becomes 2457.2799 kW and 2032.9191 kVAR.
Real power loss variation with DG at unity and lagging power
factor is shown in Fig. 27. The voltage prole obtained with unity
and lagging power factor is shown in Fig. 28. It is observed from
the gures that the DG size obtained is higher at lagging power fac-
tor compared to the size obtained at unity power factor, however,
the losses are found lower with DGs at lagging power factor rather
than DGs at unity power factor. This is due the reason of reactive
power available locally for the loads and thereby decrease in the
reactive power available from substation. The voltage prole also
improves with DGs at lagging power factor and it is observed from
the gure. The minimum voltage obtained for the system is better
compared to the voltage obtained with DGs at unity power factor.
Thus, it is essential to consider the reactive power available from
DGs for its size calculations and its impact on losses reduction
and voltage prole improvement. The results obtained are also
given in Table 8 using voltage sensitivity based method. The results
obtained with consideration of reactive power are better than the
results obtained with DGs at unity power factor.
7. Comparison of results
Based on the results obtained with different sensitivity meth-
ods, the comparison of results corresponding to real and reactive
power loss, minimum voltage level, and the cost of energy loss as
well as cost of power obtained from DGs have been provided for
both 33 and 69 bus test systems at unity and 0.9 pf lag. This has
been presented to compare the pros and cons of sensitivity based
approaches for optimal DG location as well as optimal sizes of
DGs and thereby the savings in losses and cost of energy obtained
from DGs. The cost of energy losses and cost component of DG
power has been calculated based on the mathematical model rep-
resented as:
(i) Cost of Energy Losses (CL): The annual cost of energy loss is
given by [27]
CL=(Total Real power Loss)(Kp + KeLsf8760) $ (31).
where
Kp: annual demand cost of power loss ($/kW)
Ke: annual cost of energy loss($/kW h)
Lsf: loss factor
Loss factor is expressed in terms of load factor (Lf) as below
Lsf k Lf 1 k Lf
2
36
The values taken for the coefcients in the loss factor calculation
are:
k = 0.2, Lf = 0.47, Kp = 57.6923 $/kW, Ke = 0.00961538 $/kW h.
(ii) Cost component of DG for real and reactive power
CPdg a Pdg
2
b Pdg c $=MW h 37
Table 4
Results for 69 bus system using Combined Power Loss Sensitivity method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 61 61
DG size (kVA) 1850 2200
Total real power loss (kW) 224.8799 83.1568 27.9091
Total reactive power loss (kVAR) 102.1091 40.5373 16.4639
Minimum bus voltage (p.u.) 0.909253 0.968559 0.9726
Real power from the substation (kW) 4026.2799 2176.2799 2046.2799
Reactive power from the substation (kVAR) 2795.7091 2795.7091 1836.7731
Table 5
Results for 33 bus system using Index Vector method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 30 30
DG size (kVA) 1550 1950
Total real power loss (kW) 210.9761 125.1516 78.4193
Total reactive power loss (kVAR) 143.0171 89.3088 58.9994
Minimum bus voltage (p.u.) 0.9040 0.92752 0.9391
Real Power from the Substation (kW) 3925.9761 2375.9761 2170.9761
Reactive power from the substation (kVAR) 2443.0171 2443.0171 1593.0511
464 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
Cost coefcients are taken as:
a 0 b 20 c 0:25
Cost of reactive power supplied by DG is calculated based on max-
imum complex power supplied by DG as [8]
CQdg CostSgmax Cost

Sgmax
2
Qg
2
_ _ _ _ _
k 38
Sgmax
Pgmax
cos /
39
P max 1:1 Pg, the power factor, cos/ has been taken 1 at unity
power factor and 0.9(lag) at lagging power factor to carry out the
analysis.
k 0:050:1:
In this paper work, the value of factor k is taken as 0.1. The com-
parison of results for both the test systems at unity and lagging
power factor are given in Tables 912.
Based on the results obtained for two test systems at unity and
lagging power factor, it is observed that different locations of DGs
are obtained with different methods. The sizes obtained are differ-
ent and accordingly the losses obtained for the different sensitivity
methods are different with different minimum voltage prole. We
observe that there is much reduction in real, reactive power losses
and improvement in voltage prole with installation of DG at lag-
ging pf than at unity pf due to the reactive support obtained from
DG. Cost of energy losses for IEEE 33 bus test system are obtained
lower with modied Novel method compared to the other meth-
ods due to higher reduction in the real power losses in the system.
For IEEE 69 bus test system, the cost of energy losses are obtained
lower with proposed combined power loss sensitivity method
compared to the other methods. With DG at lagging power factor,
the reduction in the losses are considerable and thereby the cost is
found lower due to the huge reduction in the losses with the reac-
tive power supplied by the DG. For 69 bus system novel, combined
loss sensitivity, and index vector methods are giving almost similar
results. Looking into all aspects of power losses, voltage prole,
cost component for losses, and cost component for power obtained
Table 6
Results for 69 bus system using Index Vector method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 61 61
DG size (kVA) 1850 2200
Total real power loss (kW) 224.8799 83.1568 27.9091
Total reactive power loss (kVAR) 102.1091 40.5373 16.4639
Minimum bus voltage (p.u.) 0.909253 0.968559 0.9726
Real power from the substation (kW) 4026.2799 2176.2799 2046.2799
Reactive power from the substation (kVAR) 2795.7091 2795.7091 1836.7731
Table 7
Results for 33 bus system using VSI method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 16 16
DG size (kVA) 1000 1200
Total real power loss (kW) 210.9761 136.7553 112.7864
Total reactive power loss (kVAR) 143.0171 92.6599 77.449
Minimum bus voltage (p.u.) 0.9040 0.9318 0.9378
Real power from the substation (kW) 3925.9761 2925.9761 2845.9761
Reactive power from the substation (kVAR) 2443.0171 2443.0171 1919.9611
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69
V
o
l
t
a
g
e

(
p
.
u
.
)
Bus Number
WITHOUT DG WITH DG OF 1450 KVA AT UNITY PF
WITH DG OF 1750 KVA AT 0.9 PF LAG
Fig. 28. voltage prole for 69 bus system using VSI method.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 465
Table 8
Results for 69 bus system using VSI method.
Without DG With DG at unity pf With DG at 0.9 pf lag
DG location bus 65 65
DG size (kVA) 1450 1750
Total real power loss (kW) 224.8799 112.0217 65.4502
Total reactive power loss (kVAR) 102.1091 55.1172 35.62506
Minimum bus voltage (p.u.) 0.909253 0.9660621 0.969302
Real power from the substation (kW) 4026.2799 2576.2799 2451.2799
Reactive power from the substation (kVAR) 2795.7091 2795.7091 2032.9191
Table 9
Results for 33 bus system with DG at unity power factor.
Without
DG
With DG
Novel method
[17]
Combined power loss sensitivity
method
Index vector
method
Voltage sensitivity index method
[18]
DG location 6 8 30 16
DG size (kW) 2494.8 1800 1550 1000
Total real power loss (kW) 210.9761 111.1462 118.1293 125.1516 136.7533
Total reactive power loss
(kVAR)
143.0171 81.6507 82.9355 89.3088 92.6599
Minimum bus voltage (p.u.) 0.9040 0.9412 0.9433 0.92752 0.9318
Cost of energy losses ($) 16982.5724 8946.7403 9508.8466 10074.1084 11007.9901
Cost of Pdg ($/MW h) 50.1460 36.2500 31.2500 20.2500
Table 10
Results for 33 bus system with DG at 0.9 pf lag.
With DG
Modied Novel method Combined power loss sensitivity method Index vector method Voltage sensitivity index method [18]
DG location 6 8 30 16
DG size (kVA) 3011.3 2100 1950 1200
Total real power loss (kW) 70.9072 84.472 78.4193 112.7864
Total reactive power loss (kVAR) 56.7162 62.085 58.9994 77.449
Minimum bus voltage (p.u.) 0.9566 0.9534 0.9391 0.9378
Cost of Energy losses ($) 5707.6922 6799.5941 6312.3806 9078.7686
Cost of PDG ($/MW h) 54.4534 38.0500 35.3500 21.8500
Cost of QDG ($/MVAR h) 5.2889 3.6952 3.4328 2.1207
Table 11
Results for 69 bus system with DG at unity power factor.
Without
DG
With DG
Novel method
[17]
Combined power loss sensitivity
method
Index vector
method
Voltage sensitivity index method
[18]
DG location 61 61 61 65
DG size (kW) 1832.536 1850 1850 1450
Total real power loss (kW) 224.8799 83.1951 83.1568 83.1568 112.0217
Total reactive power loss
(kVAR)
102.1091 40.5795 40.5373 40.5373 55.1172
Minimum bus voltage (p.u.) 0.909253 0.9685 0.968559 0.968559 0.9660621
Cost of energy losses ($) 18101.7621 6696.8097 6693.7268 6693.7268 9017.2139
Cost of PDG($/MW h) 36.9007 37.2500 37.2500 29.2500
Table 12
Results for 69 bus system with DG at 0.9 pf lag.
With DG
Modied Novel method Combined power loss sensitivity method Index vector method Voltage sensitivity index method [18]
DG location 61 61 61 65
DG Size (kVA) 2236.704 2200 2200 1750
Total real power loss (kW) 27.3795 27.9091 27.9091 65.4502
Total reactive power loss (kVAR) 16.1734 16.4639 16.4639 35.6250
Minimum bus voltage (p.u.) 0.972844 0.9726 0.9726 0.969302
Cost of energy losses ($) 2203.9195 2246.5497 2246.5497 5268.4297
Cost of PDG ($/MW h) 40.5107 39.8500 39.8500 31.7500
Cost of QDG ($/MVAR h) 3.9343 3.8701 3.8701 3.0830
466 V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467
from DGs, combined power loss sensitivity based method is giving
better results for both the test systems based on the lower kVA and
kVAr requirements compared to the other methods. Therefore, the
consideration of reactive losses is essential looking into the voltage
prole and additional burden on the substation which may lower
the efciency due to higher reactive component of current. Reac-
tive capability of DGs to supply reactive power shall be considered
during voltage prole management and losses reduction that will
lower reactive burden on the input side of distribution station
which may help utilities deferring higher kVA requirements and
lower sizes of the equipments.
8. Conclusions
This paper presents comparison of four methods for optimally
allocating distributed generation for reduction of losses and
improvement in voltage prole. Combined power loss sensitivity
method is proposed and Novel method is modied for distributed
generation location and calculating their sizes for minimum loss
and improved voltage proles. The study is carried out on two
types of DG that are: DG operating at unity power factor and DG
operating at 0.9 power factor lag. The results have been obtained
for real and reactive power losses, voltage prole, cost component
for real power and reactive power obtained from DGs, and their
sizes. We can conclude that there is much reduction in real, reac-
tive power losses and improvement in voltage prole with DG at
0.9 pf lag due to its reactive power supply to the system. Therefore,
DG operating at lagging power factor and supplying reactive power
to the system is giving better results than DG at unity power factor.
From the results, it can be concluded that modied Novel method
and Combined Power Loss Sensitivity are giving overall better re-
sults for both the test systems in terms of lower reactive burden
on the input side of distribution station which may help utilities
deferring higher kVA requirements and lower sizes of the equip-
ments. The proper reactive power management and voltage prole
with DGs will help cost reduction due to the lower kVA require-
ments from the substation.
Acknowledgements
This work has been carried out under the Department of Science
and Technology, DST, New Delhi under the Project Grant: SR/S3/
EECE/0035/2012, SERB, New Delhi. The authors acknowledges the
DST, New Delhi for the grant of the project.
References
[1] Ackermann T, Andersson G, Sder L. Distributed generation: a denition. Electr
Power Syst Res 2001;57(3):195204.
[2] CIGRE, 2000 Technical broucher on modelling new forms of generation and
storage Tf38.01.10. <http://microgrids.power.ece.ntua.gr/documents/CIGRE-
TF-38010.pdf> [Retrieved 03.02.05].
[3] McDermott T. Behavioral models of DR technologies for feeder-level analysis.
In: Power energy society general meeting, IEEE, July 2009.
[4] Barker PP, de Mello RW. Determining the impact on distributed generation on
power systems: Part 1. Radial distribution systems. In: IEEE PES summer
meeting, vol. 3; 2000. p. 164556.
[5] Rahman TKA, Rahim SRA, Musirin I. In: Optimal allocation and sizing of
embedded generators national power energy conference (PECon) 2004, Kula
Lumpur, Malaysia; 2004. p. 28894.
[6] Kashem MA, Le An DT, Negnevitsky M, Ledwich G. Distributed generation for
minimization of power losses in distribution systems. In: Proc. IEEE PES; 2006.
[7] Singh RK, Goswami SK. Optimal allocation of distribution generations based on
nodal pricing for prot, loss reduction, and voltage improvement including
voltage rise issue. Electr Power Energy Syst 2010;32:63744.
[8] Acharya N, Mahat P, Mithulananthan N. An analytical approach for DG
allocation in primary distribution network. Int J Electr Power Energy Syst
2006;28:66978.
[9] Gozel Tuba, hakan Hocaoglu m. An analytical method for the sizing and sitting
of distributed generators in radial distribution systems. Electr Power Syst Res
2009;79:9128.
[10] Hasanpour S, Ghazi R, Javidi MH. A new approach for cost allocation and
reactive power pricing in a deregulated environment. Electr Eng, vol.
91. Springer- Verlag; 2009. p. 2734.
[11] Arya LD, Koshti Atul, Choube SC. Distributed generation planning using
differential evolution accounting voltage stability consideration. Electr Power
Energy Syst 2012;42:196207.
[12] Gautam Durga, Mithulananthan Nadarajah. Optimal DG placement in
deregulated electricity market. Electr Power Syst Res 2007;77:162736.
[13] El-khattam W, Bhattacharya K, Hegazy Y, Salama MMA. Optimal investment
planning for distributed generation in a competitive electricity markets. IEEE
Trans Power Syst 2004;19(3):167484.
[14] Nagaraju K, Sivanagaraju S, Ramana T, Satyanarayana S, Ramana PV. A novel
method for optimal distributed generator placement in radial distribution
system. J Distrib Gener Altern Energy 2011;26(1):719.
[15] Borges Carmen LT, Falca~o Djalma M. Optimal distributed generation
allocation for reliability, losses, and voltage improvement. Int J Electr Power
Energy Syst 2006;28:21320.
[16] Jain Naveen, Singh SN, Srivastava SC. Particle swarm optimization based
method for optimal siting and sizing of multiple distributed generators. In Proc
16th national power systems conference, 1517th December, 2010. p. 66974.
[17] Padma Lalitha M, Veera Reddy VC, Sivarami Reddy N, Usha Reddy V. DG source
allocation by fuzzy and clonal selection algorithm for minimum loss in
distribution system. Distrib Gener Altern Energy J 2011;26(4):1735.
[18] Qian Kejun, Zhou Chengke, allan Malcolm, Yuan Yue. Effect of load models on
assessment of energy losses in distribution generation planning. Electr Power
Res 2011;2:124350.
[19] Sadighizadeh M, Rezazadeh A. Using genetic algorithm for distributed
generation allocation to reduce losses and improve voltage prole. Proc
World Acad Sci Eng Technol 2008;27:2516.
[20] Rueda-Medina AugustoC, Franco JohnF, Rider MarcosJ, Padilha-Feltrin Antonio,
Romero Rubn. A mixed-integer linear programming approach for optimal
type, size and allocation of distributed generation in radial distribution
systems. Electr Power Syst Res 2013;97:13343.
[21] Celli G, Ghiani E, Mocci S, Pilo F. A multi-objective evolutionary algorithm for
the sizing and siting for distributed generation. IEEE Trans Power Syst
2005;20(2):7507.
[22] Gopiya Naik S, Khatod DK, Sharma MP. Optimal Allocation of Distributed
Generation in Distribution System for Loss Reduction. In: 2012 IACSIT
Coimbatore conferences. p. 426.
[23] Abou El-Ela AA, Allam SM, Shatla MM. Maximum optimal benets of
distribution generation using genetic algorithms. Electr Power Res
2010;80:86977.
[24] Ramana T, Ganesh V, Sivanagaraju S. Distributed Generator Placement and
Sizing in Unbalanced Radial Distribution System. Cogener Distrib Gener J
2010;25(1):5271.
[25] Ramachandra Murthy KVS, Ramalinga Raju M, Govinda Rao G, Narasimha Rao
K. Comparison of loss sensitivity factor & index vector methods in determining
optimal capacitor locations in agricultural distribution. In: Proc. 16th national
power systems conference, 1517th December, 2010. P. 2630.
[26] Kumar Injeti Satish, Navuri Prema Kumar. An efcient method for optimal
placement and sizing of multiple distributed generators in a radial distribution
systems. J Distrib Gener Altern Energy 2012;27(3):5271.
[27] Srinivasa Rao R, 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.
[28] A. Parizad, A. Khazali and M. Kalantar, Optimal placement of distributed
generation with sensitivity factors considering voltage stability and losses
indices, in Proc. IEEE of ICEE 2010, May 1113, 2010, pp. 18.
[29] Ghosh Sudipta, Ghoshal SP. Saradindu Ghosh Optimal sizing and placement of
distributed generation in a network system. Int J Electr Power Energy Syst
2010;32:84956.
[30] de Souza ARR, Fernades TSP, Aoki AR, sans MR, Oening AP. Sensitivity analysis
to connect distributed generation. Int J Electr Power Energy Syst
2013;46:14552.
[31] Rotaru F, Chicco G, Grigoras G, cartina G. Two stage optimal distributed
generation sizing with clustering based node selection. Int J Electr Power
Energy Syst 2012;40:1209.
[32] Sivanagaraju S, Sreenivasulu N, Vijayakumar M, Ramana T. Optimal conductor
selection for radial distribution systems. Electr Power Syst Res
2002;63:95103.
[33] Das D, Kothari DP, Kalam A. Simple and efcient method for load ow solution
of radial distribution networks. Electr Power Energy Syst 1995;17(5):33546.
[34] Nagaraju K, Sivanagaraju S, Ramana T, Prasad PV. A novel load ow method for
radial distribution system for realistic loads. Electr Power Compon Syst
2011;39(2):12841.
[35] Eminoglu U, Hocaoglu MH. A new power ow method for radial distribution
systems including voltage dependent load models. Electr Power Syst Res
2005;76(2):10614.
[36] Kashem MA, Ganapathy V, Jasmon GB, Buhari MI. A novel method for loss
minimization in distribution networks. In: International conference on electric
utility deregulation and restructuring and power technologies 2000, City
University, London. IEEE; 47 April, 2000. p. 2516.
[37] Baran ME, Wu FF. Optimal sizing of capacitors placed on a radial distribution
system. IEEE Trans, Power Deliv 1989;4(1):73543.
V.V.S.N. Murthy, A. Kumar / Electrical Power and Energy Systems 53 (2013) 450467 467