Electrical Power and Energy Systems 53 (2013) 967–973

Contents lists available at SciVerse ScienceDirect

Electrical Power and Energy Systems

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

Optimal allocation of combined DG and capacitor for real power loss

minimization in distribution networks
S. Gopiya Naik ⇑, D.K. Khatod, M.P. Sharma
Alternate Hydro Energy Centre, IIT Roorkee, Roorkee 247 667, India

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

Article history: Optimal siting and sizing of Distributed Generation (DG) and shunt capacitor at the distribution networks
Received 26 November 2012 for the purpose of real power loss minimization is drawing much attention of electric power utilities in
Received in revised form 1 June 2013 the present days. Some inherent benefits of power loss minimization includes: reduction of power flow in
Accepted 6 June 2013
feeder lines, releases stress on feeder loading, and hence increases their life time, adds opportunity to
using the existing facility to serve any increased load demand, avoidance of power purchased from the
grid and also the cost of loss compensating devices, reduction in customer bill, etc. In this paper, a
Keywords:
method based on analytical approach for optimal allocation (sizing and siting) of DG and capacitor with
Power distribution networks
Distributed generation
the objective to minimize the total real power loss subjected to equality and inequality constraints in the
Shunt capacitors distribution network is presented. A sensitivity analysis technique has utilized to identify the optimal
Optimal siting candidate locations for DG and capacitor placement and the heuristic curve fitting technique is used to
Optimal sizing determine their optimal capacity in the networks. To validate the suitability of the proposed method,
Loss minimization it has been applied to 12-bus and IEEE 33-bus test distribution systems. The obtained simulation results
and comparison of different cases considered reveals that allocation of DG and capacitor combination
results in significant loss reduction with good voltage profile and also release in the line loading in the
power distribution networks.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction power distribution. It is reported that as much as 13% of total

The electric power distribution networks usually operated at
low voltage are connected to the high voltage transmission sys-
tems and finally supply power to customers at low voltage. The to-
tal power losses in the distribution network lines is high because of
low voltage and high current in comparison to high voltage net-
work, which in turn, causes increase in the cost of power and poor
voltage profile along the distribution feeder. The total power loss in
the distribution network is composed of two parts: real power and
reactive power loss. The former loss is due to the flow of active
component of current required by the load and the later is due to
the flow of reactive component of current required to compensate
the reactive power requirement of network components and hence
to control of the system voltage. Among these losses, the effect of
active power loss is very important because it reduces the effi-
ciency of power transfer and deteriorates the voltage profile. The
minimization of real power loss in the distribution networks is
therefore of much significance compared to the transmission sys-
tem. The task of power loss reduction and enhancement of energy
efficiency of electric power delivery system mostly goes to electric
⇑ Corresponding author. Tel.: +91 9411149509.
E-mail address: gopiya_s@yahoo.co.in (S. Gopiya Naik).
power generated is wasted in the form of power losses at the dis-
tribution level [1]. The capacity of radial line is often limited, it is
therefore, necessary to consider some alternatives methods so that
the future load demands can be supplied ensuring supply quality
and reliability.
Most distribution network components like motors and trans-
formers are inductive in nature, so the network power factor will
be lagging, and results in the reduction of the system’s capacity, in-
crease the system losses, and reduces the voltage. Shunt capacitors
are used to alleviate some of these problems [2–4]. Apart from
reduction in power losses, the shunt capacitors enhances the volt-
age profile, improves power factor and voltage stability of the sys-
tem. Distributed Generation (DG) units can play a major role in
distribution system planning in recent years as DG integration into
the distribution system defers major system upgrade, reduces
overall energy loss and improves the supply quality and reliability
[5]. Even though, DG technologies have positive impacts on distri-
bution system, there might be certain technical challenges with the
inclusion of active DG units in conventional passive system. It is
important to mention that DG units should be applied in an effec-
tive manner without causing degradation of reliability, system
operation and supply quality. On the other hand, shunt capacitors,
commonly deployed for reactive power compensation, can also be

0142-0615/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijepes.2013.06.008
968 S. Gopiya Naik et al. / Electrical Power and Energy Systems 53 (2013) 967–973

considered in parallel with DG units for distribution system expan-

sion planning. It is evident that any loss reduction is beneficial to
distribution utilities, which is generally the entity responsible to
keep the losses at low levels. Loss reduction is therefore most
important factor to be considered in planning and operation of
DG [6].
The literature reports that the reactive power injected by shunt
capacitors can effectively reduce system energy loss, relieve feeder
loading and improve supply reliability [2,7]. Siting and sizing of
shunt capacitors needs to be investigated carefully to avoid voltage
rise problems, and hence to reduce the operating cost of DG units. Fig. 1. Model of a branch connected between buses p and q.
Various methods may be used to reduce the power losses in distri-
bution networks: capacitor placement, network reconfiguration, as well as multiple DG capacitor allocation. Finally, based on the
DG placement, and network reconfiguration in the presence of obtained results conclusions are given in Section 5.
capacitor, network reconfiguration in the presence of DG and DG
capacitor placement [8]. The optimal allocation of capacitors in 2. Problem formulation
power distribution networks with the objectives like minimizing
power loss, system capacity release, minimization of capacitor cost The problem of combined DG and capacitor allocation in distri-
and voltage profile improvement was presented in [9–13] using bution network with their suitable size is very important, because
the optimization techniques, viz. GA [10,12,13], heuristic strategies their improper allocation causes an increase in the system power
[11], and analytical approach [9]. Optimal siting and sizing of DG in loss and operating costs, reduces the energy efficiency. The main
the distribution networks to minimize power loss were suggested objective of the proposed study is to minimize the total real power
in [14–17,20,18,19,21,22]. The optimization techniques like ana- loss (PL) at peak load condition in the distribution network as given
lytical approaches [14,15,18,21], genetic algorithm (GA) [17], arti- in Eq. (5), subjected to equality and inequality constraints in Eqs.
ficial bee colony (ABC) algorithm [20], evolutionary programming (1), (2), (3), (7)–(10). During DG allocation, the voltage at various
(EP) [19] and GA–TS [22] were employed to arrive at the optimal buses should be maintained at proper limits for safe and reliable
solution. operation of the power distribution system and the current flow
Distribution network reconfiguration in the presence of DG to in the line conductor must be within the permissible limit.
reduce power loss was presented in [23–26]. The optimization Consider a branch connected between nodes p and q of a radial
techniques utilized are GA, ant colony optimization (ACO), Tabu distribution network as shown in Fig. 1. The real and reactive
search (TS) and modified honey bee optimization (MHBMO), power flow through the branch and the terminating node (q) volt-
respectively. age (neglecting shunt conductance and susceptance) are given by
Further, optimal allocation of DG and capacitor in the distribu- Eqs. (1)–(3), respectively as [19,31]:
tion network considering power loss reduction, minimizing the
Rpq
cost of DG and capacitor along with voltage profile improvement Ppq ¼ PFq þ P Lq  PDG
q þ ðP2pq þ Q 2pq Þ ð1Þ
were proposed in [27–29]. The optimization methods used are par- V 2p
ticle swarm optimization (PSO), PSO, and differential evolutionary
algorithm (DEA), respectively. X pq
Q pq ¼ Q Fq þ Q Lq  Q DG C
q  Qq þ ðP2pq þ Q 2pq Þ ð2Þ
In [27–29], the optimal siting and sizing of the DG injecting V 2p
only the real power and the capacitor combination is presented.
It is reported in [30], that DG units with reactive power control R2pq þ X 2pq
can provide better voltage profile and lower losses. An analytical V 2q ¼ V 2p  2ðPpq Rpq þ Q pq X pq Þ þ ðP 2pq þ Q 2pq Þ ð3Þ
approach based on exact loss formula was presented to find the V 2p
optimal size and location of DG to minimize the real power loss X X
[22], however voltage constraint has not been considered. Re- where PFq ¼ P ij and Q Fq ¼ Q ij
cently, another fast analytical approach to find the optimal size 8jji¼q 8jji¼q

of DG at optimal power factor has been proposed to minimize Here Ppq(Qpq) are the sending end active (reactive) power flows and
the power loss [18]. DG
Rpq(Xpq) are the series resistance (reactance). P DG
q ðQ q Þ are the active
Most of the approaches presented so far model the optimal allo-
cation of DG only. However, very few approaches [27–29] consid- (reactive) power injections by DG; Q Cq is the reactive power injec-
ered the optimal allocation of combined DG and capacitor, but tion by capacitor and P Lq ðQ Lq Þ are the total active (reactive) demand
release of line loading is not being considered and therefore needs load at bus q. PFq ðQ Fq Þ are the sum of active (reactive) power flows
further attention. The present paper considers the optimal alloca-
through all the downstream branches connected to bus q. Vq is
tion of DG and capacitor considering that DG is also capable to sup-
the magnitude of voltage at bus q. SB is a set of buses containing
ply reactive power in addition to real power. The optimal
all the buses in the system. The value of current flowing through
allocation of DG and capacitor with the key objective to minimize
a branch connected between nodes p and q is given as:
the real power loss besides satisfying the network constraints of vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u 2
the distribution network is presented. Heuristic sensitivity analy- uPpq þ Q 2pq
sis, i.e. both real and reactive power loss is being utilized to iden- Ipq ¼t ð4Þ
V 2p
tify the optimal candidate locations for DG and capacitor
placement, and the quadratic curve fitting technique is employed Mathematically, the objective function is given as:
to determine their optimal capacity in the distribution networks. X
Min PL ¼ I2pq Rpq ð5Þ
This paper is organized as follows: followed by the brief intro-
8p;qjp;q2SB
duction in Section 1, Section 2 gives the problem formulation,
and the methodology proposed is discussed in Section 3. In Sec- The above objective function is subjected to the set of equality and
tion 4, results and discussions are given for single DG capacitor inequality constraints as given below:
 S. Gopiya Naik et al. / Electrical Power and Energy Systems 53 (2013) 967–973 969

1. Equality constraints 1. Once the optimal locations for DG and capacitor are identified,
– Power balance: The flow of active and reactive power in all the the maximum DG capacity is fixed as 50% of the total feeder
branches of the system must satisfy Eqs. (1) and (2), respectively. loading capacity and the capacitor capacity is fixed as 100% of
– Voltage equation: For all branches of the system, the voltage total MVAr loading of the network under consideration. For
magnitudes at sending and receiving end nodes must satisfy the combined DG–capacitor case both the DG and capacitor
Eq. (3). are varied in small incremental steps simultaneously. At each
– DG power factor: The power factor of DG connected at bus q step perform the load flow using Eqs. (1)–(3) and get the total
must satisfy the following eqn. real power loss using Eqs. (4) and (5) and check the constraints
described in Eqs. (7) and (8). If, at least one of the constraints is
PDG not satisfied, decrease/increase the capacity by 5% (one at a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiq ¼ cos uq ð6Þ time) and again perform the load flow and after ensuring both
2 DG 2
ðPDG q Þ þ ðQ q Þ constraints are satisfied the corresponding capacity gives the
optimal size(s).
2. Inequality constraints
2. Step 1 is repeated for all sizes and finally, the minimum loss
– Bus voltage: The voltage at each bus must lie within the pre-
point is obtained.
scribed limits
3. The DG and/or capacitor size corresponding to the minimum
V min
q 6 V q 6 V max
q q 2 SB ð7Þ loss point are the optimal capacities.
4. Place the optimal sizes obtained in step 3 at optimal locations
– Line current: The flow of current through each branch should and once again perform load flow and obtain the total real
not exceed its thermal limit power loss, bus voltage values and line current value in the
Ipq 6 Irated 8pandq 2 SB ð8Þ network.
pq

– DG capacity: The DG capacity should not exceed certain per-

3.2. Multiple DG capacitor allocation
centage of total feeder load of network
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X 2 DG 2 X 2 2 The algorithm of single DG capacitor allocation given in the
ðP DG
q Þ þ ðQ q Þ 6 0:5  ðPLq Þ þ ðQ Lq Þ ð9Þ above section can be extended for multiple DG capacitor allocation
q2SB q2SB
too with the objective to minimize the total real power loss in 33-
– Capacitor capacity: The capacity of capacitor should not exceed bus IEEE test distribution network subject to the constrains (1), (2),
the total reactive power load of network (3), (7)–(10). For multiple DG capacitor allocation it is assumed
X X L here that the maximum number of DG or capacitor located at each
Q Cq 6 1:0  Qq ð10Þ candidate bus is limited to one. Further, only the high sensitive
q2SB q2SB
buses whose values are approximately equal are considered as
where V min
q and V max
q are the minimum and maximum values of the candidate buses for multiple DG capacitor allocation.
voltages at bus q (i.e. 0.95–1.05), respectively. Irated
pq is the thermal
limit of a branch between nodes p and q. cos /q is the power factor
of DG at qth bus. 4. Simulation results and discussion

The proposed algorithm is tested on two types of radial distri-

bution networks, namely, 12-bus and IEEE 33-bus networks. The
3. Proposed methodology
Backward–Forward sweep load flow method [31] is used to find
voltage magnitude and phase angle at various buses, line flow
3.1. Single DG capacitor allocation
and line losses. The type of DG model considered in this study is
a controllable synchronous generator based which delivers both
Sensitivity analysis based heuristic method [32] has been used
real and reactive power at fixed power factor mode [37] as given
to reduce the search space by selecting most sensitive buses only
by Eq. (8). The algorithm given above is implemented in the matlab
for identifying location for DG and capacitor and for finding the
environment. To demonstrate the effectiveness of the proposed
size quadratic curve fitting technique [11,33] has been used. The
algorithm, the following cases are considered for each test system:
complete algorithm is as given in the following section:
Following are the broad steps involved in the proposed method:
Case 1: Only the capacitor is placed at the optimal location.
Case 2: Only the DG operating at unity power factor (upf) is
3.1.1. Optimal location of DG and capacitor
placed at the optimal location.
To identify the optimal location for the combined DG and capac-
Case 3: Combined DG operating at upf and the capacitor were
itor, the following steps are used:
placed at their optimal location.
Case 4: Combined DG operating at 0.9 power factor (considered
1. Determination of real power loss sensitivity with respect to
as optimal pf) lag and the capacitor were placed at their optimal
(w.r.t.) active and reactive power injection at each bus [32].
location.
2. Arrange the buses in descending order of their sensitivities.
Case 5: Combined DG operating at the pf equal to total feeder
3. The bus having highest real power loss sensitivity w.r.t. real
load pf and the capacitor were placed at the optimal location.
power injection is the candidate bus for DG siting and the bus
with highest real power loss sensitivity w.r.t. reactive power
injection is considered as the candidate bus for capacitor 4.1. 12-Bus network
placement.
This is an 11 kV radial distribution systems consisting of 12
3.1.2. Optimal size of DG and capacitor buses and 11 branches with total feeder load of 435 kW and
To obtain the optimal size of combined DG and capacitor, the 395 KVAr. The line and load data are given in [34]. The combined
following steps have been taken [11,33]: load power factor of this system is 0.75.
970 S. Gopiya Naik et al. / Electrical Power and Energy Systems 53 (2013) 967–973

Bus Number
0

Real power loss sensitivity

-0.01 1 2 3 4 5 6 7 8 9 10 11 12
-0.02
-0.03
-0.04
-0.05
-0.06
-0.07
-0.08
-0.09
-0.1
Real Power Loss Sensitivity Real Power Loss Sensitivity
w.r.t Real Power Injection w.r.t Reactive Power Injection

Fig. 2. Real power loss sensitivity at various buses of 12-bus network.

Table 1
Optimal results of 12-bus network.

Particulars Total real power % Loss reduction Capacity/optimal location Minimum voltage
loss (kW) in the system (p.u.)
Base case 198.9 – – 0.94414 at bus 12
Case 1: Only capacitor 134.3 32.47 0.16 MVAr capacitor at bus 12 0.95596 at bus 11
Case 2: Only DG at upf 109.2 45.09 0.2 MW DG bus 12 0.98032 at bus 8
Case 3: DG with upf and capacitor 71.93 63.8 0.12 MW DG at bus 12 and 0.24 MVAr capacitor at bus 12 0.9815 at bus 8
Case 4: DG at 0.9 Pf with Qc = 0 57.41 71.14 0.2 MW and 0.1 MVAr DG at bus 12 0.9847 at bus 7
Case 5: DG at 0.75 pf with Qc = 0 47.56 76.09 0.2 MW and 0.16 MVAr DG at bus 12 0.9867 at bus 7

By sensitivity analysis as shown in Fig. 2, the optimal bus to
place the DG and capacitor was found to be 12 where the real
power loss sensitivity w. r. t real and reactive power injection is
high.
Case 1: Only the capacitor is placed at the optimal location.
In this case the maximum capacity of capacitor is taken equal to
total MVAr loading of the network, i.e. 0.395  0.4 MVAr and
varied in small incremental steps from 0 to 0.4 MVAr.
Case 2: Only DG operating at unity power factor (upf) is placed
at the optimal location
In this case the maximum capacity of the DG is taken equal to
50% of the total MW loading of the network, i.e.
0.5  0.435  0.25 MW and is varied in small incremental steps
from 0 to 0.25 MW.
Case 3: Combined DG operating at upf and capacitor were
placed at their optimal location
In this case the maximum capacity of DG and capacitor are
selected same as in Case 1&2 and are simultaneously varied
in small incremental steps from 0 to 0.25 MW and 0 to
0.4 MVAr, respectively.
Case 4: Combined DG operating at 0.9 pf and the capacitor were
The optimal DG and capacitor sizes obtained for the different
cases considered are 0.16 MVAr, 0.2 MW, (0.12 MW, 0.24 MVAr),
(0.2 MW, 0.1 MVAr with Qc = 0) and (0.2 MW, 0.16 MVAr with
Qc = 0), respectively. The results of comparison of various cases
considered are given in Table 1. From the table, it may be noted
that more loss reduction is possible if DG capable to supply reac-
tive in addition to real power (i.e. when the DG power factor is
0.75) and is placed at the optimal location with optimal sizes in
the network compared to other considered cases.
The comparisons of voltage profile at various buses for the dif-
ferent cases considered in a 12-bus radial distribution network are
depicted in Fig. 3. It can be seen from the figure that improvement
in the voltage profile by optimal placement of combined DG oper-
ating at upf and capacitor (Case 3) is somewhat superior compared
to DG alone. Alternatively, if DG capable of supplying reactive
power in addition to real power is placed at optimal location where
improvement in the voltage profile is found to be better than pre-
vious cases.
1.01
1
Voltage Profile (p.u)

0.99
placed at their optimal location.
0.98
In this case the maximum DG capacity is selected equal to 50%
0.97
of the total MVA loading of the network at 0.9 pf. So, the max-
0.96
imum DG capacity is 0.3 MW and 0.15 MVAr (with capacitor
0.95
value set to zero, i.e. Qc = 0), respectively, and are varied in small
0.94
incremental steps from 0 to 0.3 MW and 0 to 0.15 MVAr. 0.93
Case 5: Combined DG operating at power factor selected equal 0.92
to the total feeder load power factor and the capacitor were 0.91
placed at their optimal location. 1 2 3 4 5 6 7 8 9 10 11 12
In this case the maximum DG capacity is selected equal to 50% Bus Number
of the total MVA loading of the network at 0.75 pf lag (i.e. Base Case Case3
PFDG = PFD). So, the maximum DG capacity is 0.25 MW and Case 1 Case 4(0.9 pf)
0.2 MVAr (with capacitor value set to zero), respectively, and Case 2 Case5(0.75 pf)
are varied in small incremental steps from 0 to 0.25 MW and
0 to 0.2 MVAr. Fig. 3. Comparison of voltage profile for different Cases in 12-bus network.
 S. Gopiya Naik et al. / Electrical Power and Energy Systems 53 (2013) 967–973 971

Fig. 3 also shows the comparison of voltage at various bus of 12- Case 4: Combined DG operating at 0.9 pf (considered as optimal
bus network for DG operating at 0.9 pf (Case 4) and 0.75 pf (Case pf) lag and the capacitor were placed at their optimal location
5), respectively. It can be seen that the voltage profile improve- In this case the maximum DG capacity is selected equal to 50%
ment with DG operating at 0.75 pf lag is better compared to the of the total MVA loading of the network at 0.9 pf. So, the max-
one at 0.9 pf lag, but the DG capacity required is more which re- imum MW and MVAr capacity of DG is 2.0 and 1.0, respectively,
sults in increased cost of the DG with additional loss reduction and the maximum capacitor capacity is selected as 2.3 MVAr
benefit of about 5%. and is simultaneously varied in small incremental steps from
0 to 2 MW, 0 to 1 MVAr and 0 to 2.3 MVAr.
4.2. 33-Bus network Case 5: Combined DG operating at the pf equal to total feeder
load pf and the capacitor were placed at their optimal location.
This is a 12.66 kV radial distribution system having 33 buses In this case the maximum DG capacity is selected equal to 50%
and 32 branches. The total load of the feeder is 3.72 MW and of the total MVA loading of the network at 0.85 pf lag (i.e.
2.3 MVAr. The line and load data are given in [35]. The combined PFDG = PFD). So, maximum MW and MVAr capacity of DG is
load power factor of this system is 0.85. 2.0 and 1.2, respectively, and the maximum capacity of capaci-
By sensitivity analysis given in Fig. 4, the optimal locations ob- tor is selected as 2.3 MVAr and is simultaneously varied in small
tained for DG and capacitor are 18 and 33 where the real power incremental steps from 0 to 2 MW, 0 to 1.2 MVAr and 0 to
loss sensitivity w. r. t real and reactive power injection are high, 2.3 MVAr.
respectively.
The optimal sizes obtained for different cases considered are
Case 1: Only capacitor is placed at the optimal location. 1 MVAr, 1 MW, (1 MW, 1 MVAr), (1.0 MW, 0.5 MVAr, 1 MVAr),
In this case the maximum capacity of capacitor is taken equal to and (0.8 MW, 0.4 MVAr, 0.8 MVAr), respectively. Table 2 compares
total MVAr loading of the network, i.e. 2.3 MVAr and varied in the results of different cases considered. It is seen from the table
small incremental steps from 0 to 2.3 MVAr. that if optimal size of DG that supplying both real and reactive
Case 2: Only DG operating at upf is placed at the optimal power at bus 18 and the capacitor at bus 33 were placed simulta-
location. neously, more loss reduction can be achieved compared to the
In this case the maximum capacity of the DG is taken equal other cases considered.
to 50% of the total MW loading of the network, i.e. The comparisons of voltage profile at different buses for the var-
0.5  3.72  2.0 MW and is varied in small incremental steps ious cases considered in a 33-bus network are depicted in Fig. 5. It
from 0 to 2.0 MW. can be seen from the figure that if combined DG and capacitor,
Case 3: Combined DG operating at upf and capacitor were where DG is capable to supply reactive power in addition to real
placed at their optimal location. power are optimally allocated can result in the substantial
In this case the maximum capacity of DG and capacitor are improvement in the voltage profile along the feeder compared to
selected same as cases 1 and 2 and are simultaneously varied
in small incremental steps from 0 to 2 MW and 0 to 2.3 MVAr,
respectively.
1.04
1.02
Voltage Profile (p.u)

Bus Number 1
0 0.98
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
Real power loss sensitivity

-0.02 0.96
-0.04 0.94
-0.06 0.92
0.9
-0.08
0.88
-0.1
0.86
-0.12 0.84
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33
-0.14
-0.16 Bus Number
-0.18 Base Case Case 3

Real Power Loss Sensitivity Real Power Loss Sensitivity Case 1 Case 4 (0.9 pf)
w.r.t Real Power Injection w.r.t Reactive Power
Injection Case 2 Case 5 (0.85 pf)

Fig. 4. Real power loss sensitivity at various buses of a 33-bus network. Fig. 5. Comparison of voltage profile for different cases in 33-bus network.

Table 2
Optimal results of 33-bus network.

Particulars Total real power % Loss Capacity/optimal location Minimum voltage

loss (kW) reduction in the system (p.u.)
Base Case 213.3 – – 0.9065 at bus 18
Case 1: Only capacitor 164.6 22.83 1.0 MVAr Capacitor at bus 33 0.91654 at bus 18
Case 2: Only DG with upf 142.34 33.29 1.0 MW DG at bus 18 0.9311 at bus 33
Case 3: DG with upf and capacitor 96.70 54.66 1.0 MW DG at bus 18 and 1.0 MVAr capacitor at bus 33 0.96003 at bus 30
Case 4: DG at 0.9 pf and capacitor 90.9 57.38 1.0 MW, 0.5 MVAr DG at bus 18 and 1.0 MVAr capacitor at bus 33 0.9646 at bus 30
Case 5: DG at 0.85 pf and capacitor 89.72 57.94 0.8 MW, 0.4 MVAr DG at bus 18 and 0.8 MVAr capacitor at bus 33 0.9566 at bus 30
972 S. Gopiya Naik et al. / Electrical Power and Energy Systems 53 (2013) 967–973

Before DG-Capacitor
Current magnitude (p.u)
After DG-Capacitor

Branch Number

Fig. 6. Feeder loading before and after DG–capacitor allocation for Case 5 of 33-bus network.

Table 3 Table 4
Comparison of optimal results of different approaches for 33-bus network. Comparison between single and multiple DG capacitor allocation.

Method [21] [20] [36] Proposed Particulars Single DG capacitor Multiple DG capacitor
allocation allocation
Optimal location(s) 6 6 6 18 and 33
DG capacity (MW) 2.49 3.38 2.4 0.76 DG capacity (MW) 0.76 0.855
Capacitor capacity (MVar) – – – 0.8 Capacitor capacity (MVar) 0.80 0.90
Percentage loss reduction 47.33 44.83 48.19 57.94 Power loss 89.72 84.28
% Loss reduction 57.94 60.49

1.01
1 [21,36], the DG units are limited to supply real power only where
as in the method [20] the DG unit is capable to supply both real and
Voltage profile (p.u)

0.99
reactive power. But, in the presented work the combined DG and
0.98 capacitor (Case 5) where DG unit is capable to deliver both real
0.97 and reactive power to the network is resulted in more loss reduc-
0.96 tion and better voltage profile in comparison to other methods.

0.95
0.94
0.93
1 3 5 7
Bus Number
Single DG Capacitor Multiple DG Capacitor
Allocation allocation
Fig. 7. Voltage profile of single/multiple DG capacitor allocation.
the case when DG supplying only real power and capacitor combi-
nation was placed at their optimal locations.
Fig. 5 also shows the comparison of voltage at various bus of 33-
bus network for 0.9 pf (Case 4) and 0.85 pf (Case 5). It can be seen
that there is voltage limit violation with DG operating at 0.9 pf lag
compared to that one at 0.85 pf lag, also the DG capacity required
is more which results in increased cost of the DG with less loss
reduction of about 0.56% from Case 5.
Feeder loading before and after DG–capacitor allocation for the
Case 5 of 33-bus network is given in Fig. 6. It can be seen from the
figure that there results in significant release of (32.64%) line load-
ing in comparison to base case (without DG). So, the life time of the
feeder is increased and the existing system can be used to supply
any possible future load growth.
To validate the results obtained by the proposed method, the
proposed analytical method is compared with the solutions ob-
tained by analytical method [21,36], and ABC algorithm [20] and
the same is given in Table 3. In those methods proposed in
4.3. Multiple DG capacitor allocation
9 11 13 15 17 19 21 23 25 27 29 31 33
Optimal locations: From Fig. 4 it s clear that bus numbers 18
and 17 have got highest values of real power loss sensitivity in
decreasing order with respect to real power injection and these
values are very close to each other. Therefore, bus number 18
and 17 are considered as the candidate buses for DG siting. Further,
bus number 33 and 32 have got highest values of real power loss
sensitivity in decreasing order with respect to reactive power
injection and are approximately same therefore, these buses are
considered as the candidates for capacitor siting.
Optimal size(s): For determining the optimal size of multiple
DG and capacitor, simulation is conducted only for the case 5 of
IEEE 33-bus network. Applying the algorithm detailed in Sec-
tion 3.1.2, optimal sizes of multiple DG and capacitor required
for total possible minimum real power loss in the system satisfying
the system constraints are found to be 0.447 MVA, 0.559 MVA and
0.4 MVAR, 0.5 MVAR at locations 18, 17 and 33, 32, respectively.
With this capacity the resulted total real power loss is 84.28 kW
which is lesser than single DG capacitor allocation by 5.44 kW,
but at the cost of more DG and capacitor size requirements. Com-
parison of single and multiple-DG capacitor allocation is given in
Table 4.
Fig. 7 shows the voltage profile for single and multiple DG
capacitor allocation cases. It is observed that voltage profile is
somewhat improved in the case of multiple DG capacitor
allocation.
 S. Gopiya Naik et al. / Electrical Power and Energy Systems 53 (2013) 967–973
5. Conclusions
This paper was presented a method for optimal siting and sizing
of combined DG and capacitor in the distribution network. It can be
concluded that allocation of capacitor alone can improve the volt-
age but may not reduce the loss as expected. On the other hand,
allocation of DG can reduce the power losses and also improves
the system voltage profile. If the DG, with both real and reactive
power generation capability is combined with the capacitor and
is optimally allocated, can result in substantial reduction in real
power loss of the distribution network and also improves the volt-
age profile along the feeder. In comparison to the conventional
planning methods, the application of combined DG unit and capac-
itor can largely reduce the investment besides improving supply
quality and reliability. The proposed methodology has been further
extended for multiple combined DG–capacitor (MCDGC) allocation
and it is found that some improvement in the loss reduction and
voltage profile is possible with MCDGC compared to the single
DG capacitor case.
Acknowledgement
The authors would like to acknowledge AICTE’s Quality
Improvement Programme (QIP) Scheme and the Management
P.E.T. (R), Mandya (Karnataka), for allowing pursuing the Ph.D.
work.
References
[1] Song YH, Wang GS, Johns AT, Wang PY. Distribution network reconfiguration
for loss reduction using fuzzy controlled evolutionary programming. Proc Gen
Trans Distrib 1997;144:345–50.
[2] Gallego RA, Monticelli AJ, Romero R. Optimal capacitor placement in radial
distribution networks. IEEE Trans Power Syst 2001;16(4):630–7.
[3] Ng HN, Salama MMA, Chikhani AY. Classification of capacitor allocation
techniques. IEEE Trans Power Deliv 2000;15:387–92.
[4] Chians HD, Wang JC, Cockings O, Shin HD. Optimal capacitor placement in
distribution systems: Part I, Part II. IEEE Trans Power Deliv 1990;5(2):643–7.
[5] Pecas JA, Lopes, Hatziargyriou N, Mutale J, Djapic P, Jenkins N. Integrating
distributed generation into electric power systems: a review of drivers,
challenges and opportunities. Electr Power Syst Res 2007;77:1189–203.
[6] Singh D, Singh D, Verma KS. Multi-objective optimization for DG planning with
load models. IEEE Trans Power Syst 2009;24(1):427–36.
[7] Etemadi AH, Fotuhi-Firuzabad M. Distribution system reliability enhancement
using optimal capacitor placement. IET Gener Trans Distrib 2008;2(5):621–31.
[8] Gopiya Naik S, Khatod DK, Sharma MP. Planning and operation of distributed
generation in distribution networks. Int J Emerg Technol Adv Eng
2012;9(2):381–8.
[9] Haque MH. Capacitor placement in radial distribution systems for loss
reduction. IEE Proc Gen Trans Distrib 1996;146(5).
[10] Sundhararajan S, Pahwa A. Optimal selection of capacitors for radial
distribution systems using a genetic algorithm. IEEE Trans Power Syst
1994;9(3):1499–507.
[11] Mekhamer SF, El-Hawary ME, Soliman SA, Moustafa MA, Mansour MM. New
heuristic strategies for reactive power compensation of radial distribution
feeders. IEEE Trans Power Deliv 2002;17(4):1128–35.
[12] Levitin G, Kalyuzhny A, Shenkman A, Chertkov M. Optimal capacitor allocation
in distribution systems using a genetic algorithm and a fast energy loss
computation technique. IEEE Trans Power Deliv 2000;15(2):623–8.
973
[13] Park JY, Sohn JM, Park JK. Optimal capacitor allocation in a distribution system
considering operation costs. IEEE Trans Power Syst 2009;24(1):462–8.
[14] Hung DQ, Mithulanathan N, Bansal RC. Multiple distributed generators
placement in primary distribution networks for loss reduction. IEEE Trans
Ind Electron 2011;99:1–8.
[15] Wang C, Nehir MH. Analytical approaches for optimal placement of distributed
generation sources in power system. IEEE Trans Power Syst
2004;19(4):2068–76.
[16] Quezeda VH Mendez, Jua-Rivier Abbad, Gomez T. Assessment of energy
distribution losses for increasing penetration of DG. IEEE Trans Power Syst
2006;21(2):533–40.
[17] Mithulanathan N, Oo Than, Van Phu Lee. Distributed generator placement in
power distribution system using genetic algorithm to reduce losses,
Thammasat. Int J Sci Technol 2004;9(3):55–62.
[18] Hung DQ, Mithulananthan N, Bansal RC. Analytical expressions for DG
allocation in primary distribution networks. IEEE Trans Energy Convers
2010;25(3):814–20.
[19] Khatod DK, Pant Vinay, Sharma Jaydev. Evolutionary programming based
optimal placement of renewable distributed generators. IEEE Trans Power Syst
2013;28(2):683–95.
[20] Abu-Mouti FS, El-Hawary ME. Optimal distributed generation allocation and
sizing in distribution systems using artificial bee colony algorithm. IEEE Trans
Power Deliv 2011;26(4):2090–101.
[21] Acharya Naresh, Mahat Pukar, Mithulananthan N. An analytical approach for
DG allocation in primary distribution networks. Electr Power Energy Syst
2006;28:669–78.
[22] Gandomkar M, Vakilian M, Ehsan M. A genetic based tabu search algorithm for
optimal DG allocation in distribution networks. Electr Power Compon Syst
2007;33(12):1351–62.
[23] Choi JH, Kim JC. Network reconfiguration at the power distribution system
with dispersed generations for loss reduction. In: Proc of IEEE power
engineering society winter meeting, vol. 4; 2000. p. 2363–7.
[24] Wu YK, Lee Ching-Yu, Liu Le-Chang, Tsai Shao-Hong. Study of reconfiguration
for the distribution system with DG. IEEE Trans Power Deliv
2010;25(3):1678–84.
[25] Khorshidi R, Niknam T, Nayeripour M. Distribution feeder reconfiguration
considering distributed generators. World Acad Sci Eng Technol 2008:46.
[26] Niknam T, Fard Abdollah Kavousi, Fard Alireza. Distribution feeder
reconfiguration considering fuel cell/wind/photovoltaic power plants. Renew
Energy 2010;37:213–25.
[27] Kai Zou, Agalgaonkar AP, Muttaqi KM, Perera S. Voltage support by distributed
generation units and shunt capacitors in distribution systems. In: IEEE power
and energy society general meeting; 2009. p. 1–8.
[28] Zou Kai, Agalgaonkar AP, Muttaqi KM, Perera S. Optimization of distributed
generation units and shunt capacitors for economic operation of distribution
systems. In: Australasian universities power, engineering conference; 2008. p.
1–7.
[29] Gunda Jagadeesh, Khan Nasim Ali. Optimal location and sizing of DG and shunt
capacitors using differential evolution. Int J Soft Comput 2011;6(4):128–35.
[30] Quezeda VH Mendez, Abbad Jua-Rivier, Gomez T. Assessment of energy
distribution losses for increasing penetration of DG. IEEE Trans Power Syst
2006;21(2):533–40.
[31] Haque MH. Efficient load flow method for distribution systems with radial or
mesh configuration. IEE Proc Gen Trans Distrib 1996;143(1):33–8.
[32] Khatod DK, Pant V, Sharma JD. A novel approach for sensitivity calculations in
the radial distribution systems. IEEE Trans Power Deliv 2006;21(4):2048–57.
[33] Abu-Mouti FS, El-Hawary ME. Heuristic curve-fitted technique for distributed
generation optimization in radial distribution feeder systems. IET Gen Trans
Distrib 2011;5(2):172–80.
[34] Das D, Nagi HS, Khotari DP. Novel method for solving radial distribution
networks. IEE Proc Gen Trans Distrib 1994;141(4).
[35] Baran M, Wu FF. Network reconfiguration in distribution systems for loss
reduction and load balancing. IEEE Trans on Power Deliv 1989;4(2):1401–7.
[36] Shukla TN, Singh SP, Srinivasarao V, Naik KB. Optimal sizing of distributed
generation placed on radial distribution systems. Electr Power Compon Syst
2010;38:260–74.
[37] Teng JH. Modeling distributed generations in three-phase distributed load
flow. IET Gen Trans Distrib 2008;2(3):330–40.