Proceedings of 2015 RAECS UlET Panjab University Chandigarh 21_22nd December 2015

Capacitor Placement in Unbalanced Radial

Distribution System for Loss Reduction
Ravi Teja Bhimarasetti Ashwani Kumar
Department of Electrical Engineering Department of Electrical Engineering
NIT Kurukshetra NIT Kurukshetra
Kurukshetra, India Kurukshetra, India
e-mail [email protected] e-mail [email protected]

Abstract-Reduction of active power loss in distribution and size of shunt capacitors in three phase unbalanced RDS by
systems is very important to improve the overall efficiency of taking cost of energy loss and cost related to capacitor
electrical power distribution systems. The active power loss due purchase and capacitor installation as main objective function
to reactive component of branch currents can be reduced by to be minimized. The Power Loss Index has been used to
supplying part of the reactive power demands locally with the select the candidate node for capacitor placement and then
help of capacitors. This paper proposed a simple method for
variational technique has been used to find the optimal sizes.
optimal placement and sizing of capacitor while minimizing
Based on bacterial foraging oriented by particle swarm
power losses and improving the voltage profile in an unbalanced
optimization algorithm (BFPSO), a new method was proposed
radial distribution system (URDS). The performance of the
to fmd the optimal location and the size of fixed and switching
proposed method have been tested on two case studies 19-bus
UBRS and 2S-bus UBRS. It was found that a significant loss
capacitor banks [14]. In [15], optimal capacitor placement was
saving can be achieved by placing optimal capacitors in the
done by using a hybrid honey bee colony optimization
system.
algorithm with the objective function of minimizing power
system losses and unbalances and maximizing the net savings
Keywords- Load flow; radial distribution systems; capacitor while maintaining voltage and total harmonic distortion of
placement buses in an acceptable. The Index Vector method was
implemented for optimal capacitor placement on three phase
1. INTRODUCTION unbalanced radial distribution system [16]. Haque M.H.
A distribution system connects consumers to the high proposed loss reduction technique in balanced radial
voltage transmission system. The losses in the distribution distribution system by supplying the reactive power locally
system are about 5-13% of total generated power. So the loss with the help of capacitors [7]. However, this paper proposes
reduction in the distribution network is the most important single capacitor placement method for loss reduction in an
priorities in the design and operation of electric power unbalanced radial distribution system. The basic idea of the
network. There are many ways to reduce the losses as like proposed method has been taken from [7]. In this proposed
capacitor placement, Distributed Generation placement, load method, the optimal size of capacitor at each node can be
management, Network Reconfiguration and so on [1]. determined through optimizing the loss saving equation. The
node at which the loss saving is maximum will be taken as
Several methods for loss reduction on balanced RDS by candidate node for capacitor placement and the corresponding
placing capacitor in an optimal location have been reported in size is the optimal size of capacitor. The load flow of
the literature [2-8]. In [9], Decision theory criteria approach unbalanced radial distribution has been implemented from
was applied to fmd the optimal allocation and the sizing of [17]. The results have been obtained for unbalanced
capacitors in unbalanced systems with the presence of distribution network of IEEE 19 bus and IEEE 25 bus systems
harmonic sources, also taking into account the uncertainties [18].
due to the presence of unbalanced loads. Carpinelli G. et al.
[10] proposed a new probabilistic method for the optimal IT. PROPOSED LOAD FLOW ApPROACH IN DISTRIBUTION
location of capacitors in an unbalanced distribution networks SYSTEMS
and the micro genetic algorithm was applied to reduce the Let us consider the following
computational efforts. In [II], a Hybrid Particle Swarm Ploss Total real power loss.
=

Optimization (HPSO) combined with a Hybrid Power Flow

algorithm (HPF) was used to find the optimal locations and
Qloss Total reactive power loss.
=

Sloss Ploss + j Qloss

=

sizes of shunt capacitors in an unbalanced RDS while taking

the harmonics into account. Carpinelli G. et al. [12] presented la' lb' Ie are the branch currents in three phases.
a method for placement of shunt capacitors in three phase
Ian IbT' leT are the active component of branch currents in
three phases.
unbalanced distribution systems with harmonic sources while
considering capacitor costs, energy costs as well as costs lai' lbi' lei are the reactive component of branch currents m
three phases.
associated with voltage harmonics. Subrahmanyam JBV, et al.
Z is the branch impedance.
[13] proposed a simple method for choosing optimal location

978-1-4673-8253-3/15/$31.00 ©2015 IEEE

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SILCHAR. Downloaded on October 02,2021 at 01:21:44 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of 2015 RAECS UlET Panjab University Chandigarh 21_22nd December 2015

a is the set of branches from source bus to the rh capacitor After placing the capacitor, the new branch currents in three
bus. phases are given by
The total power loss in an unbalanced distribution system is lai(new) = lai +D *ICa
given by [19] lbi(new) = lbi +D *ICb
Sloss Lbranehes Zaa *la *(la)* + Zab *lb *(la)* + Zae *
=
lei(new) lei +D *ICc
Ie *(la)* + Zba *la *(lb)* + Zbb *lb *(lb)* + Zbe *Ie *
=

Where D=1 ; if branch E oc

(lb)* + Zca *la *CO* + Zeb *lb *(le)* + Zee *Ie *CO* (1) =0 ; otherwise
Where la lar + jIai After capacitor placement, the Ploss for the compensated
(-� - j '7)
=

system can be written as

lb (lbr + jIbJ * a
Ploss� p LbranehesRaa *(lai +D *ICa)2 + Rbb *
=

(lbi +D *ICb)2 + Ree *(lei +D *ICJ2 + Rab (-(lai +

Ie (ler + jIeJ * ( -� + j '7)
D *ICa)(lbi +D *ICb) + V3lar(lbi +D *ICb) -
=

Sloss Ploss + j Qloss

V3lbr(lai +D *ICa) ) + Rbe (-(lbi +D *ICb)(lei +
=

Ploss associated with both active and reactive component of

branch currents can be given as: D *ICc) + V3lbr(lei +D *ICJ - V3ler(lbi +D *ICb) ) +
Ploss LbranehesRaa *(l�r + l�i) +Rbb *(lEr + lEi) +
Rca (-(lei +D *ICJ(lai +D *ICa) + V3ler(lai +
=

Ree *(l�r + 11J + Rab(-larlbr - lailbi + V3larlbi -

V3laJbr) + Rbe(-lbrler - IbJei + V3lbrlei - V3lbJer) + D *ICa) - V3lar(lei +D *ICe) ) (4)
Rca(-lerlar - leilai + V3lerlai - V3leJar) (2) The loss saving Plosssaving is the difference between eqn 0-0
and is given by
For a given configuration of single source unbalanced Plosssaving Plossr - Ploss�ap
= (5)
radial network, Ploss associated with active component of Plosssaving LbranehesRaa *(D *ICa2 +D * 2lai *ICa) +
=

branch current cannot be minimized because all active power Rbb *(D *ICb2 +D * 2lbi *ICb) + Ree *(D *ICc2 +D *
must be supplied by the source at the root bus. However, the
Ploss associated with the reactive component of branch 2lei *ICJ + Rab[-CD *lai *ICb +D *lbi *ICa +D *ICa *
currents can be minimized by supplying part of the reactive ICb) +D *V3lar *ICb - D *V3lbr *ICa] + Rbe[-CD *lbi *
power demands locally. [7] ICc +D *lei *ICb +D *ICb *ICc) +D *V3lbr *ICc - D *
Ploss associated with reactive component of current is V3ler *ICb] + Rea[-CD *lei *ICa +D *lai *ICc +D *
Plossr LbranehesRaa *l�i + Rbb *lEi + Ree *
=
ICa *ICc) +D *V3ler *ICa - D *V3lar *ICc] (6)
11i + Rab *(-laJbi + V3larlbi - V3lailbr) + Rbe * The capacitor current IC that provides the maximum loss
(-lbJei + V3lbrlei - V3lbJer) + Rca * saving can be obtained by differentiating the equation 0 and is
(-leJai + V3lerlai - V3leJar) (3) given by
aPlosssaving "
alca
--.:..:.'-'.�L..branehesRaa *CD * 2lCa +D * 2lai) +Rab *
=

After placing the capacitor at a particular node, it draws a

reactive current Ie. The reactive component of branch [-CD dbi +D dCb) - D *V3lbr] + Rae *[-CD dei +D *
currents will be affected in the branch set oc. The current of ICc) +D *V3ler] 0 (7)
aPlosssaving "
=

other branches ((toc) is unaffected by the capacitor. The single --== - L..branehesRbb *CD * 2lCb +D * 2lbi) + Rba *
_

line diagram of 19 bus unbalanced distribution system is alcb

shown in Figure l. If the capacitor is placed at bus 7, then the [-CD *lai +D *ICa) +D *V3lar] +Rbe *[-CD *lei +D *
set oc consist of 1,3,5 and 6 branches. ICc) -D *V3ler] 0 (8)
aPlosssaving "
=

18 13
--=.:::.= - L..branehes Ree *CD * 2lCe +D * 2lei) + Reb *
_

alCc
17 12 [-CD *lbi +D *ICb) +D *V3lbr] + Rca *[-CD *lai +D *
3 7
16 13
II
ICa) - D *V3lar] 0 = (9)
17 14 Arranging the equations (-) in matrix form,
2 6 10 The capacitor currents for maximum loss savings cab be
calculated as
3 5 7 8 9
10
2 4 6 8 9

4
II (10)
18 14
12
5 19 15

15

16

Fig. 1. Single line diagram of 19 bus unbalanced RDS Here a is the set of branches from source bus to the jth
capacitor bus.

978-1-4673-8253-3/15/$31.00 ©2015 IEEE

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SILCHAR. Downloaded on October 02,2021 at 01:21:44 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of 2015 RAECS UlET Panjab University Chandigarh 21_22nd December 2015

MATRlX2 = �r
V3RabU) * IbrU) - V3RacU) * IcrU) - 2Raa(j) * lai(j) + Rab(j) * lbi(j) + Rac(j) * lci(j)
-V3Rba�) * lar�) + V3RbC�) * lcr�) + Rba�) * lai�) - 2Rb (j) * IbiU) + RbC�) * ICi�)
�
1
]
V3Rca(j) * lar(j) - V3Rcb(j) * lbr(j) + Rca(j) * lai(j) + Rcb(j) * lbi(j). - 2Rcc(j) * lci(j)
voltages for the 19-bus URDS before and after compensation
The corresponding capacitor size is, are given in Table l..
QCa lCa * Vam
= (11)
B. Case study 2: IEEE 25-bus URDS
QCb =QCc QCa= (12)
Here Vam is the voltage magnitude of phase-A at capacitor The proposed algorithm is also tested on a 4.l6-kV, 25-bus
bus. The above procedure can be repeated for all nodes to get URDS shown in Fig. 2. The line and load data are obtained
the highest possible loss saving for a single capacitor from [18]. The base load of the system in phases a, b, and c
placement and the node at which loss savings are maximum are 1073.3+j*792 kVA, 1083.3+j*800 kVA, 1083.3+j*800
will be considered as candidate node for capacitor placement kVA respectively. Highest loss savings obtained at bus 7 with
and the corresponding size (A phase only) is the optimal size corresponding size of 484.773kVAR in each phase. The total
of capacitor. active power losses reduced from 150.11795kW to
Algorithm: 109.6207kW, total reactive power losses reduced from
Step 1. Run the base case load flow and obtain the branch 167.27536kVAR to 121.4229kVAR and the minimum
currents. voltages in phases a, b, and c are improved from 0.9284,
Step 2. Select a bus and then find the maximum loss saving 0.9283, and 0.9365 p.u. to 0.9505, 0.9484, and 0.9573 p.u.
using equation (6) and the corresponding size by using after installing the capacitor banks. The voltage, the total
equation (11). active power loss (TPL) and reactive power 10ss(TQL),
Step 3. Repeat step 2 for all buses except at source bus and minimum voltages for the 25-bus URDS before and after
identify the bus at which the loss saving is highest. compensation are given in Tablel!.
Step 4. Compensate the selected bus with the 19

corresponding capacitor size. 21

20
TTT. RESULTS AND DISCUSSIONS
19
A. Case study I: 19-bus URDS 20 21 22 22
18
The load flow of unbalanced radial distribution system has 5
been implemented from [17]. The proposed algorithm is tested 2 2 4 4 6 5
on an ll-kV, 19-bus URDS. The line and load data are 3
obtained from [18]. The base load of the system in a-phase, b­
phase, c-phase are 126.33+j*61.23 kVA, 116.34+j*56.339
3 16 2
23
;1
24
24
25

12
kVA, 123.27+j*59.7 kVA respectively. Highest loss savings 8 10 9 1:\ 10 14 11 15 12
6
obtained at bus 10 with corresponding size of 23.1517 kVAr 7
in each phase. The total active power losses reduced from 9 11 16
13.4709 kW to 11.5281 kW, total reactive power losses 18
14 17 13
8
reduced from 5.7956 kVAr to 4.9597 kVAr and the minimum 17
voltages in phases a, b, and c are improved from 0.9515,
0.9494, and 0.9504 p.u. to 0.9560, 0.9542, and 0.9549 p.u. 15

after installing the capacitor banks. The voltage, the total Fig. 2. Single line diagram of 25 bus unbalanced RDS.
active loss (TPL) and reactive power loss (TQL), minimum
Table T . TEEE19 bus unbaIanced rad·JaI system
Node Before Capacitor Placement After Capacitor Placement
Va (p.u) Vb (p.u) Vc (p.u) Va (p.u) Vb (p.u) Vc (P.t!)
1 1 1 1 1 1 1
2 0.98746 0.98910 0.98798 0.98829 0.98994 0.98881
3 0.98542 0.98869 0.98633 0.98625 0.98952 0.98717
4 0.98235 0.98390 0.98301 0.98360 0.98515 0.98425
5 0.98201 0.98366 0.98283 0.98325 0.98491 0.98408
6 0.97928 0.98078 0.98005 0.9808 0.98231 0.98158
7 0.97861 0.98029 0.97957 0.98013 0.98182 0.98109
8 0.97282 0.97382 0.97347 0.97502 0.97604 0.97569
9 0.96592 0.96598 0.96575 0.96895 0.96903 0.96880
10 0.95626 0.95549 0.95501 0.96067 0.95993 0.95944
11 0.95499 0.95430 0.95331 0.95942 0.95874 0.95774
12 0.95478 0.95377 0.95358 0.95920 0.95822 0.95802
13 0.9544 0.95344 0.95211 0.95882 0.95788 0.95655
14 0.95449 0.95388 0.95283 0.95891 0.95833 0.95727

978-1-4673-8253-3/15/$31.00 ©2015 IEEE

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SILCHAR. Downloaded on October 02,2021 at 01:21:44 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of 2015 RAECS UlET Panjab University Chandigarh 21_22nd December 2015

15 0.95275 0.95122 0.95126 0.95718 0.95568 0.95571

16 0.95339 0.95148 0.95218 0.95782 0.95593 0.95662
17 0.95366 0.95337 0.95232 0.95808 0.95782 0.95677
18 0.95380 0.95319 0.95209 0.95822 0.95764 0.95654
19 0.95160 0.94976 0.95047 0.95603 0.95423 0.95492

Min V 0.95160 0.94976 0.95047 0.95603 0.95423 0.95492

Capacitor Size - - -
23.1517 23.1517 23.1517
(kVAr)
location 10
Active power loss 4.4539 4.4532 4.5637 3.8179 3.7969 3.9133
(kW)
Active power loss - - -
14.279 14.738 14.252
reduction (%)
Reactive power loss 1.9405 1.8964 1.9587 1.6724 1.6165 1.6708
(kVAr)
Reactive power loss - - -
13.816 14.76 14.70
reduction (%)
TPL (kW) 13.4709 1l.5281
TQL (kVAr) 5.79562 4.9597

Table IT ResuIts f,or IEEE 25 bus unbaIanced rad·JaI sys em

Node Before Capacitor Placement After Capacitor Placement
Va (p.u) Vb (p.u) Vc (p.u) Va (p.u) Vb (p.u) Vc (p.u)
1 1 1 1 1 1 1
2 0.97020 0.9711 0.97545 0.97958 0.97939 0.98418
3 0.96323 0.96444 0.96984 0.97267 0.97278 0.97862
4 0.95978 0.96129 0.96739 0.96926 0.96965 0.97619
5 0.95872 0.96025 0.96644 0.96821 0.96862 0.97525
6 0.95498 0.95587 0.96148 0.97056 0.96987 0.97606
7 0.94191 0.94283 0.94923 0.96373 0.96256 0.96971
8 0.95286 0.95379 0.95957 0.96847 0.96781 0.97418
9 0.93589 0.93668 0.94379 0.95785 0.95654 0.96439
10 0.93150 0.93186 0.93953 0.95356 0.95182 0.96022
11 0.92941 0.92963 0.93763 0.95153 0.94963 0.95836
12 0.92841 0.92840 0.93660 0.95055 0.94842 0.95735
13 0.92871 0.92872 0.93682 0.95084 0.94874 0.95757
14 0.93594 0.93699 0.94338 0.95790 0.95685 0.96399
15 0.93377 0.93487 0.94144 0.95578 0.95476 0.96209
16 0.94083 0.94177 0.94826 0.96268 0.96153 0.96876
17 0.93473 0.93595 0.94203 0.95672 0.95582 0.96267
18 0.95732 0.95864 0.96432 0.96682 0.96702 0.97315
19 0.95241 0.95443 0.95998 0.96196 0.96285 0.96885
20 0.95482 0.95634 0.96201 0.96435 0.96475 0.97087
21 0.95379 0.95487 0.96053 0.96333 0.96328 0.9694
22 0.95184 0.95246 0.95852 0.9614 0.96090 0.96740
23 0.95647 0.95838 0.96479 0.96598 0.96677 0.97361
24 0.95444 0.95651 0.96311 0.96397 0.96492 0.97195
25 0.95202 0.95469 0.96117 0.96158 0.96311 0.97003
Min V 0.92841 0.92840 0.93660 0.95055 0.94842 0.95735
Capacitor
" Size 484.773 484.773 484.773
(kVAr) - - -

location 7
Active power loss
(kW) 52.8133 55.4431 41.8615 38.6027 40.4463 30.5715
Active power loss
reduction (%) - - -
26.907 27.049 26.969
Reactive power loss 58.2902 53.2941 55.6911 42.3424 38.6939 40.3866

978-1-4673-8253-3/15/$31.00 ©2015 IEEE

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SILCHAR. Downloaded on October 02,2021 at 01:21:44 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of 20l5 RAECS UlET Panjab University Chandigarh 2l_22nd December 2015

(kVAr)
Reactive power loss
reduction (%) - - - 27.359 27.396 27.481
TPL (kW) 150.11795 109.6207
TQL (kVAr) 167.27536 121.4229
[7] Haque MH. Capacitor placement III radial dlstnbutJon systems for loss
reduction. lEE Proc- Gener Trans and Distrib 1999; 146(5): 50 1-505.
[8] EI-Fergany AA Abdelaziz A Y. Capacitor allocations in radial
IV. CONCLUSIONS
,

distribution networks using cuckoo search algorithm. lET Gener Trans

A simple method for optimal placement and sIZIng of and Distrib 2014; 8(2): 223-232.
[9] Crispino M, Di Vito V, Russo A, Varilone P. Decision theory criteria for
capacitor while minimizing power losses and improving the
capacitor placement in unbalanced distribution systems. 2005 IEEE/PES
voltage profile in an unbalanced radial distribution system has Transmission and Distribution Conference & Exhibition; 2005; Dalian,
been presented in this paper. The method first fmds the China: IEEE. pp. 1-6.
possible loss savings by placing optimal capacitor size in each [ 10] Carpinelli G, Noce C, Proto D, Russo A, Varilone P. Single-objective
probabilistic optimal allocation of capacitors in unbalanced distribution
phase at every node (except source node). The node at which
systems. Elect Power Syst Res 20 12; 87: 47-57.
the loss savings are maximum will be taken as candidate node [II] Eajal AA El-Hawary ME. Optimal capacitor placement and sizing in
,

and the corresponding capacitor size is the optimal size. unbalanced distribution systems with harmonics consideration using
Proposed method has been implemented on two case studies particle swarm optimization. IEEE Trans on Power Del 20 10; 25(3):
1734-1741.
i.e IEEE 19-bus UBRS and IEEE 25-UBRS. The result shows
[ 12] Carpinelli G, Varilone P, Di Vito V, Abur A. Capacitor placement in
that with the help of single capacitor placement in an three-phase distribution systems with nonlinear and unbalanced loads.
unbalanced radial distribution system, the voltage profile has lEE Proc- Gener Trans and Distrib 2005; 152( 1): 47-52.
been improved and the burden on the substation has been [ 13] Subrahmanyam JBV, Radhakrishna C. A simple method for optimal
capacitor placement in unbalanced radial distribution system. Elect
reduced due to reduction in losses.
Power Comp and Syst 20 10; 38( 1 1): 1269-1284.
[ 14] Mohkami H, Hooshmand R, Khodabakhshian A. Fuzzy optimal
References placement of capacitors in the presence of nonlinear loads in unbalanced
distribution networks using BF -PSO algorithm. Applied Soft
[I] Esmailnezhad B, Shayeghi H. Simultaneous distribution network Computing 20 1 1; 1 1: 3634-3642.
reconfiguration and DG allocation for loss reduction by invasive weed [ 15] Taher SA, Bagherpour R. A new approach for optimal capacitor
optimization algorithm. In: Smart Grid Conference; 17- 18 Dec 20 13; placement and sizing in unbalanced distorted distribution systems using
Tehran, Iran: IEEE. pp. 166-172.
hybrid honey bee colony algorithm. Elect Power and Energy Syst 20 13;
[2] Ng HN, Salama MMA, Chikhani AY. Capacitor allocation by 49: 430-448.
approximate reasoning: fuzzy capacitor placement. IEEE Trans on [ 16] Murthy KVSR, Raju MR. Capacitive compensation on three phase
Power Del 2000; 15( 1): 393-398. unbalanced radial distribution system using index vector method.
[3] Huang YC, Yang HT, Huang CL. Solving the capacitor placement International Journal of Engineering and Technology 20 12; 2(2): 284-
problem in a radial distribution system using tabu search approach. IEEE 29 1.
Trans on Power Syst 1996; 1 1(4): 1868-1873.
[ 17] Teng JH. A direct approach for distribution system load flow solutions.
[4] Xu Y, Dong ZY, Wong KP, Liu E. Optimal capacitor placement to IEEE Trans on Power Del 2003; 18(3): 882-887.
distribution transformers for power loss reduction in radial distribution [ 18] Ganesh V, Sivanagaraju S, Ramana T. Feeder reconfiguration for loss
systems. IEEE Trans on Power Syst 20 13; 28(4): 4072-4079. reduction in unbalanced distribution system using genetic algorithm.
[5] Baran ME, Wu FF. Optimal capacitor placement on radial distribution World Academy of Science, Engineering and Technology 2009; 3(4):
systems. IEEE Trans on Power Del 1989; 4(I): 725-734. 626-634.
[6] Huang SJ. An immune-based optimization method to capacitor [ 19] McDermott TE, Drezga I, Broadwater RP. A heuristic nonlinear
placement in a radial distribution system. IEEE Trans on Power Del constructive method for distribution system reconfiguration. IEEE Trans
2000; 15(2): 744-749. on Power Syst 1999; 14(2): 478-483.

978-1-4673-8253-3/15/$31.00 ©2015 IEEE

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SILCHAR. Downloaded on October 02,2021 at 01:21:44 UTC from IEEE Xplore. Restrictions apply.