Please do not remove this page

Power Quality Improvement of Distribution

Network Using BESS and Capacitor Bank
Islam, F. R.; Lallu, A.; Mamun, K. A.; et.al.
https://research.usc.edu.au/discovery/delivery/61USC_INST:ResearchRepository/12148469750002621?l#13149199370002621

Islam, Lallu, A., Mamun, K. A., Prakash, K., & Roy, N. K. (2021). Power Quality Improvement of Distribution
Network Using BESS and Capacitor Bank. Journal of Modern Power Systems and Clean Energy, 9(3),
625–632. https://doi.org/10.35833/MPCE.2019.000304
Document Type: Published Version

Link to Published Version: https://doi.org/10.35833/MPCE.2019.000304

USC Research Bank: https://research.usc.edu.au

[email protected]
CC BY V4.0
This article is distributed under the terms of the Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/).
Downloaded On 2022/12/18 05:45:21 +1000

Please do not remove this page

JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, VOL. 9, NO. 3, May 2021 625

Power Quality Improvement of Distribution

Network Using BESS and Capacitor Bank
F. R. Islam, A. Lallu, K. A. Mamun, K. Prakash, and N. K. Roy

Abstract—
—The power demand around the world is increasing rent, phase angle and impedance.
rapidly. The aging distribution network architectures are used Placing compensators at the reactive load center can elimi‐
by the existing utility companies to deliver power to the con‐
nate the reactive load of distribution systems, and the reac‐
sumers, which significantly affects the reliability, stability and
quality of the delivered power. Different techniques such as tive power delivered from the generator will be reduced. The
compensation devices have been used by power system engi‐ active power loads will be served if the generator is avail‐
neers and researchers to maintain the quality of power trans‐ able at full capacity [1]. The reactive load that lags current
mitted to end users. In this paper, wattage and volt-amp reac‐ demand can be effectively eliminated by the leading current
tive (VAR) planning scheme has been proposed by using the
which can be provided by the capacitors. The compensator
combination of battery energy storage systems (BESS) and com‐
pensators to deal with the vulnerability of networks to voltage will be able to provide support for the reactive load and volt‐
drop and system inefficiency. The cost-effective combination of age improvement, while the performance of the compensator
BESS and shunt capacitor bank will then be analyzed to indi‐ will depend upon the location of the load center [2]. In addi‐
cate the benefit of the proposed scheme. tion, the harmonic distortion reduces the efficiency of the en‐
Index Terms——Battery energy storage system (BESS), shunt ergy transfer or power factor (PF) [3], [4].
capacitor bank, reliability, stability, quality, volt-amp reactive In distribution systems, the structure and configuration of
(VAR) planning, distribution network. networks can differ significantly depending on the type of
loads that define whether the electric power supplied to the
end user will be medium or low voltage (LV). For instance,
I. INTRODUCTION in rural areas, the overhead lines with short interconnection

E LECTRICAL utility companies around the world are un‐

der enormous societal and political pressure while de‐
signing more efficient electric power grids that implement
configuration are used, whereas in the urban areas, many lat‐
eral connections are used for alternative route of supplies
[5]. In this paper, a VAR planning scheme will be carried
strategies to reduce energy losses. One of the strategies is to out by power flow analysis to determine the optimal place‐
reduce the amount of reactive power supplied to the loads ment of battery energy storage system (BESS) and capacitor
through transmission and distribution lines. In others words, in four different distribution networks, i.e., radial, ring, mesh
the higher the reactive power demand, the less efficient the and aromatic [5], using Electrical Transient and Analysis Pro‐
power grids become for the utilities. gram (ETAP) software.
In recent years, volt-amp reactive (VAR) compensators
such as static synchronous compensator (STATCOM) and se‐ II. ECONOMIC ANALYSIS OF VAR PLANNING
ries/shunt capacitors have been used to compensate power
Several factors can affect the active and reactive energy
on the power grid. These compensators provide faster time
prices. Therefore, accurate cost-based pricing is highly diffi‐
response and are quite useful as they inject and/or absorb ac‐
cult, and there is a lack of appropriate exact payment exam‐
tive and reactive power simultaneously to support loads rath‐
ples internationally. The following chargeable indicators can
er than overload the main lines. The implementation of such
be used to categorize the applied reactive energy payment
devices helps increase the apparent power through transmis‐
techniques around the world: ① peak demand of apparent
sion lines and improves the stability in the network by allow‐
power; ② apparent energy consumption in kVAh; ③ peak
ing adjustment of parameters, i. e., frequency, voltage, cur‐
demand in active and reactive power; ④ energy consump‐
tion and PF; ⑤ adjusting active power or active energy bills.
Manuscript received: May 7, 2019; accepted: November 26, 2019. Date of Several researches have been conducted and discussions
CrossCheck: November 26, 2019. Date of online publication: April 29, 2020.
This article is distributed under the terms of the Creative Commons Attribu‐ have been made to optimize the capacity of distributed gen‐
tion 4.0 International License (http://creativecommons.org/licenses/by/4.0/). eration (DG) [6]. For instance, genetic algorithm (GA) based
F. R. Islam (corresponding author) is with University of Sunshine Coast, Sip‐ optimization model for a standalone photovoltaic (PV)-wind
py Downs, Australia (e-mail: [email protected]).
A. Lallu, K. A. Mamun, and K. Prakash are with University of South Pacific, hybrid system including PV battery chargers has been devel‐
Suva, Fiji (e-mail: avneellallu@gmail. com; kabir. mamun@usp. ac. fj; krishneel. oped in [7], [8]. For larger-scale isolated and grid-connected
[email protected]). wind-solar-storage hybrid systems, the optimal DG size and
N. K. Roy is with Khulna University of Engineering and Technology, Khulna,
Bangladesh (e-mail: [email protected]). BESS capacity have been addressed in [9]. In [10], the opti‐
DOI: 10.35833/MPCE.2019.000304 mal placement of BESS is evaluated using GA for minimiz‐
626 JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, VOL. 9, NO. 3, May 2021

ing net present values related to power losses to determine es and installation of larger battery capacity with shunt ca‐
the best operation cost. The STATCOM BESS is studied for pacitor are highlighted.
nonlinear load and hysteresis controller for the grid opera‐ This research will highlight and compare the existing and
tion in [11]. Reference [12] discusses the uses of ultra-capac‐ patent network structure called aromatic network [5]. The
itor and BESS with shunt capacitor that has been recently IEEE 13-bus network has been modified to form the radial,
used as auxiliary devices for large-scale PV generator sys‐ ring and mesh networks. The single-line diagram for radial,
tem. In [13], BESS and capacitor are discussed and support‐ ring and mesh networks is shown in Fig. 1 and distinguished
ed by dynamic voltage restorer (DVR) integrated with ultra- in Table I. The dichloro-diphenyl-trichloroethane (DDT)
capacitor-based rechargeable energy storage to improve its structured aromatic network is shown in Fig. 2.
voltage restoration capabilities. In [14], the cost, system loss‐

Open switch BESS station 2

Utility
T2
Closed switch
Line 2 Line 1 S9 Line 3 Line 4 500 kVA

Bus 646
Bus 645
Bus 632 Bus 633 S5 Bus 634 S7
S3 S10
S1 Motor T5
Line 12 Line 13 500 kVA
Load 1 Lump 9 Lump 1
Gen 1 S4 S6 Gen 2
Line 11 0.5 MW 0.5 MW Line 14
Line 7 Line 5 Line 9 Line 10
S2
S11 S8
Bus 671
Bus 685 Bus 684 Bus 692 Bus 675

Line 6 Line
Bus 652 8
Lump 6
Lump 3 Lump 7
Lump 8 Lump 4
BESS BESS
Bus 680
Load 2 station 3 station 1
Fig. 1. Single-line diagram of IEEE 13-bus network connected to radial, ring and mesh networks.

TABLE I
SWITCHES FOR IEEE 13-BUS NETWORK

Switch status
Network
S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11
Radial Open Open Open Open Open Open Open Open Closed Closed Closed
Ring Closed Closed Open Open Open Open Closed Closed Closed Open Open
Mesh Closed Closed Closed Closed Closed Closed Closed Closed Closed Open Open

Utility Underground line

Bus 101 Bus 107
Bus 100 Overhead line
Bus 102 Bus 106 Bus 108 Bus 112
Lump Cap 2 Gen 2
Gen 1 Motor Lump
Lump

Cap 1

Load Load Lump

Bus 103 Bus 105 Bus 109 Bus 110

Bus 104 Bus 111
Lump Lump
Fig. 2. DDT structured aromatic network.

A. Radial Network Architecture distribution system is radial network, which uses only one
The most common type of distribution network for power path between each customer and the substation. The radial
network topology has a tree-shape structure, where closed
ISLAM et al.: POWER QUALITY IMPROVEMENT OF DISTRIBUTION NETWORK USING BESS AND CAPACITOR BANK 627

loop does not exist. Therefore, the power can be delivered fast power charging station [19]. According to [20], in
from one bus to another without tracking down the original ETAP, the optimal location of capacitors is defined using the
bus. However, the power flow need find the original bus loss sensitivity factor (LSF). The active power loss Ploss in
while turning backwards [1], [15]. the network of branch m is given by:
B. Ring Network Architecture r m ( P m2 + Q 2m )
P loss = (1)
An alternative to the design of a purely radial feeder is V m2
the ring network, which has two paths between the power where rm is the resistance in branch m; V m is the voltage pro‐
sources (substations and service transformers) [16]. The file of branch m; and Pm and Qm are the active and reactive
equipment is sized so that the service can be maintained power drawn from bus m, respectively.
with a single fault. Concerning the complexity of this net‐ The LSF of the network branches and the net system loss
work, a loop feeder system is slightly more complex than a of the active power TPloss in the network can be identified us‐
radial network [17]. ing:
C. Mesh Network Architecture ¶P loss 2Q m r m
LSF = = (2)
A mesh network follows the radial structure, but includes ¶Q loss V m2
redundant distribution lines in addition to the main lines to rm
act as the backup routes during failures or faults. In a mesh Wm = (3)
network, a wide range of power transfer paths is available, V m2

TP loss = ∑W m ( P m2 + Q 2m )
which guarantees significant flexibility in the event of re‐ N

quired maintenance or a fault on part of the system [15]. (4)

m= 1

D. Aromatic Network Architecture where N is the number of branches; Qloss is the reactive pow‐
Recently, a novel smart distribution network, i.e., aromatic er loss; and Wm is the the complex magnitude of branch m.
network, has been developed [5], [18]. The design of this The net active power loss TP loss
cap
after the optimal installa‐
network follows the structure of aromatic molecule DDT tion of capacitors in the network is given by:
that comprises of two benzene networks connecting each oth‐
er through the slack bus as shown in Fig. 2. Each benzene
network has six nodes and each node has connections of sin‐
TP loss
cap
= ∑W
m Î B cap
m
é 2
ë
(
êP m + Q 2m - ∑B mk Q cap
z

k=1
k
ù
ú+
û
) (5)
gle and double bonded wiring. The single bond represents ∑ W m ( P m2 + Q 2m )
m Ï B cap
the overhead lines, whereas the double bonds represent the

{
combination of overhead and underground wiring. The sin‐ 1 Q ck at the k th node flows through m
gle and double bonded wiring of the aromatic network pro‐ B mk = (6)
0 otherwise
vides self-healing characteristics.
where m ∈ B cap depicts that branch m is for Bcap; Bcap is the
III. PF PROBLEMS capacitor bank branch; Q cap k is the reactive power of bus k; z
is the number of capacitors; B mk is the element of a binary
In rural power distribution systems, the PF usually drops
matrix B mk with the dimension of B cap ´ z; and Q ck is the reac‐
as the demand increases. With an increase in the load and
tive power of capacitor at branch k.
decrease in the PF, the following problems may arise [1],
The net active power loss saved after optimal installation
[16]: ① voltage fluctuation; ② the increase of system loss‐
of capacitors Q cap
k opt in the network is given by:
es; ③ PF penalties to the customers for high reactive power
usage; ④ the reduction of system capacity. DTP loss = TP loss - TP loss =
cap

é ù
( )
2

∑ m∑ mk k ∑
PF is defined as the ratio between active power (W) and z z
cap ú
W ê
mê 2Q B Q cap
- B Q (7)
total power (VA). To improve the system PF, battery storage mk k ú
m Î B cap
ë k=1 k=1
û
systems and capacitors can also offer voltage drop correction
with leading current of capacitor. The large amount of induc‐ Differentiating (7) with respect to Q cap i at bus i will give
tive VAR current requires bulk power facilities carried to the the maximum active power loss as:

( )
distribution system, which leads to losses on the bulk facili‐
∑ ∑
¶DTP loss é z
ù
ties and introduces unnecessary cost. =2
ê B mW m - Q m - B mk Q cap
k ú i Î z (8)
¶Q i
cap
mÎ B ë cap k=1 û
IV. CAPACITOR SIZING AND PLACEMENT FOR LOSS The net maximum active power loss saved at the first dif‐
REDUCTION ferentiation equals zero:
The optimal placement of capacitor is generally a hard- ¶DTP loss ||
= 0 = ∑ B miW m∑B mk Q cap k opt = ∑ B mW m Q m
z

combinatorial optimization problem that can be formulated |

¶Q i | Q = Q
cap
cap mÎ B
cap k 1 cap mÎ B = cap
as a nonlinear multi-objective problem. A method to reduce k k opt

the probability of overload in local distribution power net‐ (9)

work has been achieved by offsetting the electricity transmis‐ A matrix representation of the sizes of capacitors at multi‐
sion congestion and employing local energy storage item in ple locations in a network is given in (10), or in an expand‐
628 JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, VOL. 9, NO. 3, May 2021

ed form in (11): following should be taken into consideration: ① bank size;

opt ] Z ´ 1 =[X ] Z ´ Z [Y ] Z ´ 1
[Q cap -1
(10) ② the location for connection; ③ the connection type of the
cap
transformer (Y or Δ).
éQ ù1opt é X 11 X 12  X 1Z ù é Y 1 ù
ê cap ú êX
êê Q 2opt úú ê 21 X 22  X 2Z úú êê Y 2 úú Start
êê ú= ê   úê  ú
(11)
 ú ê

úê ú
ê cap ú X ZZ û ëY Z û
Perform load flow and PQ analysis
ëQ Zopt û ëX Z1 X Z2 
where Z is a positive integer. The elements of [X ] and [Y ] in Choose a combination of shunt capacitor
(10) are calculated as: X h = ∑ B mg W m Q mh, Y h = bank and battery bank size
m Î B cap

∑ B mh W m Qm, where Bmg (or Bmh) is set to 1 if Q cg (or Q ch) at

m Î B cap
Perform load flow, capacitor placement,
battery bank size again
Increase
the gth (or hth) node flows through m; otherwise, it is set to battery
0. Q cg and Q ch are the reactive power of capacitor at branch g size
0.98 p.u.<Vbus<1.02 p.u., N
and h, respectively. PF≥0.85, and do losses decrease?
Hence, by using this LSF method, the amount of capacitor
Change
bank has been distributed over each network based on the
capacity value Y
amount of active and reactive losses. The capacitor sizing
can be done using: N
Do VAR losses decrease?
Qc = Vm Ic (12)
where Qc is the capacitor size; and Ic is the capacitor current. Y
Based on a given set of battery characteristic curves avail‐ Print economical cost-effective shunt
capacity bank and battery bank size
able in the ETAP library, the required battery size for the
specified duty cycle can be determined using:
End
WHT
Wh = (13)
7 Fig. 3. VAR planning scheme for optimal placement of compensators.

Wh
Ah = (14) Tables II and III present the load data of the aforemen‐
0.8V
where Wh is the total energy required; W is the total load; H tioned distribution networks. The bus numbers identified
is the hours of usage; T is the number of days; Ah is the bat‐ with asterisk * have PFs which are unreliable or within un‐
tery capacity; and V is the system voltage. stable limits.
TABLE II
V. OPTIMUM PLACEMENT OF BESS AND CAPACITOR IEEE 13-BUS LOAD DATA FOR RADIAL, RING AND MESH NETWORKS

The battery size has been determined for low-level pene‐

Active power Reactive power
tration of battery energy storage supply for three different Bus No. Bus ID
(kW) (kvar)
PF
buses, which are considered as charging stations or BESS 1 - - - -
near heavily loaded nodes around the distribution networks. 2 632 100 50 87
Each DC station can supply 1%, 2% and 3% of the total ca‐
3 645 183 79 92
pacity of the networks, respectively.
4 646 230 132 87
The BESS stations are installed near high-dense distribut‐
5 633 - - -
ed loads. The battery capacity should satisfy the minimum
6 634* 400 290 81
voltage requirements to supply DC loads and be converted
to AC power. The requirements include: ① the charging volt‐ 7 671 1254 317 97
age applied to the battery should not be more than the maxi‐ 8 684 - - -
mum system voltage; ② the discharging battery voltage 9 652* 128 86 83
should not be less than the minimum system voltage. 10 685 170 80 90
The permanent locations of capacitors should be 1/2 to 11 680 - - -
2/3 of the total length of the line from the substation (consid‐ 12 692* 170 151 75
ering “rule of thumb”) to obtain the maximum benefits of 13 675 843 462 88
quality improvement and loss reduction due to uniformly dis‐ Total 3478 1647
tributed loads. The main constraints for the placement of ca‐
pacitor have been explained in Fig. 3 [21]. The constraints
are as follows: ① the PF should be greater than a threshold; The objective of optimal placement of battery storage and
② all voltage magnitudes of load (PQ) buses should be with‐ capacitor is to minimize the cost of the system. The cost in‐
in the lower and upper bars; ③ the minimum PF bar is met. cludes five parts for 10-year period that are listed as fol‐
To place BESS and capacitor banks in power systems, the lows: ① the energy cost of electric vehicle (EV) Li-ion bat‐
ISLAM et al.: POWER QUALITY IMPROVEMENT OF DISTRIBUTION NETWORK USING BESS AND CAPACITOR BANK 629

tery is 150 $/kWh; ② the installation cost of capacitor is 0.83 lagging, the expected value is 0.85 lagging).
$3000; ③ the purchase cost of capacitor is 350 $/kvar; ④ Table IV shows different penetration levels of battery re‐
the operation (maintenance and depreciation) cost of capaci‐ quired to support 24 hours for different capacity levels of
tor bank is 300 $ per bank year; ⑤ the cost of active power the power grid.
losses is 0.16 $/kWh.
TABLE IV
TABLE III BATTERY ENERGY STORAGE WITH DIFFERENT PENETRATION LEVELS
LOAD DATA FOR AROMATIC BUS NETWORKS
Radial, ring and mesh Aromatic network
Storage capacity (%)
Active power Reactive power network (kWh/day) (kWh/day)
Bus No. Bus ID PF
(kW) (kvar) 1 834.72 835.44
1 100 - - - 2 1669.44 1670.88
2 106* 170 151 75 3 2504.16 2506.32
3 101 183 79 92
4 102 170 80 90
Considering 1% of the total demand (wattage), i. e., 3478
5 103 - - -
kW of active power demand, an intermittent supply of 34.78
6 104 100 58 87 kWh battery energy storage per hour has been installed for
7 105* 128 86 83 radial, ring and mesh networks. Similarly, for the aromatic
8 108 100 58 87 network, the active energy storage has been sized to 1% of
9 107 - - - the total capacity, i. e., 3481 kW of active power demand,
10 112 843 462 88 and an intermittent supply of battery with the capacity of
11 111* 400 290 81 34.81 kWh per hour has been installed.
12 110 1157 659 87 Table V demonstrates that the ring network has the high‐
13 109 230 132 87 est investment cost. However, over a 10-year analysis, the
Total 3481 2055 ring network provides benefits to increase energy efficiency
and profit by reducing the loss. The Li-ion batteries [22]
used in this application are EV batteries with an average
The flowchart highlights the two types of constraints cost of 150 $/kWh [23] to power the loads. Three test cases
which have been considered for the lower and upper percent‐ with up to 3% of energy storage penetration into the net‐
age limits in the following area: ① voltage range (98%- works are carried out for the batteries.
102%, the expected value is 100%); ② PF range (more than
TABLE V
SUMMARY OF ECONOMIC ANALYSIS FOR ALL NETWORK STRUCTURES

Minimum voltage Maximum voltage Battery and capacitor Operation cost

Penetration level Network structure Benefit ($/year)
(%) (%) cost ($/year) ($/year)
Radial 91.25 98.36 - - -
Before battery and capacitor Ring 86.86 98.09 - - -
placement Mesh 93.51 98.44 - - -
DDT 91.06 101.35 - - -
Radial 99.60 101.15 513312 568835 596366
With 3% battery and 50% Ring 99.42 101.52 1004624 321365 1657150
capacitor Mesh 99.76 101.01 495499 491481 434348
DDT 99.55 100.00 481187 453433 742566
Radial 99.79 101.39 650999 568835 90536
With 3% battery and 100% Ring 99.58 101.73 1633624 321365 1028150
capacitor Mesh 99.96 101.15 615374 491481 314473
DDT 99.83 100.52 586749 453433 637003

VI. SIMULATION RESULTS back period of around 2 years using 50% of the required ca‐
pacitor demand for the buses, whereas the radial, ring and
From the comparison of the networks, all networks with‐
mesh networks will take 3.5 years to clear up the initial pay‐
out any compensators have huge line losses, as each line is ment into profits.
loaded with constant current load and impedance load. After In addition, the aromatic network is the most resilient net‐
performing the load flow analysis and optimal placement of work. To deal with losses, this network is efficient and re‐
capacitor, and applying the VAR planning scheme, the aro‐ quires less operation cost and the lowest need of compensa‐
matic network has demonstrated an economical benefit/pay‐ tors. This is demonstrated in Figs. 4 and 5. The ring network
630 JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, VOL. 9, NO. 3, May 2021

has the highest capital investment and will return with the which is the lowest in the four networks.
most benefits, generate more profit than other networks in Year
the 10-year period of investment. The comparison of power 10
losses before and after battery placement and capacitor 1.0 9
8
banks for all distribution networks are shown in Table VI. 7
Year 6
2.0 10 0.5 5
9 4
8 3
1.5
7 2

Cost (M$)
6 1
5 0
1.0
Cost (M$)

4
3
0.5 2
1 -0.5

-1.0
-0.5 DDT Mesh Radial Ring
DDT Mesh Radial Ring
Network
Network
Fig. 5. Accumulative profit in 10 years with 100% capacitor and 3% bat‐
Fig. 4. Accumulative profit in 10 years with 50% capacitor and 3% battery. tery.

Based on the power losses in the system for all four net‐ The proposed VAR planning scheme reflects how the com‐
works using battery energy storage and capacitor banks for bination of 50% capacitor and 3% of battery will generate
penetration of 0%, 1%, 2% and 3% of energy and with 0%, less losses and more benefits in terms of revenue based on
50% and 100% of reactive power, it can be concluded that the 10-year reliability of the system, which has been summa‐
an excessive amount of battery and capacitor penetration has rized and presented in Fig. 6 and Fig. 7, respectively. It is
less impact to reduce losses, but increases the investment observed that the voltages of the networks increase when the
factor, and the system becomes unstable. system losses decrease and the power system becomes more
The most significant impact of battery penetration is ob‐ reliable. When the load in the lines decreases, the amount of
served with ring network of around 86% in loss reduction in power transmitted to the consumer in each phase has less
active and reactive power with only 1% battery penetration. voltage drop, which causes the voltage to rise towards the
However, with 1% implementation of battery energy storage desired target.
on the mesh network, the system reduces 67% of losses,
TABLE VI
ENERGY LOSSES FOR DIFFERENT PENETRATION LEVELS IN RADIAL, RING, MESH AND DDT NETWORKS

Active power loss (kW) Reactive power loss (kW)

Battery (%) Capacitor (%)
Radial Ring Mesh DDT Radial Ring Mesh DDT
0 159.6 262.6 136.3 119.9 470.7 705.6 404.9 363.1
0 50 114.2 174.4 99.8 75.5 337.1 468.8 297.1 228.0
100 112.5 191.6 97.8 79.4 332.8 517.3 289.3 239.2
0 43.0 37.6 44.6 159.5 117.7 100.1 124.0 494.5
1 50 36.6 32.5 39.4 115.1 99.4 84.6 108.8 359.3
100 37.5 38.8 42.1 119.0 100.7 98.4 113.4 370.5
0 37.7 34.6 37.9 61.2 103.0 91.3 104.9 194.0
2 50 35.1 31.7 35.4 59.7 93.2 81.1 96.2 188.0
100 36.4 34.5 36.2 61.2 93.8 86.4 96.7 191.5
0 34.2 31.2 36.7 61.6 89.7 75.7 96.6 193.7
3 50 33.0 30.2 35.6 59.5 83.8 70.6 91.5 187.7
100 34.1 31.6 36.8 61.6 83.3 70.7 91.7 191.2

VII. CONCLUSION itor banks in the networks has been proved to be beneficial
The dynamic analysis of stability and economic factors us‐ for the networks. The capacitors compensate VAR power for
ing a new planning scheme combination of BESS and capac‐ the excessive reactive loads in the network, while the battery
banks assist to deliver active power to the network from dif‐
ISLAM et al.: POWER QUALITY IMPROVEMENT OF DISTRIBUTION NETWORK USING BESS AND CAPACITOR BANK 631

ferent stations, which reduces the load of line.

Active power loss (kW) 300
250 Radial
Ring
200 Mesh
150 DDT
100
50
0
0% capacitor 100% capacitor 50% capacitor 0% capacitor 100% capacitor 50% capacitor
with 0% battery with 0% battery with 1% battery with 2% battery with 2% battery with 3% battery
50% capacitor 0% capacitor 100% capacitor 50% capacitor 0% capacitor 100% capacitor
with 0% battery with 1% battery with 1% battery with 2% battery with 3% battery with 3% battery
Battery and capacitor combination
Fig. 6. Active power losses for all networks.

800
Reactive power loss (kvar)

700 Radial
600 Ring
500 Mesh
400 DDT
300
200
100
0
0% capacitor 100% capacitor 50% capacitor 0% capacitor 100% capacitor 50% capacitor
with 0% battery with 0% battery with 1% battery with 2% battery with 2% battery with 3% battery
50% capacitor 0% capacitor 100% capacitor 50% capacitor 0% capacitor 100% capacitor
with 0% battery with 1% battery with 1% battery with 2% battery with 3% battery with 3% battery
Battery and capacitor combination
Fig. 7. Reactive power losses for all networks.

The most efficient and cost-effective is the aromatic struc‐ REFERENCES

ture network using 50% capacitor placement and 3% BESS [1] M. Balci and M. Hocaoglu, “Effects of source voltage harmonic distor‐
combination. Aromatic network structure demonstrates the tion on power factor compensation in triac controlled AC chopper cir‐
quickest benefit/payback period of around 2 years using cuits,” in Proceedings of International Conference on Power Electron‐
ics and Drives Systems, Kuala Lumpur, Malaysia, Dec. 2005, pp. 1199-
50% of the required capacitor demand for the buses, while 1204.
radial, ring and mesh networks will take 3.5 years to clear [2] T. Gonen, Electrical Power Transmission System Engineering, 3rd
up the payback. edn. Hoboken: CRC Press, 2015.
As the power networks are expanding around the world, [3] A. Anand and B. Singh, “Modified dual output cuk converter-fed
switched reluctance motor drive with power factor correction,” IEEE
this research work aligns with the engineering solution of Transactions on Power Electronics, vol. 34, no. 1, pp. 624-635, Jan.
adding storage systems to improve the power quality, while 2019.
considering the reliability of combining different network [4] J. Grainger and S. Civanlar, “Volt/var control on distribution systems
with lateral branches using switched capacitors and voltage regulators,
routing structures with the traditional capacitor bank system. part I, part II and part III,” IEEE Power Engineering Review, vol. 5,
The research improves the PF, enhanes voltage profile and no. 11, pp. 52-54, Nov. 1985.
increases the feeder capacity with less investment and opera‐ [5] F. R. Islam, K. Prakash, K. Mamun et al., “Aromatic network: a novel
structure for power distribution system,” IEEE Access, vol. 5, pp.
tion cost. The following conclusions can be made.
25236-25257, Oct. 2017.
1) The optimal value of the battery bank will reduce the [6] M. J. Hadidian-Moghaddam, S. Arabi-Nowdeh, and M. Bigdeli, “Opti‐
load of the overhead lines for active power, and capacitor mal sizing of a stand-alone hybrid photovoltaic/wind system using
banks will add reactive power to the lines. new grey wolf optimizer considering reliability,” Journal of Renew‐
able and Sustainable Energy, vol. 8, no.3, pp. 1-10, May 2016.
2) The algorithm flowchart finds the estimated size and [7] R. E. Precup, Solar Photovoltaic Power Plants: Advanced Control and
placement of the BESS and capacitor. Optimization Techniques. Berlin: Springer, Feb. 2019.
3) There is an improvement in PF and voltage, which [8] L. Xu, X. Ruan, C. Mao et al., “An improved optimal sizing method
for wind-solar-battery hybrid power system,” IEEE Transactions on
helps increase the feeder capacity. Sustainable Energy, vol. 4, no. 3, pp. 774-785, Mar. 2013.
4) The installation of the battery bank improves PF be‐ [9] J. Eyer and G. Corey, “Energy storage for the electricity grid: benefits
cause of the availability of active and reactive power com‐ and market potential assessment guide,” Sandia National Laborato‐
pensator near the inductive loads. ries, vol. 20 , no. 10, pp. 1-5, Feb. 2010.
[10] M. Farsadi, T. Sattarpour, and A. Nejadi, “Optimal placement and op‐
5) The planting capacitors near demand load centers can eration of BESS in a distribution network considering the net present
be limited. value of energy losses cost,” in Proceedings of 9th International Con‐
632 JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, VOL. 9, NO. 3, May 2021

ference on Electrical and Electronics Engineering (ELECO), Bursa, [21] N. Roy, H. Pota, and M. Hossain, “Reactive power management of
Turkey, Nov. 2015, pp. 434-439. distribution networks with wind generation for improving voltage sta‐
[11] S. Kirmani and B. Kumar, “Power quality improvement by using bility,” Renewable Energy, vol. 58, pp. 85-94, Oct. 2013.
STATCOM control scheme in wind energy generation interface to [22] R. Brown, Electric Power Distribution Reliability. Boca Raton: CRC
grid,” IOP Conference Series: Earth and Environmental Science, vol. Press, 2009.
114, no.1, pp. 1-10, Jan. 2018. [23] O. Van Vliet, A. Brouwer, T. Kuramochi et al., “Energy use, cost and
[12] R. Shah and N. Mithulananthan, “A comparison of ultracapacitor, CO2 emissions of electric cars,” Journal of Power Sources, vol. 196,
BESS and shunt capacitor on oscillation damping of power system no. 4, pp. 2298-2310, Feb. 2011.
with large-scale PV plants,” in Proceedings of Universities Power En‐
gineering Conference (AUPEC), Brisbane, Australia, Sept. 2011, pp.
1-6. F. R. Islam received the Ph. D. degree in electrical engineering from the
[13] K. Mounika and C. Sridhar, “Control schemes for improving power University of New South Wales, Canberra, Australia, in 2013. He is work‐
quality with BESS and ultra-capacitor supported dynamic voltage re‐ ing as a Senior Lecturer of the University of Sun Shine Coast, Sippy
storer,” International Journal of Advance Engineering and Research Downs, Australia. He was the Publication and Publicity Chair of IEEE Fiji.
Development, vol. 4, no. 4, pp. 1-8, Apr. 2017. He is an Associate Editor of IEEE Access. His research interests include mi‐
[14] S. Karanki and D. Xu, “Optimal capacity and placement of battery en‐ crogrid, smart grid, supervisory control and data acquisition (SCADA), elec‐
ergy storage systems for integrating renewable energy sources in distri‐ trification of transportation, and control system.
bution system,” in Proceedings of 2016 National Power Systems Con‐
ference (NPSC), Bhubaneswar, India, Dec. 2016, pp. 1-6. A. Lallu received the B.Eng. and M.Sc.Eng. degrees in electrical and elec‐
[15] K. Prakash, A. Lallu, F. R. Islam et al., “Review of power system dis‐ tronics engineering from the University of South Pacific, Suva, Fiji, in 2016
tribution network architecture,” in Proceedings of 2016 3rd Asia-Pacif‐ and 2019, respectively. His current research interests include reactive power,
ic World Congress on Computer Science and Engineering, Nadi, Fiji, distributed generation, energy markets, power supplies, renewable energy
Dec. 2016, pp. 1-7. sources, and new energy storage devices.
[16] S. Iwamoto and Y. Tamura, “A load flow calculation method for ill-
conditioned power systems,” IEEE Transactions on Power Apparatus
K. A. Mamun received the Ph.D. degree in mechanical and manufacturing
and Systems, vol. 100, no. 4, pp. 1736-1743, Apr. 1981.
engineering from Dublin City University (DCU), Dublin, Ireland, in 2007
[17] Y. Amrane, R. Gueddouche, A. Elmaouhab et al., “Optimal capacitor
and 2012, respectively. He is working as a Senior Lecturer of the University
placement and sizing in radial distribution feeders using evolutionary
of the South Pacific, Suva, Fiji. His current research interests include renew‐
algorithm: a case study of the Algerian-Djanet distribution system,” in
able and sustainable energy generation.
Proceedings of International Conference on Electrical Sciences and
Technologies in Maghreb (CISTEM), Algiers, Algeria, Oct. 2018, pp.
1-6. K. Prakash received the B. Eng. and M. Sc. degrees in electrical and elec‐
[18] F. R. Islam, K. Mamun, K. Prakash et al., “Aromatic network for pow‐ tronic engineering from the University of South Pacific, Suva, Fiji, in 2016
er distribution system,” Australia Patent 2 017 100 293, Jan. 2019. and 2019, respectively. His current research interests include real power, dis‐
[19] I. Bayram, G. Michailidis, M. Devetsikiotis et al., “Local energy stor‐ tributed generation, renewable energy sources, and new energy storage de‐
age sizing in plug-in hybrid electric vehicle charging stations under vices.
blocking probability constraints,” in Proceedings of 2011 IEEE Inter‐
national Conference on Smart Grid Communications, Brussels, Bel‐ N. K. Roy received the Ph. D. degree from the University of New South
gium, Oct. 2011, pp. 1-6. Wales, Canberra, Australia, in 2014. He is an Associate Professor of Khulna
[20] Operation Technology, Inc. (2011, May). ETAP user guide. [Online]. University of Engineering and Technology, Khulna, Bangladesh. His current
Available: https://www. scribd. com/document/219496381/Etap-User- research interests include power system stability and renewable energy.
Guide