Optimal Capacity and Placement of Battery Energy

Storage Systems for Integrating Renewable Energy

Sources in Distribution System
Srinivas Bhaskar Karanki Member, IEEE , David Xu Member, IEEE
School of Electrical Sciences, Department of Electrical and Computer Engineering
IIT Bhubaneswar, India Ryerson Univeristy, Toronto

Abstract—Battery energy storage can bring benefits to multiply centers. Therefore, it is necessary to consider an appropriate
stakeholders in the distribution system. The integration of the location in the residential distribution system to install BESS
Battery Energy Storage System (BESS) and renewable energy to obtain an optimal effect. The optimal sizing and placement
sources with the existing power system networks has many
challenges. One of the major challenges is to determine the of the BESS in the distribution power system is an important
capacity and connection location of the BESS in the distribution aspect to maximize the benefits of the BESS in the system.
system. The installation of BESS units at suboptimal places The BESS should be located at a bus, where it provides a
may increase the cost, including system losses and installation higher reduction in losses without any violation to the voltage
of larger battery capacity. So, it is essential to have a method profile. Various algorithms for optimal sizing and placements
capable of analyzing the influence of BESS allocation and sizing
on power distribution system performance. In this paper, a loss of the BESS are proposed in the literature [2]. Although
sensitivity based algorithm is proposed for optimal placement most of the investigators state that increasing BESS capacity
of the BESS in the distribution system to reduce the distribu- improves its capabilities and hence power system performance,
tion system losses. This paper also presents an algorithm for no suggestion has been made clearly regarding choosing the
determining the optimal size of the BESS using particle swarm optimal placement of the BESS in the power distribution
optimization technique. An electrical distribution utility system
data in Ontario have been used to show the performance of the network [3].
proposed algorithm. A method supporting non-radial distribution system for volt-
age regulation has been proposed to determine the placement
Index Terms—Distribution system, battery energy sources,
of the energy storage units in [4]. In [5], a method based on
particle swarm optimization, optimal battery capacity.
classification of the distribution substation’s main transformer
(MTr) is proposed to determine the optimal placement of the
I. I NTRODUCTION BESS in the distribution system. In [6], a method is proposed
Variable output of the renewable energy sources presents to reduce the feeder losses by optimally placing the BESS in
many integration challenges, especially at high level of pene- the network. The major drawback of these methods are, they
tration. Similarly, the uncertainty in the weather-based gen- may not be suitable for complex distribution system.
erating facilities affects the decisions of different activities In this paper, a loss sensitivity index based method is
related to the operation of distribution systems. Electric utility proposed for determining the optimal placement of the BESS
infrastructure costs are driven primarily by the need to serve in the power distribution system. Even though the BESS is
the load during the peak demand period. Therefore, it is placed at the optimal location, the size of the BESS has a
desirable to shave peak demand in order to defer transmission considerable impact on the power distribution system perfor-
and distribution equipment upgrades, and reduce or avoid mance. If inappropriate size of the BESS is installed at the
the necessity to purchase much higher cost generation assets. optimal location, the loss in the system may decrease but the
An effective way to achieve reduction in peak load is the voltage or power flow violations may increase in the system.
application of Battery Energy Storage Systems (BESS). The Thus in this paper, once the optimal location is determined,
integration of battery energy storage in a distribution grid the optimal sizing of the BESS to be installed is obtained
could mitigate some of the problems of a high penetration by optimizing the losses in the system using Particle Swarm
of the distributed generation. In addition to the load leveling, Optimization (PSO) technique.
BESS can be widely used for frequency control, voltage reg- Typically BESS is governed by five important parameters,
ulation, improvement of the power quality in the distribution energy capacity (MWh), power capacity (MW), round trip
system. All these applications serve to increase the reliability efficiency (η), discharge rate and state of charge (SOC). State
and stability of the grid [1]. of Charge (SOC) has been considered in sizing and power
The suboptimal BESS location and sizing can cause under management of the energy storage in [7]. Similarly various
or over-voltages in the distribution network. BESS also causes algorithms are proposed in the literature for optimal sizing
an impact in the losses due to its proximity to the load of the BESS. But most of the proposed algorithms have not
978-1-4799-5141-3/14/$31.00 
c 2016 IEEE considered all the parameters of the BESS while designing

Authorized licensed use limited to: Indian Institute of Technology - BHUBANESWAR. Downloaded on July 29,2024 at 13:52:37 UTC from IEEE Xplore. Restrictions apply.
 the algorithm. In this paper, a heuristic based Particle Swarm PB, err
Optimization (PSO) technique is proposed to determine the
optimal sizing of the BESS by considering all the governing
PW SOC
parameters in the objective function of BESS in the distribu- PL PB, ref
Energy Storage- PB, act Energy Storage-
Controller converter Battery
tion system. SOC
PG

II. O PTIMAL P LACEMENT OF THE BESS JB, rated

PB, rated
An optimal location for BESS has to be identified in the
system such that the distribution system losses are minimized. Fig. 1. Block diagram for proposed methodology.
In this work, system loss sensitivity index with respect to the
BESS control parameters has been used to optimally place
the BESS. In case of optimal BESS placement, the change in
The Controller block determines the desired energy stor-
the power system performance and the BESS parameter are
age power output at the current simulation instant, PB,ref ,
considered for evaluating the sensitivity index. Total power
based on the distributed power generation (PW ), Load power
loss is considered as the parameter for the system performance
(PL ) and grid power (PG ). The energy storage power rating
and the real power injection is the BESS parameter. The power
(PB,rated ) is optimized as following.
loss in a distribution system is given in (1) [8]. ⎧
⎪
⎪ 0, SOC = 0 & PB,ref > 0
⎪
⎪

n ⎪
⎪
PLoss = [αjk (Pj Pk + Qi Qk ) + βjk (Qj Pk − Pj Qk )] (1) ⎨0, SOC = 1 & PB,ref < 0
PB,act = PB,rated , SOC = 0 & PB,ref > PB,rated
j=1 ⎪
⎪
k=1 ⎪
⎪ −PB,rated , SOC = 1 & PB,ref < −PB,rated
⎪
⎪
where ⎩P Else
B,ref
rjk rjk (3)
αjk  cos(δj − δk ) βjk  sin(δj − δk )
|Vj ||Vk | |Vj ||Vk | The discharge from the energy source corresponds to the
The loss sensitivity index with respect to the BESS control positive PB,ref and the charging of the energy source to
parameter is given in (2). negative PB,ref . The SOC of the battery pack is updated as
a function of the power into or out of the battery. The rated
∂PLoss  n
energy capacity JB , rated is optimized as following.
=2 (Pk αik − Qk βik ) (2)
∂PB η.T.PBact (T − 1)
k=1 SOC(T ) = SOC(T − 1) − (4)
60.JB,rated
where, αjk , βjk are the loss co-efficients. Pj , Pk are the real
power injections at the buses j and k. Qj , Qk are the reactive The term T refers to the sample time at which the real time
power injections at the buses j and k. Vj , Vk are the voltage data has been obtained. The round efficiency is defined as
magnitudes at the buses j and k. rjk is the resistance of the following. 
transmission line connected between bus i and j. δi , δj are the ηout , PBact,−1 > 0
η= (5)
voltage phase angles at the buses i and j respectively. ηin , PBact,−1 < 0
Since the loss sensitivity on each bus is calculated, this
method is suitable for distributed BESS allocation in the power The terms ηout and ηin are the efficiency during discharge and
system network, when compared to other methods available charge respectively. The energy storage unit will not source
in the literature. This method allows the utility to determine power if the SOC is equal to 0, or sink power if the SOC
various bus locations based on the sensitivity instead of single is equal to 1. The expression for the instantaneous state of
centralized BESS in the distribution system. charge is given below.
 t
III. O PTIMAL R ATING C APACITY OF THE BESS SOC(t) = QT − i(τ )dτ (6)
to
The BESS is governed by following parameters, energy
capacity (MWh), power capacity (MW), round trip efficiency Where QT is the state of charge at the initial time to . The
(η ), state of charge (SOC) and discharge rate (DR). The round Discharge rate (DR) of the battery is defined as the current at
trip efficiency depends on the chemical reaction inside the which a battery is charged or discharged. The rate is expressed
battery storage and a fixed value is assumed for convenience. in terms as C/h rate, where C is the rated capacity and h is the
The algorithm should determine an optimal energy capacity discharge/ charging time in hours. The instantaneous discharge
and the power capacity of the BESS with an inequality on rate at given time sample can be expressed in terms of SOC
SOC and the discharge rate. Figure 1 shows the block diagram as given below.
of the proposed methodology. PL , PW and PG are the real DR(t) = [SOC(t + 1) − SOC(t)]/Δt (7)
time data of the load, distributed power generated and the
power available from the grid at the given substation in the According to Peukret’s law, as the rate of discharge/ charg-
distribution system. Typically, the power data sample time is ing increases, the available capacity of the battery decreases.
for one hour. So in order to increase the life span of the battery the dis-

Authorized licensed use limited to: Indian Institute of Technology - BHUBANESWAR. Downloaded on July 29,2024 at 13:52:37 UTC from IEEE Xplore. Restrictions apply.
 charge/charging rate should be within specified limit. This is has the advantage of not being stuck to a local minimum.
considered as the second inequality constraint in determining A detailed step by step procedure of the PSO technique is
an optimal energy capacity and the power capacity of the discussed in [10]
BESS.
At every simulation step, PB,err the error between PB,ref IV. S IMULATION R ESULTS
andPB,act is evaluated and verified whether it is within the Electrical Distribution Utility Substation (X)
specified tolerance limit . Tf rac is the fraction of the samples The proposed algorithm for the optimal placement of the
in the simulation duration in which the error is within the BESS in the distribution system is tested on real time data from
tolerance limit. A reasonable cost model for flow cell battery an electrical utility in Ontario. Substation (X) is considered as
energy storage device is defined as following [9]. the test system. The CYME model of the substation has been
Cost = f (Tf rac ).(Cp .PB,rated + Cj .JB,rated ) (8) provided by the utility. A total of 99 spot loads are connected
to the substation through 161 overhead balanced lines and
Where Cp and CJ are the cost per Watt and cost per Wh of the 92 cables. The substation has 278 nodes and 6 buses. The
energy storage device. The term is added to penalize energy system also contains 17 shunt capacitors, one regulator, 12
storage systems in the solution time that do not meet the error two winding transformers. Two wind farms of 9.9 MW rating
tolerance band and defined as following of each are connected to the system. Each wind farm has five

1, Tf rac ≥ 0.85 1.98 MW rating generators. The wind farms are connected to
f (Tf rac ) = (9) the substation X approximately through 20 miles transmission
inf, Tf rac < 0.85
line and the load is concentrated around 15 miles away from
The objective is to minimize the cost of the battery energy the wind farms.
storage system and satisfying the inequality constraints on the Using the basic network equations, the CYME model has
SOC and DR by finding an optimal value of PB,rated and been reduced to fewer nodes and buses. The reduced CYME
JB,rated . Particle swarm optimization (PSO) is a heuristic model contains 63 nodes and 45 spot loads connected to the
based optimization technique which has been employed to substation through 62 overhead transmission lines and cables.
power system for its simplicity in mathematical formulation. The reduced model parameters are used for Matlab simulation
Once the optimal size and location of the BESS is deter- study to find the loss sensitivity of each bus.
mined, the power delivered/observed by the BESS at given Newton Raphson Iterative technique has been used to solve
operating condition should be determined by optimizing the the power flow in the three phase unbalanced distribution
loss in the system. To determine this, an objective function system in Matlab. The loss sensitivity index for each bus in
consisting of loss minimization is formulated. Voltage limit the system is determined using equations (1) and (2). The
violation at buses, line thermal limit violation are taken as the bus voltages, voltage phase angles and the power flows are
operational constraints and the BESS injected power at the obtained by solving the load flow analysis. The overall system
most sensitive bus is the control variable as given in (10). losses under various operating conditions are given in Table
I. From the table it is clear that the maximum losses in the
system occur when the load is minimum (3.01 MW) and
n
F un = min{ [αjk (Pj Pk + Qi Qk ) + βjk (Qj Pk − Pj Qk )]} the wind generation is maximum (19.8 MW). Similarly the
j=1 number of phases which are violating the voltage limits are
k=1 maximum when the load and wind generation are minimum.
(10)
TABLE I
Such that the following constraints are satisfied. L OAD F LOW A NALYSIS U NDER VARIOUS O PERATING C ONDITIONS
Equality Constraint


n 
c Load Wind Generation Losses Voltage Variation
ΔPip = Pip − |Vip | |Vkm |[Gpm pm pm in MW in MW in kW < 0.94 > 1.06
ik cos δik + Bik sin δpm ]
ik

k=1 m=a 8.84 0 341.96 0 106

3.01 0 56.09 0 162

n 
c
8.84 19.80 1624.06 0 66
ΔQpi = Qpi − |Vip | |Vkm |[Gpm pm pm
ik sin δik − Bik cos δpm ]
ik
3.01 19.80 1896.54 0 105
k=1 m=a

where i, k = 1,2,3...n buses. p, m = a,b,c phases. G and B The operating condition in which the losses are highest is
represent the conductance and susceptance. considered for calculating the loss sensitivity index in this
Inequality constraint work. Figure 2 shows the bus number versus total real power
Vipmin ≤ Vip ≤ Vipmax (11) loss sensitivity index value. It can be observed that 41st Bus
(Bus ID 753 in CYME model) is the most negative sensitive
Particle swarm optimization technique is employed in this bus and it is the best location to place BESS to reduce
work for solving the objective function. The PSO is best the losses in the system. In this particular test system, the
suitable for optimizing non-smooth and non-linear functions most negative sensitive bus coincides with the most negative
as compared to classical methods like gradient search. It also sensitive bus when individual phase sensitivity is considered.

Authorized licensed use limited to: Indian Institute of Technology - BHUBANESWAR. Downloaded on July 29,2024 at 13:52:37 UTC from IEEE Xplore. Restrictions apply.
 
Whereas, for the next best position the most negative sensitive 

/RDG3RZHULQ0:
bus does not coincides with the most negative sensitive bus 

when individual phase sensitivity is considered. In this case 


bus 29 in a-phase and b phase, bus 54 in c phase are the best 
location after the 41st bus for the placement of the BESS. Bus 
ID 753 in CYME model is the most negative sensitive bus, for 

all the above mentioned conditions, but the next best option

varies with the operating condition.         
+RXUV


(a)
[
 

:LQG3RZHU*HQHUDWLRQLQ0:
 :LQG*HQHUDWRU


:LQG*HQHUDWRU
6HQVLWLYLW\,QGH[

 

 



 

 
        
        +RXUV
%XV1XPEHU
(b)
Fig. 2. Bus number versus sensitivity index plot Fig. 3. a) Load Demand on Malden M7 Substation for 7 days (168 hours)
b) Wind Power Generation.
The optimal rating of the battery can be obtained using
the load and generation profile for specific period of time. TABLE II
S IMULATION PARAMETERS FOR OPTIMAL SIZING
The performance of the proposed algorithm in the Section
III is evaluated by running the Matlab simulation of the
combined system shown in Fig. 1 over a week of wind power Variable Value
data. The energy storage power rating (PB,rated ) and energy Discharge efficiency (ηout ) 95 %
storage energy capacity (JB,rated ) are optimized using the Charging efficiency (ηin ) 90 %
algorithm given in section III. The load profile and wind Error Limit ±4
power generation profile on substation (X) from March 1st Cp 0.20 $/W
to March 7th 2013 are shown in Fig. 3. The peak load on the Cj 0.48 $/W h
substation is 8.91 MW and the peak wind power generation Number of Particles 50
is 19.8 MW. Similarly the substation has lowest load of 3.01 Number of Iterations 100
MW and lowest generation of 0.20 MW. From the load and Sampling time 1 Hour
generation profile it is clear that high amount of power has Data duration 7 Days
been exchanged from the grid, which will eventually increase
the losses in the system.
Simulation parameters for determining the optimal sizing to be exchanged between the grid and wind generation and
of the BESS are given in Table II. The SOC of the BESS at second method is to take this power to be exchanged as a
the start of the simulation is considered as 50 % in this study. optimization variable. When the exchanged power is fixed to
The two optimizing variables are PB,rated and JB,rated , when 4 MW, the optimized rating of the battery is obtained as 7.41
the substation is run on stand alone mode i.e., without any MW and 127.9 MWh power and energy rating respectively.
exchange of power from the grid. Keeping the error PB,ref The energy exchange to the grid is 348 MWh for one week
within the given tolerance band of ±4%, the optimization duration. When the same is taken as optimization variable, the
algorithm determined the PB,rated and JB,rated as 11.52 MW objective function has been modified. The modified objective
and 198 MWh respectively. Out of 168 sample duration 145 function will minimize the energy exchange with the grid
times (86%) the error is within the specified limit. Since the along with the cost minimization of the BESS. The optimized
energy rating is higher than the power rating, the discharge power to be exchanged to the grid is 2.74 MW, with a power
rate is always less than 1C. SOC is maintained within the and energy rating of 6.26 MW and 93.89 MWh respectively.
specified limits and 23 times out 168 samples it is reaching The energy exchange to the grid is 306.85 MWh for one week
the depth of the discharge (DOD). duration. The convergence curve and the variation of State of
In case of the stand alone system operation, the energy Charge is shown in Fig. 4.
rating of the battery is very high and it is not economical. If Table III gives the values of the PB,rated and JB,rated
you consider power exchange from the grid, the energy rating for various fixed power exchanged limits and the optimized
can be decreased. In this work, two methods are discussed to solution. It can be observed from the table that the PSO
find the optimized rating of the battery when power exchange based optimized solution simultaneously reduced the energy
is allowed from the grid. The first method is to fix the power exchange with the grid and the rating of the BESS . The

Authorized licensed use limited to: Indian Institute of Technology - BHUBANESWAR. Downloaded on July 29,2024 at 13:52:37 UTC from IEEE Xplore. Restrictions apply.
 
3%UDWHG 0: of the losses in the system. PSO algorithm is used for solving

-%UDWHG 0:K
the optimization problem. For this, two optimization criteria
)XQFWLRQ9DOXH


 are considered. The first one is loss optimization and equation
 (10) is minimized without considering any constraints. The

second one is loss optimization along with minimizing the

number of phases violating the voltage profile (i.e. considering


the constraints). Table IV gives the optimization results with
          
,WHUDWLRQ1XPEHU 4 different cases. When the wind generation is at its peak, the
(a)
BESS will be in the charging mode (i.e., it acts as load). It
can be observed that, when BESS is not installed, case 4 gives

 3%UDWHG 0: the maximum power losses. But when the BESS is installed
-%UDWHG 0:K
6WDWHRI&KDUJH 62&

 at an optimal location, the losses have been reduced by 10.33



% when compared to losses without BESS (under light load
 conditions), even though the BESS is acting as an additional


load.
 The variation of the total system losses with respect to the
 loading on the system is shown in Fig. 6. It can be observed

         
6DPSOHVLQ+RXU from the figure that, at 0.7 per unit (pu) loading the losses
with and without BESS are same. The BESS injects power,
(b)
when the loading on the system is above 0.7 pu and it absorbs
Fig. 4. Results for Malden substation a) PSO Convergence curve b) Variation power when the loading is below 0.7 pu. When optimal power
of SOC . is injected into or absorbed from the system, the losses in the
system would be reduced with the installation of the BESS at
all the operating points.
Variation of the SOC for various operating conditions is shown
in Fig. 5. The exchange of power is reduced when the power 
exchange is considered as an optimization variable. This also 
:LQG*HQHUDWLRQ 0:

:LWKRXW%(66
reduce the losses that occur during the transmission and the 
/RVVHVLQSX


problem of overloading also reduces. 

TABLE III
 :LWK%(66
C OMPARISON OF O PTIMIZATION R ESULTS



       
Power Exchange PB,rated JB,rated Energy Exchange T /RDGLQSX
with Grid (MW) (MW) (MWh) with Grid (MWh)
Fig. 6. The variation of total losses w.r.t to load changes.
1 2.0 6.69 105.52 248 143
2 3.0 7.41 104.90 309 143 Table V gives the optimization results using CYME model.
3 4.0 6.41 127.70 348 144 Bus number 753 is chosen as the optimal location for the BESS
4 5.0 6.59 132.44 290 149 installation based on the loss sensitivity index. The optimal
5 2.74 6.26 93.88 306.85 143 power injected/absorbed at various operating conditions ob-
tained from Matlab simulations are used in CYME model. The
losses in the system have reduced with the BESS installation.

Even in Case 3 and Case 4, where BESS is in charging mode


(acts as extra load on the system) the losses have decreased

6WDWHRI&KDUJH 62&

  and the loading on the system is also reduced.



 V. C ONCLUSIONS

 An algorithm based on loss sensitivity index is proposed for


 optimal placement of the BESS in the unbalanced distributed

system. The proposed algorithm has been verified using the


data from the Ontario electrical distribution utility system.
         
6DPSOHVLQ+RXUV
PSO based controller is proposed to determine the optimal
rating of the BESS to be installed in the given substation. The
Fig. 5. State of charge under varying operating conditions. optimal power to be injected/absorbed by the BESS at various
operating conditions is determined by loss minimization crite-
Once the optimal location and sizing of the BESS is deter- ria and voltage profile variations. It is possible to conclude that
mined, the optimal power to be injected/observed by the BESS optimally-allocated and sized BESS will potentially reduce
at a particular operating condition is obtained by minimization the losses in the system even when the BESS is in charging

Authorized licensed use limited to: Indian Institute of Technology - BHUBANESWAR. Downloaded on July 29,2024 at 13:52:37 UTC from IEEE Xplore. Restrictions apply.
 TABLE IV
O PTIMAL BATTERY POWER DELIVERED / OBSERVED UNDER VARIOUS OPERATING CONDITION

Operating Optimization Optimal Battery Power Losses Voltage variations

Condition criteria observed/delivered in kW (kW) < 0.94 > 1.06
a b c
Case 1
Wind Generation = 0 MW Basic No BESS 341.96 0 106
Load Power = 8.84 MW Loss Minimization 1552.7 1603.5 141.74 108.02 0 150
Loss +Voltage Profile 226.7 1802 76.90 216.60 0 106
Case 2
Wind Generation = 0 MW Basic No BESS 56.02 0 162
Load Power =3.01 MW Loss Minimization 418 548 507 30.49 0 162
Loss +Voltage Profile 480 1000 34.7 39.69 0 140
Case 3
Wind Generation =19.8 MW Basic No BESS 1624.06 0 66
Load Power =8.84 MW Loss Minimization -468 -216 -131 1612.54 0 55
Loss +Voltage Profile -1930 -1733 -243 1771.58 0 0
Case 4
Wind Generation =19.8 MW Basic No BESS 1896.54 0 105
Load Power =3.01 MW Loss Minimization -1414.8 -1271.8 -1085.3 1700.37 0 104
Loss +Voltage Profile -1953.7 -766.2 -1828.7 1736.67 0 79

TABLE V
O PTIMIZATION R ESULTS USING CYME M ODEL

Operating BESS Grid Optimal Battery Power Losses Over loading

Condition Status Power (MW) observed/delivered ( kW) (kW)
a b c a b c
Case 1 Without
Wind Generation = 0 MW BESS 9.19 0 0 0 316 1 1 1
Load Power =8.84 MW With BESS 4.37 1552.7 1603.5 141.74 130 0 0 0
Case 2 Without
Wind Generation = 0 MW BESS 3.13 0 0 0 42.57 0 0 0
Load Power =3.09 MW With BESS 1.64 418 548 507 22.16 0 0 0
Case 3 Without
Wind Generation =19.8 MW BESS -9.59 0 0 0 1336.6 11 11 11
Load Power =8.84 MW With BESS -8.8 -468 -216 -131 1280.6 11 11 11
Case 4 Without
Wind Generation = 19.8 MW BESS -15.23 0 0 0 1479.58 14 14 14
Load Power =3.09 MW With BESS -11.45 -1448.8 -1271.8 -1085.3 1319.48 11 11 11

mode. The losses in system has been reduced by 10.35% when [5] K. Jung, H. Kim, and D. Rho, “Determination of the installation site and
compared to losses without BESS. Similarly the number of optimal capacity of the battery energy storage system for load leveling,”
IEEE Transactions on Energy Conversion,, vol. 11, no. 1, pp. 162–167,
buses violating the voltage profile limits also reduced by 24.76 1996.
% in the system. [6] F. Alvarado, “Locational aspects of distributed generation,” in Power
Engineering Society Winter Meeting,, vol. 1. IEEE, 2001, pp. 140.
[7] C. Abbey and G. Joos, “Sizing and power management strategies for
R EFERENCES battery storage integration into wind-diesel systems,” in 34th Annual
Conference of IEEE Industrial Electronics, IECON,. IEEE, 2008, pp.
[1] R. Dell and D. Rand, “Energy storage a key technology for global energy 3376–3381.
sustainability,” Journal of Power Sources, vol. 100, no. 1, pp. 2–17, 2001. [8] O. Elgerd, Electric energy systems theory: an introduction. Tata
[2] S. Dutta and R. Sharma, “Optimal storage sizing for integrating wind McGraw-Hill Education, 2007.
and load forecast uncertainties,” in Innovative Smart Grid Technologies [9] T. Brekken, A. Yokochi, A. von Jouanne, Z. Yen, H. Hapke, and
(ISGT),. IEEE PES, 2012, pp. 1–7. D. Halamay, “Optimal energy storage sizing and control for wind power
applications,” IEEE Transactions on Sustainable Energy,, vol. 2, no. 1,
[3] K. Divya and J. Østergaard, “Battery energy storage technology for
pp. 69–77, 2011.
power systems an overview,” Electric Power Systems Research, vol. 79,
no. 4, pp. 511–520, 2009. [10] S. B. Karanki, D. Xu, B. Venkatesh, and B. N. Singh, “Optimal location
of battery energy storage systems in power distribution network for
[4] A. Barnes, J. Balda, A. Escobar-Mejia, and S. Geurin, “Placement of
integrating renewable energy sources,” in 2013 IEEE Energy Conversion
energy storage coordinated with smart pv inverters,” in Innovative Smart
Congress and Exposition, Sept 2013, pp. 4553–4558.
Grid Technologies (ISGT), 2012 IEEE PES, jan. 2012, pp. 1 –7.

Authorized licensed use limited to: Indian Institute of Technology - BHUBANESWAR. Downloaded on July 29,2024 at 13:52:37 UTC from IEEE Xplore. Restrictions apply.