WK,QWHUQDWLRQDO&RQIHUHQFHRQ5HQHZDEOH(QHUJ\5HVHDUFKDQG$SSOLFDWLRQV 3DULV)5$1&(2FW

Cost Savings Oriented Microgrid Control Strategy

Considering Battery Degradation
Ram Shankar Yallamilli Leela Krishna Vedulla Mahesh K. Mishra
Application Engineer Student Member, IEEE Senior Member, IEEE
ABB India Limited Department of Electrical Engineering Department of Electrical Engineering
Bengaluru, India Indian Institute of Technology Madras Indian Institute of Technology Madras
Chennai, India Chennai, India

Abstract—In this paper a cost savings oriented centralised employing appropriate energy management system (EMS) [5],
control strategy for grid connected hybrid microgrid consisting [6]. Apart from power balancing, the other important aspects
of renewable energy sources and energy storage systems is of energy management system according to latest research are
proposed. In this work, cost savings are achieved by a power
management algorithm (PMA) which takes battery degradation to operate batteries efficiently and to minimize microgrids
cost and grid pricing into account. In-addition the proposed operating costs such as fuel costs, maintenance costs, and
PMA always operates battery within SOC limits and also ensures purchased cost of electricity from the main grid [7].
microgrid reliability. The cost savings are computed as the Grid-connected microgrids can either send/receive power
difference between battery degradation cost and electricity cost. from the grid. This electric power trading with the grid has tra-
However, the battery degradation cost is calculated using depth
of discharge (DOD) which is computed on a real time basis. In ditionally been based on a fixed tariff. However, with the incor-
this work, hybrid energy storage system (HESS) is introduced to poration of smart meters, time-varying tariffs are introduced.
absorb the sudden transients and oscillations in battery current Due to rapid development in ESS technologies and power
thereby increasing battery performance and life span. Apart from electronics, these days microgrids are equipped with wide
cost savings analysis, power quality aspects such as harmonic variety of RES powered DG units and ESSs, which allows
mitigation, reactive power support and unity power factor are
achieved. The proposed control strategy is simulated using a the consumers to sell power during surplus generation/peak
MATLAB based simulink environment during both excess and grid pricing periods, and buy power in case of generation
deficit modes of microgrid by taking PV variations into account. scarcity/off-peak grid pricing periods. To maximize the cost
savings [8] an effective EMS is required which includes the
Index Terms—Battery, supercapacitor, grid side converter, electricity price of the grid and the battery degradation cost.
state of charge, power management, cost savings, battery degra- In [9], a supervisory power management system for a
dation cost, depth of discharge, microgrid. grid interactive microgrid with HESS is proposed and the
objectives of PMA formulated is to minimize/maximize power
I. I NTRODUCTION drawn/supplied from/to grid during peak pricing. In [10], a
Renewable energy sources such as PV and wind are pen- grid adaptive power management strategy for an integrated
etrating more into present day power system in the name microgrid with HESS is proposed and the objectives of this
of microgrids, which are considered as the preferred way control strategy are grid frequency based scheduling of HESS.
to promote integration of renewable energy sources to grid. In [11], a low complexity energy management for DC-coupled
Production of electric power from renewable energy sources hybrid microgrid is proposed and the features of the proposed
results in volatile and intermittent power, to address these energy management scheme are use of sharing coefficient
problems energy storage systems (ESSs) are incorporated in for effective power sharing between energy storages and
microgrids [1], [2]. The primary application of ESSs consisting grid. In [12] a grid connected hybrid microgrid system with
of different energy storage devices (ESDs) is to coordinate dual voltage source inverter is proposed with reduced control
with RES to ensure microgrid generation adequacy. These complexity and overall system cost. From the above literature
ESDs have to be operated within upper and lower SOC [9]–[12], it is clear that none of the EMS/PMA proposed did
limits otherwise, it results in rising of internal impedance not consider battery degradation.
and decomposing of electrolyte in the battery which leads to Moreover, in [13] an EMS is proposed to minimize the
capacity loss and power fade [3]. global energy cost in a microgrid, where the global cost
Microgrids in which both AC and DC voltages are used, includes the cost of energy purchased from grid, energy
called as Hybrid AC/DC Microgrids and these are often generation cost, and the negative cost obtained by selling
implemented [4]. The microgrid can operate both in grid energy to the grid. In [14] an optimal energy management
connected and in islanded mode. The power balancing among strategy for the operation of multiple energy storage units
different sources and sinks is one of the most important (batteries) in grid-connected industrial microgrids is presented.
tasks of microgrid management, which can be achieved by The approach is based on a regrouping particle swarm opti-

,&5(5$ 
‹,(((
WK,QWHUQDWLRQDO&RQIHUHQFHRQ5HQHZDEOH(QHUJ\5HVHDUFKDQG$SSOLFDWLRQV 3DULV)5$1&(2FW

Pin(t) Pout(t)

Unbalanced and Non

HGC

linearAC Loads
BC
DC Loads

Pdc(t)

GRID
DC Link

HESS

PV
Supercapacitor Battery

Grid Side Converter

Fig. 1. Proposed hybrid microgrid structure.

mization. In [15] a control strategy is proposed to minimize impedances and the ac bus consists of grid and unbalanced
the industrial facility electricity bill costs and this paper deals non-linear ac loads.
with the energy management of battery storage systems in the
frame of real time pricing. From the above literature [13]–[15], III. C OST A NALYSIS
it is clear that none of the papers has considered both battery
degradation cost as well as battery SOC simultaneously for Batteries are one of the most common forms of energy
cost savings. Moreover, in the above works cost savings are storage used in microgrids. In general the battery life cycle
computed by considering constant DOD which is impractical is defined as the charging/discharging cycles performed by
because the batteries are always cycled at various levels of a battery before its capacity fades to 80% of the original
DOD in real-time operation. In-addition, these works do not capacity, the depth of discharge (DOD) is a key variable
consider a combined energy storage system such as HESS that relates to the battery life cycle. As batteries have limited
which will especially improve the transient response of battery cycle life under different DOD scenarios, the battery capacity
as well as DC link voltage control is also not explored. degradation caused by charging and discharging should be a
Considering the above challenging aspects, in this work constituent part of the energy management algorithms. In this
authors have proposed a cost savings oriented centralized work, cost savings are calculated by considering electricity
EMS for grid connected hybrid microgrids considering bat- purchase costs/selling income from/to the grid and battery
tery degradation cost, grid electricity cost and battery SOC. degradation cost. Battery degradation cost is of three types.
Moreover, the proposed energy management system considers However, degradation cost due to SOC and temperature are
the following vital aspects: a) supercapacitor is introduced to not considered in this work and only DOD related degradation
absorb the sudden transients and oscillations in battery current cost is considered.
thereby increasing battery performance and life span [16], The equivalent full cycles are used for assessing the actual
b) compensation of unbalanced and non-linear load [17] is battery cycle life. But to evaluate the degradation cost of a
achieved to improve the power quality, and c) better dc link single time charge and discharge cycle on a real time basis,
voltage regulation. in this work DOD is determined for a duration of every one
hour. To determine DOD, initial and final SOCs of battery are
II. S YSTEM D ESCRIPTION required. If the initial SOC i.e., SOCi of the battery is known
the final SOC i.e., SOCf at the end of every one hour can
The proposed hybrid microgrid structure considered for the be calculated using coulomb counting method as given below
study is shown in Fig. 1. Renewable energy sources such as [3].
PV modules are interfaced to common dc bus (dc link) using 
a high gain converter (HGC). Due to the intermittent nature 1
SOCf = SOCi − ib dt. (1)
of renewable energy sources, storage systems have become 3600Cb
an integral part of the microgrid. In this work, a combination
Where ib , SOCi , and Cb are current, initial state of charge
of battery and supercapacitor called as hybrid energy storage
and nominal capacity of battery respectively. After computing
system (HESS) is considered and it is interfaced to common
the final SOC of battery, the battery degradation cost for a
dc bus using a bidirectional converter (BC), which enables
particular duration i is calculated using the (2), where the
power flow into and from the energy storage elements. Various
factor 2 in the denominator represents both charging and
topologies of grid side converters are available for interfacing
discharging energy throughputs are considered in the cost [18],
ac and dc sides. However, a neutral point clamped (NPC),
[19].
voltage source converter (VSC) topology is used for both
 
bidirectional power flow and for load compensation. The DOD(i)
grid side converter (GSC) is interfaced to ac bus using filter BDODcost (i) = Cbp (2)
2Ntotal (i)

,&5(5$ 
WK,QWHUQDWLRQDO&RQIHUHQFHRQ5HQHZDEOH(QHUJ\5HVHDUFKDQG$SSOLFDWLRQV 3DULV)5$1&(2FW

18000 0.3
where, pavg (t) is the average power difference between PV
16000 0.26
and ac, dc loads. The average power difference has to be

Degradation cost ($/cycle)

14000 0.23
shared between grid and battery, otherwise it results in varia-
Number of cycles

12000 0.2
10000 0.16 tions in the dc link voltage. The transient power ptr (t) and the
8000 0.13 oscillatory power pos (t) are provided by the supercapacitor.
6000 0.1
4000 0.06 B. Reference current generation
2000 0.03
0 0 The reference currents for various power converters can be
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Depth of discharge derived from the following effective current (ief (t)) equation
using PI controller as given below
Fig. 2. DOD vs Battery lifetime vs Degradation cost. 
ief (t) = Kp ve + Ki ve dt (11)
Where DOD = |SOCi − SOCf |, Ntotal is number of cycles = iavg (t) + itr (t) + ios (t) (12)
and Cbp is battery purchase price. Fig. 2 shows the variations
in battery lifetime (number of cycles) and battery degradation where, ve = vdcref − vdc is the difference between reference
cost with respective to DOD where the battery purchase price and actual dc link voltage, which arises due to input and
is taken as 500 $ by considering 10 packs of 12 V, 26 Ahr output power difference in the dc link. In (11) Kp and Ki are
lead acid EXIDE batteries. The purchase cost/selling income proportional and integral gains of PI controller. The average
from/to grid is calculated by considering two types of tariffs component of effective current is obtained by passing through
namely time of use tariff and feed in tariff. The electricity a low pass filter whose cut-off frequency is chosen to be 5
purchase cost during time period i can be given as Hz.
1
Cp (i) = Pμgrequired (i).λ(i) (3) iavg (t) = ief (t) (13)
(1 + sτ )
Where Pμgrequired and λ are power required in microgrid and The average component of current has to be deliv-
unit electricity price in time period i. Similarly, the income ered/absorbed by grid/battery. Similarly, the transient and
from selling electricity to the grid during time period i can be oscillating components supplied by supercapacitor are given
written as follows below.
Cs (i) = Pμgavailable (i).ϑ(i) (4) iosc (t) = ios (t) + itr (t) = ief (t) − iavg (t) (14)
Where Pμgvailable and ϑ are power available in microgrid and
C. Grid Side Converter Control
feed in tariff during time period i. Considering (2), (3), and
(4) the total cost over a time period i can be given as Instantaneous symmetrical component theory (ISCT) [20]
is one of the time domain control theories used to generate
Ct (i) = BDODcost (i) + Cp (i) − Cs (i) (5) reference currents for active power filters, which are used
IV. DC L INK C ONTROL , R EFERENCE C URRENT for power quality improvement. According to this theory, the
G ENERATION , & GSC C ONTROL compensator/filter reference currents are as given below.
 
A. DC Link Control ∗ vg(abc) + β(vg(bca) − vg(cab) )
if (abc) = il(abc) −  2 Plavg (15)
A dc storage capacitor is used for implementing dc link and i=a,b,c vgi
the instantaneous power flow relationship across the dc link
+
)
can be given as where β, φ+ , i∗f (abc) , il(abc) , vg(abc) , and Plavg are tan(φ
√
3
,
power factor angle between voltage and current positive se-
pin (t) = pdc (t) + pout (t) (6)
quence components, filter reference currents, load currents,
pin (t) = ppv (t) ± pb (t) ± psc (t) − pdcl (t) (7) grid voltages, and average load power on ac side. For unity
pout (t) = ±pg (t) − pacl (t) (8) power factor operation β = 0. Therefore, the above equation
is written as given below.
where, pacl (t), pg (t), pdcl (t), psc (t), pb (t), and ppv (t) are  
instantaneous ac load, grid, dc load, supercapacitor, battery, vg(abc)
∗
and PV powers. The bidirectional power flow is denoted by if (abc) = il(abc) −  2 Plavg (16)
i=a,b,c vgi
the notation  ± . The dc link power is divided into three
components i.e., average power component pavg (t), transient To incorporate bidirectional power flow feature of grid side
power component ptr (t), and oscillatory power component converter in a microgrid environment (16) is modified as
pos (t) as shown below. follows [21].
 
pdc (t) = pin (t) − pout (t) (9) ∗ vg(abc)
if (abc) = il(abc) −  2 (Plavg − Pμg )(17)
= pavg (t) + ptr (t) + pos (t) (10) i=a,b,c vgi

,&5(5$ 
WK,QWHUQDWLRQDO&RQIHUHQFHRQ5HQHZDEOH(QHUJ\5HVHDUFKDQG$SSOLFDWLRQV 3DULV)5$1&(2FW

where Pμg is the net power available in the microgrid and it is Start
the difference between source and sink powers in microgrid.
Read iavg(t), iosc(t),vsa, il(abc), SOCb, SOCbl, SOCbu, immp
Hence, using this formulation of ISCT (17) for microgrid
applications requires measurement as well as processing of
all source and sink power signals in a microgrid. This results Microgrid Deficit Mode
iavg(t) 0
Microgrid Excess Mode
N Y
in increased sensor requirement, increased control complexity
Compute Energy Compute Energy
and high communication bandwidth for centralized control. Required in µg Available in µg
ib = ibrated, isc = iosc, ipv = immp, SOCbl = 0.5
To address the above-mentioned issues, authors have pro- ifj(abc) = il(abc) -(vsa/Ʃi = abc v2si)(1.5Vmiavg)
posed an improved version of (17) for microgrid applications N SOCb < Y
as given below. SOCbu
  SOCb <
SOCbl
Y Peak
N
Pricing
∗ vg(abc)
if (abc) = il(abc) −  2 (1.5iavg Vm ) (18) Compute Battery

i=a,b,c vgi
N Y Degradation Cost
ib = 0, isc = iosc, ipv = immp, SOCbl = 0.2
For regulating DC link voltage, the effective current average Compute Battery ifj(abc) = il(abc) -(vsa/Ʃi = abc v2si)(1.5Vmiavg)
Y BDODcost
Degradation Cost
component has to be delivered/absorbed by grid/battery. But < Cs
using (18) the effective current average component is deliv- ib = -ibrated, isc = iosc, ipv = immp, SOCbu = 0.5
ifj(abc) = il(abc) -(vsa/Ʃi = abc v2si)(1.5Vmiavg) N
ered/ absorbed by grid only. However, in some cases like when
grid is in peak pricing mode battery has to supply the average BDODcost N
< Cp ib = 0, isc = iosc, ipv = immp, SOCbu = 0.9
component of effective current and grid current should be zero. ifj(abc) = il(abc) -(vsa/Ʃi = abc v2si)(1.5Vmiavg)
To incorporate this feature (18) is modified as given below. Y
  Y ib = iavg, isc = iosc, ipv = immp, SOCbl = 0.2
ifj(abc) = il(abc) Peak
∗ vg(abc)
if (abc) = il(abc) −  2 (1.5Kiavg Vm )(19) ib = iavg, isc = iosc, ipv = immp, SOCbu = 0.6
Pricing
i=a,b,c vgi ifj(abc) = il(abc) N

Where, K is a constant which is equal to 1 if grid has to

supply/absorb average component of effective current other-
wise K is equal to 0 if battery has to supply/absorb average Fig. 3. Power Management Algorithm.
component of effective current.
V. P OWER M ANAGEMENT S TRATEGY costlier. Hence, power is taken from the grid only to meet
Power management strategy is required for continuous and the microgrid deficit power requirement.
reliable operation of microgrids. The PMA proposed in this Subcase-II: Peak Pricing Disabled: Here, grid is in off-
work uses average current based mode detection i.e., based peak pricing mode, which means power taken from the grid is
on the value of iavg (either positive/negative/zero) microgrid cheaper. Hence, power is taken from the grid not only to meet
modes are classified as deficit/excess/floating modes [22]. the microgrid deficit power requirement but also to charge
The proposed PMA shown in Fig. 3, tries to minimize the battery till SOCb = 0.5.
microgrid operational costs (i.e., maximizing cost savings) Case-B: SOCb > SOCbl : Here, the deficit power required
as well as ensures reliable microgird operation by regulating by microgrid is either given by battery/grid depending on the
DC link voltage. In this work, only battery SOC limits are battery degradation cost and grid electricity cost. In this case
considered where, the initial upper and lower SOC limits are there are two subcases based on the relation between battery
0.9 and 0.2 and supercapacitor is provided only to absorb degradation cost and grid electricity cost. Grid electricity cost
transient and oscillating current components. After identifying is computed by multiplying the deficit power with electricity
the microgrid mode, the required/available amount of power purchase cost.
in a microgrid is calculated using dc link voltage and average Subcase-I: BDODcost < Cp : As the cost required to supply
value of effective current. the microgrid deficit energy from battery is less than that from
1) Deficit Mode: In this mode, PV power is less than ac and grid, the entire deficit power will be supplied by battery and
dc loads, hence to maintain energy balance the deficit power the power taken from grid is zero.
has to be given by battery or grid based on SOC limits of Subcase-II: BDODcost > Cp : As the cost required to supply
battery. the microgrid deficit energy from battery is more than that
Case-A: SOCb < SOCbl : So battery is kept idle to from grid, the entire deficit power will be supplied by grid
avoid deep discharging and grid will supply the entire deficit and the power taken from battery is zero.
power. In this case there will not be any savings as battery is 2) Excess Mode: In this mode, PV power is more than ac
unavailable and to maintain DC link voltage grid has to supply and dc loads, hence to maintain energy balance the excess
deficit power. In this case there are two subcases based on the power has to be taken by battery or grid based on SOC limits
grid pricing mode. of battery.
Subcase-I: Peak Pricing Enabled: Here, grid is in peak Case-A: SOCb > SOCbu : Here, the battery is kept
pricing mode, which means power taken from the grid is idle/discharged depending on the battery degradation cost. The

,&5(5$ 
WK,QWHUQDWLRQDO&RQIHUHQFHRQ5HQHZDEOH(QHUJ\5HVHDUFKDQG$SSOLFDWLRQV 3DULV)5$1&(2FW

grid will take the entire excess power. In this case there will 8 Ipv
250

be significant amount of savings as power is fed back to grid. 6 Defict Mode Defict Mode
240

Current (A)

Voltage (V)
Subcase-I: BDODcost < Cs : Here, rather than keeping bat- 4 Excess 230 Vdc
Mode
tery idle, it is discharged till SOCb = 0.5 to maximize savings. 2
220

The microgrid excess power along with battery discharged 0

210
(a) (b)
power is taken by grid to maintain DC link voltage. 0 4 8 12 16 Time (hrs) 24 200
0 4 8 12 16 Time (hrs) 24
Subcase-II: BDODcost > Cs : Here, battery is kept idle and

Time Of Use Tarrif ($/Kwhr)

0.5

grid will take the entire excess power available in microgrid. 0.4 Pg Pacl

Active Power (W)

1000 Pdcl Pgsc
Case-B: SOCb < SOCbu : In this case, battery is either kept 0.3
Off-Peak Off-Peak
idle or charged based on grid pricing mode. 0.2 Peak 0

Subcase-I: Peak Pricing Enabled: Here, grid is in peak 0.1

-1000
(c) (d)
pricing mode, which means, the income earned by supplying 0
0 4 8 12 16 Time (hrs) 24 0 4 8 12 16 Time (hrs) 24
power to the grid is higher. Hence, battery is kept idle and the Ib 15
4 Ig(a,b,c)
entire excess power available in microgrid is fed back to grid.

Current (A)

Current (A)
2
5
Subcase-II: Peak Pricing Disabled: Here, grid is in off-peak 0 0
pricing mode which means, the income earned by supplying -2 Isc -5

power to the grid is lesser. Hence, the entire excess power -4

(e) -15 (f)
available in microgrid is used to charge battery till SOCb = 0 4 8 12 16 Time (hrs) 24 0 4 8 12 16 Time (hrs) 24
0.6, and power fed back to grid is zero. 0.4 4

Current (A)
VI. S IMULATION R ESULTS 0.3 2
Battery SOC

0
The proposed system is simulated using MATLAB based 0.2
-2
simulink environment. The system parameters considered for 0.1 (g)
-4
(h)
simulation studies are mentioned in Table I. Tests are con- 0 4 8 12 16 Time (hrs) 24 4 Time (hrs)
ducted to verify the steady state and transient state perfor- 80 Vsa
mance of the proposed system and the detailed simulation 5 40

Current (A)
Current (A)

1
results are as follows. 0 0
The variations in PV current for a duration of 24 hrs is -40
-5
shown in Fig. 4(a), where the PV output increases till midday
(i) -80 (j) Isa
and decreases thereafter. The variations in dc link voltage are 4 Time (hrs) 9 Time (hrs)
shown in Fig. 4(b), where the dc link voltage settles at rated
value of 220 V. The pricing scheme considered in this work Fig. 4. Simulation results: (a) PV current, (b) dc link voltage, (c) Time
is time-of-use pricing, where the electricity price is set as of use tariff, (d) Various active powers as per labelling, (e) variations in
supercapacitor and battery currents, (f) variations in Grid current, (g) Battery
0.1$/kWh during off-peak hours and 0.25$/kWh at peak hours SOC, (h) Unbalanced and non-linear load currents, (i) Zoomed view of GSC
as shown in Fig. 4(c) and the selling price is considered as currents, (j) unity power factor (UPF) operation.
80% of the electricity price [23]. The variations in grid, GSC,
AC and DC loads powers are shown in Fig. 4(d). The battery
and supercapacitor currents are shown in Fig. 4(e), where the
supercapacitor absorbs the sudden transients and oscillations
TABLE I thereby smoothening the battery current.
S YSTEM PARAMETERS
The grid current variations for the entire duration of op-
Supercapacitor specifications Values eration are shown in Fig. 4(f), after t = 13 hrs grid currents
Terminal voltage (Vsc ) 16 V become zero since battery supplies the deficit power . Battery
No. of packs in series 11
Battery specifications Values SOC is shown in Fig. 4(g), where battery is charged till t =
Ah Capacity 26 Ah 9 hrs from grid. Battery kept idle from t = 9-13 hrs, since
Terminal voltage (VB ) 12 V the grid is in peak pricing mode. From t = 13-17, battery
No. of batteries in series 10
DC load parameters Rdcl =320 Ω
is discharged to supply microgrid deficit power as the battery
AC load parameters Values degradation cost is less than the price required to supply deficit
Unbalanced linear load Ra = 20 Ω, La = 65 mH power from grid and from t = 17-24, battery is charged as the
Rb = 10 Ω, Lb = 50 mH grid is in off-peak mode. The zoomed version of unbalanced,
Rc = 30 Ω, Lc = 80 mH
Nonlinear load 3-Φ, diode bridge rectifier with nonlinear load currents are shown in Fig. 4(h). The zoomed
Rl = 100 Ω, version of grid currents shown in Fig. 4(i) are balanced and
Grid side converter parameters Values sinusoidal, the in phase operation of grid currents with grid
DC link voltage Vdc =220 V
DC link capacitance Cdc = 1650 μF voltages are shown in Fig. 4(j) thus demonstrating unbalanced
Hysteresis band h = ±0.1A and nonlinear load compensation.
Grid/source voltages 3-Φ, 50 V (P-N) The microgrid operational cost analysis is shown in table-II,

,&5(5$ 
WK,QWHUQDWLRQDO&RQIHUHQFHRQ5HQHZDEOH(QHUJ\5HVHDUFKDQG$SSOLFDWLRQV 3DULV)5$1&(2FW

TABLE II [5] S. T. Kim, S. Bae, Y. C. Kang, and J. W. Park, “Energy management

M ICROGRID O PERATIONAL C OST A NALYSIS based on the photovoltaic hpcs with an energy storage device,” IEEE
Transactions on Industrial Electronics, vol. 62, no. 7, pp. 4608–4617,
Time BDODcost Cp Cs Ct (Proposed) Ct [9] July 2015.
0-9 hrs 0.047 $ 0.570 $ - 0.617 $ 0.617 $ [6] M. Grami, M. Rekik, and L. Krichen, “Power dispatch strategy for
interconnected microgrids based hybrid renewable energy system,” In-
9-13 hrs - - 0.002 $ -0.002 $ 0.007 $
ternational Journal of Renewable Energy Research, vol. 8, no. 2, June
13-17 hrs 0.020 $ - - 0.020 $ 0.454 $ 2018.
17-24 hrs 0.015 $ 0.463 $ - 0.478 $ 0.478 $ [7] Q. Jiang, M. Xue, and G. Geng, “Energy management of microgrid in
Total microgrid operational cost (Ct ) for 24 hrs 1.113 $ 1.556 $ grid-connected and stand-alone modes,” IEEE Transactions on Power
Systems, vol. 28, no. 3, pp. 3380–3389, Aug 2013.
[8] A. Nadimi and F. Adabi, “Optimized planning for hybrid microgrid
in grid-connected mode,” International Journal of Renewable Energy
where the 24 hour duration is divided into four parts. From Research, vol. 6, no. 2, 2016.
[9] S. Kotra and M. K. Mishra, “Energy management of hybrid microgrid
t = 0-9 hrs, the microgrid is in deficit mode, grid is in off- with hybrid energy storage system,” in 2015 International Conference
peak mode and the battery is allowed to charge as its SOC on Renewable Energy Research and Applications (ICRERA), Nov 2015,
is less than the lower limit, the pp. 856–860.
 associated
0.34
battery degradation
[10] N. Korada and M. K. Mishra, “Grid adaptive power management
cost from (2) is given as 500$ 2∗1780 = 0.047$. The average strategy for an integrated microgrid with hybrid energy storage,” IEEE
power required by the microgrid during this duration is 600W Transactions on Industrial Electronics, vol. 64, no. 4, pp. 2884–2892,
and the associated cost to purchase this amount of power from April 2017.
[11] U. Manandhar, A. Ukil, H. B. Gooi, N. R. Tummuru, and S. K.
grid is 0.57$. Hence, the total microgrid operational cost for Kollimalla, “A low complexity control and energy management for dc-
the duration of 9 hrs is 0.617$. The total microgrid operational coupled hybrid microgrid with hybrid energy storage system,” in 2017
cost for entire day using the proposed control strategy is IEEE Power Energy Society General Meeting, July 2017, pp. 1–5.
[12] Ram. Shankar. Yallamilli and Mahesh. K. Mishra, “Power management
1.113$ which is less than the existing control strategy [9] by of grid connected hybrid microgrid with dual voltage source inverter,”
0.443$. in 2016 IEEE International Conference on Renewable Energy Research
and Applications (ICRERA), Nov 2016, pp. 407–412.
[13] G. Celli, F. Pilo, G. Pisano, and G. G. Soma, “Optimal participation
VII. C ONCLUSION of a microgrid to the energy market with an intelligent ems,” in 2005
International Power Engineering Conference, Nov 2005, pp. 663–668
In this work, a microgrid operation cost savings oriented Vol. 2.
power management algorithm is proposed for the cetralised [14] A. T. Eseye, D. Zheng, H. Li, and J. Zhang, “Grid-price dependent op-
timal energy storage management strategy for grid-connected industrial
control of grid connected hybrid microgrid. Cost savings are microgrids,” in 2017 Ninth Annual IEEE Green Technologies Conference
achieved by creating a trade off between battery degradation (GreenTech), March 2017, pp. 124–131.
cost and grid electricity cost considering variations in PV. The [15] G. Carpinelli, F. Mottola, and L. Perrotta, “Energy management of
storage systems for industrial applications under real time pricing,”
battery degradation cost is computed on a real time basis for in 2013 International Conference on Renewable Energy Research and
every one hour using initial and final SOCs with in that time Applications (ICRERA), Oct 2013, pp. 884–889.
interval. Apart from minimizing microgrids operational costs, [16] K. Nikhil, M. K. Mishra, and S. Kotra, “Power management based
on the operating conditions of grid, mircogrid and hybrid storage,”
the battery packs are always operated with in a safe SOC limits in 2015 International Conference on Renewable Energy Research and
and microgrid reliability is enhanced by regulating DC link Applications (ICRERA), Nov 2015, pp. 1437–1441.
voltage. In addition the proposed system uses HESS which [17] R. Majumder, A. Ghosh, G. Ledwich, and F. Zare, “Load sharing
and power quality enhanced operation of a distributed microgrid,” IET
diverts the transient and oscillating components from battery Renewable Power Generation, vol. 3, no. 2, pp. 109–119, June 2009.
thereby increasing its performance and lifetime. Moreover, the [18] Y. Zhang and M.-Y. Chow, “Microgrid cooperative distributed energy
power quality features like current harmonic compensation, scheduling (codes) considering battery degradation cost,” in 2016 IEEE
25th International Symposium on Industrial Electronics (ISIE), June
reactive power support and unity power factor operation are 2016, pp. 720–725.
achieved. [19] H. Li, S. Su, L. He, and W. Gao, “An analysis on plug-in electric vehi-
cle’s operating cost considering cost of battery capacity degradation,” in
2017 IEEE International Conference on Industrial Technology (ICIT),
R EFERENCES March 2017, pp. 1388–1392.
[20] A. Ghosh and A. Joshi, “A new approach to load balancing and power
[1] S. Vazquez, S. M. Lukic, E. Galvan, L. G. Franquelo, and J. M. Carrasco, factor correction in power distribution system,” IEEE Transactions on
“Energy storage systems for transport and grid applications,” IEEE Power Delivery, vol. 15, no. 1, pp. 417–422, Jan 2000.
Transactions on Industrial Electronics, vol. 57, no. 12, pp. 3881–3895, [21] M. V. M. Kumar, Mahesh. K. Mishra, and C. Kumar, “A grid-connected
Dec 2010. dual voltage source inverter with power quality improvement features,”
[2] V. Rallabandi, O. M. Akeyo, and D. M. Ionel, “Modeling of a multi- IEEE Transactions on Sustainable Energy, vol. 6, no. 2, pp. 482–490,
megawatt grid connected pv system with integrated batteries,” in 2016 April 2015.
IEEE International Conference on Renewable Energy Research and [22] S. Kotra and Mahesh. K. Mishra, “A supervisory power management
Applications (ICRERA), Nov 2016, pp. 1146–1151. system for a hybrid microgrid with HESS,” IEEE Transactions on
[3] P. E. Hartz, L. Liu, and G. Zhu, “State of charge estimation for Industrial Electronics, vol. 64, no. 5, pp. 3640–3649, May 2017.
lion-lithium batteries using extended kalman theorem,” in Industrial [23] C. Ju and P. Wang, “Energy management system for microgrids in-
Informatics - Computing Technology, Intelligent Technology, Industrial cluding batteries with degradation costs,” in 2016 IEEE International
Information Integration (ICIICII), 2015 International Conference on, Conference on Power System Technology (POWERCON), Sept 2016,
Dec 2015, pp. 295–298. pp. 1–6.
[4] F. Nejabatkhah and Y. W. Li, “Overview of power management strategies
of hybrid ac, dc microgrid,” IEEE Transactions on Power Electronics,
vol. 30, no. 12, pp. 7072–7089, Dec 2015.

,&5(5$