This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

Enhancing Quality of Power to Sensitive Loads

with Microgrid
Ambarnath Banerji, Member, IEEE, Sujit K. Biswas, Sr. Member, IEEE & Bhim Singh, Fellow, IEEE

AbstractThis paper presents the effectiveness of a Microgrid

(MG) in a distribution system. The Combined Heat and Power
(CHP) system is used to enhance the overall efficiency of the
power system. Microturbines are used for combined heat and
power operation in this paper. The various methods of control of
the MG are discussed. A simple low cost model is proposed for
MG based on frequency droop control for active power
generation. The proposed model can be used in plug and play
mode of operation of the MG. The model of MG is developed and
its performance is simulated in MATLAB platform. Performance
of the model is studied under grid connected and under islanded
conditions.
Index TermsCombined Heat and Power applications (CHP),
Distributed Generation (DG), droop control method, Microgrid
(MG), Microsource (MS), Power Quality.

I. INTRODUCTION

ODAY, several mission critical loads are present on the

power system which require quality power for their proper
functioning [1], [2]. Popular technologies which provide
quality power, during the failure of grid, include use of
Microturbines [3], fuel cells, as well as renewable energy
sources like solar Photo Voltaic (PV) [4] and wind power [5 &
20]. These small generators, dispersed throughout the power
system, have been used primarily for backup and were not
synchronized to the grid power supply. Furthermore, they were
not interconnected. There has been a trend to change the role
of this Distributed Generations (DG) from backup to primary
energy supply and to have flexible interconnection strategies.
The concept of Microgrid (MG) has thus grown out of this
desire for a flexible interconnected system [6]-[8]. A MG on
the low voltage (LV) distribution is an appropriate solution
for:
a) The customer with mission critical loads receiving
quality power
b) The Distribution Network Operators (DNOs) as relief
for their already overloaded system
Manuscript received May 9, 2014.
Ambarnath Banerji is with the Electrical Engg. Dept., Meghnad Saha
Institute of Technology, Kolkata 700150, INDIA (phone: +91-9830235168; email: ambarnathbanerji@gmail.com).
Sujit K. Biswas is with the Electrical Engg. Dept., Jadavpur University,
Kolkata 700032, INDIA (e-mail: sujit_biswas@hotmail.com).
Bhim Singh is with the Electrical Engg. Dept., Indian Institute of
Technology
Delhi,
New
Delhi
110016,
INDIA
(e-mail:
bhimsinghr@gmail.com).

DGs are emerging as new paradigm to produce onsite,

highly reliable and good quality power. The power system is
going through a rapid growth with the connection of DGs in
the distribution system. Transmission through pipes is slowly
and surely replacing the age old concept of transmission
through wires [1]. One of the most promising applications of
this new concept corresponds to the Combined Heat and
Power (CHP) application leading to an increase of the overall
energy utilization in the total system [7]. This is due to the fact
that compared to the electricity transmission, transporting low
grade recovered heat is prohibitively expensive relative to its
net economic value. Thus generating close to potential user of
waste heat has a compelling attraction [9]. The installation of
DG has enabled the DNOs to postpone incurring expenditure
on expansion of the transmission and distribution network
[10]. To install a new DG unit at a particular point in the
distribution network, the DNO needs to quantify the capacity
of the DG. If the customer has a DG of definite capacity, the
DNO should be able to identify the optimal location for it [11].
A variety of approaches have been suggested for achieving the
flexibility of operation of the MG. An overview is provided in
literature [12]-[19]. Depending on the type of load
requirements, Microgrids can operate with ac or dc power,
resulting in ac or dc Microgrids. The coexistence of ac and dc
subgrids in a hybrid Microgrid is proposed by Loh et al [14].
Control strategies for MG islanded operation have been
discussed by Lopes et al [15]. Controller design and
optimization method of a Microgrid is presented by Chung et
al [16]. An ac-ac matrix converter has been utilized to
interface high speed Microturbine generator to utility grid as
distributed generation unit [18]. Lately, the concept of virtual
flux has been introduced by Hu et al [19] for droop control of
Microgrid voltage and frequency. In this paper, an attempt has
been made to model the MG in a manner which is simple, with
control schemes that has matured in the industry and involves
less expenditure. Any addition or removal of a DG from the
MG is a plug and play matter. Furthermore, the performance
of the proposed MG under both grid-connected and islanded
condition is also investigated.
II. CONFIGURING THE MICROGRID
A. Control of MG interfaced with Rotating Machines
Of the various alternatives available, the diesel engine and

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
synchronous generator have received acceptance with small
and medium sized consumers and are the most widely used
back up source of power. If such machines are connected to
the MG, three modes of control are possible [12]:
(a) Fixed power control
(b) Fixed speed control
(c) Droop Control
The fixed power control logic is adopted for plants
connected to utility grid with no obligation regarding load
regulation. However, when such a plant has to work under
isolated condition and if there is more than one unit on the
MG, then droop control method has to be adopted.
B. Control of MG with Inverter Interface
There are several energy sources like fuel cells, solar PV
units, Microturbines etc. (popularly called MicroSources or
MS) which produce dc output or produce electrical output at
frequencies not compatible with the grid frequency. In such
cases, the system block shown in Fig. 1 is adopted. The MS is
made to charge a battery, which in turn feeds power to the MG
through an inverter [12]. Moreover, it will be assumed that, if
the power demand is within the capability of the device, the dc
voltage is kept constant by the primary generator controls.
Im

Converter
AC / DC or
DC/DC

Microsource

Ich

Inverter
AC

Vdc
3

Microgrid

Idc

Fig. 1. Block diagram of a Microsource connected to the Microgrid.

III. MICROGRID CONTROL ARCHITECTURE

MG control architecture can be broadly classified into
following types.
A. Centrally Controlled
A MG Central Controller (MGCC) installed on grid
substation at the PCC, centrally controls the MG operation.
It sends control signals to the second level of controllers,
such as those at group of loads (Load Controller or LC) or
controllers located at MS (Microsource Controller or MC).
Depending on the economic managing function, stability
control and other functionalities that are built into the MGCC,
it communicates with the LC and MC and provides set points
to both LC and MC [15]. Schematic diagram of such a MG is
shown in Fig. 2.
FEEDER A

PCC

STATIC
SWITCH

LC

MC
Heat
+
Load

MGCC
DG

SENSITIVE LOADS

FEEDER B

GRID

LC

MC
DG
CIRCUIT BREAKER

LC

Heat
+
Load

LC

SENSITIVE LOADS

FEEDER C
NONSENSITIVE LOADS

POINT OF COMMON
COUPLING

POWER & VOLTAGE

CONTROLLER

Fig. 2. Schematic of centrally controlled Microgrid.

B. Autonomous Control
Each MS on the MG has its own autonomous control. There
is no component, such as master controller or central storage
unit which may be critical for the operation of the MG. This
model has an advantage that the MS can be connected at any
point on the MG, which will allow a plug and play operation
model for each component of the MG. Further, with one MS
more than the quantity required for meeting the total load on
the MG, greater security and reliability of the MG can be
ensured in the event of loss of a MS. This control architecture
is discussed later for the control of the MG. A single line
layout of the proposed MG is shown in Fig. 3.
FEEDER A
Heat
+
Load

FEEDER B

PCC

DG

SENSITIVE LOADS

Heat
+
Load

STATIC
SWITCH
DG

GRID

SENSITIVE LOADS

FEEDER C
NONSENSITIVE LOADS

POWER & VOLTAGE POINT OF COMMON

CONTROLLER
COUPLING

CIRCUIT BREAKER

Fig. 3. Single line diagram of the proposed Microgrid.

C. Unit Power Control

In this mode of operation, each MS regulates the voltage at
its output terminal, which is also the point where it is
connected to the grid. The MS sources a constant power into
its connection point in grid connected mode. If the load
increases anywhere in the MG, the extra power is drawn from
the utility grid. This mode supports the CHP mode of
operation because production of power depends on the heat
demand. Electricity is then generated at higher efficiencies.
When the MG islands itself, the MSs frequency droop ensures
that the power requirement of all loads is met within the MG.
IV. MICROSOURCE CONTROL
A. Inverter Control
The model of the inverter connecting the MS to the MG is
shown in the Fig. 4. The plug and play operation of the MS in
a MG imposes conditions that new MS can be added to the
system without modification of the existing equipments. The
MSs autonomously choose their set points. Further, the MG
has the capability to isolate itself from the utility grid or
connect to it in a rapid and seamless fashion. The MS follow
the dynamic changes of the local loads and autonomously
control the active power produced by them. Each MS have
autonomous control without requiring data from the loads, the
static switch or other sources.
Measurement of the active power, reactive power and the
voltage magnitude at connection point to the MG are done by
the blocks on the left side of Fig. 4. After the control action,
they are presented as voltage magnitude Vi and phase angle

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
to the PWM control of the inverter, which in turn issues the
correct pulses to the devices inside the inverter so as to track
the controllers response.
MG
BUS

Vreq
LOAD
CURRENT

LOAD
VOLTAGE
LOCAL
LOAD

LOAD
VOLTAGE

Q
REACTIVE
POWER
MEASUREMENT

VOLTAGE
MEASUREMENT

ACTIVE
POWER
MEASUREMENT

BATTERY
INVERTER
IGBT
BASED

E vs Q
DROOP

Vg

Vi
DYNAMIC
CONTROLLER

AC
3
To
MG

PWM
CONTROL
SIGNAL
GENERATOR

P vs f
DROOP

where Poi & Pi are the power supplied by the ith MS under
grid-connected and islanded operation respectively.
The characteristics shown in Fig. 5 are under steady state.
They have a fixed slope in the region where the unit is
operating within its power range. The slope becomes vertical
as soon as any limit is reached.
Now, if identical inverters are used, Pmax of both the
inverters are the same and they have same droop
characteristics. Then,
versus P characteristics of both the
inverters will overlap. Any increase in load P causes the
same change in for both. In other words the load is equally
shared by the two inverters. This is shown in Fig. 6.

Fig. 4. Block diagram of Microsource controller.

exp
o

Po1
imp

Po2

Exporting
to grid
Importing
From grid

Pmax

P
Fig. 5. Frequency versus active power droop.

Thus the new frequency

1 = o M(Poi Pi)

Exporting
to grid

Po1 = Po2

B. Frequency versus Active Power Droop

The frequency versus active power droop characteristic is
shown in Fig. 5. During the period the MG is connected to the
utility grid, the inverter operates in constant power mode. The
loads on MG receive power both from the main grid and the
local MS, depending on the customers situation. In Fig. 5
power set points are Po1 and Po2 for two units. This is the
amount of power injected by each source when connected to
the grid, with frequency held at grid frequency o.
If the grid power is lost either because of power quality
event or because of forced islanding by opening the static
switch (SS), the MG can automatically transfer to island
operation. In the island mode, each MS has a constant negative
slope droop on the -P plane. The slope M is achieved by
allowing the frequency to drop by a given amount, , as the
power increases by P as,
M = - /P
(1)
Fig. 5 also shows that if the system transfers to island mode
when importing from the grid, then the generation needs to
increase power output to balance power in the MG. The new
operating point will be at a frequency that is lower than the
grid frequency o. In this case, both sources have to increase
their power output. If the system transfers to island mode when
exporting power to the grid, then the new frequency is higher,
corresponding to a lower power output from the sources.

o
'

Importing
From grid

Pmax

Fig. 6. Frequency versus active power droops with identical inverters.

V. MODELING OF MICROGRID
Fig. 7 shows the model of a MG on the MATLAB platform.
The MG is modeled with the help of two identical MSs in
parallel. The MG is further connected with the utility grid
(modeled by a synchronous generator) through a Static Switch
(SS). Each MS is modeled using a Microturbine and an
inverter. Lf and Rf are the per-phase value of inductance and
resistance of the line transformer while Ll is the inductance of
the line.
Since the purpose of this simulation is to investigate the
performance of the inverter under grid connected and islanded
condition, the slow dynamics of the gas fired turbine and the
permanent magnet generator are not represented. The TurbineGenerator-Rectifier group is modeled as a simple dc current
source, injecting the requested dc power into the dc bus.
The battery bank is modeled by its Thevenin equivalent
circuit [20]-[24] and is represented as such in Fig. 7. Vdc is the
dc bus voltage, Voc is the no load open circuit voltage of the
battery, R2 is the internal resistance. The resistance R2 is
usually small. The over voltage condition is represented by
parallel combination of R1 and C1. Since the self-discharging
current of a battery is small, the value of the resistance R1 is
large. The terminal voltage of the equivalent battery [21] is
given by
Vb = (2 2 Vrms)/( 3 m)
(3)
where m is the modulation index, with a maximum value
of 1.0 and Vrms is the line voltage on the ac side of the inverter.

in the islanded mode is given by:

(2)

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
Power
Demand

R1
R2

R2

Vdc
Turbine
Generator
& Rectifier

B2

C
g

PWM
Controller

Inverter 1
R1

Rf

Vdc
Turbine
Generator
& Rectifier
Power
Demand

Lf

B1

C1

B6

C1

Lf

B3

Filter
Load2

b c

Ll

30kW

Load1

PI T

Ll

B7

Rf

SYNCHRONOUS
GENERATOR

Pm

A
B
C

E
N

1pu

B4 STATIC

SWITCH

C
g

Inverter 2

PWM
Controller

B8

Filter
Load2

b c

30kW

Load1

MICROGRID

GRID

Fig. 7. Simulation of a Microgrid connected to the utility grid

The electrical energy is stored in the battery. This energy

can be represented in kWh. For the model of the battery, the
equivalent capacitance can be determined from [20]-[25].
` C1 = (2*kWh*3600*103)/(Vocmax2 Vocmin2)
(4)
Vocmax and Vocmin are the maximum and the minimum open
circuit voltage of the battery during its operation.
From the above equations, different parameters of the
battery (R1, C1) are selected and are given in the Appendix. It
is considered that the battery has energy storage of 24kW for
4hrs peaking capacity and its variation in terminal voltage is
from 740V to 760V.
The model for generating frequency versus active power
droop characteristic is presented in Fig. 8.
f

INVERTER
FREQUENCY

Droop
Constant
1 pu

3Ph
Active Power
measurement

50Hz

Fig. 8. Simulation of frequency versus active power droop.

The active power output of the inverter is compared with the

reference value and the variation in load is measured. The
variation in load is then multiplied with the droop constant of
the system to get the change in frequency. This is then
subtracted from the rated frequency i.e. 50Hz.
VI. CONTROL ALGORITHM OF INVERTER
The Voltage Source Inverter (VSI), shown in Fig.7, is
modeled in discrete mode using ode23tb on MATLAB
platform. The discrete-time integrator block is used to
implement the PI controller. Forward Euler method is used for
integration. In the discrete-time integrator block, 1/s is

approximated by T/(z-1), which results in the following

expression for the output at the nth step:
Y(n) = Y(n-1) + KT*U(n-1)
(5)
where Y(n) is the output and U(n) is the input to the
controller.
The scheme for controlling the ac terminal voltage of the
VSI is shown in Fig. 9. The method of generating the control
signals is explained below. The Vd and Vq signals, thus
generated, are converted into modulation index m and phase
angle , which are used by the PWM modulator for producing
the required pulses for triggering the IGBTs of the VSI.
Vtref

+_

Calculation of Vt

PI Controller
(outer)

Iabc

Iqref

PI Controller
(inner)

+_

Vq

Iq

abc to dq I
d
Generator Converter
Bus
Va Vb Vc
Pe /Vt

_
+

PI Controller
(inner)

Id_ref

PWM
pulse
generator

Vd
Pulses to IGBTs of VSI

Fig. 9. Schematic of the control system of inverter.

A. AC Terminal Voltage Control of VSI

The three phase voltages at the VSI terminal (Vsa Vsb and
Vsc) are considered sinusoidal and hence their amplitude can
be determined as
Vt = [(2/3)(Vsa2 + Vsb2 + Vsc2)]
(6)
This value of Vt is compared with the desired value of the
terminal voltage Vtref. This results in generation of the ac
voltage error and its value at the nth sampling instant is,
Ver(n) = (Vtref Vt(n) )
(7)
where Vt(n) is the amplitude of the sensed three phase ac
voltage at the VSI terminal at nth instant.
The ac voltage error Ver(n), generated above, is fed to an
outer PI controller, using discrete time integration, to generate

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
the quadrature axis reference current Iqref as:
Iqref(n) = Iqref(n-1) + Kap{ Ver(n) - Ver(n-1)} + Kai Ver(n)
(8)
where Kap and Kai are the proportional and integral gain
constants of the outer PI controller for the ac terminal voltage
control.
The actual value of Iq flowing out of the inverter is
generated by an abc to dq conversion of inverter output
current. The error Iqer(n) at the nth instant is generated by
comparing Iqref(n) and Iq(n) as,
Iqer(n) = (Iqref(n) - Iq(n) )
(9)
This error is fed to an inner PI current controller to generate
quadrature axis voltage Vq as,
Vq(n) = Vq(n-1) + Kbp{Iqer(n) Iqer(n-1)} + Kbi Iqer(n)
(10)
where Kbp and Kbi are the proportional and integral gain
constants of the inner PI controller for the ac terminal voltage
control of the inverter.
B. Active Current Output Control of the VSI
The active power Pe(n) of the load is divided by Vt(n) to get
Idref(n), the direct axis reference current as,
Idref(n) = Pe(n)/ Vt(n)
(11)
The actual value of Id flowing out of the inverter is
generated by an abc to dq conversion of inverter output
current. Idref(n) and Id(n) are compared to generate an error Ider(n)
as:
Ider(n) = (Idref(n) - Id(n) )
(12)
The error is then fed to a PI controller to generate V d(n) at
the nth instant as:
Vd(n) = Vd(n-1) + Kbp{Ider(n) Ider(n-1)} + Kbi Ider(n)
(13)
where Kbp and Kbi are the proportional and integral gain
constants of the PI controller of the active load current control
of the inverter.

generator. Simulated transient waveforms of the grid bus

voltage (VUG), grid bus current (IUG), Microsource1 output
voltage (VMS1), Microsource1 output current (IMS1), voltage at
local load of Microsource1 (VLL1), current flowing into the
local load of Microsource1 (ILL1), Microsource2 output voltage
(VMS2), Microsource2 output current (IMS2), voltage at local
load of Microsource2 (VLL2), current flowing into the local
load of Microsource2 (ILL2), electrical power output Pe of each
Microsource, are all presented along with Microgrid frequency
f. Parameters of the proposed system are presented in the
Appendix.
A. Operation of MG with Equal Local Load on both MS
Fig. 10 shows the performance of the MG connected to the
utility grid with equal local load on both MS. The entire
system starts in grid connected mode. A local load of 30kW is
placed on each MS. Both the inverters operate in constant
power mode. The frequency of operation is dictated by the
grid at 50Hz. A part of the local load is supplied by the MS.
The balance power required by the local load is provided by
the grid.
At 0.15s, the local load on each inverter is increased by
130kW. The characteristic shows that the output power of both
inverters has increased. The inverters still operate in constant
power mode. The frequency of operation is again dictated by
the grid at 50Hz. A part of the local load is supplied by the
MS. The balance power required by the local load is provided
by the grid.
The MG is isolated from the utility grid at 0.2s. The
operation of the MG now depends on the frequency droop
characteristic. The power which was being supplied by the grid
has now to be provided by the MS. Hence, there is an increase
in the power drawn by the load from each MS.

VII. RESULTS AND DISCUSSIONS

Iabc Vabc Iabc Vabc Iabc Vabc Iabc Vabc Iabc Vabc
B2
VLL2 B3
f Pe
Pe B7
IMS2 B7
ILL2 B8
ILL! VB3
LL1 IB2
VMS2 B8
MS1 V
MS1 IB4
UG VB4
UG
Wn
(V) (Amp)
(V) (Amp)
(Hz)
(kW)
(A) (Volt)
(A) (Volt)
(A) (Volt)
(A) (Volt)
(kW) (amp)
(V) (Amp)
(V) (Amp)
(A) (Volt)
(V)
(Hz)

Fig.10-12 shows the performance of the proposed MG

comprising of Microturbines and voltage source inverters and
further connected to the utility grid modeled by a synchronous
300

-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
0
50
48

0.1

0.15

0.2

0.25

Time (s)

0.3

0.35

0.4

0.45

Fig. 10. Performance of the Microgrid connected with equal load on both MS

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

0.5

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

Iabc Vabc Iabc Vabc Iabc Vabc Iabc Vabc IabcVabc

B7 B8 B8 B3 B3 B2 B2 B4 B4
f
P
Pee IB7
MS2 VMS2 ILL2 VLL2 ILL! VLL1 IMS1 VMS1 IUG VUG
Wn
(V) (amp)
(V) (amp)
(kW) (amp) (volt)(amp) (volt)
(Hz)
(A) (volt)
(A) (volt)
(V) (amp)
(A) (volt)
(V)
(Hz) (kW) (A) (V) (A)

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
0
50
48

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Time (s)

Fig. 11. Performance of the Microgrid connected with unequal loads on both MS.

There is a corresponding decrease in frequency of the

inverters because of the frequency versus active power droop
characteristic. The inverters settle at the new frequency of
49.5Hz in 0.02s or within half a cycle.
The MG operates in the islanded mode from 0.2s to 0.5s. At
0.3s the load is increased on each MS. An additional load of
100kW is placed on each MS. Using the frequency versus
active power droop mode, the frequency of the inverter falls to
48.5Hz in 0.015s (or a little over half cycle) to cater for the
increased load. Thereafter, the MG maintains constant power
flow at a stable frequency.
TABLE I
LOAD SHARING BY THE INVERTERS DURING SIMULATION
Both inverters
Inverter 1 has (30 + 130
connected to equal
+ 100) kW,
local load
Inverter 2 has (30 + 130
(30 +130 + 100) kW
+ 50) kW
Inverter 1 Inverter 2
Inverter 1
Inverter 2
Grid connected
96 A
96 A
96 A
96 A
mode with
30kW load
Grid connected
270A
270A
270A
270A
mode with
(30+130)kW
After Islanding
354A
354A
354A
354A
& Before Load
increase
After Islanding
534A
534A
499A
499A
& After Load
increase

B. Operation of MG with Unequal Local Loads on both MS

Fig. 11 shows the performance of the MG connected to the
utility grid with unequal local load on both MS. The MG again
starts with equal local load of 30kW placed on each MS. At
0.15s, the local load on each of the grid connected inverter is
increased by 130kW. The frequency is still maintained at 50Hz

as the inverters are still connected to the grid. The MG is

islanded at 0.2s. The MG operates in the islanded mode from
0.2s to 0.5s.
At 0.3s, an additional load of 100kW is placed on MS-1 and
50kW is placed on MS-2. Due to the similar frequency droop
characteristics, the inverters of the MG share the load equally,
which is depicted by equal increase of the output currents of
the inverters. The inverters now operate at slightly lower
frequency of about 48.8Hz, as shown in the Fig. 11.
Table I shows the load current of the inverters under, both
grid connected and islanded mode, with equal and unequal
local loading. It is evident from the Table I that the load
current is equally divided between the inverters both after
islanding and after the subsequent load increase.
C. Reconnection of MG to the Main Grid
Fig. 12 shows the performance of the MG when reconnected
to the main grid. At 0.5s the MG is reconnected to the main
grid. The loads on the MG are now shared by main grid. Thus
the loads on the MS are reduced. The frequency of the MG is
now dictated by the main grid.
D. Voltage Waveforms at Different Instances
Fig. 13 shows the voltage waveform of MS-1 at different
instances of the simulation. Since the two MS are identical, the
waveforms are presented for MS-1 only. MS-2 behaves in an
identical manner.
In the grid connected mode, during the period 0.15s to 0.2s,
the wave form of MS-1 output voltage is shown in Fig. 13(a).
The output voltage is a sine wave without any distortion or
presence of any harmonic. At 0.2s the MG gets islanded,
causing the entire load to appear on the two Microsources. A
step change of load is observed. The waveform Fig. 13(b)
depicts this islanding.

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

B7 B7 B8 B8 B3 B3 B2 B2 B4 B4
Wnf Pe
VLL2(amp)
VLL1(amp)
Pe (amp)
IMS2 (Volt)
ILL2(Volt)
ILL! (Volt)
IMS1 (Volt)
VMS1(amp)
VMS2(amp)
IUG (Volt)
VUG
(kW) (A) (V) (A) (V) (A) (V) (A) (V) (A) (V)
(Hz) (kW)
(Hz)

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
300
-300
500
-500
400
-400
300
-300
240
120
50
47
0.4

0.42

0.44

0.46

0.48

0.5

0.52
Time (s)

0.54

0.56

0.58

Fig. 12. Performance of the Microgrid on reconnection to the main grid.

VVabc
MS1B2

200

0.17
0.18
Time (s)

VMag
MS1 Volts
B2
(% of Fundamental)
VabcVabc
B2

0.16

0.19

FFT window: 1 of 30 cycles of selected signal

10
0

-200
5
-400
0.19
0

0.195

0.21

0
-400

Volts
Vabc V
B2MS1

20
150
10

0.26

0.27
2
3
Time (s)
Frequency (Hz)

0.28

0.38 2
0.4 3
Frequency
Time (s)(Hz)

0.42 4

0.44 5
4

x 10

FFT window: 1 of 30 cycles of selected signal

200
10
0

5
-200

-4000
0.49 0

0.25

400
15

FFT window: 2.5 of 30 cycles of selected signal

0.36

20

35

30
500
25

5
-200

.
Fig. 13(e). Output voltage of MS during steady state after load increment.

1000

(V)

Volts
Vabc B2
VMS1
(% of Fundamental)
Mag
Vabc B2

0.2
0.205
Time (s)
0
200
400
600
800
Fundamental (50Hz)
= 325.3
THD= 37.73%
Frequency
(Hz),islanding.
Output voltage of MS (1 cycle)
during

0.29

5
4

x 10
Fig. 13(c). Output voltage of MS in islanding mode before load increase.

of Fundamental)

0.305

200
10

(V)

400
15
200

200
0.495

4000.5
600 0.505 800
Frequency
Time (s) (Hz)

1000

Fundamental (50Hz) = 285.1 , THD= 69.83%

Fig. 13(f). Output voltage
of MS (1 cycle) while reconnecting to main grid.

Mag (% of Fundamental)

VabcVMS1
B2

Volts

20
600

15

0.3
Time (s)

FFT window: 5 of 30 cycles of selected signal

15
400

(V) (% of Fundamental)
Vabc B2 Mag

(V)
(% of Fundamental)
Vabc B2 Mag(V)

Fig. 13(a). Output voltage of MS in grid connected mode

20

0.295

Fig. 13(d). Output voltage

of MS(50Hz)
(1 cycle)
during
increment under
Fundamental
= 399.2
, THD=load
22.75%
islanded mode

-400
0.15

5
-500
0

-500
0.29

-200

Fig. 13(b).

FFT window: 1 of 30 cycles of selected signal

500

(V)

FFT window: 2.5 of 30 cycles of selected signal

400

Fig. 13(e) shows the waveform of the output voltage during

the period 0.3s to 0.5s with a steady load on each MS. No
visible distortion in the output waveform is observed.

Volts
Vabc V
B2MS1

The step change of load causes distortion of the waveform.

In the islanded mode, under steady state, the waveform is
shown in Fig. 13(c). The output voltage waveform of the MS
is a sine wave without any distortion. At 0.3s the load on the
MS is increased in a step and the change is reflected in the
output voltage waveform with a corresponding distortion as
shown in Fig. 13(d).

60
50
40

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution

requires IEEE permission. See
30
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TIA.2015.2478884, IEEE Transactions on Industry Applications

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <
FFT window: 2.5 of 30 cycles of selected signal

MS1

200

(V)

Vabc VB2 Volts

400

[5]

[6]

-200
-400
0.54

0.55

0.56
0.57
Time (s)

0.58

[7]

Vabc B2 Mag (% of Fundamental)

Fig. 13(g). Output voltage of MS in steady state after reconnection to main

grid.
20

[8]

At 0.5s the MG is again connected to the main grid. There is

15
again a redistribution
of the load between the MS and the main
grid. This grid connection being a step change, results in
10
distortion of the output waveform as depicted in Fig. 13(f).
After reconnection,
a steady state is reached at 0.54s. Beyond
5
0.54s, no distortion is observed in the MS output voltage, as
presented in0Fig.
13(g).200
0
400
600
800
1000

[9]
[10]
[11]

Frequency (Hz)

VIII. CONCLUSION
A simple and low cost control method using parallel
inverters on a MG has been proposed. The proposed control
method has also allowed plug and play operation when the MS
is added or removed from the MG. The use of the frequency
droop method has provided for the control of the MG without
the use of expensive and complex communication system
between the MSs. Simulation results have shown that the MG
can function properly with the proposed type of droop control.
The results of simulation have further demonstrated that the
proposed MG works in appropriate manner for mission critical
loads, where it is essential to provide quality power and
accommodate load changes within a very short time. Future
work can be attempted to reduce the distortions at the time of
step load changes.

[12]
[13]

[14]

[15]
[16]

[17]

APPENDIX
Inverter & system specification:
415V ac rated VSI using IGBT
Lf = 800 H, Rf = 0.004 , C = 1500 F, Ll = 100 H
Kap = 5, Kai = 20, Kbp = 0.05, Kbi = 2
Battery specification:
R1 = 10 K, R2 = 0.1, C1 = 25000F, Voc = 750V.
Synchronous machine specification:
2MVA, 415V, 50Hz, Pole pairs = 2,
H = 0.6s, R = 0.0036 pu, X = 0.16 pu.
REFERENCES
[1]
[2]
[3]

[4]

R. C. Dugan, M. F. McGranaghan, S. Santoso and H. W. Beaty,

Electrical Power Systems Quality, 2nd ed. Tata McGraw-Hill
Education, New Delhi, 2008.
A. Ghosh and G. Ledwich, Power Quality Enhancement using Custom
Power Devices, Kluwer Academic Publisher, Boston, 2002.
A. K. Saha S. Chowdhury, S. P. Chowdhury and P. A. Crossley,
Modeling and performance analysis of a microturbine as a distributed
energy resource, IEEE Trans. Energy Conversion, vol.24, no. 2, pp.
529 -538, Jun. 2009.
P. Xuewei, A. K. Rathore, U. R. Prasanna, Novel soft-switching
snubberless naturally clamped current-fed full-bridge front-end-

[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]

converter-based bidirectional inverter for renewables, microgrid, and

UPS applications, IEEE Trans. Industry Applications, vol. 50, no. 6,
pp. 4132 - 4141, Nov/Dec 2014.
S. N. Bhadra, D. Kastha and S. Banerjee, Wind Electrical Systems
2nd ed. Oxford University Press, New Delhi, India, 2007.
P. Chiredeja and R. Ramakumar, An approach to quantify the technical
benefits of distributed generation, IEEE Trans. Energy Conversion,
vol. 19, no. 4, pp. 764-773, Dec. 2004.
R. Lasseter, A. Akhil, C. Marnay, J. Stephens, J. Dagle, R. Guttromson,
A. S. Meliopoulous, R. Yinger and J. Eto, White paper on Integration
of Distributed Energy Resources, The CERTS MICROGRID Concept
Consultant Report California Energy Commission, U.S. Department of
Energy, Berkeley, CA, LBNL-50829, Apr. 2002.
N. Hatziargyiou, H. Asano, R. Iravani and C. Marnay, Microgrids,
IEEE Power and Energy Magazine, vol. 5, no. 4, pp.78-94, Jul/Aug
2007.
A. S. Siddique, C. Marnay, R. M. Firestone and N. Zhou, Distributed
generation with heat recovery and storage, ASCE Journal of Energy
Engineering, pp 181-210, Sep. 2007.
R. E. Brown, J. Pan, X. Feng and K. Koutlev, Siting distributed
generation to defer T&D expansion, in Proc. 2001 IEEE PES
Transmission and Distribution Conf. Expo. vol. 2, 2001, pp. 622627.
G. P. Harrison, P. Siano, A. Piccolo and A. R. Wallace, Distributed
generation capacity evaluation using combined genetic algorithm and
OPF, International Journal of Emerging Electric Power Systems, The
Berkeley Electronic Press, vol.8, no.2, art.7, 2007.
S. Barsali, M. Ceraolo, P. Pelacchi and D. Poli, Control techniques of
dispersed generators to improve the continuity of electricity supply, in
Proc IEEE Power Engr. Soc. Winter Meeting, 2002, pp. 789-794.
F. Blaabjerg, R. Teodorescu, M. Liserre and A. V. Timbus, Overview
of control and grid synchronization for distributed power generation
systems, IEEE Trans. Industrial Electronics, vol.53, no. 5, pp. 13981408, Oct. 2006.
P. C. Loh, D. Li, Y. K. Chai and F. Blaabjerg, Autonomous control of
interlinking converter with energy storage in hybrid ACDC microgrid,
IEEE Trans. Industry Applications, vol.49, no.3, pp. 1374-1382,
May/June 2013.
J. A. P. Lopes, C. L. Moreira and A. G. Madureira, Defining control
strategies for microgrids islanded operation, IEEE Trans. Power
Systems, vol. 21, no. 2, pp.916-924, May 2006.
I. Y. Chung, W. Liu, D. A. Cartes, E. G. Collins Jr. and S. I. Moon,
Control methods of inverter-interfaced distributed generators in a
microgrid system, IEEE Trans. Industry Applications, vol.46, no.3, pp.
1078-1088, May/June 2010.
K. D. Brabandere, B. Bolsens, J. V. d. Keybus, A. Woyte, J. Driesen and
R. Belmans, A voltage and frequency droop control method for parallel
inverters, IEEE Trans. Power Electronics, vol. 22 no. 4, pp. 11071115, Jul. 2007.
H. Nikkhajoei and M. R. Iravani, A matrix converter based
microturbine distributed generation system IEEE Trans. Power
Delivery, vol.20, no. 3, pp. 2182-2192, Jul. 2005.
J. Hu, J. Zhu, D. G. Dorrell, J. M. Guerrero, Virtual flux droop method
- a new control strategy of inverters in microgrids, IEEE Trans. Power
Electronics, vol.29, no.9, pp. 4704 4711, Sept 2014.
B. Singh and G. K. Kansal, Solid state controller for an isolated wind
energy conversion system feeding dynamic loads, in 8th International
Power Engineering Conference, (IPEC 2007), 2007, pp.532-537.
Z. M. Salameh, M. A. Casacca and W. A. Lynch, A mathematical
model for lead acid batteries, IEEE Trans. Energy Conversion, vol.7,
no.1, pp. 93-97, 1992.
Z. M. Salameh, M. A. Casacca and W. A. Lynch, Determination of
lead-acid battery capacity via mathematical modeling technique, IEEE
Trans. Energy Conversion, vol.7, no. 3, pp. 442-446, 1992.
M. Chen and G. A. Rincon-Mora, Accurate electrical battery model
capable of predicting runtime and I-V performance, IEEE Trans.
Energy Conversion, vol.21, no.2, pp. 504-511, 2006.
M. Ceraolo, New dynamic models of leadacid batteries, IEEE Trans.
Power System, vol.15, no. 4, pp. 1184-1190, 2000.
N. Mohan, T. M. Undeland and W. P. Robbins, Power Electronics
Converters, Application and Design, 3rd ed., Wiley India, New Delhi,
2007.

0093-9994 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.