IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 24, NO.

1, MARCH 2009

137

Microgrid Dynamic Performance Improvement

Using a Doubly Fed Induction Wind Generator
M. Shahabi, Student Member, IEEE, M. R. Haghifam, Senior Member, IEEE, M. Mohamadian, Member, IEEE,
and S. A. Nabavi-Niaki, Senior Member, IEEE

AbstractThis paper describes a control approach applied on a

doubly fed induction generator (DFIG) to provide both voltage and
frequency regulation capabilities, and hence, an improvement in
the dynamic behavior of a microgrid system. The microgrid system
is assumed to be a portion of a medium voltage distribution feeder
and is supplied by two distributed generation (DG) units, i.e., a gasturbine synchronous generator and a variable-speed wind turbine
with DFIG. A control approach algorithm is proposed for the DFIG
unit to improve both voltage and primary frequency controls. Two
distinct operation modes, i.e., grid-connected and islanding mode,
are used in the proposed approach for proper transfer from normal to islanding operation. Case studies are simulated based on
both planned and unplanned islanding scenarios to evaluate the
performance of the control approach. The study results show that
the proposed control approach for DGs in the microgrid increase
the microgrid systems dynamic performance, reduce frequency
changes, and improve bus voltages regulation during islanding and
autonomous operations.
Index TermsDistributed generation (DG), doubly fed induction generator (DFIG), frequency and voltage controls, islanding,
microgrid.

NOMENCLATURE
Ar
Cpopt
d, q
et
Kopt
Lm
Ls
Popt
r,
R
s, r
Te , Te opt

Area covered by the turbine blades.

Optimal power coefficient.
First subscript indicates direct and quadrature
axes quantities.
Terminal voltage of the DFIG.
Optimal torque/speed constant of the wind
turbine.
Magnetizing inductance of the DFIG.
Stator leakage inductance of the DFIG.
Optimal extraction power.
Blade radius and pitch angle of the wind turbine.
Droop value of frequency-real power
characteristic.
Second subscript indicates a stator and a rotor.
Electromagnetic and optimal electromagnetic
torque.

Manuscript received March 12, 2008; revised July 10, 2008. Current version
published February 19, 2009. Paper no. TEC-00085-2008.
M. Shahabi, M. R. Haghifam, and M. Mohamadian are with the Department of Electrical Engineering, Tarbiat Modares University, Tehran 14115-111,
Iran (e-mail: [email protected]; [email protected]; mohamadian@
modares.ac.ir).
S. A. Nabavi-Niaki is with the Department of Electrical and Computer
Engineering, Toronto University, Toronto, ON M5S 3G4, Canada (e-mail:
[email protected]).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TEC.2008.2006556

opt
w

s , r
t

Optimal tip speed ratio.

Wind speed.
Air density.
Synchronous and rotor angular speed.
Wind turbine rotor speed.
I. INTRODUCTION

ISTRIBUTED power generation system is emerging as

a complementary infrastructure to the traditional central
power plants. This infrastructure is constructed on the basis
of decentralized generation of electricity close to consumption
sites using distributed generation (DG) sources [1]. The increase
in DG penetration depth and the presence of multiple DG units in
electrical proximity to one another have brought about the concept of the microgrid [2], [3]. A microgrid is defined as a cluster
of DG units and loads, serviced by a distribution system, and
can operate in the grid-connected mode, islanded (autonomous)
mode, and ride-through between the two modes. The idea supporting the formation of the microgrid is that a paradigm consisting of multiple generators and aggregated loads is far more
reliable and economical than a single generator serving a single
load.
Planned (intentional) islanding and autonomous operations
of the microgrid supplied by local distributed energy resources
(DERs) has recently attracted major utilities interest worldwide [4], [5]. The concept can potentially improve reliability and
supply security of the distribution network by reducing system
downtime. It also allows the utility company to perform maintenance on upstream medium-voltage (MV)/high-voltage feeders
without supply interruption of the low-voltage customers. The
formation of an islanded microgrid, due to an islanding process,
can be caused by disturbances, such as a fault, or as a result of
intentional islanding events.
After disconnection from the main grid, microgrid experiences frequency and voltage deviations. The amount of these
deviations is highly dependent on: 1) the preislanding operating
conditions; 2) the type of the event that initiates islanding; and
3) the type of DG units within the microgrid. The microgrid is
expected to remain operational after islanding and meet the corresponding load requirements during the autonomous operation.
Transients and the small-signal dynamic model of a microgrid
including conventional and electronically interfaced distributed
resources are investigated and presented in [6] and [7].
The wind-turbine-based DG unit is one of the fastest growing sources of power generation in the world mainly due to:
1) strong worldwide available wind resources; 2) environmentfriendly power generation source especially suitable for remote

0885-8969/$25.00 2009 IEEE

138

Fig. 1.

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 24, NO. 1, MARCH 2009

Schematic diagram of the study system.

areas; and 3) rapid technological development [8]. The

continuous trend of increase in the rate of DG connection and
penetration depth of wind-turbine-based DG units can provoke
several technical concerns and adverse impact on the operation
of distribution systems [9].
Among the different wind energy conversion systems, double
fed induction generators (DFIGs) are capable of providing the
required capabilities without the need of large costs on the power
electronics hardware, provided that adequate control strategies
are added [10]. The control flexibility provided by the DFIG
electronic converter makes it possible to define various control
strategies for participation in primary frequency control and
voltage regulating support.
In this paper, the developed dynamic model of a variablespeed wind turbine with DFIG is employed to provide additional
support to primary frequency control and voltage regulation
of a microgrid. This can improve the dynamic behavior of a
microgrid during islanding and autonomous mode operations.
The proposed control approach consists of using an additional
frequency control loop and a voltage regulating control loop,
which are integrated into the rotor-side controller of a DFIG.
To complete the control approach, a pitch control strategy is
also included, and it dominates the power control for high wind
speeds. In order to evaluate the effectiveness and advantages of
the proposed control scheme, two types of disturbances, planned
switching event (intentional islanding) and unplanned switching
event (unintentional islanding), e.g., a short-circuit event on an
upstream section of the feeder, are applied in a study system. The
study system is designed and the corresponding digital computer
simulation model is developed using PSCAD/EMTDC software
package.
II. STUDY SYSTEM DESCRIPTION
Fig. 1 shows a single-line diagram of a 20-kV local distribution feeder used to investigate possible microgrid operational
scenarios. In this distribution system, the last section of feeder1
that consists of DG units and loads is capable of operating
in islanded mode from the main grid. This section constitutes
the microgrid system. This microgrid system includes two DG
units. DG1 is a 1.8-MVA conventional gas-turbine generator
equipped with excitation and governor control systems. A full
description of such a gas-turbine generator can be found in [11].

Fig. 2.

Schematic diagram of a distribution system with the proposed zones.

DG2 represents a variable-speed wind turbine set with rated capacity of 2 MW, which is interfaced through a DFIG with rated
capacity of 2.5 MVA. The DG1 and DG2 parameters are given
in the Appendix. The microgrid system is separated from the
rest of the MV feeder by a circuit breaker.
III. ISLANDING DETECTION AND THE SYSTEM
PROTECTION SCHEME
Under the regulation governing distribution system operation,
an islanding scenario is permitted only for loads with dedicated
generation units. As a result, DG units must be equipped with
specific islanding detection and prevention schemes to disconnect the unit within 2 s of an islanding event [12].
However, in the future, to realize the full benefit of high DG
penetration depth, the islanding operation of microgrids needs
to be considered.
In the case of future microgrid applications, with the potential
of islanding operation, a fast and reliable detection algorithm is
required to effectively distinguish between an islanding condition and other types of disturbances.
In this research, the islanding and formation of a microgrid
is assumed to be detected within 100 ms (five cycles in 50 Hz).
After detection, the islanding mode operation of the microgrid
is activated by changing the DGs control mode.
To consider the islanding operation of microgrids, a protection scheme in distribution network can be suggested as
follows. The distribution system is divided into several zones in
such a way that in each zone, there is no DG, or if there is any,
balance of generation and consumption in that zone is possible
regardless of the main grid and by using only the power generated by DGs that exist in that zone (microgrid concept) [13]
(see Fig. 2). In other words, the distribution system is divided
into two categories: the first category includes those zones that
have no DG and their loads are fully supplied through the main
grid, and the second category includes those zones that have
one or more DGs and are capable of operating in the islanded

SHAHABI et al.: MICROGRID DYNAMIC PERFORMANCE IMPROVEMENT USING A DOUBLY FED INDUCTION WIND GENERATOR

Fig. 3.

139

Schematic representation of the proposed control approach.

mode. The zones are interconnected using breakers with fast

and consecutive open and close capabilities as well as receiving
remote command. Besides, these breakers must be equipped
with synchro-check relays, to be able to maintain zones
synchronization when it is needed to connect two island zones.
Using an online fault diagnosis algorithm and coordination
of protection devices, such as recloser and fuses and remote
control of breakers in the system, the faulted part of the network
can be detected and isolated from the rest of the grid. In this
scheme, out-of-phase (asynchronous) reclosure will not occur.
This is an important feature, because the out-of-phase reclosure
can impose heavy damages on distribution system devices.
IV. MICROGRID OPERATIONAL MODES
In the proposed microgrid system (see Fig. 1), the islanding
incident that results in the islanding operation can be due to
either a planned (intentional) or an unplanned (unintentional)
switching event. In the case of planned islanding operation, appropriate microgrid load sharing among the DG units and the
main grid may be scheduled prior to islanding. Thus, the islanding process results in minimal transients and the microgrid
continues to operate. An unplanned islanding and microgrid
formation is due to either a fault and its subsequent switching
events or some other unexpected switching process. Prior to
islanding, the operating conditions of the microgrid could be
widely varied, e.g., the DG units can share load in various manners. In the proposed system, planned islanding and unplanned
islanding due to a fault in upstream of the microgrid will be
investigated.
In the grid-connected mode, DG units are expected to supply
their local demand or operated in their optimal operating point
from efficiency point of view. This can be due to the economics
aspects.
Also to avoid interference with the voltage regulator devices
in the network, DGs normally operate at constant power factor
(close to unity) and do not control the grid voltage actively.
In the study system of Fig. 1, DG1 generates constant real
power output (close to unity power factor) and DG2 delivers

maximum power extracted from the wind with unity power

factor (DFIG capability). In this mode, the grid acts as a slack bus
that dominantly supports the real/reactive power requirements
during transients or due to the power fluctuations caused by
DG2, and also stabilizes the frequency.
In the islanding-mode operation, fast and flexible real/reactive
power control strategies are required to minimize dynamics of
the system and damp out transient power oscillations where
no infinite source of power is available. In the proposed system, DG1 is equipped with excitation and governor systems.
After islanding detection, DG1 switches to islanding mode by
regulating its terminal voltage at a reference value, e.g., 1 per
unit (p.u.), through excitation system [using automatic voltage
regulator (AVR)] and by controlling deviations in the generator speed through a governor system that is implemented based
on a frequency-active power relation (droop characteristic) and
a frequency restoration loop (using an integrator). Also, DG2
switches to islanding mode by contributing to primary frequency
control and regulating its terminal voltage through both q- and
d-axis of rotor currents.
The proposed algorithm and the control approach, in general,
are illustrated in Fig. 3.
V. SYSTEM MODEL
The PSCAD/EMTDC software package is used to develop a
time-domain simulation model of the study system of Fig. 1.
The component models used for simulation are as follows.
The main grid (63-kV side of the distribution substation) is
represented by a 63-kV three-phase voltage source with the
short-circuit capacity of 1600 MVA and X/R ratio of 15. The
distribution substation 63-/20-kV transformer is represented as
a linear transformer with appropriate winding connections (see
Fig. 1). Each feeder is represented as a three-phase overhead line
with lumped RL elements per phase or a cable with equivalent
circuits (R, L, and C elements) per phase.
DG1 is modeled as a single-mass machine. The machine electrical system is represented in the dqo frame with two rotor
windings on each axis. The excitation and governor systems of
the machine are also included in the model.

140

Fig. 4.

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 24, NO. 1, MARCH 2009

PI controllers of the rotor-side converter.

DG2 is modeled as a variable-speed wind turbine with a

DFIG. The mechanical power generated by the rotor is converted into electrical power using a DFIG. In the doubly fed
(wound rotor) induction generator, the stator winding is coupled directly to the grid. The rotor winding is connected to a
back-to-back voltage source converter. Speed controller, primary frequency controller, pitch angle controller, and terminal
voltage regulator are included in the model, and described in the
following sections.
A. Variable-Speed Wind Turbine With DFIG
The equations describing a doubly fed induction machine can
be found in [14]. When modeling the DFIG, the generator convention will be used, which means that the current direction is
toward the output instead of inputs. The generator equations in
a dq reference frame used in this research is given in [15]. Note
that the dq reference frame is rotating at synchronous speed
with the q-axis 90 ahead of the d-axis. The position of the
d-axis coincides with the maximum of the stator flux, which
means that Vq s equals the terminal voltage et and Vds equals
zero [15]. For simplicity, dynamic phenomena and switching
transients in a back-to-back voltage source converter are neglected. Instead, the rotor-side converter is modeled as a controllable voltage source, and the grid-side converter is modeled
as a controllable current source.
The rotor-side converter is used to control speed and both
active and reactive power outputs through Vq r and Vdr components obtained from two separate sets of PI controllers, as
shown in Fig. 4. To consider the rated capacity of the backto-back converter, the limits are imposed on both rotor d- and
q-axis currents.
The grid-side converter is used to take into account the balance of active power between the rotor and the grid.

Optimal extraction power curve.

can be obtained from the following function [15]:


116
Cpopt (opt , ) = 0.22
0.4 5 e12.5/ i
i
0.035
1
1
3
.
=
i
opt + 0.08 + 1
The optimal tip speed ratio is defined as
topt r
opt =
vw

The maximum (optimal) power obtained from a given wind

speed is commonly expressed by the following equation:
(1)

with Cpopt being the optimal power coefficient of the wind turbine for a given wind speed, Ar (in square meters) the effective
area covered by the turbine blades, w (in meters per second)
the wind speed, and (in kilograms per cubic meter) is the air
density. The calculation of the optimal power coefficient Cpopt

(2)
(3)

(4)

where r (in meters) is the blade radius and topt is the optimal
wind turbine rotor speed (in radians_mechanical per second) for
a given wind speed. From (1) and (4), an optimum (maximum)
power value Popt can be obtained as a function of the shaft speed,
referred to as the generator side of the gearbox, as follows:
Popt = Kopt r3

(5)

where Kopt = (/2)(Cpopt /3opt )r5 , r = (P/2)Gt , P being the number of poles and G the gear ratio. From (5), the
wind turbine predefined optimal extraction power curve can be
established for a given Kopt associated to a fixed blade angle.
In this research, Kopt is defined for a fixed blade angle equal
to 0 . In Fig. 5, the maximum power extraction curve adopted
for a variable-speed wind turbine is shown.
The speed controller is controlling the electromechanical
torque (Te ). The reason for such control is that the torque is
directly dependent on the quadrature component of the rotor
current (iq r ), when stator resistant is neglected [16]. The following relation between torque and iq r holds, in which et is the
terminal voltage:
Te =

B. Maximum Power Extraction and Speed Controller

Popt = 0.5Cpopt (opt , )Ar vw3

Fig. 5.

Lm et
iq r .
Ls + Lm

(6)

From (5), the relation between the electromechanical torque

and the rotor speed can be derived as follows:
Te opt = Kopt r2 .

(7)

From (6) and (7), the rotor speed controller is implemented by

measuring the actual rotor speed and then the value of iq r needed
to realize the desired electromechanical torque is calculated. The
resulting iq r (iq r ref ) is fed into the DFIG through Vq r component
using a PI controller (see Fig. 6). This control algorithm is
applied up to the maximum mechanical power of the turbine.

SHAHABI et al.: MICROGRID DYNAMIC PERFORMANCE IMPROVEMENT USING A DOUBLY FED INDUCTION WIND GENERATOR

Fig. 6.

141

Rotor speed controller (maximum power extraction).

Fig. 8.

Terminal voltage controller of the DFIG.

control will control the rotor speed and operate until the wind
cutoff speed limit is reached. A typical pitch control can be found
in [15]. The input speed reference for a pitch control is defined
according to if Pm ec > Pm ax , and then r -ref = (Pm ax /Tm ec ).
Fig. 7.

Integrated active power and primary frequency control of the DFIG.

E. Reactive Power Control and Terminal Voltage Regulator

C. Primary Frequency Control
Variable-speed wind turbines are controlled to capture as
much power as possible from the wind. However, the large
blades of the turbine give this type of DG unit a significant inertia. The kinetic energy stored in their inertia gives the turbine the
possibility to support primary frequency control for a short period. One particular advantage of variable-speed wind turbines
is that unlike thermal or gas turbine power plants, they can increase their output power almost instantaneously [17]. This is
an important feature, especially for a microgrid system when it
goes to islanding-mode operation.
The primary frequency control integrated into the rotor-side
speed (active power) control loop is similar to the one usually
used in a synchronous generator. In this case, the droop (characterized by a regulation R, expressed in p.u. in the system
base) is used to produce a change in the active power injected
by the DFIG when a system frequency variation occurs. The
active power increment is therefore proportional to the system
frequency variation and is defined as follows:


1
(8)
(sys sys -ref ).
Pinjected = Pref
R
This generated active power requires a new operating point of
the DFIG. The primary frequency control that integrated into the
rotor-side speed control loop is implemented by converting the
required active power to equivalent electromechanical torque
(see Fig. 7). Note that this feature (primary frequency control)
is activated when an islanding condition occur. The variable sys
is obtained from the grid connection bus voltage of the DFIG
using a phase lock loop (PLL).
D. Pitch Control
For wind speeds below maximum rated speed (according to
the operational optimal power curve), the input reference electromechanical torque to the active power (rotor speed) control
loop in DFIG follows (6) and (7), as discussed in the previous
section.
However, for wind speed larger than the maximum rated
speed, the blade pitch regulation dominates the control and limits the energy captured from the turbine. In this case, the pitch

A variable-speed wind turbine with DFIG has the capability

to control its terminal voltage. It is an important feature, especially for a microgrid islanding operation. By this feature, the
DFIG can contribute the reactive power management and voltage regulation of the microgrid. A terminal voltage controller is
being developed here. It can be applied in two operation modes
of a microgrid (grid-connected mode and islanding operation
mode). When stator resistance is neglected, the reactive power
generated by the wind turbine is directly dependent on idr [16].
It can be shown that the relation between the reactive power
generated and idr is as follows:

e2t
s (Ls + Lm )
(9)
where the direct component of the rotor current has been split
into a part that magnetizes the generator, idr,m agn , and a part
that determines the net reactive power exchange with the grid,
idr,gen . The direct component of the rotor current necessary to
magnetize the generator itself has the following value:
Qgrid =

Lm
Ls + Lm

(idr,m agn + idr,gen )

idr,m agn =

e2t
.
s Lm

(10)

The value of idr ,gen determines whether the net reactive power
is generated or consumed. The terminal voltage will increase
when more reactive power is delivered to the grid. The voltage
controller should have the following features.
1) In the grid-connected mode operation of the microgrid,
the DFIG should be operated at unity power factor.
2) Terminal voltage of the DFIG should be adjusted appropriately when the microgrid switches to islanding mode.
A voltage controller that satisfies the aforementioned features
is proposed and depicted in Fig. 8.
From Fig. 8, it can be seen that during the grid-connected
mode of operation, idr,ref provides the required magnetization
current for the DFIG. Also, in the islanding mode of operation,
in addition to providing the required magnetizing current, the PI
controller regulates idr,ref in a way that it provides the required
reactive power.

142

Fig. 9.

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 24, NO. 1, MARCH 2009

Real wind speed pattern.

VI. TEST CASES AND SIMULATION RESULTS

Two case studies have been demonstrated to show the ability
of DG controllers (especially for a DFIG) to continue the
microgrid operation during separation and in the islanded mode.
Case studies are chosen to illustrate both the steady-state and
dynamic responses to the changes in the system operating point.
A. Planned (Intentional) Islanding Event
The objective of this study is to investigate dynamic behavior
of the microgrid with DGs due to a planned islanding scenario.
Prior to islanding, DG1 supplies 1.3 MW/0.05 Mvar (close to
unity power factor) and the difference between the power generated by DG2 and the load demand of 3.14 MW/1.63 Mvar is
imported from the grid (0.45 MW/1.57 Mvar). DG2 is operated
at unity power factor, and a real wind pattern [18] with average
wind speed of 10.3 m/s, as shown in Fig. 9, was used as input
to the wind turbine.
In this situation, about 86% of the total load in microgrid is
supplied by DG units, and the rest is imported from the main
grid. Both the gas-turbine synchronous generator unit (DG1)
and the variable-speed wind turbine with DFIG unit (DG2) use
a 5% droop value (in each machine base).
At t = 4.0 s, the microgrid is disconnected from the main
grid by initiating an intentional islanding command that opens
CB2 in Fig. 1.
Fig. 10 shows the system transients due to the planned islanding event. Fig. 10(d) illustrates that the bus voltages recovered
to its normal range in 70 ms. It is due to the contribution of
both DFIG terminal voltage controller and excitation system of
the synchronous generator unit in reactive power compensation
and voltage regulation of the islanded network (the microgrid
system).
Fig. 10(c) also shows power sharing of the DG units after
islanding based on power control strategy of the DG units in
microgrid. DG1 adjusts its real power output through the action
of its governor, according to its dynamic response time. Also,
the restoration of microgrid frequency is accomplished by DG1
through frequency restoration control loop (integral frequency
control) that integrated in its governor system.

Fig. 10.

Planned islanding event with primary frequency control of the DFIG.

DG2 supports the primary frequency control of the islanded

system (microgrid) through its proposed controller. Hence, the
frequency deviation in the islanded system is acceptable, and it
can be seen that the microgrid system frequency restores to the
reference value of 50 Hz [see Fig. 10(a)]. During the islanding
operation, the real power output of DG1 varies based on the
changes in the output power of DG2 due to variation in the wind
power that changes the speed of synchronous generator unit [see
Fig. 10(b)].
Fig. 11 shows the case that primary frequency control capability of the DG2 is deactivated. From Figs. 10 and 11, it is clear
that due to primary frequency control feature of the DG2 unit, an
improved frequency control of the microgrid can be achieved.
The results of this case study indicate that the proposed DFIG
controllers can contribute to both voltage and frequency controls
of a microgrid during separation and islanding operation. This
leads to a good dynamic behavior of the microgrid during and
subsequent to a planned islanding event.
B. Unplanned (Unintentional) Islanding Event
In this case study, it is assumed that an unplanned islanding event of the microgrid system takes place due to a fault
occurrence in the upstream section of the MV feeder. Prior to
islanding, the operating conditions of the system are the same
as those of the planned islanding case.
At t = 4.0 s, a three-phase fault occurs on bus 2 in the upstream section of the supply MV feeder. Considering the proposed system protection scheme, after about 150 ms, the fault
is diagnosed, and an opening command is issued for CB2 from
a distribution network operator (DNO). After opening the CB2,
synchronous reclosing operation of the upstream section can

SHAHABI et al.: MICROGRID DYNAMIC PERFORMANCE IMPROVEMENT USING A DOUBLY FED INDUCTION WIND GENERATOR

Fig. 11.
DFIG.

Planned islanding event without primary frequency control of the

Fig. 13.
DFIG.

143

Unplanned islanding event without primary frequency control of the

During the fault, bus voltages drop severely [see Fig. 12(d)].
After the islanding detection, fast control actions of the DG
units terminal voltage controllers eventually return the voltages
to their normal range. Fig. 13 illustrates the system response
when the DFIG primary frequency control is disabled.
From Figs. 12 and 13, it is observed that the primary frequency
control responds quickly and helps to improve the system frequency response slightly.
A better damping performance is obtained for the case where
DFIG has a primary frequency control capability, as shown in
Fig. 12(a), because after islanding detection, the DFIG participating in the primary frequency control leads to an increase
in active power injection. This fact makes it possible for DG1
(synchronous generator unit) to respond by injecting less active
power into the microgrid, thereby improving the system frequency smoothly. It can be seen that when power increases, the
DFIG rotor active power also increases, and consequently, the
rotor speed tends to fall a little more after islanding detection
balancing the total active power (stator plus rotor active power)
injected by the DFIG, as illustrated in Figs. 12(b) and 13(b).
Fig. 12.
DFIG.

Unplanned islanding event with primary frequency control of the

be made available by CB1, and also islanding operation of the

microgrid system will be allowed.
The islanding phenomena is detected 100 ms after CB2 opens,
i.e., at t = 4.25 s. At this moment, the operation mode of DG
units in the microgrid system switches to the islanding mode.
Fig. 12 shows the system transients during the unplanned
islanding scenario.

VII. CONCLUSION
In this paper, the impact using a DFIG with capability to
contribute to primary frequency control and terminal voltage
regulation in a microgrid system was addressed. Two control
modes were adopted for DGs in a microgrid system, and both
planned and unplanned islanding scenarios were examined to
evaluate the control approach. The microgrid is supplied by
two DG units, i.e., a synchronous machine and a variable-speed
wind turbine with DFIG technology. The latter unit includes

144

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 24, NO. 1, MARCH 2009

both voltage and primary frequency controls with a fast response

time.
A detailed description of the applied control approach was
provided and discussed. The simulation results show that the
proposed controllers for DFIG can lead to a better dynamic
behavior of the microgrid when islanded from the main grid and
operate in islanding mode.
It was shown that having a variable-speed wind turbine with
DFIG in a microgrid participating in the frequency and voltage
controls may improve the microgrid systems dynamic performance, reducing the frequency changes following disturbances
that are subsequent to islanding.

C. Control parameters:

APPENDIX

Kopt = 0.628; Kpd = Kpq = 0.5, Kid = Kiq = 100;

Kpv = 0.1, and Kiv = 100.
It should be noted that the parameters of PI controllers are
obtained based on the trial-and-error approach. Different values
for different system conditions were investigated, and finally,
the proposed values were selected.

REFERENCES

A. DG1 (Gas-Turbine Synchronous Generator) Parameters

1) Synchronous machine characteristics:
Sn = 1.8 MVA;
Vn = 11 kV;
Xd = 1.56 p.u.;
Xd = 0.29 p.u.;
Xd = 0.17 p.u.;
Xq = 1.07 p.u.;
Xq = 0.35 p.u.;
Xq = 0.17 p.u.;
Xl = 0.055 p.u.;
Td = 3.5 s;
Td = 0.035 s;
Tq = 0.037 s;
Rs = 0.0036 p.u.;
H = 1.04 s.
2) AVR parameters: KA = 400, TA = 0.02 s.
3) Turbine-governor total time constant: about 450 ms.
B. DG2 (Variable-Speed Wind Turbine With DFIG) Parameters
1) Wind Turbine characteristics:
Pn = 2 MW;
number of blades = 3;
rotor diameter = 80 m;
air density = 1.255 kg/m3 ;
gear box ratio = 1:85;
inertia time constant = 3.5 s;
nominal wind speed = 12 m/s;
cut-in wind speed = 4.5 m/s;
cutoff wind speed = 25 m/s.
2) DFIG parameters:
Sn = 2.5 MVA;
Vn = 0.69 kV;
Rs = 0.01 p.u.;
Xs = 0.1 p.u.;
Rr = 0.01 p.u.;
Xr = 0.08 p.u.;
Xm = 3.0 p.u.;
number of poles = 4;
s = 1500 r/min.

[1] M. W. Davis, Distributed resource electric power systems offer significant advantages over central station generation and T&D power systems,
part I, in Proc. 2001 IEEE Transmiss. Distrib. Conf. Expo., pp. 54
61.
[2] N. D. Hatziargyriou and A. P. S. Meliopoulos, Distributed energy sources:
Technical challenges, in Proc. 2002 IEEE PES Winter Meeting, pp. 1017
1022.
[3] C. L. Smallwood, Distributed generation in autonomous and nonautonomous microgrid, in Proc. 2002 IEEE Rural Elect. Power Conf.,
pp. D1-1D1-6.
[4] H. You, V. Vittal, and Z. Yang, Self-healing in power systems: An approach using islanding and rate of frequency decline-based load shedding,
IEEE Trans. Power Syst., vol. 18, no. 1, pp. 174181, Feb. 2002.
[5] R. Fulton and C. Abbey, Planned islanding of 8.6 MVA IPP for BC
Hydro system reliability, in Proc. 2004 First Int. Conf. Integr. RE DER,
pp. 19.
[6] F. Katiraei, M. R. Iravani, and P. W. Lehn, Micro-grid autonomous operation during and subsequent to islanding process, IEEE Trans. Power
Del., vol. 20, no. 1, pp. 248257, Jan. 2005.
[7] F. Katiraei, M. R. Iravani, and P. W. Lehn, Small-signal dynamic model
of a microgrid including conventional and electronically interfaced distributed resources, IET Gener. Transmiss. Distrib., vol. 1, no. 3, pp. 369
378, May 2007.
[8] The European Wind Energy Association (EWEA). (1999). Green peace
and wind industry unveil global energy blue print [Online]. Available:
http://www.ewea.org/src/press.htm
[9] S. K. Salman and A. Teo, Windmill modeling consideration and factors
influencing the stability of a grid-connected wind power-based embedded
generator, IEEE Trans. Power Syst., vol. 18, no. 2, pp. 793802, May
2003.
[10] J. Ekanayake and N. Jenkins, Comparison of the response of doubly fed
and fixed-speed induction generator wind turbines to changes in network
frequency, IEEE Trans. Energy Convers., vol. 19, no. 4, pp. 800802,
Dec. 2004.
[11] W. J. Rowen, Simplified mathematical representations of heavyduty gas turbines, ASME J. Eng. Power, vol. 105, pp. 865869,
1983.
[12] IEEE Standard for Interconnecting Distributed Resources with Electric
Power Systems, IEEE Standard 1547, Jul. 2003.
[13] S. M. Brahma and A. A. Girgis, Development of adaptive protection
scheme for distribution systems with high penetration of distributed
generation, IEEE Trans. Power Del., vol. 19, no. 1, pp. 5663, Jan.
2004.
[14] S. A. Papathanassion and M. P. Papadopoulos, Dynamic behavior of
variable speed wind turbines under stochastic wind, IEEE Trans. Energy
Convers., vol. 14, no. 4, pp. 16171623, Dec. 1999.
[15] J. G. Slootweg, H. Polinder, and W. L. Kling, Dynamic modeling of a
wind turbine with doubly fed induction generator, in Proc. 2001 IEEE
Power Eng. Soc. Summer Meeting, pp. 644649.
[16] S. Heier, Grid Integration of Wind Energy Conversion Systems. Chicester,
U.K.: John Wiley, 1998.
[17] H. Knudsen and J. N. Nielsen, Introduction to the modeling of wind
turbines, in Wind Power in Power Systems, T. Ackerman, Ed. Chicester,
U.K.: Wiley, 2005, pp. 525585.
[18] Database of wind characteristics located at DTU, Denmark. (2006).
[Online]. Available: http://www.winddata.com

SHAHABI et al.: MICROGRID DYNAMIC PERFORMANCE IMPROVEMENT USING A DOUBLY FED INDUCTION WIND GENERATOR

M. Shahabi (S07) received the B.Sc. degree in electrical engineering from

Tabriz University, Tabriz, Iran, in 1998, and the M.Sc. degree in power engineering in 2001 from Tarbiat Modarres University, Tehran, Iran, where he is
currently working toward the Ph.D. degree.
His current research interests include distribution system operation and control in the presence of embedded generation resources.

M. R. Haghifam (M98SM06) received the B.Sc. degree from Tabriz University, Tabriz, Iran, in 1988, the M.Sc. degree from Tehran university, Tehran,
Iran, in 1990, and the Ph.D. degree from Tarbiat Modares University, Tehran,
in 1995.
In 1995, he joined Tarbiat Modares University, where he is currently a
Professor of electric power systems. His current research interests include power
system reliability, electric distribution systems, and soft computing application
in power systems.
Prof. Haghifam is a Research Fellow of the Alexander Von Humboldt
Foundation.

145

M. Mohamadian (M05) received the B.Sc. degree from Amirkabir University

of Technology, Tehran, Iran, in 1989, the M.Sc. degree from Tehran University,
Tehran, in 1992, and the Ph.D. degree from the University of Calgary, Calgary,
AB, Canada, in 1997, all in electrical engineering.
From 1997 to 2000, he was a Software Engineer with OPISystems, Inc.,
Calgary, designing real-time microprocessor programs for industrial control
and remote data acquisition systems. From 2000 to 2005, he was with the
Scientific Research Laboratories, Iranian Research Organization for Science
and Technology (IROST), Tehran, where he worked on electric drives. Since
2005, he has been with Tarbiat Modares University, Tehran. His current research
interests include DSP simulation and implementation of control algorithms on
electric drives and application of power electronics in distributed generation.

S. A. Nabavi-Niaki (M92SM04) received the B.Sc. and M.Sc. degrees from

Amirkabir University of Technology, Tehran, Iran, in 1987 and 1990, respectively, and the Ph.D. degree from the University of Toronto, Toronto, ON,
Canada, in 1996, all in electrical engineering.
In 1996, he joined the University of Mazandaran, Babolsar, Iran. He is currently with the University of Toronto. His current research interests include
analysis, operation and control of power systems, and flexible ac transmission
system (FACTS) controllers.