Electric Power Systems Research 80 (2010) 4652

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Modeling and control of PMSG-based variable-speed wind turbine

Hong-Woo Kim a , Sung-Soo Kim b , Hee-Sang Ko a,
a
b

Wind Energy Research Center, Korea Institute of Energy Research, Yuseong-gu Jang-Dong 71-2,305-343 Daejeon, Republic of Korea
Chungbuk National University, Republic of Korea

a r t i c l e

i n f o

Article history:
Received 19 July 2008
Received in revised form 1 August 2009
Accepted 5 August 2009
Available online 22 October 2009
Keywords:
Permanent-magnetic synchronous
generator
Variable speed
Wind turbine
Wind farm

a b s t r a c t
This paper presents a control scheme of a variable-speed wind turbine with a permanent-magnetic
synchronous generator (PMSG) and full-scale back-to-back voltage source converter. A comprehensive
dynamical model of the PMSG wind turbine and its control scheme is presented. The control scheme
comprises both the wind-turbine control itself and the power-converter control. In addition, since the
PMSG wind turbine is able to support actively the grid due to its capability to control independently active
and reactive power production to the imposed set-values with taking into account its operating state and
limits, this paper presents the supervisory reactive power control scheme in order to regulate/contribute
the voltage at a remote location. The ability of the control scheme is assessed and discussed by means of
simulations, based on a candidate site of the offshore wind farm in Jeju, Korea.
2009 Elsevier B.V. All rights reserved.

1. Introduction
VARIABLE-SPEED power generation enables the operation of the
turbine at its maximum power coefcient over a wide range of
wind speeds, obtaining a larger energy capture from the wind with
a power converter which allows variable-speed operation. One of
the problems associated with variable-speed wind systems today
is the presence of the gearbox coupling the wind turbine (WT) to
the generator. This mechanical element suffers from considerable
faults and increases maintenance expenses. To improve reliability
of the WT and reduce maintenance expenses the gearbox should
be eliminated.
Megawatt (MW) class wind turbines equipped with a
permanent-magnetic synchronous generator (PMSG) have been
announced by Siemens Power Generation and GE Energy. In this
concept, the PMSG can be directly driven or have smaller gearboxes
or even gearless and is connected to the ac power grid through the
power converter. Use of the power converter is essential because
it allows the linkage of the generator operating at variable speed
to the ac power grid at a xed electrical frequency. The converter
rating must be similar to or even larger than the rated power of the
generator. Permanent-magnetic excitation allows to use a smaller
pole pitch than do conventional generators, so these machines can
be designed to rotate at rated speeds of 20200 rpm, depending on
the generator rated power [1].

Corresponding author. Tel.: +82 550 630 1670.

E-mail address: heesangko@gmail.com (H.-S. Ko).
0378-7796/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsr.2009.08.003

However, the electromagnetic construction of the PMSG is more

complex than in the case of conventional WT concepts such as
xed-speed with squirrel induction generators and variable speed
with doubly fed induction generators, etc. Also, the reduced gear
ratio may require an increase in the number of generator pole pairs,
which complicates the generator construction [18].
MW class wind turbines (WTs) have been commissioned in
large (offshore) wind farms connected directly to transmission
networks. However, increased wind power generation has inuenced the overall power system operation and planning in terms
of power quality, security, stability, and voltage control [914].
The local power ow pattern and the systems dynamic characteristics change when large WTs are connected to the utility
grid [15]. Thus, the compliance with the grid codes of national
Transmission System Operators (TSOs) becomes an important issue
[16].
Therefore, the interaction between wind farms (WFs) and power
systems is a research topic that needs more attention. To get a better
understanding of how the control systems of the individual WTs
and the WFs inuence each other, modeling and simulation are
essential. To investigate the interaction between controllers of WTs
or WFs and the controllers of the grid is considered a challenge.
With more advanced control algorithms, WTs and WFs can provide
ancillary services to the grid, e.g. by providing reactive power or
participate in voltage/frequency control. To study the impacts of
these advanced control strategies on a system level, more modeling
efforts are required.
Therefore, this paper presents the detail system modeling and
the control design of a PMSG-based-WT. As well as alternative
design and/or control solutions are proposed to improve the voltage

H.-W. Kim et al. / Electric Power Systems Research 80 (2010) 4652

47

Fig. 2. Permanent-magnetic synchronous generator wind turbine (PMSG-WT).

Fig. 1. Grid-connected wind-turbine system.

control at a required location such as a point-of-common coupling

(PCC).
This paper is organized as follows: the detail dynamic model
including voltage source converter (VSC) control design is presented in Section 2; in Section 3, the supervisory reactive power
control scheme is proposed; case studies are carried out in Section
4; and conclusions are drawn in Section 5.

sonable, special attention should be given to model development. In

particular, in this paper, to increase the simulation speed of various
electrical components, these components are modeled in the dqsynchronous reference frame [20]. Wherein, the d-axis is assumed
to be aligned to the stator ux, and the current coming out of the
machine is considered positive. The PMSG controllers utilize the
concept of disconnection of the active and reactive power controls
by transformation of the machine parameters into the dq-reference
frame and by separating forming of the stator voltages. Then, the
active power can be controlled by inuencing the d-axis component
of the stator current while the reactive power can be controlled by
inuencing the q-axis components of the stator current. The system
parameters and control gains, etc., are summarized in Appendix A.
2.1. Permanent-magnetic synchronous-generator (PMSG)

2. Dynamic model of PMSG-WT-based power system

The PMSG was represented by the following equations [1]:
The system considered in this paper is shown in Fig. 1. The WF
consists of 5 unit of WT. Each WT is equipped with a 0.69/22.9 kV
step-up transformer (TR). The WF is connected to the grid using
a 2 km submarine cable (Ca) and a 14 km overhead transmission
line (TL). The considered operating condition is as follows: the WF
supplies 7 MW of active power and 0.3 MVar of reactive power to
the local load, which consumes 8 MW and 1.9 MVar. The remaining active power comes through the 154 kV utility grid, which is
represented by an innite bus (Table 1).
Although the fundamental principle of a WT is straightforward,
modern WTs are very complex systems. The design and optimization of the WTs blades, drive train, and tower require extensive
knowledge of aerodynamics, mechanical and structural engineering, control and protection of electrical subsystems, etc.
The details of the WT considered in the model are shown in
Fig. 2. The WT consists of the following components: a three-bladed
rotor with the corresponding pitch controller [17]; a PMSG with
two converters, a dc-link capacitor, and a grid lter; and converter
controllers.
Although personal computers are becoming increasingly faster,
computational speed is still one of the limiting factors in dynamic
simulation of power systems [18,19]. Electrical transients have very
small time constants that require small integration time steps and
result in long computation time. To keep the simulation speed rea-

1 d ds
= vd1 + Rs ids + e
b dt

qs ,

1 d qs
= vq1 + Rs iqs e
b dt

ds

with
ds

= Lds ids

m,

qs

= Lqs iqs

Load
Bus: 4 (PCC); v (pu): 1.1435.

p (pu)

q (pu)

0.419
3.562
3.981

0.8039
0.1687
0.9725

Resistance

Reactance

0.3099

0.0757

(2)

where v is the voltage, R is the resistance, i is the current, e is

the stator electrical angular speed, b is the base angular speed in
rad/s, Ls is the stator leakage inductance, m is the exciter ux of
the PMSG, and
is the ux linkage. The subscripts d and q indicate the direct and quadrature axis components, respectively. The
subscripts s indicates stator quantities. The electrical active and
reactive power delivered by the stator are given by
Ps = vd1 ids + vq1 iqs ,

Qs = vd1 iqs vq1 ids

(3)

The mathematical model of a TL, a TR, a cable, and a load can be

obtained from the description of the R, L, C segment [20] into the
dq-synchronous reference frame. The equations of the TL, the TR,
the cable, and the RL load are given in (4)(7) based on Fig. 3 where
superscript s and e stand for the sending-end and the receiving-end.

Table 1
Operating conditions.

IB
WF
Total

(1)

Fig. 3. Lumped-parameter  equivalent-circuit description in the dq-domain.

48

H.-W. Kim et al. / Electric Power Systems Research 80 (2010) 4652

For example, in the TL model (see. Eq. (4)) vsd corresponds to vd3 at
bus 3 and ved corresponds to vd4 at bus 4.
2.2. Transmission line (TL)
LTL didl
= vd4 vd3 RTL idl + e LTL iql ,
b dt
LTL diql
= vq4 vq3 RTL iql e LTL idl ,
b dt
CTL dvd3
s
+ e CTL vq1 ,
= idc
b dt

CTL dvq3
s
e CTL vd3 ,
= iqc
b dt

CTL dvd4
e
+ e CTL vq4 ,
= idc
b dt

CTL dvq4
e
e CTL vd4
= iqc
b dt

Fig. 4. Block diagram of the VSC controller showing the input/output variables.

(4)

2.3. Transformer (TR)

Ltr didl
= vd2 vd1 Rtr idl + e Ltr iql ,
b dt
Ltr diql
= vq2 vq1 Rtr iql e Ltr idl ,
b dt
Co dvd1
= idl + e Co vq1 ,
b dt

Co dvq1
= iql e Co vd1
b dt

(5)

where Co is the dummy capacitor to obtain voltage for modeling

purpose.
2.4. Cable

2.7. Voltage source converter controller

Lca didl
= vd3 vd2 Rca idl + e Lca iql ,
b dt
Lca diql
= vq3 vq2 Rca iql e Lca idl ,
b dt
Cca dvd2
s
+ e Cca vq2 ,
= idc
b dt

Cca dvq2
s
e Cca vd2
= iqc
b dt

(6)

2.5. RL load
The RL load in the dq-domain can be described as
Lload didL
= vd4 Rload idL + e Lload iqL ,
b dt
Lload diqL
= vq4 Rload iqL e Lload idL ,
b dt
Co dvd4
= idL + e Co vq4 ,
b dt

Fig. 5. WT maximum energy-harvesting curve.

Co dvq4
= iqL e Co vd4
b dt

(7)

Fig. 4 presents the detailed block diagram of the VSC controller

depicting the respective input and output variables. Here, Pgset is
the set-value for the active power for the WT terminal. The value
of Pgset is determined from the WT energy-harvesting characteristic as shown in Fig. 5, which is represented here as a look-up
table Pgset (r ) determined in terms of generator rotational-speed r .
Since variable-speed WTs are traditionally operated in the power
factor control (PFC) mode to achieve the unity power factor at the
terminal of the WT, the reactive power set-points Qgset is set to zero.
The VSC control module consists of the generator-side, the dclink, and the grid-side converter controller. These controllers utilize
proportional-integral (PI) controllers. These PI controllers are tuned
using the Nyquist constraint technique to deal with model uncertainties [21,22]. Each of the controllers is briey described below.
Generator-side converter controller: Fig. 6 shows a block diagram
of the generator-side converter controller module, which includes
four internal PI controllers, PI1 through PI4. The controller is implemented as two branches, one for the active power (PI1 and PI2) and
one for the reactive power (PI3 and PI4) with the corresponding
de-coupling terms between the d and q axes, respectively.

As shown in Fig. 2, the grid-side converter is connected to the

grid through the lter. The voltage equations for the lter in the
dq-synchronous reference frame are as follows.
2.6. RL-lter on the grid-side converter
Llt didg
= vd1 Rlt idg + e Llt iqg ,
b dt
Llt diqg
= vq1 Rlt iqg e Llt idg
b dt
where subscript lt stands for lter.

(8)
Fig. 6. Block diagram of the generator-side converter controller.

H.-W. Kim et al. / Electric Power Systems Research 80 (2010) 4652

49

Fig. 8. DC-link model and its controller.

Fig. 7. Block diagram of the grid-side converter controller.

The transfer function from the stator voltage to the stator current is approximated as

Ids (s)
 (s)
Vds

Iqs (s)
 (s)
Vqs

T

1
Rs + s(Lds /b )

1
Rs + s(Lqs /b )

T
(9)

Similarly, the transfer function from the stator current to reactive and active power is approximated as

Qg (s)
Iqs (s)

Ps (s)
Ids (s)

T
=

Lqs
Rs + s
b

L
Rs + s ds
b

T
(10)

Then, (9) is used to tune PI2 and PI4, and (10) is used to tune PI1
and PI3.
Grid-side converter controller: Fig. 7 shows a block diagram of
the grid-side converter controller module, which also includes two
internal PI controllers PI5 and PI6, with corresponding de-coupling
terms between the d and q axes.
The voltage equation for the grid-side converter RL-lter can be
expressed as

 L  di
dg
lt
b

dt

dt

 L  di
qg
lt

= vd1 Rlt idg + e Llt iqg ,

= vq1 Rlt iqg e Llt idg

Idg (s)

T 
Iqg (s)

Vd1 (s)

Vq1 (s)

1
Rlt +s(Llt /b )

1
Rlt +s(Llt /b )

T
(12)

The inputs to the grid-side controller are the set-values for the
currents, which ows to the grid through the VSC. The set-values of
the input currents are calculated by the active and reactive power
commands Psset and Qgset as follows:

set
iqg
set
idg


=

vq1

vd1

vd1

vq1

1 

Psset
Qgset


(13)

where Psset and Qgset are the set-point of the active and reactive
power commands. The value for Psset is provided by the dc-link controller, which determines the ow of active power and regulates the
dc-link voltage by driving it to a constant reference value.
DC-link dynamic model and its controller: The capacitor in the
dc-link is an energy storage device. Neglecting losses, the time
derivative of the energy in this capacitor depends on the difference
of the power delivered to the grid lter, Pg , and the power provided
by the stator circuit of the PMSG, Ps , which can be expressed as
2
1 Cdc dvdc
= Pg Ps
2 b dt

controller PI7. The set-point for the output active power by Psset =
set .
vdc idc,s
3. The supervisory reactive power control
The purpose of the supervisory reactive power control presented in this section is to regulate the voltage at the specied
remote PCC (see Fig. 1) by adjusting the reactive power produced by
the grid-side converter, taking into account its operating state and
limits. As shown in Fig. 9, the control objective is to utilize Qj from
the grid-side VSC to control the voltage at the PCC to the predened
value by the reactive power set-point control signal Qjset .
When controlling WT, it is important that the operating limit of
WT is not exceeded. The reactive power required from an individual
grid-side converter of the VSC can be computed as

(11)

from which the transfer function from the lter voltage to current
is

Fig. 9. Schematic diagram of the supervisory reactive power control.

Qjset = min

Qjmax ,

Qjmax

max Qpcc

Q1max + + Q5

where j = 1, . . . , 5, Qjmax is the maximum reactive power (limit)

that the jth grid-side converter can provide, and Qpcc is the total
reactive power required to support the voltage at the PCC.
Fig. 10 shows the active and reactive power operating limits, wherein it is assumed that the grid-side converter should not
exceed its apparent power limit Sjmax depicted by the half-circle.
Suppose that at a given time each grid-side converter is delivering
the active power denoted herein by Pj . Then, in addition to the active
power, the converter can supply or absorb a maximum of Qjmax of
the reactive power. Therefore, the reactive power available from the
grid-side converter lies within the limits [Qjmax ; + Qjmax ], which
are operating-condition dependent.

(14)

The dc-link controller regulates the capacitor voltage by driving

it to the reference value vref
, and outputs the set-point for the active
dc
power Psset needed in (13). Fig. 8 shows the dc-link model with its

(15)

Fig. 10. VSC active and reactive power operating limits.

50

H.-W. Kim et al. / Electric Power Systems Research 80 (2010) 4652

Fig. 11. Implementation of PI controller with the distributed anti-windup.

Fig. 13. Voltage observed at the PCC due to the wind-speed variation.

Thus, the maximum available reactive power from the each gridside converter can be expressed as
Qjmax =

(Sjmax )2 Pj2

(16)

where it is assumed that the nominal apparent power of the each

converter is Sjmax , dened here as the WT rating. Based on Fig. 10,
it also follows that Sjmax Pj Sjmax . Thus, the maximum reactive

power set-point of Qjset (see Fig. 4) can be determined by (15) and

(16).
Finally, a proportional-integral (PI) controller is designed for a
controller shown in Fig. 9. The PI gains are summarized in Appendix
A. Since limiting control action should be implemented together
with the integrator-anti-windup scheme that would stop integrating the error when the limit is being reached, a PI controller
with the proposed distributed anti-windup is implemented in Matlab/Simulink [23] as shown in Fig. 11 for case studies.
4. Case studies

The system depicted in Fig. 1 was implemented in detail using

the Matlab/Simulink [21]. Computer studies considering the windspeed variations, the local-load variations, and the voltage sag due
to the fault were conducted to compare the dynamic responses of
the system with different controls. In comparison, Mode 1 indicates the PFC-mode operation of the grid-side converter of the WT,
which Qgset is set to zero. As the proposed operation, Mode 2 actively
utilizes Qgset from the grid-side converter for voltage control at the
PCC.
4.1. Wind-speed variation
In this study, the wind speed shown in Fig. 12 was considered
for the WTs. Fig. 13 shows the voltage at the PCC, predicted by the
model with different controls, respectively. As shown in Fig. 13,
Mode 1 operation caused the voltage deviation about 3%, which
is much higher than the permissible voltage range of HV power

Fig. 12. Wind speed (m/s).

Fig. 14. Active and reactive power from WF to PCC for wind-speed variation.

system network 2% while Mode 2 operation achieved the voltage

regulation at the PCC. Fig. 14 shows the measured data of the active
power and the reactive power from the WF to the PCC (see Fig. 1).
The reactive power contribution from the WTs is the difference
between Mode 2 and Mode 1.
4.2. Local-load variation
For this study, the local-load impedance is decreased by 20%
with wind speed 12 m/s. The comparison of the voltage transients
observed at the PCC was showed in Fig. 15. As can be noticed, when

Fig. 15. Voltage observed at the PCC due to the 20% impedance decrease.

H.-W. Kim et al. / Electric Power Systems Research 80 (2010) 4652

51

Fig. 19. Voltage observed at the PCC due to the fault.

Fig. 16. Active and reactive power from WF to PCC for the load variation.

at the PCC from Mode 1 operation. Fig. 16 shows the measured data
of the active power and the reactive power from the WF to the PCC.
The reactive power contribution from the WTs is the difference
between Mode 2 and Mode 1.
4.3. Voltage sag in the innite bus
To emulate this scenario, it is assumed that there was a fault at
t = 0.5 s in the network that caused 10% voltage drop at the innite
bus with wind speed 12 m/s. As can be noted in Fig. 17, Mode 1 operation showed the signicant voltage drop by 14.5% while Mode 2
operation resulted in the voltage recovery to its predened voltage
at the PCC. Fig. 18 shows the measured data of the active power
and the reactive power from the WF to the PCC.
4.4. Fault ride-through study

Fig. 17. Voltage observed at the PCC due to voltage sag in the innite bus.

Mode 1 was in operation, the load impedance changes resulted in

noticeable drop of the bus voltage by 4.5% which does not satisfy the
permissible voltage range 2%. When the WT operated in Mode 2,
the voltage recovery to its predened value was achieved. Thus, the
performance in Mode 2 operation has been signicantly improved

Fig. 18. Active and reactive power from WF to PCC for the voltage sag.

A three-phase symmetrical fault was assumed in the middle of

TL with wind speed 12 m/s. To emulate this fault scenario, the fault
was assumed at t = 0.2 s and was subsequently cleared at t = 0.36 s
by restoring the initial TL impedance. As can be noted in Fig. 19,
the fault resulted in signicant voltage swings that can undesirably interfere with the protection circuitry and possibly trip the

Fig. 20. Active and reactive power from WF to PCC due to the fault.

52

H.-W. Kim et al. / Electric Power Systems Research 80 (2010) 4652

WT. From this point of view, it is desirable to minimize and/or suppress the voltage swings. During the fault, the voltage drop has
been slightly improved in Mode 2. After the fault was cleared, faster
voltage recovery to reach to its predened voltage at the PCC was
noticed in Mode 2.Fig. 20 shows the measured data of the active
power and the reactive power from the WF to the PCC. The reactive
power contribution from the WTs is the difference between Mode
2 and Mode 1.
5. Conclusion
The paper presented the modeling and the control design of
the variable-speed wind turbine with a permanent-magnetic synchronous generator. The detailed models of a candidate industrial
site with multiple wind turbines were developed and were used to
perform simulation studies and evaluate alternative control solutions. The goal of the investigation was to make use of available
wind-turbine technology, namely the variable-speed permanentmagnetic synchronous generator with power electronic converters,
to actively participate in improving voltage control in the system.
To ensure reliable operation of the supervisory reactive power control scheme, the operating-point-dependent reactive power limit of
each grid-side converter was taken into account. The overall supervisory reactive power control scheme can be applied to larger wind
farms and network congurations.
Appendix A.
Base values
Sb = 2 MVA,
Vdc = 800 V,
Jb =

Sb
b2

Vb = 690 V, b = 2f (rad/s), f = 60 Hz,

V / 3
Z
1
S
Zb = b
, Lb = b , Cb =
, Tb = b ,
b
Zb b
b
ib

idc =

Sb
,
Vdc

Zdc =

Vdc
,
idc

Ldc =

Zdc
,
b

Cdc =

1
Zdc b

Innite bus voltage and maximum operating limit of VSC (pu)

vdq,inf = 1.1 0.5 ,

Smax = 1

Line parameter (pu)

RTL = 0.0059,
Lca = 0.003,
Rtr1 = 0.0003,

LTL = 0.1132,
Cca = 0.042,
Ltr1 = 0.001,

CTL = 0.025,
Rlt = 0.014,

Rca = 0.006,
Llt = 0.175,

Rtr2 = 0.0005,

Ltr2 = 0.004

PMSG (pu)
Rs = 0.042,

Lds = 1.05,

Lqs = 0.75,

[2] P. Lampola, J. Perho, J. Saari, Electromagnetic and thermal design of a lowspeed permanent magnet wind generator, in: Proceedings of the International
Symposium on Electric Power Engineering, 1995, pp. 211216.
[3] A. Grauers, Design of direct-driven permanent magnet generators for wind turbines, in: Technical Report 292, Chalmers University of Technology, Gteborg,
1996.
[4] V. Akhmatov, Variable-speed wind turbines with multi-pole synchronous permanent magnet generators. Part 1. Modeling in dynamic simulation tools, Wind
Eng. 27 (2003) 531548.
[5] F.M. Gonzales-Longatt, Dynamic model of variable speed WECS: attend of simplication, in: Proceedings of the Fifth International Workshop on Large-scale
Integration of Wind Power and Transmission Networks for Offshore Wind
Farms, 2005.
[6] L.H. Hansen, L. Helle, F. Blaabjerg, E. Ritchie, S. Munk-Nielsen, H. Binder, P.
Srensen, B. Bak-Jensen, Conceptual Survey of Generators and Power Electronics for Wind Turbines, Ris National Laboratory, 2001.
[7] A. Cavagnino, M. Lazzari, F. Profumo, A. Tenconi, A comparison between the
axial and the radial ux structures for PM synchronous motors, IEEE Ind. Appl.
Conf. Proc. 3 (2000) 16111618.
[8] E. Spooner, A.C. Williamson, Direct coupled, permanent magnet generators for
wind turbine applications, IEE Electric Power Appl. Proc. 143 (1996) 18.
[9] R. Doherty, E. Denny, M. OMalley, System operation with a signicant
wind power penetration, IEEE Power Eng. Summer Meeting Proc. 1 (2004)
10021007.
[10] K.S. Salman, A.L.J. Teo, Windmill modeling consideration and factors inuencing the stability of a grid-connected wind power-based embedded generator,
IEEE Trans. Power Syst. 18 (2003) 793802.
[11] Z. Litipu, K. Nagasaka, Improve the reliability and environment of power system
based on optimal allocation of WPG, IEEE Power Syst. Conf. Exposition Proc. 1
(2004) 524532.
[12] N. Dizdarevic, M. Majstrovic, S. Zutobradic, Power quality in a distribution network after wind power plant connection, IEEE Power Syst. Conf. Exposition
Proc. 2 (2004) 913918.
[13] E. ON Netz, Wind power report, Available: http://www.nowhinashwindfarm.
co.uk/EON Netz Windreport e eng.pdf (April 22, 2005).
[14] N. Protter, G. Garnett, S. Pai, BC Wind Integration System Expansion Study,
Elsam Engineering, 2004.
[15] T. Gjengedal, Large scale wind power farms as power plants, in: Proceedings
Nordic Wind Power Conference, 2004.
[16] J. Matevosyan, T. Ackermann, S.M. Bolik, Technical regulations for the interconnection of wind farms to the power system, Wind Power Power Syst. Proc.
(2005) 115142.
[17] H.S. Ko, Supervisory voltage control scheme for grid-connected wind farms,
PhD Dissertation, Dept. Elect. and Comp. Eng., Univ. of British Columbia, Vancouver, BC, Canada, 2006.
[18] C.H. Ong, Dynamic Simulation of Electric Machinery, Prentice Hall, New Jersey,
1998.
[19] J. Morren, S.W.H. de Haan, P. Bauer, J.T.G. Pierik, Comparison of complete and
reduced models of a wind turbine using doubly-fed induction generator, in:
Proceedings of the 10th European Conference on Power Electronics and Applications, 2003.
[20] P.C. Krause, O. Wasynczuk, S.D. Sudhoff, Analysis of Electric Machinery and
Drive Systems, John Wiley & Sons Inc, New Jersey, 2002.
[21] K. strm, T. Hgglung, PID Controllers, Lund Institute of Technology, 2004.
[22] H.-S. Ko, S. Bruey, G. Dumont, J. Jatskevich, A. Ali, A PI control of DFIG-based
wind farm for voltage regulation at remote location, IEEE Power Eng. Soc. Gen.
Meet. Proc. (2007) 16.
[23] MATLAB and Simulink , MathWorks, January 2000.

= 1.16

Controller gains (pu)

Generator-side converter:
Controllers PI1 and PI3: kp = 0.2953, ki = 12.4832.
Controllers PI2 and PI4: kp = 21.5, ki = 11.5.
Grid-side converter: controllers PI5 and PI6: kp = 0.7147,
ki = 7.1515.
DC-link module: vref = 1.16, Cdc = 0.1, kp = 0.9544, ki = 7.8175.
dc
Reactive power controllers: kp = 0.001, ki = 120.
References
[1] T. Ackermann, Wind Power in Power Systems, John Wiley & Sons, Ltd, UK, 2005.

Hong-Woo Kim received his B.S. degree in mechanical engineering from Daejeon
Engineering University, Daejeon, Korea, in 1990, his M.Sc. degree in energy system
engineering from Sunggyun University, Daejeon, Korea, in 1998, and his Ph.D. candidate in electrical engineering at Chungbuk National University, Korea since 2005. He
is a senior engineer in the wind energy research center at Korea Institute of Energy
Research (KIER), Daejeon, Korea since 1990.
Sung-Soo Kim received his M.S. degree in electrical engineering from the University
of Arkansas-Fayetteville in 1989 and his Ph.D. from the University of Central Florida
in 1997. He is presently a professor of electrical engineering at Chungbuk National
University. Prof. Kims interests include signal processing, communication theory,
and articial intelligence.
Hee-Sang Ko received his B.Sc. degree in electrical engineering from Jeju National
University, Jeju, Korea, in 1996, his M.Sc. degree in electrical engineering from Pennsylvania State University, University Park, USA, in 2000, and his Ph.D. in electrical and
computer engineering from the University of British Columbia, Vancouver, Canada,
in 2006. He is a researcher in the wind energy research center at Korea Institute of
Energy Research (KIER), Daejeon, Korea. His research interests include wind power
generation, power systems voltage and transient stability, data processing for power
systems security analysis, electricity market analysis, control design, and system
identication.