-
WIND ENERGY
Wind Energ. (2013)
Published online in Wiley Online Library
(wileyonlinelibrary.com). DOI: 10.1002/we.1637
RESEARCH ARTICLE
A fuzzy control strategy for power smoothing andgrid dynamic
response enrichment of agrid-connected wind energy conversion
systemAbdul Motin Howlader1, Naomitsu Urasaki1, Alok Pratap1,
Tomonobu Senjyu1 andAhmed Yousuf Saber2
1 Department of Electrical and Electronics Engineering,
University of the Ryukyus, 1 Senbaru, Nishihara-cho, Nakagami,
Okinawa903-0213, Japan
2Operation Technology, Inc., California, USA
ABSTRACT
This paper concentrates on the output power smoothing and the
grid dynamic response enhancement of a grid-interactiveMW-class
permanent magnet synchronous generator-based wind energy conversion
system (WECS). A simple fuzzy con-troller method is applied to
improve the overall performance of the WECS. The proposed method
can retrieve the storingkinetic energy from the inertia of a wind
turbine, perfectly. As a result, it can ensure a proficient power
smoothing ofthe variable speed WECS. On the other hand, the grid
side inverter is controlled by the fuzzy controller. This
approachcan reduce the fluctuation of DC link voltage and can
deliver a smooth power to the power grid. The proposed methodis
compared with two other methods such as the maximum power point
tracking control method and the without fuzzycontroller method. A
simple shunt circuit also includes in the DC link circuit.
Therefore, during the system fault condition,the WECS can perform a
stable operation. Effectiveness of the proposed method is verified
by numerical simulations.Copyright 2013 John Wiley & Sons,
Ltd.
KEYWORDS
wind turbine; power smoothing; kinetic energy; maximum power
point tracking; fuzzy controller; permanent magnet
synchronousgenerator
Correspondence
Tomonobu Senjyu, Department of Electrical and Electronics
Engineering, University of the Ryukyus, 1 Senbaru,
Nishihara-cho,Nakagami, Okinawa 903-0213, Japan.E-mail:
[email protected]
Received 9 October 2011; Revised 20 November 2012; Accepted 21
April 2013
1. INTRODUCTION
In recent years, the penetration of the wind power into electric
grids has been enlarged. Wind energy is one of the rapidgrowing
industries among preferred sources (e.g., solar, wave, hydro and
biomass) of the green energy.1 It can meet theelectricity demand
with high reliability, economically and environmental friendly.
Depending on the rotational speed, thereare two types of wind
energy conversion systems (WECSs): (i) fixed speed wind turbine,
and (ii) variable speed wind tur-bine (VSWT). Due to the variation
of wind velocity, the VSWT is the most popular WECS because it can
utilize the windenergy efficiently.2 VSWT systems are based on
doubly-fed induction generators (DFIGs), permanent magnet
synchronousgenerators (PMSGs) and synchronous generators with
electromagnets. The PMSG-based VSWT has been increased owingto the
reliability, availability, simple structure and high energy
efficiency. Moreover, the PMSG can operate without a gear-box.
Hence, it can release from all difficulties of a gear box. Usually,
the PMSG-based WECS is connected to the powergrid via an ACDCAC
converter system. In this system, the WECS is not required to
synchronize with the rotationalspeed of the grid frequency.
However, wind velocity is inconstant, and the wind turbine
output power is proportional to the cube of wind speed.Thus, the
output power of the WECS is fluctuated, which incurs the frequency
deviation from the rated value.35Various methods have been proposed
to generate a smooth output power of the WECS such as ultra
capacitors,68
Copyright 2013 John Wiley & Sons, Ltd.
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
super conducting magnetic energy storage,9,10 flywheel energy
storage system,11,12 storage batteries,13 fuel cell,14and backup
generator15 and so on. But, installation and maintenance costs of
these active devices are extremelyhigh. Therefore, several
researches regarding to the output power smoothing have been
published by using of thepitch angle control or the inertia of
VSWTs.1622 Extra energy storages or devices are not required for
these powersmoothing techniques.
In the works of Sato and Saitoh and Abedini and Nasiri16,17,
output power smoothing methods have been proposedby using the rotor
inertia. Output power fluctuations of the VSWT can be smoothed at
wind speed variations; how-ever, detail characteristics of the
power smoothing and the power efficiency of a VSWT are not
addressed. Therefore,large wind speed variations may put the WECS
in unstable operating regions. In the works of Banakar et al. and
Luoet al.,18,19 the steady-state stability of a DFIG is analysed,
and stable operating regions are proposed. However, therotor speed
of the VSWT is controlled by the power converter, but a stable
operation at the system fault condition isnot proposed. In the work
of Chang-Chien,20 a control method to regulate the output power of
the DFIG under variouswind conditions is proposed. The output power
smoothing of the WECS is achieved by using the pitch angle
controland the rotor inertia. However, the blade stress of the VSWT
is increased drastically. Output power smoothing methodsof the
pitch angle control are proposed in the works of Senjyu et al. and
Rashad et al.21,22 However, these methodsreduce the WECS output
power and increase the wind turbine blade pitching. In the
aforementioned methods, thereare no discussions of the power
smoothing efficiency. Therefore, an efficacy power smoothing method
is required tothe WECS.
This paper presents a fuzzy controller-based output power
smoothing method of the WECS by controlling the kineticenergy in
inertia of a wind turbine. Further, a fuzzy controller-based grid
side inverter control system is applied to inject asmooth power to
the power grid. The WECS adopts an ACDCAC converter system, which
is composed of a generatorside converter and a grid side inverter.
The generator side converter controls the torque of the PMSG,
whereas the gridside inverter controls the DC link voltage and the
grid voltage, respectively. To achieve an efficient power smoothing
ofthe WECS by using the inertia, one should consider two thingswind
speed and the change of wind speed. Therefore, theproposed method
includes the fuzzy logic23,24 to determine the speed command for
the generator side converter. Inputsof the fuzzy controller are
wind velocity and the change of wind velocity. From these inputs,
the fuzzy controller decidesan adaptive averaging time to retrieve
the kinetic energy from the inertia perfectly. As a result, the
WECS can generate aconstancy smooth output power. The grid side
inverter is controlled by the fuzzy controller-based PI controller
to deliver asmooth power to the power grid. To perform a stable
operation of the WECS at the fault condition, a shunt circuit
includesin the DC link circuit. By using the proposed method,
mechanical stresses of the wind turbine shaft, blades and
genera-tor can be mitigated as compared with the conventional
maximum power point tracking (MPPT) control method. It alsoensures
an efficient power smoothing and transports a smooth power to the
power grid. The proposed method is verified bynumerical simulations
using MATLAB/Simulink (MathWorks, Inc., Natick, Massachusetts,
United States).
2. WIND ENERGY CONVERSION SYSTEM
2.1. System configuration
The WECS integrates with a permanent magnet synchronous
generator, a generator side converter and a grid side inverter.The
configuration diagram of the WECS is shown in Figure 1. Wind energy
is obtained from the wind turbine and is sent tothe PMSG. To
generate the maximum output power, the rotational speed of the PMSG
is controlled by a pulse width mod-ulation (PWM) converter. The
output power of the PMSG is supplied to the power grid through a
generator side converterand a grid side inverter.
PMSG~
Generator side converter
~
Grid side inverter Power system
PMSG control system
Figure 1. Wind energy conversion system configuration.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
2.2. Wind turbine system
The wind turbine output power, Pw and the wind turbine output
torque, Tw are defined by the following equations:
Pw D 12Cp.; /R
2oV
3w (1)
Tw D 12Cp.; /R
3oV
2w = (2)
where Vw is the wind speed, is the air density, Ro is the wind
turbine blade radius, !w is the angular speed of the windturbine,
Cp is the power coefficient, is the tip speed ratio, can be defined
as D Ro!wVw , !w is the angular speed of thewind turbine and is the
pitch angle. The power coefficient Cp is defined by the following
equation.25
Cp D 0:22116
0:4 5
exp 12:5 (3)
D 11
C0:08 0:0353C1(4)
The wind turbine output power characteristics are depicted in
Figure 2, from which it can be seen that, for any par-ticular wind
speed, there is a rotational speed !opt, called the optimum
rotational speed, which generates the maximumpower Pmax. In this
way, the MPPT control for each wind speed increases the energy
generation in the variable speed windgeneration system. !opt is
calculated by differentiating Cp with respect to !w. Therefore,
!opt is approximated by26
!opt D 0:2Vw 0:2: (5)
If !w D !opt, the wind turbine maximum output power Pmax can be
obtained. The MPPT control is applied when the windspeed is less
than the rated wind speed Vw_rated D 12 m/s to generate the maximum
power. Above the rated wind speed, thepitch angle control system is
applied to maintain the rated power of the PMSG. The pitch angle is
operated in the followingcases. When the wind speed is within 0
< Vw < 5 m/s, there is no power generation, and the pitch
angle is fixed at D 90.When the wind speed is within 5 < Vw <
12 m/s, the pitch angle operates at D 2 so that the power can be
maximized.When the wind speed is within 12 < Vw < 24 m/s, the
pitch angle operates at 2 90 so that the PMSG can maintainthe rated
power 2 MW. Above the wind speed Vw > 24 m/s, the pitch angle is
again fixed at D 90 for safety reasons.The different operating
regions of the pitch angle control laws are shown in Figure 3(a).
Figure 3(b) shows the pitch anglecontrol system that determines the
pitch angle , where output power error Pg is used as the input of
the PI controller.Actually, the pitch angle control system includes
a hydraulic servo system. The system has nonlinear characteristics,
butcan be modeled as a first-order lag system.27,28 Therefore, in
this paper, the first-order lag system is used where the
timeconstant is 1 s. Moreover, the pitch angle is limited by a
limiter within 2 90, and the maximum rate of change is10/s.2931
2.5
0 1 2 3 40
0.5
1.0
1.5
2.0
Win
d tu
rbun
e ou
tput
pow
er P
g [M
W]
Wind turbine speed w [rad/s]
11m/s
12m/s
10m/s
9m/s
8m/s7m/s
6m/s
Pmax
Figure 2. Wind turbine output power characteristics.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
Wind speed Vw
[m/s] 0
0.2
0.4
0.6
0.8
1.0
1.2
0 5 10 15 20 25 30
cut-in
rated cut-out
=90 deg =2 deg =2~90 deg =90 deg
(a)
+
-
PIs+11
90 deg
2 deg
90 deg
2 deg Rate limiterPref
Pg
Vw
P g 1s+1
CMD
+10 deg
-10 deg
Pitch angle control systemHydraulic servo systemAngle
selector
(b)
Win
d tu
rbun
e ou
tput
po
wer
Pw
[p.u]
Figure 3. (a) Pitch angle control law for all operating regions
and (b) pitch angle control system.
1/JeqD -
-
+Te
Tw
e
epg
s1
s1
Figure 4. Model of drive train.
dq
abc
e
Eq. (6)
Eq. (7)
Eq. (8)
vd
vq iq ic
ib
iaid
Te
dq
abc
e
va
vb
vc
Figure 5. Model of PMSG.
2.3. PMSG model
The mathematical model of a PMSG is the same as a permanent
magnet synchronous motor. The voltage andtorque equations of the
permanent magnet synchronous motor in the synchronous reference
frame are given by thefollowing equations:
vd D .Ra C PLd/id !eLqiq (6)
vq D !eLdid C .Ra C PLq/iq C !eK (7)
Te D pfKiq C .Ld Lq/idiqg (8)where vd and vq are the dq-axis
voltages, id and iq are the dq-axis currents, Ra is the stator
resistance, Ld and Lq are thedq-axis inductances, !e is the
electrical rotational speed, K is the permanent magnetic flux, P is
the differential operatorand p is the number of pole pairs. In
addition, the motion equation of the PMSG is given as
Te D Jeq d!gdt C D!g C Tw (9)
where D is the rotational damping, Jeq.Jeq D Jg CJw/ is the
equivalent inertia, and !g is the mechanical rotational speed.The
models of the drive train and the PMSG are given in Figures 4 and
5, respectively.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
PMSG~
Converter
~
InverterPower system
e
abc dq
i1abc
Convertercontroller
v1dq
i1dq
abc dq
PWM
v1abc
sPLL
abc dq
Invertercontroller
i2dq
abc dqv2dq
PWM
v2abc
Shunt
i2abc
Figure 6. Power converter control system.
+_
i1d*
vi1d
PI+
_
*1d
+_
i1q*
vi1q PI ++
*1q
PI +*
_
Figure 7. Generator side converter control system.
2.4. Configuration of power converter system
The PMSG is connected to the grid through two six switches-based
hard switched converters, with a DC link capacitor.The power
converter control system is shown in Figure 6. The generator side
converter is controlled through a PI controllersuch that the d
-axis current is zero to obtain the maximum torque. The generator
side converter achieves the variable speedoperation by controlling
the rotational speed of the PMSG. On the other hand, the grid side
inverter supplies the electricalpower, which is synchronized with
the power system frequency. The DC link includes a shunt circuit to
overcome the DClink overvoltages at the system fault condition.
Therefore, a stable operation of the WECS can be achieved under a
systemfault condition. Each configuration of the power converter
control systems is described below.
2.4.1. Generator side converter.The generator side converter
controls the rotational speed of the PMSG to perform the variable
speed operation for the
MPPT controller. The vector control scheme is used and is shown
in Figure 7. The speed control of the PMSG is real-ized on a
rotating frame, where the rotational speed error is used as the
input of the speed controller, which produces theq-axis stator
current command i1q. Generally, a cylindrical pole type synchronous
machine is considered to control thed -axis stator current, and i1d
is set to zero. Therefore, the d -axis stator current command, i1d,
is set to zero. The errorsbetween the dq-axis current commands and
the actual dq-axis currents are used as the inputs of the current
controllers.The current controller outputs produce the dq-axis
voltage commands v1d and v1q after decoupling. The rotor position,
e,is used for the transformation from abc to dq variables and is
detected from the rotational speed of the PMSG by using amechanical
sensor.
2.4.2. Grid side inverter.The grid side inverter controls the
constant DC bus voltage Vdc and performs the unity power factor
operation. The
control system of the grid side inverter is shown in Figure 8.
The d -axis current can control the DC bus voltage Vdc, and
theq-axis current can control the reactive power Qi . The DC bus
voltage reference V dc is set to 3.5 kV, whereas the reactivepower
command Qi is set to zero for the unity power factor operation. The
phase angle s, for the transformation betweenabc and dq frame, is
detected from the three-phase low voltage side of the grid-side
transformer by using the phase-lockedloop system.32,33 The angular
position of the dq reference frame is controlled by a feedback
loop, which regulates theq-axis component to be zero, where the d
-axis component depicts the voltage vector amplitude, and phase is
determinedby the output of the feedback loop.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
+_
i2d*
vi2d PI*2d
+_
i2q*
vi2q PI*2q
PI+
*_
+*
_
PI
Figure 8. Grid side inverter control system.
3. PROPOSED POWER SMOOTHING CONTROL SYSTEM
The power fluctuations should be minimized to prevent any kind
of interruptions. There are many power generators that areequipped
with the control system, which is called the automatic generation
control (AGC).17 The AGC regulates the outputpower of the generator
so that the power fluctuations attenuate. Since the AGC activates
some mechanical and electricalactuators, which have a large time
constant. It is effective for the low frequency power fluctuations.
This section presentsthe proposed fuzzy-based output power
smoothing method and the grid side inverter control techniques.
Above the ratedwind speed, the output power is smoothed by the
pitch angle control system, and below the rated wind speed, the
outputis smoothed by the kinetic energy in inertia. Finally, at the
fault condition, the DC link over voltage control method isalso
described.
3.1. Fuzzy controller-based power smoothing scheme
This subsection presents the proposed fuzzy-based output power
smoothing method by utilizing the kinetic energy in iner-tia of a
wind turbine. The concept behind the proposed power smoothing
method depends on two events:
Increase in wind turbine kinetic energy due to the acceleration
of the turbine rotational speed at the high wind speed. A discharge
of wind turbine kinetic energy due to the deceleration of the
turbine rotational speed at the low
wind speed.
These two incidents can generate a smooth output power of the
WECS. The maximum output power of the wind turbine,Pmax is
calculated as
Pmax D Tw!opt (10)The average value of the maximum wind turbine
output power is calculated as
P D 1T
Z ttT
Pmaxdt (11)
where t denotes the present time and T is the averaging time.
Then, the power difference P between the maximum outputpower and
the average value of the maximum output power is determined. The P
gives the storing and restoring power ininertia of the wind
turbine. Therefore, the
RP gives the wind turbine storing and restoring energy. Wind
turbine kinetic
energy, E, is determined as
E D 12Jeq!
2g (12)
The wind turbine rotational speed command, !g , is determined
from the kinetic energy command, E (the sum ofRP
and E), as
!g Ds
2EJeq
(13)
Finally, the generator electrical reference speed !e for the
proposed method is determined from the number of pole pairsp as
!e D !g p (14)
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
2 4 6 8 10 1230
35
40
45
50
55
Sum
mat
ion
of g
ener
ated
po
wer
(MW
)
Averaging time T (s)
Maximum output power
(a) Averaging time Vs. total output power characteristics.
0 10 20 30 400
20
40
60
Time (s)
Ener
gy fu
nctio
n *P m
ax
(MJ) Constant averaging time
MPPT method
T =5 s
T =2 s
T =8 s
(b) Maximum output power function.
0 10 20 30 400
20
40
60
Time (s)
Pow
er sm
ooth
ing
func
tion
*P l
evel
(M
W)
Constant averaging timeMPPT method
T =5 s
T =2 s
T =8 s
(c) Power smoothing function.
Figure 9. Characteristics of averaging time. (a) averaging time
versus total output power characteristics, (b) maximum output
powerfunction and (c) power smoothing function.
and for the MPPT control method,
!e D !optp: (15)
The averaging time, T , in equation (11) is an important
parameter for the output power smoothing control system.Figure 9(a)
shows the characteristic between the averaging time and the
summation of output power of the WECS. Thedotted line shows the
summation of output power for the MPPT controller during a
simulation period. The output powermoves to the maximum power at
the short averaging time. From this figure, (Figure 9(a)), it can
be realized that a constantaveraging time may decrease the system
efficiency. At the short averaging time, the total generated power
is similar to theMPPT control method, but the power fluctuation is
also the same as the MPPT control method. On the other hand,
thelarge averaging time can generate a smooth power, but total
generated power of the WECS is decreased. There are
othercharacteristics of the averaging time and the power smoothing
for the detailed analyses. Evaluation of the output power(Pg)
smoothing is represented by the maximum energy function Pmax and
the smoothing function Plevel, which areexpressed as21
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
Pmax DZ t
0Pg.t/dt (16)
Plevel DZ t
0
dPg.t/dt
dt : (17)
If Pmax is large, the efficiency of the WECS increases. If
Plevel is small, the output power fluctuation decreases.Figure 9(b)
and (c) shows the maximum energy function and the output power
smoothing function for various averag-ing times. There is a
comparison between the MPPT control method and the power smoothing
method at the constantaveraging time. From these two figures,
efficiency of the WECS is increased (Figure 9(b)), and smoothing
performance ofthe WECS is decreased (Figure 9(c)) during a short
averaging time. At the high averaging time, efficiency of the
WECSis decreased, and the smoothing performance of the WECS is
increased. When the wind velocity fluctuates rapidly, thelarge
averaging time is suitable for the WECS to generate a smooth output
power. Alternatively, when the wind velocitybecomes steady, the
short averaging time can increase the efficiency of the WECS. So,
an adaptive averaging time is nec-essary to improve the overall
performance of the output power smoothing method (instead of a
constant averaging time). Itcan be implemented through the fuzzy
controller. Wind velocity and the fluctuation of wind velocity are
determined withan appropriate averaging time instantaneously.
The fuzzy logic is described by a set of if then-based fuzzy
rules.23,34 It is effective when the mathematical expres-sions are
difficult due to the inherent complexity, nonlinearity or
non-clarity. There are two inputs of fuzzy reasoning suchas wind
velocity .Vw/ and the fluctuation of wind velocity d=dt .Vw/. The
fuzzy input membership functions (i.e., windvelocity and the change
of wind velocity) are shown in Figure 10(a) and (b), respectively.
Output membership functionsand fuzzy rules are shown in Figure
10(c) and (d), respectively. The membership functions can be tuned
according to thespecifications of generators and turbines. The i th
fuzzy rule is expressed as35
Rule i : if Vw is wvx and d=dt .Vw/ is cwvy then T is ATlx D 1;
2; : : : ; 8; y D 1; 2; : : : ; 8; l D 1; 2; : : : ; 8
where wvx, cwvy denote the antecedents, and ATl represents the
consequent part. The fuzzy reasoning output (i.e.,averaging time) T
is calculated by
T DP64
iD1 !iATlP64iD1 !i
(18)
where !i denotes the grade for the antecedent and is obtained
by
!i D .! wvi /.! cwvi / (19)where ! wvi and ! cwvi are grades of
the antecedents for each rule. The proposed fuzzy controller-based
powersmoothing control system is shown in Figure 11. The final
outcome of this controller is the electrical reference speed !e
,and is the input of the converter controller. It can control the
speed and the generator output power, and delivers an
efficientoutput power.
The rotational speed of wind turbine !r is proportional to the
wind speed. Therefore, the wind turbine rotational speedand the
change of wind turbine rotational speed can be inputs of the fuzzy
controller rather than the wind speed and thechange of wind speed.
If the measurement of the wind speed is difficult in a place, the
rotational speed and the change ofrotational speed of wind turbine
will utilize as the inputs of fuzzy controller for the proposed
system.
3.2. Fuzzy controller-based grid side inverter control
method
Due to the non-linear behavior of the power system and the
linearization problems, the control of the variable speed WECSis
difficult by using the conventional PI controller methods. For
example, the PI controller design requires to identify thewind
turbine transfer function, the linear model of the network and
defining an accurate tuning process. Because of thesereasons, to
implement a PI controller is not an easy task. The fuzzy controller
is a rule-based non-linear control technique.The fuzzy controller
can overcome these problems and presents some advantages as
compared with a PI controller. It iseasy to obtain the variable
gains depending on the errors and overcome the problems, which are
affected by an uncertainmodel.3639 In this study, a fuzzy
controller-based inverter control system is shown in Figure 12(a).
In this figure, theconventional PI controller is replaced by the
fuzzy controller. Figure 12(b) shows the detail of the fuzzy-based
PI con-trol system. It is a multi-input multi-output-based
system.40 The proposed controller is designed to control the
smoothDC link voltage and the unity power factor operation.
Consequently, it can deliver the smooth power to the power
grid.Figure 13 shows fuzzy input membership functions (i.e, error
and change of error) and the output membership functions(i.e.,
proportional gain Kp and integral gain Ki ), and fuzzy rules.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
wv1 wv2 wv3 wv4 wv5
4 100
0.5
1
Wind Velocity
wv6 wv7 wv8
16
(a) Membership function of wind velocity.
cwv1 cwv2 cwv3 cwv4 cwv5
00
0.5
1
d/dt (Wind Velocity)
cwv6 cwv7 cwv8
2.2
(b). Membership function of change of wind velocity.
AT1 AT2 AT3 AT4 AT5
6 00
0.5
1
Averaging Time T (s)
AT6 AT7 AT8
8
(c) Membership function of averaging time.
cwv1d/dt (V_ )
V_
cwv2 cwv3 cwv4 cwv5 cwv6 cwv7 cwv8wv1wv2wv3wv4wv5wv6wv7wv8
AT1AT1AT1AT1AT1AT1AT1AT2
AT1AT1AT1AT1AT1AT1AT2AT3 AT4
AT2AT1AT1AT1AT1AT1
AT5AT4AT2AT2AT1AT1AT1AT1 AT1
AT2AT2AT2AT2AT3AT5AT6
AT2AT2AT2AT3AT3AT3AT5AT7
AT2AT2AT3AT3AT4AT5AT6AT7
AT3AT3AT3AT4AT4AT6AT7AT8
AT3
(d) Table I (fuzzy rules).
Figure 10. Fuzzy controller for power smoothing enhancement. (a)
membership function of wind velocity, (b) membership functionof
change of wind velocity, (c) membership function of averaging time
and (d) Table I (fuzzy rules).)
T P
P*P
E
E*g+
+
+
-
Eq. (10)Eq. (11)
Eq. (13)Eq. (2)
Eq. (5)V
opt
max
Eq. (12)Pitch AngleController
Td/dt Fuzzy controller
Eq. (14)e
Figure 11. Power smoothing control system for WECS.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
+
_
i2d*
vi2d PI*
2d
+
_
i2q*
vi2q PI*
2q
+*
Qi
_
+
*
_
V dc
V dc
Qi
FuzzyController
FuzzyController
e
e
(a) Fuzzy controller based inverter control system
e
ddt
Fuzzy LogicController
Kp
Ki
x
x
++
ref i 2d*
2q, i* }{
ce
(b) Fuzzy controller based PI control system.
Figure 12. Proposed inverter control system. (a) fuzzy
controller-based inverter control system and (b) fuzzy
controller-based PI controlsystem.
The fuzzy rules and membership functions presented in fuzzy
controller (i.e., power smoothing control and invertercontrol) are
determined by the trial-and-error process. However, it is possible
to tune the parameters of the controllers andmembership functions
of the fuzzy controller to achieve the smoothing and maximum
acquisition of the available windpower. Various methods have been
proposed for tuning the fuzzy controller, such as self-tuning
algorithm based on anexperimental planning method,41 where the
scaling factors of optimal parameters can be determined efficiently
accordingto the desired performance indexes, Taguchi tuning
method,42 and tuning the membership functions.43 However, in
thepresent literature, the selection of scaling factors is still
based on the trial-and-error method.
3.3. Configuration of the DC link circuit
The DC link voltage, Vdc is controlled by the grid side
inverter. Therefore, under the system fault condition, it is
verydifficult to control the over DC link voltage through the
conventional PI controller. Hence, a shunt circuit includes in
theDC link capacitor as shown in Figure 14(a). The DC link voltage
control system is shown in Figure 14(b). From this figure,when the
grid voltage, Vt, is less than 0.8 pu (i.e., fault occurred) then
the error between the DC link voltage command V dc,and the DC link
voltage Vdc, is the input of the PI controller, and the switching
command of the shunt circuit is determinedby the triangular PWM
law.
4. SIMULATION RESULTS
The power system model is shown in Figure 15. The grid side
inverter is connected to an infinite bus through an LCfilter, a
three-phase transformer and a transmission line. The parameters of
the PMSG and the wind turbine are listed inAppendix. To evaluate
effectiveness of the proposed method, the WECS operations are
assessed in the steady-state and thesystem fault conditions. The
evaluation of the steady-state condition is verified through the
fuzzy controller-based powersmoothing method and the grid side
inverter control system.
4.1. Operation in fuzzy controller-based power smoothing
system
Simulation results of power smoothing are shown in Figure 16.
Simulation results among the MPPT control method, thefuzzy
controller-based method (i.e., adaptive averaging time) and the
without fuzzy controller (i.e., constant averaging timeT =6 s)
method are compared. In these figures, the fuzzy controller-based
method is referred as the proposed method, and
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
PMPS PBZONSNMNB
-800
-60 -40 -20 0 20 40 60 80
0.5
1
Error, e(a) Membership function of error.
PMPS PBZONSNMNB
00 20
0.5
1
-20Change of error, ce
(b) Membership function of change of error.
PMPS PBZONSNMNB
0
0.5
1
1 1.5 2Proportional gain, Kp
(c) Membership function of Kp.
PMPS PBZONSNMNB
0
1
0.5
30 701Integral gain, Ki
(d) Membership function of Ki.
K p i NB NM NS ZO PS PM PBNBNMNSZOPSPMPB
NBNBNM
NB NM
NM
ZO
NSZOPS
NS
ZOPSPM
NSZOPSPMPB
NMNSZO
PS
PMPB
PB
NSZOPS
PMPB
ZO
PM
e
ce
(e) Table II(fuzzy rules).
K&
Figure 13. Fuzzy PI control system. (a) membership function of
error, (b) membership function of change of error, (c)
membershipfunction of Kp and (d) membership function of Ki.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
RdcSW
C
idc
Power systemGenerator
(a) Configuration of DC-link circuit.
+
_
Vdc
Vdc
*
PI1
00
SW+1
1
Grid voltage v
0.8 put
v t 0.8 1v t 0.8 0
_
Carrier wave
(b) DC-link voltage control technique.
Figure 14. DC link voltage control for WECS. (a) configuration
of DC link circuit and (b) DC-link voltage control technique.
PMSG AC-DC-AC LCfilter
1.5kV/6.6kV
Grid P , Qtt
Fault
AC load
vt
5 MW
Inf. BUS
Figure 15. Power system configuration.
the without fuzzy controller based is referred as the
conventional method. Wind velocity is shown in Figure 16(a).
Windspeed data contain significant fluctuation and high turbulence
after 25 s. The change of wind speed is derived by the
firstderivative of wind speed. It is shown in Figure 16(b). Figure
16 (a) and (b) shows the inputs of the fuzzy controller. Theoutput
of the fuzzy controller is the averaging time, which is depicted in
Figure 16(c). The characteristics of the kineticenergy of the wind
turbine is shown in Figure 16(d). The kinetic energy comes up when
wind speed increases, and comesdown the kinetic energy when wind
speed decreases. The generator electric torque is shown in Figure
16(e). The fuzzycontroller-based method can reduce the torque
fluctuation significantly. As such, it can reduce the mechanical
stress of agenerator. A comparison of the pitch angle is shown in
Figure 16(f). From this figure, the proposed method can lessen
theblade pitching as compared with the MPPT control system.
Therefore, the proposed fuzzy controller-based method candiminish
the stress of the wind turbine blades. Figure 16(g) shows the
generated output power of the WECS. The proposedmethod can reduce
the power fluctuation extensively as compared with the MPPT control
method. The proposed methodcan generate efficient output power
smoothing at different phases of wind velocity as compared with the
without fuzzycontroller method. The without fuzzy controller method
can generate a smooth output power but it reduces the outputpower
extensively as compared with the MPPT control method. As a result,
the efficiency of the output power is reduced.The proposed method
generates a tiny fluctuation as compared with the without fuzzy
controller method, but it can ableto generate an efficient output
power of the WECS. The torque difference of the WECS is shown in
Figure 16(h). Theproposed method can reduce the torque difference
significantly as compared with the MPPT control method. As a
result, itcan mitigate the shaft stress of the WECS.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
0 10 20 30 40
0 10 20 30 40
8
10
12
14
Time [s]
Win
d sp
eed
V w[m
/s]
(a)
Time [s](b)
0
0.5
1
1.5
2
Chan
ge o
f win
d sp
eed
d/dt
( Vw
)
0 10 20 30 406
6.5
7
7.5
8
Time [s]
Ave
ragi
ng ti
me
T
[s]
(c)
0 5 10 15 20 25 30 35 4002468
1012 x 104
Time [s]
Kin
etic
ene
rgy
E[J]
(d)
108642
02 x 105
PMSG
ele
ctric
torq
ue T
e[Nm]
Proposed methodConventional methodMPPT method
0 10 20 30 400
5
10
15
20
25
Time [s]Pi
tch
angl
e
[deg
](f)
0 10 20 30 40Time [s]
(g)
0
0.5
1
1.5
2
2.5
Gen
erat
ed o
utpu
t pow
er P
g [M
W]
5 10 15 20 25 30 35 40
0.20.1
00.10.20.3
Time [s]Torq
ue d
iffer
ence
Tdi
ff [M
Nm]
(h)
Proposed methodConventional methodMPPT method
Proposed methodConventional methodMPPT method
Proposed methodConventional methodMPPT method
Proposed methodConventional methodMPPT method
0 5 10 15 20 25 30 35 40Time [s]
(e)
Figure 16. Simulation results (a) wind velocity Vw, (b) change
of wind velocity, (c) averaging time T , (d) wind turbine kinetic
energy E,(e) generator electric torque Te, (f) pitch angle , (g)
output power of WECS Pg and (h) torque difference Tdiff.
Figure 17 shows simulation results of the proposed control
system at a high wind velocity. Wind velocity is shown inFigure
17(a). From this figure, wind velocity exceeds the cut out speed.
Outcomes of the proposed controller are shownin Figure 17(b)(d).
Figure 17(b) shows the kinetic energy of the wind turbine. Due to
the cut out wind speed, the pitchangle increases rapidly to protect
the wind turbine (as shown in Figure 17(c)). The output power of
the PMSG is shownin Figure 17(d). From this figure, the output
power of the PMSG decreases swiftly due to the pitch actions of the
cut outspeed. When wind velocity reduces from the cut out speed,
the pitch angle also decreases, and the PMSG delivers theoutput
power again.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
0 5 10 15 200
10
20
30
Time [s]
Win
d sp
eed
V w
[m
/s]Cutout wind speed
(a)
0 5 10 15 200
0.5
1
1.5
2 x 106
Time [s]
Kin
etic
ene
rgy
E
[J]
(b)
0 5 10 15 200
50
100
Time [s]
Pitc
h an
gle
[
deg]
(c)
0 5 10 15 201
0
1
2
3
Time [s]
Out
put p
ower
Pg
[MW
]
(d)
Figure 17. Simulation results (a) wind velocity Vw, (b) kinetic
energy E, (c) pitch angle and (d) output power of WECS Pg.
5 6 7 8 9 10 11 12 13 14 1510
11
12
13
14
Time [s]
Win
d sp
eed
V w [m
/s]
(a)
5 10 153560
3580
3500
3520
3540
Time [s]
DC-
link
volta
ge V
dc[V
]
Conventional methodProposed method
6 7 8 9 11 12 13 14
(b)
5 6 7 8 9 10 11 12 13 14 151.2
1.4
1.6
1.8
2
2.2
Time [s]Act
ive
outp
ut o
f grid
sid
e in
verte
r P t
[M
W]
Conventional methodProposed method
(c)
5 6 7 8 9 10 11 12 13 14 15
0.016
0.018
0.02
0.022
0.024
Time [s]Rea
ctiv
e ou
tput
of g
rids
ide
inve
rter
Q t [M
var] Conventional method
Proposed method
(d)
Figure 18. Simulation results (a) wind velocity Vw, (b) DC link
voltage Vdc, (c) active power of inverter Pt and (d) reactive power
ofinverter Qt.
4.2. Operation in fuzzy controller-based grid side inverter
control system
Simulation results of the fuzzy-based inverter control system
are shown in Figure 18. Simulation results are shown asa comparison
between the conventional PI controller method and the fuzzy-based
PI controller method. Wind speed isshown in Figure 18(a). As shown
in Figure 18(b), the proposed fuzzy-based PI controller can reduce
the DC link voltagefluctuation as compared with the conventional PI
controller. So, it ensures the smooth operation of the DC link
capacitorand increases the lifetime of capacitor. The active and
reactive powers are delivered to the power grid through the
inverter,which are shown in Figure 18(c) and (d). The proposed
method can deliver smooth power to the power grid as comparedwith
the conventional method.
4.3. Operation in fault condition
Simulation results of the fault condition are shown in Figure
19. The three-line-to-ground fault occurs at the middle ofthe
transmission line in Figure 15, and electric power supply to the
load is stopped. The sequences of the simulation aredescribed as
follows:
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
2 3 4 5 6 7 810
11
12
13
Time [s]
Win
d sp
eed
V w
[m
/s]
(a)
2 3 4 5 6 7 80
2
4
6
8
Time [s]
Pitc
h an
gle
[de
g]
(b)
1 2 3 4 5 6 7 80
0.5
1
1.5
2
2.5
Time [s]
Gen
erat
ed o
utpu
tpo
wer
Pg
[MW
]
(c)
2 3 4 5 6 7 8Time [s]
3.504.00
DC-
link
volta
ge V
dc[k
V]
4.50
5.50
3.00
2.50
2.00
Conventional methodProposed method
(d)
2 3 4 5 6 7 8-2
0
2
4
6
Time [s]
Act
ive
outp
ut p
ower
of
grid
-sid
e in
verte
r P t
[M
W]
Without shuntWith shunt
(e)
2 3 4 5 6 7 80
0.5
1
1.5
2
2.5
Time [s]
Rea
ctiv
e ou
tput
pow
er o
fgr
id-s
ide
inve
rter
Q t [M
Var]
Without shuntWith shunt
(f)
Figure 19. Simulation results (a) wind velocity Vw, (b) pitch
angle , (c) WTG system generated output power Pg, (d) DC link
voltageVdc, (e) active power of grid side inverter Pt and (f)
reactive power of grid inverter Qt.
At simulation time 6.0 s, the three-line-to-ground fault occurs
at the middle of the transmission line. When the AC grid voltage is
within vt
-
Fuzzy control strategy for power smoothing and grid control A.
M. Howlader et al.
a shunt circuit uses in the DC link circuit. Therefore, a stable
operation of the WECS can achieve during the system faultcondition.
From simulation results, effectiveness of the proposed methods are
verified.
APPENDIX
Simulation parameters of the WECS are as follows:
Wind turbine: blade radius Ro D 41 m, inertia Jeq D 8000 kgm2,
air density D 1:205 kg/m3, rated wind speedVw_rated D 12 m/s,
cut-in speed Vw;cutin D 5 m/s and cut out speed Vw;cutout D 25
m/s.
Parameters of generator and grid network: rated power Pg_rated D
2 MW, number of poles pair p D 80, statorresistance Ra D 0:1 ,
inductance L D 0:005 H, field flux K D 10:68 Vs/rad, rotational
damping D D 0, gridfilter impedance, D 0:05 C j 0:3 pu.
Parameters of power converter: PWM carrier frequency fp D 10
kHz, rated DC link voltage Vdc_rated D 3:5 kV,DC link capacitor C D
15 mF, shunt circuit resistance Rs D 2:5 .
REFERENCES1. Economic impact of wind energy development in
Illinois, June 2010. [Online]. Available: http://www.
renewableenergy.ilstu.edu/wind/. (Accessed 2 September 2011).2.
Carlin PW, Laxson AS, Muljadi EB. The history and state of the art
of variable-speed wind turbine technology. Wind
Energy 2003; 6: 129159.3. Uehara A, Senjyu T, Yona A, Funabashi
T. Frequency control by coordination control of WTG and battery
using load
estimation, The 8th Eighth International Conference on Power
Electronics and Drive Systems (PEDS), Taipei, Taiwan,2009;
216221.
4. Kaneko T, Uehara A, Yona A, Senjyu T, Kim CH. A new control
methodology of wind turbine generators for frequencycontrol of
power system in isolated island. Wind Energy 2011; 14: 407423.
5. Uehara A, Senjyu T, Kaneko T, Yona A, Muhando EB, Urasaki N,
Kim CH. Output power dispatch control for a windfarm in a small
power system. Wind Energy 2010; 13: 671684.
6. Abbey C, Joos G. Supercapacitor energy storage for wind
energy applications. IEEE Transactions on IndustryApplication 2007;
43: 769776.
7. Muyeen SM, Shishido S, Ali MH, Takahashi R, Murata T, Tamura
J. Application of energy capacitor system to windpower generation.
Wind Energy 2008; 11: 335350.
8. Mandic G, Nasiri A. Modeling and simulation of a wind turbine
system with ultracapacitors for short-term powersmoothing, IEEE
International Symposium on Industrial Electronics (ISIE), Bari,
2010; 24312436.
9. Sheikh MRI, Muyeen SM, Takahashi R, Tamura J. Smoothing
control of wind generator output fluctuations byPWM voltage source
converter and chopper controlled SMES. European Transactions on
Electrical Power 2011;21: 680697.
10. Kinjo T, Senjyu T, Urasaki N, Fujita H. Terminal-voltage and
output-power regulation of wind-turbine generator byseries and
parallel compensation using SMES. IET Proceedings on Generation
Transmission & Distribution 2006;153: 276282.
11. Cardenas R, Pena R, Asher G, Clare J, Blasco-Gimenez R.
Control strategies for power smoothing using a flywheeldriven by a
sensorless vector-controlled induction machine operating in a wide
speed range. IEEE Transactions onIndustrial Electronics 2004; 51:
603614.
12. Jerbi L, Krichen L, Ouali A. A fuzzy logic supervisor for
active and reactive power control of a variable speedwind energy
conversion system associated to a flywheel storage system. Electric
Power Systems Research 2009; 79:919925.
13. Baran ME, Teleke S, Anderson L, Huang A, Bhattacharya S,
Atcitty S. STATCOM with energy storage for smoothingintermittent
wind farm power, IEEE PES General Meeting, Pittsburgh, PA, 2008;
16.
14. Iqbal MT. Modeling and control of a wind fuel cell hybrid
energy system. Renewable Energy 2003; 28: 223227.15. Takano T.
Natural energy power and energy storing technology. IEE Japan 2006;
126-B: 857860.16. Sato D, Saitoh H. Smoothing control of wind farm
output by using kinetic energy of variable speed wind power
generators. IEE Japan 2009; 129-B: 580590. (Japanese).17.
Abedini A, Nasiri A. Output power smoothing for wind turbine
permanent magnet synchronous generators using rotor
inertia. Electric Power Components and Systems 2008; 37:
119.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we
-
A. M. Howlader et al. Fuzzy control strategy for power smoothing
and grid control
18. Banakar H, Luo C, Ooi BT. Steady-state stability analysis of
doubly-fed induction generators under decoupled P-Qcontrol. IEE
Proceedings on Electric Power Applications 2006; 153: 300306.
19. Luo C, Banakar H, Shen B, Ooi BT. Strategies to smooth wind
power fluctuations of wind turbine generator. IEEETransactions on
Energy Conversion 2007; 22: 341349.
20. Chang-Chien LR, Yin YC. Strategies for operating wind power
in a similar manner of conventional power plant. IEEETransactions
on Energy Conversion 2009; 24: 926934.
21. Senjyu T, Sakamoto R, Urasaki N, Funabashi T, Fujita H,
Sekine H. Output power leveling of wind turbine generatorfor all
operating regions by pitch angle control. IEEE Transactions on
Energy Conversion 2006; 21: 467475.
22. Rashad MK, Aymen C, Ken N. Wind power smoothing using fuzzy
logic pitch controller and energy capacitor systemfor improvement
Micro-Grid performance in islanding mode. Energy 2010; 35:
21192129.
23. Tsoukalas LH, Uhrig RE. Fuzzy and Natural Approaches in
Engineering. Wiley: New York, 1997.24. Jauch C, Cronin T, Sorensen
P, Jensen BB. A fuzzy logic pitch angle controller for power system
stabilization. Wind
Energy 2007; 10: 1930.25. Yin M, Li G, Zhou M, Zhao C. Modeling
of the wind turbine with a permanent magnet synchronous generator
for
integration, IEEE PES General Meeting, Tampa, FL, 2007; 16.26.
Senjyu T, Nakasone N, Yona A, Yousuf SA, Funabashi T, Sekine H.
Operation strategies for stability of gearless wind
power generation systems, IEEE PES General Meeting, Pittsburgh
PA, 2008; 17.27. Hu W, Chen Z, Wang Y, Wang Z. Wind power
fluctuations mitigation by DC-link voltage control of variable
speed
wind turbines, Universities Power Engineering Conference, Proda,
2008; 15.28. Matsuzaka T, Tsuchiya K. Study on stabilization of a
wind generation power fluctuation. Transaction on Institute of
Electrical Engineers, Japan 1997; 117-B: 625633.29. Muyeen SM,
Takahashi R, Murata T, Tamura J. A variable speed wind turbine
control strategy to meet wind farm grid
code requirements. IEEE Transactions on Power Systems 2010; 25:
331340.30. Kaneko T, Uehara A, Senjyu T, Yona A, Urasaki N. An
integrated control method for a wind farm to reduce frequency
deviations in a small power system. Applied Energy 2011; 88:
10491058.31. Bossanyi EA. The design of closed loop controllers for
wind turbines. Wind energy 2000; 3: 149163.32. Blaabjerg F,
Teodorescu R, Liserre M, Timbus AV. Overview of control and grid
synchronization for distributed power
generation systems. IEEE Transactions on Industrial Electronics
2006; 53: 13981409.33. Rodriuez P, Pou J, Bergas J, Candela JI,
Burgos RP, Boroyevich D. Decoupled double synchronous reference
frame
PLL for power converters control. IEEE Transactions on Power
Electronics 2007; 22: 584592.34. Lee CC. Fuzzy logic in control
systems: fuzzy logic controller-part I and part II. IEEE
Transactions on Systems, Man,
and Cybernetics 1990; 20: 404435.35. Datta M, Senjyu T, Yona A,
Funabashi T, Kim CH. A co-ordinated control method for levelling PV
output power fluc-
tuations of PV-Diesel hybrid systems connected to isolated power
utility. IEEE Transactions on Energy Conversion2009; 24:
153162.
36. Galdi V, Piccolo A, Siano P. Exploiting maximum energy from
variable speed wind power generation systems byusing an adaptive
Takagi-Sugeno-Kang fuzzy model. Energy Conversion And Management
2009; 50: 413421.
37. Calderaro V, Galdi V, Piccolo A, Siano P. A fuzzy controller
for maximum energy extraction from variable speedwind power
generation systems. Electric Power Systems Research 2008; 78:
11091118.
38. Galdi V, Piccolo A, Siano P. Designing an adaptive fuzzy
controller for maximum wind energy extraction. IEEETransactions On
Energy Conversion 2008; 23: 559569.
39. Cecati C, Ciancetta F, Siano P. A multilevel inverter for PV
systems with fuzzy logic control. IEEE Transactions onIndustrial
Electronics 2010; 57: 41154125.
40. Ruey-Jing L, Shiuh-Jer H. A mixed fuzzy controller for MIMO
systems. Fuzzy Sets and Systems 2001; 120: 7393.41. Wu TF, Yang CH,
Chen YK, Liu ZR. Photovoltaic inverter systems with selftuning
fuzzy control based on an
experimental planning method, Proc 34th IAS Annual Meeting,
Phoenix, USA, Vol. 3, 2001; 18871894.42. Chen YK, Yang CH, Wu YC.
Robust fuzzy control photovoltaic inverter with Taguchi method.
IEEE Transactions on
Aerospace and Electronic Systems 2002; 38: 940954.43. Burkhardt
D, Bonissone P. Automated fuzzy knowledge base generation and
tuning, Proceedings of the FUZZ-IEEE,
San Diego, CA, 1992; 179188.
Wind Energ. (2013) 2013 John Wiley & Sons, Ltd.DOI:
10.1002/we