中国科技论文在线 http://www.paper.edu.

cn

Energy Conversion and Management 88 (2014) 332–347

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Review

The state of the art of wind energy conversion systems and technologies:
A review
Ming Cheng ⇑, Ying Zhu
Research Center for Wind Power Generation, School of Electrical Engineering, Southeast University, Nanjing 210096, China

a r t i c l e i n f o a b s t r a c t

Article history: This paper gives a comprehensive review of the state of the art of wind energy conversion systems
Received 6 September 2013 (WECS) and technologies, with an emphasis on wind power generator and control. First, different types
Accepted 15 August 2014 of common WECSs are classified according to their features and drive train types. The WECSs are com-
pared on the basis of the volume, weight, cost, efficiency, system reliability and fault ride through capa-
bility. The maximum power point tracking (MPPT) control, which aims to make the generator speed meet
Keywords: an optimum value to ensure the maximum energy yield, plays a key role in the variable speed WECSs. A
Wind energy
comprehensive review and comparison of the four most popular MPPT control methods are carried out
Wind power generation
Generator
and improvements for each method are presented. Furthermore, the latest development of wind energy
MPPT control conversion technologies is introduced, such as the brushless doubly fed induction generator (BDFIG), the
Review stator permanent magnet synchronous generators, the magnetic-geared generators, dual power flow
DFIG WECS with the electrical variable transmission (EVT) machine, and direct grid-connected WECS. Finally,
Permanent magnet the future trends of the technologies are discussed.
Direct-drive Ó 2014 Elsevier Ltd. All rights reserved.
Stator PM
EVT
Magnetic-gear

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
2. Traditional types of WECSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
2.1. Constant speed WECS with multiple-stage gearbox and the SCIG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
2.2. Limited variable speed WECS with multiple-stage gearbox and the WRIG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.3. Variable speed WECS with multiple-stage gearbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.3.1. With DFIG and a partial-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.3.2. With SCIG and a full-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.3.3. With SG and a full-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.4. Variable speed direct-drive WECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.4.1. With the EESG and a full-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
2.4.2. With the PMSG and a full-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
2.5. Variable speed WECS with single-stage gearbox. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
2.5.1. With the PMSG and a full-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
2.5.2. With the DFIG and a partial-scale power converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
3. Comparison of the popular WECSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
3.1. Comparison of the volume and weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
3.2. Comparison of the cost and efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
3.3. Comparison of the system reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
3.4. Comparison of the fault ride through capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
4. MPPT control strategies for the variable speed WECSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
4.1. Optimum TSR control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

⇑ Corresponding author.
E-mail address: mcheng@seu.edu.cn (M. Cheng).

http://dx.doi.org/10.1016/j.enconman.2014.08.037
0196-8904/Ó 2014 Elsevier Ltd. All rights reserved.

转载
中国科技论文在线 http://www.paper.edu.cn

M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347 333

Nomenclature

WECS wind energy conversion system C-SCIG–MG constant speed WECS with a SCIG and multiple-
SCIG squirrel-cage induction generator stage gearbox
DFIG doubly fed induction generator DFIG–MG variable speed WECS with a DFIG and multiple-stage
SG synchronous generator gearbox
PMSG permanent magnet synchronous generator EESG–DD direct-drive WECS with a EESG
EESG electrically excited synchronous generator PMSG–DD direct-drive WECS with a PMSG
MPPT maximum power point tracking PMSG–1G WECS with a PMSG and a single-stage gearbox
MPP maximum power point DFIG–1G WECS with a DFIG and a single-stage gearbox
TSR tip speed ratio DSPM doubly salient permanent magnet
HCS hill climb searching FRPM flux reversal permanent magnet
BDFIG brushless doubly fed induction generator FSPM flux switching permanent magnet
WRIG wound rotor induction generator MG magnetic gear
FL fuzzy-logic HTS high temperature superconductor
NN neural network EVT electrical variable transmission
MC matrix converter DPF dual power flow
P&O perturb & observe CFDR constant-frequency double-rotor
WRBFN Wilcoxon radial basis function network PMIG permanent magnet induction generator
TSK Takagi–Sugeno–Kang SS-PMG slip-synchronous permanent magnet generator
k tip speed ratio O&M operation and maintenance
P wind mechanical power q air density
v wind velocity R radius of turbine blades
x rotor speed Cp power coefficient
xopt optimum rotor speed kopt optimum tip speed ratio
Popt optimum wind mechanical power Cp max maximum power coefficient
kopt factor determined by the wind turbine characteristics Topt optimum wind mechanical torque
FRT fault ride through

4.2. Power feedback control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

4.3. Hill climb searching control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
4.4. Fuzzy-logic and neural network based control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
4.5. Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
5. Latest development of wind energy conversion technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
5.1. With BDFIG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
5.2. With stator permanent magnet generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
5.3. Magnetic-geared generator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
5.4. Superconducting generator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
5.5. Dual power flow WECS with electrical variable transmission (EVT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
5.6. Variable speed WECS with constant-frequency double-rotor generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
5.7. Direct grid-connected WECS without converters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
6. Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
7. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

1. Introduction
Fossil fuel is the main energy resource of the worldwide econ-
omy, but the recognition of it as being a major cause of environ-
ment problems makes the mankind look for alternative energy
resources. Then wind energy is becoming the world’s fastest grow-
ing renewable energy, recorded an average growth of 21% in the
past decades (Fig. 1). The new worldwide wind capacity at the
end of 2013 reached 318,105 MW and contributed about 3% of
the global electricity demand [1], due to the fact that it can be eas-
ily captured by wind generators with higher power capacity com-
pared to other renewable energy sources [2,3]. Without doubt
wind power has become a pillar of the energy systems in many
countries and is recognized as a reliable and affordable source of
electricity [4].
The development of wind energy conversion technology has
been going on since 1970s and the rapid development has been
seen from 1990s [5], especially in the last decade, stimulated by
the high growth rate in the wind energy market. Hence, various
new wind energy conversion technologies have been emerging,
aiming at reducing cost, enhancing efficiency and reliability. The
early wind energy conversion system (WECS) type is the constant
speed WECS with a multiple-stage gearbox and a squirrel-cage
induction generator (SCIG) which is connected to the grid directly.
Due to the development of the power electronic technique, the var-
iable speed WECS with a multiple-stage gearbox, a doubly fed
induction generator (DFIG) and a partial-scale converter has been
proposed to expand the system operation range of wind speed
and increase the system efficiency [6,7]. Since the multiple-stage
gearbox is vulnerable, the gearless direct-drive WECS with a syn-
chronous generator (SG) and a full-scale converter has been devel-
oped. Then the permanent magnet synchronous generator (PMSG)
is adopted to replace the traditional electrically excited SG (EESG)
to improve the efficiency and reliability [3]. Currently, the variable
中国科技论文在线 http://www.paper.edu.cn

334 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347

World Total Installed Capacity (MW)

318105
282275
350000

236749
300000

196944
250000

159742
120894
200000

93919
150000

74112
59012
47681
39295
31181
100000

24322
18039
13700
9667
7480
50000

0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Year

Fig. 1. Total installed capacity of wind generators in the world.

speed WECS with a DFIG and the direct-drive WECS with a PMSG generators and new system topologies are introduced. In addition,
are the most popular systems in the market which can be seen the future trends of wind energy conversion technologies are dis-
from Table 1. cussed, such as the fault diagnosis, the fault tolerant control, and
In order to capture as much energy as possible, the maximum the transformerless concept.
power point tracking (MPPT) control is essential for the variable
speed WECS and many MPPT control methods, such as optimum
2. Traditional types of WECSs
tip speed ratio (TSR) control, power feedback control, hill climb
searching (HCS) control and fuzzy-logic, neural network based con-
According to the rotation speed and drive train types, the tradi-
trol, have been developed over these years [8]. In addition, many
tional WECSs can be classified into the following types: constant
advanced measures have been proposed to improve the perfor-
speed WECS with multiple-stage gearbox, limited variable speed
mance of the MPPT control.
WECS with multiple-stage gearbox, variable speed WECS with
On the other hand, to solve the problems existing in the tradi-
multiple-stage gearbox, variable speed direct-drive WECS and
tional generators, novel types of generators, such as brushless
variable speed WECS with a single-stage gearbox.
DFIGs (BDFIG), the stator PM generators, and the superconducting
generators, have been proposed by researchers. Furthermore, some
novel wind energy conversion technologies have also been pre- 2.1. Constant speed WECS with multiple-stage gearbox and the SCIG
sented, such as the dual power flow WECS based on the electrical
variable transmission (EVT) machine to eliminate the mechanical The constant speed WECS has been used with a multiple-stage
gearboxes, and the direct grid-connected WECS without converters. gearbox and a SCIG which is connected to the grid through a trans-
The objective of this paper is to provide a comprehensive former directly, as shown in Fig. 2. This concept was applied by
review of the state of the art of WECSs and technologies, including many Danish wind turbine manufactures during the 1980s and
the various typical types of existing WECSs, several MPPT control 1990s due to its simplicity, reliability and low cost [2]. However,
methods and detailed comparisons of the different WECSs and one disadvantage of this WECS is the requirement for a reactive
MPPT methods. The latest development of WECSs with new type power source to establish the magnetic field. Then a capacitor bank

Table 1
Typical products of the traditional WECS.

Speed range Gearbox Converter Generator type Capacity Manufacturer Reference

Variable speed Multi-stage Partial scale DFIG 1.5 MW, 3.0 MW, 5 MW Sinovel [15]
2 MW, 2.6 MW, 3 MW Vestas [16]
1.5 MW GE wind [17]
0.85 MW, 2 MW, 2.5 MW Gamesa [19]
2.4 MW, 2.5 MW, 3.0 MW, 3.3 MW Nordex [21]
1.0 MW, 1.5 MW, 2.0 MW, 2.5 M Dong Fang [22]
2.0 MW, 3.0 MW, 6.0 MW REpower [23]
2.1 MW Suzlon [27]
Full scale PMSG 3.3 MW Vestas [16]
2.5 MW GE wind [17]
3.0 MW WinwinD [18]
4.5 MW, 5.0 MW Gamesa [19]
SCIG 6.0 MW Sinovel [15]
Direct-drive Full scale PMSG 8.0 MW Vestas [16]
4.0 MW GE wind [17]
3 MW, 6 MW Siemens wind [20]
1.5 MW, 2.0 MW STX wind [28]
1.5 MW, 2.5 MW Goldwind [24]
EESG 0.8 MW-7.58 MW Enercon [25]
Single-stage Full scale PMSG 1.0 MW WinwinD [18]
5.0 MW Multibrid (Areva) [26]
Limited variable speed Multi-stage – WRIG 1.5 MW, 2.1 MW Suzlon [27]
Constant speed Multi-stage – SCIG 0.6 MW, 1.25 MW Suzlon [27]
2.3 MW, 3.6 MW, 4 MW Siemens wind [20]
中国科技论文在线
M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
Gearbox SCIG
Soft starter Grid
Capacitor
Fig. 2. Scheme of a constant speed WECS with multiple-stage geared SCIG.
for reactive power compensation was extended during 1980s.
Besides, a soft starter is usually used to reduce the inrush currents
during start up [9].
The disadvantages of this WECS are:
(1) The speed of the SCIG for the constant speed WECS is not
controllable and only varies in a narrow range around the
synchronous speed. Then the maximum energy cannot be
captured from the wind over a wide range of wind speeds.
(2) The fluctuations of wind speed are directly transmitted into
the electromechanical torque, and then into the electrical
power, degrading the quality of electrical energy. Further-
more, the fluctuations may cause high mechanical and fati-
gue stresses on the whole system and may result in swing
oscillations between wind turbine and generator shaft.
(3) The multiple-stage gearbox used in the system inevitably
has some drawbacks, such as heat dissipation from friction,
regular maintenance and audible noise.
2.2. Limited variable speed WECS with multiple-stage gearbox and the
WRIG
The limited variable speed WECS is composed of a multiple-
stage gearbox and a wound rotor induction generator (WRIG) as
shown in Fig. 3. In this WECS, the rotor windings are connected
with variable resistor which can be adjusted by the electronic con-
trol system. The slip of the WRIG is higher as the resistance of the
rotor windings is higher. However, the high slip means a high
power extracted by the rotor, in turn low generator efficiency.
With the increase in variable speed range, the rating of the resis-
tors must be high. Therefore, the variable speed range of the WRIG
is limited by the size of the rotor resistors. A typical limited vari-
able speed range is less than 10% above the synchronous speed
[2]. In addition, both the SCIG and the WRIG need to operate at a
speed higher than the synchronous speed to generate electricity.
Conventionally, the resistors are connected to rotor windings
via brushes and slip rings, which is a drawback compared to the
conventional SCIG. Furthermore, the capacitor for reactive power
compensation and the soft starter are also required for the limited
variable speed WECS.
2.3. Variable speed WECS with multiple-stage gearbox
2.3.1. With DFIG and a partial-scale power converter
The scheme of the variable speed WECS with a multiple-stage
gearbox, a DFIG and a partial-scale power converter is shown in
Gearbox WRIG Grid
Soft starter
Capacitor
Variable resistance
Fig. 3. Scheme of a limited speed WECS with multiple-stage geared WRIG.
http://www.paper.edu.cn
335
Fig. 4. The rotor of the DFIG is connected to the grid through a
back-to-back converter while the stator is directly connected to
the grid. The Crowbar is connected to the rotor-side converter to
protect the converters. The DFIG can deliver wind energy to the
grid in a wide speed range from supersynchronous to subsynchro-
nous speeds. Conventionally, the variable speed range of the DFIG
is ±30% around the synchronous speed. Then the nominal power of
the back-to-back converter is only about 30% of the wind turbine
rated power [10].
The generator speed can be regulated to optimize the power
extraction from the wind by controlling the active power of the
rotor-side converter. Besides, the reactive power delivered to the
grid can also be controlled by the rotor-side converter through
the decoupling of active power and reactive power. However,
several disadvantages of this WECS exist such as [11]:
(1) The slip rings and brushes used in the DFIG require frequent
maintenance and may lead to machine failures and electrical
losses.
(2) The fault ride through (FRT) capability of the WECS with
DFIG is relatively weak due to the direct connection of the
DFIG stator and grid. Thus, the corresponding control strate-
gies for the system used to ride through the grid faults are
quite difficult and complicated.
2.3.2. With SCIG and a full-scale power converter
The SCIG has the advantages of simple structure, low cost, high
reliability and so on. Then in order to expand the speed operation
range of the SCIG, an alternative variable speed WECS with SCIG,
which is different from the Danish concept, is proposed as shown
in Fig. 5. The multiple-stage geared SCIG is connected with the grid
through a full-scale back-to-back converter. The generator speed
can vary by controlling the generator side converter while the DC
bus voltage and the reactive power transmit to the grid can be
controlled through the grid-side converter.
The disadvantages of this concept are the high cost and the
losses of the full-scale converter compared to the constant speed
WECS with SCIG.
2.3.3. With SG and a full-scale power converter
The variable speed WECS with a multiple-stage geared SG and a
full-scale power converter is shown in Fig. 6. The usage of the mul-
tiple-stage gearbox can reduce the volume and weight of the SG
while comparing with the direct-drive system, and then decrease
the difficulty of designing and manufacturing the generator. In this
system, although both electrically excited synchronous generator
(EESG) and permanent magnet synchronous generator (PMSG)
can be used, the later is more popular due to the fact that the PMSG
offers advantages of high efficiency and robust structure. However,
the cost of the PMSG is higher than that of the EESG, SCIG and DFIG
due to the expensive permanent magnet materials.
Gearbox DFIG Grid
AC DC
Crowbar DC AC
Fig. 4. Scheme of a variable speed WECS with multiple-stage geared DFIG and
partial-scale converter.
中国科技论文在线 http://www.paper.edu.cn

336 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347

Grid Grid
Gearbox AC DC
SCIG DC AC AC DC
PMSG DC AC

Fig. 5. Scheme of a variable speed WECS with multiple-stage geared SCIG and full-
scale converter. Fig. 8. Scheme of a direct-drive variable speed WECS with PMSG and full-scale
converter.

Grid
Gearbox PMSG AC DC
DC AC
Fig. 6. Scheme of a variable speed WECS with multiple-stage geared PMSG and full-
scale converter.
2.4. Variable speed direct-drive WECS
Multiple-stage gearboxes are responsible for the losses in the
WECS with a high speed generator. The rule of thumb is that 1%
of the power applied at the input shaft is lost for each stage. Today
many large wind turbines use three-stage gearboxes, and then only
about 97% of the input power is transmitted to the output shaft
[12]. Moreover, the stages of the gearboxes will be further
increased with the increasing of wind turbine power capacity thus
the more gearbox losses will be produced. Besides, the multi-stage
mechanical gearbox is rather vulnerable and poses many difficul-
ties to maintain, particularly for high-capacity offshore wind farms
[13]. Therefore, the direct-drive WECSs are more attractive at pres-
ent. Hence, the direct-drive WECS has got increasing acceptance in
recent years due to high efficiency, high reliability and less fre-
quent maintenance.
2.4.1. With the EESG and a full-scale power converter
The direct-drive variable speed WECS with the EESG and full-
scale converter is shown in Fig. 7. The EESG is built with a rotor
carrying the field system supplied with a DC excitation. The gener-
ator must be designed with high number of poles to make the gear-
less system possible. Consequently, the volume and weight of
these low speed generators are much greater than those of the
multiple-stage geared generators.
The slip rings and brushes are required in the EESG for the excit-
ing windings which increase the system maintenance requirement.
In addition, the field winding will produce power losses, thus
degrading the system efficiency.
and reliability are high. Besides, the weight and size of the PMSG
are smaller than those of the EESG due to the usage of high perfor-
mance permanent magnets. An additional advantage of this system
is the noise reduction due to the absence of the gearbox and inde-
pendent excitation system.
In direct-drive PMSG, the outer rotor integrated with hub may
be adopted so that the drive train can be much shortened. Fig. 9
illustrates a direct-drive WECS with outer rotor PMSG.
At this moment, it is not practical for the market wind genera-
tors to have an external diameter larger than 8 m because of the
logistics and construction technology limitations, which limit the
development of the direct-drive WECSs with larger capacity [12].
2.5. Variable speed WECS with single-stage gearbox
2.5.1. With the PMSG and a full-scale power converter
For the increasing power levels and decreasing turbine speeds,
the direct-drive generators are becoming larger, heavier and more
expensive. As a trade-off between the generator size and reliability,
the WECS with a single-stage gearbox, a medium-speed PMSG and
a full-scale power converter was first introduced by Multibrid in
Germany [14]. Hence this system is also called Multibrid system.
However, this system combines some of the disadvantages of
both the geared and direct-drive systems, as the system has a gear-
box, a relatively expensive generator and a full-scale converter.
Compared to the traditional direct-drive system, the significant
decrease in the generator cost is obtained, but the cost for the gear-
box is increased. Besides, considering the gearbox cost and gener-
ator cost, the system optimization is necessary to determine the
gearbox ratios and generator diameters at various power levels
for the Multibrid WECS.
2.5.2. With the DFIG and a partial-scale power converter
The system with a single-stage gearbox, a DFIG and a partial-
scale power converter was proposed in [3]. The rated power of
the converter is only 30% of the system which offers an important

2.4.2. With the PMSG and a full-scale power converter

The direct-dive WECS with PMSG as the most promising system
currently is shown in Fig. 8, where external excitation and slip
rings are no needed compared to the EESG, thus the efficiency

Grid
AC DC
EESG DC AC

DC
AC

Fig. 7. Scheme of a direct-drive variable speed WECS with EESG and full-scale
converter. Fig. 9. Direct-drive WECS with outer rotor PMSG.
中国科技论文在线 http://www.paper.edu.cn

M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347 337

benefit in cost and efficiency compared to the system with full-
scale converter. However, due to the only one-stage speed increas-
ing, the generator speed is still low and the torque is still high, so
the DFIG can be expected to have a large diameter and air gap, in
turn, high magnetizing current and high losses.
Table 1 lists the typical products of the aforementioned tradi-
tional types of WECSs [15–28].
is about two times larger and heavier than that in DFIG–3G system,
but is smaller than that in the other three systems.
Considering the wind turbine blade, the gearbox, the generator
and the converter, there is no big difference among the total
volume and weight of the five variable speed systems.
3.2. Comparison of the cost and efficiency

3. Comparison of the popular WECSs
The six aforementioned WECSs, which are the constant speed
WECS with a SCIG and multiple-stage gearbox (C-SCIG–MG), the
variable speed WECS with a DFIG and multiple-stage gearbox
(DFIG–MG), the direct-drive WECS with a EESG (EESG–DD), the
direct-drive WECS with a PMSG (PMSG–DD), the WECS with a
PMSG and a single-stage gearbox (PMSG–1G) and the WECS with
a DFIG and a single-stage gearbox (DFIG–1G), are compared here
and several criteria are taken into account.
3.1. Comparison of the volume and weight
The quantitative comparisons of five 3 MW systems except for
the C-SCIG–MG system are given in [3]. On the basis of this paper,
the comparisons are modified according to the present state of
WECS products shown in Table 1. The volume and weight compar-
isons of the six typical systems are listed in Table 2. Assuming that
the volume and weight of the DFIG–MG is the value 1 which is
dimensionless, and then the other five systems regard it as the
standard value to evaluate their volume and weight.
As shown in Table 2, the DFIG–MG generator is the smallest and
lightest one in the variable speed systems, which explain why it is
the most popular WECS in the commercial market. The SCIG is a bit
smaller and lighter than the DFIG due to the absence of brushes
and slip rings. Besides, the back-to-back converter is not needed
in the C-SCIG–MG system and then the system volume and weight
is smaller and lighter than that of DFIG–MG system. However, the
constant speed system is having a less and less market share
nowadays because the system cannot make the best use of wind
energy.
Due to the omission of the multi-stage gearbox, the EESG–DD
generator seems to be the biggest and heaviest solution, so only
Enercon uses this system in practical. The volume and weight of
the PMSG–DD generator are smaller and lighter than those of
EESG–DD due to the high power density of the permanent magnet
materials. The weight and size of the generator in the PMSG–1G
system are obviously decreased compared to those in PMSG–DD
system. Therefore, the PMSG–1G system is now attractive for the
high power wind turbines. The generator in the DFIG–1G system
The comparisons of the six typical systems considering cost and
energy yield are shown in Table 3. Similarly, assuming that the cost
and energy yield of the DFIG–MG are the value 1 which is dimen-
sionless, then this value is regarded as the standard to judge the
other five systems. The average annual wind speed for the yield
calculation is 7 m/s with a Weibull distribution.
The ratio of the energy yield and cost of a WECS is the essential
criterion which is in direct relation to its commercial value. By
comparing the structures of the six systems, the cost of the C-
SCIG–MG seems to be the least due to the almost free maintenance,
no need of converter and relatively cheap SCIGs, but the energy
yield of the constant speed system is much lower than that of
the variable speed system. Therefore, the market penetration of
this concept has decreased dramatically over the twenty years [2].
The EESG–DD appears to be the most expensive system accord-
ing to the research results in [3], while the PMSG–DD is the second
most expensive system because of the full-scale converter and
large multi-pole generator. And yet, the efficiency of the PMSG-
DD is the highest owing to the high efficiency of PM materials
and absence of the gearbox. The energy yield of EESG–DD is lower
than that of PMSG–DD due to the low efficiency of field windings
but still higher than the other three systems owing to the rid of
gearboxes. However, the value of energy yield/cost of EESG–DD
is still lowest due to the high cost.
The efficiency of the PMSG–1G is higher than that of DFIG–3G
and DFIG–1G, while that of the DFIG–1G is higher than DFIG–3G.
Meanwhile, the cost of PMSG–1G is higher than that of the system
with DFIG. Finally, it is interesting that the DFIG–1G has the higher
energy yield/cost than PMSG–1G and DFIG–3G according to the
research [3].
3.3. Comparison of the system reliability
The reliability of the WECS is influenced by many factors. The
multiple-stage gearbox is the most vulnerable part in high speed
systems such as the C-SCIG–MG and DFIG–MG. The brushes and
slip rings used in the DFIG and EESG should be replaced regularly
which increase the maintenance cost and reduce the system reli-
ability. Besides, the power electronics failures cannot be ignored
during the consideration. Usually, the failures of the full-scale

Table 2
Comparisons of the six WECSs considering volume and weight.

C-SCIG–MG DFIG–MG EESG–DD PMSG–DD PMSG–1G DFIG–1G

Generator volume 0.95 1 5.5 5 2 3
Generator weight 0.95 1 6 4.5 1.3 2.2
System volume 0.9 1 1.1 1.1 1.1 1.1
System weight 0.9 1 1.1 1.1 1.1 1.1

Table 3
Comparisons of the six WECSs considering cost and energy yield.

C-SCIG–3G DFIG–3G EESG–DD PMSG–DD PMSG–1G DFIG–1G

System cost 0.9 1 1.13 1.06 1.02 0.98
Energy yield 0.8 1 1.02 1.04 1.02 1.01
Energy yield/cost 0.89 1 0.9 0.98 1 1.03
中国科技论文在线
338 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
Table 4
Comparisons of the six WECSs considering the system reliability and FRT capability.
C-SCIG–3G DFIG–3G EESG–DD
Reliability 3 2 3 5
FRT capability 2 2 5 5
converter in direct-drive system are more than that of the partial-
scale converter in the system with DFIG [29].
Variable speed control and pitch control enable load reductions
[30] while constant speed stall controlled WECSs may have signif-
icant overshoots in power and load which also influence the reli-
ability [31]. The simple system like C-SCIG–MG with few
components is more reliable than the complex systems.
Combining with the above mentioned factors, the reliability
comparison of the six systems are given in Table 4 roughly, where
the reliability is graded from 1 to 5 points. 5 Point is the best, the
higher the point, the better. The PMSG–DD and PMSG–1D are con-
sidered as enjoying the relatively high reliability while the DFIG–
3G is relatively weak. The reliability of the other systems is in
the middle.
3.4. Comparison of the fault ride through capability
The increasing penetration of wind energy into the power sys-
tem is reshaping the way wind farms operate directly. The most
demanding requisite for the wind farms is FRT capability [32].
Wind farms must remain be connected to the grid when grid volt-
age dip occurs, and then ride through the grid faults and contribute
to system stability after fault clearance. The comparison of the six
systems considering the FRT capability is listed in Table 4. The FRT
capability is graded from 1 to 5 points while 5 point is the best, the
higher the point, the better.
The DFIG-based WECS is vulnerable to grid faults because the
stator of the generator is directly connected to the grid. The stator
flux cannot follow the sudden change in stator voltage and a DC
component in the stator flux appears when a voltage dip occurs.
The rotor keeps rotating and a high slip occurs, which tends to
cause an overvoltage and overcurrent in the rotor. Besides, asym-
metrical faults can cause higher overvoltages and overcurrents
since a negative sequence component exists in the stator voltage
and the slip of this negative sequence component is very high
[33]. A large number of research projects are carried out to improve
the FRT capability of the DFIG [34–36]. However, it is still difficult
to achieve the FRT of the DFIG-based WECS perfectly.
The C-SCIG–MG also has a high sensitivity to voltage sags due to
the fact that the stator is directly connected to the grid. But this
system can stay be connected to the grid during a grid fault and
produce high currents because the stator currents can be much lar-
ger than the rated value for a short time. Unfortunately, the reac-
tive power cannot be controlled for voltage recovery after a fault
which influences the FRT capability. Consequently, the FRT capabil-
ity of the C-SCIG–MG is weak [37].
The WECSs based on the SGs such as EESG–DD, PMSG–DD,
PMSG–1G are connected with the grid thorough full-scale convert-
ers, hence grid faults have no direct effect on the generators.
Besides, these systems have a larger capability of reactive power
supply than the systems based on DFIGs. Therefore, these three
systems have a stronger FRT capability than C-SCIG–MG and
DFIG-based systems.
4. MPPT control strategies for the variable speed WECSs
To maximize the use of wind energy, the MPPT control for the
variable speed WECSs is indispensable when the wind speed is
http://www.paper.edu.cn
PMSG–DD PMSG–1G DFIG–1G
4 3
5 2
below the rated speed. The mechanical power extracted by a wind
turbine can be given as [38]:
P ¼ qpR2 C p ðk; bÞv 3 =2 ð1Þ
where q is the air density, R is the radius of turbine blades, v is the
wind velocity, and Cp is the power coefficient which is a nonlinear
function of the tip speed ratio k and the blade pitch angle b. k is
defined as:
k ¼ xR=v ð2Þ
Fig. 10 shows the curve of Cp versus k. It is clear that there is a
value of k, for which Cp is maximized, thus maximizing the power
for a given wind speed [39]. Then the variable speed WECS follows
the Cp max to capture the maximum power by varying the rotor
speed to keep the system at the optimum TSR kopt.
To keep the wind turbine at the optimum TSR, the speed of the
wind turbine is controlled by regulating the torque of the genera-
tor. Hence, various MPPT algorithms have been developed and can
be categorized into four types: optimum TSR control, hill climb
searching (HCS) control, power feedback control and fuzzy-logic-
based control [40–42]. Each of the methods has their own advan-
tages and disadvantages, hence many variations of these methods
have been proposed to solve the relevant weaknesses using differ-
ent techniques over the past years.
4.1. Optimum TSR control
The optimum TSR control regulates the wind turbine rotor speed
to maintain an optimum TSR, as shown in Fig. 11. This MPPT
method requires the measurement of both the wind speed and tur-
bine speed. The wind speed is usually measured by an anemometer
which increases the cost of the system and the accurate value is
almost impossibly obtained in practice. Besides, the wind turbine
characteristic which varies from system to system is also indispens-
able. Therefore, the optimum TSR MPPT method is rarely applied in
the actual WECSs though the control strategy is very simple.
4.2. Power feedback control
The power feedback MPPT control is implemented based on the
calculations of the generator speed without the need of wind
0.5
C p max
0.4
0.3
Cp
0.2
0.1
λopt
0 2 4 6 8 10 12
λ
Fig. 10. Curve of power coefficient versus tip speed ratio k.
中国科技论文在线 http://www.paper.edu.cn

M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347 339

overtime. (2) Even when the kopt can be accurately determined via
simulation or experiments, wind speed fluctuations make the gen-
load erator speed difficult to be measured exactly and force the wind
Generator Converter
turbine to operate off the peak of its Cp curve much of the time.
Due to the disadvantages existing in the traditional power feed-
back control, many effective measures for improvement have been
proposed by the researchers. The sliding mode power feedback
MPPT method was proposed in [43], which can reduce the negative
+-
opt
v Control effects of both the uncertainty regarding kopt and the change in
opt R system optimal operating point due to wind speed turbulence. In [44], a
mechanical speed-sensorless power feedback control with a matrix
Fig. 11. Control diagram of optimum TSR MPPT method. converter (MC) was proposed without requiring a shaft speed sen-
sor. A novel MPPT control with an adaptive compensation control
was proposed in [45] based on the optimum torque feedback con-
speed, but the wind turbine characteristics and the generator
trol. Based on the adaptive control, the dynamic response can be
speed are indispensable. This method can be divided into two cat-
improved and more wind energy can be captured during the wind
egories: the optimum power feedback control and the optimum
speed variations. The optimum current MPPT method based on the
torque feedback control. Nonetheless the two control methods
method in [46] was proposed in [39] with considering the system
are essentially the same.
torque losses, consequently the efficiency of wind energy conver-
According to Eqs. (1) and (2), the optimum power and torque
sion is increased.
output by the wind turbine are given by [39]

Popt ¼ qpR5 C p max x3opt =2k3opt ¼ kopt x3opt ð3Þ

5 2 3 2
T opt ¼ qpR C p max x opt =2kopt ¼ kopt x opt
where kopt is a factor determined by the wind turbine
characteristics.
The optimum power and torque are corresponding to an
optimum generator speed with a relationship of (3) and (4), which
can be seen in Fig. 12 intuitively [39]. In other words, the maxi-
mum power and torque can be achieved by the wind turbine when
the generator speed is at the optimum value. The control diagram
of the traditional optimum power feedback and optimum torque
feedback MPPT control strategies are shown in Fig. 13(a and b)
respectively.
Two significant drawbacks of the traditional power feedback
control method are: (1) There is no accurate way to determine
the factor kopt since blade aerodynamics can change significantly
4.3. Hill climb searching control
ð4Þ
The hill climb searching (HCS) control can be employed by per-
turbing the control variable and observing the resulting increase or
decrease in power, hence also known as perturb & observe (P&O)
method, as shown in Fig. 14. If it results in the increase in power,
then the same perturbation is applied for the next control instance.
Otherwise the sign of the perturbation is reversed in order to track
in the direction of increasing power [47]. The HCS control has the
advantages of no need of the measurement of wind speed or gen-
erator speed and independent of the system characteristics. So the
HCS seems to be the top choice for MPPT control in the WECSs due
to these features.
However, serious disadvantages exist in the HCS method. As
shown in Fig. 14(a), a large perturbation step size increases the
speed of convergence but deteriorates the efficiency of MPPT by

P T Topt
Popt
P3 v3 > v2 > v1 T3
v3 v3 v3 > v2 > v1
P2 T2
v2
P1 T1 v2
v1 v1
0 0

(a) power versus generator speed (b) torque versus generator speed

Fig. 12. Characteristics of wind turbine.

load Generator load

Generator Converter Converter

Control system
P T T
Popt + - Po + - Control
Topt system

(a) Optimum power feedback (b) Optimum torque feedback

Fig. 13. Control diagram of power feedback MPPT method.

中国科技论文在线 http://www.paper.edu.cn

340 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347

P Pmax P Pmax

Uphill Downhill
Uphill
Downhill

P
P

(a) a large perturbation step (b) a small perturbation step

Fig. 14. Principle of hill climb searching control.

amplifying the oscillations around the maximum power point noisy and inaccurate signals [53]. Three fuzzy-logic (FL) steps are
(MPP) Pmax. The smaller step size as shown in Fig. 14(b) boosts adopted in [54]. The first FL step which is based on the HCS varies
the efficiency but reduces the MPPT control speed which may generator speed, observes power output and then tracks the gener-
result in incapable of tracking the exact MPP under rapidly varying ator speed to extract the maximum power. The second FL step
wind speed. Furthermore, the sign of the next perturbation, which programs the machine flux by an on-line search so as to optimize
is decided by the increase or decrease in power due to the previous the machine-converter system efficiency. The third FL step is used
perturbation, might be mislead when the wind speed change rap- to provide robust speed control against turbine oscillatory torque
idly. Then the wrong decision can lead to the failure of the whole and wind vortex.
MPPT control. The papers [55,56] present a data driven design methodology
Therefore, traditional HCS control works well only when the able to generate a Takagi–Sugeno–Kang (TSK) fuzzy model for
wind turbine inertia is small so that the turbine speed reacts to the MPPT control. In order to obtain the TSK model, fuzzy cluster-
wind speed almost instantaneously. For wind turbines with large ing methods for partitioning the input–output space, combined
inertia, the system output power is interlaced with the turbine with genetic algorithms, and recursive the least-squares optimiza-
mechanical power and rate of change in the mechanically stored tion methods for model parameter adaptation are used. The TSK
energy, which can always render the HCS MPPT control a failure fuzzy system can be significant for wind turbine manufactures
especially under rapidly changing wind conditions. because of its adaptive and learning ability.
Thus some improved measures based on this method have been The neural network (NN) based wind velocity estimator was
proposed. In [48], an advanced intelligent HCS control method, developed in [57,58] to provide fast and accurate velocity informa-
which can overcome the difficulties of the previous conventional tion to avoid using anemometers. Besides, the NN based scheme
HCS methods caused by wind turbine inertia, was proposed. By was proposed to compensate the potential drift of wind turbine
recording the HCS searching results through an on-line training power coefficient without extra sensors.
process in an intelligent memory, the algorithm can record the The MPPT method based on an adaptive neuro-fuzzy inference
optimum system operating conditions, and then use direct current system was presented in [59], which was designed as a combina-
demand to find the maximum power points rapidly and effectively. tion of the Sugeno fuzzy model and neural network. From the var-
Another improved MPPT method in [49] can reduce the turbine iation of the wind speed, a neuro-fuzzy model made up of radial
mechanical stress with regard to conventional techniques, so that basis functions computes the optimal rotation speed and thus
both the maintenance needs and the medium time between fail- the aerodynamic torque.
ures are expected to be improved. The proposed approach is simi-
lar to the traditional HCS method, but a ramp signal instead of a 4.5. Comparison
stepped signal is used to modify the reference speed of the gener-
ator, obtaining a softer response of the mechanical variables than Based on the above description, the comparisons of the four
the typical one of conventional HCS methods. MPPT methods are given in Table 5 by considering the important
The authors in [50] proposed a fast HCS control algorithm with- criteria—tracking speed, complexity and performance under
out the need of any extra hardware, which is much quicker than varying wind, where ‘‘+’’ denotes strong, ‘‘’’ denotes weak, and
most of the existing HCS MPPT schemes. The algorithm is divided ‘‘+/’’ denotes medium. The optimum TSR MPPT method needs
into two stages: In the first stage, large iterative steps are used the most system parameters thus it is not practical in the applica-
to move within a close range of MPP. In the second stage, tradi- tion, though the method provides many advantages, including fast
tional HCS method is used to track the exact MPP corresponding tracking speed, implementation simplicity and minor effect of
to the current wind speed. An intelligent HCS MPPT with Wilcoxon varying wind. The most commonly used method in practical is
radial basis function network (WRBFN) strategy was proposed in
[51,52]. The WRBFN with modified particle swarm optimization
Table 5
algorithm forces the system to reach its equilibrium quickly where
Comparisons of the typical four MPPT methods (+: strong, : weak, +/: medium).
the turbine inertia effect is minimized. The proposed HCS control
can be fat and effective in spite of variations in wind speeds and Optimum Power HCS FL, NN
TSR feedback based
the presence of turbine inertia.
Wind speed Need No No No
Rotor speed Need Need Need Need
4.4. Fuzzy-logic and neural network based control Wind turbine characteristics Need Need No No
Tracking speed + +/  
Complexity + + + 
The fuzzy-logic-based MPPT control methods have the advanta- Performance under varying wind + +/  +/
ges of fast convergence, parameter insensitivity and acceptance of
中国科技论文在线
M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
the power feedback control due to no need of the wind speed, sim-
ple implementation and good dynamic response. The HCS MPPT
control methods are studied and verified in small wind turbines
by many research papers such as [60] but still not proved to be
applied in large-scale wind turbines due to the poor dynamic
response under wind turbulence. The fuzzy-logic and neutral net-
work based MPPT methods are too complex thus a lot of time is
spend on the algorithm computing which will affect the tracking
speed and performance under varying wind. Therefore, they are
still difficult to be realized in practice at this moment. At present,
as more and more improvement measures are proposed based on
these control methods, the applied range of each method is
extended thus they will probably be available in the future.
5. Latest development of wind energy conversion technologies
5.1. With BDFIG
The brushless DFIG (BDFIG) is attractive for use in the WECS to
replace the traditional DFIG, especially in offshore places, as it
offers high reliability and low maintenance requirements by virtue
of the absence of brushes and slip rings [61]. The scheme of the
variable speed WECS with a special cage rotor based BDFIG is
shown in Fig. 15.
The stator of the BDFIG is furnished with two separate windings
which differ in pole pair numbers to avoid direct coupling between
the windings. The rotor of the BDFIG is designed to couple the two
stator windings [62]. One stator winding is connected to the grid
directly as called power winding. Another stator winding is con-
nected to the back-to-back converter as called control winding.
Similar to the DFIG system, the nominal power of the converter
is only a fraction of the wind turbine rated power. The control
winding side converter can regulate the active power and obtain
the maximum wind energy while the grid-side converter can
regulate the reactive power transmit to the grid. The control strat-
egy is similar to the strategy for the WECS with the DFIG but is
more complicated.
The rotor configuration of the BDFIGs can be divided into
special cage type and reluctance type [63,64]. According to the
rotor bar connection rules the special cage rotor type can be clas-
sified into two kinds, independent concentric cage type and inde-
pendent concentric cage with a public end-ring type. The bar
Back-to-back
f1 converter
A Stator power winding C
(pole 2p) b
fc1
a c
Stator control winding
Rotor (pole 2pc)
Fig. 15. Scheme of a variable speed WECS with multiple-stage geared BDFIG.
http://www.paper.edu.cn
341
connection of the two type cage rotors can be seen in Fig. 16(a
and b) respectively. The reluctance rotor type can be classified into
salient pole type, axially-laminated type and radial-laminated
type. The sheet of radial-laminated reluctance type rotor with axial
flux barriers is shown in Fig. 16(c).
At present, the research of the WECS with BDFIG is still at the
early-stage. A lot of work remains to be done in devising a simpler
method of determining the balance of electric and magnetic
loadings in the machine, especially in the rotor. Furthermore, the
control of the BDFIG remains a challenge which limits the practical
application [65].
5.2. With stator permanent magnet generators
In recent years, many types of generators are proposed to
improve the performance of WECSs, especially PM generators
[66]. Traditional PM generators all have magnets on the rotor.
Hence, the temperature rise in magnets may be a problem due to
poor thermal dissipation, which may cause irreversible demagneti-
zation of magnets and may ultimately limit the power density of
the generator. Recently, in contrast, a new type of PM generator
having magnets on the stator, so called the stator-PM machines,
has attracted more and more attention [67]. Basically, there are
three types of modern stator-PM machines, namely, doubly salient
PM (DSPM) machine, flux reversal PM (FRPM) machine, and flux-
switching PM (FSPM) machine. Fig. 17 shows the schematics of
the DSPM and FSPM generators. It can be seen that both the PMs
and windings are placed on the stator, and consequently, the
temperature rise of the magnets is much easier to control. Concen-
trated windings are employed with shorter end parts, resulting in
(a) independent concentric cage rotor
(b) independent concentric cage rotor with a public end ring
(c) radial-laminated reluctance type rotor
Fig. 16. Different type rotors of BDFIGs.
中国科技论文在线
342 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
PM
Winding
Rotor
Stator
(a) Doubly salient PM generator
Fig. 17. Stator permanent magnet generators.
Outer rotor Permanent
magnets
Stationary
ring
Airgap
Stator
Fig. 18. Magnetic-geared permanent magnet generator.
low copper consumption and winding resistance. The salient rotor
is made up of iron with neither windings nor magnets, being
mechanically robust. Several types of stator PM generators have
been proposed for the WECSs [68–72], showing good potential
for wind power application. Moreover, it is convenient for the sta-
tor PM generator to use modular stator design which is attractive
for high power direct-drive PM generators so as to solve the prob-
lem of manufacture and transportation.
Moreover, by replacing the permanent magnets with electrical
field windings, the stator electrical excited generator can be
obtained [73,74]. If both the permanent magnets and field wind-
ings are mounted in the stator, the stator hybrid excited generator
can be achieved [75,76]. Both the stator electrical excited and
hybrid excited generators offer the possibility of regulating the
flux, in turn the output voltage of the generator.
5.3. Magnetic-geared generator
The magnetic gear (MG) offers significant advantages of
reduced acoustic noise, minimum vibration, free from mainte-
nance, improved reliability, inherent overload protection and
physical isolation between the input and output shafts, as com-
pared with traditional mechanical gear. By adopting the coaxial
topology, the utilization of PMs can be greatly improved, leading
to the offering of a torque density comparable with that of the
mechanical gear [77]. Recently, the MG is integrated into a PM
machine to offer high-torque low-speed operation [78]. Fig. 18
illustrates a magnetic-geared PM generator with outer rotor which
can be directly coupled with wind turbine. It can be seen from
Fig. 18 that there is a stationary ring between the rotor and stator,
which modulates the magnetic field. Thus, the rotor magnetic field
can be modulated from low speed in the rotor to high speed in the
stator. Hence, the armature windings in the stator can be designed
based on high speed magnetic field.
http://www.paper.edu.cn
(b) Flux switching PM generator
It has shown that the proposed magnetic-geared PM generator
is smaller, lighter and with lower material cost than the conven-
tional direct-drive PM generators.
5.4. Superconducting generator
As aforementioned, the direct-drive WECS is very promising
due to the advantages of high efficiency and reliability. However,
since the direct-drive generator rotates at a low speed, the
direct-drive WECS has the drawbacks of bulky volume and heavy
weight associated with high cost, which limits the development
of the direct-drive system.
In recent years, the development of high-temperature super-
conductor (HTS) machines provides an opportunity for the
direct-drive WEC system having the merits of high power density
and high efficiency. A 10 MW superconducting direct-drive gener-
ator with weight of 160 tons, which is roughly only 50% of that of a
direct-drive PMSG generator, has been designed by American
Superconductor [79].
In general, the HTS machines excited by field windings can be
sorted into two categories, namely the rotor excitation and stator
excitation according to the location of the HTS field windings,
respectively. For the rotor excitation, the excitation current has
to feed through slip rings and brushes due to HTS field windings
mounted in the rotor, which make the refrigeration complicated
and difficult. In [80], a flux-switching HTS machine for the WECS
was proposed based on the traditional FSPM by replacing the per-
manent magnets with HTS windings, as shown in Fig. 19. Obvi-
ously, it is much easier for refrigeration than its rotor excitation
counterpart because the HTS field windings locate in the stator.
Stator
Copper armature
windings
HTS field
windings
Rotor
Fig. 19. Flux switching HTS machine.
中国科技论文在线
M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
5.5. Dual power flow WECS with electrical variable transmission (EVT)
In a geared WECS with a high speed generator, the multiple-
stage mechanical gearbox is the most vulnerable part especially
in high capacity offshore wind farms and poses a great difficulty
to maintain. Recently, electrical variable transmission (EVT), an
electrical continuous variable gearbox (CVG), has been proposed
[81] and studied to replace the vulnerable mechanical gearbox in
wind power application. Hence, a novel dual power flow (DPF)
WECS based on the EVT machine was proposed [82–87].
The scheme of the DPF–WECS based on the EVT is shown in
Fig. 20. The EVT enclosed in the dash line is composed of three
parts, namely, a wound stator (#1), a PM outer rotor (#2) and a
wound inner rotor (#3). The wind turbine is connected to the outer
rotor of the EVT machine while the synchronous generator is con-
nected with the inner rotor. The power supplied by the wind tur-
bine can be divided into two parts, namely mechanical power Pd
and electric power Pes. The mechanical power is transmitted
through the coupling of the two rotors with high efficiency, while
the electrical power Pes is transmitted to the inner rotor by the
back-to-back converter with relatively low efficiency.
The available research results show that DPF–WECS based on
EVT has the ability to absorb more wind power than the rated
capacity of the generator, to eliminate fluctuation of wind power
and to ride through the low-voltage fault. Moreover, the system
can produce constant voltage constant-frequency electrical power
with sinusoidal waveform due to the absence of converters after
the synchronous generator.
However, the disadvantages of the novel type WECS are:
(1) The structure of the EVT machine is more complicated than
the conventional generators and then increases the manu-
facturing cost. Besides, the brushes and slip rings, which
need regular maintenance, are used in the inner rotor of
EVT machine.
(2) The rating power of the converter is approximately 70% of
the system power which is higher than that in a DFIG based
WECS.
5.6. Variable speed WECS with constant-frequency double-rotor
generator
In [88], a novel WECS based on the EVT machine was proposed.
However, the extra SG is not needed in this WECS by comparing
with the DPF–WECS in Fig. 20. The EVT machine, named as con-
stant-frequency double-rotor generator (CFDR) in this WECS, can
#1
#2
#3
Pm1 Pd
Pes
Pes
AC Pei
DC
Pbs
Fig. 20. Scheme of the dual power flow WECS based on the EVT.
http://www.paper.edu.cn
343
be seen as a combination of the PM generator and the DFIG system.
The structure of the WECS is shown in Fig. 21, the inner rotor is
connected with the wind turbine directly without gearboxes and
the stator is directly connected with the grid. The partial scale
back-to-back converter is used to connect the inner rotor windings
and the grid which is similar to the DFIG based WECS. By adjusting
the frequency of current of the inner rotor windings, the rotation
speed of inner rotor can be controlled with the variation of the
wind speed to achieve the MPPT. The detailed control strategies
for this WECS are presented in [89].
The CFDR generator owns some advantages of both the DFIG
and PM generator systems, such as high power density, low speed
direct-driven, and partial scale converter.
5.7. Direct grid-connected WECS without converters
Induction generators are widely used due to the advantages of
no requiring frequency control equipment, less initial investment
and maintenance costs. However, induction generators require a
magnetizing current from a power grid, which essentially causes
a decrease in the power factor and efficiency. Then the induction
machine with a permanent magnet excitation is proposed as a per-
manent magnet induction generator (PMIG) shown in Fig. 22,
which can decrease the magnetizing current and improve the
Turbine Double-rotor generator
Stator
Outer rotor
Inner rotor g
r
shaft
PWM converter
Grid
Fig. 21. Scheme of the WECS based on the CFDR generator.
Pm2
SG
DC
AC
中国科技论文在线
344 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
Stator
Primary winding
Outer air-gap Squirrel-cage rotor
Secondary winding
PM
PM rotor
Shaft
Inner air-gap
Fig. 22. Structure of the PMIG.
power factor and efficiency [90]. The stator of PMIG is the same as
that of the usual induction machines equipped with a three-phase
winding. The rotor consists of two parts: an outer squirrel-cage
rotor is linked to the shaft while the inner PM rotor is free to rotate
against the shaft [91]. When the squirrel-cage rotor is driven by the
wind turbine, the PM rotor rotates with this rotor. When the speed
of the squirrel-cage rotor accelerates more than the synchronous
speed, the PM rotor rotates synchronously with the stator’s rotat-
ing field. Then since the slip of the squirrel-cage rotor is a negative
value, the PMIG can feed the electrical power to the power grid
[92].
To overcome the disadvantages in the WECS with PMIG, namely
difficult construction and the need of extra set of bearings, a new
concept direct grid-connected WECS with a slip-synchronous per-
manent magnet generator (SS-PMG) was proposed in [93,94],
which is similar to the WECS with the PMIG.
The SS-PMG concept shown in Fig. 23 consists of two PM
machine units. It differs from the conventional PMIG system where
the two machine units are magnetically separated. The two
machine units are mechanically linked by a common PM rotor.
The one generator, as a slip PM generator, operates on a principle
similar to that of an induction generator, whose short-circuited
rotor is mechanically connected to the wind turbine and runs at
slip speed with respect to the synchronously rotating PM rotor.
The other generator unit is a normal PMSG with its stationary sta-
tor connected to the grid.
Several limiting factors exist in the direct-online systems, such
as the fixed speed disadvantage, the stability of the generator
under grid faults and the ability to control the reactive power.
Therefore, none of this type WECSs have been installed in practice
so far.
sm
Common PM-rotor
Cage Stator
tm
S-PMG
= tm -
slm sm PMSG
Cage Stator
Common PM-rotor
Fig. 23. Scheme of the WECS based on the SS-PMG.
http://www.paper.edu.cn
6. Future trends
Owing to the stochastic nature of wind speed, the wind power
generated by WECSs is fluctuating rapidly. Several serious prob-
lems will be caused by the power fluctuations, such as the grid fre-
quency fluctuation, the active power fluctuation and the voltage
flicker at the buses of the power grid. Then the poor quality and
instability problems in power system will be created [95]. Various
power smoothing methods have been proposed to solve these
problems. Most of the methods are based on the energy storage
system, such as batteries, flywheel [96]. However, the cost of the
energy storage system is high. Hence, the low cost power smooth-
ing methods without energy storage system are the future trends.
Right now, a double objectives control scheme was proposed in
[97] to minimize the torque variations of turbine shaft based on
the frequency separation principle. The active current control
method in [98], the generator torque control strategy in [99] and
individual pitch control in [100] were proposed to smooth the gen-
erator output power oscillations.
Wind turbines are often located in remote areas and easily sub-
jected to extreme environment. These factors not only increase
operation and maintenance (O&M) cost, but also reduce availabil-
ity of wind power due to downtime of WECS [101,102]. In order to
make wind power competitive to conventional power technolo-
gies, a critical issue is focusing on reducing the O&M cost and
improving the availability of WECS. An effective way to achieve
this improvement is to apply fault diagnosis, which is the require-
ment of the fault tolerant control. In the field of fault diagnosis of
WECS, several general methods have been put forward in recent
years [103–106]. With the increasing proportion of wind power
generation in power grid and offshore wind power generation,
fault diagnosis for the WECS will become more and more impor-
tant in the future.
Due to the high cost of unexpected maintenance, the develop-
ment of a reliable WECS is necessary. The power converters used
in the WECSs, especially full-scale converters, have a high probabil-
ity of failures which are associated to semiconductor or control cir-
cuit faults [107]. Along with the growing adoption of modern wind
turbines with full scale converters, fault tolerance of converters is
becoming more and more important in order to increase the sys-
tem reliability. A fault tolerant converter is intended to maintain
its operation after an internal fault occurrence until a maintenance
operation can be scheduled. Various topologies have been pro-
posed to endow a standard three-phase converter with fault toler-
ant capabilities [108,109]. The more effective fault tolerant
topologies and control methods for the converters in WECSs are
worth studying in the future.
Offshore wind power has emerged as one of the most promising
renewable energy resources. The cost of harvesting offshore WECSs
is high compared to that of traditional onshore WECSs. The nacelle-
to-sea cables do also become an issue due to the stiffness and
weight associated with high current, low voltage cables. Then the
transformer has to be placed in the nacelle, which will drive the
nacelle weight further up [110]. Therefore, the transformerless
generator or converter is beneficial in the case of large offshore
WECSs. During recent years, several concepts have been proposed
to increase the output voltage of the WECS without the distribu-
tion transformer [111,112]. More effective concepts may come true
and be applied in the future.
p To increase the current and voltage capability of the wind tur-
=(
2)
s sm bine, the parallel connection of converters is used. Then this corre-
sponds to an increase of number of legs and consequently leads to
the use of multi-phase generators. Multi-phased structures offer
the advantage of modularity, with immediate consequences for
the fabrication process, assembly, transportation and maintenance
中国科技论文在线
M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
[113]. Besides, the fault tolerant capability of multi-phase
machines is better than traditional three-phase machines. When
faults occurs in one or more of the phases in multi-phase machines,
the machines can still continue operation using the remaining
healthy phases without additional hardware [114,115]. Therefore,
the multi-phase generators have great potential to be adopted in
the future WECSs.
7. Conclusions
This paper has reviewed the common types of WECSs, the MPPT
methods for the variable speed WECSs and the prospect of poten-
tial WECS technologies.
The variable speed WECS with DFIG and multiple-stage gearbox
is still the most popular system in the present market. With the
increasing requirement of the system reliability and FRT capability,
the direct-drive system with full-scale converter and PMSG
becomes more and more attractive. However, the size and weight
of PMSG is increasing with the increase in the wind turbine rated
power, thus the so-called Multibrid system with the PMSG and a
single-stage gearbox is presented to reduce the size and weight
of the PMSG. The comparisons of the common WECSs have been
presented by considering the criteria such as the size, weight, cost,
efficiency, and reliability.
The typical four MPPT methods have been summarized in this
paper, while the power feedback control and HCS control are the
most popular ones. Besides, the several improvement measures
based on the two MPPT methods have been introduced. A table with
comparisons of the several MPPT methods has been presented.
The several potential wind energy conversion technologies have
been introduced in this paper. The BDFIGs are studied to replace
the traditional DFIGs to improve the system reliability. Several
novel PM generators are researched to improve the performance
of WECSs. The EVT machine adopted in the dual power flow WECS
can replace the mechanical gearboxes and then improve the sys-
tem reliability. The direct grid-connected WECS with a PMIG can
eliminate the converters which are indispensable in common var-
iable speed WECSs.
Finally, the future trends of WECSs are discussed, such as the
smoothing methods for power fluctuations and effective fault tol-
erant methods for WECSs.
Acknowledgements
This work was supported in part by the Grants (Project
51,137,001, 51320105002) from the National Natural Science foun-
dation of China, a Grant (Project BK2010013) from Innovative
Scholar Ascent Program of Jiangsu Province, China and a grant
(Project 20120092130008) from Ph. D. Programs Foundation of
Ministry of Education of China.
References
[1] Global wind report – the global status of wind power in 2013. Global Wind
Energy Council; April 2014.
[2] Li H, Chen Z. Overview of different wind generator systems and their
comparisons. IET Renew Power Gener 2008;2(2):123–38.
[3] Polinder H, van der Pijl FFA, de Vilder GJ, Tavner PJ. Comparison of direct-
drive and geared generator concepts for wind turbines. IEEE Trans Energy
Convers 2006;21(3):725–32.
[4] BoroumandJazi G, Rismanchi B, Saidur R. Technical characteristic analysis of
wind energy conversion systems for sustainable development. Energy
Convers Manage 2013;69:87–94.
[5] Chen Z, Guerrero JM, Blaabjerg F. Review of the state of the art of power
electronics for wind turbines. IEEE Trans Power Electron 2009;24(8):
1859–75.
[6] Blaabjerg F, Teodorescu R, Liserre M, Timbus AV. Overview of control and grid
synchronization for distributed power generation systems. IEEE Trans Industr
Electron 2006;53(5):1398–409.
http://www.paper.edu.cn
345
[7] Blaabjerg F, Liserre M, Ma K. Power electronics converters for wind turbine
systems. IEEE Trans Industr Electron 2012;48(2):708–19.
[8] Abdullah MA, Yatim AHM, Tan CW, Saidur R. A review of maximum power
point tracking algorithms for wind energy systems. Renew Sustain Energy
Rev 2012;16(5):3220–7.
[9] Slootweg JG, de Vries E. Inside wind turbines-fixed vs. variable speed. Renew
Energy World 2003;6(1):30–40.
[10] Xu L, Cartwright P. Direct active and reactive control of DFIG for wind energy
generation. IEEE Trans Energy Convers 2006;21(3):750–8.
[11] Ibrahim AO, Nguten TH, Lee DC, Kim SC. A fault ride-through technique of
DFIG wind turbine systems using dynamic voltage restorers. IEEE Trans
Energy Convers 2011;26(3):871–82.
[12] Semken RS, Polikarpova M, Roytta P, Alexandrova J, Pyrhonen J, Nerg J, et al.
Direct-drive permanent magnet generators for high-power wind turbines:
benefits and limiting factors. IET Trans Renew Power Gener 2012;6(1):1–8.
[13] Ribrant J, Bertling LM. Survey of failures in wind power systems with focus on
Swedish wind power plants during 1997–2005. IEEE Trans Energy Convers
2007;22(1):167–73.
[14] Siegfriedsen S, Bohmeke G. Multibrid technology—a significant step to multi-
megawatt wind turbines. Wind Energy 1998;1:89–100.
[15] Sinovel. <http://www.sinovel.com/en/procducts.aspx> [accessed May 2014].
[16] Vestas. <http://www.vestas.com/en/wind-power-plants/procurement/
turbine-overview.aspx#/vestas-univers> [accessed May 2014].
[17] GE wind. <http://www.ge-energy.com/wind> [accessed May 2014].
[18] WinwinD. <http://www.winwind.com/> [accessed May 2014].
[19] Gamesa. <http://www.gamesacorp.com/en/products-and-services/wind-
turbines/> [accessed May 2014].
[20] Siemens. <http://www.energy.siemens.com/hq/en/ renewable-energy/wind-
power/platforms/> [accessed May 2014].
[21] Nordex. <http://www.nordex-online.com/en/products-services/wind-
turbines.html> [accessed May 2014].
[22] Dong Fang. <http://dqfd.dfstw.com/cpjs/cpxl.htm> [accessed May 2014].
[23] REpower. <http://www.senvion.com/wind-energy-solutions/wind-turbines/>
[accessed May 2014].
[24] Gold wind. <http://www.goldwindglobal.com/web/solutions.do?action=
products> [accessed May 2014].
[25] Enercon. <http://www.enercon.de/en-en/Windenergieanlagen.htm>
[accessed May 2014].
[26] Multibrid (Now Areva). <http://www.areva.com/EN/global-offer-713/wind-
turbines-for-offshore-wind-farms-renewable-energies-solutions.html> [accessed
May 2014].
[27] Suzlon. <http://www.suzlon.com/> [accessed May 2014].
[28] STX wind. <http://www.stxwind.com/products/models> [accessed May
2014].
[29] McMillan D, Ault GW. Techno-economic comparison of operational aspects
for direct drive and gearbox-driven wind turbines. IEEE Trans Energy Convers
2010;25(1):191–8.
[30] Zhang Y, Chen Z, Cheng M. Proportional resonant individual pitch control for
mitigation of wind turbines loads. IET Renew Power Gener 2013;7(3):
191–200.
[31] Polinder H. Overview of and trends in wind turbine generator systems. In:
IEEE power and energy society general meeting 2011, San Diego, CA, USA. p.
1–8.
[32] Liserre M, Cardenas R, Molinas M, Rodriguez J. Overview of multi-MW
wind turbines and wind parks. IEEE Trans Industr Electron 2011;58(4):
1081–95.
[33] Lin Y, Cai X, Wang N. Rotor current transient analysis of DFIG-based wind
turbines during symmetrical voltage faults. Energy Convers Manage
2013;76:910–7.
[34] Ibrahim AO, Nguyen TH, Lee DC, Kim SC. A fault ride-through technique of
DFIG wind turbine systems using dynamic voltage restorers. IEEE Trans
Energy Convers 2011;26(3):871–82.
[35] Okedu KE, Muyeen SM, Takahashi R, Tamura J. Wind farms fault ride through
using DFIG with new protection scheme. IEEE Trans Sust Energy
2012;3(2):242–54.
[36] Mohammadi J, Afsharnia S, Vaez-Zadeh S. Efficient fault-ride-through control
strategy of DFIG-based wind turbines during the grid faults. Energy Convers
Manage 2014;78:88–95.
[37] Pourmohammad S, Fekih A. Fault-tolerant control of wind turbine systems-a
review. In: IEEE green technologies conference 2011, Baton Rouge, LA, USA. p.
1–6.
[38] Dahbi A, Hachemi M, Nait-Said N, Nait-Said M. Realization and control of a
wind turbine connected to the grid by using PMSG. Energy Convers Manage
2014;84:346–53.
[39] Zhu Y, Cheng M, Hua W, Wang W. A novel maximum power point tracking
control for permanent magnet direct drive wind energy conversion systems.
Energies 2012;5(5):1398–412.
[40] Shirazi M, Viki AH, Babayi O. A comparative study of maximum power
extraction strategies in PMSG wind turbine system. In: IEEE electrical power
& energy conference (EPEC), Montreal, QC, Canada; 2009. p. 1–6.
[41] Musunuri S, Ginn HL. Comprehensive review of wind energy maximum
power extraction algorithms. In: IEEE power and energy society general
meeting 2011, San Diego, CA, USA. p. 1–8.
[42] Aissaoui AG, Tahour A, Essounboucil N, Nollet F, Abid M, Chergui MI. A fuzzy-
PI control to exact an optimal power from wind turbine. Energy Convers
Manage 2013;65:688–96.
中国科技论文在线
346 M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347
[43] Beltran B, Ahmed-Ali T, Benbouzid MEH. Sliding mode power control of
variable-speed energy conversion systems. IEEE Trans Energy Convers
2008;23(2):551–8.
[44] Barakati SM, Kazerani M, Aplevich JD. Maximum power tracking control for a
wind turbine system including a matrix converter. IEEE Trans Energy Convers
2009;24(3):705–13.
[45] Pan C, Juan Y. A novel sensorless MPPT controller for a high-efficiency
microscale wind power generation system. IEEE Trans Energy Convers
2010;25(1):207–16.
[46] Morimoto S, Nakayama H, Sanada M, Takeda Y. Sensorless output
maximization control for variable-speed wind generation system using
IPMSG. IEEE Trans Ind Appl 2005;41(1):60–7.
[47] Datta R, Ranganathan VT. A method of tracking the peak power points for a
variable speed wind energy conversion system. IEEE Trans Energy Convers
2003;18(1):163–8.
[48] Wang Q, Chang L. An intelligent maximum power extraction algorithm for
inverter-based variable speed wind turbine systems. IEEE Trans Power
Electron 2004;19(5):1242–9.
[49] Gonzalez LG, Figueres E, Garcera G, Carranza O. Maximum-power-point
tracking with reduced mechanical stress applied to wind-energy-conversion-
systems. Appl Energy 2010;87(7):2304–12.
[50] Agarwal V, Aggarwal RK, Patidar P, Patki C. A novel scheme for rapid tracking
of maximum power point in wind energy generation systems. IEEE Trans
Energy Convers 2010;25(1):228–36.
[51] Lin W, Hong C. Intelligent approach to maximum power point tracking
control strategy for variable-speed wind turbine generation system. Energy
2010;35(6):2440–7.
[52] Hong C, Chen C, Tu C. Maximum power point tracking-based control
algorithm for PMSG wind generation system without mechanical sensors.
Energy Convers Manage 2013;69:58–67.
[53] Hilloowala RM, Sharaf AM. A rule-based fuzzy logic controller for a PWM
inverter in a stand alone wind energy conversion scheme. IEEE Trans Ind Appl
1996;32(1):57–65.
[54] Simoes MG, Bose BK, Spiegel RJ. Fuzzy logic based intelligent control of a
variable speed cage machine wind generation system. IEEE Trans Power
Electron 1997;12(1):87–95.
[55] Galdi V, Piccolo A, Siano P. Designing an adaptive fuzzy controller for
maximum wind energy extraction. IEEE Trans Energy Convers
2008;23(2):559–69.
[56] Galdi V, Piccolo A, Siano P. Exploiting maximum energy from variable speed
wind power generation systems by using an adaptive Takagi–Sugeno–Kang
fuzzy model. Energy Convers Manage 2009;50(2):413–21.
[57] Ro K, Choi H. Application of neural network controller for maximum power
extraction of a grid-connected wind turbine system. Electr Eng
2005;88(1):45–53.
[58] Li H, Shi KL, McLaren PG. Neural-network-based sensorless maximum wind
energy capture with compensated power coefficient. IEEE Trans Ind Appl
2005;41(6):1548–56.
[59] Meharrar A, Tioursi M, Hatti M, Stambouli AB. A variable speed wind
generator maximum power tracking based on adaptive neurons-fuzzy
inference system. Expert Syst Appl 2011;38(6):7659–64.
[60] Dalala ZM, Zahid ZU, Yu W, Cho Y, Lai J. Design and analysis of an MPPT
technique for small-scale wind energy conversion systems. IEEE Trans Energy
Convers 2013;28(3):756–67.
[61] Basic D, Zhu J, Boardman G. Transient performance study of a brushless
doubly fed twin stator induction generator. IEEE Trans Energy Convers
2003;18(3):400–8.
[62] Protsenko K, Xu D. Modeling and control of brushless doubly-fed induction
generators in wind energy applications. IEEE Trans Power Electron
2008;23(3):1191–7.
[63] Xu L, Guan B, Liu H, Gao L, Tsai K. Design and control of a high-efficiency
doubly-fed brushless machine for wind power generator application. In: IEEE
energy conversion congress and exposition (ECCE); 2010. p. 12–6.
[64] Liu H, Zhang Y, Xu L. Theoretical analysis of doubly excited brushless
reluctance machine used in wind power generation system. Int Conf Sust
Power Gener Supply 2009:1–4.
[65] Shao S, Abdi E, Barati F, McMahon R. Stator-flux-oriented vector control for
brushless doubly fed induction generator. IEEE Trans Industr Electron
2009;56(10):4220–8.
[66] Chau KT, Li W, Lee CHT. Challenges and opportunities of electric machines for
renewable energy. Progr Electrom Res B 2012;42:45–74.
[67] Cheng M, Hua W, Zhang J, Zhao W. Overview of stator-permanent magnet
brushless machines. IEEE Trans Industr Electron 2011;58(11):5087–101.
[68] Zhang J, Chen Z, Cheng M. Design and comparison of a novel stator interior
permanent magnet generator for direct-drive wind turbines. IET Renew
Power Gener 2007;1(4):203–10.
[69] Zhang J, Cheng M, Chen Z. Optimal design of stator interior permanent
magnet machine with minimized cogging torque for wind power application.
Energy Convers Manage 2008;49(8):2100–5.
[70] Fan Y, Chau KT, Cheng M. A new three-phase doubly salient permanent
magnet machine for wind power generation. IEEE Trans Ind Appl
2006;42(1):53–60.
[71] Lin M, Hao L, Li X, Zhao X, Zhu Z. A novel axial field flux-switching permanent
magnet wind power generator. IEEE Trans Magn 2011;47(10):4457–60.
[72] Yan J, Lin H, Huang Y, Liu H, Zhu Z. Magnetic field analysis of a novel flux
switching transverse flux permanent magnet wind generator with 3-D FEM.
http://www.paper.edu.cn
In: IEEE international conference on power electronics and drive systems
(PEDS); 2009. p. 332–5.
[73] Wei J, Zhou B, Cheng F, Zhang Z, Hu C, Wu X. A new three-phase doubly
salient electrical magnet generator for direct-driven wind power generation.
In: IEEE international conference on electrical machines and systems
(ICEMS); 2008. p. 4250–2.
[74] Zhang Z, Yan Y, Tao Y. A new topology of low speed doubly salient brushless
DC generator for wind power generation. IEEE Trans Magn
2012;48(3):1227–33.
[75] Fan Y, Chau KT, Niu S. Development of a new brushless doubly fed doubly
salient machine for wind power generation. IEEE Trans Magn
2006;42(5):3455–7.
[76] Liu X, Lin H, Zhu Z, Yang C, Fang S, Guo J. A novel dual-stator hybrid excited
synchronous wind generator. IEEE Trans Ind Appl 2009;45(3):947–53.
[77] Li X, Chau KT, Cheng M, Hua W. Comparison of magnetic-geared permanent-
magnet machines. Progr Electrom Res 2013;133:177–98.
[78] Jian L, Chau KT, Jiang J. A magnetic-geared outer-rotor permanent-magnet
brushless machine for wind power generation. IEEE Trans Ind Appl
2009;45(3):954–62.
[79] Qu R, Liu Y, Wang J. Review of superconducting generator topologies for
direct-drive wind turbines. IEEE Trans Appl Supercond 2013;23(3):5201108.
[80] Wang Y, Sun J, Zou Z, Wang Z, Chau KT. Design and analysis of a HTS flux-
switching machine for wind energy conversion. IEEE Trans Appl Supercond
2013;23(3):5000904:1–04:4.
[81] Xu L. A new breed of electric machines- basic analysis and applications of
dual mechanical port electric machines. In: International conference on
electrical machines and systems (ICEMS); 2005. p. 24–31.
[82] Cheng M, Hua W. Dual power flow wind power generation system based on
electrical variable transmission. Chinese Patent: 200710019235.X; January
2007.
[83] Sun X, Cheng M, Hua W, Xu L. Optimal design of double-layer permanent
magnet dual mechanical port machine for wind power application. IEEE Trans
Magn 2009;45(10):4613–6.
[84] Sun X, Cheng M. Thermal analysis and cooling system design of dual
mechanical port machine for wind power application. IEEE Trans Industr
Electron 2013;60(5):1724–33.
[85] Zhu Y, Cheng M, Hua W. Sensorless maximum power point tracking in dual
power flow wind energy conversion system. In: IEEE international symposium
on power electronics for distributed generation systems (PEDG); 2012. p. 58–63.
[86] Sun X, Cheng M, Zhu Y, Xu L. Application of electrical variable transmission in
wind power generation system. IEEE Trans Ind Appl 2013;49(3):1299–307.
[87] Zhu Y, Cheng M, Hua W, Zhang B. Sensorless control strategy of electrical
variable transmission machines for wind energy conversion systems. IEEE
Trans Magn 2013;49(7):3383–6.
[88] Zhang J, Cheng M, Zhu Y. A constant-frequency double-rotor generator for
wind power generation. In: IEEE international conference on electrical
machines and systems (ICEMS); 2010. p. 583–7.
[89] Wang Y, Zhang J, Qu X, Cheng M. Control strategy of variable-speed wind
power generation system with constant-frequency double-rotor generator.
In: IEEE international symposium on power electronics for distributed
generation systems (PEDG); 2012. p. 87–92.
[90] Fukami T, Nakagawa K, Kanamaru Y, Miyamoto T. A technique for the steady-
state analysis of a grid-connected permanent-magnet induction generator.
IEEE Trans Energy Convers 2004;19(2):318–24.
[91] Tsuda T, Fukami T, Kanamaru Y, Miyamoto T. Effects of the built-in permanent
magnet rotor on the equivalent circuit parameters of a permanent magnet
induction generator. IEEE Trans Energy Convers 2007;22(3):798–9.
[92] Vermaak R, Potgieter JHJ, Kamper MJ. Grid-connected VSC-HVDC wind farm
system and control using permanent induction generators. In: IEEE
international conference on power electronics and drive systems (PEDS);
2009. p. 897–902.
[93] Potgieter JHJ, Kamper MJ. Design of new concept permanent magnet
induction wind generator. In: IEEE international conference on energy
conversion congress and exposition (ECCE); 2010. p. 2403–8.
[94] Potgieter JHJ, Kamper MJ. Design of new concept direct grid-connected slip-
synchronous permanent-magnet wind generator. IEEE Trans Ind Appl
2012;48(3):913–22.
[95] Howlader AM, Urasaki N, Yona A, Senjyu T, Saber AY. A review of output
power smoothing methods for wind energy conversion systems. Renew
Sustain Energy Rev 2013;26:135–46.
[96] Jiang Q, Hong H. Wavelet-based capacity configuration and coordinated
control of hybrid energy storage system for smoothing out wind power
fluctuations. IEEE Trans Power Syst 2013;28(2):1363–72.
[97] Xu F, Cheng M, Zhang J, Wang Z. Double objectives control for variable-speed
wind energy conversion system with permanent magnet synchronous
generator. In: IEEE international conference on electrical machines and
systems (ICEMS); 2010. p. 493–8.
[98] Zhang Y, Hu W, Chen Z, Cheng M. Mitigation of wind power fluctuation by
active current control of variable speed wind turbines. Int J Smart Grid Clean
Energy 2013;2(2):252–7.
[99] Zhang Y, Hu W, Chen Z, Cheng M, Hu Y. Flicker mitigation strategy for a
doubly fed induction generator by torque control. IET Renew Power Gener
2014;8(2):91–9.
[100] Zhang Y, Chen Z, Hu W, Cheng M. Flicker mitigation by individual pitch
control of variable speed wind turbines with DFIG. IEEE Trans Energy Convers
2014;29(1):20–8.
中国科技论文在线 http://www.paper.edu.cn

M. Cheng, Y. Zhu / Energy Conversion and Management 88 (2014) 332–347 347

[101] Hameed Z, Hong YS, Cho YM, Ahn SH, Song CK. Condition monitoring and [108] Errabelli RR, Mutschler P. Fault-tolerant voltage source inverter for
fault detection of wind turbines and related algorithms: a review. Renew permanent magnet drives. IEEE Trans Power Electron 2012;27(2):500–8.
Sustain Energy Rev 2009;13(1):1–39. [109] Freire NNA, Cardoso AJM. A fault-tolerant direct controlled PMSG drive wind
[102] Márquez FPG, Tobias AM, Pérez JMP, Papaelias M. Condition monitoring energy conversion systems. IEEE Trans Industr Electron 2014;61(2):821–34.
of wind turbines: techniques and methods. Renew Energy 2012;46: [110] Ng C, Parker M, Ran L, Tavner P, Bumby J, Spooner E. A multilevel modular
169–78. converter for a large, light weight wind turbine generator. IEEE Trans Power
[103] Freire NMA, Estima JO, Marques Cardoso AJ. Open-circuit fault diagnosis in Electron 2008;23(3):1062–74.
PMSG drives for wind turbine applications. IEEE Trans Industr Electron [111] Parker M, Ng C, Ran L, Tavner P, Spooner E. Power control of direct drive wind
2013;60(9):3957–67. turbine with simplified conversion stage transformerless grid interface. In:
[104] Xiang G, Wei Q. Bearing fault diagnosis for direct-drive wind turbines via Universities power engineering conference (UPEC); 2006. p. 65–8.
current demodulated signals. IEEE Trans Industr Electron [112] Yuan X, Li Y, Chai J, Ma M. A modular direct-drive permanent magnet wind
2013;60(8):3419–28. generator system eliminating the grid-side transformer. In: Power electronics
[105] Gritli Y, Zarri L, Rossi C, Filippetti F. Advanced diagnosis of electrical faults in and applications (EPE); 2009. 1–7.
wound-rotor induction machines. IEEE Trans Industr Electron [113] Vizireanu D, Kestelyn X, Brisset S, Brochet P, Milet Y, Laloy D. Polyphased
2013;60(9):4012–24. modular direct-drive wind turbine generator. In: European conference on
[106] Hang J, Zhang J, Cheng M. Fault diagnosis of wind turbine based on multi- power electronics and applications; 2005. p. 1–9.
sensors information fusion technology. IET Renew Power Gener [114] Levi E. Multiphase electric machines for variable-speed applications. IEEE
2014;8(3):289–98. Trans Industr Electron 2008;55(5):1893–909.
[107] Yang S, Bryant A, Mawby P, Xiang D, Ran L, Tavner P. An industry-based [115] Li F, Hua W, Cheng M. Analysis of fault tolerant control for a nine-phase flux-
survey of reliability in power electronic converters. IEEE Trans Ind Appl switching permanent magnet machine. IEEE Trans Magn [in press], http://
2011;47(3):1441–51. dx.doi.org/10.1109/TMAG.2014.2325060.