High-Power Wind Energy Conversion Systems: Emerging Technologies and State-of-the-Art

CONTRIBUTED
P A P E R
High-Power Wind Energy
Conversion Systems:
State-of-the-Art and
Emerging Technologies
Wind energy installed capacity increased exponentially over the
past three decades,
and has become a real alternative to increase renewable energy
penetration
into the energy mix.
By Venkata Yaramasu, Member IEEE, Bin Wu, Fellow IEEE,
Paresh C. Sen, Life Fellow IEEE,
Samir Kouro, Member IEEE, and Mehdi Narimani, Member
IEEE
ABSTRACT | This paper presents a comprehensive study on the
state-of-the-art and emerging wind energy technologies from
the electrical engineering perspective. In an attempt to de-
crease cost of energy, increase the wind energy conversion
efficiency, reliability, power density, and comply with the strin-
gent grid codes, the electric generators and power electronic

converters have emerged in a rigorous manner. From the mar-
ket based survey, the most successful generator-converter
configurations are addressed along with few promising topol-
ogies available in the literature. The back-to-back connected
converters, passive generator-side converters, converters for
multiphase generators, and converters without intermediate
dc-link are investigated for high-power wind energy conver-
sion systems (WECS), and presented in low and medium voltage
category. The onshore and offshore wind farm configurations
are analyzed with respect to the series/parallel connection of
wind turbine ac/dc output terminals, and high voltage ac/dc
transmission. The fault-ride through compliance methods used
in the induction and synchronous generator based WECS are
also discussed. The past, present and future trends in megawatt
WECS are reviewed in terms of mechanical and electrical tech-
nologies, integration to power systems, and control theory. The
important survey results, and technical merits and demerits of
various WECS electrical systems are summarized by tables. The

list of current and future wind turbines are also provided along
with technical details.
KEYWORDS | ac-ac; ac-dc; dc-ac; dc-dc power conversion;
doubly fed induction generator (DFIG); fault-ride through
(FRT);
grid codes; low voltage (LV); medium voltage (MV); multilevel
converters; permanent magnet synchronous generator (PMSG);
power electronics; squirrel cage induction generator (SCIG);
wind energy conversion systems (WECS); wind farms; wound
rotor induction generator (WRIG); wound rotor synchronous
generator (WRSG)
I . I N T R O D U C T I O N
Due to depleting fossil fuels and environmental concerns
about global warming, renewable energy sources have
emerged as a new paradigm to fulfill the energy needs of
our society. In recent years, electricity production from the
hydro, solar, wind, geothermal, tidal, wave and biomass
energy sources has come under increasing attention [1],
[2]. By 2012, the power production from renewable energy
sources worldwide exceeded 1470 gigawatt (GW) repre-

senting approximately 19% of global energy consumption
[3]–[5].
Manuscript received May 7, 2014; revised September 10, 2014;
accepted November 26,
2014. Date of publication May 18, 2015; date of current version
May 22, 2015. This
work was supported by the Natural Sciences and Engineering
Research Council
of Canada (NSERC) through Wind Energy Strategic Network
(WESNet) Project 3.1,
by Fondecyt 1131041, and by SERC Chile (FONDAP/15110019)
and AC3E (FB0008)
of Conicyt.
V. Yaramasu, B. Wu, and M. Narimani are with the Department
of Electrical and
Computer Engineering, Ryerson University, Toronto, Ontario,
Canada, M5B 2K3
(e-mail: [email protected]; [email protected]; [email protected]).
P. C. Sen is with the Department of Electrical and Computer
Engineering, Queen’s
University, Kingston, Ontario, Canada, K7L 3N6 (e-mail:
[email protected]).
S. Kouro is with the Electronics Engineering Department,
Universidad Técnica

Federico Santa Marı́a, Valparaı́so, Chile, 2390123 (e-mail:
[email protected]).
Digital Object Identifier: 10.1109/JPROC.2014.2378692
0018-9219 � 2015 IEEE. Personal use is permitted, but
republication/redistribution requires IEEE permission.
See
http://www.ieee.org/publications_standards/publications/rights/i
ndex.html for more information.
740 Proceedings of the IEEE | Vol. 103, No. 5, May 2015
Among all the renewable energy sources, wind energy is
increasingly becoming mainstream and competitive with
conventional sources of energy [6], [7]. The cumulative
installed wind power capacity increased exponentially from
6100 megawatt (MW) in 1996 to 282.6 GW by 2012. It is
anticipated that, following the current trend, the cumula-
tive wind capacity would reach 760 GW by 2020. In 2012,
approximately 45 GWs of new wind power was added
which represents investments of about euros 56 billion.
The wind energy industry is also providing many direct or
indirect job opportunities leading to a significant stimulus
to economic development. This industry demonstrated an

excellent growth rate of more than 19%, and represents
1.9% of the world’s net electricity production. Currently
83 countries are using wind energy on a commercial basis
to generate electricity [3]. Approximately 24 countries
have more than 1 GW cumulative installed wind power
capacity, including 16 in Europe, 4 in the Asia-Pacific area
(China, India, Japan, Australia), 3 in North America
(Canada, Mexico, US) and 1 in Latin America (Brazil) [4].
The cost reduction, government incentive programs, and
technological advancements are some of the key reasons
behind this impressive growth rate.
The energy from the wind has been harnessed by
mankind for millennia to carry ships across oceans and
later to pump water and grind grain. The conversion of
wind kinetic energy to electrical energy has started during
1887 with an automated wind turbine equipped with a
12-kW dc generator. To generate electricity from wind
turbines more efficiently and reliably, and to compete
against fossil fuel based power plants, many improvements

have been made in the design of wind turbine mechanical
and electrical components. The wind turbine technology
has reached a sufficient maturity level by 1980s leading to
the commissioning of first 50-kW utility-scale wind
turbines [8], [9].
According to the aerodynamic properties, the power
output of a wind turbine is proportional to the square of a
rotor diameter and a cubic of wind speed [6], [7]. The large
turbines can capture higher wind power with lower instal-
lation and maintenance costs compared to the group of
small turbines. Owing to this fact, the size of commercial
wind turbines has exponentially increased over the past
30 years as demonstrated in Fig. 1. The turbine size has
increased from 50 kW in 1980 to 7.5 MW in 2010 [10],
[11]. The wind turbine rotor diameter also increased from
15 m in 1980 to 126 m in 2010. The largest wind turbine
reported by 2014 is 8 MW with a diameter of 164 m (Vestas
V164), and it is currently in testing stage at Osterild,
Denmark [12], [13]. The 10 MW wind turbines have been

announced by the Clipper, Sway Turbine AS, and Windtec-
AMSC, and the GE Energy has ambitious plans to develop
15 MW turbines (refer to Table 11 for details). The offshore
technology is another important driving force behind this
amazing growth size in wind turbines [14]. The market
survey indicates that the rotor diameter and power ratings
of offshore (located in the sea) wind turbines are higher
compared to the onshore (located on the land) wind
turbines. In 2013, the average size of onshore and offshore
wind turbine are reported as 1.926 and 3.613 MW,
respectively [5]. The market trend also indicates that 10–
20 MW turbines will be operational in near future with
rotor diameters exceeding 150 m, which is approximately
twice the length of a Boeing 747 airplane.
The wind energy industry has gone through much
technological advancement in terms of aerodynamic
design, mechanical systems, electric generators, power
electronic converters, integration to power systems and
control theory. From the electrical engineering perspec-
tive, the electric generators and power electronic con-

verters are two major components in the operation of wind
energy conversion systems (WECS). Since the beginning
Fig. 1. Evolution in the size of commercial wind turbines.
Yaramasu et al.: High-Power Wind Energy Conversion Systems
Vol. 103, No. 5, May 2015 | Proceedings of the IEEE 741
of grid-connected operation in 1980s, various combina-
tions of electric generators and power electronic con-
verters have been developed in commercial wind turbines
to achieve fixed-speed, semi-variable-speed and full-
variable-speed operation [15], [16].
A group of wind turbines are often placed over an ex-
tended area to form the wind farm, and they are connected
collectively to a national electric grid. The wind farms can
be located on the land (onshore) or in the sea (offshore)
[14]. Traditionally onshore wind farms have been devel-
oped to take the advantage of easy access, lower initial and
maintenance costs and better proximity to the transmis-
sion lines [17]. The initial and maintenance costs of off-

shore wind farms are higher compared to the onshore
farms for same power levels because stronger foundations
are needed and the connection to the onshore grid is per-
formed by submarine cables. To connect the onshore and
offshore wind farms to electric power system, various
series/parallel and ac/dc configurations, and high voltage
ac and dc (HVAC and HVDC) transmission systems have
been developed by the wind turbine manufacturers, off-
shore operators and academic researchers [18]–[22].
Due to the rapid integration of wind power into the
electric grid, many concerns emerged related to the stable,
secure and efficient operation of the existing electric power
system. The grid codes have been updated and enforced in
many countries on the grid-connection of large-scale wind
turbines and wind farms [23]–[27]. To increase wind
energy conversion efficiency, reduce mechanical stress on
wind turbines, improve grid power quality and to meet the
grid codes, the high-power wind turbine technology has
upgraded from fixed-speed to full-variable-speed operation.

As of 2013, more than 90% of global wind turbine manu-
facturers have announced their future projects based on
semi-variable-speed and full-variable-speed technology
[12], [13], [28], [29].
Reflecting the research and development activities by
both industry and academia, many books or book chapters
[30]–[44]; reports [45]–[48]; and excellent survey papers
on wind energy conversion systems [6]–[11], [14]–[18],
[49]–[69] have been published over the past few decades.
These works have discussed some specific aspects of the
wind energy, electric generators, power converters, grid
integration and grid code compliance. The complete list of
electrical technologies applicable for the high-power wind
turbines have not been addressed yet. Moreover, in recent
years, researchers all around the world contributed through
large number of research articles to enable the wind energy
further. This paper is aimed to extensively review the state-
of-the-art and emerging MW wind generator-converter
configurations, wind farm configurations and grid code

compliance methods with respect to the previously pub-
lished research articles, survey papers, books and industrial
repots. To aid our discussion, in this paper, the generator-
converter configurations are classified into four categories
and compared based on component count, modularity, re-
liability, converter/control complexity, device voltage
stress, operation voltage, and achievable power levels.
The organization of paper is shown in Fig. 2. The first
three sections are introductory in nature, while Sections IV
to X contain in depth technical details, and the last two
sections provide concluding remarks.
I I . O V E R V I E W O F H I G H P O W E R W E C S
In this section, an overview of high power WECs is pre-
sented. The major components of grid-connected WECS,
classification of WECS operating voltages and grid code
requirements are presented in detail. The information pro-
vided in this section will be used in the subsequent sections.
A. Major Components of Grid-Connected WECS
The basic configuration of grid-connected MW-WECS

is depicted in Fig. 3. The WECS is composed of several
components that convert wind kinetic-energy into electric-
energy in a controlled, reliable and efficient manner. The
major components of a WECS can be broadly classified as
mechanical, electrical and control systems. The mechan-
ical components include tower, nacelle, rotor blades, rotor
hub, gearbox, pitch drives, yaw drives, wind speed sensors,
drive-train, and mechanical brakes [42]. The electrical
components include electric generator, possible power
electronic converter along with generator- and grid-side
harmonic filters, step-up transformer and three-phase grid
or collection-point [30]. The control related components
are used with both the mechanical and electrical energy
conversion systems [7], [37], [70]. The most visible parts in
the large wind turbines are tower, nacelle and rotor blades,
and rest of the components are housed inside the wind
turbine.
1) Mechanical Components: The wind kinetic-energy is
first converted to mechanical-energy with the help of
airfoil-shaped rotor blades. For the current MW wind

turbines, three-bladed design is most efficient and common
[71], [72]. The tower, nacelle and rotor hubs provide me-
chanical support to the rotor blades. The kinetic to mecha-
nical energy conversion efficiency depends on many factors
such as the shape of rotor blades, angle of blades, wind
speed velocity, air density, etc. [36]. The wind velocity and
direction are measured with the help of sensors, and a yaw
drive is used to move the rotor blades along with nacelle
towards the wind to extract the maximum possible energy.
As per the aerodynamic properties, any particular turbine
generates maximum electricity at or above the rated wind
speed only. When the wind speed is more than the rated
value, the angle of blades is changed such that the electric
power output is limited to the rated value [16], [73], [74].
The MW wind turbines usually run at very low speed
(typically 6–20 rpm) and high torque.
To couple the low-speed, high-torque wind turbine
shaft with the high-speed, low-torque generator shaft, a

multistage gearbox is used. A three-stage gearbox
containing one planetary stage and two helical stages is
usually employed in high-speed generators. For a three-
stage gear box, the gear ratios usually vary between 60 and
120. The gearbox presents with several serious issues such
as high initial cost, high audible noise, extensive wear and
Fig. 2. Organization of content in this paper.
Fig. 3. Basic configuration of a grid-connected megawatt wind
turbine.
tear, reduced life span, reduced efficiency and need for
regular maintenance [30], [50]. By matching the generator
speed with the wind turbine speed, the need for the
gearbox can be eliminated. The omission of gearbox (often
referred to as gearless or direct-drive technology) helps to

overcome the aforementioned problems, especially in off-
shore wind turbines [49], [75]. This concept was first in-
troduced in 1992 by a German manufacturer, Enercon
through E-40/500 kW wind turbine model. In recent years,
many turbine manufacturers (Avantis, GE Energy,
Goldwind, Vensys to name a few) have used direct-drive
technology in their commercial products.
On the other hand, the direct-drive operation leads to
some drawbacks in the design of generator such as large
diameter and more weight [66]. To make a compromise
between the high-speed and low-speed operation, a
medium-speed generator can be used in conjunction
with the single or two-stage gear box. The single-stage
gearbox with a gear ratio of 10 was first introduced by
Multibrid (now Areva Wind) in their M5000 wind turbine
models [76]. Few other turbine manufacturers such as
MingYang and WinWinD developed two-stage gearboxes
with a gear ratio of 20–30 [13]. The list of commercial
turbines along with the gearbox details are given in

Appendix. The commercial wind turbines have many al-
ternative designs for the drive-trains and they will be dis-
cussed in Section III. The mechanical brakes are mounted
directly on the generator drive-train (high-speed shaft)
to stop the wind turbine during fault conditions or high
wind gusts.
2) Electrical Components: An electric generator is used to
convert rotational mechanical-energy into electric-energy.
Over the past 30 years, many generators such as the
squirrel-cage induction generator (SCIG), wound rotor
induction generator (WRIG), doubly-fed induction gener-
ator (DFIG), permanent magnet synchronous generator
(PMSG) and wound rotor synchronous generator (WRSG)
have been developed for wind turbines [46], [47], [63]–
[65]. The first generation of wind turbines were based on
SCIG only, but the present generation turbines incorpo-
rates both induction and synchronous generators. The in-
duction generators (IGs) usually operate at high rotational
speeds, while the synchronous generators (SGs) can ope-
rate at low, medium, or high speeds [68], [77], [78]. In

order to achieve lower operational speed, the generator
needs to be equipped with a large number of poles, which
is a feasible solution with SG’s. In accommodating the
large number of poles, the stator radius becomes 6 times
larger and 4.5 times heavier compared to the three-stage
gearbox based induction generators [50], [54].
The generator output voltage and frequency change
with respect to the wind speed. The generator can be di-
rectly coupled to the grid or it can be interfaced through a
power electronic converter. By arranging the power switch-
ing devices in different ways, possibly with the dc-link
elements such as capacitors or inductors, numerous power
converter topologies can be derived. As shall be detailed in
Section III, these power electronic converters can be com-
bined with the electric generators to form a wide variety of
WECS configurations. Again by connecting the wind tur-
bines in different manner, various wind farm configura-
tions can be obtained. The switching harmonics are
inevitable when using power converters, and to solve this

issue, harmonic filters are used in generator- and grid-side
converters [57]. The harmonic filter on the generator-side
helps to reduce harmonic distortion of the generator cur-
rents and voltages. This leads to a reduction in harmonic
losses incurred in the generator’s magnetic core and wind-
ing. The harmonic filter in the grid-side converter helps to
meet strict harmonic requirements specified by the grid
codes [79]–[81]. The output of the grid-side harmonic filter
is connected to a three-phase grid (or collection-point)
through a step-up transformer, electric switch gear and a
circuit breaker. By operating the power electronic con-
verter at collection-point voltage level, the need for the
step-up transformer can be avoided.
3) Control System: The wind turbine system also consists
of several slave control systems for the mechanical/
electrical components and a master control system (not
shown in Fig. 3) to achieve desired dynamic and steady-
state performance for the WECS. The controller usually
monitors various variables such as wind speed velocity,
wind direction, generator voltages/currents, filter/dc-link

voltages if any, grid voltages and currents, and adjusts the
system operating states or variables at the reference value
or in the set boundaries [7], [70]. For example, when the
wind speed is more than the rated value, the master con-
trol system initiates passive stall, active stall or pitch con-
trol systems to respond and change the angle of blades
such that the turbine output power can never exceed the
rated value [16], [73], [74]. The control systems have been
with limited functions in the first generation of wind
turbines, and now they perform large number of functions
with respect to the turbine, generator and power converter
operation, grid integration, protection standards, and wind
farm operation, to name a few [32], [37], [82]. The control
systems are usually implemented using a computer, micro-
controller, digital signal processor (DSP) or field program-
mable gate array (FPGA) [83], [84]. With the modern
control platforms, the control actions can be taken very
fast (in less than 100 microseconds) and repeatedly.
B. WECS Operating Voltages

The definition of WECS operating voltages in the North
American and European market is summarized in Table 1
[45]. These operating voltages are further classified accord-
ing to low voltage (LV) and medium voltage (MV) opera-
tion. The LV class includes voltages below 1000 V, where as
voltages in the range of 1–34.5 kV belong to the MV class.
The most standard low voltages used for electric generators
and power converters in the North American and European
market are 575 and 690 V, respectively. The current MV
generators and power converters are in the range of 3–4 kV.
Due to the participation of European manufacturers in
North America and vice versa, these regional classifications
are becoming less important. The reason is that the com-
mercial wind turbines can be connected to the collection
points or transmission lines through step-up transformers
irrespective of the regional voltage classes. The most com-

mon collection point voltages are 34.5 and 33 kV in North
America and Europe, respectively.
C. Grid Code Requirements
The steady growth in the power levels of wind turbines
and wind farms have led to significant penetration of wind
energy systems in the existing electric power system. To
ensure the grid stability and consumer power quality, many
specific technical requirements often called as ‘‘grid codes’’
have been developed and regularly updated [23], [24]. The
main elements in grid codes include active power control so
as to adjust the grid frequency, reactive power control to
regulate the grid voltage, grid power quality, flickers, har-
monic oscillations, fault ride-through (FRT) operation, and
system protection. The correct interpretation of these
codes is crucial for wind turbine manufacturers as well as
utility operators.
1) FRT: The grid disturbances might lead to disconnec-
tion of large-scale wind power generation units. The sud-
den disconnection of generation units stimulates instability
of the utility network. The grid codes have dictated some

special requirements such as FRT operation to overcome
the aforementioned scenario. The FRT requirement is a
broad category covering zero voltage ride-through (ZVRT),
low-voltage ride-through (LVRT) and high-voltage ride-
through (HVRT). The ZVRT and LVRT requirements are
essentially same: during grid faults, the grid voltage be-
comes zero in ZVRT profile, while in LVRT profile the grid
voltage becomes 15%–25% of its nominal value [24].
Among all the grid codes, the FRT is a major concern for the
wind turbine and power converter manufacturers. A
detailed discussion on various methodologies to comply
with the FRT requirements will be discussed in Section X.
The transmission and distribution system operators
(TSOs and DSOs) of diverse countries issued different FRT
profiles [23]–[26]. Among all the FRT profiles, the German
Transmission and Distribution Utility (E.ON) regulation
introduced in early 2003 is likely to set the standard [27].
The ZVRT and HVRT profiles according to E.ON regulation
are shown in Fig. 4. They specify that the wind turbines

must ‘‘ride-through’’ instead of ‘‘trip off’’ during transmis-
sion faults. According to this code, the FRT function should
start when the grid voltage falls below 90% of its nominal
value. The wind turbine must be connected to electric
network if the grid voltage profile is above the ZVRT limit
line specified by the utility operator. It is allowed to dis-
connect from grid if the magnitude of grid voltage falls
below the ZVRT limit line. Similar interpretation can be
applied to the HVRT function. Recently considerable re-
search has been carried out addressing this issue [85]–[91].
2) Reactive Power Generation: Apart from the FRT ope-
ration, another important requirement for WECS is that it
should perform ‘‘reactive power control’’ similar to the
Table 1 Regional Classification of Low and Medium Voltages
[Source: NREL (2012)]
Fig. 4. Voltage ride-through requirements according to the
E.ON regulation.

conventional power plant [16]. The reactive power control
helps to compensate the transmission equipment such as
cables and transformers in addition to maintaining voltage
stability. Thus it became an important grid code regulation
to maintain reliable and efficient transmission and distri-
bution grids.
Many reactive power profiles are defined by diverse
TSOs similar to the FRT profiles [23]–[26]. As an example,
according to Danish grid code, when the WECS delivers
rated (1.0 p.u.) active power, it should be able to supply
�0.33 p.u. reactive power to support grid voltage. This
case implies that the WECS should be able to adjust
the grid power factor anywhere from 0.95 lagging to
0.95 leading when it delivers rated active power. Though
this is a stringent grid code, it can be easily fulfilled in
variable-speed WECS by properly designing the power
converter and its control system [92].
I I I . C O M M E R C I A L C O N F I G U R A T I O N S
O F M W - W E C S
The major electrical components in WECS are the gene-

rator and power electronic converter. As shown in Fig. 5,
using different designs and combinations with these two
components, a wide variety of WECS configurations can be
achieved such as [15], [48], [51], [60], [65], [67]:
• Type 1: Fixed-speed (�1%) WECS with SCIG,
• Type 2: Semi-variable speed (�10%) WECS with
WRIG,
• Type 3: Semi-variable speed (�30%) WECS with
DFIG,
• Type 4: Full-variable speed (0–100%) WECS with
SCIG, PMSG or WRSG, and
• Type 5: Full-variable speed (0–100%) WECS with
WRSG.
The above five configurations have been analyzed, do-
cumented and commercialized extensively over the past
three decades.
A. Type 1 WECS Configuration
A fixed-speed SCIG-based WECS without power con-
verter interface (Type 1 turbine) is illustrated in Fig. 5(a),
where the generator is connected to the grid through a soft
starter and step-up transformer [49], [66], [93]. This is the

oldest and very first technology (‘‘Danish’’ concept) devel-
oped for the wind turbines. In high-power WECS, the
SCIG contains 4 or 6 poles for 50- or 60-Hz operation,
respectively. The generator speed varies within 1% around
the corresponding synchronous speed at different wind
speeds, and thus this configuration is called fixed-speed
WECS. A gearbox is normally required to match the speed
difference between the turbine and generator. After the
start-up procedure, the soft-starter is bypassed by a switch,
and the system essentially works without any power con-
verter. The SCIG draws reactive power from the grid and
to compensate for this, three-phase capacitor banks are
usually employed [30], [94].
This configuration features simplicity, low initial
costs, and reliable operation. The major drawbacks include:
(i) lower wind energy conversion efficiency; (ii) changes in
the wind speed are reflected to the grid; and (iii) the grid
Fig. 5. State-of-the-art electric generator and power converter
configurations for the commercial WECS.

faults cause severe stress on the mechanical components of
the wind turbine [95]. The fixed-speed wind turbines are
equipped with additional hardware, such as STATCOM, to
comply with the grid codes [96], [97]. Despite its draw-
backs, this configuration has been accepted by the wind
industry and commercial solutions are available in MW
range such as: (i) Vestas V82, 1.65 MW; and (ii) Siemens
SWT 2.3–101, 2.3 MW. It should be noted that the fixed-
speed turbines were popular until a decade back and this
technology is slowly becoming seldom due to its inherent
disadvantages. The fixed-speed turbines which have been
installed already are still in operation to generate the
electricity.
B. Type 2 WECS Configuration
The variable-speed operation of the wind turbine in-
creases the energy conversion efficiency, and reduces
mechanical stress caused by wind gusts, reduces the wear-
and-tear of gearbox and bearings, reduces the maintenance

requirements, and thus increases the life cycle. The semi
variable-speed WECS using WRIG and partial rated (10%)
power converter is shown in Fig. 5(b) (Type 2 turbine).
The change in the rotor resistance affects the torque/
speed characteristic of the generator, enabling variable-
speed operation of the turbine, and this configuration is
often called Optislip control [98]. The rotor resistance is
normally made adjustable by a power converter composed
of a diode-rectifier and chopper [30]. The speed adjust-
ment range is typically limited to about �10% of its rated
speed. With variable-speed operation, the system can cap-
ture more power from the wind, but also has energy losses
in the rotor resistance. This configuration also requires a
gearbox, soft starter, and reactive power compensation.
The WRIG with variable rotor resistance has been on
the market since the mid 1990’s with a power rating up to a
couple of megawatts. A few examples of commercial solu-
tions are: (i) Vestas V66-2.0 MW; and (ii) Suzlon Energy
S88-2.1 MW. This configuration is also becoming less

important among the wind turbine manufacturers due to
limited speed range and low energy conversion efficiency.
C. Type 3 WECS Configuration
Another semi variable-speed WECS using DFIG is shown
in Fig. 5(c) (Type 3 turbine) [99]–[102]. As the name im-
plies, the power from the generator is fed to the grid through
both stator and rotor windings. A partial rated (30%) power
converter is employed in the rotor circuit to process the slip
power, which is approximately 30% of the rated generator
power. Similar to those in Type 1 and 2 turbines, this con-
figuration also uses the gearbox, but there is no need for a soft
starter and reactive power compensation [103].
The use of the power converters allows bidirectional
power flow in the rotor circuit and increases the speed
range of the generator. This system features improved
overall power conversion efficiency by performing maxi-
mum power point tracking (MPPT) [104], [105], extended
speed range (�30%), enhanced dynamic performance and
robustness against power system disturbances compared to
the Type 1 and 2 turbines [106]–[108]. These features have

made the DFIG WECS one of the dominating technologies
in today’s wind industry with a market share of approx-
imately 50% [54], [58].
The FRT capability is limited due to the partial scale
power converter. The gearbox increases overall turbine
cost, weight and as well demands regular maintenance.
Moreover, the power converter is connected to the rotor
windings through slip rings and brushes. The average life
time of brushes is 6–12 months only, and thus regular
maintenance is essential in these turbines. These major
drawbacks impeded these turbines being applied in off-
shore wind farms where maintenance cost is quite expen-
sive. A few high power DFIG turbines are: (i) Repower 6M,
6.0 MW; (ii) Bard 5.0, 5 MW; and (iii) Acconica AW-100/
3000, 3 MW.
D. Type 4 WECS Configuration
The performance of WECS can be greatly enhanced
with the use of full-scale (100%) power converters as
shown in Fig. 5(d) (Type 4 turbine) [109]–[113]. The

PMSG, WRSG, and SCIG have all found applications in this
type of configuration with a power rating of up to several
megawatts. Since the power converters must be rated same
as generator capacity, the size, cost and complexity of
overall system increases. Moreover the losses in power
converter are higher leading to lower efficiency. However,
with the full-scale power converter, the generator is fully
decoupled from the grid, and can operate at full speed range
(0 to 100%).
The power converters also enable the system to per-
form reactive power compensation and smooth grid con-
nection. The wind energy conversion efficiency is highest
in these turbines compared to other types of turbines
[114]–[116]. The best FRT compliance can also be achieved
without any external hardware. Though the cost of power
converter is high, it only a small fraction (approximately
7%–12%) of total wind turbine cost [53]. The need for the
gearbox can be eliminated by using a high-pole number
PMSG/WRSG. This configuration is more robust against
power system faults compared to the Type 1, 2, and 3

turbines [117]. The typical commercial turbines include:
(i) Enercon E126, 7.5 MW; (ii) Multibrid M5000, 5 MW;
and (iii) Vestas V-112, 3 MW.
The distributed drive-train concept is used in recent
megawatt Type 4 wind turbines. Though SCIG and WRSG
can be used in this concept, the PMSG is most suitable
because it eliminates the need for slip rings/brushes and
there by gives simple design [68]. The gearbox drives
multiple generators at higher speeds. Due to the distrib-
uted drive-train and multiple generators, a higher power
density can be achieved [118]. One of the commercial ap-
plications is Clipper Liberty which uses a quantum drive-
train, 4 generators and 4 converters as shown in Fig. 6 [46].
The higher torque is distributed among the four drive trains.
The power rating of the converters is one-fourth of the
system rating. This configuration also offers effective fault-

tolerant operation. When one converter fails, the other three
converters can still deliver the power to the grid [30]. To
minimize the circulating currents, multi-winding trans-
former is used on the grid-side. The main disadvantage with
this configuration is complicated drive-train.
E. Type 5 WECS Configuration
The Type 5 wind turbine with direct grid-connected
WRSG with speed/torque converter is shown in Fig. 5(e).
This is rather an old concept for wind turbines where the
variable speed operation is achieved by mechanical con-
verter rather than the electrical converter [46], [65]. The
torque/speed converter, also known as variable ratio trans-
mission (VRT) converts variable speed of wind turbine to
constant speed. The generator operates at a fixed-speed
and it is directly connected to the grid through a synchro-
nizing circuit breaker.
The overall system cost and space becomes lower than
Type 4 turbine as no power electronic converter is needed.
The generator can be directly connected to MV collection

point without any step-up transformer as there is no re-
striction imposed by the power electronic converter unlike
in Type 4 turbine. Despite the advantages of this configu-
ration, it is rarely used in the wind energy industry due to
the limited knowledge, and issues related to the mechan-
ical converter. The commercial solutions using this
technology are: (i) DeWind D8.2, 2.2 MW, 4.16/13.8 kV,
(ii) AMSC-Windtec SuperGear (SG), 2.0 MW, 11 kV, and
(iii) Wikov W2000, 2.0 MW, 6.3/11 kV.
F. Comparison of WECS Configurations
The top 10 wind turbine manufacturers as of December
2012 and their main turbine configurations are summa-
rized in Fig. 7. They account for approximately 77% of the
45 GWs installed wind power capacity in 2012 [3]–[5]. The
details about the turbine configurations are obtained from
the respective company product brochures [12], [13], [28],
[29] and details from the survey papers listed before in
Section I. The Type 3 turbines (DFIG) hold the highest
market share and this technology have been used by 7

manufacturers among the top 10. Approximately 100 dif-
ferent DFIG turbine models are available from all the wind
turbine manufacturers. The Type 4 turbines are produced
by 6 manufacturers, while 4 of them are offering direct-
drive solutions. This implies that the best selling wind
turbines in the present market use Type 3 and 4 technol-
ogies. The future projects announced by the wind turbine
manufacturers indicate that the Type 4 technology would
take over the wind energy market in coming years.
The summary of all five types of turbines is given in
Table 2. They are compared using generator, power con-
verters employed; capacity of power converter; speed-range
achievable; requirement for soft-starter, gearbox and exter-
nal reactive power compensation; and maximum power
point tracking (MPPT) ability; aerodynamic power control,
compliance with the fault ride-through requirement; tech-
nology status; and market penetration. Overall, the Type 3
and 4 turbines are most favorable for MW-level application.
In this paper, the generator-converter configurations are

investigated in detail for these two types of wind turbines.
I V . O V E R V I E W O F P O W E R C O N V E R T E R S
F O R M W - W E C S
The five types of WECS discussed earlier show that, since
1980s, the power electronics technology has an important
collaboration with the commercial grid-connected wind
turbines [32]. This technology has gone through much ad-
vancement, and the state-of-the-art solutions are available
in the form of full-scale converters. The current technol-
ogy uses power electronics at wind turbine and wind farm
level for the energy conversion and grid integration. This
section is dedicated to discuss the power electronics
technology briefly. The general overview of power
converters, technical requirements for power converters
in WECS and classification of power converters for MW-
WECS are addressed.
Fig. 6. Type 4 WECS with distributed drive-train and quantum
generators.

A. General Classification of Power Converters
The general classification of power converters which
are more relevant to the wind energy application is shown
in Fig. 8. The objective of the power converters is to enable
variable-speed operation in Type 3 and 4 WECS, while
eliminating the need for soft-starter and reactive power
compensation. To enable the grid connection of these
wind turbines, the variable voltage/frequency of the wind
Table 2 Summary of Five Types of Wind Energy System
Configurations
Fig. 7. Summary of top 10 wind turbine manufacturers and their
market share by December 2012. [source: REN21 and GWEC]
(2S/3S ¼ two-/three-stage gearbox, DD ¼ direct-drive).
generator should be converted to fixed voltage/frequency.
To accomplish this, a wide variety of power conversion
stages can be employed as summarized in Fig. 8 [70]. Most

of the power conversion stages have found commercial
applications, and some have been proposed in literature
with promising features for future development.
The converter topologies are broadly classified as direct
and indirect according to the power conversion performed.
The direct conversion uses single-stage ac/ac converters,
while indirect conversion uses two-stage (ac/dc+dc/ac) or
three-stage (ac/dc+dc/dc+dc/ac) conversion. Some of
these converters are adopted from the electric drives in-
dustry, while some other are solely developed for wind
power application. The direct ac/ac converters and current
source converters are also shown in the classification be-
cause they are main competitors to the voltage source
converters in the electric drives industry [119]–[121]. The
successful converters from the electric drive industry are
also presented even though they have not been used in
wind industry yet.
B. Technical Requirements for MW-WECS
Power Converters

The power converters in the first generation of wind
turbines (Type 1) were used for smooth grid connection
only. They are disconnected from the circuit once the tur-
bine is tied to the grid. But the modern Type 3 and 4 WECS
demand the power converters to meet several technical
and operational requirements [30], [33], [69], [122]. The
most important requirements are listed below:
• Initial Cost: This factor has prime importance in
achieving low cost of energy (COE) and competing
with other energy sources [37]. The initial cost of
power converter is only a fraction (approximately
7%–12%) of overall wind turbine cost [53]. Though
it is a small part, huge cost savings can be accom-
plished for a wind farm which consists of hundreds
of wind turbines.
• Reliability, Modularity and Maintenance Cost:
In addition to the initial cost, the maintenance cost
(replacement cost for components and salary for
technicians) must also be very less to achieve low
COE. According to the latest reports on wind tur-

bine faults, the electric generators and power con-
verters are on the top of list with an average failure
rate of 13%–20% [123], [124]. The power convert-
er faults increase the downtime of wind turbine
operation, and overall cost of energy. For this rea-
son, the power converters for wind turbines, espe-
cially offshore turbines must have high reliability
[52]. The power converters having a modular
structure are preferable because even one power
converter fails, the wind turbine can still work
with reduced capacity, and thus the downtime can
be mitigated [125].
• Efficiency: At MW power level, efficiency is an
important factor in reducing COE [126]. Even 1%
improvement in the efficiency of power converter
can save millions of dollars at the wind farm level,
which consists of hundreds of power converter
based wind turbines. The power losses which di-
rectly affect the efficiency should be minimized by
using highly efficient switching devices, optimal

arrangement of switching devices (also called
Fig. 8. General classification of state-of-the-art high power
converters (LCI: load commutated inverter, PWM: pulsewidth
modulated).
power converter topology), cooling system,
modulation/control schemes, etc.
• Power Quality: The power quality can be attrib-
uted to various parameters of power converter. The
output voltage waveform must be close enough to
sinusoidal waveform. This is also described as
number of steps in the output voltage waveform
ðdv=dtÞ. As the number of steps in the waveform
increases, the dv=dt decreases and thus the require-
ment for output filter also decreases. Moreover,
the electromagnetic interference becomes lower as
dv=dt decreases [127], [128]. The total harmonic
distortion (THD) of generator and grid currents
must be lower to decrease generator shaft oscilla-

tions and to feed quality currents to the grid, re-
spectively [70].
• Grid Code Compliance: It is one of the important
requirements for the grid-connected MW wind
turbines [33]. The power converters must feed
currents to grid with low THD (less than 5%),
provide reactive power whenever requested by the
grid operator, ride-through during grid faults, and
provide voltage/frequency support among other
requirements. These requirements must be accom-
plished by the power converter itself, without re-
questing support from the external hardware/
components such as STATCOM or FACTS.
• Footprint and Weight: Unlike in the electric
drives, a limited space is available in the nacelle of
wind turbine. The power converter (also electric
generator) must have high power density to achieve
small footprint and weight. This is an important
requirement especially for the offshore wind
turbines.

• Cable Size and Losses: The generator/converter
ac output is connected to the step-up transformer/
substation through ac cables. The typical hub
heights for the modern wind turbines are in the
range of 60–150 m, and thus the cost of cables and
associated losses become high. The power conver-
sion system must take this factor into account and
decrease the cable cost and losses to the maximum
possible extent.
An ideal power converter must possess all the above
features. In practice it is impossible to design a power
converter embedding all the above technical merits. The
engineering approach would be to satisfy most crucial
requirements while sacrificing the least important condi-
tions. A best-selling power converter obviously incorpo-
rates most of the above features for wind turbines.
C. Classification of Power Converters for MW-WECS
In pursuit of achieving the previously mentioned tech-
nical requirements, various power converter configura-
tions have been developed by wind turbine manufacturers

and their supporting power converter companies. The
classification of state-of-the-art WECS power converters is
of a complex subject matter and it is not possible to classify
all the converters based on one parameter/operation. In
this paper, we have classified the generator-converter con-
figurations into four different groups as summarized in
Fig. 9 to facilitate easier discussion.The power converter
topologies which have been commercialized by several
wind turbine manufacturers, and also proposed in litera-
ture with promising features, belong to these four distinct
categories. The main features and drawbacks of each con-
figuration are discussed in the following sections with
important survey results being tabulated (refer to Table 8
in the Appendix). To simplify the diagrams of various
WECS configurations in this paper, the generator-side fil-
ters are not shown, and the grid-side filters are represented
by equivalent block diagram.
V . B A C K - T O - B A C K C O N N E C T E D
P O W E R C O N V E R T E R S
The power converters, which are identical on both the

generator- and grid-side, and linked through a dc-link, are
classified as back-to-back (BTB) connected converters.
Different BTB converters which can be used in the com-
mercial WECS are summarized in Fig. 10. They perform a
conversion of variable voltage/frequency output of the
generator to dc, and then dc to ac, with fixed voltage/
frequency for the grid connection. The power flow is bi-
directional, and thus the BTB converters can be used with
SCIG, DFIG, PMSG, and WRSG. The BTB converters are
Fig. 9. Classification of power converters for MW-WECS.
classified as low voltage (G 1 kV) and medium voltage (1–
35 kV) converters according to IEC 60038 standard given
in Table 1.
A. LV Converters
In this subsection, four LV converter configurations are
analyzed as follows. The most standard voltages used by

many commercial wind turbine manufacturers for the LV
grid connection are 690 and 575 V.
1) Full-Scale BTB 2L-VSCs: A typical Type 4 WECS using
BTB connected full-scale two-level (2L) voltage source
converters (VSCs) is shown in Fig. 11. The voltage source
rectifier (VSR) and voltage source inverter (VSI) are linked
by a dc-link capacitor. The VSR and VSI are realized by LV
Insulated Gate Bipolar Transistors (LV-IGBTs) arranged in
a matrix form. The dc-link unit is usually realized using
series/parallel string of capacitors to achieve required
voltage and capacitance level. The dc-link provides decou-
pling between the generator and grid, and thus the
transients in the generator do not appear on the grid-side.
The wind generators, PMSG [109], [110], WRSG [129],
and SCIG [82], [113] can be used with this configuration.
In terms of technology status and market penetration, it is
a mature power converter topology and being used by 90%
Type 4 wind turbines rated below 0.75 MW.
The power rating of the converter is usually equal to the
generator output power. For example, a 0.75-MW electric

generator is connected to the grid through a 0.75 MW power
converter. The VSR controls the generator torque and speed,
while the VSI controls the net dc-bus voltage and grid
reactive power. The net dc-bus voltage is maintained higher
than the peak of grid line-line voltage to ensure proper
operation of grid-side converter. The switching frequency of
VSR and VSI is maintained at 1–3 kHz to achieve lower
switching losses and higher power density [130], [131].
The grid current contains higher total harmonic distor-
tion, and to meet the grid codes, LCL filters are used on the
grid-side [79]. The generator-side harmonic filter is not
shown. The entire power converter including generator-
side filters, VSR, dc-link, VSI, grid-side harmonic filter are
packed in a cabinet and placed in the nacelle. The output
of grid-side LCL filter is connected through three-phase ac
cables to the step-up transformer which is located at the
Fig. 10. Classification of back-to-back connected converters.
(VSC: voltage source converter, NPC: neutral-point clamped,
FC: flying capacitor, 2L:
two-level, 3L: three-level, 4L: four-level).
Fig. 11. Type 4 WECS with two-level BTB voltage source
converters. (Mainstream commercial power converter

configuration).
bottom of tower. The ac cables present significant cost and
losses as they are rated for low voltage and high current
operation. Due to the mass production of three-phase con-
verter modules, the cost of these converters is low. One of
the most widely used commercial VSC modules is SKiiP by
Semikron which is equipped with its own heat sink, semi-
conductor switching devices and gate drivers, and thus they
feature compact design with high power density [132].
2) Partial-Scale BTB 2L-VSCs: The Type 3 semi-variable
speed WECS using BTB voltage source converters is shown
in Fig. 12. The stator of the generator is directly connected
to the grid, while rotor of the generator is connected
through power converter [99]–[101], [133]. The power
rating of the converter is usually 30% of the generator
rated power [58]. For example, a 2.5 MW DFIG requires
only a 0.75 MW power converter. Similar to the full-scale

converter, the partial scale converter in DFIG WECS can
control the generator active/reactive power, dc-link volt-
age and grid power factor [103], [134]–[137].
The speed range achievable is decided according to the
rating of power converter. A power converter with 30%
rated capacity can control the DFIG speed in �30% which
is sufficient to perform variable-speed operation. The BTB
converters in rotor circuit enable bidirectional power flow
and thus power can flow from rotor to grid or vice versa.
The use of partial-scale converter decreases the size and
weight of Type 3 turbine in comparison to Type 4 turbine.
The lower size and weight of power converter allows more
room in nacelle and thus step-up transformer can be placed
in nacelle to decrease the cable costs and losses. This ap-
proach has been applied in ECO110 model of Alstom wind
turbines, among few others. The FRT operation of Type 3
wind turbines is complicated and challenging compared to
the Type 4 wind turbines, and a significant research has
been carried out in this area [58], [138]–[140].

3) Parallel BTB 2L-VSCs With Common DC-Link: For
power ratings greater than 0.75 MW in Type 4 turbines
(2.5 MW in Type 3 turbines), the current carrying capabi-
lity can be increased by connecting the three-phase VSC
converters along with harmonic filters in parallel. For
example, as shown in Fig. 13, two BTB VSC modules can be
connected in parallel to achieve a power rating of 1.5 MW
and 5 MW in Type 4 and 3 turbines, respectively. The
dashed lines represent the connection of Type 3 turbine,
where stator is directly connected to the grid, and the
power converter is connected to the rotor windings. For
higher power ratings, more converter modules can be
connected in parallel. The dc-link is configured as a
common element for all the converters to reduce cost and
space [141]. This configuration offers energy efficiency and
Fig. 12. Type 3 WECS with two-level BTB voltage source
converters. (Mainstream commercial power converter
configuration).
Fig. 13. Type 3 and 4 WECS with parallel connected BTB 2L-
VSCs and common dc-link. (Mainstream commercial power
converter configuration).

redundancy. For example, when the wind speed is low, one
or more converters can be turned-off leading to higher
system efficiency. When a converter fails, other converters
can still deliver the power, but with reduced capacity.
Moreover, by operating converters in interleaving mode,
the equivalent converter switching frequency can be in-
creased, therefore producing less total harmonic distortion
in generator- and grid-side currents.
Due to a mismatch in converter and grid-side filter pa-
rameters, the circulating currents exist in both the generator
and grid-side converters and this issue should be considered
in the design of the controller. On the generator side, L filters
are connected between each converter to reduce the
circulating currents. On grid-side, in addition to the L filters,
a CL filter is used to reduce the THD of grid currents. Since
the dc-link is common, the generator-side converters along
with harmonic filters can be placed close to the generator in

nacelle, while the dc-link, grid-side converters and step-up
transformer can be placed at the bottom of tower. The
generator-side converter is connected to the dc-link through
dc cables leading to lower cables costs and losses, in addition
to the reduced nacelle weight. However, this configuration
leads to lower reliability as dc-link capacitors are more prone
to failure. In Enercon E-126 model, more than 10 power
converters are connected in parallel to reach a power rating
of 7.5 MW. The large number of modules leads to complex
control as well.
4) Parallel BTB 2L-VSCs With Individual DC-Links: To
solve the issue of circulating currents, as well as reliability
issue, the dc-links can be configured as individual elements
as shown in Fig. 14 without losing the best qualities such as
power handling capability, modularity, redundancy and
efficiency. But the individual dc-link in each converter
module leads to higher system cost. Moreover, the com-
plete power converter must be placed in nacelle to decrease
the voltage drop in generator-side cables, and thus overall
nacelle space requirement increases. Despite these dis-

advantages, this configuration is still dominant technology
in Type 3 and 4 WECS.
To minimize the circulating currents, open winding
transformers can also be used at the grid-side. The open-
winding transformer provides isolation between the con-
verters, but with a drawback of high cost and size. The
configuration of harmonic filters on grid-side is simpler
compared to the previous topology in Fig. 13. The LCL
filters are used separately for each grid-side converter. The
circulating currents still exist on the generator-side for
which open-winding generators should be used. The dis-
cussion on open-winding generators will be presented in
Section VII.
B. MV Converters
The LV converters discussed before are efficient and
cost effective at power levels lower than 3 MW in Type 4
turbines. As the power rating increases, the number of
converter modules increases and as a result, the size, cost,
and complexity of the system increases [61], [131]. A
summary is provided between the LV and MV operation of

a 6- MW Type 4 wind turbine in Table 3, where it has been
noticed that the MV operation of WECS is the most suitable
and economical approach for power ratings greater than
3 MW [70], [142]. The MV operation is a mature technol-
ogy in the electric drives industry [122], but wind turbine
manufacturers are reluctant to move from LV to MV tech-
nology due to the limited availability of MV generators and
less knowledge available for the MV operation of turbines.
In [45], a detailed cost analysis has been carried out be-
tween the LV and MV operation of wind turbines, and the
results show that the cost of energy production can be
decreased by 2%–4% with MV operation. Currently only
4-MV wind turbines are operational, but 7 manufacturers
have announced their future projects based on MV tech-
nology (refer to Table 11 given at the end of this paper).
The power converter configurations for the MV operation
of wind turbines are discussed as follows.
1) Series Connected Switches: The two-level VSC shown
in Fig. 11 can also be used for MV applications by connect-

ing the switching devices in series [69]. This is a simple
Fig. 14. Type 3 and 4 WECS with parallel connected BTB 2L-
VSCs, individual dc-link’s and open winding transformer.
(Mainstream commercial
power converter configuration).
solution for MV operation, but due to the mismatch in
IGBT characteristics, the converter capacity decreases.
The Converteam VDM5000 with a maximum power and
voltage rating of 7.2 MW and 4.16 kV, respectively, fea-
tures this converter for the MV drives application [122].
2) BTB Neutral-Point Clamped Converters: As an alterna-
tive solution, the three-level diode clamped converter (3L-
DCC) also known most popularly as neutral-point-clamped
(NPC) converter has been widely studied in literature for
Type 4 turbines [90], [111], [143]–[145]. In this configu-
ration, two 2L-VSC’s are stacked one over the other using
split dc-link capacitors and clamping diodes [146]. With
this arrangement, the converter output phase voltage

contains three levels leading to reduced dv=dt and
electromagnetic interference than 2L-VSCs [126], [127],
[146], [147]. As shown in Fig. 15, the NPC converters
enable MV operation, and commercial wind turbines have
reached 6 MW power rating without connecting switching
devices in series or parallel [143]. In the commercial
solutions offered by ABB, the switching device is realized
by Integrated Gate-Commutated Thyristor (IGCT) with a
Table 3 Comparison of LV and MV Operation for a 6-MW Wind
Turbine [142]
Fig. 15. Type 4 WECS with BTB connected neutral-point
clamped (NPC) converters. (Commercial power converter
configuration).
voltage rating of 4.5–6.5 kV [143]. The other manufac-
turers such as Converteam and Ingeteam use MV-IGBT as
the switching device [144].
In the current MV wind turbines, offered by the Areva,

Shandong, XEMC-Darwind and Zephyros, the NPC con-
verters have been used with the PMSG, but they can also
be used with WRSG [145], SCIG [111], [148] and DFIG
[149]–[151]. One of the future wind turbine projects,
Condor6 proposed to use MV SCIG in conjunction with the
BTB NPC converters. In the current WECS market, DFIGs
with MV stator voltages (6.6 to 12 kV) and LV rotor voltages
are being offered by Acciona, China Creative Wind Energy
(CCWE) and Senvion. The MV stator voltages can eli-
minate the wind turbine step-up transformers (and asso-
ciated losses), and contributes to significant savings in the
collector system costs. As stated by these manufacturers,
this configuration is ideal for wind farms which are in close
proximity to the substation (less than 5 km). The rotor
circuit voltage ratings are lower and thus BTB NPC con-
verters can be employed with LV switching devices, or
simply BTB 2L-VSCs can be used.
The switching actions of the semiconductor
switches lead to the drift in the capacitor voltages. If the

capacitor voltages are not balanced, it leads to higher stress
on the semiconductor switches and damages them. This
issue has been extensively studied by industry and
academia [152]–[154]. The use of external hardware can
mitigate the problem of capacitor voltages imbalance
[152]. It is also possible to use carrier-based pulse with
modulation (PWM) with zero-sequence voltage injection
[155] or space vector modulation (SVM) with redundant
switching states selection [153], [154] for the balancing of
capacitor voltages. Since the high-power NPC converters
have been marketed by many manufacturers, it can be
assumed that this problem has been solved [126]. To mini-
mize the switching losses and also to allow proper heat
dissipation, the semiconductor device switching frequency
is limited to few hundred Hertz [143], [144]. The outer
switching devices (those connected to the positive and
negative dc-bus) operate at higher switching frequency
while the inner switches (those connected to the split-dc
bus) work with low switching frequencies. This phenom-
enon leads to uneven power losses, heat dissipation, and

possible derating of converter. This leads to a difficulty in
the design of the mechanical layout of the semiconductor
switches [128].
3) Other Voltage Source Multilevel Converters: The uneven
power loss challenge associated with the NPC can be
solved by using active neutral-point clamped (ANPC) con-
verters as shown in Fig. 16(a). In this configuration the
clamping diodes are replaced by the IGBT switches giving
more redundancy to maintain equal switching frequency
(and thus switching losses) among all the IGBT’s [126],
[156]. Under similar operations, the BTB 3L-ANPC con-
verters are capable of handling of 32% higher power (up to
7.12 MW) and 57% higher switching frequency (1650 Hz)
compared to the BTB 3L-NPC converters [157]. This con-
figuration is applied more recently in the MV drives in-
dustry, and it can be used in the WECS industry as well.
One of the top manufacturers, Vestas is currently re-
searching this power converter topology.
The flying capacitor converter configuration is similar

to the NPC converter, where the clamping diodes are re-
placed by the flying capacitors (FC) as shown in Fig. 16(b).
This configuration offers a simplified structure and more
redundant switching states in order to achieve easier con-
trol for the capacitors voltage balancing [130], [158], [159].
The power distribution among the switching devices be-
comes more even in FC converters compared to the NPC
converters [160]. This configuration requires a large num-
ber of capacitors and each of them requires precharging
circuit [16], [69]. The clamping capacitors in FC are less
reliable compared to the diodes in NPC converter. The
control scheme requires a greater number of sensors to
send feedback signals from the FC’s and these add cost and
complexity to the system. Moreover, the average switching
frequency of the FC converter should be high (1200 Hz) to
ensure the balancing of capacitor voltages and this causes
higher switching losses [128]. This configuration has not
found its commercial application in the wind energy
industry yet, even though it was commercialized in the MV
drives industry (with less market penetration).

The NPC converters are most suitable for 3–4 kV class
MV operation. To connect to the MV collection point of
6.9 kV (North America Standard) or 6.6 kV (Europe
Standard) without using a step-up transformer, the
switching devices in a NPC converter should be connected
in a series [126], [161], but this approach leads to derating
the converter. As shown in Fig. 16(c), a four-level (4L)
diode-clamped converter (DCC) can be used to achieve
higher levels of MV operation [70], [162]. In comparison
to the 3L-DCC (NPC converter), the 4L-DCC offers added
benefits such as: MV operation with greatly reduced de-
vice voltage rating, increased equivalent switching fre-
quency, lower values for the grid-side filter, grid friendly
waveforms, and better grid code compliance [92]. How-
ever, the dc-link capacitor voltages balancing becomes
more complex and sophisticated, and therefore control
techniques or external hardware should be used [163],
[164]. The semiconductor device count also increases. In
particular, the clamping diodes increase from 12 to 36 in

comparison to the BTB NPC converters. Another disad-
vantage is that the uneven power loss and heat dissipation
among outer and inner switching devices becomes predo-
minant compared to the NPC converters. The four-level
converters are not commercialized yet, but the basic
power semiconductor switches are readily available in the
market.
To achieve higher operating voltages, decrease grid-
side filter size and eliminate the wind turbine transformer,
multilevel ANPC converters can be used as shown in
Fig. 16(d). For a 5L-ANPC converter, the 3L-FC converter
is connected between the internal switching devices of
the 3L-ANPC converter, there by producing high number
of levels in the output voltage waveform. Though the
power handling capacity remains same as 3L-ANPC con-

verter, the higher number of output voltage levels can be
achieved by adding more number of FC cells [126]. Com-
pared to the 3L-NPC, 4L-DCC and 3L-ANPC converters
discussed earlier, the multilevel ANPC converters feature
more modular design without connecting switching or
clamping devices in series. Compared to the standard 5L
diode-clamped converter, the number of switching devices
and clamping diodes are lower in 5L-ANPC converter
[165]. The disadvantages for this converter includes com-
plex structure, pre-charging circuits for the FC voltages,
complicated control scheme to regulate the FC voltages in
addition to the split dc-link capacitor voltages.
The high grid current quality is more important (to
comply with the grid codes) compared to the generator
current quality. For this reason, multilevel ANPC converters
can be used at the grid-side, while converters which produce
moderate power quality can be used at the generator-side.
This leads to hybrid configuration with cost savings in the
wind turbine power converters. Note that the dc-link in
multilevel ANPC converters consists of two split dc-link

capacitors, therefore they can be integrated easily with the
generator-side 3L-NPC or 3L-ANPC converters. At grid-
side, the size for the filters can be further decreased by
using 7L and 9L ANPC converters as proposed in [166].
Although these converters can increase the number of
output levels and decrease output filter size, the voltage
stress of the main switches remain as the same as three-
level converters which limits the power rating of these
converters.
Fig. 16. Per-phase representation of promising multilevel
converters for medium voltage WECS: (a) BTB three-level
active NPC (ANPC) converters.
(b) BTB three-level flying capacitor (FC) converters. (c) BTB
four-level diode-clamped converters. (d) Multilevel ANPC
converters.
4) BTB Current Source Converters: The previously dis-
cussed topologies belong to voltage source converters. The
current source converters (CSC) can also be used in wind

turbines as shown in Fig. 17. The CSCs are duality of VSCs. A
comprehensive summary between the VSC and CSC con-
verters for MV-WECS is given in Table 4 [70], [161], where it
has been observed that the CSC configuration is most favo-
rable for power ratings greater than 5 MW. The topology
consists of PWM current source rectifier (CSR) and current
source inverter (CSI). The active switches are realized by
Symmetric Gate-Commutated Thyristor (SGCT). The CSR
and CSI are linked by a dc-choke and thus, similar to the
VSCs, decoupling between the generator and grid can be
achieved [167], [168]. Three-phase capacitor banks are used
on the ac-sides of CSR and CSI to assist the commutation of
semiconductor devices and also to mitigate the switching
harmonics [161], [169].
This topology features a simple structure, and reliable
short-circuit protection. In VSCs, the dc-link capacitors
are bulky components, whereas in CSC, the dc-choke is
the bulky component. The dynamic response of the CSC
converters is slower compared to the VSC converters due

Fig. 17. Type 4 WECS with BTB connected current source
converters. (Promising power converter configuration).
Table 4 Comparison of Voltage and Current Source Converters
for MV-WECS [70], [161]
to bulky dc-choke and lower switching frequency opera-
tion. The net dc-bus voltage is maintained lower than the
peak of grid line-line voltage to ensure proper operation
of current source inverter. In MV drives industry, the
CSCs are competent to the multilevel converters due to
transformerless operation, but in WECS a step-up
transformer is mandatory. The cost of CSC production
is 1%–2% lower compared to the VSC converters [45].
The CSC technology is successfully applied to the multi-
megawatt MV drives [121], but it is not yet used in the
MV-WECS.
C. Comparison of BTB Power Converters
The summary of comparison between BTB converters

is given in Table 5 with respect to power/voltage rating,
semiconductor/passive component count, voltage stress of
switches, reliability, power quality, converter and control
complexity, grid code compliance, technology status and
market penetration. The commercially practiced power
converters in the present wind industry are limited to
2L-VSC, parallel 2L-VSCs and 3L-DCC only. The analysis
given here summarizes the feasibility of applying various
power converters in LV and MV WECS. The advantages
and disadvantages of these converters are listed in Table 8
in the Appendix. The multilevel converters are promising
for next-generation wind turbines due to the technical
merits they exhibit.
V I . P A S S I V E G E N E R A T O R - S I D E
C O N V E R T E R S
The previously discussed BTB topologies enable a four-
quadrant operation. But in the WECS, the power flow is
unidirectional i.e., from the generator to the grid. For this
reason, passive (diode-bridge) converters can be employed
on the generator side instead of pulse width modulated

(PWM) active converters [170], [171]. The diode-bridge
rectifiers are less expensive and inherently more reliable
compared to the PWM converters. In the PMSG and
WRSG, the rotor flux is generated by permanent magnets
and rotor field excitation, respectively. For this reason, the
generator-side power conversion system in the PMSG/
WRSG wind turbines can be realized using passive con-
verters [172]. The induction generators (SCIG/DFIG) re-
quire magnetizing current during its operation and thus
they cannot allow passive converters on the generator side.
The use of passive generator-side converters is asso-
ciated with a few disadvantages. The generator currents
contain significant 5th (14%) and 7th (7%) harmonics and
this leads to 6th harmonic distortion (10%) in the electro-
magnetic torque [30], [173]. However, due to the decou-
pling offered by a second dc-link, these torque ripples and
generator current distortion do not cause any conflict to
the grid-codes. The passive generator-side converters have
been used in practical WECS ranging from a few kilowatts

(kWs) to megawatts (MWs). Few examples of MW wind
Table 5 Comparison of Back-to-Back Connected Power
Converters for Megawatt Wind Turbines
turbines include the Enercon E82, Clipper Liberty C89,
Vensys V70/77, and Gold Wind GW70/77. The summary of
comparison between the active and passive generator-side
converters for MW-WECS is given in Table 6 [70], [85].
1) Diode Rectifier + 2L-VSC: The power converter con-
figuration for PMSG/WRSG WECS with diode rectifier and
2L-VSC is shown in Fig. 18 [46]. This topology offers low
cost, light weight solution compared to the BTB 2L-VSCs.
The generator output voltage is converted to dc by the
diode-rectifier which is then converted back to ac by 2L-
VSC. During low wind speeds, the diode rectifier output
voltage becomes significantly lower. As discussed earlier, to
transfer the generated power to grid, the dc-link voltage

must be higher than the grid line-line voltage. To ensure
this condition, the generator should be over rated [174].
This configuration is being used in the Clipper Liberty
2.5 MW wind turbines along with the quantum drive-train
shown in Fig. 6 [118]. The absence of intermediate dc/dc
converter decreases the degree of control freedom by one.
Table 6 Comparison of Active and Passive Generator-Side
Converters for LV-WECS [70], [85]
Fig. 18. Type 4 WECS with diode rectifier and 2L-VSI.
(Commercial power converter configuration).
In other words, the control system cannot incorporate
either MPPT operation or regulation of dc-link voltage. In
the Clipper Liberty wind turbines, MPPT is achieved by
controlling the grid-side inverter, while dc-link voltage is
allowed to vary with respect to the wind speed. The grid-
side inverter is then designed according to the maximum
possible dc-link voltage.

2) Diode Rectifier + 2L-Boost Converter + 2L-VSC: To
increase the control freedom by one, a boost converter can
be employed as an intermediate stage. The WECS with
diode-rectifier, boost converter and PWM inverter is
shown in Fig. 19. The variable output voltage of the gene-
rator is converted to dc by the diode-rectifier, and the
boost converter increases the unregulated dc-link voltage
to a higher level that is suitable for the PWM inverter. The
capacitor in the first dc-link filters the ripple in diode
rectifier output dc voltage, and it is an optional component
[53], [175]. The boost converter also enables variable-
speed operation for the PMSG/WRSG WECS by perform-
ing maximum power point tracking (MPPT) [176], [177].
With this scenario, the wind energy conversion efficiency
improves, especially during low wind speeds.
The control system development for the MPPT is less
complicated with the passive converters. The cost and
complexity associated with the gate drivers also decreases
as the semiconductor switches decrease from 6 to 1 [171].
To increase power handling capacity, more number of

boost converters and 2L-VSCs can be connected in parallel
similar to the BTB converters (refer to Figs. 13 and 14).
Due to the interleaving operation of boost converters, the
inductor current ripple becomes lower and thus the size of
dc choke becomes smaller. This configuration has been
implemented with WRSG in Enercon E82 turbines.
3) Diode Rectifier + 3L-Boost Converter + 3L-VSC: The
diode-rectifiers can also be used in MV WECS. As shown in
Fig. 20, the generator-side NPC rectifier can be replaced
with a diode-rectifier and 3L-boost converter. With this
topology, the advantages of generator-side passive con-
verters and grid-side multilevel converters are combined
[178], [179]. The voltage rating for the IGBT/IGCT
switches and diodes is half of the net dc-bus voltage. The
3L boost converter offers many benefits compared to the
standard boost converter: reduced switching and reverse
recovery losses, and balancing of capacitor voltages [180].
Fig. 20. Type 4 WECS with diode rectifier, 3L boost converter
and 3L-VSI. (Promising power converter configuration).

Fig. 19. Type 4 WECS with diode rectifier, 2L boost converter
and 2L-VSI. (Commercial power converter configuration).
Compared to the BTB NPC converters, the number of
active switches in generator-side converter decreases from
12 to 2, and it leads to a cost-effective solution for MV
turbines. This configuration leads to enhanced power
quality and better grid code compliance compared to the
BTB-NPC converters [85], [142].
Similar to the 3L boost converter, a 4L boost converter
can be used in conjunction with the 4L diode-clamped
converter [112], [164]. As mentioned earlier, the balancing
of the dc-link capacitor voltages with the 4L-DCC is very
complicated compared to the NPC converter [181], [182].
But the intermediate 4L-boost converter ensures balancing
of the capacitor voltages during all the operating conditions.
Even with the higher number of clamping diodes in the grid-
side converter, the cost of this configuration becomes

approximately equal to the BTB-NPC converters [112], [164].
4) Diode Rectifier + Buck Converter + CSC: The current
source converters can also be designed with passive
generator-side converters. The configuration of WECS
with diode rectifier, 2L buck converter and PWM CSI is
shown in Fig. 21 [183]. This configuration is simple and
less complicated compared to the previous configurations
[184], [185]. In VSC’s the dc-link voltage is boosted, while
in CSCs the dc-link current is boosted. Like others, this
configuration also suffers from the high torque ripple.
The summary of passive generator-side converters
along with technical merits and demerits is provided in
Table 8 in the Appendix. The present and future wind
turbine projects are also listed in Table 8 in the Appendix.
V I I . C O N V E R T E R S F O R
M U L T I P H A S E G E N E R A T O R S
As discussed in the previous two sections, the parallel
power converters increase the power handling capacity.
The wind energy industry also practiced multiphase gene-

rators and distributed converters as an alternative approach
to increase the power handling capacity. A generator
having two-sets of three-phase windings is denoted as a
six-phase generator. The generators which have more than
six-phases are expressed as open-winding generators. In
this section, the distributed and cascaded converter config-
urations for six-phase and open-winding generators are
presented.
A. Converters for Six-Phase Generators
The six-phase machines, both induction and synchro-
nous, are widely used in the electric drives industry. The
wind energy industry also adopted the use of six-phase
generators in Type 4 turbines. With the two sets of wind-
ings, the insulation level for each winding decreases. The
power handling capacity also increases by two-times as
each set of winding carries half the rated current [186].
1) Distributed Converters: A six-phase generator based
WECS with BTB 2L-VSCs is shown in Fig. 22 [67], [131].
The stator windings are separated by 30�, and thus the stator
voltages are phase shifted by 30�. The phase-shift causes
cancelation of 5th and 7th harmonics in the stator currents
with which the torque ripples are minimized [172]. This

feature leads to lower size or even elimination of generator-
side harmonic filters. Since the two sets of windings are
separated, no circulating current flows through the gener-
ator-side converters, and thus the power density (high-power
per ampere) of system increases. The Envision E128-3.6 MW
wind turbines use such a six-phase configuration with
DD-PMSG. The present industry is also investigating the use
of BTB-NPC converters in place of BTB 2L-VSCs for MV
operation [144]. For the details of few other six-phase con-
figurations, refer to Tables 9 to 11 in the Appendix.
A similar six-phase configuration with passive gener-
ator-side converters is shown in Fig. 23. The disadvantage
of high torque ripples with the passive front-end can be
mitigated by employing six-phase configuration with 30�
phase shift. A six-phase configuration employed by the
Vensys V70/77 and Goldwind GW70/77 wind turbines is
shown in Fig. 24 [187]. The power converter is realized by
six-phase diode rectifier, three-channel boost converter
and two-channel VSCs. The dc-link is configured as a

Fig. 21. Type 4 WECS with diode rectifier, 2L buck converter
and CSI. (Promising power converter configuration).
common element for VSCs. The boost converters and VSCs
operate in interleaving mode to decrease the inductor and
grid current ripple, respectively.
2) Cascaded Converters: The MV operation of wind tur-
bines is cost effective and a promising solution for mega-
watt wind turbines. In the standard approach, the output
of MV generator is connected to the MV grid through a
MV power converter [131], [143], [145]. But, one of the
shortcomings in the current wind energy industry is the
lack of availability of the MV generators. The use of
LV converters at the generator-side and MV converters at
the grid side represents a very promising approach. The
off-the-shelf LV generators and MV converters can be re-
arranged to increase the efficiency of wind energy con-

version while reducing the overall cost [188].
Fig. 22. Type 4 WECS with six-phase generator and parallel 2L
VSC modules. (Commercial power converter configuration).
Fig. 23. Type 4 WECS with six-phase generator and parallel
boost converter + 2L-VSC modules. (Promising power converter
configuration).
Fig. 24. Type 4 WECS with six-phase generator and three-
channel boost converter + parallel 2L-VSC. (Commercial power
A six-phase configuration with two series-connected LV
converters on the generator-side, and an NPC converter on
the grid-side is shown in Fig. 25 [189]. The generator-side
2L converter outputs are connected in a series to achieve
higher dc-link voltage, and also to provide midpoint for the
grid-side NPC. The power output of the generator is dis-
tributed among the two converters and thus the ampere-
per-phase ratio decreases. The midpoint voltage control is a
challenging issue with this particular configuration. This
configuration may not work during the whole wind speed

range due to the fact that the dc-link voltage becomes
insufficient during low wind speed conditions.
As shown in Fig. 26, passive generator-side converters
can also be used to decrease the cost of the turbine and to
operate during the whole wind speed range [190], [191].
The torque ripples can be minimized because of the phase-
shifted windings, and meanwhile the LV-MV operation can
be accomplished. This configuration is more promising
compared to the active generator-side converters because
during low wind speed operation, the boost converters can
operate with a higher duty cycle to maintain sufficient dc-
link voltage [190]. Similar to the VSC converters, the LV to
MV conversion can also be achieved by current source
converters. The PWM-CSI configuration is much simpler
and there are no issues related to the balancing, unlike
those in the NPC converters presented in Fig. 26. This
configuration has already been introduced in the MV
drives industry [192], and it can be adopted for high power
wind turbines.

B. Converters for Open-Winding Generators
The wind generators with multiple windings (also
named open-winding structure) are also used in the present
wind turbines. The generator is equipped with more than
two sets of windings, and offers new possibilities to connect
the converters in different configurations [52], [57], [67],
[193], [194].
1) Distributed Converters: The distributed converters are
similar to the one shown in Fig. 22, except that the
number of phases or converter channels are more than
two. One of the practical wind turbines, which use six sets
of three-phase windings, is shown in Fig. 27 [195], [196].
The Gamesa G10x 4.5 MW wind turbines use this concept
Fig. 25. Type 4 WECS with six-phase generator, series
connected 2L-VSRs, and 3L-VSI. (Promising power converter
configuration).
Fig. 26. Type 4 WECS with six-phase generator, series
connected 2L-boost converters, and 3L-VSI. (Promising power

with 6 BTB 2L-VSCs modules in parallel. Each module is
rated for 690 V with its own harmonic filters, circuit
breaker, measurement and control boards. The design also
includes the step-up transformer in the nacelle to decrease
the cable cost and losses [195]. The works in [131], [144],
[197] also investigated the use of different number of con-
verter channels with the open-winding generators. In
comparison to the parallel VSC modules discussed earlier
in Section V, the circulating currents in the generator-side
converters are eliminated. As a result, the power rating of
the overall system increases, in proportion to the number
of channels employed. The distributed windings also pro-
vide insulation between the converters. These generators
are specially designed at a higher cost. To minimize the
circulating current in grid-side converters, open-windings
transformers can also be used.
2) Cascaded Converters: A high-power wind turbine
using cascaded converters is shown in Fig. 28. This

converter configuration is similar to the cascaded H-bridge
converter (CHB) used in the electric drives industry. The
converter requires isolated dc sources which are generated
by the open winding generator. The generator contains
multiple sets of two-phase windings with a phase displace-
ment of 90�. When more number of converter cells is
connected in series, the system voltage and power rating
increases. Each module can be realized using the power
converters introduced in Fig. 29 [57], [193], [194]. The
transformerless grid connection is also possible with this
configuration as system voltages of 10–35 kV can be
reached by connecting more modules in series. Moreover,
the LV switching devices can be used in the power con-
verter to achieve MV operation levels. The modularity,
redundancy and fault-tolerant operation are other added
advantages of this configuration [146].
A three-level H-bridge converter is shown in Fig. 29(a).
It is a simple structure and its output contains three levels.
This converter can be realized by mass produced two-level
VSCs [171]. With three cells in series as shown in Fig. 28,

and 3L H-bridge converter, the output voltage contains
7 levels and there by the dv=dt and harmonic filter size
decreases significantly. To achieve the higher system volt-
age operation with a lower number of modules, NPC
H-bridge modules can be used as shown in Fig. 29(b) to
form a 5L-HB converter. In this case, the generator wind-
ings should also be designed for medium voltage operation.
The cost associated with the generator-side active
converters can be decreased by using single-phase diode-
rectifier and boost converter similar to the topology pre-
sented in Fig. 19 [193], [198]. Another possibility is to use
medium frequency transformers (MFTs) in the BTB
H-bridge modules (Fig. 29(c)). This concept has been
Fig. 27. Type 4 WECS with multiphase generator, multiple 2L
VSC modules and open-winding transformer. (Commercial
power converter
configuration).
Fig. 28. Type 4 WECS with multiphase generator and cascaded
2L VSC modules. (Promising power converter configuration.)

proposed in [199] for UNIFLEX-PM project. The dc/dc
converters provide galvanic insulation and thus the grid-
side transformer can be eliminated by connecting more
modules in series as shown in Fig. 28. The MFT operates at
several kHz and there by reduces overall size of the
converter significantly [57]. The large number of compo-
nents decreases the overall reliability of this configuration.
The pros and cons for various converter configurations
discussed in this section are summarized in Table 8 in the
Appendix. The market penetration of these converters is
also highlighted.
V I I I . P O W E R C O N V E R T E R S
W I T H O U T D C - L I N K
The matrix converters (MC) provide direct ac-ac conver-
sion without any intermediate dc-link, leading to more
silicon based conversion with low cost and small foot print
[120], [200], [201]. The cycloconverters also perform di-
rect ac-ac conversion, but their output voltage/frequency is
just a fraction of input voltage/frequency [121]. Compared

to the cycloconverters, the MCs can produce wide ranges
for output voltage/frequency. The MC is able to convert
variable voltage/frequency to fixed voltage/frequency to
connect the wind turbine to grid [59], [119]. Due to the
elimination of dc-link capacitors, it offers reliable solution
for wind turbines, especially when they are employed in
offshore. The summary of comparison between the voltage
source converters and matrix converters is given in Table 7
for MV operation of wind turbines [202]–[204]. Since
there is no intermediate dc-link, the variations in the
generator-side variables strongly affect the grid-side va-
riables and vice versa [205]. The development of proper
control system is crucial to ensure the correct operation of
MC when employed in wind turbines.
A. LV Matrix Converters
The direct matrix converter (DMC) based LV wind
turbine is shown in Fig. 30, where the generator and grid
are rated for LV operation. It employs bidirectional
switches (with common emitter or collector) and CL filter
on the grid-side similar to the current source converters

discussed before. The power flow is bidirectional and thus
this converter can be employed with SCIG, DFIG, WRSG,
and PMSG turbines [119], [206]–[208]. The low voltage
gain is a major challenge for DMCs compared to the VSCs
which lead to poor semiconductor device utilization. The
large number of semiconductor devices and gate drivers is
another drawback of the matrix converter functional. Cur-
rently, the DMCs are used in low power and LV industrial
applications, but not practiced in WECS yet. The use of
indirect matrix converters is also reported in literature for
wind energy application [209].
B. MV Matrix Converters
Another possibility of using MCs for the MV turbines is
to arrange them in a modular way. A MV wind turbine
with nine modules is shown in Fig. 31. The basic power
cell is realized with 3-phase to 1-phase matrix converter
(also called single-phase matrix converter) as shown in
Fig. 32 [210]. To increase the MV operation further more
modules can be connected in series [211]–[213]. This
configuration is called multi-modular matrix converter

(MMMC) or cascaded matrix converter (CMC), and it
offers sinusoidal input and output currents. In addition
to the step-up transformer, this configuration requires
Fig. 29. Sub-module configurations for the cascaded power
conversion system (refer to Fig. 28).
phase-shifting transformer to ensure series connection of
matrix converter modules. The phase-shifting transformer
cancels the lower order harmonics there by the grid cur-
rent quality is improved. The phase-shifting transformers
are more expensive and bulky compared to the standard
transformers.
The semiconductor device voltage rating is fraction of
the system operating voltage, but the component count is
significantly higher. For example, the 9-module MMMC
requires 108 switches in contrast to 72 switches used in
9-module cascaded voltage source converters shown in

Fig. 28. The output waveform contains 7 levels similar to the
one obtained by Fig. 28, therefore the output voltage wave-
form is close to sinusoidal. Due to large number of switches,
the reliability of system becomes lower. This configuration
also requires complex modulation scheme and control sys-
tem. This technology is not being applied in the wind tur-
bines yet, but a 9 module MMMC is available in market for
wind turbine applications (Yaskawa Enewin-MX1) [119].
The advantages and disadvantages of matrix converters
for wind energy conversion application are given in Table 8
in the Appendix.
Table 7 Comparison of Voltage-Source and Matrix Converters
for MV-WECS
Fig. 30. Type 3 and 4 WECS with three-phase to three-phase
direct matrix converter. (Promising power converter
configuration.)
I X . W I N D F A R M C O N F I G U R A T I O N S

As a contrary to the conventional power plants, a group of
wind turbines form the wind power generation units, often
called wind farms. The early wind farms have been on land
to take the advantage of low initial and maintenance costs.
Nowadays, the offshore wind farms are gaining more
attention because the power production can be increased
and stabilized with the help of stronger and steady winds,
the impact on land use and landscapes can be reduced,
audible noise and visual impacts can be mitigated and less
opposition by the ‘‘Not In My Back Yard (NIMBY)’’ move-
ment [14], [17], [31].
The most of the offshore projects (nearly three quarters)
are located in the European countries. The United Kingdom
(U.K.), Denmark, Belgium, Netherlands, Germany, Sweden,
Finland and Ireland are the key players for offshore wind in
Europe [4]. The offshore wind capacity account only 2% of
installed wind power capacity by 2012. The offshore project
proposals and the present trends indicate that by 2020, the
offshore wind power capacity would reach 40 GW [214].

The largest offshore wind farm to date is London Array
with 630 MW installed capacity. The future offshore wind
farms are proposed in the range of 1200–2500 MW. Few
such projects are Blekinge Offshore, Sweden (2500 MW);
Korea Offshore, South Korea (2500 MW); and Moray Firth,
United Kingdom (1300 MW) [13].
A. Overview of Wind Farm Interconnection Methods
The optimal interconnection of wind turbines play a
crucial role in decreasing the cost, and increasing
Fig. 31. Type 4 WECS with multimodular (9 module version)
matrix converter. (Promising power converter configuration).
Fig. 32. Configuration of 3 � 1 matrix converter. (Sub-module
of Fig. 31.)
efficiency, reliability and performance of wind farm [31].
Many configurations have been proposed in literature, but
only few of them have made their way to the practical
implementation. In this section, we will study the practical

and most promising wind farm configurations. In Fig. 33,
four different configurations are shown. These configura-
tions exhibit three distinct features as defined here [15],
[18], [52], [54]:
• Series or parallel connection of WT output terminals,
• Coupling of WT output ac or dc terminals, and
• Connection of wind farm to utility grid by ac or dc
transmission lines.
Fig. 33. Onshore and offshore wind farm configurations.
All the wind farm configurations which have been re-
searched and commissioned till date combine these three
discrete classes in different manner. The parallel connec-
tion of wind turbines increases the current and power
capacity, while series connection increases the voltage and
power rating. The parallel connection is most widely in the
present wind farms, while some works propose that the
series connection could increase the power density and

decrease the transmission losses. However, this configu-
ration causes high power losses in the converters [54].
The ac coupling enables use of all the five types of wind
turbines depicted in Fig. 5 because the WT step-up trans-
former is a common element to all these. This is the most
commonly used configuration with both ac and dc trans-
mission lines. The dc coupling is possible with the Type 4
wind turbines only.
The power to be delivered and distance of wind farm to
the nearby utility grid are two important factors that play a
crucial role in deciding between ac and dc transmission
systems. The high voltage ac (HVAC) systems are favorable
for low power wind farms which are located close to the
utility grid. For power ratings and distances greater than
400 MW and 60 km, the HVDC transmission is the most
preferable choice [38], [215]. The Fig. 33 shows only one
collector point, however in the practical wind farms more
number of collector systems are used to increase the power
handling capacity and also reliability [31].
B. Parallel AC Configuration + HVAC Transmission

The parallel-ac configuration of wind farm with high
voltage ac (HVAC) transmission is shown in Fig. 33(a) [31].
As the name implies, it uses parallel connection, ac cou-
pling and HVAC transmission from the aforementioned
classification. This configuration offers low initial cost for
the wind farm substation. For the present wind turbines,
the most standard output voltages are 690 or 3000 V. The
WT output voltages are converted to medium voltage level
(33 or 34.5 kV) by the wind turbine step-up transformer.
The three-phase output terminals from the step-up trans-
formers are connected in parallel to form a medium voltage
ac (MVAC) collection system. The MVAC is then stepped-
up to HVAC in the range of 60–245 kV by the wind farm
substation [38], [215]. The HVAC transmission lines con-
nect the wind farm substation to the national electric grid.
A step-down transformer may be employed at the receiving
end to connect to the distribution lines.
It should be noted that though the Fig. 33(a) shows only
Type 4 wind turbine, all other wind turbines depicted in

Fig. 5 can be used to form the wind farm because the step-
up transformers are common elements in all these con-
figurations. The earlier offshore wind farms were located
close to the shore and they used same technology as on-
shore farms shown in Fig. 33(a) along with submarine
three-phase ac cables. An example offshore wind farm em-
ploying this interconnection approach is Horns Rev,
Denmark with a capacity of 160 MW. This interconnection
approach is associated with few disadvantages. To improve
the transmission efficiency, additional reactive power com-
pensators such as static compensator (STATCOM) or static
VAr compensator (SVC) should be connected at both-sides
of HVAC transmission lines [16], [22]. The faults on the
HVAC lines adversely affect the wind farm and vice versa.
C. Parallel AC Configuration + HVDC Transmission
The parallel-ac configuration of wind farm with HVDC
transmission is shown in Fig. 33(b) [216]. This is the most
promising interconnection approach for far located
offshore wind farms with larger power capacity [15]. The
MVAC of WT collection system is first converted to HVAC

High-Power Wind Energy Conversion Systems: Emerging Technologies and State-of-the-Art

High-Power Wind Energy Conversion Systems: Emerging Technologies and State-of-the-Art

Recommended

Recommended

More Related Content

Similar to High-Power Wind Energy Conversion Systems: Emerging Technologies and State-of-the-Art

Similar to High-Power Wind Energy Conversion Systems: Emerging Technologies and State-of-the-Art (20)

More from donnajames55

More from donnajames55 (20)

Recently uploaded

Recently uploaded (20)

High-Power Wind Energy Conversion Systems: Emerging Technologies and State-of-the-Art