Professional Documents
Culture Documents
Abstract—In this paper, a near-unity-power-factor front-end Since the majority of the literature related to small-scale
rectifier employing two current control methods, namely, aver- wind turbine systems deals with grid-connected systems and the
age current control and hysteresis current control, is considered. analysis presented cannot be extended to stand-alone systems,
This rectifier is interfaced with a fixed-pitch wind turbine driv-
ing a permanent-magnet synchronous generator. A traditional there is a need for study of wind energy as a principal power
diode-bridge rectifier without any current control is used to com- source for isolated sites, i.e., stand-alone wind systems [3].
pare the performance with the proposed converter. Two constant The focus of this work is on the front-end converter design
wind speed conditions and a varying wind speed profile are used for a stand-alone wind energy conversion system (WECS).
to study the performance of this converter for a rated stand-alone Such a system design is relevant to locations where conven-
load. The parameters under study are the input power factor and
total harmonic distortion of the input currents to the converter. tional generation is not practical or connection to the grid is not
The wind turbine generator–power electronic converter is mod- feasible. A very pertinent problem in such systems is that of
eled in PSIM, and the simulation results verify the efficacy of the harmonic generation and poor input power factor. High levels
system in delivering satisfactory performance for the conditions of harmonic distortion in line currents and voltages reduce
discussed. The efficacy of the control techniques is validated with the overall system efficiency and lead to interference with the
a 1.5-kW laboratory prototype, and the experimental results are
presented. interdependent components, equipment malfunction, and high
losses [4]. A number of standards and regulations such as
Index Terms—Average current control (ACC), hysteresis cur- IEEE-519 and IEC 61000-3 have been introduced to determine
rent control (HCC), permanent-magnet synchronous generator
(PMSG), unity-power-factor (UPF) converter. permissible levels of harmonic content in currents that are
injected into the utility grid.
I. I NTRODUCTION The aim of the designed converter is to facilitate high-power-
factor operation and achieve acceptable harmonic content of
Manuscript received June 30, 2011; revised March 2, 2012 and October 17,
2012; accepted May 12, 2013. Date of publication July 10, 2013; date of II. S TAND -A LONE WECS
current version March 17, 2014. Paper 2011-IACC-193.R2, presented at the
2011 IEEE Industry Applications Society Annual Meeting, Orlando, FL, USA, Wind is a highly stochastic energy source. There is also a
October 9–13, and approved for publication in the IEEE T RANSACTIONS strong interdependence between the aerodynamic characteris-
ON I NDUSTRY A PPLICATIONS by the Industrial Automation and Control tics of the wind turbine, the generator’s rotor speed, and the
Committee of the IEEE Industry Applications Society.
The authors are with the School of Electrical and Electronic Engineering, amount of power that can be extracted from the wind. Hence,
Nanyang Technological University, Singapore 639798 (e-mail: aditya1@ntu. it becomes necessary to implement a control method that will
edu.sg; nirn0001@ntu.edu.sg; eamaswood@ntu.edu.sg; hpooi1@e.ntu.edu.sg). enable the extraction of the maximum power from the system
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. under all possible operating conditions. The following sections
Digital Object Identifier 10.1109/TIA.2013.2272606 highlight the major components of a stand-alone WECS.
0093-9994 © 2013 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1422 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 50, NO. 2, MARCH/APRIL 2014
A. Wind Turbine
Wind turbines are classified based on the number of blades
and the axis about which they are mounted. Typically, the three-
blade horizontal-axis wind turbine is preferred due to better
performance as well as the even distribution of variations in
wind speed from the rotors to the drive shaft. It is also capable
of achieving better power coefficient [7].
The output mechanical power of the wind turbine is given by
the cube law equation
coefficient, A is the area swept by the wind turbine rotor (in Owing to the bell-shaped nature of wind turbine power
square meter), Uw is the wind speed (in meters per second). curves, the WTG is operated in the variable speed mode in
The power coefficient is a function of the tip speed ratio λ order to achieve maximum power from the incident wind. The
and the blade pitch angle β. It describes the efficiency of the parameters of the wind turbine and the direct driven PMSG
wind turbine in converting the energy present in wind into employed are provided in the Appendix.
mechanical power. The tip speed ratio λ may be defined as the
ratio of the speed at which the outer tip of the turbine blade is
C. Front-End Power Converter
moving to the wind speed. It is given by the equation
In most existing small-scale wind turbine systems, the pre-
λ = (rωm )/Uw . (2) ferred choice for a front-end converter is a diode-bridge rectifier
because of its inherent simplicity. However, due to their nonlin-
The blade pitch angle β is defined as the angle at which
ear nature, diode-bridge rectifiers inject harmonic components
the wind contacts the blade surface. The expression for power
into the system, leading to an increased total harmonic distor-
coefficient is given as [8]
tion (THD) of input current, which is expressed as increased
Cp (λ, β) = C1 ((C2 /λi ) − C3 β − C4 ) e(−C/λi) + C6 λ. (3) losses due to heating, malfunction of equipment, and reduced
overall efficiency of the system [15], [17].
Here, the coefficients C1 = 0.5176, C2 = 116, C3 = 0.4, Along with input current THD, the PMSG operating power
C4 = 5, C5 = 21, and C6 = 0.0068, and factor is another parameter of interest. If the PMSG operates
at a lagging power factor, it leads to production of reactive
(1/λi ) = 1/(λ + 0.08β) − 0.035/(β 3 + 1). (4) power at the stator terminals, thereby reducing the maximum
usable (real) power that the generator is capable of producing.
It is valid to assume pitch angle to be zero for low to medium
Hence, it is desirable that the front-end converter is capable of
wind velocities [9]. Hence, in this work, β = 0◦ . To achieve
maintaining high power factor and good quality currents, i.e.,
maximum extraction of power from the system, the tip speed
show good current control ability.
ratio and power coefficient have to be maintained close to the
The converter employed in this work, based on a topology
optimal values and varied according to the variations in the
suggested by Mehl and Barbi [18], is shown in Fig. 1. The
wind speed conditions. The mechanical torque of the system
selection of line inductance values (La, Lb, and Lc) plays a
is given by the equation
critical role in maintaining unity-power-factor (UPF) opera-
Tm = Pm /ωm (N · m). (5) tion and is defined as the sum of source inductance and the
transformer leakage reactance [19], [20]. In a transformer-less
system, the value of line inductance is based on the PMSG
stator inductance alone. It employs a diode-bridge rectifier
B. Choice of WTG
with three current-controlled static bidirectional switches. If the
There are many generator types which find application in a bidirectional switch connected to a particular phase is turned
WECS, such as squirrel-cage induction generator, doubly fed on, the corresponding phase is connected to the voltage central
induction generator (DFIG), and the PMSG [10]. Of these, the point, causing a rise of the associated phase current. Turning
last two are most popularly used in WECS. off the switch leads to conduction of the associated diode in
In this paper, a PMSG is used as the wind turbine generator the upper or lower half bridge (depending on the direction of
(WTG) because of its compact size, higher power density, current flow) and, therefore, to a reduction of the phase current.
reduced losses, high reliability, and robustness [9]. The most Thus, the switching of the bidirectional switches enables us
significant advantage is the elimination of the gearbox, in to achieve the possibility of a sinusoidal line current control
comparison to the DFIG. Thus, the wind turbine can be di- (improved power factor). Maswood and Fangrui proposed a
rectly coupled to the generator. Such systems are called direct number of modifications to this topology in [21], where ac-
drive systems and are most suitable for low-speed operating tual variations in load level on the rectifier were taken into
conditions. account.
VENKATARAMAN et al.: EFFICIENT UPF RECTIFIER FOR A STAND-ALONE WIND ENERGY CONVERSION SYSTEM 1423
Fig. 4. Switching operation for phase “a”: Sa is the bidirectional switch for
phase “a,” and D1 and D4 are the upper and lower bridge diodes, respectively.
Fig. 3. Average current controller.
For example,
The dc link has two identical capacitors with a mid-point
1, if ia ≥ 0
connection to the current control circuit, to maintain the voltage sign (ia ) = (8)
−1, if ia < 0.
balance [22]. The bidirectional switches are typically assem-
bled using four diodes and an IGBT, as shown in Fig. 2. The Here, sa , sb , and sc are the switching states for the three
voltage stress across the bidirectional switches in this topology bidirectional switches Sa, Sb, and Sc. “1” is for switch-on, and
is clamped to half of the dc output voltage by the dc-link “0” is for switch-off. An outer dc voltage regulator has been
capacitors, without the need for additional clamping circuits. implemented to maintain the dc-link voltage balance across the
This provides a significant advantage over a single-switch boost capacitors [23]. However, there is a condition which states that,
topology, which suffers from the problem of high-voltage and when half of the dc-link voltage is lesser than the phase voltage,
high-current stresses across the single switch. the high-power-factor condition is not achieved [24].
E. HCC Method
D. ACC Method
In this current control method, the bidirectional switches are
The average current controller employs the ramp comparison
controlled based on the phase current error falling within the
method. The instantaneous current error is calculated and fed
hysteresis band. The choice of the hysteresis band is critical
to a proportional–integral (PI) regulator, as shown in Fig. 3.
in determining the switching states and the shaping of the
The output of the PI regulator is fed to a comparator along
rectifier currents [25]. Fig. 4 shows the switching operation for
with a saw-tooth carrier. If intersections are obtained, the error
phase “a” of the UPF converter employing HCC. The switching
is forced to remain within the band specified by the carrier
signals for the bidirectional switches are as follows, where h is
waveform which is common to all the three phases. The integral
the hysteresis band:
term reduces the steady-state error between the reference and
⎧
actual currents. ⎪ 1, if (ix > 0 and ix > ix∗ − h)
⎨
Referring to Fig. 1, the following equations have been or (ix < 0 and ix > ix∗ + h)
Sx = ∗ (9)
obtained: ⎪
⎩ 0, if (ix > 0 and ix > ix ∗ + h)
⎧ or (ix < 0 and ix < ix − h).
⎨ L(dia /dt) = va − (vAM + vM O )
L(dib /dt) = vb − (vBM + vM O ) (6)
⎩
L(dic /dt) = vc − (vCM + vM O ) III. C ONVERTER T OPOLOGY
where ia , ib , and ic are the inductor currents; va , vb , and vc A traditional wind system typically employs an uncontrolled
are the source voltages from the PMSG; vM O is the voltage diode-bridge rectifier–inverter system. It shows a vast degrada-
of mode M in reference to the neutral point N ; vAM , vBM , tion in performance under varying wind conditions. The line
and vCM are the voltages at nodes A, B, and C referring to current at the input of the traditional rectifier exhibits very high
node M. These voltages can be expressed as follows, where harmonic distortion, contributing to increased losses and poor
sign(ia ), sign(ib ), and sign(ic ) depend on the polarity of power factor at the PMSG–rectifier interface. This implies that
inductor currents: the power extracted by the traditional ac–dc–ac converter is
⎧ lower than the rated power of the wind-turbine–PMSG system.
⎨ vAM = sign (ia ) (1 − sa ) (V dc/2) This highlights the importance of a current control method to
v = sign (ib ) (1 − sb ) (V dc/2) (7) ensure good power extraction, improved quality of current and
⎩ BM
vCM = sign (ic ) (1 − sc ) (V dc/2). voltage waveforms, and high efficiency. Thus, in this work, the
1424 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 50, NO. 2, MARCH/APRIL 2014
Fig. 5. Schematic of the UPF converter in the wind generator system employ- Fig. 6. Schematic of the UPF converter in the wind generator system employ-
ing the ACC method. ing the HCC method.
Fig. 7. Performance parameters of the UPF rectifier using ACC at a rated wind
speed of 12 m/s. (a) Input power factor of the front-end rectifier employing
ACC at a rated wind speed of 12 m/s. (b) FFT of phase “a” current to front-
end rectifier employing ACC at a rated wind speed of 12 m/s. (c) Mechanical,
PMSG, and dc output powers of the system employing ACC at a rated wind of Fig. 9. Wind speed variation and performance coefficient of wind turbine for
speed 12 m/s. (d) DC bus capacitor voltages of the system employing ACC at a system operating with ACC.
rated wind speed of 12 m/s.
Fig. 9 shows the sudden variation in wind speed from 12 to
case and measures 389.64 V. It is lesser than the ideal value 14 m/s and the corresponding variation in the power coefficient
of 395 V. However, this difference is still acceptable. To avoid Cp . It is seen that, initially, the Cp value is 0.474, and after the
redundancy, the capacitor voltage waveforms are not shown for step change in wind speed, it increases to 0.479, which is closer
this wind speed condition. to the optimum value of 0.48.
1426 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 50, NO. 2, MARCH/APRIL 2014
Fig. 11. Performance parameters of the UPF rectifier using HCC at a wind
speed of 14 m/s. (a) Input power factor of the front-end rectifier employing
HCC at a higher wind speed of 14 m/s. (b) FFT of phase “a” current to front-
end rectifier for HCC at a higher wind speed of 14 m/s. (c) Mechanical, PMSG,
and dc output powers of the system for HCC at a higher wind speed of 14 m/s.
the input power factor is high (0.994). Fig. 10(b) shows the FFT
spectrum of phase “a” current. The fundamental frequency is at
36 Hz. The dominant components are close to 180 and 250 Hz,
and their magnitudes are well below 10% of the fundamen-
tal current component’s magnitude. The THD of this current
Fig. 10. Performance parameters of the UPF rectifier using HCC at a rated is 8.26%.
wind speed of 12 m/s. (a) Input power factor of the front-end rectifier employing The wind turbine mechanical power produced is 2.525 kW.
HCC at a rated wind speed of 12 m/s. (b) FFT of phase “a” current to front-
end rectifier for HCC at a rated wind speed of 12 m/s. (c) Mechanical, PMSG, The PMSG produces 2.47 kW, which is transferred entirely
and dc output powers of the system for HCC at a rated wind speed of 12 m/s. to the dc-link output, as can be seen in Fig. 10(c). Fig. 10(d)
(d) DC bus capacitor voltages for HCC at a rated wind speed of 12 m/s. shows the voltages across the two output dc-link capacitors.
The wind speed is increased to 14 m/s. Fig. 11(a) shows that
B. Performance of the System With HCC
the input power factor of the UPF rectifier is 0.994. Fig. 11(b)
Fig. 10(a) shows the input voltage and current of phase “a” of shows the dominant harmonics at 225 and 315 Hz. However,
the UPF rectifier. The quality of current and voltage is good, and their magnitudes are lower than 10% of the fundamental’s
VENKATARAMAN et al.: EFFICIENT UPF RECTIFIER FOR A STAND-ALONE WIND ENERGY CONVERSION SYSTEM 1427
Fig. 12. Wind speed variation and performance coefficient of wind turbine for
system operating with HCC.
TABLE I
C OMPARISON OF P ERFORMANCE PARAMETERS OF THE
WECS W ITH ACC AND HCC M ETHODS
Fig. 15. Laboratory prototype of UPF converter used to compare the ACC and
HCC methods.
front-end diode-bridge rectifier with a current-controlled UPF
TABLE II
rectifier improves the line current quality and increases the real UPF PARAMETERS U SED IN H ARDWARE P ROTOTYPE
power extracted from the PMSG.
Table I lists the key performance parameters of the system
for the two different wind speed conditions with two current
control methods. The comparison yields that, while both current
control methods are excellent in transferring all the generator
power to the load, the ACC method is more effective in achiev-
ing high power factor, consistently above 0.995, while the HCC
method achieves an input power factor of 0.994. The THDs
of the line current for phase “a” of the converter have been
compared. While the ACC method achieves THDs around 7%
and 6%, the HCC achieves values around 8% and 6%.
The ability of the system to extract maximum possible real
power with minimal reactive power at the converter input is
particularly demonstrated. The power that is available at the A. Experiment Results With ACC
input of the front-end rectifier is transferred entirely to the The obtained voltage and current waveforms of the UPF rec-
dc-link output, indicating low converter losses and high effi- tifier employing the ACC technique are shown in Fig. 16(a)–(c).
ciency. The dc-link voltage balance is also excellent. Fig. 16(a) shows the ability of the system to maintain near UPF
For rated load conditions in the variable wind speed opera- at the input terminals (0.994). However, as Fig. 16(c) shows,
tion, the controller is capable of extracting the maximum possi- the voltage tracking is not perfectly achieved, and there is some
ble real power with an efficiency of 97%–99%. The efficiency ripple in the dc-link voltage. This leads to a higher THD (15%)
of power transfer is excellent at the input and output of the than the values obtained in the simulations; however, a slight
front-end rectifier. The ACC achieves a better power factor of degradation in the hardware prototype is expected.
0.996 in comparison to the input power factor of 0.994 for the
HCC. Thus, it is seen that both current control methods are
B. Experiment Results With HCC
extremely well suited for satisfactory operation of this system
under both rated load and variation in wind speed. The hardware results of the UPF rectifier employing the
HCC technique are shown in Fig. 17(a)–(c). Fig. 17(a) shows
the ability of the system to maintain near UPF at the input
VI. H ARDWARE R ESULTS terminals. The converter input current is practically sinusoidal,
in phase with the input voltage and only has a THD of 6%
To validate the simulation results, the performance of the
[Fig. 17(b)]. This leads to high power factor (0.996). The
UPF drive was tested experimentally using a 1.5-kW labora-
dc-link voltage tracking is accurately achieved using HCC. The
tory prototype that can be readily configured to employ HCC
test shows that, while both the HCC method and the ACC
or ACC.
method are able to achieve UPF at the input terminals, the HCC
The hardware prototype was implemented using a Dspace
is easier to implement and is able to track the reference dc-link
DS1103 controller and developed under the MATLAB–
voltage perfectly accurately compared to ACC.
Simulink real-time workshop provided by Math Works envi-
ronment to generate the switching pattern for the front-end
VII. C ONCLUSION
rectifier and was developed with the available components
in the Electric Power Research Laboratory. The prototype is In this paper, a WECS interfaced with a UPF converter feed-
shown in Fig. 15, and the key parameters are given in Table II. ing a stand-alone load has been investigated. The use of simple
VENKATARAMAN et al.: EFFICIENT UPF RECTIFIER FOR A STAND-ALONE WIND ENERGY CONVERSION SYSTEM 1429
Fig. 17. Results obtained from UPF employing the HCC method. (a) Per-
phase ac voltage and current with 20 V/div and 2.5 A/div from the input
Fig. 16. Results obtained from UPF employing the ACC method. (a) Per- terminal. (b) THD of the input ac currents. (c) DC-link voltage V dc.
phase ac voltage and current with 20 V/div and 2.5 A/div from the input
terminal. (b) THD of the input ac currents. (c) DC-link voltage V dc.
Ali I. Maswood (S’85–M’88–SM’96) received the Ooi H. P. Gabriel (S’11) received the B.Eng. de-
B.Eng. and M.Eng. degrees (with first class honors) gree in electrical and electronics engineering from
from Moscow Power Engineering Institute, Moscow, the School of Electrical and Electronic Engineering,
Russia, and the Ph.D. degree from Concordia Uni- Nanyang Technological University, Singapore, in
versity, Montreal, QC, Canada. 2010, where he is currently working toward the Ph.D.
Having taught in Canada for a number of degree in the Electric Power Research Laboratory.
years, he joined Nanyang Technological University, His research interests include multilevel con-
Singapore, in 1991, where he is currently an Asso- verter topologies for solar photovoltaic application
ciate Professor. He is the author/coauthor of nearly and power factor correction techniques for grid-
200 IEEE/Institution of Engineering and Technology connected systems.
publications, chapter author of the Power Electronics
Handbook, and coinventor of two U.S. patents. His field of interest is in optimal
converter design for renewable energy sources and has secured over $2 million
of grants in this sector.
Dr. Maswood is actively involved in the local IEEE Industry Applications
Society/IEEE Power Electronics Society chapter and in the Steering Committee
of the IEEE Power Electronics and Drives Conference.