228 IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 25, NO.

1, MARCH 2010

A Novel Scheme for Rapid Tracking of Maximum

Power Point in Wind Energy Generation Systems
Vivek Agarwal, Senior Member, IEEE, Rakesh K. Aggarwal, Pravin Patidar, and Chetan Patki

Abstract—This paper presents a novel maximum power point

(MPP) tracking (MPPT) algorithm for grid-connected wind energy
generation systems (WEGS). This is a rapid tracking algorithm
that uses the fact that the value of “β,” an intermediate variable,
especially defined for the purpose, remains constant (=βM P P ) for
a given WEGS at the MPP irrespective of the wind velocity. The
value of βM P P is known in advance. The algorithm works in two
stages. In the first stage, it uses large steps to quickly drive the
operating point to lie within a narrow band with limits βm a x and
βm in . In the second stage, exact MPP is tracked using the “per-
turb and observe” method. No extra hardware or measurements
(sensors) are required compared to the existing algorithms. Hence,
the cost is not increased. Application of the proposed algorithm
to an example WEGS shows that the time taken by the system to
reach MPP is much smaller compared to most of the existing algo-
rithms. A prototype matrix converter has been developed for grid
interfacing and the proposed MPPT scheme has been implemented
in conjunction with Venturini and space-vector-modulation-based
switching schemes. All the results of this study are presented.
Index Terms—Matrix converter (MC), maximum power point
tracking (MPPT) algorithm, space vector modulation (SVM),
squirrel cage induction generator (SCIG), Venturini, wind energy
generation system (WEGS).
I. INTRODUCTION
HE DEMAND for electric energy is increasing rapidly.
T Since the conventional fuels are depleting fast and their
prices are going up, the attention has shifted to nonconventional
energy sources, like wind, solar, fuel cell, etc. In this context,
wind is a particularly attractive option. Electric energy is gener-
ated from wind using a wind turbine and an electric generator.
The generated energy can be used either for standalone loads or
fed into the power grid through an appropriate power electronic
interface, such as a matrix converter (MC).
Different types of electric generators are used for the gener-
ation of electric energy from wind. These include the squirrel
cage induction generator (SCIG), the doubly fed induction gen-
erator (DFIG), and the synchronous generator (SG) [1]–[3]. Out
of these, the SCIG is most commonly used because of several
advantages it offers, viz., it is robust, economical, involves low
maintenance cost, and is easy to control [4]. The work reported
in this paper is based on SCIG.
At a particular wind velocity, the amount of power generated
by the turbine depends upon the speed of the turbine, turbine
parameters, and the air density. The air density is usually as-
Manuscript received December 28, 2007; revised February 11, 2009. First
published December 8, 2009; current version published February 17, 2010.
Paper no. TEC-00501-2007.
The authors are with the Applied Power Electronics Laboratory, Department
of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai
400 076, India.
Digital Object Identifier 10.1109/TEC.2009.2032613
0885-8969/$26.00 © 2009 IEEE
Fig. 1. Turbine power versus turbine speed for different wind velocities.
sumed to be constant. Turbine parameters are determined by its
design and are constant. Therefore, for a fixed blade pitch angle
turbine, the output power of the turbine is mainly dependent
upon the turbine speed. Fig. 1 shows the nonlinear power–speed
characteristics of a turbine. The characteristics shift as the wind
velocity (VW ) varies. Each power–speed curve is characterized
by a unique turbine speed (ωr∗ ) corresponding to the maximum
power (P ∗ ) point (MPP) for that wind velocity [5]. This ef-
fectively means that for a given wind velocity, if the turbine
is rotated at (ωr∗ ), maximum power can be extracted from the
wind.
Conventionally, the energy from the wind is extracted by
using a constant speed wind energy generation system (WEGS).
The extracted energy is converted into electric energy by using
an SCIG or DFIG and is supplied to the grid or a standalone load.
The main drawback of this system is its poor efficiency because
it cannot track the MPP [6], [7] as the wind velocity changes.
This situation is depicted by segment T–V–Q–U in Fig. 1. Let
the constant speed system be set to correspond to MPP “Q”
for a wind velocity of 9 m/s. This would result in the system
running at points U, V, and T for other wind velocities, which are
far away from the actual MPP points P, R, and S, respectively,
for the corresponding wind velocities. With the advent of high
speed, high power converters, variable-speed operation of the
WEGS has now become possible and the system can be made
to run at a speed corresponding to MPP for the current wind
velocity, i.e., the system, represented by Fig. 1, can run at P, Q,
R, and S. The amount of energy captured from the wind in this
case is much higher than a fixed speed system.
Several MPP tracking (MPPT) algorithms have been pro-
posed in the past [5]–[10], such as perturb and observe (P&O),
anemometer-based method, calculation-based method, fuzzy-
logic-based scheme, etc. In the P&O algorithm, the turbine
speed is varied in small steps and the corresponding change
in power is observed. Step changes are effected in a direction

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
AGARWAL et al.: NOVEL SCHEME FOR RAPID TRACKING OF MAXIMUM POWER POINT IN WIND ENERGY GENERATION SYSTEMS 229

Fig. 2. Block diagram of a typical grid-connected WEGS.

Fig. 3. C P versus tip speed ratio curve.

so as to move toward MPP [7], [9]. This process is continued
until MPP is reached. By using this algorithm, maximum power
corresponding to any wind velocity can be captured. But the
time taken to reach MPP is long and a considerable amount of
power loss takes place during the tracking phase.
In the anemometer-based MPPT algorithm, the wind velocity
is measured and a reference speed for the induction generator
(IG) corresponding to the MPP of the present wind velocity is
set [8]. Although this is a fast MPPT scheme, the overall cost of
the system increases because anemometer is expensive. Fuzzy-
control-based scheme [6] is good, but is complex to implement.
The algorithm proposed by Wang and Chang [10] is independent
of the turbine characteristics and has good dynamic tracking
speed. However, this scheme also results in slow MPPT because
it needs to compute dVdc /dt for its control action.
Use of MPPT is not beneficial for capturing maximum power
in standalone applications [8], [11]. In fact, in this case, an
arrangement is also required to satisfy the reactive power de-
mand of the WEGS. In the grid-connected system, however, any
amount of power generated by the WEGS can be injected into
the grid. Hence, at any wind velocity, the system can be operated
at MPP to maximize the generation and utilization of power [4].
The block diagram of a typical grid-connected WEGS is shown
in Fig. 2.
When an SCIG with power converter [12]–[14] or DFIG with
rotor side control is used, the speed of the IG can be varied
over a wide range by changing the frequency at the generator
terminals [15]. However, this frequency may be different from
the grid. Hence, a power converter is needed to interface the IG
to the grid [16]. In the past, the commonly used configuration of
power converter for the WEGS was the back-to-back connection
of two power converters along with a large capacitor serving as
the dc link [4]. The main disadvantage of this configuration
is the requirement of a bulky capacitor for the dc link, which
also reduces the life of the converter. The MC is an emerging
alternative to the two-stage ac–dc–ac power converter [17], [18].
The MC provides a single stage ac to ac conversion with the
control of output voltage, output frequency and input power
factor. It also eliminates the requirement of the bulky dc-link
capacitor, hence making the system compact [19]. Also, the
MC is inherently a bidirectional power converter.
In this paper, a new and fast MPPT algorithm is proposed,
which is much quicker than most of the existing schemes and
yet, it does not require any extra hardware. The algorithm drives
the system in two stages. In the first stage, large iterative steps are
used to move within a close range of MPP. In the second stage,
conventional P&O method is used to track the exact MPP corre-
sponding to the current wind velocity. Use of MC as the interface
between the grid and the WEGS is investigated. The operating
frequency of MC is governed by the MPPT scheme used. The
proposed MPPT scheme has been tuned in conjunction with
both Venturini and space vector modulation (SVM) switching
schemes. The control logic for the MC is implemented using
Texas Instrument’s DSP (TMS320F2812) [20]. All the details
of this study are presented in the subsequent sections of this
paper.
II. PROPOSED MPPT SCHEME
The output power of the wind turbine is given by the following
equation [4]:
P = 0.5ρ CP AVw3 (1)
where P is the turbine output power (in watts), ρ is the air density
(in grams per cubic meters), CP is the power coefficient (dimen-
sionless), A (πRr2 ) is the cross-sectional area of the turbine (in
square meters), VW is the wind velocity (in meters per second),
and Rr is the radius of the turbine shaft. CP is a function of λ
and θ, and is given by [15]


151
−13.2 e−18.4/ λi
CP (λ, θ) = 0.73 − 0.58 θ − 0.002 θ 2.14
λi
(2)
where
 −1
1 0.003
λi = − 3 (3)
λ − 0.02 θ θ + 1
with
ωr R r
λ= (4)
Vw
where θ is the turbine blade pitch angle, ωr is the turbine rota-
tional speed (in radians per second), and λ is the tip speed ratio.
The parameter CP signifies the component of wind energy,
which is converted to mechanical energy by the wind turbine.
Fig. 3 shows the CP versus tip speed ratio curve. As per
this plot, if the system operates at the peak point of the curve,
irrespective of the wind velocity, the power captured from the
wind is maximum. For this purpose, the turbine speed should

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
 230 IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 25, NO. 1, MARCH 2010

be adjusted in such a way that the tip speed ratio corresponds to

MPP.
To determine the turbine speed corresponding to MPP for a
particular wind velocity, (1) is differentiated with respect to the
turbine speed and equated to zero assuming the air density and
turbine radius to be constants. This yields
dP 1 dCP
∵ = ρAVw3 (5)
dωr 2 dωr
Fig. 4. (a) Power versus β curves for different wind velocities. (b) Turbine
using dCP /dωr = (dCP /dλi )(dλi /dωr ), (5) can be rewritten speed versus β curves for different wind velocities.
as
dP 1 dCP dλi
= ρAVw3 . (6)
dωr 2 dλi dωr
Differentiating (2) with respect to λi , keeping the blade pitch
angle θ constant, yields


dCP −110.23 2028.23 13.43ψ −18.40 /λ
= + − e i (7)
dλi λ2i λ3i λ2i
where ψ = 0.58θ + 0.002θ2.14 + 13.2.
Differentiating (3) with respect to ωr gives
 
dλi Vw Rr η(θ3 + 1) − 0.003σ
= (8)
dωr (Vw η − 0.003Rωr )2
where σ = 0.02(1 + θ3 )θ and η = 1 + 0.00006θ + θ3 .
Fig. 5. Turbine power and speed versus β curves (not to scale). The various
Using (5)–(7), we have operating sectors are shown in different shades.


dP 1 −110.23 2028.23
= ρAVw3 +
dωr 2 λ2i λ3i At this point, a new variable β is introduced, which is defined
 
13.43ψ −18.4 /λ dλi as β = ωr3 P . The value of β corresponding to MPP is given
− e i . (9) by
λ2i dωr
At MPP, dP/dωr = 0. Applying this condition to (9) provides ωr3M P P 476.20
βM PP = = . (17)
the value of turbine speed corresponding to the MPP (ωr M P P ), PM PP ρRr5
as follows: The right-hand term of (17) involves quantities that are con-


Vw 2028.23η + σ(110.23 + 13.40ψ) stant for a particular wind turbine system and are known from
ωr M P P = . (10)
Rr (θ3 + 1)(110.23 + 13.43ψ) + 6.08 the specifications of the turbine. By substituting these values,
βM PP can be predetermined for a given system.
Putting θ = 0, ψ = 13.2, σ = 0, and η = 1 in (10) gives
The turbine output power versus β curves for different wind
Vw velocities are shown in Fig. 4(a). It is observed that the maximum
ωr M P P = × [6.91] (11)
Rr power increases with an increment of the wind velocity, but the
ωr M P P R r value of β at MPP, i.e., βM PP remains constant irrespective of
λM PP = = 6.91. (12) the wind velocity. This is an important observation and forms
Vw
the basis of the proposed MPPT algorithm. Turbine speed versus
Using (2), (3), and (12) yields λi M P P and CP M P P as follows: β curves are shown in Fig. 4(b).
λi M P P ≈ 7.05 (13) Averaging the curves shown in Fig. 4(a) and (b) over the wind
velocity range, results in the set of curves shown in Fig. 5. The
CP M P P = 0.44. (14) entire operating region is divided into three sectors. Sector 1
is further divided into subsectors 1-A and 1-B as shown. It
Substituting the value of VW from (11) and using values of
should be noted that βM PP corresponds to the MPP only at the
CP M P P and A = πRr2 in (1) yields the power corresponding to
junction of sectors 2 and 3. The junction of sectors 1-A and 1-B,
the MPP as follows:

3 although also denoted by βM PP , does not represent MPP. The
R r ωr M P P current operating sector of the system can be identified using the
PM PP = 0.5 × 0.44ρπRr2 . (15)
6.90 slopes of the turbine power versus β and turbine speed versus
β curves. The slopes of both the curves are negative in sector 1
Rearranging (15) results in
and positive in sector 2. In sector 3, the power versus β curve
PM PP = 2.10 × 10−3 ρRr5 ωr3M P P . (16) has a negative slope, whereas the speed versus β curve has a

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
AGARWAL et al.: NOVEL SCHEME FOR RAPID TRACKING OF MAXIMUM POWER POINT IN WIND ENERGY GENERATION SYSTEMS
Fig. 6. Flow chart of the proposed algorithm.
positive slope. Based on these observations (see Fig. 5), a novel,
fast MPPT algorithm is described next.
Proposed MPPT algorithm.
1) Measure the speed of the wind turbine (ωr k ) apart from
voltage, current, and frequency (fk ) at stator terminals of
the IG.
2) If the present frequency (fk ) at the stator terminals is
equal to the reference frequency (Fk −1 ), calculate the
present turbine output power (Pk ), βk , ∆Pk = Pk −
Pk −1 , ∆ωk = ωk − ωk −1 , and ∆βk = βk − βk −1 .
3) Identify the operating sector (see Fig. 5) depending upon
the value of ∆Pk /∆βk and ∆ωk /∆βk . If both are nega-
tive then the sector is 1, if both are positive then the sector
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
231
Fig. 7. Basic block diagram of an MC.
Fig. 8. Basic block diagram of implementation of MC using Venturini scheme.
is 2, and if ∆Pk /∆βk is negative but ∆ωk /∆βk is positive
then the sector is 3.
4) If βk value (i.e., current value of β) lies within the
“βM PP ± ∆β” band and the operating sector is either
2 or 3, set the reference frequency equal to the present
frequency, i.e., Fk = fk .
5) If the current sector is 1 and βk > βM PP , set the refer-
ence frequency Fk = fk + fm in1 . If βk < βM PP , set the
reference frequency Fk = fk + fm in2 .
6) If the current sector is 2 and βk < βm in , set the refer-
ence frequency, Fk = fk + (βM PP −βk )Gf , else set the
reference frequency, Fk = fk + fm in3 .
7) If the current sector is 3 and βk > βm ax , set the refer-
ence frequency, Fk = fk + (βM PP −βk )Gf , else set the
reference frequency Fk = fk − fm in3 .
8) Go to step 1) and start again. In this manner, continuously
track the maximum power at any wind velocity.
The flowchart corresponding to this algorithm is shown in
Fig. 6. The operating point may lie in any of the sectors shown
in Fig. 5. If the operating point is in sector 1, with βk > βM PP
(sector 1-B), the stator frequency is altered by fm in1 to drive
the operating point to MPP. On the other hand, if βk < βM PP
(sector 1-A), stator frequency is altered by fm in2 to attain MPP.
If the condition is βm in < βk < βm ax , a frequency step size
of fm in3 is used. The values of fm in1 and fm in2 are governed
by WEGS parameters and the wind velocity range at a given
location (average wind velocity has been used in this study).
 232
Fig. 9. Simulation results using Venturini scheme. (a) Grid voltage and current (magnified ten times) injected into the grid. (b) Line voltage at SCIG terminals.
(c) Phase voltage at SCIG terminals. (d) Phase current at SCIG terminals.
Fig. 10. (a) Output voltage sector and (b) input current sector.
To begin with approximate values are used. Let point “X” be
the farthest point on the speed versus β curve (see Fig. 5).
Then, fm in1 is chosen so as to drive the turbine speed from
ωr (X ) to ωr (M PP) . Similarly, fm in2 is approximately obtained
by judging the difference between ωr (Y ) and ωr (M PP) . Both
fm in1 and fm in2 might need some fine tuning at the time of
system installation fm in3 is a very fine frequency step change
used to attain the MPP as precisely as possible. If the operating
point lies in sector 2 or 3 such that βk > βm ax or βk < βm in , the
stator frequency change is decided by Gf (βM PP −βk ), where
Gf is defined as the frequency gain factor and is used to reduce
the number of steps required to reach the MPP. The values of
βm in and βm ax are tuned in such a way that the band across the
βM PP is sufficiently narrow so that the time taken to reach MPP
is minimum. At the same time, if (βm ax −βm in ) is too narrow,
the system may oscillate between sectors 2 and 3 before entering
the (βm ax −βm in ) band. Hence, there is a tradeoff between the
stability of the system and the MPPT speed while selecting
the values of βm in and βm ax . It is practically impossible for the
system to attain the exact βM PP point. Hence, some computation
error tolerance must be provided to the controller. ∆β represents
this small error tolerance around βM PP .
III. SIMULATION OF WEGS
The speed of the wind turbine is usually very low. Therefore,
if the turbine is directly connected to the IG, then the frequency
at the IG terminals will also be very low and for a given power
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 25, NO. 1, MARCH 2010
output, the current at the IG terminals will become very large.
Due to this reason, the turbine shaft is connected to the IG
through a gear box, with the gear ratio adjusted in such a way
that maximum power can be extracted over a wide range of wind
velocity with reasonable frequency at the IG terminals.
The whole system, comprising of the wind turbine, gear box,
SCIG, and an MC is simulated in the MATLAB software. The
simulation is done for a 15-kW WEGS with the following pa-
rameters: for wind turbine: ρ = 1.225 kg/m3 and Rr = 2.5 m;
for IG: rs = 0.2147 Ω, rr = 0.2205 Ω, Ls = 0.991 mH, Lr =
0.991 mH, Lm = 0.06419 H, P = 4, and J = 0.102 kg·m2 . The
wind turbine, which is mechanically coupled to the SCIG, is
simulated using (1), (2), (3), and (4). Simulation of the SCIG
is done using the d–q model [21], [22]. Each block is modeled
separately in MATLAB/Simulink. Special computations and al-
gorithms, such as the MPPT algorithm, are written as MATLAB
programs (functions). These functions are invoked by the main
Simulink program whenever required.
To extract maximum power corresponding to a given wind
velocity, the frequency at the terminals of the IG is adjusted in
such a way that IG runs at a speed corresponding to the MPP.
To interface the variable frequency terminals of the IG to the
fix frequency grid, an MC (see Fig. 7) is connected between
the SCIG and the grid. The implemented MC consists of nine
bidirectional switches with each input phase connected to each
output phase through a bidirectional switch.
The control of MC is implemented using Venturini algorithm
and SVM technique discussed later in this paper. Both the ends
of the MC can operate at a different frequency and different
voltages. One end of the MC operates at a frequency required
to operate the wind turbine at MPP, while the other end of the
MC is connected to the grid that operates at grid frequency. The
reference frequency input to the MC is decided by the MPPT
algorithm. The MC is simulated in MATLAB/Simulink software
using bidirectional switches (devices).
The simulated MC was controlled using both the Venturini
and the SVM schemes. These two schemes for MC control are
briefly described next.
AGARWAL et al.: NOVEL SCHEME FOR RAPID TRACKING OF MAXIMUM POWER POINT IN WIND ENERGY GENERATION SYSTEMS 233

Fig. 11. Simulation results using SVM scheme. (a) Grid voltage and current injected into the grid. (b) Line voltage at SCIG terminals. (c) Phase voltage at SCIG
terminals. (d) Phase current at SCIG terminals.

Fig. 12. Typical curves during the tracking of the MPP using the proposed MPPT algorithm.

Fig. 13. Comparison of MPPT speeds of the proposed scheme and the con-
ventional P&O method. In the example considered, the MPP is approximately
5.3 kW for the given wind velocity.

A. Venturini Scheme
Fig. 14. (a) Laboratory prototype of the MC for grid interfacing. (b) Realiza-
The block diagram of the implementation of Venturini algo- tion of a bidirectional switch. Nine such switches are used in the MC shown
rithm is shown in Fig. 8. in (a).

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
 234 IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 25, NO. 1, MARCH 2010

Fig. 15. Experimental results using Venturini strategy. (a) Grid voltage and current injected into the grid. (b) Line voltage at SCIG terminals. (c) Phase voltage
at the SCIG terminals. (d) Phase current at SCIG terminals (maximum power of 240 W injected into the grid corresponding to a wind velocity of 8 m/s).

Fig. 16. Experimental results using SVM strategy. (a) Grid voltage and current injected into the grid. (b) Line voltage at SCIG terminals. (c) Phase voltage at
SCIG terminals. (d) Phase current at SCIG terminals (maximum power 270 W injected into the grid corresponding to a wind velocity of 8 m/s).

The duty cycle of each switch is given by [23] Tα δ π

 dα δ = = mv sin − θv · sin (θi ) (20)
Tk j 1 2Vk Vj T 3
mk j = = 1+ (18) Tβ γ π
T 3 Vm2 dβ γ = = mv sin (θv ) · sin − θi (21)
T 3
for k = A, B, C, and j = a, b, c.
The results of MC for Venturini scheme are shown in Fig. 9. Tβ δ
dβ δ = = mv sin (θv ) · sin (θi ) (22)
T
B. SVM Scheme T0
d0 = = 1 − dα γ − dα δ − dβ γ − dβ δ . (23)
The block diagram of the implementation of SVM scheme is T
same as shown in Fig. 8. The only difference is in the method The results of MC for SVM scheme are shown in Fig. 11.
of calculating the switching sequence and turn-ON time. The The value of β for the example system considered is 3.97
switching sequence and turn-ON time of the different switches from (17). Initially, the wind velocity of the system is assumed
in MC depend upon the present sector of the rotating input to be 8 m/s and the algorithm tries to move the system toward
and output vectors [24], [25]. Fig. 10 shows the output voltage MPP. The time taken by the system to reach MPP is nearly 3 s.
and input current sectors in their corresponding space vector At t = 4 s, the wind velocity changes from 8 to 11 m/s in a step-
planes. The duty cycles of the various switches depend upon the wise manner [Fig. 12(a)] and the system tends to move toward
position of the output voltage and input current vectors in the the new MPP corresponding to the wind velocity of 11 m/s.
space vector planes. Accordingly, the reference output frequency of the MC changes
The duty cycles are calculated using the sector information with variable steps as shown in Fig. 12(b). The corresponding
and the following equations: change in output power is shown in Fig. 12(d). Fig. 12(c) shows
Tα γ π π the variation of β during the tracking of the MPP. It can be con-
dα γ = = mv sin − θv · sin − θi (19) firmed from Fig. 12(c) that as the operating point tends toward
T 3 3

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
AGARWAL et al.: NOVEL SCHEME FOR RAPID TRACKING OF MAXIMUM POWER POINT IN WIND ENERGY GENERATION SYSTEMS
MPP, the value of β approaches the constant value of 3.97
(for the example system considered), corresponding to the MPP
for any wind velocity. At t = 6 s, the wind velocity changes
from 11 to 8 m/s, and then, again, to 14 m/s at t = 9 s. The
corresponding variations of the different parameters (reference
frequency, power, etc.) during the tracking can be verified from
Fig. 12.
Fig. 13 shows the time taken by the P&O scheme and the
proposed scheme for tracking of the MPP for a wind velocity
of 10 m/s. Initially, the system is running at 20 Hz frequency.
At t = 3 s, the wind velocity changes in a stepwise manner
and both algorithms start tracking the MPP. The time taken by
the conventional P&O scheme to track the MPP is nearly 4 s,
while the proposed scheme takes approximately 1 s. Thus, the
proposed scheme tracks much faster and is able to capture more
energy from the wind during the transient tracking phase.
IV. HARDWARE RESULTS
Hardware implementation of a 1-kW prototype of WEGS
was carried out in the laboratory, as shown in Fig. 14(a). The
following parameters were used:
Wind turbine: ρ = 1.225 kg/m3 , Rr = 1.5 m.
IG: rs = 0.2147 Ω, rr = 0.2205 Ω, Ls = 0.991 mH, Lr =
0.991 mH, Lm = 0.06419 H, P = 4, and J = 0.102 kg·m2 .
The bidirectional switch used for implementing the MC is
shown in Fig 14(b). The switch was realized using two antipar-
alleled power MOSFETs (IRFP450). A power diode (U10A60)
was connected in series with the power MOSFET in both the
paths to increase the reverse voltage blocking capability of the
bidirectional switch.
The wind turbine was simulated using a dc motor. The torque
of the dc motor is controlled to match that of a wind turbine for
a given shaft speed. The armature control is used for the shaft
speeds below rated speed and the field control is used for the
speeds above rated speed.
Besides the main power stage consisting of an MC, Fig. 14(a)
also shows the snubber circuits, driver circuits, and control cir-
cuits. The control and MPPT logic was implemented using the
DSP (TMS320F2812). The voltage level of the output signal
from the DSP is 3.3 V. This signal was fed into a noninverting
amplifier (OP-AMP TL084) to minimize the loading of DSP and
to step up the signal to 5 V level. These noninverting signals are
given to the gate driver IC (HPCL-3120) for driving the device
and to isolate the power and control circuits.
The proposed MPPT algorithm is tested for a wind velocity
of 8 m/s, using both Ventruni and SVM algorithms to drive the
MC. The experimental results are shown in Figs. 15 and 16,
respectively.
V. SUMMARY AND CONCLUSION
A new MPPT algorithm for WEGS has been presented and
implemented on a grid-connected system. Comparison with the
existing schemes, such as the P&O method, shows that the new
MPPT algorithm provides much faster tracking, resulting in
more optimal usage of the source.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
235
An MC has been used for interfacing the WEGS with the
power grid. The MC is a better alternative than the combination
of two back-to-back connected (ac–dc–ac) converters. The MC
facilitates the change of frequency and voltage at the generator
terminals and is able to maintain unity power factor at the grid
terminals. MC also satisfies the reactive power requirement of
the IG. Computer simulations have shown encouraging results
and it is felt that the proposed topology has good potential
for applications in distributed generation as well as standalone
systems. The latter will, of course, need an energy storage source
like a battery to take full advantage of the MPPT scheme.
A notable drawback of the proposed MPPT algorithm, how-
ever, is its dependence on the system parameters. This is the
focus of further research, which is being carried out by the
authors. These results will be presented in a future paper.
REFERENCES
[1] F. Blaabjerg, Z. Chan, R. Teodoreson, and F. Lov, “Power electronics
in wind turbine systems,” in Proc. IEEE Conf. Power Electron. Motion
Control, Aug. 2006, vol. 1, pp. 1–11.
[2] J. G. Slootweg, H. Polinder, and W. L. Kling, “Representing wind tur-
bine electrical generating systems in fundamental frequency simulations,”
IEEE Trans. Energy Convers., vol. 18, no. 4, pp. 516–524, Dec. 2003.
[3] K. Tan and S. Islam, “Optimum control strategies in energy conversion of
PMSG wind turbine system without mechanical sensors,” IEEE Trans.
Energy Convers., vol. 19, no. 2, pp. 392–399, Jun. 2004.
[4] S. M. Barakati, M. Kazerani, and X. Chen, “A new wind turbine generation
system based on matrix converter,” in Proc. IEEE/PES General Meeting,
Jun. 2005, vol. 3, pp. 2083–2089.
[5] J. Yaoqin, Y. Zhongqing, and C. Binggang, “A new maximum power
point tracking control scheme for wind generation,” Int. Conf. Power
Syst. Technol., vol. 1, pp. 144–148, Oct. 2002.
[6] M. G. Simoes, B. K. Bose, and R. J. Spiegel, “Fuzzy logic based intelligent
control of a variable speed cage machine wind generation system,” IEEE
Trans. Power Electron., vol. 12, no. 1, pp. 87–95, Jan. 1997.
[7] R. Datta and V. T. Ranganathan, “A method of tracking the peak power
points for a variable speed wind energy conversion system,” IEEE Trans.
Energy Convers., vol. 18, no. 1, pp. 163–168, Mar. 2003.
[8] G. D. Moor and H. J. Beukes, “Maximum power point trackers for wind
turbines,” in Proc. IEEE 35th Annu. Power Electron. Spec. Conf., 2004,
vol. 3, pp. 2044–2049.
[9] Q. W. T. Tanaka, T. Toumiya, and T. Suzuki, “Output control by hill-
climbing method for a small scale wind power generating system,” Else-
vier Int. J. Renewable Energy, vol. 12, no. 4, pp. 387–400, Dec. 1997.
[10] Q. Wang and L. Chang, “An intelligent maximum power extraction al-
gorithm for inverter-based variable speed wind turbine systems,” IEEE
Trans. Power Electron., vol. 19, no. 5, pp. 1242–1249, Sep. 2004.
[11] Bagen and R. Billinton, “Evaluation of different operating strategies in
small stand-alone power systems,” IEEE Trans. Energy Convers., vol. 20,
no. 3, pp. 654–660, Sep. 2005.
[12] T. Ahmed, K. Nishida, and M. Nakaoka, “Advanced control of PWM
converter with variable-speed induction generator,” IEEE Trans. Ind.
Appl., vol. 42, no. 4, pp. 934–945, Jul./Aug. 2006.
[13] W. L. Chen and Y. Y. Hsu, “Controller design for an induction generator
driven by a variable-speed wind turbine,” IEEE Trans. Energy Convers.,
vol. 21, no. 3, pp. 625–635, Sep. 2006.
[14] R. Cárdenas and R. Peña, “Sensorless vector control of induction ma-
chines for variable-speed wind energy applications,” IEEE Trans. Energy
Convers., vol. 19, no. 1, pp. 196–205, Mar. 2004.
[15] F. Blaabjerg, Z. Chan, and S. B. Kjaer, “Power electronics as efficient
interface in dispersed power generation systems,” IEEE Trans. Power
Electron., vol. 19, no. 5, pp. 1184–1194, Sep. 2004.
[16] A. Miller, E. Muljadi, D. S. Zinger, and H. J. Beukes, “A variable speed
wind power control,” IEEE Trans. Energy Convers., vol. 12, no. 2,
pp. 181–186, Jun. 1997.
[17] T. Kume, K. Yamada, T. Higuchi, E. Yamamoto, H. Hara, T. Sawa, and
M. M. Swamy, “Integrated filters and their combined effects in matrix
converter,” IEEE Trans. Ind. Electron., vol. 43, no. 2, pp. 571–581,
Mar./Apr. 2007.
 236
[18] J. W. Kolar, F. Schafmeister, S. D. Round, and H. Ertl, “Novel three-phase
ac–ac sparse matrix converters,” IEEE Trans. Power Electron., vol. 22,
no. 5, pp. 1649–1661, Sep. 2007.
[19] M. Imayavaramban, A. V. K. Chaithanya, and B. G. Fernandes, “Analysis
and mathematical modeling of matrix converter for adjustable speed ac
drives,” in Proc. IEEE Power Syst. Conf. Expo., 2006, pp. 1113–1120.
[20] D. Ramırez, C. Veganzones, and F. Bl’azquez, “Adaptation of float-
ing point DSP-based technology for small variable-speed wind turbine,”
IEEE Trans. Energy Convers., vol. 22, no. 2, pp. 376–382, Jun. 2007.
[21] J. G. Slootweg, S. W. H. de Haan, H. Polinder, and W. L. Kling, “General
model for representing variable speed wind turbines in power system
dynamics simulations,” IEEE Trans. Power Syst., vol. 18, no. 1, pp. 144–
151, Feb. 2003.
[22] A. Mullane and M. O. Malley, “The inertial response of induction-
machine-based wind turbines,” IEEE Trans. Power Syst., vol. 20, no. 3,
pp. 1496–1503, Aug. 2005.
[23] P. W. Wheeler, J. Rodriguez, J. C. Clare, L. Empringham, and
A. Weinstein, “Matrix converters: A technology review,” IEEE Trans.
Ind. Electron., vol. 49, no. 2, pp. 276–287, Apr. 2002.
[24] J. C. Han and P. N. Enjeti, “An approach to reduce common-mode voltage
in matrix converter,” IEEE Trans. Ind. Appl., vol. 39, no. 4, pp. 1151–
1159, Aug. 2003.
[25] L. Huber, D. Borojevic, X. F. Zhuang, and F. C. Lee, “Design and im-
plementation of a three-phase to three-phase matrix converter with input
power factor correction,” in Proc. Appl. Power Electron. Conf. Expo.,
1993, pp. 860–865.
Vivek Agarwal (S’93–M’93–SM’01) received the
Bachelor’s degree in physics from St. Stephen’s Col-
lege, Delhi University, Delhi, India, the integrated
Master’s degree in electrical engineering from In-
dian Institute of Science, Bangalore, India, and the
Ph.D. degree from the Department of Electrical and
Computer Engineering, University of Victoria, BC,
Canada.
After a brief stint with Statpower Technolo-
gies, Burnaby, Canada as a Research Engineer
during 1994–1995, he joined the Department of Elec-
trical Engineering, Indian Institute of Technology-Bombay, where he is cur-
rently a Professor. His main field of interest is power electronics. He works on
the modeling and simulation of new power converter configurations, intelligent
and hybrid control of power electronic systems, power quality issues, electro-
magnetic interference (EMI)/electromagnetic compatibility (EMC) issues, and
conditioning of energy from nonconventional sources.
He is an Associate Editor of the IEEE TRANSACTION ON POWER ELECTRON-
ICS, a Fellow of IETE, and a Life Member of the Indian Society for Technical
Education.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on May 22,2010 at 23:46:06 UTC from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 25, NO. 1, MARCH 2010
Rakesh Kumar Aggarwal was born in Rajasthan,
India, in 1983. He received the B.E. degree in electri-
cal engineering from M. B. M. Engineering College,
Jodhpur, India, in 2004 and the M.Tech. degree in
power electronics and power systems from Indian In-
stitute of Technology-Bombay in 2007.
He is currently working as an Engineer in R&D
Department at CA, India. His research interests in-
clude power electronics, embedded systems, cryptog-
raphy, computer security, and secure coding.
Pravin Patidar received the B.Tech degree in elec-
trical engineering from Samrat Ashok Technological
Institute, Vidisha, India, in 2004 and the M.Tech. de-
gree from the Indian Institute of Technology-Bombay
in 2007.
He is currently working as a Product and Test
Engineer in Cypress Semiconductor India Pvt. Ltd.
His research interests include power electronics, non-
volatile and volatile memories, microcontrollers, and
reliability of semiconductor devices.
Chetan Patki was born in Mumbai, India, in 1984.
He received the B.E. degree in electrical engineering
from Mumbai University, Mumbai, in 2005 and the
M.Tech. degree in electrical engineering from Indian
Institute of Technology-Bombay (IIT-B) in 2009.
He is currently working with Prof. Vivek Agarwal
as a Project Engineer in the Department of Electrical
Engineering, IIT-B. His technical interests include
power electronics, electric drives, renewable energy,
and power quality issues.