14.

A Master-Slave Fuzzy Logic Control Scheme

for Maximum Power Point Tracking in Wind
Energy Systems
Joanne Hui, Member, IEEE, Alireza Bakhshai, Senior Member, IEEE, and Praveen K. Jain, Fellow,
IEEE

power and perturb the operating point accordingly by

Abstract— A novel master-slave fuzzy logic control (FLC) generating a new reference signal and controlling the system
scheme for maximum power point tracking (MPPT) is proposed via conventional controllers such as the proportional integral
in this paper for wind energy conversion power system. The (PI) control. It has been pointed out in various literatures [7]-
master fuzzy logic controller tracks the maximum power point
(MPP) by generating appropriate generator speed references for
[8] that fuzzy logic control (FLC) is more robust over the
the slave fuzzy logic engine (FLE). The slave FLE then forces the conventional perturb and observe (P & O) control. As a
system to operate towards the dictated speed reference. The result, various FLC based schemes have been documented [7]-
slave communicates to the master FLE whenever a reference has [12].
been reached. Each reference is generated only when the The MPPT control scheme presented in this paper consists
previous speed reference has been reached; as a result, actions of two fuzzy logic engines (FLE). The first FLE acts as an
due to measurement inaccuracies are avoided. In addition, the
proposed scheme can distinguish between atmospheric change
overall master controller for moving the system to the
and intentional system perturbation, without the aid of a wind maximum power point (MPP) by generating the appropriate
speed sensor. Detailed explanations of the proposed control logic speed references for the slave FLE. The slave FLE is the
are discussed and illustrated in this paper. Simulation results speed controller that forces the system towards the dictated
have verified the proposed concept and are provided in this speed reference. FLC is widely used in motor speed control
paper. due to its robustness and accuracy. To avoid discrepancies in
Index Terms—Renewable energy, wind, solar, hybrid system,
the MPP determination logic, the master-slave configuration is
maximum power point tracking, power management
used. Each time power and speed change measurements are
I. INTRODUCTION made by the master FLE, the system is guaranteed by the slave
FLE to have reached the previous set point; the master FLE
As energy sources are depleting due to the energy needs of will not be accidentally triggered (e.g. overshoots: speed
the world, alternative energy sources are being sought after. reference has not been achieved). In addition, this structure
In particular, renewable energy sources, such as wind, are also allows the system to determine whether the change in
being placed in the spotlight because of their inexhaustible power is due to a change in wind speed or due to a change in
and environmentally friendly nature. There are two main operating point. The master FLE will detect power change
types of wind energy systems: fixed-speed and variable speed. due to wind speed by observing the change in power and the
Although fixed-speed systems are robust and are cost change in reference speed. If the master did not signal a
effective, the unpredictable nature of wind causes these change in reference and the power has changed (the slave
systems to waste potential power that could have been constantly monitors the system speed and maintains it at the
otherwise extracted by their variable-speed cousins. For each designated speed), then there has been a change in speed. A
wind speed, a turbine has a different operating point (dictated positive power change indicates the wind speed has increased
by the relationship between the wind speed and the turbine and a negative power change indicates the wind speed has
rotational speed) where the system extracts maximum power decreased.
from the wind. For this reason, it has been reported that
variable-speed wind systems, coupled with control
mechanisms (mechanical or electrical control II. WIND ENERGY SYSTEM WITH THREE-PHASE BOOST
methods/algorithms) are capable of boosting the annual RECTIFIER
energy capture [1]. Fig. 1 represents the wind energy conversion system used
Various maximum power point tracking (MPPT) algorithms for the verification of the algorithm. A three-phase boost
for wind energy systems have been presented in literature [2]- rectifier is used to simplify the control process and thus allows
[6]. The algorithms presented in [2]-[3] observe the change in easy verification of the algorithm [13, 16]. As for the
generator, a permanent magnet synchronous generator

978-1-4244-3384-1/10/$25.00 ©2010 IEEE

 14.3
(PMSG) is used due to its high efficiency, small size and no
slip rings are necessary [13, 16]. In Fig. 1, ωgen is the
generator angular speed; d is the duty ratio, Vdc and Idc are the
average voltage and current of the load respectively. The
MPPT control in this system is therefore obtained by changing
the duty cycle of the switch of the boost converter.
The use of the boost converter in Fig. 1 also allows power
factor correction (PFC) to be achieved at the output terminals Fig 2: Power stage of the wind energy system
of the PMSG. As discussed in [16]-[19], PFC is essential to
provide high quality sinusoidal voltage and current waveforms Vdc ⎛ dTs ⎞
C= ⎜ ⎟ (1)
at the output of the generator to minimize both the copper and
core losses of the generator. In this design, the boost
Δv ⎝ R ⎠
converter is operated in discontinuous conduction mode TsVdc
(DCM) to minimize the number of control loops required in I OB = d (1 − d ) 2 (2)
the control design. 2L
When the boost converter operates in DCM, a filter is
required at the input of each phase of the diode rectifier to 1
ωc = i = a , b, c (3)
remove the high frequency content in the boost inductor Li C i
current. The resultant combination of the filter, diode rectifier
and the boost converter is known as a single-switch three-
III. MPPT IN WIND ENERGY SYSTEM
phase boost rectifier [16, 17], which is illustrated by Fig 2. In
order to design a proper DC-link capacitor (C) that ensures the In a wind energy system, the theoretical mechanical power
Vdc voltage ripple to be within the designed ripple voltage extracted by a wind turbine is governed by (4), where ρ, A,
range, the value of this capacitor is chosen according to (1), and vw represents the air density, turbine blade cross sectional
where Ts is the switching period of the boost converter; R is area and wind speed respectively.
the load resistance and Δv is the ripple voltage of the DC-link
1
capacitor. As for the boost inductor (L), it should be designed pm = ρAv w3 (4)
according to the DCM boundary operating condition as given 2
by (2). The input filter components (i.e. Lf,a, Lf,b, Lf,c, Cf,a, Cf,b, The aerodynamic efficiency of a wind turbine is represented
Cf,c) are then designed according to the cut-off angular by the power coefficient function, Cp(β, λ), where β is the
frequency of the input filter as shown in (3). pitch angle and λ is the tip-speed ratio (TSR). The pitch angle
Since the main objective of this paper is to develop a MPPT is the angle at which the blade is twisted along its longitudinal
algorithm rather than optimizing the whole wind energy axis. The TSR is the ratio between the wind speed and the
conversion system, details regarding the design optimization
rotor speed, and is represented by (6), where R and ω
aspects of the three-phase boost rectifier are not provided in
represent the rotor blade radius and the turbine angular
this paper. In the following sections, the mechanism behind
rotational speed respectively.
MPPT for typical wind energy system is explained. Then the
detailed descriptions of the proposed FLC algorithm will be
provided. Actual _ measured _ power pw
C p (λ , β ) = = (5)
Theoretica l _ power _ in _ wind 1
ρAv w3
2

Rω
TSR = λ = (6)
vw
The power coefficient function is a non-linear curve that
illustrates power conversion efficiency of a wind turbine. Fig.
3 shows the typical shape of the power coefficient (Cp) curve
for a horizontal-axis fixed pitch wind turbine. It can be seen
from Fig. 3 that there is a particular optimal TSR value that
will result in the highest power conversion efficiency (this
Fig. 1: General diagram of wind energy system with boost converter and fuzzy
logic control occurs at Cp,max). Given Eq. (6), it implies that there is a
unique angular rotational speed (ω) for each vw that results in
the optimal TSR value. By taking into the account of the
turbine’s aerodynamic efficiency, the actual power extracted
from the wind (pw) can be expressed by (7).
 14.3
fluctuating wind conditions, the algorithm is able to quickly
1
pw = ρAC p (β , λ )vw3 (7) determine and store the optimum operating points of the
2 system. Fig. 5 illustrates the complete system with the
If one wind speed is considered at a time, the extracted proposed master-slave FLC configuration.
wind power is dependent mainly on the Cp function since the
rotor swept area and air density can be considered constant.
Therefore, the shape of the power curve for each wind speed
has a shape similar to that of power coefficient function. This
leads to the observation that for each wind speed, there is a
particular TSR that results in maximum power. Fig. 4 is an
illustration of a typical power curve of a fixed-pitch
horizontal-axis wind turbine system.

Fig. 5: Overall system diagram with proposed master-slave fuzzy logic control

In Fig. 5, the command signal for the slave FLE is the

reference speed determined by the master FLE. The slave
FLC rule matrix and the membership functions are illustrated
by Figs. 6 and 7 respectively. The two input parameters are
Fig. 3: Typical power coefficient curve for a horizontal-axis fixed-pitch wind
turbine the speed error (current speed – reference speed) and the rate
of change of the speed error. The boundary values for the
speed error (B1, and B2) can be adjusted to obtain the desired
accuracy level of the speed control. The speed error has five
different values to achieve good speed regulation (negative
big (NB), negative small (NS), zero (Z), positive small (PS)
and positive big (PB)). In order to minimize the complexity
of the control, the rate of change of the speed error uses three
values in the logic (negative (N), zero (Z), and positive (P)).
The slave FLC operational logic is as follows:

Fuzzification:

1. According to input (speed error, and rate of change) assign

membership values (MV) based on the membership functions.
Fig. 4: Typical power curve of a fixed-pitch horizontal-axis wind turbine 2. Determine the rule strengths (which rules (1-15) should
fire) as per the rule matrix.
To track the MPPT of the wind system, there are many
Examples:
existing methods [13]-[15]. Many of the strategies for wind
- Rule 1 (R1) = min(MV of speed error to NB, MV
MPPT are variations of the P&O method due to the wind’s
largely unpredictable behavior. The essence of the P&O of rate of change to NB)
method is to measure the output power of the wind and - Rule 2 (R2) = min (MV of speed error to NS, MV of rate
change the loading of the system (through changing the output of change to NB)
current) to increase/decrease the rotor speed.
Defuzzification:

IV. MASTER-SLAVE FUZZY LOGIC CONTROL SCHEME 3. For each particular action (increase big (IB), increase small
OPERATION (IS), no change (X), decrease small (DS), decrease big (DB)),
determine the weight of each action using the “Root Sum
To overcome the drawbacks seen in the P&O MPPT control
Square” approach using the appropriate rules (see rule
method in wind energy systems, a novel MPPT control
matrix).
algorithm implemented with FLC is developed. The main task
Example: IB Æ Rules 1, 6, 11 Æ R 1 2 + R 6 2 + R 11 2
of the proposed control algorithm is to ensure that under
 14.3
the slave signals that the reference has been reached, the
4. Determine output value based on the action weights master determines if there is a change in wind speed
determined in step 3 and the centriod value (or weight) of (reference speed is the same, yet there is a change in power),
each action in the output membership function in Fig. 7. The or to adjust the reference as per the rule matrix and
final output value is calculated by: membership functions. The new reference speed is calculated
using the output of the FLC logic, the change in the system
( IBcentriod ) * IB + ( IS centriod ) * IS + ( Z centriod ) * Z + ( DS centriod ) * DS + ( DBcentriod ) * DB power, and the current reference speed.
IB + IS + Z + DS + DB

5. If the speed error is within a defined threshold, then the

system has reached the reference. Signal to the master FLE to
determine next speed reference.

speed error

NB NS Z PS PB
1 2 3 4 5
N IB Is Ds Ds DB
rate of
change of 6 7 8 9 10
speed error Z IB Is X Ds DB Fig. 8: Master FLE Rule Matrix

11 12 13 14 15

P IB Is Is Ds DB

(a) (b)

Fig. 6: Slave FLE Rule Matrix

(c)
Fig. 9: (a, b) Master FLE input membership function, (c) Output
membership function

V. SYSTEM PERFORMANCE
To verify the functionality of the proposed control
algorithm, the system that consists of a wind turbine model,
(a) (b)
PMSG, diode rectifier, a boost converter and the proposed
FLC algorithm are simulated in PSIM 9.0. Table I lists the
design specifications of the design example. Fig. 10
illustrates the performance of the proposed control scheme
under a fixed wind speed of 6.5m/s. The designed generator
speed for the turbine that results in maximum power capture is
2518 rpm and the system was able to successfully obtain the
generator speed reference of 2481 rpm (actual system
(c) generator speed: 2452 rpm).
Fig. 7: (a, b) Slave FLE input membership function, (c) Output
membership function Table I. Design specifications of the design example
The operational logic for the master FLE is similar to the Number of poles in PMSG: 4
slave FLE with a few modifications. The rule matrix and the Rated Power: 20kW
membership functions (input and output) are given by Figs. 8 Switching frequency: 5 kHz
and 9 respectively. The inputs to the master FLE are: the
system power, current speed reference, and slave signal. Once
 14.3

Fig. 10: System performance wind speed step change at 20s from 6.5 m/s
to 7 m/s: Fig. 12: System performance with wind speed step change at 26s from 6.4
(Top): Speed Reference and Generator Speed (rpm); m/s to 7.7 m/s:
(Bottom): Master FLE Decision (0: waiting for slave signal, 2: wind speed (Top): wind speed value (m/s);
increase detected, 3: wind speed decrease detected, 4: normal FLC operation) (Bottom): Speed Reference and Generator Speed (rpm)

Fig. 11 illustrates the performance of the algorithm under a

sudden wind speed change from 6.5 m/s to 7 m/s, and it was The resultant current waveform of the system is illustrated by
able to detect the change in wind speed. The last reference for Fig. 13. It can be observed that the input filter decreases the
6.5 m/s was 2561 rpm. The designed optimal speed was 2707 harmonic content of the generator output current significantly.
rpm and the system was able to obtain a speed (both reference A high quality sinusoidal current waveform can be observed
and actual) of 2680 rpm. Fig. 12 shows the response of the from Fig. 13. Fig. 14 shows the DCM boost inductor current.
generator speed when the wind speed is step changed from
6.4m/s to 7.7m/s. It can be observed that the generator speed
is able to again follow very close to the reference signal.

Fig. 13: System with filter: Permanent magnet synchronous generator line
current (phase A)

Fig. 11: System performance wind speed step change at 20s from 6.5 m/s to 7
m/s:
(Top): Speed Reference and Generator Speed (rpm);
(Bottom): Master FLE Decision (0: waiting for slave signal, 2: wind speed
increase detected, 3: wind speed decrease detected, 4: normal FLC operation)

Fig. 14: DCM boost inductor current

 14.3
VI. CONCLUSION [16] F. S. dos Reis, K. Tan, and S. Islam. “Using PFC for Harmonic
Mitigation in Wind Turbine Energy Conversion Systems” The 30th
In conclusion, a novel master-slave FLC scheme has been Annual Conference of the IEEE Industrial Electronics Society (IECON)
presented in this paper for MPPT application in wind energy 2004. Nov. 2004. pp. 3100-3105.
[17] Prasad, A.R.; Ziogas, P.D.; Manias, S.; “An active power factor
systems. The proposed master-slave fuzzy logic MPPT
correction technique for three-phase diode rectifiers” IEEE
control was then combined with a three-phase boost rectifier Transactions on Power Electronics, vol. 6, no. 1,1991, pp. 83 – 92.
to form the complete wind energy conversion system. [18] IEEE Std. 519-1992, IEEE Recommended Practices and Requirements
Simulation results have verified that the scheme is capable of for Harmonic Control in Electrical Power Systems.
[19] Adegas, F. D., Villar, J. A., Tonkoski, R., da Silva Simioni, G. C. and
forcing the system towards the MPP and can distinguish dos Reis, F. S., “Maximum power point tracker for small wind turbines
between changes due to intentional system perturbation and including harmonic mitigation” in Proceedings of the 2006 European
changes due to atmospheric change (wind change). Wind Energy Conference

REFERENCES
[1] Yongqiang Lang, Bin Wu, Navid Zargari; “Wind Power Technology-
State of the Art” IEEE Canadian Review, December 2007, no. 56, pp.
10-17.
[2] Koutroulis, E.; Kalaitzakis, K.; “Design of a maximum power tracking
system for wind-energy-conversion applications” IEEE Transactions on
Industrial Electronics, Vol. 53, no. 2, April 2006 pp. 486 – 494
[3] Sinha, A.; Kumar, D.; Samuel, P.; Gupta, R.; “A two-stage converter
based controller for a stand alone wind energy system used for remote
applications” in Proceedings of the IEEE 2008 Telecommunications
Energy Conference, 2008. INTELEC 2008. pp. 1 – 5
[4] Patsios, C.; Chaniotis, A.; Rotas, M.; Kladas, A.G.; “A comparison of
maximum-power-point tracking control techniques for low-power
variable-speed wind generators” in Proceedings of the 2009 IEEE
Advanced Electromechanical Motion Systems & Electric Drives Joint
Symposium, 2009 pp. 1 – 6
[5] Abo-Khalil, A.G.; Dong-Choon Lee; “MPPT Control of Wind
Generation Systems Based on Estimated Wind Speed Using SVR” IEEE
Trans. On Industrial Electronics, Vol. 55, no. 3, March 2008, pp. 1489 –
1490
[6] Mirecki, A.; Roboam, X.; Richardeau, F.; “Comparative study of
maximum power strategy in wind turbines” Proceedings of the 2004
IEEE International Symposium on Industrial Electronics, pp. 993 - 998
[7] Abo-Khalil, A.G.; Dong-Choon Lee; Jul-Ki Seok; “Variable speed wind
power generation system based on fuzzy logic control for maximum
output power tracking” in Proceedings of the IEEE 2004 Power
Electronics Specialists Conference. PESC 04. pp. 2039 – 2043
[8] Simoes, M.G.; Bose, B.K.; Spiegel, R.J.; “Design and performance
evaluation of a fuzzy-logic-based variable-speed wind generation
system” IEEE Trans. On Industry Applications, Vol. 33, no. 4, July-Aug.
1997, pp. 956 - 965
[9] Hilloowala, R.M.; Sharaf, A.M.; “A rule-based fuzzy logic controller for
a PWM inverter in a stand alone wind energy conversion scheme” IEEE
Trans. On Industry Applications, Vol. 32, no. 1, Jan.-Feb. 1996 pp. 57 –
65.
[10] El Mokadem, M.; Courtecuisse, V.; Saudemont, C.; Robyns, B.; Deuse,
J.; “Fuzzy Logic Supervisor-Based Primary Frequency Control
Experiments of a Variable-Speed Wind Generator” IEEE Trans On
Power Systems, Vol. 24, no. 1, Feb. 2009, pp. 407 – 417
[11] Qingrong Zeng; Liuchen Chang; Riming Shao; “Fuzzy-logic-based
maximum power point tracking strategy for Pmsg variable-speed wind
turbine generation systems” in Proceedings of the IEEE 2008 Canadian
Conference on Electrical and Computer Engineering, pp. 405 – 410
[12] Chen, Z.; Gomez, S.A.; McCormick, M.; “A fuzzy logic controlled
power electronic system for variable speed wind energy conversion
systems” in Proceedings of the International Conference on Power
Electronics and Variable Speed Drives, 2000, pp. 114 – 119
[13] Marques J, Pinheiro H, Grundling H, Pinheiro J, Hey H. “A Survey on
Variable-Speed Wind Turbine System” Proceedings of Brazilian
Conference of Electronics of Power, vol. 1, 2003. pp. 732–8.
[14] R. Datta and V.T. Ranganathan “A Method of Tracking the Peak Power
Points for a Variable Speed Wind Energy Conversion System” IEEE
Transactions on Energy Conversion, vol. 18. no. 1, March 2003. pp.
163-168.
[15] Y.D. Song, B. Dhinakaran, and X.Y. Bao. “Variable speed control of
wind turbines using nonlinear and adaptive algorithms” Journal of Wind
Engineering and Industrial Aerodynamics. Vol. 85, Issue 3, pg 293-308.
April 2000.