Modelling and Fuzzy Logic Control of DFIG Based

Wind Energy Conversion Systems

Karim Belmokhtar, Mamadou. L Doumbia and Kodjo Agbossou
Hydrogen Research Institute
Industrial Electronics Research Group
Department of Electrical and Computer Engineering
Université du Québec à Trois-Rivières
C.P. 500 3351 boul. des Forges, Trois-Rivières (Québec) Canada G9A 5H7
[email protected], [email protected], [email protected]

Abstract—This paper proposes the modelling and control of energy system characteristics [8]. The (P&O) method does not
Wind Energy Conversion Systems (WECS) based on Doubly Fed require prior knowledge of the maximum wind turbine power
Induction Generator (DFIG). The fuzzy logic control is used to at different wind velocities and electric machine parameters
improve the extracted wind power. At given wind velocity, the
mechanical power available from a wind turbine is a function of [10], [11]. The (P&O) method is suitable for wind turbines
its shafts speed. Then, the rotor side converter (RSC) is controlled with small inertia, but not for medium and large inertia wind
in the aim to follow the optimal torque for given maximal wind turbine systems, since the (P&O) method adds a delay to
power. A new maximum power point tracking (MPPT) strategy the system control [9]. The (P&O) based strategy is simple;
based on the fuzzy logic controller was developed. Moreover, a however, it could be slow in response to fast changes in wind
new approach to estimate the speed from the measurement of
the power is presented. The effectiveness of the proposed control speed. Consequently, the WECS may not be capable to operate
strategies is validated by theoretical analysis and simulation at the maximum power point all throughout the possible wind
carried out in Matlab/Simulink environment. speed variations. Hence, the benefits of the WECS will not
Index Terms—Wind turbine, DFIG, MPPT, Fuzzy Logic Con- be exploited, despite the increased capital investment due
trol. to the power electronics converter [8]. To implement PSF
control, the controller uses information about the wind turbine
I. I NTRODUCTION maximum power curve (maximum power versus shaft speed).
Wind energy is the source of the renewable energy that The maximum power is then tracked by shaft speeds control.
has been the largest growing rate in the world in the last Wind velocity measurement is not required with this method
two decades, mainly because it is considered to be a green [12]. Usually, experimentally measurement of the power ver-
energy and economically viable [1], [2], [3]. Variable speed sus turbine speed curves is required.
operation of wind turbine is usually used to provide energy
with best efficiency. Wind Energy Conversion System (WECS)
based on DFIG has several advantages [4]. It reduces stresses
of the mechanical structure, acoustic noise reduction and the
possibility to regulate both active and reactive power [5].
Another advantage of the DFIG system is that the back-to-
back PWM converters, connected between the grid and the
induction machine rotor circuit, are sized only for a part of
the full power of the generator (about 30The wind turbine
generators (WTGs) can achieve the maximum wind power Fig. 1. Structure of the DFIG based wind system.
provided at various wind speeds by adjusting correctly the
shaft speed [7]. There are three commonly used Maximum
Power Point Tracking (MPPT) techniques which are: the In the industry, Proportional (P), Proportional Integral (PI)
Wind Speed Measurement (WSM) method, Perturbation and and Proportional Integral Derivative (PID) controllers are
Observation (P&O) method, and Power Signal Feedback (PSF) widely used due to their simple structures and ease of use
method [8]. In the WSM method, both wind velocity and in wide range of applications. However, theses controllers
shaft speed should be measured. Also, the tip-speed-ratio must are tuned for specific linearized models. They may provide
be determined. There are two drawbacks of implementing insufficient performance for different operating point caused
the WSM method [9]. Firstly, obtaining accurate value of by disturbances and parameter uncertainties. This may lead to
the wind velocity is complicated, and using the mechanical undesirable and sustained oscillations in power and voltage.
sensor to measure the shaft speed increases the systems cost. Furthermore, large and slow wind deviations may even lead
Secondly, the optimal tip speed ratio is dependent on the wind to systems instability [1].

978-1-4673-0158-9/12/$31.00 ©2012 IEEE 1888

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:02:58 UTC from IEEE Xplore. Restrictions apply.
 II. M ODELLING AND C ONTROL OF THE DFIG W IND Where Ls , and Lr are respectively the inductances of the
T URBINE stator and rotor windings, and Lm are the mutual inductance.
Fig. 1 shows the basic structure of a DFIG wind turbine. The The electromagnetic torque of the DFIG can be expressed as
DFIG is a wound rotor induction machine which is connected follow:
to the grid through both stator and rotor terminals. The stator
is directly connected to the grid, while the rotor is connected Tem = P (ψds ıqs − ψqs ıds ) (6)
via a PWM back-to-back converter, which consists of rotor Where P is the number of pole pairs.
side converter (RSC) and a grid side converter (GSC) [10]. By neglecting the power losses associated with the stator re-
A. Modelling of the Wind Turbine sistances, the active and reactive stator power can be expressed
as [7]:
The mechanical power Pm captured by the turbine from the
wind for a given wind speed vw is computed by the following Ps = 32 (vds ids + vqs iqs )

expression [13]-[14]: (7)
Qs = 23 (vqs ids − vds iqs )
1 III. C ONTROL OF THE DFIG W IND T URBINE
Pm =ρπR2 Cp Vw3 (1)
2 The control system of the DFIG wind turbine consists of
Where ρ is the air density in (kg/m3); A=πR2 is the area in electrical control of the DFIG and the mechanical control of
m2 swept by the blade, and R the radius of the blade in (m). both blade pitch angle and yaw system. The electrical control
The aerodynamic model of a wind turbine can be determined is achieved by controlling the RSC and the GSC. RSC is
by the Cp (λ, β) curves. Cp is the power coefficient, which is controlled in the aim to regulate independently the stator active
function of both tip speed ratio λ and the blade pitch angle β. power Ps and reactive power Qs . The objectives of controlling
The tip speed ratio is given by: the GSC are to keep constant the dc-link voltage Vdc , and to
regulate the reactive power Qg which is exchanged between
ωr R
λ= (2) the GSC and the grid [10].
Vw
Where ωr represents the rotational speed of the wind turbine A. Control of the RSC
in (rad/sec). In our study, the mathematical representation of Fig. 2 shows the overall vector control scheme of the RSC,
the power coefficient used for a wind turbine is given by: in which the independent control of the rotor active power Ps
and reactive power Qs is achieved by means of rotor current
! regulation. The control of RSC consists of regulating the stator
π(λ − 3)
Cp = 0.398 sin − 0.00394(λ − 2)β (3) active power Ps and reactive power Qs independently [10].
15 − 0.3β In this paper, we assume that both frequency and voltage of
If the wind speed is below the rated value, the WTG grid are constant, and then in our control system, there is one
operates in the variable speed mode, and Cp is keep at its control loop as shown in Fig. 2. Fuzzy logic current controllers
maximum value. In this operating mode, the pitch control is are used to generate the reference signals of the direct and
deactivated. When the wind speed is above the rated value, the quadratic axis current components.
pitch control is activated, in the aim to reduce the generated
mechanical power [7].
1) Modelling of the DFIG: Application of Concordia and
Parks transformations to the three-phase model of the DFIG
allows writing the dynamic voltages and fluxes equations in
an arbitrary d-q reference frame [15]:

v = rs ids + dψdtds − ωs ψqs


 ds

dψ
vqs = rs iqs + dtqs + ωs ψds

(4)
v = rs idr + dψdtdr − ωr ψqr
 dr


dψ
vqs = rs iqr + dtqr + ωr ψdr
Where rs and rr are respectively the resistance of the
stator and rotor windings, and s is the rotational speed of the
synchronous reference frame.
The flux linkages are given by: Fig. 2. Overall vector control of the RSC.


 ψds = Ls ids + Lm idr In the order to achieve a power decoupling control, the
ψqs = Ls iqs + Lm iqr vector control strategy is adopted, with stator field orientation.

(5)

 ψ dr = Lr idr + Lm ids Reactive power Qs and the active power Ps will be propor-

ψqr = Lr iqr + Lm idqs tional to iqr and idr respectively and can be expressed as [7]:

1889

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:02:58 UTC from IEEE Xplore. Restrictions apply.
 vs Φds vs Lm idr
Qs = − (8)
Ls Ls
Lm iqr
Ps = vs iqs = −vs (9)
Ls
Where vs and Φs are respectively the RMS voltage and the Fig. 4. Block diagram of the dc-link voltage regulation.
rated flux of the grid.
B. Control of the GSC
C. Design of the Fuzzy Logic Controller
Fig. 3 shows the overall vector control scheme of the GSC,
in which the control of the dc-link voltage Vdc and the To regulate the DFIG wind turbine, fuzzy logic controllers
reactive power Qg exchanged between the GSC and the grid are used because of the nonlinearity of the system. The fuzzy
is achieved by means of current regulation in a synchronously controller includes three parts: fuzzification, fuzzy rule base
rotating reference frame. and defuzzification as shown in Fig. 5. There are two input
In the d-q reference frame, the active power Pg and reactive signals to the FLC, the speeds error e and the change of the
power Qg flows are respectively expressed as follow: error ce. The error and change of error are converted by gain
factors. To obtain the output of the FLC, the defuzzication used
3 is based on the center of gravity method, and the triangular
Pg = (vgd igd + vgq igq ) (10)
2 membership with overlap is used for the inputs and for the
3 output of each FLC. G1, G2 and G3 are the gain factors.
Qg = (vgd igq − vgq igd ) (11)
2
In the Park reference frame, the q-axis of the reference
frame is aligned with the stator-voltage position vtd . Since
the amplitude of the grid voltage is constant, vgq is constant.
Consequently, the active and reactive power will be propor-
tional respectively to igq and igd .
3
Pg = vgq igq (12) Fig. 5. Structure of the fuzzy logic controller.
2
3
Qg = − vgq igd (13) The mechanical sensorless technique to estimate the DFIG
2 speed is used in this paper. If the stator and rotor copper losses
of the DFIG are neglected, the relation between stator active
power and rotor active power can be expressed as follow:

Pr = −sPs (16)

Pr
s=− (17)
Ps

Where s is the slip of the DFIG. If the PWM back-to-back

converter and filter losses are neglected, the rotor active power
Pr is equal to the active power Pg sent to the grid. This power
Fig. 3. Overall vector control of the GSC.
can be measured. In the other hand, the relation between the
The reference values of direct current and quadratic current rotor speed Ωm and synchronous speed of the DFIG Ωs can
components can be expressed respectively as follow: be expressed as follow:

2 Q∗g
i∗gq = − (14) Ωm = (1 − s)Ωs (18)
3 vgq
2 Pg∗ From (17) and (18), the estimated wind generator speed can
i∗gq = − (15)
3 vgd be expressed as follow:
The DC-link voltage Vdc , is regulated as shown in the Fig.
4. Fuzzy logic controllers are proposed, to regulate the filter Pg
Ω̂m = (1 + )Ωs (19)
currents and the dc-link voltage as shown in Fig. 4. Ps

1890

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:02:58 UTC from IEEE Xplore. Restrictions apply.
 IV. W IND T URBINE C ONTROL Where Cpopt is the optimal power coefficient to extract the
A. Blade Pitch Control maximum wind power, vw is the wind speed, and Prated is
the power rated of the WECS.
The non-linear variation of the pitch angle versus wind
speed for large wind speed for large wind speed implies the
necessity of using non-linear controllers [16]. The structure
of the Pitch Control System using fuzzy logic controller is
shown in Fig. 6. When the wind speed is below the rated
value, both the rotating speed and output power of the DFIG
are set to their reference values so, the WTG provides the
maximum available power. In this case, the blade pitch control
is deactivated. If the wind speed is above the rated value, the
pitch control is activated in the aim to maintain the output
wind power constant at its rated value [7].

Fig. 8. Fuzzy membership functions for power coefficient error, TSR and
DFIGs speed.
Fig. 6. Structure of the FLC Pitch Control.
Using the relation (20) and the value of blade angle (which
is determined by the Pitch Control system), the FLC calculates
B. New MPPT Approach the value of tip speed ratio (λopt ) which gives the optimal
In the aim to extract the maximum active power, the speed wind power regardless the wind speed. The mathematical
of the WTG must be adjusted to achieve the optimal value representation of the Cp curves used in Fig. 7 is given in (3).
of the tip speed ratio. In this paper, a Fuzzy Logic Controller Then, the optimal generator speed command Ω∗r for maximum
(FLC) is proposed for determining the optimum value of the wind power tracking is given by:
tip speed ratio. The structure of the proposed MPPT FLC is λopt vw
shown in Fig. 7. Ω∗r = (21)
R
For the MPPT method, the speed controller continuously
adjusts the speed of the rotor in order to impose the reference
torque of the DFIG in the aim to follow the calculated
turbines torque. The turbine speed is then regulated to give the
maximum power coefficient. The MPPT method significantly
increases the efficiency of the wind turbine.
V. S IMULATION R ESULTS AND D ISCUSSIONS
Fig. 7. The structure of the MPPT FLC.
To verify the effectiveness of the proposed control strategies
In this paper, we use a 3-input 1-output fuzzy logic con- of the DFIG wind system, simulation studies have been
troller, in which the power coefficient error e = Cpopt − Cpmes , performed using Matlab/Simulink software. The time step and
tip speed ratio and DFIGs speed are fed as inputs. The control the simulation time have been chosen to be respectively 0.001s
output provided by the FLC computes the tip speed ratio which and 120 s. The wind speed model used is described in [13],
is used to gives the optimal DFIGs speed. The membership [18], [19].
functions for input variables are shown in Fig. 8. The linguistic The strong and medium wind speed profiles used are shown
terms used for input output membership function are labeled in Fig. 9. Fig. 10 and Fig. 11 show that the DFIG speed is
as BN (Big Negative), SN (Small Negative), Z (Zero), Small accurately estimated with the estimation errors within 1.5 rad/s
Positive (SP), and BP (Big Positive). In the aim to improve for both medium and strong wind speeds.
the performance of the FLC, an empirical analysis is used to Fig. 12 and Fig. 13 represent the measured and reference dc-
set the membership functions. Note than the defuzzification link voltages for respectively medium and strong wind speeds
method for the output (TSR) is chosen to be centroid of area. using FLC. The control strategies provide good performance.
For the adopted MPPT strategy, the optimal value of the power The average values of the tip speed ratio of the wind turbine
coefficient is given by: for both medium and strong wind speeds are respectively 11
and 7 (Fig. 13).
Prated These variations are the consequence of the fast variations of
Cpopt (λ, β) = (20)
2ρAvw3 the wind speed and the relatively slow dynamic of the WTG

1891

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:02:58 UTC from IEEE Xplore. Restrictions apply.
 Fig. 9. The strong and medium wind speed profiles.

Fig. 12. The DC-link measured and reference voltage for strong wind speed.

Fig. 10. The DFIG measured and estimated speeds and tracking errors for
medium wind speed.

Fig. 13. The DC-link measured and reference voltage for strong wind speed.

Fig. 11. The DC-link measured and reference voltage for medium wind
speed.
Fig. 14. Tip speed ratio for medium and strong wind speeds.

system. The variation of the power coefficient is shown in the blade pitch angle is kept at lower values for both cases.
Fig. 15. With medium wind speed, the power coefficient is So, the new MPPT strategy permits to extract the maximum
kept at its optimal value, while it is decreased if the wind wind power and adjust the value of the power coefficient in
speed increases above the rated value. Then, the wind power the aim to limit the output power generation when the wind
generation is limited. speed increases above the rated value.
Fig.16 shows the output electrical power provided by the
DFIG for strong and medium wind speeds. For the strong wind VI. C ONCLUSION
speed, the electrical power is limited at the rated power by both The overall control scheme of the WECS based on the
MPPT strategy and the pitch control. Fig. 17 shows the blade DFIG is investigated in this paper. To achieve the maximum
pitch angle for medium and strong wind speed. As we can see, wind power generation, a new MPPT strategy using fuzzy

1892

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:02:58 UTC from IEEE Xplore. Restrictions apply.
 in Matlab/Simulink environment shows that the mechanical
sensorless DFIG speed has good performance, with small
tracking errors. The simulation results show that the whole
control strategy gives good performance.
ACKNOWLEDGMENT
This work has been supported by LTE of Hydro-Quebec,
H2CAN Network and Natural Resources Canada and the Nat-
ural Sciences and Engineering Research Council of Canada.
R EFERENCES
[1] H. H. Lee, P. Q. Dzung, L. M. Phuong, L. D. Khoa and N. H. Nhan
A New Fuzzy Logic Approach for Control System of Wind Turbine
with Doubly Fed Induction Generator, International Forum Strategic
Fig. 15. The power coefficient for medium and strong wind speeds. Technology (IFOST) , 2010, pp. 135-139.
[2] X. Yao, C. Yi, D. Ying, J. Guo and L. Yang The Grid-side PWM Converter
of the Wind Power Generation System Based on Fuzzy Sliding Mode
Control, IEEE/ASME International Conference on Advanced Intelligent
Mechatronics, 2008, pp. 973-978.
[3] X. Li, D. Hui, X. Lai, and M. Ouyang Control Strategy of Wind
Power Output by Pith Angle Control Using Fuzzy Logic, 2010 IEEE
International Symposium on Industrial Electronics (ISIE), pp. 120-124,
Jul 2010.
[4] R. Pena, R. Cardenas, J. Proboste, J. Clare, and G. Asher, Wind-Diesel
Generation Using Doubly Fed Induction Machines, IEEE Trans. Energy.
Convers., vol. 23, no. 1, pp. 202-213, March. 2008.
[5] T. Burton, D. Sharpe, N. Jenkins, and E. Bossanyi, Wind Energy Hand-
book. John Wiley & Sons, Ltd, 2001.
[6] J. Shi, Y. Tang, Y. Xia, L. Ren, and J. Li, SMES Based Excitation System
for Doubly-Fed Induction Generator in Wind Power Application, IEEE
Trans. Appl.Superconductivity, vol. 21, no. 3, pp. 1105-1108, June. 2011.
[7] W. Qiao, W. Zhou, J.M. Aller, and R.G. Harley, Wind Speed Estimation
Based Sensorless Output Maximization Control for a Wind Turbine
Driving a DFIG. IEEE Trans. Power Electron., vol. 23, no. 3, pp. 1156-
1169, May 2008.
Fig. 16. The output power generation for medium and strong wind speeds.
[8] S. M. Barakati, M. Kazerani, and J. D. Aplevich, Maximum Power Track-
ing Control for a Wind Turbine System Including a Matrix Converter.
IEEE Trans. Energy. Convers., vol. 24, no. 3, pp. 705-713, Sep. 1999.
[9] Q. Wang and L. Chang, An Intelligent Maximum Power Extraction
Algorithm for Inverter-based Variable Speed Wind Turbine System, IEEE
Trans. Power Electron., vol. 19, no. 5, pp. 12421249, Sep. 2004.
[10] L. Qu and W. Qiao, Constant Power Control of DFIG Wind Turbines
With Supercapacitor Energy Storage. IEEE Trans. Ind. Appl., vol. 47, no.
1, pp. 359-367, Jan/Feb. 2011.
[11] S. R. Kalantarian, and H. Heydari, An Analytical Method for Selecting
Optimized Crowbar for DFIG with AHP Algorithm. 2nd Power Electron-
ics, Drive Systems and Technologies Conferences, pp. 1-4, Feb. 2011.
[12] R. Pena, J. C Clare, and G. M. Asher, Doubly fed induction generator
using back-to-back PWM converters and its application to variable speed
wind-energy generation. IEE Proc.-Electr. Power Appl., Vol. 143, No 3,
May 1996.
[13] I. Munteanu, A. I. Bratcu, N-A. Cutululis and E. Ceanga, Optimal
Control of Wind Energy Systems. Springer 2008.
[14] N. M. Miller, W. W. Price, and J. J. Sanchez-Gasca, Dynamic Modelling
Fig. 17. Blade pitch angle for medium and strong wind speeds. of GE 1.5 and 3.6 Wind Turbine-Generators. GE-Powwer Systems Energy
Consulting, General Electric International, Inc, Schenectady, NY, USA,
Oct, 27, 2003.
[15] T. Ghennam. E. M. Berkouk and F. Bruno, Modeling and Control of a
logic controller is presented. This strategy presents many Doubly Fed Induction Generator (DFIG) Based Wind Conversion System
Power Engineering, Energy and Electrical Drives, 2009, POWERING09,
advantages like using only the measured wind speed; not pp. 507-512, Lisbon, 18-20 March 2009.
required significant memory space; and not required complex [16] R. Dettmer, Revolutionary energyA wind/diesel generator with flywheel
and time consuming calculations. Also, this strategy reduces storage. Inst. Electr. Eng. Rev., vol. 36, pp. 149151, Apr. 1990.
[17] E. Welfonder, R. Neifer, and M. Spanner, Development and experimental
the stress in the pitch control system by adjusting the value of identification of dynamic models for wind turbines. Control Eng. Practice,
the power coefficient according to the wind speed variations Vol. 5, No 1, pp. 63-73. 1997.
to extract the maximum wind power and limit it at its rated [18] C. Nichita, D. Lucas, B. Dayko, and E. Ceanga, Large Band Simulation
of the Wind Speed for Real Time Wind Turbine Simulators. IEEE Trans.
value. Using two different wind profiles, with fast variations, Energy. Convers., vol. 17, no. 4, pp. 523-529, Dec. 2002.
we showed that by the intermediary of rapid changes in TSR,
the generator comes to answer these variations. The simulation

1893

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:02:58 UTC from IEEE Xplore. Restrictions apply.
Powered by TCPDF (www.tcpdf.org)