See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/272802069

Backstepping control of DFIG-Generators for Wind Turbines Variable-Speed

Conference Paper · October 2014

CITATIONS READS

26 675

8 authors, including:

Badre Bossoufi Alami Hala

Faculty of Sciences; Sidi Mohamed Ben Abdellah University Université Mohammed Premier
159 PUBLICATIONS 2,134 CITATIONS 29 PUBLICATIONS 389 CITATIONS

SEE PROFILE SEE PROFILE

Mohammed Karim Ahmed Lagrioui

Sidi Mohamed Ben Abdellah University National High School of Arts and professions in Meknes (ENSAM)
105 PUBLICATIONS 1,374 CITATIONS 54 PUBLICATIONS 1,010 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Badre Bossoufi on 26 February 2015.

The user has requested enhancement of the downloaded file.

 Backstepping control of DFIG-Generators for Wind
Turbines Variable-Speed
Badre Bossoufi1, Mohammed Larbi El Hafyani1, Hala Mohammed Karim2, Ahmed Lagrioui3, Mohammed
Alami Aroussi1 Taoussi2, Aziz Derouich4
1 2
Laboratory of Electrical Engineering and Maintenance, STIC Team, Faculty of Sciences Dhar El Mahraz, Sidi
Higher School of Technology, EST-Oujda, Mohamed Ben Abdellah University, Fez, Morocco.
3
University of Mohammed First, Morocco. Department of Electrical and Computer Engineering,
[email protected] Higher National School of Arts and Trades, Moulay Ismail
University Meknes, Morocco
4
Department of Electrical and Computer Engineering,
Higher School of Technology, Sidi Mohamed Ben Abdellah
University Fez, Morocco

Abstract — In this paper, we present a nonlinear robust control the GSC converter, it controls the DC bus voltage and power
of active and reactive power by the use of the technique factor network side.
Backstepping a double-fed asynchronous generator (DFIG) The not adaptive backstepping approach offers a choice of
system incorporated in a wind. The power transfer between the design tools for accommodation of uncertainties
stator and the network is carried out by acting on the rotor via a
nonlinearities. And can avoid wasteful cancellations.
bidirectional signal converter.
A new control technique for wind systems is presented. This However, the not adaptive backstepping approach is
control system is based on an adaptive pole placement control capable of keeping almost all the robustness properties of the
approach integrated to a Backstepping control system. The mismatched uncertainties. The not adaptive backstepping is a
stability of the system is shown using Lyapunov technique. The rigorous and procedure design methodology for nonlinear
performances and robustness of DFIG and wind-turbine are feedback control. The principal idea of this approach is to
analyzed and compared by simulation based Matlab / Simulink recursively design controllers for machine torque constant
software. uncertainty subsystems in the structure and ‘‘step back’’ the
Keywords-component; DFIG-Generator, Backstepping Control, feedback signals towards the control input. This approach is
Matlab/Simulink, Wind Turbine.
different from the approach of the conventional feedback
I. INTRODUCTION linearization in that it can avoid cancellation of useful
nonlinearities in pursuing the objectives of stabilization and
Today, wind energy has become a viable solution for the
tracking. A nonlinear backstepping control design scheme is
production of energy, in addition to other renewable energy
developed for the speed tracking control of DFIG that has
sources. While the majority of wind turbines are fixed speed,
exact model knowledge. The asymptotic stability of the
the number of variable speed wind turbines is increasing [1].
resulting closed loop system is guaranteed according to
The Doubly-Fed Asynchronous Generator (DFIG) with
Lyapunov stability theorem.
Backstepping control is a machine that has excellent
A nonlinear backstepping control design scheme is
performance and is commonly used in the wind turbine
developed for the speed tracking control of DFIG that has
industry [2-3]. There are many reasons for using an Doubly-
exact model knowledge. The asymptotic stability of the
Fed Asynchronous Generator (DFIG) for wind turbine a
resulting closed loop system is guaranteed according to
variable speed , such as reducing efforts on mechanical parts,
Lyapunov stability theorem.
noise reduction and the possibility of control of active power
In this paper, we present a technique to control two power
and reactive.
converters which is based on the backstepping control. We
The wind system using DFIG generator and a "back-to-
analyze their dynamic performances by simulations in
back" converter that connects the rotor of the generator and
Matlab/Simulink environment. We start by modeling of the
the network has many advantages. One advantage of this
wind turbine, and then a tracking technique operating point at
structure is that the power converters used are dimensioned to
maximum power point tracking (MPPT) will be presented.
pass a fraction of the total system power [5-6]. This allows
Thereafter, we present a model of the DFIG in the dq
reducing losses in the power electronics components. The
reference, and the general principle of control of both power
performances and power generation depends not only on the
converters which is based on backstepping technique.
DFIG generator, but also the manner in which the two parts of
The general structure of an electric control device of
"back-to-back" converter are controlled.
machines which is show in Fig.1:
The power converter machine side is called "Rotor Side
Converter" (RSC) and the converter Grid-side power is called
"Grid Side Converter" (GSC). The RSC converter controls the
active power and reactive power produced by the machine. As
 speed ratio λ which is 8 (λopt). Fixing β and λ respectively to
their optimal values, the wind system provides optimal power.

Figure 1. Architecture of the Control Figure 2. Power coefficient as a function of λ and β

II. MODELLING OF THE WIND-TURBINE

By applying the theory of momentum and Bernoulli's
theorem, we can determine the incident power, the power
extracted from the turbine rotor due to wind, the aerodynamic
torque, the total inertia J and the evolution of the mechanical
speed from Cmec total torque applied to the rotor of the DFIG
[7-8]:
1
Pincident = .ρ .S .v 3 (1)
2
1
Pextracted = .ρ .S .C p (λ , β ).v3 (2) Figure 3. Wind-turbine DFIG characteristics
2
Ω .R The above equations are used to prepare the block diagram
λ= t (3) of the model of turbine (Fig.4).
v
Peol 1 1
Cal = = .ρ .S .C p (λ , β ).v3 . (4)
Ωt 2 Ωt
J tur
J= + Jg (5)
G2 Figure 4. Wind-turbine model
dΩ mec
J = Cmec = Car − Cem − f .Ω mec (6)
III. EXTRACTION OF MAXIMUM POWER
dt In order to capture the maximum power of the incident
S: the area swept by the pales of the turbine [m2], energy of the wind-turbine, must continuously adjust the
ρ : the density of the air (ρ =1.225kg / m3), rotational speed of the wind turbine. The optimal mechanical
v :the wind speed [m/s]. turbine speed corresponds has λopt and β = 0°. The speed of
Ωt : Rotational speed of the turbine the DFIG is used as a reference value for a controller
Cal : Torque on the slow axis (turbine side) proportional-integral type (PI phase advance). The latter
Jtur : turbine inertia determines the control set point which is the electromagnetic
J g : inertia of the generator. torque that should be applied to the machine to run the
Ωmec : Mechanical speed of DFIG generator at its optimal speed.
Car : Aerodynamic torque on the fast axis of the turbine The torque thus determined by the controller is used as a
Cem : Electromagnetic torque reference torque of the turbine model (Fig.5). The system of
The Cp (λ, β) is called the power coefficient, which variation of the angle of orientation of the blades (variation of
expresses the aerodynamic efficiency of the turbine. It the angle of incidence) to change the ratio between the lift and
depends on the ratio λ, which represents the ratio between the drag. To extract the maximum power (and maintain constant),
speed at the end of the blades and the wind speed and the the adjusted angle of incidence of the blades to the wind
angle of orientation of the blades β. speed. The "Pitch Control" is a technique that mechanically
The Fig.2 and Fig.3 shows the curves of the power adjusts the blade pitch angle to shift the curve of the power
coefficient as a function of λ for different values of β. A coefficient of the turbine [13].
coefficient of maximum power of 0.498 is obtained for a However, it is quite expensive and is generally used for
wind turbines and high average power. For our model, the
"Stall Control" technique, which is a passive technique that The active and reactive powers stator and rotor of the
allows a natural aerodynamic stall (loss of lift when the wind DFIG are written as follows [10-14]:
speed becomes more important). Otherwise, β is zero. The Ps = vds .ids + vqs .iqs (18)
synthesis of the PI controller requires knowledge of the
transfer function of our system. This is especially difficult Qs = vqs .ids − vds .iqs (19)
because of the power coefficient. A simple proportional
correction (P) is obtained after testing. Note that β may vary
Pr = vdr .idr + vqr .iqr (20)
from 0 ° to 90 ° characterized by saturation. Qr = vqr .idr − vdr .iqr (21)

Figure 5. Block diagram with control of the speed

IV. DFIG MODEL SYSTEM Figure 6. DFIG Model for wind-turbine from Matlab (Table I)
The Power equations in the dq reference DFIG can be V. BACKSTEPPING ADAPTATIVE CONTROLLER APPLIED TO
written [10-14]: DFIG GENERATOR
d
vds = Rs .ids + φds − ω s .φqs (7) The approach control adaptive Backstepping is a control
dt technique that can effectively linearize a nonlinear system as
d DFIG in the presence of uncertainties. Unlike other
vqs = Rs .iqs + φqs + ω s .φds (8) information linearization techniques, adaptive backstepping
dt has the flexibility to keep useful nonlinearity intact during
d stabilization. The essence of backstepping is to stabilize the
vdr = Rr .idr + φdr − ω s .φqr (9)
state of the virtual control. Therefore, it generates a
dt
corresponding error variable that can be stabilized by carefully
d
vqr = Rr .iqr + φqr + ωr .φdr (10) selecting the appropriate control inputs. These inputs can be
dt determined from the analysis of Lyapunov stability
With: From (7) it is clear that the dynamic model of the PMSM is
ωr = ω s − P.Ω (11) highly non-linear due to the coupling between the speed and
the electric currents. According to the vector control principle,
φds = Ls .ids + M .idr (12) the direct axis current id is always forced to be zero in order to
φqs = Ls .iqs + M .iqr (13) orient all the linkage flux in the d axis and achieve maximum
torque per ampere.
φdr = Lr .idr + M .ids (14) disd r Lsq V
= − s .isd + pΩisq + sd (22)
φqr = Lr .iqr + M .iqs (15) dt Lsd Lsd Lsd
Ls = l s − M s , Lr = lr − M r (16) disq r L Φf Vsq
Ls, Lr: Cyclic inductances of stator and rotor phase; = − s .isq − sd pΩisd − pΩ + (23)
ls, lr : Inductances of stator and rotor phase;
dt Lsq Lsq Lsq Lsq
Ms, Mr : Mutual inductances between stator and rotor phases dΩ 3 p f C
respectively; = (Φ f isq + (Lsd − Lsq )isd isq ) − Ω + r (24)
dt 2J J J
[ ]
M: Maximum mutual inductance between stator and rotor
stage (the axes of the two phases coincide). The vector [x ] = [x1 x2 x3 ] = isd isq Ω choice as
T T

The expression of the electromagnetic torque of the DFIG state vector is justified by the fact that currents and speed are
depending on flow and stator currents can be written as measurable and that the control of the instantaneous torque
follows: can be done comfortable via the currents isd and/or isq. And
Cem = P(φds .iqs − φqs .ids ) (17) stator voltages as control variables [
u = Vsd Vsq . ] T

With P: number of pole pairs of the DFIG.

 The main objective of the backstepping control is to B. Backstepping Current controller
regulate the speed of the machine to its reference value Ωref We have the asymptotic stability of the origin of the
whatever external disturbances. It is assumed that the engine system (24). We defined:
parameters are known and invariant.
[ ]
By choosing isd isq Ω T as variable states and equation (22)
ed = isdref − isd with isdref = 0
 (31)
the mathematical model of the machine. The objective is to eq = isqref − isq
regulate the speed to its reference value. Their dynamics can be written:
A. Backstepping speed Controller r L sq V
e& d = i&sdref − i&sd = s .i sd − p Ω i sq − sd
We begin by defining the tracking errors: L sd L sd L sd
eΩ = Ω ref − Ω (25) 2 (32)
e& q = i&sqref − i&sq = ( fΩ + C r + J .k Ω .e Ω )
The derivative of (25) is: 3 pΦ f
deΩ &
e&Ω = = Ωref − Ω& rs L pΦ f V
dt + .i sq + sd p Ω i sd + Ω − sd
(26) L sq L sq L sq L sd
& − 1 3 p (Φ i + (L − L )i i ) − fΩ − C 
=Ω To analyze the stability of this system we propose the
J  2
ref f sq sd sq sd sq r
 following Lyapunov function:
We define the following quadratic function: 1
1 V2 = (eΩ2 + ed2 + eq2 ) (33)
V1 = eΩ2 (27) 2
2
The expression (20) found above requires the following
Its derivative along the solution of (27), is given by:
control laws:
V&1 =eΩe&Ω 3pLsd
Vsd = kd Lsded + rsisd − LsqΩisq + (Lsd − Lsq)eΩisq
& 1 3p  (28) 2J
= eΩΩ ref −  (Φf isq +(Lsd −Lsq)isdisq) − fΩ−Cr 
 J2  2L (k J − f )  3pΦf 3p 
Vsq = sq Ω  eq + (Lsd − Lsq)edisq −kΩeΩ  (34)
Using the backstepping design method, we consider the dq 3pΦf  2J 2J 
axes current components isd and isq our virtual control elements
and specify their desired behavior, called the stabilization 3pΦf Lsq
function in the terminology of the backstepping design as + eΩ +rsisq + LsdΩisd +ΩΦf + kqLsqeq
follows: 2J
With this choice the derivatives of (21) become:
 i sdref = 0
 V&2 = −kΩ eΩ − k d ed − k q eq ≤ 0 (35)
 2 (29)
 i sqref = 3 p Φ ( f Ω + C r + J . k Ω .e Ω )
VI. SIMULATION RESULTS
 f

Where kΩ is a positive constant The overall model of the wind system was simulated in
Substituting (29) in (28) the derivative of V1: Matlab/ Simulink environment. The model includes: wind
turbine, Doubly-Fed Asynchronous Generator (DFIG), two
V&1 = −kΩeΩ2 ≤ 0 (30) power converters that connect the rotor to the network (Fig.7).

Figure 7. Simulation scheme of adaptive Backstepping Control applied for DFIG-wind-turbine

 2) Disturbance rejection
A. DFIG performances
The following results are obtained by choosing the
following values:

Gains of the control law: k Ω = 0.15 , k d = 0.01 ,
k q = 0.01 .

Adaptation gains: γ 1 = 0.15 , γ 2 = 0 .01 , γ 3 = 0.015 .
1) Follow of the trajectory

(a)

(a)

(b)

(b)

(c)
Figure 9. Test performance of the adaptive controller for rejecting
disturbance torque load applied at t = 0.3s.

B. Wind-turbine performances
The following graph shows the function of a wind turbine
with asynchronous generator dual power scheme (Fig.10).
(c)

(d)
Figure 10. DFIG-Generator with wind turbine (Control interface)
Figure 8. Test performance of the adaptive controller for trajectory tracking
 Using the reduced model, we applied a profile closer to the
evolution of the real wind was filtered to suit the slow
dynamics of the system studied random wind. The objective is
to see the degree of continuing point of maximum power and
efficiency of the speed control provided by the backstepping
controller.
The Fig.11 shows the Benchmark used for completion of this
work:

Figure 11. Benchmark of Wind Turbine System

The Fig.12 shows the wind profile filtered and applied to

the system in this case.

Figure 12. Profile applied to random wind Wind-turbine.

The Fig.13 shows the results obtained for this application,

where the following observations can be distinguished:
 The specific speed λ and the power coefficient Cp does
not change a lot of values, they are almost equal to their
optimal values references 9 and 0.4999 successively;
 The wind power captured follows its optimal reference
and has the same shape as the wind profile applied, this
rate is also consistent with the wind torque side of the
DFIG;
 The speed of the DFIG is the image of wind causing the
wind, it properly follows its reference;
 The shapes of the electromagnetic torque of the DFIG
and its reference, are virtually identical, but different
from the shape of the profile of the wind speed due to the
dynamic torque due to inertia;
 The shape of the components of the stator flux
orientation shows a good flow to ensure vector control
well decoupled from the DFIG.
 [3] W. L. Kling and J. G. Slootweg, « Wind Turbines as Power Plants». in
Proceeding of the IEEE/Cigré workshop on Wind Power and the impacts
on Power Systems, 17-18 June 2002, Oslo, Norway.
[4] D. Seyoum, C. Grantham, «Terminal Voltage Control of a Wind Turbine
Driven Isolated Induction Generator using Stator Oriented Field
Control». IEEE Transactions on Industry Applications, pp. 846-852,
September 2003.
[5] A. DAVIGNY, Participation aux services système de fermes éoliennes à
vitesse variable intégrant un stockage inertiel d’énergie, Thèse de
Doctorat, USTL Lille (France), 2007.
[6] K. GHEDAMSI, Contribution à la modélisation et la commande d’un
convertisseur direct de fréquence. Application à la conduite de la
machine asynchrone, Thèse de Doctorat, ENP Alger (Algérie), 2008.
[7] B.Bossoufi, M.Karim, S.Ionita, A.Lagrioui, “Nonlinear Non Adaptive
Backstepping with Sliding-Mode Torque Control Approach for PMSM
Motor” Journal of Journal of Electrical Systems JES, pp236-248. Vol.8
No.2, June 2012.
[8] X. YU, K. STRUNZ, « Combined long-term and shortterm access
storage for sustainable energy system », 2004 IEEE Power Engineering
Society General Meeting, vol.2, pp.1946-1951, 10 June 2004.
[9] S. E. Ben Elghali, « Modélisation et Commande d’une hydrolienne
Equipée d’une génératrice Asynchrone Double Alimentation »,
JGGE’08, 16-17 Décembre 2008, Lyon (France).S
[10] E. AIMANI, Modélisation de différentes technologies d’éoliennes
intégrées dans un réseau de moyenne tension, Thèse de Doctorat, École
Centrale de Lille (France), 2004.
[11] 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 », Advanced Intelligent Mechatronics, IEEE
2008, Xian (Chine).
[12] B.Bossoufi, M.Karim, S.Ionita, A.Lagrioui, “DTC control based
artificial neural network for high performance pmsm drive” Journal of
Theoretical and Applied Information Technology JATIT, pp165-176,
Vol. 33 No.2, 30th November 2011.
[13] B.Bossoufi, M.Karim, S.Ionita, A.Lagrioui, “Indirect Sliding Mode
Control of a Permanent Magnet Synchronous Machine: FPGA-Based
Implementation with Matlab & Simulink Simulation” Journal of
Theoretical and Applied Information Technology JATIT, pp32-42, Vol.
29 No.1, 15th July 2011.
[14] A. BOYETTE, Contrôle-commande d’un générateur asynchrone à
double alimentation avec système de stockage pour la production
éolienne, Thèse de Doctorat, Nancy (France), 2006.
Figure 13. The Simulation results of the asynchronous wind generator dual [15] B.Bossoufi, M.Karim, S.Ionita, A.Lagrioui, “Low-Speed Sensorless
power and stator flux oriented, with a Backstepping controller (Case scale Control of PMSM Motor drive Using a NONLINEAR Approach
model of the DFIG and profile of random wind). BACKSTEPPING Control: FPGA-Based Implementation” Journal of
Theoretical and Applied Information Technology JATIT, pp154-166,
Vol. 36 No.1, 29th February 2012.
CONCLUSION
This work has been devoted to modeling, simulation and APPENDIX
analysis of a wind turbine operating at variable speed. A stable Table I: DFIG-Generator parameters
operation of the wind energy system is obtained with the
application of nonlinear Backstepping Adaptive control. The Rated Voltage VN 400/230V, 50Hz
overall operation of the wind turbine and its control system Rated Current IN 2.0A/3.5A
were illustrated by responses to transient and permanent Rated Speed 1400/1500 rpm
control systems. The Generator supplied power to the network Rated Power PN 0.8KW
with an active power whatever the mode of operation. The Cosφ 0.75
wind generator has been tested and modeled with a variable Excitation Voltage 130 V CA / 24 V CC
speed operation for a power of 200 kW. Simulation results Excitation Current 4 A CA / 11 A CC
show that the proposed wind system and is feasible and has
many advantages.
REFERENCES

[1] T. Ackermann and Soder, L. « An Overview of Wind Energy-Status

2002 ». Renewable and Sustainable Energy Reviews 6(1-2), 67-127
(2002).
[2] T. Burton, D. Sharpe, N. Jenkins and E. Bossanyi, Wind Energy
Handbook. John Wiley&Sons, Ltd, 2001.

View publication stats