International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp.

3524-3532
© Research India Publications. http://www.ripublication.com

Adaptive Super-Twisting Sliding Mode Control for Wind Energy

Conversion System

Saïd Boubzizi1,2, Hafedh Abid1, Ahmed El hajjaji2, Mohamed Chaabane1

1
Lab-STA Laboratory, National Engineering School of Sfax, University of Sfax, Route de la Soukra 3029 Sfax, Tunisia.
2
MIS Laboratory, University of Picardie Jules Verne, 33, Rue Saint-Leu, 80039 Amiens, France.

Abstract the nominal power, which contributes considerably to reduce

costs.
This paper deals with control strategy of a Wind Energy
Conversion System (WECS) equipped with Doubly-Fed- To effectively extract wind power, the tip-speed-ratio of the
Induction-Generator (DFIG). The novelty of this controller WT should be maintained at its optimum despite wind-speed
concerns the Adaptive-Super-Twisting (ASTW) based design variations. However, the mechanical and electrical parts of
algorithm for the DFIG control. The proposed controller wind energy conversion system (WECS) work generally as a
ensures free-chattering response, finite-time-convergence and nonlinear-system where parametric variations and unmodeled-
improves accuracy without needing time derivative of its dynamics are the major problems which could deviate the
sliding surface. An interesting point comes from the proposed system from its optimal operation. In addition, VSWTs are
adaptive gain strategy which permits to reduce by 50% the expected to work under high wind-speed variations, making
number of gains to be adapted which is interesting for systems their control design a serious challenge, particularly, with large
with several controllers such as WECS (Turbine + generator + wind turbines. In this context, several research studies have
grid). The stability analysis of the proposed control is proved been developed to improve the control strategies for the WT-
using the Lyapunov candidate function approach. The DFIG system. Field-Oriented-Control using Proportional–
designed ASTW algorithm has been compared to conventional Integral (PI) controllers is one of the famous methods due to its
STW-SMC and tested using 1.5MW -wind turbine simulator simple design [2]. However, this control scheme may have low
FAST (Fatigue, Aerodynamics, Structures, and Turbulence) performance under medium and high wind speed variation
which is developed by the National Renewable Energy because of its sensitivity to system parameters and external
Laboratory (NREL), USA. disturbances. Therefore, to overcome these drawbacks further
nonlinear-control techniques have been used, such as fuzzy
Keywords: Wind turbine (WT), Doubly-fed_induction
logic control [3], [4], predictive control [5] and neural networks
generator (DFIG), Power generation, ASTW (Adaptive_Super
control [6]. Sliding-mode approach is an effective solution for
twisting), Sliding mode control (SMC)
controlling perturbed-systems [7]. Although Sliding mode
control (SMC) is robust, insensitive to external perturbations
and uncertainties of the system, it unfortunately generates
INTRODUCTION undesirable chattering [7]. High order sliding mode control
Wind power has been one of the world's leading renewable (HOSMC) techniques are among the appropriate alternatives
energy suppliers in the last years. This source of power for avoiding chattering effects [8], [9], [10]. This approach
generation which produces clean energy plays a more ensures driving of sliding surface and its successive derivatives
important role in the way our world is powered, as it can also to zero as well as increases the stabilization of the system.
solve the problem of importing energy around the world and Super-Twisting (STW) algorithm is a second-order SMC. The
meet the energy needs of isolated sites. By the end of 2016, idea consists in generating a continuous control function which
global wind installed capacity was 486.7 GW which represents ensures the convergence of the variable surface and its
an increase of 12% market growth from the number in 2015, derivative in finite time under a gradient of delimited
mainly in china 145.3 MW, United States 74.48 MW, Germany perturbations, at a known limit. Indeed, standard STW
44.947 MW, India 25 MW, Spain 23 MW, United Kingdom 13 algorithm includes a discontinuous term under its integral, so
MW, Canada 11.2 MW, and France 10,4 MW [1]. This growth the chattering is attenuated but not removed. This standard
is due to the astonishing increase of global new installations control requires knowledge of the limits of the disturbance
63,467 MW in 2015 which represents a growth of 22% of the gradients which is not an easy task in practical cases. This
annual market [1]. results in an overestimation of the disturbance limit while
developing the STW laws which leads to higher control gains.
Variable-speed-wind-turbine (VSWT) and double fed induction The main contribution of this paper, which deals with WT-
generator have been taking increasing attention during the last DFIG is the control approach, which is original due to the
decades, since they can work in a wide range of wind speed and proposed adaptive algorithm system that permits not only an
have four-quadrant-power capabilities. Furthermore, wind online adaptation of control gains but also, reduces by 50%
turbine (WT) based on double fed induction generator (DFIG) number of gains to be adjusted compared to strategy proposed
has the advantage of reducing converter prices to only 25% of by [11]. The proposed strategy is interesting for systems with

3524
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

several controllers such as WT-DFIG system (Turbine +

generator + grid). The controller is fully based on an Adaptive
Super Twisting control (ASTW) that ensures the convergence
of the sliding surface to zero without over-estimating the
algorithm gains with unknown bounded uncertainties /
perturbations.
This paper is organized as follows section 2 describes the wind
energy conversion system based on DFIG. Section 3 presents
the proposed control strategy of the DFIG. Finally, section 4
gives simulated results of the proposed controller.

DESCRIPTION OF WECS Figure 2. WT power coefficient

This note focuses on one of the most popular configurations in
wind energy conversion systems with a fixed pitch and variable The slow angular velocity of the turbine ωr is adapted to the
speed wind turbine. A comprehensive scheme of WT-DFIG
structure is shown in figure 1. fast speed of the generator ω by the following relation

ng  (3)
r
The global mechanical dynamics of the rotor can be
expressed as follows, [13]
J r  Ta  Kr  Tg   (4)

 J  J r  ng2 J g

with  K  K r  ng2 K g
Figure 1. WECS based on DFIG T  n T
 g g em

The expression of aerodynamic power extracted from the wind

 denotes the unmodeled dynamics and external disturbances.
is given by, [12]
1
Pa   R 2  C p    3 (1)
2
where R , C p ,  and  denote consecutively the radius and
the power coefficient of the WT, the wind speed and the air
density.
The tip-speed-ratio is defined as
Rr
 (2)

The power extraction of a wind turbine is a function of the
available wind power, the power curve of the machine and the
ability of the machine to respond to wind fluctuation as shown
in figure 2 and figure 4. Figure 3. Wind turbine drive-train
To achieve maximum efficiency, rotor speed should track the
reference given by the Maximum Power Point Tracking
(MPPT), under different wind speed patterns.
1 C p max 
Tref  kopt r2 , kopt   R5 3 , ref  opt (5)
2 opt R

3525
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

oscillations of torque. This results in maximum power

extraction.
Based on relation (9), the stator-side reactive power of the
DFIG given by (8) can be written as follows
Vs
Q s  MI dr  (12)
Ls

By setting equation (12) to zero, the rotor reference current

I rd , ref is obtained as
s Vs
I rd ,ref   (13)
Figure 4. Cp–λ characteristics M s M
The dynamic model of the DFIG is described in a The main objectives in this work, is to maximize wind power
synchronously rotating frame d-q, [14] generation by reaching optimal torque Tref , ensuring zero
Vsd  Rs I sd   sd  ssq 
 Vrd  Rr I rd   rd  rq reactive power and improving quality of the generated energy.
V  R I  ,
V  Rr I rq   rq  rd
(6) Those objectives can be successfully fulfilled through a robust

 sq s sq  sq  ssd   rq control strategy.
where s is synchronous speed. Let’s consider the following tracking errors of rotor current and
electromagnetic torque
The flux linkages are given by
 S1  I rd  I rd_ref

sd  Ls I sd  MI rd rd  Lr I rd  MI sd S  (14)
  L I  MI ,   L I  MI (7)  S2  Tem  Tref

 sq s sq rq  rq r rq sq
Systems of equations (6) can be described by the following
The electromagnetic torque, the active and the reactive power block diagram figure 5.
generated by the generator can be expressed as follows [10]

 P  Vsd I sd  Vsq I sq
 M
Q  V I  V I , Tem  n p
 Ls

I rd sq  I rqsd  (8)
 sq sd sd sq

By applying the following vector alignment [10], we obtain

sd  s Vsd  0
  0 , V  V    (9)
 sq  s sq s sd

and
 1  M 
 I rd   Vrd  Rr I rd  ss Lr I rq   s 
  Lr  Ls 
 (10)
  1  V  R I  s  L I  s M  
 I rq  L  rq r rq s r rd s 
Ls s 
 r 

M2
where,   1  , n p is the pole pair number and s is the Figure 5. System to be controlled
Ls Lr
slip speed.
Hence, the electromagnetic torque becomes  Adaptive Super-Twisting sliding mode control

M High order sliding mode control is an effective method to

Tem  ke I rq with ke  n p s (11) overcome the chattering problem. In fact, a control order “n”
Ls acts on “n” derivatives. This feature helps to attenuate the
undesirable chattering while preserving the prosperity of the
robustness of the SMC approach [15]. However, the
CONTROL OF THE DFIG implementation of an n-th order controller requires knowledge
The control of WT-DFIG must be a compromise between of the n-th consecutive sliding surfaces. The exception is the
maintaining the optimum performance and limiting the STW algorithm which needs only information on the sliding
surface S ,[16]. In the following, an adaptive Super_Twisting

3526
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

(ASTW) control strategy is proposed to improve the Vrd controls the reactive power ( I rd ) and Vrq controls the
performance and effectiveness of the WT-DFIG system.
active power ( Tem ).
The derivative of system (14) is
Assumption 1. We assume that G1 and G2 are bounded
 1  M  functions which satisfy the following conditions:
 S1   Vrd  Rr I rd  ss Lr I rq  s   I rd ref
  Lr  Ls  and G2 (t , S2 )   2 S2
G1 (t , S1 )  1 S1
12 12
, where 1 and  2
 (15)
 S  n M   V  R I  s  L I  s M    T are some positive constants.
 2 p s  rq r rq s r rd s s ref
  Ls Lr  Ls 
The theorem below which is inspired from [17], provides
sufficient conditions to ensure the stability and robustness of
Let’s consider functions G1 and G2 as follows:
current and electromagnetic torque control of the DFIG.
 1  M  Theorem 1: Consider the system described by equation (10)
G1    Rr I rd  ss Lr I rq  s   I rd_ref
  Lr  Ls  and the sliding surface given by (14). If we apply the control
 (16) law (18) based on Second Order Sliding Mode Controller to the
G  n M    R I  s L  I  s M    T system. The closed-loop system (15) is ensured with finite-time
 2 p s r rq s r rd s s ref
  Ls Lr  Ls  convergence with the following adaptive gains
(ai ,bi for i  1, 2)
Assumption 1: We assume that functions G1 , G2 are bounded
G1  1 S1
12
, G2   2 S2
12   
 ki i Si , if Si  i
where 1 ,  2 are positive unknown constants. a
 i (t , S i )   2
 0, if Si  i
(20)
 
Then, equation (15) can be written as: b (t , S )  2 a (t , S )    4 2
 i i i i i i i
 1
 S1   L Vrd  G1 where  i , ki ,i , i and  i are arbitrary positive constants. The
r
 (17) parameter i is introduced in order to get only positive values
 S2  n p M sVrq  G2
  Ls Lr for ai (t , Si )  0 .

The control laws are given by: Proof of convergence

Let us consider the case of Vrd
V  y  a ( S , t ) S 1 2 sgn S
 rd 1 1 1 1  1
 y1  b1 ( S1 , t ) sgn  S1  S1 
1
Vrd  G1 (22)
 (18)  Lr
Vrq  y2  a2 ( S2 , t ) S2 2 sgn  S2 
1


 y2  b2 ( S2 , t ) sgn  S2 
and
 S  qa S 1 2 sgn S +  G
where, the adaptive gains are
 1 1 1  1 * 1
 *  qb1 sgn  S1 
ai ( Si , Si , t ), (23)
 , for i  1, 2 (19) 1
bi ( Si , Si , t )
 q , *  qy1
 Lr

where a1 (t , S1 ) and b1 (t , S1 ) are adaptive gains:

  
 k1 1 S1 , if S1  1
a
 1 (t , S1 )   2
 0, if S1  1
(24)
 
b (t , S )  2 a (t , S )    4 2
1 1 1 1 1 1 1

Let’s introduce the following new vector:

T
z   z1 , z2    S1 sgn  S1  , * 
T 1
2 (25)
 

Figure 6. Block-scheme of the proposed WT-DFIG control We can easily deduce the derivative of (25)
strategy

3527
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com


 z1 
1  qa1
 
1 1 
z1  z2  G1  a1 
 Lr1  1  412   21 1  12   (35)
 z1  2 2 2 
 (26) 1
 z   qb1 z
 2 z1
1 From (33), we have
1 T 2 
V0  z    z Qz   1 zT z   1 z
2
Using assumption A1, we get (36)
2 z1 2 z1 z1
G1 ( x, t )  1 ( x, t ) S1 sgn(S1 )  1 ( x, t ) z1
12
(27)
We note min  P  and max  P  as the minimal and
where 1 ( x, t ) is a bounded function, 0  1 ( x, t )  1 : maximal eigenvalues of matrix P , then we can write:
z  A  z1 , z2  z (28) min  P  z  zT Pz  max  P  z
2 2

with In view of (25), we deduce

1  qa1 +1 1
A  z1 , z2  
2
  qb 0  (29) z  z12  z22  S1  z22
2 z1  1 
12 V01 2 ( z )
The stability analysis of (28), can be proved using and z1  S1  z 
Lyapunov candidate function (30), inspired from [17] min
12
( P)

T 1
  1
  Then, we can conclude
2 2
V  z, a1 , b1   V0  a1  a1*  b1  b1* (30)
2 1 2 1
V0 ( z )  rV01 2 (37)
where
min
12
( P)
where, r 
 
V0 ( z )  1  412 z12  z22  4 z1 z2  zT Pz (31) max ( P)

  412  21  Incorporating with (32) and (37), we obtain

P 1
 21 1  k1 k1 1
V   rV0
12
  a1 S1   b1 S1   a1a1
where P is a positive definite matrix 1  0,  1  0 , 2 1 2 1 1
1 k1 k1
a1 (t , S1 )  a1* and b1 (t , S1 )  b1* , t  0 . +  b1b1   a1 S1   b1 S1 (38)
1 2 1 2 1
Hence, the derivative of (30)
1 1
V  z, a1 , b1   z T  AT P  PA z   a1a1   b1b1
T Consider the well-known inequality
(32)
 
x 
1 1 12
2
 y2  z2  x yz
where  a1  a1  a1* and  b1  b1  b1* .
and in view of (38), we can derive
The first term of equation (32)
k1 k1
T 1 T rV01 2   a1 S1   b1 S1
V0  z, a1 , b1   zT  AT P  PA z   z Qz (33) 2 1 2 1
2z 1 (39)
1
 1 1  2
where  1  V0   a1S1 2   b1S1 2 
 2 1 2 1 
Q Q 
Q   11 12 
Q21 4  with 1  min r, k1 , k1  
Q11  21qa1  4q1  21a1  b1   2 1 1  412 .

with 
  Let’s note term V1 as follows


 12 
Q  Q21  qb1  2q1a1  1  412  21 1.  V1  V0 
1
2 1
1
 a1S1 2   b1S1 2
2 1
By choosing
From (39), we can rewrite (38) as
b1  21a1  1  412 (34)

Then, Q will be positive definite matrix if a1 is large

enough satisfying:

3528
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

1 1 simulator is validated by the German company Germanischer-

V  z , a1 , b1   1V11 2   a1a1   b1b1 Lloyd-WindEnergie, which specializes in certification and
1 1
(40) testing of WT’s. FAST have many input for turbine control
k1 k1
  a1S1   b1S1 settings, environmental conditions, blade models,etc, to help
2 1 2 1 researchers evaluate wind energy offshore technology figure 7.

Now we assume that adaptation adaptive gains a1 (t )

and b1 (t ) are bounded such that a1 (t )  a1* and b1 (t )  b1* ,
t  0 , then (40) becomes

1 k 
V  z , a1 , b1   1V11 2   a1  a1  1 S1 
 1 2 1 
 
(41)
1 k 
  b1  b1  1 S1 
 1 2 1 
 
Thus, we obtain

V  z, a1 , b1   1V1  z, a1 , b1  
12
(42)

where Figure 7. FAST wind turbine model in Simulink

1 k  1 k 
    a1  a  1 S1    b1  b1  1 S1 
 1 1 2 1   1 2 1 
   
By choosing   0 , we can ensure the finite time
convergence through the following dynamic adaptation
gains

1
a1  k1 S1 (43)
2
From adaptation law (24) and (43), we have

1
b1  21a1  1  412  b1  21a1  k1 S1
2

k1  1
By selecting 1  , we ensure that   0 .
2k1  1

Hence, the derivative of system (30) is guaranteed definite Figure 8. Block diagram of the global system
negative and thus convergence of S1  0 which can be finally
written as follows: SIMULATION RESULTS AND DISCUSSION
V  z, a1 , b1   1V11 2 (45) The proposed control strategy has been validated with
numerical validation using 1.5 MW 3-blade-WT FAST
with 1  0 . simulator, developed by National Renewable Energy
Laboratory (NREL), figure 8.
Similar development can be used to prove the stability
analysis in the case of Vrq . The wind inflow for the simulation generated by FAST
simulator is shown in figure 9. The wind speed profile
 Simulation using FAST Code simulator represents a turbulent wind speed variation between 8 m/s and
14.5 m/s. Figure 10 shows the mechanical speed of the
The proposed Adaptive Super-Twisting algorithm has been generator until 1800 rpm. The desired DFIG reference are
tested for validation using FAST (Fatigue-Aerodynamics- directly tracked through an Adaptive Super-Twisting control.
Structures-Turbulence) software which is an effective aero- As can be clearly seen in figure 11 and figure 14, the
elastic simulator that can predict extreme and fatigue loads of a electromagnetic torque and rotor current of the DFIG show
spread of WT’s configurations (two and three-blade horizontal- good tracking performances under turbulent wind-speed
axis) [18]. FAST relies on advanced engineering models variation. The rotor current reference in figure 11 is determined
derived from basic laws with appropriate assumptions. This
as a function of stator voltage Vs . Figure 14 represents the

3529
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

electromagnetic torque response and the aerodynamic torque of

the simulator, its maximum value is around -8.4 KN.m. The
proposed ASTW control strategy does not induce increased
mechanical stress and ensures smooth aerodynamic torque
variations. The simulation results displayed in figure 12 and
figure 13 show that the gains a1 (S1 , t ) and a2 ( S2 , t ) are
growing quickly as the disturbances are quicker eliminated,
which permits faster convergence of sliding surface to zero.
b1 ( S1 , t ) and b2 (S2 , t ) are deduced from a1 (S1 , t ) and
a2 ( S2 , t ) according to the proposed adaptation laws. Figures 9-
15 prove the efficiency of the proposed ASTW control
algorithm in terms of tracking error, free chattering and better
power quality generation figure 15 compared to standard super-
twisting SMC with fixed gains presented in figure 16. The Figure 11. Current I rd tracking.
proposed control overcomes the problem of undesirable
chattering of standard control while ensuring the robustness
features of SMC.
The presented results show that a standard SOSMC control
method can be an interesting solution for systems based DFIG
technology for wind power conversion. However, chattering
effect remains a significant problem with this technique. The
results obtained show that ASTW strategy effectiveness is
more attractive in terms of power energy quality and higher
accuracy. The proposed control strategy reduced the number of
gains to be adjusted via an adaptation algorithm.

Figure 12. Gain adaptation B1 .

Figure 9. Wind speed profile (m/s)

Figure 13. Gain adaptation B3 .

Figure 10. DFIG mechanical speed.

3530
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

and satisfactory in terms of power extraction maximization,

higher accuracy, faster convergence, free chattering, no
overflow and without induced mechanical stress which
contributes to increase reliability and improve the system's
energy efficiency.

Table 1. Wind Turbine and DFIG Parameters

Turbine diameter (m) D = 70
Air-density (kg.m3)  = 1.225

Gear reduction ratio ng = 75

Figure 14. Electromagnetic torque. Stator-resistance (Ω) Rs = 0.00297

Rotor-resistance (Ω) Rr = 0.00382

Stator-inductance (H) Ls = 0.121

Rotor-inductance (H) Lr = 0.0573

Cyclic-inductance (H) M = 0.031

Number of pole pairs np = 2

Table 2. ASTW Control Parameters

k1 k2 1 2 1   2
80 100 1 12 4
Figure 15. Active power (MW), reactive power (MVar) with
proposed control.
REFERENCES
[1] Global Wind Energy Council Statistics, GWEC,
www.gwec.net/global-figures/graphs/, 2017.
[2] Saurabh, M., Amar, N., Deependra, S., 2016,
“Optimum design of Proportional-Integral controllers in
grid-integrated PMSG-based wind energy conversion
system,” International Transactions on Electrical
Energy Systems, 26, pp. 1006-1031.
[3] Mohamed, A., Eskander, M., Ghali, F., 2001, “Fuzzy
logic control based maximum power tracking of a wind
energy system,” Renewable Energy, 23, pp. 235-245.
[4] Mostafa, J., Ali , A., 2015, “Self-Tuning fuzzy PI-based
Figure 16. Active power with the standard control. controller of DFIG wind turbine for transient conditions
enhancement,” International Transactions on Electrical
Energy Systems, 25, pp 2657–2673.
CONCLUSION [5] Mohsen, D., Abolfazl, J., 2016, “Predictive control
An Adaptive-Super-Twisting algorithm based on SMC control strategy to improve stability of DFIG-based wind
strategy is developed for wind energy conversion system. The generation connected to a large-scale power system,”
proposed Lyapunov-based second order sliding mode approach International Transactions on Electrical Energy
allows an online adaptation of control gains and ensures the Systems, Online Version of Record published before
convergence in a finite-time. The boosted, robustness and inclusion in an issue.
stability of the proposed control strategy have been compared [6] Mahabuba, A., Abdullah , M., 2009, “Small signal
to conventional STW-SMC under FAST wind turbine stability enhancement of a multi-machine power system
emulator. The obtained results with ASTW are more significant using robust and adaptive fuzzy neural network-based

3531
 International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 6 (2018) pp. 3524-3532
© Research India Publications. http://www.ripublication.com

power system stabilizer,” International Transactions on

Electrical Energy Systems, 19, pp. 978–1001.
[7] Utkin, V., 1993, “Sliding Mode Control Design
Principles and Applications to Electric Drives,” IEEE
Trans on industrial electronics, 40.
[8] Patnaik, R., Dash, P., Mahapatra, K., 2016. “Adaptive
terminal sliding mode power control of DFIG based
wind energy conversion system for stability
enhancement,” International Transactions on Electrical
Energy Systems, 26, pp. 750–782.
[9] Adjoudj, M., Abid, M., Aissaoui, A., Ramdani, Y.,
Bounoua, H., 2011, “Sliding mode control of doubly
fed induction generator for wind energy turbine,”
Rev.Roum. Sci.Techn-Électrotechn. et Énerg,, 56, pp.
15-24.
[10] Benbouzid, M., Beltran, B., Amirat, Y., Yao, G., Han ,
J., 2014, “Second-order sliding mode control for DFIG-
based wind turbines fault ride-through capability
enhancement,” ISA transactions, 53, pp. 827-833.
[11] PK, D., RK, P., 2014, “Adaptive second order sliding
mode control of doubly fed induction generator in wind
energy conversion system,” Journal of Renewable and
Sustainable Energy, 6(5), AIP Publishing.
[12] Wu, B., Lang, Y., Zargari, N., Kouro, S., 2011, “Power
conversion and control of wind energy systems,“ John
Wiley & Sons, Inc., United States of America.
[13] Poitiers, F., Bouaouiche, T., Machmoum, M., 2009,
“Advanced control of a doubly-fed induction generator
for wind energy conversion”, Elect. Power Syst. Res.,
79(7), pp. 1085–1096.
[14] Justo, J., Mwasilu, F., Jung, J.W., 2015, “Doubly-fed
induction generator based wind turbines: A
comprehensive review of fault ride-through
strategies,” Renewable and sustainable energy
reviews, 45, pp. 447-467.
[15] Azar, A.T., 2016, “Advances and applications in
sliding mode control systems,” Q. Zhu (Ed.). Springer,
2016.
[16] Chiew Tsung, H., Zamberi, J., Ahmad Yusairi Bani, H.,
Lokman, A., Nur Aidawaty, R., 2017, “ Design of super
twisting algorithm for chattering suppression in
machine tools,” International Journal of Control,
Automation and Systems, 3(15), pp. 1259-1266.
[17] Shtessel. Y., Plestan, F., Taleb, M., 2012, “A novel
adaptive-gain supertwisting sliding mode controller,”
Methodology and application, Automatica, pp. 759–
769.
[18] Jason J, Bonnie J. NWTC information portal (FAST
V8). https://nwtc.nrel.gov/FAST8, 2015.

3532