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

net/publication/344460531

Wind energy conversion system control robustness based on current analysis

of IGBT open-circuit fault

Article in PRZEGLĄD ELEKTROTECHNICZNY · October 2020

DOI: 10.15199/48.2020.10.03

CITATIONS READS

2 301

4 authors:

Mohammedi Mokhtar Mahmoud Bendiabdellah Azeddine

University of Sciences & Technology of Oran Mohamed Boudiaf (USTO-MB) Oran, … Université des Sciences et de la Technologie d'Oran Mohamed Boudiaf
5 PUBLICATIONS 3 CITATIONS 171 PUBLICATIONS 1,028 CITATIONS

SEE PROFILE SEE PROFILE

Allaoui Tayeb Cherif Bilal Djamal Eddine

Université Ibn Khaldoun Tiaret University of M'Sila
140 PUBLICATIONS 785 CITATIONS 49 PUBLICATIONS 315 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mohammedi Mokhtar Mahmoud on 15 March 2022.

The user has requested enhancement of the downloaded file.

 Mohammedi Mokhtar Mahmoud1, Bendiabdellah Azzedine1, Allaoui Tayeb2 , Cherif Bilal Djamal Edine1
Group Diagnostic, Laboratory of Electrical Drives Development (LDEE), University of Sciences & Technology of Oran Mohamed Boudiaf
(USTO-MB) (1), University of Ibn Khaldun, Tiaret Algeria(2)

doi:10.15199/48.2020.10.03

Wind energy conversion system control robustness based on

current analysis of IGBT open-circuit fault
Abstract. This paper deals with the study of the performance of a wind energy conversion system (WECS) based on a doubly fed induction
generator (DFIG) under the IGBT open-circuit fault of the rotor side converter (RSC) during the application of robust control techniques, such as
backstepping control (BSC) and sliding mode control (SMC). The presence of IGBT open-circuit faults in DFIG-WECS can disrupt service continuity
resulting in financial loss. To overcome such a problem, robust control techniques are usually used as a solution. These control techniques are well
known for their ability to treat non-linear structures as power electronics converters and to maintain the performance and stability of the DFIG-WECS
connected to the network by the back-to-back converter in healthy and faulty operations. The aim of using robust non-linear control techniques is to
obtain better performance and to extend the DFIG-WECS functionality in degraded mode in the event of a failure, and consequently to increase its
reliability, unlike the proportional integral (PI) controller which shows less robustness when DFIG non-linearities are considered. The results obtained
from these control techniques illustrate well the merits and the effectiveness of each of them in the case of healthy and faulty operations, in particular
for the BSC technique, which shows a better performance compared to the SMC technique, which faces the main problem associated with
discontinuous control.

Streszczenie. Przedstawiono analizę właściwości systemu energii wiatrowej bazującego na igeneratorze DFIG w obwodzie IGBT z zastosowaniem
odpornego sterowania typu backstepping BSC I sterowania ślizgowego SMC. Zastosowano nieliniową technikę sterowania. Badania symulacyjne
wykazały odporność systemu zarówno w warunkach zdrowych jak i przy pojawieniu się błędów. System przetwarzania energii wiatrowej z
odpornym sterowaniem bazującym na analizie prądu w układzie IGBT

Keywords: open-switch fault, robust control techniques, system performance, current analyses.
Słowa kluczowe: energia wiatrowa, generator DGIG, odporne sterowanie

Introduction The present paper proposes the study of a DFIG-WECS

Using robust control techniques to keep system system, to investigate its performance under both the
operation under IGBT switch faults appearance until the healthy and the faulty cases. The existence of the open-
application of a hardware redundancy intervention to avoid circuit-fault in the RSC IGBT switch can disturb the
the interruption of service is a crucial issue. The reason is continuity of service of the system. To overcome such a
that, one should consider several constraints; among them problem, robust control techniques like BSC and SMC are
is the fault detection [1], the reduction of the fault detection proposed and used in this paper. The performance of these
needed time [2] and to define the assessment of the power control techniques have the ability to deal with nonlinear
electronic system reliability, such as the failure existence structures and variable configurations like power electronic
(the time between failure (TBF) until the time to repair (TTR) converters and also to maintain the performance and the
[3]). The fault-reconfiguration is generally founded on stability of the DFIG integrated to a wind-energy turbine
hardware redundancy design [4] [5] and most of the linked to the network by the back-to-back converter under
detection and tolerant methods are hardware based. healthy and rotor side open-switch fault operations.
It is therefore very significant for researchers to study and The robust control techniques are applied until a hardware
investigate the influence of the fault occurrence in the redundancy implementation and a future maintenance
robust control techniques presence. To carry out this study, operation are achieved. The paper also introduces a
the following system is considered: a wind turbine to extract performance comparative study between the proposed
maximum energy [6], which is integrated to a Doubly-Fed control methods and the results obtained illustrate well the
Induction Generator (DFIG). DFIG is one of the most used merits of each one of them during faulty conditions.
generator in wind energy applications with the largest part The proposed system investigated in this paper is a wind
of the installed Wind Energy Conversion Systems (WECS) system based on a DFIG, which is essentially composed of
using the variable speed technology [7] and also for the a doubly-fed induction generator, a back-to-back converter
back-to-back converter and its stability on the sub- linked to the grid by the grid side converter (GSC) and to
synchronous and super-synchronous operating modes [8]. the rotor by the rotor side converter (RSC) and a wind
The rotor side converter (RSC) control techniques are turbine supported by an MPPT control. Figure 1 represents
widely studied by researchers. The conventional control the PI controls synoptic scheme of the DFIG-WECS.
approaches for the DFIG are built on voltage and flux
oriented control techniques [9]. When the DFIG 1. DFIG-WECS description and modelisation
nonlinearities are considered, the classical vector control In the following, the model of each sub-system of the
techniques; using the proportional-integral (PI) [10]; show DFIG-WECS is to be presented and discussed.
less robustness and low performance. Hence, the nonlinear
control approaches such as the back stepping and sliding 1.1 Model of the turbine
mode control techniques have been more useful due to The turbine is a device that converts the kinetic wind
their various properties; their fast dynamic response, their power into mechanical power. The wind power is defined as
capability to deal with unmodeled dynamics and faults follows:
disturbance. These properties are appropriate to control ρ.S.vv3
both active and reactive powers [11], but both the sliding (1) Pv =
mode control (SMC) and the back stepping control (BSC) 2
techniques have been used in the WECS power regulation where: ρ –density of the air, S–circular area swept by the
mode respectively only with the objective to develop the blades of the turbine, VV–wind speed.
performance and the quality of this powers in the healthy
case [12] [13].

14 PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020

The mechanical power developed on the shaft of a wind R.Ω turbine
(3) λ=
turbine is expressed by [11] [14]: vv
ρ.S.vv3 where: Ωturbine –rotational speed of the turbine, R-radius of
(2) Paer = C p .Pv = C p  λ, β  .
2 the turbine

where: Cp(λ,β) _power coefficient, λ _specified speed given

as:

Fig.1. synoptic scheme of the DFIG-WECS based on PI controllers

The expression of the power coefficient is: The basic equation of the dynamic allows determining
21
the evolution of the mechanical speed from the overall
116 ( ) mechanical torque Cmec applied to the rotor:
(4) Cp =0.5176( -0.4β)exp λi +0.0068λi
λi dΩ mec
(7) J. = C mec
Where the coefficients of the equation (4) depend on the dt
considered turbine where: J –overall inertia of the rotor of the generator:
With: J turbine
(8) J= + J generator
(5) 1 1 0.035 G2
= -
λi λ +0.008β 1+ β 3 And G –gain of the speed multiplier.
The mechanical torque takes into account, the
electromagnetic torque Cem produced by the generator; the
The relation between Cp and λ for the given values of the
torque of the viscous frictions Cvis and the torque coming
angle of orientation of the blades β is represented by the
from the multiplier Cg.
following figure 2:
(9) Cmec = Cg - Cem - Cvis

1.2 Speed control and maximum wind power extraction

The control consists to adjust the electromagnetic torque
on the DFIG shaft. To accomplish this, one must perform
the control of the rotation speed, as shown in Figure 3.
Figure 4 depicts the speed adjustment.
Using the poles compensation method the regulator gains
are determined as follow:

(10) J
K PΩmec =
τ

(11) F
K iΩmec =
Fig.2. Pitch angle β effect on the aerodynamic coefficient of power τ
From the power, the wind turbine torque is given by:
where
3
(6) Paer ρ.S.v 1
C aer = = Cp. .
Ωturbine 2 Ωturbine

PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020 15

(12) J (16)
τ=
K pΩmec   Rs R sM sr
Φ sd = - L s Φ sd + ω s Φ sq + L ird + V sd
where F–overall friction coefficient  s
  Rs R s M sr
Φ sq = -ω sΦ sd - Φ sq + irq + V sq
 L s Ls
 
 i = R s M sr Φ - M sr ωΦ -  R r + R s M sr  i +
2

 rd σLr Ls 2 sd
σLr Ls
sq
 σLr σLr L2s 
rd

 M sr 1
+(ω s - ω)irq - V sd + V rd
 σLr L s σLr
  M sr R M
 irq = ωΦ sd + s sr Φ sq - (ω s - ω)ir d +
 σL L
r s σLr L2s

 -  R r + R s M sr  i - M sr V + 1 V
2

  σLr σLr L2s 

 rq
σLr L s
sq
σLr
rq

 P 2 M sr P f
 ω = L J (Φ sq ird - Φ sd irq ) - J Cr - J ω
 s

Fig.3. Block diagram of the turbine model With MPPT Control where: σ –Blondel dispersion coefficient given as:

(17) 
σ = 1 - M sr2 Lr .Ls 
with: Rs and Rr -stator and rotor resistances, Ls and Lr -stator
and rotor inductances, Msr -mutual inductance between
stator and rotor, ωs and ωr -stator and rotor angular speed,
Cr –load torque.

1.4 Active and reactive powers vector control

The idea of the stator flux orientation is to align with the
Fig.4. Closed loop for PI speed regulation d axis of the rotating frame the stator flux hence resulting in
Φsq=0 and Φsd=Φs. This approach helps the decoupling
1.3 Model of the DFIG between the active and the reactive powers.
The voltages equations of the DFIG in the d-q reference The equations (13) and (14) become [11] [15] [16]:
are [11]:
Φ sd = Φ s
 dΦ sd (18) 
V sd = R s i sd +
dt
- ω s Φ sq
 Φ sq = 0

 dΦ sq
(13) V sq = R s i sq + + ω s Φ sd  V sd = 0
dt 
  V sq = V s = ω s Φ s
V dΦ rd (19)
= R r ird + - ω r Φ rq  dΦ rd
 rd dt  V rd = R r ird + - ω r Φ rq
 dΦ rq  dt
V = R r irq + + ω r Φ rd  dΦ rq
 rq dt V rq = R r irq + + ω r Φ rd
 dt

 Φs = Ls isd + M sr ird
The flux equations in the d-q reference are:  0= L i +M i
(20)  s sq sr rq

Φ rd = L r ird + M sr isd

Φsd = Ls isd + M sr ird  Φrq = Lr irq + M sr isq


Φ = L i + M i
 sq s sq sr rq The rotor currents are written in terms of the stator powers
(14) 
Φ = L i + M as follow:
 rd r rd sr isd
 Φrq = Lr irq + M sr isq  V M
  Ps = - s sr i rq
(21)  Ls

Q = - V s M sr V s2
i rd +
 s Ls Ls ω s
The electromagnetic torque of the DFIG is:
M sr  Ls
(15) C em = p
Ls

Φ sq i rd - Φ sd i rq  (22) 

irq = -
V .M
.Ps
s sr

i = Vs - Ls .Q
2
The state model presenting the DFIG in the d-q  rd ωs .Ls Vs .M sr s
reference associated to the rotating field at the speed of
synchronism ωs is: The voltages control is given by:

16 PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020

   M sr2    M sr2   P
 Vrd = Rr ird +  Lr -  ird - gωs  Lr -  irq (32)  PF =
  Ls   Ls   S
(23)   S 2 = P 2 + Q 2
  M2    M2 
Vrq = Rr irq +  Lr - sr  irq + gωs  Lr - sr  ird + where: PF -power factor, P -active power measured value, S
  Ls   Ls 
 -apparent power, Q -reactive power measured value.
M
+g sr sV
 Ls 1.5.2 Rotor side control with power loop
1.5 RSC and GSC models
Figure 5 presents Block diagram of the PI indirect
The type of used RSC and GSC converters is a two-
control with power loop and Figure 6 gives the functional
level three-phase back-to-back converter, both the RSC
closed loop of the currents control.
and the GSC are controlled by the pulse width modulation
(PWM) technique. Their mathematical model is represented
by the control signals and the output phase voltages [17] as
follow:

 va   2 - 1 - 1  δ1 
(24)   vd c  -1 2 - 1   δ 
 vb  = 3   2
 vc   -1 - 1 2   δ3 

With: va, vb, vc –output phase voltages, δ1, δ2, δ3 - PWM top
switches gates control signals, vdc -DC bus voltage.

1.5.1 Grid side control

The currents grid side control leads to the following
equations:

 
v fd = R f .i fd + L f i fd + e fd Fig.5. Block diagram of the PI indirect control with power loop
(25)  
v = R .i + L i +e
 fq f fq f fq fq

with

e fd = ωs .L f .i fq
(26) 
e fq = -ωs .L f .i fd +vsq
Fig.6. Closed loop for PI currents regulation
Vfdq and ifdq–grid currents and voltages in the d-q reference,
Lf and Rf –inductance and resistance of the filter.

1.5.1.1 DC bus voltage vdc control

The dc bus powers are:
 Prectifier = vdc .irectifier

(27)  Pcapacitor = vdc .icapacitor
 Fig.7. Closed loop for PI powers regulation
 Pinverter = vdc .iinverter
The relation between these powers is: By applying the pole compensation method, the regulator
gains are determined as:
(28) Prectifier = Pcapacitor + Pinverter
Lr .σ
(33) k pc =
If the filter resistance Rf is neglected, equations (29) and τ1
(30) are respectively obtained as:

 Pf = vsq .i fq
Rr
(34) kic =
(29)  τ1
Q f = vsq .i fd
(30) Pf = Prectifier = Pcapacitor + Pinverter with:

The power reference for the capacitor is:

Lr .σ
(35) τ1 =
* *
k pc
(31) Pcapacitor = vdc .icapacitor
Figure 7 describes the functional closed loop of the powers
control.
The calculation of the power factor is one of the Again by applying the pole compensation method, the
techniques for measuring the quality of energy, the regulator gains are determined as:
definition of which is given in equation (32) [18].

PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020 17

 Lr .σ The solutions of the null Lyapunov derivative function are:
(36) k pp =
τ 2 .k pc  •   • 
(45)  e1 = -K1 .e1  and  e 2 = -K 2 .e2 
   
1
(37) kip =
τ2 With: K1 and K2 –positive gains.
By replacing (39) and (41) in (44), one obtains:
with
  Vrq - Rr irq - gωs Lr σird + 
Lr .σ  •   Ls • 
1  
(38) τ2 = e1 =   - Ps ref  -  -g M srVs 
k pp .k pc (46) M V  L σ
   sr s  r  L 
   s 
 •  
 e =   - Ls Q ref  - 1 V - R i +gω L σi 
•

2. Proposed control techniques used  2   M V s  L σ rd r rd  s r rq



  sr s  r 
The BSC and the SMC techniques are both robust
nonlinear techniques applied to the RSC DFIG-WECS Step 2:
control design with the aim to replace the PI current inner Replacing the derivatives of the error by their values, we
loop controller by the BSC or the SMC. obtain:

2.1 BSC applied to the DFIG-WECS   -Rrirq - gωs Lrσird +

  Ls • 
1 1  
The BSC technique considers the orientation of the -K e =
 1 1   - P ref
s   - Vrq - -g MsrVs 
stator flux such that Φsq= 0 and Φsd= Φs. (47)   MsrVs  Lr σ Lr σ  
   Ls 
The rotor derivatives currents are:   L •  
  
 -K e =- s Q ref  - 1 V - 1 -R i +gω L σi 

(39)
  1
 ird = Vrd - Rr ird + gωs Lr σirq L .σ
 2 2  M V s  L σ rd L σ r rd
  sr s  r r
s r rq


 r

i = V - R i - gω L σi - gω M srVs  The BSC voltage law is:
 rq  rq r rq s r rd s 
ωs Ls 
     L •  
  Lr σ  - s Ps ref +K1e1  +Rr irq + gωs Lr σird +
The derivatives of the rotor currents references are given in    M srVs  
 
the equation system (41): (48) Vrq =  
 M
+g sr s V 
  Ls 
 ref Ls ref   
 irq = - V M Ps    Ls •  
(40)  s sr  V =  L σ -
 Q ref +K2 e2  +Rr ird - gωs Lr σirq 
 rd  r  M srVs s  
i ref = Vs - Ls - Q ref     
 rd ωs M sr Vs M sr
s
The diagram represented by Figure 8 depicts the
 • ref ref application of the hybrid BSC to the DFIG.
Ls •
 i rq = - Ps
(41)  Vs M sr
 ref ref
 • Ls •
i rd = - V M Q s
 s sr

The combination of vector control with BSC allow

replacing PI regulators. The calculating of the BSC voltages
is based on Lyapunov functions [11], [12], [19], [20]. This
control is based on two steps, which are given as follow:
Step 1:
The Lyapunov function is:

(42) v=
2

1 2 2
e1 - e2 
Where the current errors e1 and e2 are given as:
Fig.8. Block diagram of the hybrid BSC

(43)

 e = i ref - i
 1 rq rq  2.2 SMC applied to the DFIG-WECS

ref
 e 2 = ird - ird
  The SMC is another powerful nonlinear controller, which
attracts many researchers and has been successfully
The derivatives of these errors are: applied in the DFIG-WECS [13], [21], with the idea to attract
the states of the DFIG-WECS in a suitable selected area
•  ref• •  then design a control law that maintain the system in this
 e1 =  irq - irq  area [22].
 
(44)   
 A) Selection of the sliding surfaces s(Ps) and s(Qs)
 •  ref
• • 
 e 2 =  ird - i rd  The rotor currents irq and ird are respectively images of the
   active and reactive powers. For that, the powers control
areas expression is:

18 PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020

(49)  s rq 
s  P  = iref -i
rq  (59)
  ref  
s  Ps  =  irq -irq 
  
  s  rd rd

s Q = i ref

-i   

 
B) Convergence condition When replacing the currents irq and ref
irq by their

The efficiency of the SMC is conditioned by verifying the expressions, we find:

following Lyapunov relationship of attraction: Vrq - Rr .irq - g.ωs .Lr .σ.ird +

(60  Ls 1  
 s  Ps  = - .Psref - VM
s  x  .s  x   0 . Vs .M sr Lr .σ  -g s sr 

 Ls 
C) SMC law calculation
 Ls 
1
The control algorithm is identified by: s  Ps  = - .Psref - .Vrq +
(61)  Vs .M sr Lr .σ

(50) u=ueq +uattr  1  Vs M sr 
- L .σ  -Rr .irq - g.ωs .Lr .σ.ird - g L 
 r  s 
uattr = umax .sign  s  x  
 L .L .σ ref
Where: u–control variable, ueq–equivalent command, uattr– Vrq = - s r .Ps +Rr .irq +g.ωs .Lr .σ.ird +
(62)  Vs .Msr
term control switch and sign(s(x)) –sign function. 
V .M
Let us now apply the SMC to the RSC DFIG-WECS.  Ls
r 1   
+g. s sr +L .σ.v .sgn s P

1. The reactive power control area and its d axis derivative
The control voltage is defined by: Vrq =Vrqeq +Vrqattr
expressions are given by:
(51) 
s  Qs  = ird
ref
-ird  Through the sliding mode and in the steady state, s(Ps) = 0,
d(s(Ps)) = 0, Vrqattr = 0 so the equivalent control is given by:

(52)   ref   
Ls .Lr .σ ref V .M
s  Qs  =  ird -ird  (63) Vrqeq = - .Ps +Rr .irq + g.ωs .Lr .σ.ird + g s sr
  Vs .M sr Ls
 

  Through the convergence mode, the condition s  x  .s  x   0
When replacing the currents ird and ref
ird by their
should be verified with the relation (64):
expressions, it is found:
(64) Vrqattr = Lr .σ.v1 .sgn  s  Ps  
  Vs Ls   1 Vrd - Rr .ird + 
(53) s  Qs  =  - .Qref -
 ωs .Msr Vs .Msr s  Lr .σ +g.ωs .Lr .σ.irq  Where the product term: (Lr.σ.v1)–a positive gain.
   

  Vs   The diagram shown in Figure 9 represents the application

Ls
s  Qs  =  - .Qsref  + of the hybrid SMC to the DFIG
(54)   
  ωs .Msr Vs .Msr 
 1
- .V -
1
 L .σ rd L .σ r rd -R .i+g.ω s r .σ.irq
.L 
 r r

  Vs Ls  

(55) Vrd = Lr .σ  - .Qref  +R .i +

 ωs .M sr Vs .M sr s  r rd
  

-g.ω .L
s r .σ.irq +L r .σ.v2 .sgn   
s Q

The control voltage is defined by the relation: Vrd =Vrdeq +Vrdattr

Through the sliding mode at steady state, s(Qs) = 0, d(s(Qs))

= 0, Vrdattr = 0, so the equivalent control is given by:
 Vs Ls  
(56) Vrdeq = Lr .σ  - .Qsref  +Rr .ird - g.ωs .Lr .σ.irq
 ωs .Msr Vs .Msr 
 

Through the convergence mode, the condition s  x  .s  x   0 Fig.9. Block diagram of the hybrid SMC
should be verified with the relation (57):
3. Simulation results
(57) Vrdattr = Lr .σ.v2 .sgn  s  Q s  
The proposed control techniques are supported by a
simulation study using the Matlab / Simulink environment to
Where the product term (Lr.σ.v2) is a positive gain.
verify their effectiveness. The simulation results are
2. Similarly, the active power control area and its q axis conducted to investigate the performance of a wind turbine
derivative expressions are given by: related to a 7.5 MW DFIG, a three-phase voltage source at
the stator with a sinusoidal network of 220/380V, 50 HZ and
(58) 
s  Ps  = irq
ref
-irq  with the RSC and GSC switches commutation frequency
fixed at 2KHZ. All simulation results related to control
techniques are studied and presented for healthy and faulty
operations.

PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020 19

3.1 DFIG-WECS in the healthy case
A) Use of the classical PI controllers

Fig.15. Current ifa and grid voltage Va

Fig.10. Mechanical speed

Fig.16.DC link voltage

Fig.11. Electromagnetic torque

Fig.17. Power coefficient Cp

Fig.12. Stator active power

Fig.18. Power factor

Fig.13. Stator reactive power The model of the applied wind profile reference [23] is
expressed by a summation of several harmonics in a
deterministic form given by:

vv (t)= 6.5+(0.5* sin(0.1047* t)+ 2* sin(0.2665* t)+

(65)
sin(1.2930* t)+0.2* sin(3.6645* t))

To extract the maximum of the wind active power, the

angle
of orientation of the blades is chosen to be β=0.
The DFIG rotor mechanical speed is determined by the
available speed at the turbine stage shown in figure 03.
Figure 10 shows a good following of the mechanical speed
Fig.14. Three-phase rotor current Ω mec to it reference Ω ref in the presence of the PI-MPPT

20 PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020

controller. From this control study, it results an
electromagnetic torque reference Cem ref. The staror active
power reference is obtained from:

(66) Ps ref = Cem ref ×Ω mec

Figures 12 and 13 depict a good tracking of the active
power Ps to it reference Ps ref and the reactive power Qs to
it reference Qs ref in the presence of the PI powers
controllers. As a result and in a closed loop system, the
electromagnetic torque Cem shown in Figure 11 follows well
its reference Cem ref.
Figure 14 depicts the sinusoidal form of the three-phase
rotor currents (ira, irb, irc) in the (a, b, c) referential included Fig.21. Measured stator active powers under steady state
between [-15, 15 A]. Whereas, the rotor current increase
between the interval [5, 6 s] is due to the optimal stator
active power reference determined by the available power
at the turbine stage shown by equation (66). This affects the
currents in relation to the mechanical speed increase
between [5, 6 s] shown in figure 10. This study is carried out
with the aim of reaching a point of maximum MPPT tracking
power. It is to be noted that in Figure 15, the grid phase
voltage Va and its current ifa are shown when the DFIG
operates at sub-synchronous mode. Figure 16 represents
the control of the DC bus voltage vdc which depicts a good
tracking of the reference vdc ref = 600v in the presence of
the PI controller. Figure 17 represents the power coefficient Fig.22. Measured stator reactive powers
Cp. Finally, Figure 18 represents the power factor PF
performance when the stator reactive power is fixed at
Qs_ref =0 Var, which is for a near unity power factor case.
This is a significant indication for the good effect of the
controller for the grid power factor.

B) Use of the robust BSC and SMC techniques in the RSC:

Fig.23. Measured stator reactive powers under transient state

Fig.19. Measured stator active powers

Fig.24. Measured stator reactive powers under steady state

The simulation results obtained when applying robust

control techniques as highlighted by the various figures
above, show a good performance and an increasing
improvement with respect to the classical PI controllers.
This is obviously observed from the overshoots and the
thickness of the measured values in both transient and
steady state conditions. The transient sates of the active
and reactive powers as shown respectively by figures 20,
23, illustrate well the enhanced convergence to the steady
state of the proposed control techniques compared to the PI
controller. On the other hand, the steady states of the active
Fig.20. Measured stator active powers under transient state
and reactive powers as depicted respectively by figures 21,
24, show the controls overshoots of the active powers
measured values. These are respectively represented by
the following intervals at time t = 1s: [-5900, -5000 Wat] for

PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020 21

the PI controller, [-5700, -5100 Wat] for the SMC technique,
[-5600, -5200 Wat] for the BSC technique. The controls
overshoots of the reactive powers measured values shown
by figure 24 are respectively represented by the following
intervals [-1400, -550 Var] for the PI controller, [-1350, -650
Var] for the SMC technique and, [-1280, -700 Var] for the
BSC technique. It can be therefore concluded that the BSC
technique gives a minimized overshoots response
compared to the PI controller and the SMC technique. This
does not prevent that one of the strong points of the robust
control techniques lies in the good decoupling between the
two components of the stator-generated powers. The
proposed control techniques using either the BSC or the Fig.28. Rotor current irc under fault
SMC allows convergence improvement to the steady state
with a close to zero steady state error.

3.2 DFIG-WECS current analysis in the faulty case

The aim of using the robust control techniques is to
enhance the performance of the DFIG-WECS in the
degraded mode, which is defined from the fault appearance
until the hardware compensation implementation. The IGBT
open-circuit fault is therefore applied twice, in the RSC S1
switch shown in figure 25, where the faulty leg correspond
directly to the current ira. The fault is applied at the
beginning of the current ira periods (see section A and B)
with the corresponding times, first=1.02s then t=1.06s. Fig.29. Measured stator active powers under fault

Fig.30. Measured stator reactive powers under fault

Fig.25. RSC during S1 switch open-circuit fault

A) Applied fault at time t = 1.02 s with a zoom of one period:

Fig.31. DFIG rotor mechanical speed under fault

Fig.26. Rotor current ira under fault B) Applied fault at time t = 1.06 s with a zoom of one period:

Fig.27. Rotor current irb under fault Fig.32. Rotor current ira under fault

22 PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020

 Figures 26, 27, 28, 29, 30 and 31 show the behavior of the
currents ira , irb and irc, the active and reactive powers and
the DFIG rotor mechanical speed measured values when
applying an open-circuit fault in the switch S1 at t=1.02s
which corresponds to a beginning of a negative alternation
(ira<0 until t=1.06s). During this time interval, in the
presence of the open-circuit fault, the diode D1 allows the
current ira to flow in it. It can be seen that the waveforms
are not affected and the DFIG-WECS system continues to
generate the electrical power in nominal conditions. Now,
from t=1.06s until t =1.1s, a positive current ira waveform
occurs in the healthy case. It should be noted that in the
presence of the inverter IGBT open-circuit fault case and
Fig.33. Rotor current irb under fault during the blocking period of the diode D1, the current will
become zero. The different figures illustrate the behavior of
the applied control techniques where it can clearly be seen
that the SMC and BSC performances are better than those
of the PI controller.
When using SMC or BSC control techniques
preceded by fault detection and isolation, the behavior
allows the selected applied redundancy to reduce the
performance of the DFIG-WECS system to nominal
conditions such as that of a healthier condition. When
applying the fault at t=1.06s which corresponds to a positive
alternation of ira, the same analyses are considered in
figures 32, 33, 34, 35, 36 and 37. It should be noted from
Fig.34. Rotor current irc under fault the figures that the differences in this case is that the diode
D1 is immediately blocked after the appearance of the fault
and the waveforms are now affected. It is clearly observed
from the applied control techniques that the SMC and the
BSC show a better performance than the PI controller does
even when the negative waveform appear at t=1.1s where it
is supposed that the diode D1 is not blocked and the DFIG-
WECS work in nominal conditions. It can be seen from the
powers and the DFIG rotor mechanical speed represented
respectively by figures 35, 36 and 37, that the DFIG-WECS
based on the PI controllers show an important overshoots
and a slow restoration into its initials values. On the other
hand by using either the BSC or the SMC techniques, we
get to minimize the open-circuit fault transient state in fault
appearance degraded mode by the enhancement of the
Fig.35. Measured stator active powers under fault
fault overshoots and the settling times resulting therefore in
a fast dynamic response. This behavior enables the
selected applied redundancy to return the DFIG-WECS
system performance to the nominal conditions as that of a
healthier state case. This confirms the robustness of our
proposed control techniques compared to the PI controller.
It is also important to know that it is crucial for the continuity
of service of the DFIG-WECS to reconfigure the fault
immediately after its occurrence by installing a hardware
redundancy.

Conclusion
The present paper proposes the study of the
Fig.36. Measured stator reactive powers under fault performance of a DFIG-WECS under healthy and faulty
conditions, where the fault is introduced as an open-circuit
fault in the RSC IGBT switches. A performance comparison
is carried out to analyze the behavior of the DFIG-WECS
under healthy and faulty cases by proposing the use of
three RSC control strategies. First a classical indirect PI
controller based on the pole compensation is presented and
considered as a reference well known approach, then both
a hybrid BSC technique based on the stability theory of
Lyapunov and a hybrid SMC technique are used. In the
healthy case, the BSC and SMC techniques show a good
power decoupling performance compared to the PI
controller. For the faulty case, the performances of the three
control techniques are proposed to investigate how to deal
Fig.37. DFIG rotor mechanical speed under fault with the RSC open-circuit fault presence. It is found that

PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020 23

both the SMC and the BSC techniques show a better International Conference on Trends in Electronics and
performance compared to the PI controller; this is due to the Informatics (ICOEI 2018).
strong power decoupling and the fast response of these [10] Van T., Nguyen T., Lee D (2015). “Advanced Pitch Angle
Control Based on Fuzzy Logic for Variable-Speed Wind
robust control techniques. It is important to note that the
Turbine Systems“. IEEE Trans. Energy Convers. 2015, PP, 1–
SMC technique faces the main problem associated with the 10.
discontinuous control, which inevitably leads to a chattering [11] Belkacem B., Tayeb A., Mohamed T., Mouloud D, (2019)
phenomenon. This behavior gives a robustness advance “Comparative study of back-stepping controller and super
margin advantage to the BSC technique compared to the twisting sliding mode controller for indirect power control of
SMC technique. The major objective of the use of the wind generator”, Int J Syst Assur Eng Manag
robust control techniques is to benefit from the performance https://doi.org/10.1007/s13198-019-00905-7.
and robustness of the DFGI-WECS and to maintain its [12] Riyadh Rouabhi ., Chouder Aissa and Ali Djerioui. (2015)
"hybride backstepping control of a doubly fed wind energy
operation in acceptable conditions after the fault occurrence
induction generator "reasearchgate, The Mediterranean Journal
until the implementation of a hardware redundancy of Measurement and Control, Vol. 11 No. 1, 2015.
guaranteeing the continuity of the service. [13] Riyadh Rouabhi ., Rachid Abdessemed., Aissa Chouder and
Authors: PHD student Mohammedi. Mokhtar. Mahmoud, Ali Djerioui (2015) “Power Quality Enhancement of Grid
University of Sciences & Technology Mohamed Boudiaf (USTO- Connected Doubly-Fed Induction Generator Using Sliding
MB), BP 1505 El-M'naouar, Oran, Algeria, Mode Control“ International Review of Electrical Engineering
Email mokhtar.mohammedi@yahoo.fr (I.R.E.E.), Vol. 10, N. 2 apr 2015.
Pr. Bendiabdellah. Azeddine, University of Sciences & Technology [14] OM PRAKASH M., NEERAJ G., MAHDI K., NILESH P, (2019)
Mohamed Boudiaf (USTO-MB), BP 1505 El-M'naouar, Oran, “Comprehensive Overview of Low Voltage Ride Through
Algeria, Methods of Grid Integrated Wind Generator “, Comprehensive
Email bendiazz@yahoo.fr Overview of LVRT Methods of Grid Integrated Wind Generator,
Pr Allaoui. Tayeb, University of Ibn Khaldun, BP P 78 zaâroura VOLUME 7, 2019, DOI 10.1109/ACCESS.2019.2930413.
14000, Tiaret, Algeria, [15] Nur S., Paul M., Judith M., Yingzhao W, (2018) “DFIG Stator
Email allaoui_tb@yahoo.fr Flux Oriented Control Scheme Execution for Test Facilities
Dr Cherif. Bilal. Djamal Edine, University of Sciences & Technology Utilising Commercial Converters “, IET Renewable Power
Mohamed Boudiaf (USTO-MB), BP 1505 El-M'naouar, Oran, Generation 2018, DOI: 10.1049/iet-rpg.2018.5195.
Algeria, [16] Bossoufi B., Karim M., Lagrioui A., Taoussi M., Derouich A.,
Email cherif.doc84@gmail.com (2015) “Observer Backstepping control of DFIG-Generators for
Wind Turbines Variable-Speed: FPGA Based Implementation”
REFERENCES Renewable Energy Journal (ELSIVER), pp 903-917, Vol. 81.
[1] Marcelo N., Yves M., Michel K., Jan H., Johan G. (2018) September 2015.
“Robust Power-Electronic Converter Fault Detection and [17] Benbouhenni H., Boudjema Z., Belaidi A., Yingzhao W, (2018)
Isolation Technique for DFIG Wind Turbines”, IOP Conf. “Direct Vector Control of a DFIG Supplied by an Intelligent SVM
Series: Journal of Physics: Conf. Series 1037 032043 2018. Inverter for Wind Turbine System “,Iranian Journal of Electrical
[2] ZHANG J., CHEN Y., CHEN Z., ZHOU A. (2019) “Open-Switch and Electronic Engineering, Vol. 15, No. 1, March 2019
Fault Diagnosis Method in Voltage-Source Inverters Based on DOI:10.22068/IJEEE.15.1.45.
Phase Currents”, IEEE access VOLUME 7 [18] Ciprian M., Adriana F., Constantin D., (2020) “Improving the
doi:10.1109/ACCESS.2019.2913164 . E_ciency and Sustainability of Power Systems Using
[3] Stefan B., Elena K., Jyotirmoy R., Ralph G. (2018) “Impact of Distributed Power Factor Correction Methods”. Sustainability
Component Reliability on Large Scale Photovoltaic Systems’ 2020, 12, 3134; doi:10.3390/su12083134.
Performance”, Energies 2018, 11, 1579; [19] Sara M., Ahmed E., Issam M., Tamou N, (2018)
doi:10.3390/en11061579 . “Backstepping Controller for a Variable Wind Speed Energy
[4] Niraj K., Tapan P., Jitendra K., Krishna K. (2019) "Open‑circuit Conversion System Based on a DFIG “,International Science
fault‑tolerance in multilevel inverters with reduced component Index, Electrical and Computer Engineering Vol:12, No:9, 2018
count" Electrical Engineering, https://doi.org/10.1007/s00202- waset.org/Publication/10009495.
019-00884-9 2019. [20] Mohamed N., Ahmed E., Tamou N, (2020) “Coordinated
[5] Ajaykumar T., Nita R. (2019) "Fault-tolerant switched capacitor– Control Using Backstepping of DFIG-Based Wind Turbine for
based boost multilevel inverter" Electrical Engineering, Int J Frequency Regulation in High Wind Energy Penetrated System
Circ Theor Appl. 2019;1–15., DOI: 10.1002/cta.2671 2019. “,Mathematical Problems in Engineering, Volume 2020, Article
[6] Ayushi S., Akhilesh K., Paulson S, (2017) “A Review of MPPT ID 8287949, 16 pages, https://doi.org/10.1155/2020/8287949.
Algorithms Employed in Wind Energy Conversion Systems”, [21] Douadi T., Harbouche Y., Abdessemed R.., Bakhti I, (2018)
Journal of Green Engineering, Vol. 6 4, 385–402, doi: “Improvement Performances of Active and Reactive Power
10.13052/jge1904-4720.643. Control Applied to DFIG for Variable Speed Wind Turbine
[7] Amina Tamer., Djilali Toumi., Azzedine Bendiabdellah., Bilal Using Sliding Mode Control and FOC “, International Journal of
Djamal Eddine Cherif.,(2017) ’’Open Switch Fault Detection Engineering (IJE), IJE TRANSACTIONS A: Basics Vol. 31, No.
and Fault Tolerant of Power Converter Fed DFIG in WECS’’, 10, (October 2018) 1689-1697.
International Review of Automatic Control (I.RE.A.CO.), Vol. [22] Larbi D., Edgar N., Mohammed B, (2020) “First and High Order
10, N. 3, 2017 Copyright © 2017 Praise Worthy Prize. Sliding Mode Control of a DFIG-Based Wind Turbine “,Electric
[8] Mahalakshmi R., Sindhu T., (2019) “Detailed modelling and Power Components and Systems, DOI:
development of a laboratory prototype for the analysis of 10.1080/15325008.2020.1758836.
subsynchronous resonance in DFIG based wind farm”, Int [23] Abdenour Abdelli., (2007) “Optimisation multicritère d’une
Trans Electr Energ Syst.2020;e12245, DOI: 10.1002/2050 chaîne éolienne passive “. PHD thesis of the national
7038.12245. polytechnic institute of Toulouse., Plasma and Energy
[9] Asha Ram K, (2018) “Vector Control Strategy to Control Active Conversion laboratory, order N°: 2519.
and Reactive Power of Doubly Fed Induction Generator Based
Wind Energy Conversion system”, Proceedings of the 2nd

24 PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 10/2020

View publication stats