9A1-4 The 2nd IEEE Conference on Power Engineering and Renewable Energy

ICPERE 2014

Control of the Doubly Fed Induction Generator in

WECS
Boumaraf Farid Abdessemed Rachide Bendaas M. Lokmane
LEB- Research Laboratory LEB- Research Laboratory LEB- Research Laboratory
Department of Electrical Engineering Department of Electrical Engineering Department of Electrical Engineering
University of Batna – Algeria University of Batna – Algeria University of Batna – Algeria
[email protected] [email protected]

Abstract— The major part of the existing research works through a frequency converter introduces itself as one of the
concerning variable speed wind turbines control is not only to most important generators for WECS because of the following
capture the maximum power from the wind but, also, to improve reasons: a higher energy yield, a reduction of mechanical loads,
the quality of power and to converge the system for operating at an extensive controllability of both active and reactive powers,
unity power factor. This paper proposes a new power control of a less fluctuation in output power, a simpler pitch control. The
variable speed wind energy conversion system based on a doubly major advantage associated with this topology is that the power
Fed Induction Generator (DFIG) using space vector modulation converters could be rated at only 30% of the overall generator
to achieve control of active and reactive powers exchanged power.
between the stator of the DFIG and the grid to ensure a
Maximum Power Point Tracking (MPPT) of a wind energy Many control schemes have been proposed, among the
conversion system and to reduce significantly Powers and Torque vector control based on rotor current decoupling, which are
ripples. The SVM control is applied to rotor side converter for either stator flux oriented [1, 2] or stator voltage oriented [3, 4].
active and reactive power control and voltage regulation of wind The scheme decouples DFIG rotor current into two parts
turbine. This new approach is validated by modeling using representing active and reactive power separately. The control
Matlab-Simulink software and simulation results can prove the of machine active and reactive power is achieved by adjusting
excellent performance of this control as improving power quality these two rotor currents using PI controllers. However, the
and stability of wind turbine.
performance of a DFIG depends not only on the induction
Keywords- Wind turbine; SVM control; Doubly fed induction
machine but also on the two PWM converters as well as how
generator; Variable speed; WECS; Pitch control. they are controlled. In order to improved the DFIG power
generation characteristics under different conditions, this paper
I. INTRODUCTION presents a detailed analysis of the SVM control technique.
During the last decade, wind energy sources have become A. Model of Turbine
more popular due to the increasing concern about the emission The aerodynamic performance of horizontal axis wind
caused by the fossil fuels and the security issues regarding turbine expressed by the power coefficient Cp is written as
nuclear. The popularity of the wind energy has always been in follows[3]:
relation with the fossil fuel prices. After World War II with 1
decreasing fossil fuel prices, wind energy lost its popularity; p t = c p ( λ , )..s.v 3 (1)
however the oil crisis in 1970s boosted interest in large scale 2
energy production via wind turbines . At present the wind The power coefficient Cp depends on the speed ratio 
power costs, together with incentives and financing options to (ratio between blade tip speed and wind speed) and  the pitch
support the renewable energy facilities make wind energy angle of rotor blades. It is possible to define the coefficient Cp
sources competitive with conventional generation sources. It is by a mathematical approximation:
believed that wind energy will be the most cost-effective -C 5
C2
source of electrical power in the next future. However and due C p (, ) = C1( -C 3 . − C 4 )e i
+ C6  (2)
to its intermittent nature, there are some problems associated i
with this type of energy, especially for grid application [1]. In With:
fact, many technologies were used in wind power generation in c1 = 0.5176, c2 = 116, c3 = 0.4, c4 = 5, c5 = 21, c6 = 0.0068
order to find the best solution for integrating a high power The ratio of the speed  is defined by the expression:
quality to the distribution network. Recently, the most widely

used topology for high power grid applications in wind energy  = t .R (3)
conversion systems (WECS) is the one based on the doubly fed v
induction generator (DFIG) [2]. The doubly fed induction Fig. 1 represents this coefficient in function of  for
generator (DFIG) which its stator winding is directly connected different values of pitch angle  of the blades. This curve is
to the grid and its rotor winding is connected to the grid characterized by the optimal point (opt=9, Cp-max=0.5, =2°)

978-1-4799-6402-4/14/$31.00 ©2014 IEEE 25

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:05:08 UTC from IEEE Xplore. Restrictions apply.
9A1-4 The 2nd IEEE Conference on Power Engineering and Renewable Energy
ICPERE 2014

corresponding to the maximum power coefficient Cp and A. 1 Pitch control algorithm

therefore the maximum mechanical power recovered. For generation of the reference angle ref, the generated
0.6 power Pg, is compared with the reference power Pref, the error
β =2
β =4
in them, and sent to the PI controller that generates the
0.5 β =6 reference value of the angle ref.
β =8
β =10 When the wind speed is above nominal, the control objective
0.4 β =12
β =14
is to maintain the stationary output power constant by means
β =16 of pitch control. However, the pitch angle is not able to reach
Cp

0.3
β =18
β =20 the set point value immediately. Accordingly, a rate limiter is
0.2 β =22 implemented in the pitch controller model [5].The
β =24
β =26 configuration is based on a closed loop structure.
0.1 β =28
β =30
β =32  ref
0
0 5 10 15 20 25 Pref 
λ
+ PI +
Figure 1. Characteristic Cp=f(opt) − −
Pitch limit Rate limiter
Knowing the speed of the turbine t, the aerodynamic P
torque is directly determined by: Figure 3. pitch controller model
P C  2 R t3 v 2
Taer = aer = p (4)
B. Model of DFIG
t 2
The DFIG voltage and flux equations, expressed in the Park
The multiplier adapts the slow speed of the blades to the reference frame, are given by:
fast speed of the machine by the multiplication ratio G. This d
element defines the mechanical torque and speed of the V sd = R s I sd +  sd −  s  sq
machine as follows: dt (9)
d
Taer V sq = R s I sq +  sq +  s  sd
Tg = (5) dt
G  sd = L s I sd + MI rd
  sq = L s I sq + MI rq
G= t (6) (10)
g  rd = L r I rd + MI sd
For our model we'll assume the stiffness and damping are  rq = L r I rq + MI sq
neglected, and then the dynamic equations of the drive-train
Moreover, the electromagnetic torque is given by:
can be obtained with a model to a mass in this case described
by:
d m Tem = p
M
(I rq sd − I rd sq ) (11)
T g − Tem =J (7) Ls
dt The stator resistance of the DFIG is neglected and the stator
The equivalent moment of inertia is: flux φs is set aligned with the d axis and assumed to be
Jt constant (it is the case of a powerful and stable grid) [12].
J = Jg + (8) Then, we can write  sd =  s , sq = 0
G2 Consequently, the stator voltages and fluxes can be
Fig. 2 can define a physical model of the turbine having as rewritten as follows:
inputs the pitch of the blades, wind speed and electromagnetic d
torque provided by the generator. V sd = s = 0 (12)
dt
Ωt

0.6

0.5  V sq =  s  s
0.4
R 1
 sd = L s I sd + MI rd
0.3
C p

0.2

V G
0.1
(13)
0 = L s I sq + MI rq
0
0 5 10 15 20 25
Lambda

T aer Hence, the torque equation can be written as :

Tg M
V 1
.C p .ρ.Π.R2.
V 2
1 +- 1 Tem = p I rq sd (14)
js + f Ls
2 Ωt G Ωm
Tem In order to provide a decoupled indirect control of the stator
active power Ps and reactive power Qs of the DFIAG by
Figure 2. Wind Turbine Model means of the rotor current regulation, the d-q components of

26
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:05:08 UTC from IEEE Xplore. Restrictions apply.
9A1-4 The 2nd IEEE Conference on Power Engineering and Renewable Energy
ICPERE 2014

the rotor currents are defined in the stator flux oriented di f 1

reference frame[5]. Vf1 = −Rf if1 − lf + v s1
dt
The main is to represent Ps and Qs as functions of the
individual rotor current components as: di f 2 (17)
V f 2 = −R f if 2 − l f + vs 2
dt
M di f 3
P = −Vs I rq
Ls Vf 3 = −Rf if 3 − lf + vs3
(15) dt
M 
Q = −Vs I rd + Vs s Io Ired
Ls Ls S1 S3 T5
The rotor voltage equations can be defined as:
lf Rf
M2 dIrd M2 ic if1
Vrd = Rr Ird +(Lr − ) − gs(Lr − )Irq
Ls dt Ls

Réseau
Vf1 Vs1
(16) Vdc
2 dIrq 2
M M MV
Vrq = Rr Irq +(Lr − ) + gs(Lr − )Ird + g s Vf2
Vs2
Ls dt Ls Ls

C. Rotor Side Converter(RSC) Vf3 Vs3

S4 S6 S2
The rotor side converter is a two-level inverter and has only
Figure 5. Grid side converter.
8 switching states. This converter controls the active and
reactive powers generated by the stator of the DFIG.
In the d, q reference frame:
The space vector modulation (SVM) is used for its
modulation strategy so as to improve its modulation di fd
performance compared with PWM [6]. A block diagram of
V fd = − R f i fd − l f + s l f i fq + vsd
dt
DFIG Control strategy is shown in Fig.4. (18)
di fq
V fq = − R f i fq − l f + s l f i fd + vsq
−Ls g
MV s dt
MV s
Ls The active and reactive powers (Pf and Qf ) can be both
expressed by using Park components of supply voltages (vfd
and vfq) and line currents (ifd and ifq) as follows[3]:
Pref
+− +
+ PI PI + P f = v fd i fd + v q i fq
− +
P Ird Vrq (19)
2
SVM Q f = v fq i fd − v d i fq
M
Irq g s (Lr − ) dq We can find the reference currents (ifd-ref, ifq-ref), which
Ls
allows setting the desired reference active and reactive powers
Ird
 (Pf-ref, Qf-ref), as follows:
M2
gs (Lr − ) Q
= f-ref
RSC

V s2 Ls i fd-ref
L s s Vrd v sq
(20)
Irq P f-ref
Q ref
+ − =
+ PI
−
+ +− PI + D F IG
i fq-ref
v sq
Q −L s The unity power factor is obtained simply by setting the
MVs reactive power reference null. The DC reference voltage Vdc-ref
is compared to the measured DC voltage across the capacitor
Figure 4. DFIG control strategy. Vdc . The DC voltage corrector regulates the DC bus and sets
the active power Pc-ref which is necessary to charge the
D. Grid Side Converter(GSC)
capacitor to the desired value [7]. The reference active power
The objective of the grid side converter is to regulate the Pf- ref is obtained after being calculated the Pc-ref. The evolution
DC-link voltage and to set a unit power factor [7]. The of the DC voltage Vdc is given by the following equation:
mathematical model of the inverter in three-phase static abc dv dc
coordinates can be drawn as follows [3]: C. = (i red − i o )
dt
i c = i red − i o (21)

27
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:05:08 UTC from IEEE Xplore. Restrictions apply.
9A1-4 The 2nd IEEE Conference on Power Engineering and Renewable Energy
ICPERE 2014

i c −ref Po Po
P c −ref i
vd c - ref Pf -ref fq-ref i fq Pf Pc ic vdc
The vector Vref is given by two components V and V, an
* ++
PI − 1 1
+ * G * +
− v dc C.s algorithm for Research angular sector is used to determine the
vd c * v sq / v sq / sector (k= 1,2,3 ....6) as shown in Figure (9),
vd c
Vα , Vβ
Figure 6. Bloc diagram of the DC bus control [3].

Fig 7 presents the basic configuration of grid side converter

control strategy: Yes No
V , V
i fd - ref vfd - ref
Q f - ref
* + PI − PI
v sq − + dq
/ i fd
Yes No No
ifq Lf.s ifd Vα > 0 Vα > 0
Lf.s
v sq / i fq - ref

vd c-ref vf q-ref
−
+ + +
− PI * + −
PI −
+ Yes No Yes No Yes No Yes No
* v sq
vd c * ifq V− 3V<0 V+ 3V<0 V+ 3V>0 V− 3V>0
* *
SVM

io vd c
GSC

vd c K=1 K = 2 K=3 K=2 K=5 K=6 K=5 K=4

ic
io i red Figure 9. Algorithm for Determining the Sector K
lf
i fd
2 i f1 Rf
i fq 3 i f2 The tensions V , V and the three variables X, Y, Z are
v sq v s1 given by the following equations:
v sd PLL
s v s2 ­ T
°X = 3 * *Vβ
Figure 7. Grid side converter control. ° VDC
°° 3 T 3 T
II. SVM CONTROL STRATEGY ®Y = * *Vβ + * *Vα (22)
° 2 VDC 2 VDC
The principle of the Space Vector Modulation (SVM) is to °
reconstruct the vector Vref from eight voltage vectors. Each of °Y = 3 * T *Vβ − 3 * T *Vα
these vectors corresponds to a state of switches of a three- °̄ 2 VDC 2 VDC
phase voltage inverter, [8], [9], [10], [11]. The 6 switches can For high switching frequencies, the reference voltage vector
take 23 different states, these 8 states are coded S0 to S7. can be considered constant during each switching period. The
Modulation "SV" is to control the switches to switch PWM on-duration of active vectors V1 and V2 that are adjacent to
gradually from one state to the next, or S1, S6 represented Vref are found by solving the vectorial equation:
active switching ([011] [001] [101] [100] [110] [010]), and T
two zero voltage vectors , i.e. S0 and S7, corresponding to V ref . = V 1 .T 1 + V 2 .T 2
2 (23)
switching states ([000] and [111]). Change of state [111]
T
refers to the three upper transistors are switched on. These T0 = − T1 − T m
statements are transformed into voltage vectors V0 to V7 as 2
Where T1, Tm, T0 represent the time widths for vectors V1,
showing in figure (8).
V2, V0. T0 is the period in a sampling period for null vectors

should be filled, as each switching period (half of sampling
V 3 (010) V 2 (110)
period) starts and ends with zero vectors i.e. there will be two
Sector2
zero vectors per T/2 or four null vectors per T, duration of
Sector V1 (100)
each null vector is To/4 [12].
Sector3 V ref α
The values of the times T1 and Tm for each sector are
V 4 (011) calculated from the values of X, Y and Z, figure (10), in the
Sector 4 Sector 6 V 0 (000) table (1).
Sector 5
K 1 2 3 4 5 6
V5 (001) V 7 (111)
T1 Z Y -Z -X X Y
V 6 (101) Tm Y -X X Z -Y -Z
Figure 8. Diagram of Voltage Space Vectors Table1 : Determination of X,Y,Z

28
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:05:08 UTC from IEEE Xplore. Restrictions apply.
9A1-4 The 2nd IEEE Conference on Power Engineering and Renewable Energy
ICPERE 2014

S The simulation results of the active and reactive powers

control of the WECS based DFIG are shown in figure 12. It is
Z clearly noticed that the measured active and reactive powers of
Y the DFIG follow respectively their references.
K For DC-link voltage regulation, the reference value for DC-
link voltage is set to 660v. It is clearly shown from Fig.12, that
S K
T1 the measured DC voltage follows its reference. This result
confirms that the modified control of the GSC of the WECS
K
can regulate the DC link voltage and set a unit power factor.
K F(n) DFIG rotor current and DFIG stator current. It is obvious, that
Tm
S the rotor current frequency is lower than the stator current
frequency.
Y
9

X 8
K

Wind speed(M/s)
7
Z K
T 6

K 5

K 4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Time(S)

Figure 10. Calculation model of times T1, Tm 350

300

The modulating signals are determined by the following 250

Ω m (Rd/s)

200
equations and figure (11) show the model simulink:
150
­Taon = (T - T 1 - Tm ) / 4
° 100

®Tbon = (Taon + T ) / 2 (24) 50

°Tcon = (Tbon + T ) / 2
¯ 0
0 0.1 0.2 0.3 0.4
Time(S)
0.5 0.6 0.7 0.8

0.5 + 0
+ T bon Ps mes
Ps ref
T1 -2000

-4000
Ps(W)

T F(n) Taon
-6000

Tm -8000

++
0.5 T con
-10000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Time(S)

Figure 11. Model of switch operation time taon,tbon,tcon 5000

Q ref
Q mes
4000

The switch times tam,tbm,tcm can be determined by relation 3000

with each sector and times taon,tbon,tcon , showing in table (2)
Qs(Var)

2000
where:
1000
K 1 2 3 4 5 6
0
tam tbon taon taon tcon tcon tbon
-1000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
tbm taon tcon tbon tbon taon tcon Time(S)

tcm tcon tbon tcon taon tbon taon 0

Table 2: Determination of tam,tbm,tcm -10

III. SIMULATION RESULTS

Tem (N.m)

-20

To testify the proposed optimal control strategy, some -30

simulations of dynamic performances for DFIG wind power -40

generating system are implemented.
The parameters of wind turbine for a simulation: P=10kW, U -50
0 0.1 0.2 0.3 0.4
Time(S)
0.5 0.6 0.7 0.8

= 230V, fs = 50 Hz, Rs=0.455 , Rr=0.19 , Rf=0.01,

lf=0.075e-3, ls=0.07H, lr=0.0213H, M=0.034H, p=2.

29
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:05:08 UTC from IEEE Xplore. Restrictions apply.
9A1-4 The 2nd IEEE Conference on Power Engineering and Renewable Energy
ICPERE 2014

100 [4] A. K. Mishra, L. Ramesh, S.P. Chowdhury,S.Chowdhury, « Review of

Is
Ir
Wind Turbine System and its Impact for Grid Stability», Journal of
50
Electrical Engineering. Vol.11, Issue 1. 2011. pp. 153–165.
[5] L. Shuhui, T.A. Haskew, « Analysis of Decoupled d-q Vector Control in
Is,Ir(A)

0
DFIG Back-to-Back PWM Converter», IEEE CNF Power Engineering.
June 2007. pp.1-7.
-50 [6] Y. Bekakra, D. B. attous, «Active and Reactive Power Control of a
DFIG with MPPT for Variable Speed Wind Energy Conversion using
Sliding Mode Control», World Academy of Science, Engineering and
-100
0 0.1 0.2 0.3 0.4
Time(S)
0.5 0.6 0.7 0.8 Technology 60, 2011.
700 [7] A, Gaillard,“ Système éolien basé sur une MADA : contribution à
600
l’étude de la qualité de l’énergie électrique et dela continuité de service”,
thèse de doctorat, Nancy-I, 2010..
500

400
[8] Badrul H. Chowdhury ,and Srinivas Chellapilla, «Double-fed induction
Vdc(V)

generator control for variable speed wind power generation», Electric

300
Power Systems Research, Volume 76, Issues 9-10, June 2006, pp.786-
200 800.
100
[9] J.W Jung, A. Keyhani, «Space Vector PWM Inverter», Departement of
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Electrical and Computer the Ohio State University, 2005.
Time(S)
[10] R. Penaa, R. Cardenasb, E. Escobarb, J. Clarec, P. Wheelerc, « Control
3000
strategy for a doubly-fed induction generator feeding an unbalanced grid
or stand-alone load, Electric Power Systems Research, Vol. 79, pp. 355-
GSC and filter voltage(V)

2000
364, 2009.
1000 [11] A. M. Eltamaly, A. I. Alolah, M. H. Abdel-Rahman, « Modified DFIG
control strategy for wind energy applications», SPEEDAM 2010,
0
International Symposium on Power Electronics, Electrical Drives,
-1000 Automation and Motion, 2010 IEEE, pp. 659-653, 2010.
[12] M. Machmoum, F. Poitiers, «Sliding mode control of a variable speed
-2000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 wind energy conversion system with DFIG», International Conference
Time(S)
and Exhibition on Ecologic Vehicles and Renewable Energies,
Figure 12. Simulation Results MONACO, March 26-29 (2009).
[13] Choon Yik Tang, Yi Guo, John N. Jiang, «Nonlinear Dual-Mode
IV. CONCLUSION Control of Variable-Speed Wind Turbines with Doubly Fed Induction
generators»,IEEE,Transactions on Control Systems Technology ,July
The paper presents a model for a wind turbine generation 2011, 19 (4), pp. 744-756
system based on a doubly fed induction generator with a stator [14] Yao Xing-jia, Liu Zhong-liang, Cui Guo-sheng , «Decoupling control of
directly connected to the grid and a rotor connected to the grid Doubly-Fed Induction Generator based on Fuzzy-PI Controller», IEEE
through SVM converter. Trans. Mechanical and Electrical Technology (ICMET 2010) , pp 226-
230 ,2010.
A detailed modeling of the mechanical dynamics, the wind
[15] Md. Rabiul Islam1, Youguang Guo, Jian Guo Zhu, «Steady State
turbine electrical system and the converter are presented. Characteristic Simulation of DFIG for Wind Power System», IEEE
Simulation results show good decoupling between stator Trans. Electrical and Computer Engineering (ICECE), pp. 151-154,
active and reactive power and proven that both control 2011.
strategies are able to offer convergence of the dynamic [16] T. Luu, A. Nasiri, «Power Smoothing of Doubly Fed Induction
Generator for Wind Turbine Using Ultra capacitors», IEEE Trans.
response of the system to the reference values despite wind IECON 2010 - 36th Annual Conference., pp. 3293-3298, 2010.
variations. The proposed scheme ensures perfect tracking of [17] C.O. Omeje, D.B. Nnadi, and C.I. Odeh, «Comparative Analysis of
maximum captured power and improves good dynamics and Space Vector Pulse-Width Modulation and Third Harmonic Injected
provides less harmonics in rotor currents. Modulation on Industrial Drives»,The Pacific Journal of Science and
Also the obtained results demonstrate that the proposed DFIG Technology, Volume 13. Number 1. May 2012 (Spring).
system control operating at the variable speed may be [18] H. Zhang, Q. Wang, E. Chu, Xi. Liu, and L. Hou, «Analysis and
Implementation of A Passive Lossless Soft-Switching Snubber for PWM
considered as an interesting way for problems solution in Inverters», IEEE Trans. on Power Electron. Vol. 26, No.2, pp. 411- 426,
renewable energy area. Feb.201.1
[19] R. Huang and S. K. Mazumder, «A Soft Switching Scheme for
REFERENCES Multiphase DC/Pulsating-DC Converter for Three-Phase High-
[1] W. Yu, J. Lai and S.Y. Park , «An Improved Zero-Voltage Switching Frequency-Link Pulse width Modulation (PWM) Inverter», IEEE Trans.
Inverter Using Two Coupled Magnetic in One Resonant Pole» IEEE on Power Electron.,Vol.25,No.7,pp.1761-1774,July.2010.
Trans. on Power Electron.,Vol.25,No.4,pp.952-961,April.2010.
[2] J. Hua, Y. Hea and L Xub, « Improved rotor current control of wind
turbine driven doubly-fed induction generators during network voltage
unbalance »,Electric Power Systems Research 80 (2010) 847–856.
[3] A, Gaillard,“ Système éolien basé sur une MADA : contribution à
l’étude de la qualité de l’énergie électrique et dela continuité de service”,
thèse de doctorat, Nancy-I, 2010.

30
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on July 15,2021 at 05:05:08 UTC from IEEE Xplore. Restrictions apply.