International Renewable Energy Congress

November 5-7, 2010 Sousse, Tunisia

Modeling and Control of a Variable Speed Wind Energy Conversion System Based on an Induction Machine
B. Bouaziz, F. Bacha and M. Gasmi
Unit de Recherche en Automatique et en Informatique Industrielle (URAII), Institut National des Sciences Appliques et de Technologie (INSAT), Universit de Carthage, Centre Urbain Nord, BP 676, 1080 Tunis Cedex bechir.bouaziz@gmail.com, Faouzi.Bacha@esstt.rnu.tn, m.gsm@enit.rnu.tn Abstract -This paper deals with the study of a Wind Energy Conversion System (WECS) based on a squirrel cage induction machine connected to an infinite power grid through two double side pulse width modulated converter systems. After modeling of all components of the conversion system, we develop a new approach to control the electrical and mechanical variables. This approach is based on the PI controller of the DC link voltage, pitch control of the wind turbine and a hysteresis controller for the grid side converter. Then, a study of the performance of our system under grid disturbances such as dip in the grid voltage will be presented. This study justifies, under specific conditions, bidirectional transfer of power between the induction machine and the grid, which confirms its application in the WECS. Simulation results are carried out to show the feasibility of the proposed approach. Keywords wind energy, back to back converters, PI DC link control, pitch control. 1. Introduction Wind Energy Conversion Systems (WECS) are on an established technology for power generation [10] and they are evolving from being an alternative energy source to providing additional functions such as reactive power supply, voltage control and active power regulation [13]. Such functions are possible owing to the improvement of the WECS components and to increasing sophistication of the control system [5]. For wind energy conversion system WECS, recent research has been focused chiefly on permanent magnet synchronous machines PMSM or doubly fed induction machines DFIM as generator of electrical power. Advantages have been variously attributed to high power density for PMSM and reduced rating of power converters for DFIM. However, the PMSM suffers from high cost of materials and manufacturing. In DFIM schemes, the slip-ring brush contacts pose a serious maintenance and reliability problem, especially for remote installation. Squirrel-cage induction machines still have an edge over these approaches in terms of cost ID105/ IREC2010 and ruggedness. Squirrel cage generators with shunt passive or active VAR (Volt Ampere Reactive) generators was proposed in [9], which generate constant frequency power through a diode rectifier and line commuted thyristor inverter. The control systems for the operation of indirect rotor flux oriented vector controlled induction machines for variable speed wind energy applications are discussed in [2]-[11]. Sensorless vector control scheme suitable to operate cage induction is considered and a fuzzy control system is used to drive the WECS to the point of maximum energy capture for a given wind velocity. The induction machine is connected to the utility using buck to buck converters. In this paper a variable speed wind turbine driven squirrel cage induction generator system with two double sided PWM converters is described. In the first section of this paper the modeling of all the components of the conversion system including inverters, one is the grid side converter (GSC) built around a hysteresis controller, and the second is the induction machine side converter (IMSC) is controlled with scalar (V/f) approach to optimize efficiency and enhance performance. In a second section of this paper a control algorithms are evaluated by simulation study. 2. Modeling of the WECS 2.1 Wind turbine model A wind turbine is characterized by the curves of power coefficient C p which is a function of velocity ratio l and the pitch angle of blades b , and the mechanical power extracted from the wind.
Pmec =

r
2

C p (l , b )S v3

(1)

l is given by: l =

vt 2 R wr = hGB v pv

154

IMSC
Turbine
Induction machine

i dc M i o iC d1

GSC

d 1' ia

d 2'

d 3'

d2

d3

Rf i1 i2 Rf Rf

Lf Lf Lf

a b
C

D
U dc

e1

nd

ib ic
Gearbox

Wi

E F

B e2 C e3

c
d 2' d 3'

i3

d 1'

d1

d2

d3

Figure 1. Synoptic presentation of the WECS

In our example, the power coefficient is given by the following relation:

c2 - c 5 C p (l , b ) = c1 c b c exp + c l . 3 4 l 6 l i i

With: x = [ ia ib
1 ls 0 B= 1 0 0 1 ls 0 1

j a j b va vb ,u = [

(2)

li=

c 1 - 38 l + c7 b b + 1

The coefficients are given by the following table 1. Table 1. Parameters of the performance coefficients

c1
0.5109

c2
116

c3
0.4

c4
5

c5
21

c6
0.0068

c7
0.08

c8
0.035

2.2. Modeling of the asynchronous machine For the purposes of transformation of Concordia, the electrical model of the induction machine is written:

1 1 1 1 - + - w s t t t s r r ls 1 1 1 w A= - w - + s ls t s t r - Rs 0 0 0 - Rs 0

ls 1 t r ls 0 0

Where:

dj s vs = Rs is + . dt

(3) . (4) (5) (6)

l s = s L s , s = 1-

0 = Rr i r +

dj r - j wj dt

L L2 L L m , mr = m , t r = r , t s = s Ls L r Lr Rr Rs w : is the electrical speed of the rotor

j s = Ls is + Lm ir .

2.3. Modeling of the IMSC converter The Induction machine side converter (IMSC) is identical in design to the converter (GSC) already studied. On the other hand, ( L s and R s ) mean respectively the inductance and resistance per phase of the machine. The output current of the IMSC expressed by the following equation: ' ' (8) idc = d1' ia + d2 ib + d3 ic . The voltage components of the induction machine in Concordia referential are given by:

j r = Lr ir + Lm is .

The modeling of the induction machine can be put in the form of an equation of state:

dx = A x + Bu . dt

(7)

ID105/ IREC2010

155

V a = d a' U dc .
V b = d b U dc .
i dc
'

(9) (10)
Power (kW)
300 200 100 0

'

3 = (i b d a' + i a d b' ) . 2
' b

Performance coefficient

Characteristic of the maximal power

0.5 v=12m/s v=10m/s 0,4 0.3 0,2 0.1 0 0 beta=16 beta=8 beta=20 10 Tip speed ratio 20 beta=0 beta=4

(11)

v=8m/s v=7m/s 0 100 200 Time (s) 300

Where (d a ) and (d

)presents the transformation of

the switching states (dk' )of the induction machine side converter. The quantities i a and i b are the direct and quadratic components of the stator current. 2.3. Modeling of the (GSC) converter The converter (GSC) represented in Figure 1, is a static converter. On the other hand L f and R f respectively denote the inductance and resistance of the line filter. The current in the capacitor is expressed as follows:
i C = i dc - i o = C dU dc . dt

Figure 2. Wind turbine power characteristics 3.1.2. Pitch Control The speed is always equal to its nominal value. The principle of this command is to limit the power generated by adjusting the angle of the blade pitch. The model incorporating the wind turbine control "Pitch Control" is given by the figure 6.

(12)

Pn

ref
R (P )

1 s

Tm

Model
r
v

With (i 0 ) , is the input current of the GSC converter:

i 0 = d 1i 1 + d 2 i 2 + d 3i 3 . (13) As a result, the full model of the grid, capacitor and converter is summarized by the expressions below:

1 2

wind turbine Pm

Pn
Pmesure

Pitch Control

Lf

3 U di k + R f i k = - e k + (U dc d k - dc d k ) . dt 3 k =1 3 dU dc C = i dc - d k i k . dt k =1 3. Control strategy of the WECS

(14) (15)

Figure 3. Model of the wind turbine incorporating the command "Pitch Control"

3.2. Control of IMSC converter The structure of the speed controller is shown in Figure 4. Such a structure is to compare the actual speed of the machine (wr ) to a reference speed (wr* ) . The reference speed is reached so as to optimize (l ) . According to the characteristic C p (l , b ) , for b = 0 the maximum value of C p is obtained when l = l opt . Attach l to its optimal value with the measurement of wind speed can be
l opt v * wr = p hGB . R

3.1. Control strategy of the wind turbine 3.1.1. Maximum Power Point Tracking (MPPT) The power of the turbine is optimized by an MPPT algorithm. The control is to vary the speed of the wind in order to remain at around (l opt = 8.1) the pitch angle is fixed (b = 0) . The power coefficient of the turbine is equal to its maximum value (C p max = 0.475) . In these conditions the maximum power is given by:

(17)

Pm max = Aopt wr3 .

Aopt 3 2R = 0.5r C p max = cste phGBl opt

(16)

As a result, from the difference between these two speeds and with a (PI) controller we estimated the slip speeds (wslip ) . To which we add the actual speed
* measured. We thus generated electrical pulse (w )

this pulse and the RMS voltage produced by the proportionality (V f ) are communicated to the control laws of IMSC.

ID105/ IREC2010

158

nd

Wi

GB

IMSC ia

MLI

idc iC

io C Udc

i dc M

io
C

GS C

IM

ib ic

U dc

iC

MLI

i1 i2 i3

Lf Lf Lf

Rf Rf Rf

e1 e2 e3

' 1

d 2'
PWM

' 3

Pitch control

e123 _ ref
p
Gnration des tension triphass

U dc
+

d1

d2

d3

Pref

Udcref
+

Pm

r
v
MPPT opt p GB
R

PI

slip

V f

PI

i 1_ ref

i 2_ref
Im

i 3 _ ref
1 Em

1 Em

1 Em

Figure 4. Block diagram of control strategy for all wind and converter (IMSC).

Figure 5. Block diagram of the structure of closed loop control 4. Application and Simulation Results

3.3. Control of the GSC converter The aim of the GSC is to regulate the DC link voltage and maintain the balance between the dc link power and the injected powers to the grid. Several approaches are used in literature. In this paper, we use a hysteresis current controller to compare the reference current (i k _ ref ) and the actual current (i k ) This controller maintains the grid currents in a hysteresis band. The grid voltage and current respectively are given by :
e k = E m sin (wt - (k - 1) i k _ ref 2p ). 3 2p = I m sin(wt - (k - 1) + qvi ) . 3

To validate our proposed approaches, we performed simulations of the WECS with nominal data and parameters provided in Appendix. The simulations are carried out with Matlab / Simulink. 4.1. Simulation of Wind Turbine and IMSC converter In Figure 7 we present the simulation results of the IMSC converter with a scalar control, MPPT algorithm control and the aerodynamic control Pitch Control . We simulate the performance of the proposed control strategy in case of a turbulent wind speed. The profile is illustrated by Figure 6, simulation results give the behavior of electrical and mechanical variables.
13
Wind speed (m/s)

(18) (19)

12.5 12 11.5 0 10 20 30 Times (s) 40 50

Where (qvi ) is the angle between the grid voltage and current. The generation of command current waveforms is achieved via a closed-loop control of the converter. Such a controller regulates the dc link voltage at the desired voltage level and adjusts the magnitude of grid reference currents. The angle of the reference current is given by the grid voltage. The structure of closed loop control is given by the block diagram of Figure 5.

Figure 6. Turbulent wind speed

13 10

Wind speed (m/s)

Pitch angle ()
0 10 20 30 Time (s) 40 50

12.5 12 11.5

10

20 30 Time (s)

40

50

Perfomence coefficient

0.5 0.4 0.3

Tip speed ratio

8.5 8 7.5 7 0 10 20 30 Time (s) 40 50

10

20 30 Time (s)

40

50

ID105/ IREC2010

159

DC-link voltage( V)

Electric speed/ reference speed (rad/s)

200

Active power (kW)

770 760 750 740

500 300 0 -300 -500 3 3.5 4 Time (s) Voltage 4.5 5

Rotor speed (rad/s)

340

180 160 140

320

300

10

Reactive power (kVAR)

20 30 Time (s)

40

50

10

20 30 Time (s)

40

50
0

6 Time (s)

10

Line voltage (V) Line current(A)

Electromagnetic torque (kN.m)

2 0 -2 -4 -6 0 10 20 30 Time (s) 40 50

Mecanic torque of the turbine (kW)

-1 .4 -1 .6 -1 .8 -2

-100 100 0 -100 3.9 3.95 4 Time (s) 4.05

500 0 -500

Curent 3.95 4 Time (s) 4.05

10

20 30 Time (s)

40

50

600

300

Magnitude (A)

400

Line currents (A)

500 0 -500 3.95 4 Time (s) 4.05

Stator active power (kW)

Mecanical power (kW)

0 -300 -500 -1 000

400 200 0

300

200

10

20 30 Time(s)

40

50

10

20 30 Time (s)

40

50

10 15 20 Harmonic order

25

30

Figure 7. Wind speed, pitch angle, performance coefficient, tip speed ratio, electric speed and reference speed, rotor speed, electromagnetic torque, mechanic torque of turbine, mechanical power and stator power As can be observed in Figure 7, the pitch angle is oscillating around 0; the performance coefficient is oscillating around C p max = 0.475 ; the tip speed ratio is oscillating around l opt = 8.1 values corresponding to the nominal wind speed of 12m/s. The rotor speed and the electromagnetic torque are both oscillating around their nominal value of (wr = 160 rad s ,Tem = 1.89kN .m ) . The speed of the machine is well controlled and follows its reference. The mechanical power and the mechanical torque of the turbine are both oscillating around their nominal value. We can finally deduce that all variables while a turbulent wind speed of 12 m s , will be influenced by the disturbance and therefore oscillating around their nominal value, we note that oscillations are not considerable. 4.2. Simulation of the GSC converter In Figure 9 we present the simulation results for the grid side converter (GSC) with a hysteresis current controller. To show the bidirectionality of the GSC converter, we consider the profile of the input power given by Figure 10.
500 300

Figure 9. DC-link voltage, active power, reactive power, line voltage and line current, line currents and current spectrum. As can be observed in Figure 9, the DC bus voltage and its reference, we note that the DC bus voltage follows its reference with a very low ripple. We observe also the inversion of the power and confirm the desired power transfer. The line current of the grid side shows the increase in their amplitude under fault operation. The spectrum of the line current shows that the THD (THD= 3.6 %) does not exceed the limit required by the standard IEEE 519, which is 5%. 4.3. Performance of the WECS under voltage dip We simulate in this last paragraph, the behavior of the system in a voltage dip. The scenario of the dip occurring in the grid is given in Figure 10.
400

Grid voltage (V)

350 300 250 200 0 1 2 Time (s) 3 4

Figure 10. Scenario of a dip in the grid voltage The Figure 10 shows a dip of 20% in the grid voltage, occurring at t=2s and disappearing at t=2.5s.

Pdc (kW)

0 -300 -500 0 2 4 6 Time (s) 8 10

Figure 8. Input power of the (GSC) converter

ID105/ IREC2010

160

200 0 -200 -400 2 2.1 2.2 2.3 Time (s) 2.4 2.5

DC-link voltage (V)

400

780 760 740 1.5 2 Time (s) 2.5 3

Line voltage (V)

petites puissances , PhD Thesis, Institut National Polytechnique de Toulouse, Avril 2005.
[2] D. Dong, Modeling and Control Design of a Bidirectional PWM Converter for single-phase Energy System, Master of Science in Electrical Engineering, Faculty of the Virginia Polytechnic Institute, May 2009. [3] E. Muljadi, CP. Butterfield, Pitch Controlled Variable Speed Wind Turbine Generation , IEEE Transactions on Industry Applications, vol. 37, no. 1, 2001. [4] J.Rodrguez, J.Pontt,P.Corea, U.Ammann, and P.Corts, Novel Control Strategy of an AC/DC/AC Converter Using Power Relations, Proceedings of PELINCEC, Warsaw, Poland, 2005. [5] J.C Smith, R. Thresher, R.Zavadil A mighty wind, IEEE power energy mag, vol 7 (2) pp 4151, 2009. [6] K.B.Mohanty, Study of Wind Turbine Driven DFIG Using AC/DC/AC Converter, PhD thesis, National Institute of Technology Rourkela, 2008. [7] N. Aouani, F. Bacha, R. Dhifaoui Control Strategy of a Variable Speed Wind Energy Conversion System Based on a Doubly Feed Induction Generator , IREACO, Vol.2, no.2, March 2009. [8] N. Abu-Tabak, "Stabilit dynamique des systmes lectriques multi machines : modlisation, commande, observation et simulation ", Thse, Novembre, 2008. [9] R.C. Bansal, T.S. Bhatti, D.P. Kothari, "On some of design aspect of wind energy conversion systems" Energy Conversion and Management 43, 2002. [10] T.Ackermaw, Wind power in power systems, John Wile yard Sons, 1er edition, 2005. [11] T. Surgevil and Eyub Akpinar Modeling of a 5-kW wind energy conversion system with induction generator and comparison with experimental results, Renewable Energy Science direct, vol 30 pp 913929, 2005. [12] T. Surgevil, "Modeling and simulation of wind energy conversion system using PWM converters", Thesis, January, 2004. [13] Z. Chen and F.Blaabjerg Wind energy: the worlds fastest growing energy source, IEEE power electr. Soc.Newselt, vol 18 (3) pp 1519, 2006.

Active power (kW)

-200

Reactive power (kVAR)

10

-300

-400

1.8

2.2 2.4 Time (s)

2.6

2.8

-10

2.1

2.2 2.3 Time (s)

2.4

2.5

2.6

Line currents (A)

500 0 -500 2 2.1 2.2 2.3 Time (s) 2.4 2.5

Line currents (A)

500 0 -500 2.48 2.49 2.5 2.51 Time (s) 2.52 2.53

Figure 11. Line voltage, DC-link voltage, active power, reactive power, line currents and zoom on line currents As can be observed in Figure 11, the magnitude of the grid line currents increase about 24% under fault operation. The DC bus voltage is also affected by this voltage dip and not exceeds 764 V. We can note that the DC bus voltage is maintained around its reference value (760V). In fact the variation of the DC link affected the active and the reactive power injected of the grid. 5. Conclusion In this paper, a control strategy of the WECS was proposed. The control has permitted to regulate the DC link voltage and the active and reactive power injected to the grid. The proposed control strategy was therefore implemented for three cases that are first a wind turbulent speed, second the bidirectionality of the GSC converter and finally the behavior of the WECS during a fault of the grid voltage dip. Simulation results show the effectiveness of the proposed techniques Appendix The parameters of the wind turbine Pmec = 300kW based on a cage induction machine are used:
R = 14m ; hGB = 23; J = 50 kg .m 2 p = 2 ; ws = 100p rad s ; N n = 1515tr min ;
R s = 0.0063W; Ls = 0.0118H ; R r = 0.0048W; L r = 0.0063W; L m = 0.0118H ; R f = 5m W; Lf = 0.5mH ; C = 100mF

6. Rfrences [1] A. Mirecki, tude comparative de chaines de conversion dnergie ddies une oliennes de

ID105/ IREC2010

161