LATVIAN JOURNAL OF PHYSICS AND TECHNICAL SCIENCES

2014, N 2
DOI: 10.2478/lpts-2014-0009

AN INDUCTION MOTOR BASED WIND TURBINE EMULATOR

A. Sokolovs1, L. Grigans1,2, E. Kamolins1, J. Voitkans1

1
Riga Technical University
1 Kalku Str., Riga, LV-1658, LATVIA
e-mail: Alvis.Sokolovs@rtu.lv
2
Institute of Physical Energetics
21 Aizkraukles Str., Riga, LV-1006, LATVIA
e-mail: Linards.Grigans@gmail.com

The authors present a small-scale wind turbine emulator based on the

AC drive system and discuss the methods for power coefficient calculation.
In the work, the experimental set-up consisting of an AC induction motor, a
frequency converter, a synchronous permanent magnet generator, a DC-DC
boost converter and DC load was simulated and tested using real-life
equipment. The experimentally obtained wind turbine power and torque
diagrams using the emulator are in a good agreement with the theoretical
ones.
Keywords: wind turbine (WT), wind turbine emulator, wind energy
conversion.

1. INTRODUCTION

Renewable energy sources are the main topic for a few past decades, with
the most rapid development of wind power systems.
In earlier works (e.g. [1-12]) the use of electric drive wind turbine emulators
for simulating the torque of a wind turbine is reported.
Wind turbine emulators with variable speed drives are commonly used, since
it is a cheaper and easier way to control the generator speed and power as
compared with a real wind turbine. The research into electric generators, converters
and control systems under laboratory conditions necessitates that there be
appropriate test equipment that can simulate natural weather conditions.
In this paper, the authors present a laboratory set-up that is built for detailed
research on the power converter control for a direct drive synchronous wind
generator on permanent magnets. The set-up can be used for tracking the maximum
power point or testing the pitch control algorithm as well as can serve the
educational purposes for studying the control and dynamics of wind power
systems. Also, some modifications of the set-up allow for testing the permanent
magnet synchronous generators (PMSGs), synchronous generators with excitation,
squirrel-cage induction generators or double-fed induction generators. In every
11

Unauthenticated
Download Date | 2/25/16 12:53 AM
case, a correct turbine model must be obtained and loaded into the system
controller.
Further in this paper, the wind turbine emulator hardware, control program
and Matlab simulation model are described and the experimental results are
presented. As stated in [13], the wind turbine emulator can be used for different
experiments – e.g., for studying the influence of wind shear and tower shadow on
the power quality or for testing the efficiency of a new algorithm for maximum
power point tracking.

2. WIND TURBINE MODEL

In general, the kinetic energy of the undisturbed upstream wind flow over a
definite section with area A can be calculated as
1 3
Pw  Av  , (1)
2
where v is the wind speed, and ρ is the density of air. In turn, the mechanical power
that can be transferred from wind to the generator is
1 3
PT  Av C p . (2)
2
Here A is the area covered by the rotor blades of turbine, and CP is the wind turbine
power coefficient. Equation (2) can be rewritten in terms of the rotor radius rrot:
1 2 3
PT   rrot v C p . (3)
2
The aerodynamic torque of the wind turbine is also a function of wind speed
v, and is calculated as
1
TA   r 3  v2CT , (4)
2
where CT is the aerodynamic coefficient that relates to CP by the tip speed ratio
(TSR) λ:
Cp
CT  . (5)

The TSR λ is defined as the ratio of rotational speed ω of the tip of turbine
rotor blade and wind speed v:
r
 . (6)
v
Power coefficient CP is considered the most important parameter in the
power regulation of a turbine. In order to capture the maximum available power at
a particular wind speed the rotational speed of the turbine rotor ω is relevant.
Aerodynamic characteristics of the WT blades define the optimum rotational speed
of the rotor, and any deviation from this speed causes a significant drop in the
captured wind power. At the same time, the rotational speed directly depends on
12

Unauthenticated
Download Date | 2/25/16 12:53 AM
the generator load and wind speed variations. To calculate the power captured from
wind, the power coefficient should be defined.
The power coefficient CP is a nonlinear function of TSR λ and blade pitch
angle β characterizing the aerodynamic behaviour of rotor blades. The CP value
differs for each turbine type. Since this value depends on the wind speed, it is given
by turbine manufactures in the look-up tables. Theoretical models of CP(λ,β) have
been developed and expressed analytically. For example, in [13] CP calculation is
done as
C5
C  
C p   ,    C1  2  C3   C4  e i  C6  , (7)
 i 
with
1 1 0.035
  3 . (8)
i   0.08   1
In (7) the C1 through C6 are coefficients that depend on the turbine type, with C1
and C6 calculated as
Cp
C1  C5
, (9)
 C2   i
  C 4 e  k
 i 
C6  k  C1 , (10)
and
C5
 C   
k    C2  5  C4  C5  C2   e i . (11)
 i 
Coefficients C2, C3, C4 and C5 are used as shown in [13].
In [14] a similar equation for CP is presented as a function of λ and β with
different coefficients:
C5
C  
C p       C1  2  C3   C7  C8  C4  e i , (12)
 i 
where C7 and C8 are additional coefficients determined by the turbine type [15],
and λi is the same as in (8). In [16] CP is approximated with the sine wave function:
      C3  
C p      C1  C2   sin    C6     C7   , (13)
 C4  C3 
where C1 – C7 are different from those used in (7) and (12)
Theoretical CP curves as function of λ are shown in Fig. 1. It can be seen that
Eqs. (7) and (12) – lines 1 and 2, respectively, give rather close results at large λ
values, though coefficients C1 – C6 in Eq.(12) are to be changed to fit the curves. In
13

Unauthenticated
Download Date | 2/25/16 12:53 AM
this case, C1=0.27, C2=150, C4=3.35 and C5=18. The coefficients C3, C7 and C8 are
not considered since β=0. Equation (12) described by line 2 gives more variance at
low λ values due to the lack of term C6λ as compared with (7). Approximation (13)
characterized by curve 3 has a significant deviation at higher λ values and the
beginning of the curve is linearized, which gives a considerable error at low λ
values.
For further calculations and CP modelling of the wind turbine Eq. (7) will be
used.

Fig. 1. Theoretical power coefficient curves 1, 2, and 3

of a wind turbine vs. λ (TSR) calculated by Eq. (7), Eq.(12), and Eq.(13), respectively.

Based on the CP curves, the turbine power Pt and torque Tt curves are
obtained as functions of the rotational speed of the rotor at different wind speed
values (Fig. 2). As can be seen from (6), for the turbine to operate at maximum
power coefficient the rotational speed must be proportional to the wind speed. This
can be achieved when in a variable speed turbine the maximum power point
tracking strategy is applied. In Fig. 2a it is seen that for each wind speed there is a
rotational speed of the rotor that allows operation at maximum power. In (2) it is
defined that the turbine power is proportional to the cube of the wind speed. Since
the rotational speed of the turbine is proportional to the wind speed, the power of
turbine is also proportional to the cube of rotational speed of the rotor, i.e.:

PTmax  k 3 . (14)

To obtain optimal operation, torque (14) is divided by rotational speed of the

rotor:

TTopt  k 2 . (15)

The maximum power and the optimum torque curves are presented in Fig. 2.
It is worth mentioning that if CP is calculated by (12), the obtained power
and torque curves (Fig. 2a and b, dashed lines) are inaccurate for several reasons:
first, at low rotor speeds the turbine would not produce any power and torque even
at significant wind speeds. Second, the power and torque curves for different wind
speeds intersect, so that at low rotor speeds a higher wind speed produces smaller
power and torque. These drawbacks should be taken into account when a wind
turbine is simulated, otherwise a more precise emulator is needed. Therefore, it is
recommended to use (7) for CP calculations.
14

Unauthenticated
Download Date | 2/25/16 12:53 AM
 a)

b)
Fig. 2. Theoretical power (a) and torque (b) curves of a wind turbine as functions of rotational
speed of rotor. Solid lines correspond to Eq. (7) and dashed lines – to Eq. (12).

Next, the dynamical behaviour of a wind turbine is considered. In general,

the WT rotational speed variation rate is determined by the total system inertia and
turbine torque. The total dynamic torque can be expressed as
d d
JS  TT  TG  BT  J S  TT  TG  BT  , (16)
dt dt
where JS is the total moment of inertia of the system,
TG is the electromagnetic torque of generator,
BT is the friction coefficient of turbine (in direct drive turbines may be
neglected).
The total system’s inertia is the sum of wind turbine inertia JT, generator rotor
inertia JG and shaft inertia Jsh:
J S  JT  J G  J sh . (17)
In the case of cylindrical shapes (the generator shaft and the rotor) the
approximate moment of inertia can be calculated as
1
J  mr 2 , (18)
2
where m, r is the rotor (or shaft) mass and radius, respectively.
The moment of inertia of wind turbine blades (assuming that these are
homogeneous bars) for a three-blade horizontal axis turbine can be calculated as
1 
J  3   mb l 2   mbl 2 , (19)
 3 
where mb is the mass of a rotor blade, and l is its length.
15

Unauthenticated
Download Date | 2/25/16 12:53 AM
 To achieve a more precise emulation of the wind turbine, its inertia and that
of the whole emulator (i.e. the moment of inertia of induction motor JIM) are to be
taken into account. Therefore, the difference should be found between these
moments of inertia:
J  JT  J IM , (20)
which is the virtual part of the WT moment of inertia included in the reference
torque calculations.

3. SIMULINK WT MODEL

In this study, a simulation model in Simulink was developed to compare the

wind turbine emulator and the Simulink WT models.

-1

v_turb_rpm
v_turb_rpm

0 beta
be ta T Discre te ,
Ts = 1e -05 s.
v_wind
v_wind
powe rgui
Wind Turbine1 wind turbine
PWM
+ rpm

L1
D1 -K-
+

Tm
D
g

rectifier1 <Rotor speed wm (rad/s)>

V1 Q1 C1
+ A A
m
S

B B

-
C C
generator with load PMSM1

turbine speed
wind speed
ref

v_turb_rpm
v_turb_rpm
|u|
Flux ref erence (Wb)
Flux table
0 beta
be ta T [0]
Torque ref erence (Nm)
v_wind
v_wind
Ctrl sat.
Wind Turbine2
Torque*
MagC
Flux*
Three-phase V_abc
Gates
N

inverter I_ab
g Rate Transition Induction Rad2Rpm
DTC
+
machine
m

A Ta I_ab
VDC V_abc Tm
B Tb
- Mta A
C Tc Mtb B m
Nm
V_Com Mtc C
Tl
wind turbine emulator Measures Shaft Nl rpm

PWM
+

L2 -K-
D2
+

Tm
D
g

-1
rectifier2 <Rotor speed wm (rad/s)>
V2 Q2 C2
+ A A
m
S

B B

-
C C
generator with load PMSM2

Fig. 3. Simulink WT model and the wind turbine

emulator model with the same generator-load system.

Both the models contain the PMSG with a rectifier and a boost converter to
300V DC grid (Fig. 3 “generator with load”). In this figure Simulink wind turbine
model is represented by Wind Turbine1 block, which calculates CP by (7) and the
turbine torque by (4) and contains the dynamic part of a wind turbine. In the
emulator the same WT model is used to obtain the reference torque for frequency
16

Unauthenticated
Download Date | 2/25/16 12:53 AM
converter (in Fig. 3 represented by three-phase-inverter and direct torque control
(DTC) units.

500
wind turbine model
wind turbine emulator model

turbine speed (rpm)

450

400

350

300

250
5 10 15 20 25 30 35
time (s)

Fig. 4. Matlab simulation results for the wind turbine and its
emulator models at wind speed step response from 6 to 8 m/s.

Figure 4 shows comparative simulation results for turbine dynamic response

at wind speed change from 6 m/s to 8 m/s. As can be seen, the results for the wind
turbine output power and torque are the same, with a negligible deviation that
might arise due to losses in the AC drive system.

4. WIND TURBINE EMULATOR

A WT emulator laboratory set-up (Fig. 5) is based on a commercial

1500 rpm 7.5 kW AC induction motor with a frequency converter which
implements direct torque control (DTC) algorithm. As the microcontroller, a
PIC32MX360F512L on EXPLORER16 development board is used, with an
additional digital-analog converter for the torque reference input to the frequency
converter. The rotational speed and torque are measured by on-shaft DRBK-100
torque transducer.

Frequency AC induction Torque-speed

converter motor sensor

Test generator
Microcontroller

a)

b)
Fig. 5. Wind turbine emulator: a) laboratory set-up; b) schematic diagram.

17

Unauthenticated
Download Date | 2/25/16 12:53 AM
 The wind speed and rotor speed ω are fed to a microcontroller in which the
reference torque is calculated by Eqs. (3), (4), (6) and (7). In the case of a fixed
blade (stall-controlled) turbine the pitch angle of the blade is fixed to 0o, and,
therefore the power coefficient depends only on λ; hence (7) and (8) can be
simplified:
C5
C  
C p     C1  2  C4  e i  C6  , (21)
 i 
1 1
  0.035 . (22)
i 
Values of coefficients C1 – C6 are mentioned before and are the same as in
simulation, so that maximum CP corresponds to nominal λ for the generator speed
of 450 rpm and the rated wind speed of 8 m/s. The rotor diameter is assumed to be
3 m.
The control of wind speed is done from the user interface of the
microcontroller. It is also possible to feed in a particular wind speed pattern from
other sources.

5. RESULTS AND DISCUSSION

Using the WT emulator set-up (Fig. 5) the power and torque diagrams have
been obtained. A schematic diagram of Fig. 5b provides details of the experimental
set-up: on the left, the wind turbine emulator is shown as consisting of the
frequency converter (f-converter), AC induction motor (IM), microcontroller
(μ-controller) and torque & speed sensor. The load shown on the right consists of a
PMSG (rated power 1 kW, rated speed 450 rpm), a diode rectifier (AC-DC
converter), a boost DC-DC converter, a voltage limiter to 300V, and a load resistor.
The load on the generator and hence the speed of the turbine is controlled by
controlling the voltage at the rectifier output. It is done by changing the duty cycle
D of the boost converter (PWM). Since the output voltage is fixed by the voltage
limiter, the input voltage is proportional to 1-D.
By changing D from 0 to 1, the programmed WT torque diagrams are
measured and corresponding power diagrams are calculated. In Fig. 6 the obtained
power and torque diagrams are compared with the theoretical ones. It can be seen
that the former are in good agreement with the latter.

1600 9 m/s
turbine output power (W)

1400
1200
8 m/s
1000
800
7 m/s
600
400 6 m/s
200 5 m/s
4 m/s
0 3 m/s
0 100 200 300 400 500 600
a) turbine speed (rpm)

18

Unauthenticated
Download Date | 2/25/16 12:53 AM
 35
9 m/s
30

turbine torque (Nm)

25 8 m/s

20 7 m/s
15 6 m/s
10 5 m/s
4 m/s
5 3 m/s
0
0 100 200 300 400 500 600
b) turbine speed (rpm)

Fig. 6. Experimental (dashed lines) and theoretical (solid lines) results:

turbine power (a) and torque (b) vs. turbine speed at different wind speeds.

6. CONCLUSIONS

From the results of work the following conclusions can be drawn:

1. To obtain reliable results for the turbine output power and torque the WT
power coefficient CP should be calculated by Eq. (7).
2. The power and torque diagrams obtained with the WT emulator
simulation model in Matlab Simulink and experimentally are in a good agreement
(with a negligible error).
3. The WT emulator built for the research (in particular, for experiments
with wind generators and wind generator control in a controlled environment) has
shown a good performance. Our emulator set-up contains the elements (such as AC
induction motor with frequency converter, PIC microcontroller and torque-speed
on-shaft sensor) that altogether allow modelling of the wind turbine output torque
taking into account the WT dynamics in the same way as under real-life conditions.
4. Depending on the microcontroller code, both stall- and pitch-controlled
wind turbines can be emulated.
5. Using a WT emulator, the algorithms of maximum power point tracking
can be tested.

ACKNOWLEDGMENTS

This work was supported by the European Regional Development Fund

project “Development of a slow-speed power generator for wind turbines”, No.
2010/0215/2DP/ 2.1.1.1.0/10/APIA/VIAA/039

REFERENCES
1. Kana, C. L., Thamodharan, M., & Wolf, A. (2001). System management of a wind-
energy converter. IEEE Transactions on Power Electronics, 16(3), 375–381.
doi:10.1109/63.923770
2. Lu, H.-C., & Chang-Chien, L.-R. (2010). Use of wind turbine emulator for the WECS
development. In: International Power Electronics Conference (IPEC),
(pp. 3188–3195).
3. Monfared, M., Madadi Kojabadi, H., & Rastegar, H. (2008). Static and dynamic wind
turbine simulator using a converter controlled dc motor. Renewable Energy,
33(5), 906–913.

19

Unauthenticated
Download Date | 2/25/16 12:53 AM
4. Abo-Khalil, A. G. (2011). A new wind turbine simulator using a squirrel-cage motor
for wind power generation systems. In IEEE Ninth International Conference on Power
Electronics and Drive Systems (PEDS), 750–755.
5. Voltolini, H., Granza, M. H., Ivanqui, J., & Carlson, R. (2012). Modeling and
simulation of the Wind Turbine Emulator using induction motor driven by torque
control inverter. In: 10th IEEE/IAS International Conference on Industry Applications
(INDUSCON), 1–6.
6. Chinchilla, M., Arnaltes, S., & Rodriguez-Amenedo, J. L. (2004). Laboratory set-up
for wind turbine emulation. In: IEEE International Conference on Industrial
Technology, IEEE ICIT’04., 1, 553–557.
7. Vaheeshan, J., Vihirthanath, V., Abeyaratne, S. G., Atputharajah, A., & Ramatharan,
G. (2011). Wind Turbine Emulator. In: 6th IEEE International Conference on
Industrial and Information Systems (ICIIS), 511–516.
8. Kumsup, S., & Tarasantisuk, C. (2010). Real-time wind turbine emulator for testing
wind energy conversion systems. In: IEEE International Energy Conference and
Exhibition (EnergyCon), 7–9.
9. Arifujjaman, M., Iqbal, M. T., & Quaicoe, J. E. (2006). An isolated small wind turbine
emulator. Canadian Conference on Electrical and Computer Engineering,
2006. CCECE’06., 1854–1857
10. Neammanee, B., Sirisumrannukul, S., & Chatratana, S. (2007). Development of a wind
turbine simulator for wind generator testing. International Energy Journal, 8(1),
21–28.
11. Cutululis, N. A., Ciobotaru, M., Ceanga, E., & Rosu, M. E. (2002). Real time wind
turbine simulator based on frequency controlled AC servomotor. Annals of “Dunarea
de Jos” University of Galati, 3, 97–101.
12. Kojabadi, H. M., Chang, L., & Boutot, T. (2004). Development of a novel wind turbine
simulator for wind energy conversion systems using an inverter-controlled induction
motor. IEEE Transactions on Energy Conversion, 19(3), 547–552.
13. Heier, S. (1998). Grid Integration of Wind Energy Conversion Systems. Wiley & Sons.
14. Ofualagba, G., & Ubeku, E. (2012). The modeling and dynamic characteristics of a
variable speed wind turbine. Journal of Energy Technologies and Policy, 1(3), 10–21.
15. Manyonge, A. W., Ochieng, R. M., Onyango, F. N., & Shichikha, J. M. (2012).
Mathematical modelling of wind turbine in a wind energy conversion system: Power
coefficient analysis. Applied Mathematical Sciences, 6(91), 4527–4536.
16. Tahar, M., & Lachguer, N. (2012). Matlab Simulink as Simulation Tool for Wind
Generation Systems Based on Doubly Fed Induction Machines. In; V. Katsikis (Ed.),
MATLAB - A Fundamental Tool for Scientific Computing and Engineering
Applications - Vol 2. InTech.

UZ ASINHRONĀ ELEKTRODZINĒJA BĀZES VEIDOTS

VĒJA TURBĪNAS EMULATORS
A. Sokolovs, L. Grigāns, E. Kamoliņš, J. Voitkāns
Kopsavilkums

Šajā rakstā parādīta mazas jaudas vēja turbīnas emulatora izveide ar

maiņstrāvas piedziņas sistēmu, kā arī analizētas vairākas turbīnas jaudas
koeficienta analītiskās aprēķina metodes. Vēja turbīnas emulatora eksperimentālais
stends, kas sastāv no asinhronā elektromotora, frekvenču pārveidotāja, sinhronā
pastāvīgo magnētu ģeneratora, līdzstrāvas paaugstinošā pārveidotāja un slodzes,
20

Unauthenticated
Download Date | 2/25/16 12:53 AM
tika pārbaudīts gan simulēšanas vidē, gan uz reālām iekārtām. Eksperimentāli
iegūtās vēja turbīnas emulatora jaudas un momenta diagrammas ir salīdzinātas ar
teorētiskajām.
27.02.2014.

21

Unauthenticated
Download Date | 2/25/16 12:53 AM