Fault tolerant switched reluctance machine for wind turbine blade pitch control

Abstract

The paper's goal is to accomplish a study regarding the usefulness of a fault tolerant machine in wind turbine pitch control. The pitch control both helps to catch the wind at the optimal angle of attack of the blade's airfoil and to turn the blades out of the wind when maximum speed is exceeded. To fulfill this assignment electrical machines are used to control the blade's pitch angle function of wind's speed. The faults that can occur in electrical machines can stop the pitch control system, decreasing drastically the efficiency of the wind turbine. Fault tolerance ability of the pitch controlling machines can avoid such issues. In the paper a fault tolerant switched reluctance machine is studied. As it being simple and cheap it is optimal for such vital applications. Numerical field computations were used to analyze the machine's behavior. An advanced cosimulation was performed by coupling two simulation programs in order to study the machine's drive system and fault tolerance capacity.

Full text

Page 1

Fault Tolerant Switched Reluctance Machine for
Wind Turbine Blade Pitch Control
Mircea RUBA, Loránd SZABÓ, Florin JURCA
Technical University of Cluj,
15, Daicoviciu str., 400020 Cluj, (Romania)
tel: (+40)-264-401827, fax: (+40)-264-593117
e-mail: Mircea.Ruba@mae.utcluj.ro
Abstract-The paper's goal is to accomplish a study
regarding the usefulness of a fault tolerant machine in wind
turbine pitch control. The pitch control both helps to catch
the wind at the optimal angle of attack of the blade's airfoil
and to turn the blades out of the wind when maximum speed
is exceeded. To fulfill this assignment electrical machines
are used to control the blade's pitch angle function of wind's
speed. The faults that can occur in electrical machines can
stop the pitch control system, decreasing drastically the
efficiency of the wind turbine. Fault tolerance ability of the
pitch controlling machines can avoid such issues. In the
paper a fault tolerant switched reluctance machine is
studied. As it being simple and cheap it is optimal for such
vital applications. Numerical field computations were used
to analyzethe machine's behavior. An advanced
cosimulation was performed by coupling two simulation
programs in order to study the machine's drive system and
fault tolerance capacity.
Index Terms—Fault
machine, wind turbine, blade control.
tolerant, switched reluctance
I.INTRODUCTION
The global market of electrical energy demands for
intensive researches in means of developing new energy
resources, mainly renewable ones. Clean energy is
generate from renewable sources with low environmental
impact.
The most frequently used clean energy is produced by
specially designed wind turbines that transform wind
power in electricity. In our days modern wind turbines
are the fastest growing and most cost-effective renewable
energy technology in the world. The usage of wind
turbine generators is increasing, especially where
customers are environmentally conscious. In Europe, the
wind turbine market is ever increasing and there are
countries that exploit wind power in high proportions, as
Spain, Germany, etc.
Wind turbine aerodynamics remains a challenging and
crucial research area for wind energy. Clearly,
steady-state aerodynamic performance is essential to
turbine energy capture, since blade aerodynamic forces
produce mechanical energy that is subsequently
converted to electrical energy. However, more recent
inquiry has focused on adverse time varying aerodynamic
loads that wind turbines frequently suffer during routine
service.
These undesirable aerodynamic
excessive stresses on turbine blades and gear boxes, and
appreciably shorten machine service life.
loads impose
In the present, the wind energy technical community is
contemplating the utility of various aerodynamic control
methodologies for mitigating adverse aerodynamic
loading [1].
On a pitch controlled wind turbine the turbine's power
monitor checks the power output of the turbine several
times per second. When the power output becomes too
high, it sends an order to the blade pitch mechanism
which immediately pitches (turns) the rotor blades
slightly out of the wind. Conversely, the blades are turned
back into the wind whenever the wind drops again.
The rotor blades thus have to be able to turn around
their longitudinal axis (to pitch) controlled by a complex
system, as that shown in Fig. 1.
During normal operation the blades will pitch a
fraction of a degree at a time, and the rotor will be turning
at the same time.
Designing a pitch controlled wind turbine requires
some clever engineering to make sure that the rotor
blades pitch exactly the amount required. On a pitch
controlled wind turbine, the computer will generally pitch
the blades a few degrees every time the wind changes to
keep the rotor blades at the optimum angle in order to
maximize output for all wind speeds.
This way, the power captured from the wind power
can be controlled by a pitch actuator [2].
There are several control schemes that can be applied
to investigate the system behaviour in a wind turbine. It is
evident from the results that active pitch regulation
Generator
Power
Monitor
Controller
SRM
Grid
50Hz
Pitch control
Fig. 1 Pitch blade control system
721978-1-4244-2544-0/08/$20.00 ©2009 IEEE

Page 2

reduces the wind generator output fluctuations.
Inverter/converter control can smooth the voltage and
power output to make it suitable for the network
connection [3].
The development of a high reliability and safe
operational system can be achieved using a fault tolerant
variant of the switched reluctance motor.
This variantrequire
modifications [4], but also for special winding
connections and intelligent
compensate the higher costs demanded by a more
complex power converter the use of a relatively simple
machine is imposed.
To reach this goal a fault tolerant (FT) switched
reluctance motor (SRM) is considered as an optimal
solution.
The proposed SRM based drive system was studied by
means of simulations. The machine's model was built up
using specific software packages. Flux 2D, was used for
modelling the machine itself via finite element method
(FEM). The MATLAB-Simulink environment was
applied for modeling the inverter's control system and to
generate the faults for the studied cases.
These two programs were coupled and worked
together using the Flux-to-Simulink Technology. This
way it was possible to study in details all the typical
faulty states of the motor.
not only for structure
controldevices. To
II. THEPURPOSEDSRMSTRUCTURE
The development of a fault tolerant variant of an
electrical machine implies both topology and phase
connection modifications. The structure of the basic fault
tolerant SRM is given in [5], with a complex phase
system and high number of poles. The torque ripples of
this SRM are diminished due to its novel design. The
drawbacks of such system are the more complex winding
scheme and its power converter.
The fault tolerant SRM in study has a 12/14 pole
structure, as shown in Fig. 2.
This topology emerges from the classical 12/8 one.
The main modifications of the studied machine involve
increased rotor pole number and new stator winding
connections.
The machine has 12 concentrated stator windings
around each pole forming 6 phases. All the windings are
split in two parallel connected coils (called channels) as
shown in Fig. 2. Hence, each phase (like A, Fig. 2) is
divided into two channels (A1 and A2, Fig. 2). Two
windings (channels) from opposite poles form a phase.
Practically each phase is doubled. Channels belonging
to same phase are fed separately.
In such rotor design at each moment two adjacent
phases are fed (like A and B, Fig. 2), which means in
total 4 energized windings (like A1, A2, B1 and B2,
Fig. 2). This is called "two phase on" feed technique.
Hence at each moment two pairs of adjacent stator poles
are contributing to the torque development. This
connection helps the motor to overrun the poles with
faulty phases and to minimize the torque ripple. By
shorting the magnetic flux paths also lower iron losses
and higher efficiency can be achieved.
III. THEPOWER CONVERTER TOPOLOGY
The intelligence of the control system has to fulfill the
assignment of isolate, mask and remedy the faults in
manner that the machine's behavior to be influenced as
less as possible [6], [7]. This will be the main task of the
control system.
The purposed converter in study has a separate
asymmetrical half bridge connection for each channel, as
seen in Fig. 4.
A1
B1
A2
Fig. 2. The purposed fault tolerant SRM
B2
Fed
Phases
Fed
Phases
Fig. 3. The flux lines in the fault tolerant SRM
722

Page 3

This topology is a quite complex one: 24 solid-state
power switches are required. Beside this each branch
needs separate control
Nevertheless this sets the independence of the windings
in manner to control and feed each channel separately, as
the fault tolerant concept demands [8], [9].
and protection circuit.
IV. THESIMULATIONMODEL
The simulations were performed using the
co-simulation technique, by coupling two simulation
environments to work together.
The model of the SRM and the electric circuits of the
power converter were built up in Flux 2D [10].
Simulation of the power converter was possible using
an electrical circuit built in Electriflux, Flux's embedded
application for circuit design. This was attached to the
FEM model of the machine, developed in Flux 2D. The
electrical circuit for one channel is given in Fig. 5.
As compound of two channels, each phase will use
two asymmetrical half bridges, built around two coils,
each representing the "come and go" sides of the
channel's winding (see in Fig. 5).
Usual coils components (like BB1PL and BB1MI) are
used to link the two faces of each channel from the
geometrical design.
The power switches in the circuit were replaced with
resistors. This replacement will allow the modeling of the
switches by changing their resistance from a high value,
corresponding to OFF state of the switch, to a low value,
corresponding to its ON state. The management of the
reverse current is handled by the diodes.
The switching signals are obtained using the coupling
between MATLAB/Simulink and Flux 2D [11], [12]. The
finite elements model of the SRM practically is
embedded in the Simulink program via an S-type
function block called "Coupling with Flux2D". All the
control signals computed in Simulink are multiplexed and
enteres in this block. The main characteristics of the
machines computed via Flux 2D (currents in all the
phases, torque, speed and position) are returned to the
main program through another multiplexed signal line.
The Simulink model is given in Fig. 6.
As it can be seen there are six blocks, corresponding to
each phase, generating the command signals in
accordance with the rotor position and current values
from the previous time step.
Six sub-systems (blocks) are simulating the control of
the phase currents upon the PWM technique (see Fig. 7).
V
A1
A2
B1
B2
C1
C2
D1
D2
E1
E2
F1
F2
Fig. 4. The FTSRM's power converter
Fig. 5. The electrical circuit of one
channel
??????
????
??????????
??????????
??????????
??????????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
?????
????????????
?????
?????
?????
?????
?????
?????
?????
?????
??????????
??????????
????
???
?????????????
?
?
??????
???
????
???????
?????????
?????????
?
????
???????
?????????
?????????
?
????
???????
?????????
?????????
?
?????????????
??????
???????
????
???????
?????????
?????????
?
????
???????
?????????
?????????
?
????
???????
?????????
?????????
?
Fig. 6. The main window of the Simulink program
?
?????????
?
?????????
??????
????????????
????
?????????
??????
??????????
???????????
??????????
??????????
?
??????
??
???????
???
????
????????????????
??????
???????
?
?
???????
?
????
Fig. 7. The phase current controller block
723

Page 4

The commutations of the power switches are set upon
the rotor position.
For each time step, there are two main computations.
First, the Simulink model sets the command pulses for
the switches. These signals are sent to the Flux 2D
numeric field model and the computations for the given
time step are preceded.
Output values like phase currents, rotor position and
torque are sent back to the Simulink model. The
computation of the firing angles is set versus the rotor
position angle and they are set from an external m-type
file together with the maximum current value and the
changing switch resistance values.
In order to perform the coupled simulations some
parameter adjustments are required. Both programs
coupled together require high computer resources.
Therefore a compromise needs to be taken by lowering
the mesh density in order to obtain affordable
computation time but, in meantime, not to decrease the
precision of the results.
The open circuit faults will be simulated by imposing
OFF state for both switches of one phase/channel. The
precision of the PWM is set by the hysteresis band width
around the imposed current.
For this study the velocity was imposed to be
600 r/min. The considered
steady-state one, so no transient start-up was imposed.
The main interest was on the machine's behavior in
passing from normal operation to a faulted regime
regime was no-load
V. SIMULATIONRESULTS
Different cases were studied in order to examine the
machine's fault tolerance capability:
? normal operating mode (reference case),
? open circuit of one channel,
? open circuit of one phase,
? open circuit of two channels from different phases,
? open circuit of one phase and one channel from a
different phase (worst case in study).
These are considered as being the most common
defects that can occur in such system.
The machine's behavior in the above mentioned cases
were characterized through the mean torque values.
The simulated machine's geometrical dimensions are:
? stator outer diameter 190 mm
? stator inner diameter 141 mm
? stator pole depth12 mm
? rotor inner diameter 120 mm
? active stack length315 mm
The obtained results are given in Fig. 8, where the
phase currents and the torque of the SRM are plotted
versus time for each case in study.
The normal operation mode is reference for all the rest
of the faulty study cases. The mean torque developed in
this case is near 20 N·m (considered as 100%). All the
mean torques developed by the machine in faulty
operation cases will be referred to this value.
A. One channel open circuit failure
For simulating the open circuit failure the resistance of
the power switch was kept constant at a high value
corresponding to the OFF state. This way, the machine
will behave as having an open circuit winding on one
channel.
The current having nil value in Fig. 8b corresponds to
the faulty channel, as the currents are plotted only for one
channel of each phase.
The torque has a small fall corresponding to the
missing current pulse. The mean developed torque is
close to the reference value (see Table. 1), hence the
machine can operate with no significant behavioral
changes.
B. One phase (two channels) open circuit failure
The fault is modeled in a similar way as in the
previous case by setting OFF state of the power switches
in all the corresponding electrical circuit branches.
The phase currents in the case of open circuit are given
in Fig. 8c.
As the simulation was performed under no-load
conditions, at the occurrence of the phase damage, the
generated torque felled to nil. In this faulty case the
machine is able to develop more than 80% of the
reference rated torque.
Naturally, the torque ripples are greater than in the
previous case, becoming a compromise to fulfill the
continuous operation assignment.
TABLE I
DEVELOPED TORQUE VALUES
Mean torques [N·m] and
percentage of the rated torque
Studied cases
Healthy case19.93 (100%)
Faulty case 119.59 (98.29%)
Faulty case 216.16 (81.03%)
Faulty case 319.28 (96.71%)
Faulty case 413.79 (69.19%)
724

Page 5

?????????
????
??
???
?
??
?????
???????????
?????????
????
??
???
?
??
?????
???????????
a) normal study case
?????????
????
??
???
?
??
?????
???????????
?????????
????
??
???
?
??
?????
???????????
b) one channel open circuit
?????????
????
??
???
?
??
?????
???????????
?????????
????
??
???
?
??
?????
???????????
c) one phase open circuit
?????????
????
??
???
?
??
?????
???????????
?????????
????
??
???
?
??
?????
???????????
d) two channel open circuit
?????????
????
??
???
?
??
?????
???????????
?????????
????
??
???
?
??
?????
???????????
e) one phase and one channel open circuit
Fig. 8. The current and torque characteristics
725

Page 6

C. Two channels from two phases open circuit failure
In the situation of the faults on two different channels
from two different phases (Fig. 8d) the torque ripples are
higher than the case of one channel open circuit. But the
machine's mean torque seems not to be influenced by
these faults, as Table 1 show that more then 96% of the
rated torque is developed in this case. Due to the phase to
channel division, the machine can operate continuously
and can develop significant torque, even in the case of
this double failure condition.
D. One phase and one channel open circuit failure
The worst case in study is the fault of an entire phase,
and a second fault on one channel from different phase.
The torque characteristic falls to nil, because of the no
load and no inertia conditions of the simulqation
program's mechanical settings (see Fig. 8e).
The torque ripples are the most significant in this case
as compared to the previous ones. In spite of the severity
of the fault the machine is able to develop nearly 70% of
its rated torque.
VI. CONCLUSIONS
The coupled simulation program connecting two
software environments (FLUX 2D and Simulink) was
useful in studying the effects of different winding faults
on the torque developing capacity of the fault tolerant
SRM in study. Thus the computing power of FLUX 2D
joined the advanced facilities of MATLAB / Simulink in
simply describing the different working regimes of the
machines and drives taken into study.
Problems were regarding the computation times. In
order to obtain precise results in reasonable time the
quality of the FEM model's mesh had to be optimized.
It was demonstrated by means of simulations that
increasing the number of rotor poles, separating the
phases/channels, setting new connections between the
existing windings and using a complex control system all
provide good solutions for the fault tolerant SRM based
electrical drive system.
It was also proved that the proposed 12/14 poles SRM
has considerable fault tolerant capacity. This increased
fault tolerance unfortunately is achieved by more
complex constructions (especially that of the power
converter).
An improvement of the performances of the SRM
under faults will be studied in the future. After detecting
the fault in a winding it has to be isolated and the current
in the remaining healthy phases has to be increased. By
this the average torque can be held at its rated value, but
of course the torque ripples will be increased.
This solution can be used only if the windings and the
cooling system of the machine were designed to support
the greater currents.
REFERENCES
[1] Schreck, S., Robinson, M., Wind Turbine Blade Flow
Fields and Prospects for Active Aerodynamic Control,
ASME 2007 Fluids Engineering Division, San Diego,
California, July 30-August 2, 2007.
[2] Mulajdi, E., Butterfield, C.P., Pitch controlled variable-
speed wind turbine generation, IEEE Transactions on
Industry Applications, vol. 37, no. 1 (Jan/Feb 2001),
pp. 240-246.
[3] Sharma, H., Pryor, T., Islam, S., Effect of pitch control and
power conditioning on power quality of variable speed
wind turbine generators, Proceedings of the Australasian
Universities Power
(AUPEC '2001), 2001, pp. 95-100.
[4] Ruba, M., Szabó, L., Strete Larisa, Viorel, I.A., Study on
Fault Tolerant Switched
Proceedings of the 18thInternational Conference on
Electrical Machines (ICEM '2008), Vilamoura (Portugal),
on CD: Fullpaper_comm_id01200.pdf.
[5] Ruba, M., Anders, M., Fault Tolerant Switched Reluctance
Machine Study, Proceedings
Conference on Power Electronics, Intelligent Motion and
Power Quality (PCIM '2008), Nürnberg (Germany), 2008,
on CD: S2d-3.pdf
[6] Mir, S., Islam, M.S., Sebastian, T., Husain, I., Fault-
tolerant switched reluctance motor drive using adaptive
fuzzy logic controller, Proceedings. of the IEEE
International Electric Machines and Drives Conference
(IEMD '03), Madison (WI, USA); vol. 2, pp. 835-841.
[7] Park, B.-G., Kim, T.-S., Ryu, J.-S., Hyun, D.-S., Fault
Tolerant Strategies for BLDC Motor Drives under Switch
Faults, Record of the 2006 IEEE Industry Applications
Conference, Tampa (FL, USA), vol. 4, pp. 1637-1641.
[8] Szabó, L., Ruba, M., Fodorean, D., Study on a Simplified
Converter Topology for Fault Tolerant Motor Drives, in
Proc. of the 11th Intern. Conf. on Optimization of Electrical
and Electronic Equipment
(Romania), pp. 197-202, 2008.
[8] Kasson, M., Eaves, S., Fault tolerant motor drive
arrangement with independent phase connections and
monitor system, Patent no. WO01/91265, 29 November
2001.
[9] Briso-Montiano, J.R., Karrelmeyer, R., Dilger, E.,
Simulation of Faults by Means of Finite Element Analysis
in a Switched Reluctance Motor, in Proc. of the COMSOL
Multiphysics User's Conference, Frankfurt (Germany),
pp. 225-231, 2005.
[10] Busi, M., Cadeau-Belliard, S., Induction Motor Drive
using FLUX to SIMULINK Technology, FLUX Magazine,
no. 47, pp. 15-17, January 2005.
[11] du Peloux, B., Lacombe, G., Engineering-focused software
for the design of the drive of electrical machines, in Proc.
of the 18th Intern. Conf. on Electrical Machines
(ICEM '2008), Vilamoura (Portugal), pp. 1-4.
Engineering Conference
Reluctance Machines,
of the International
(OPTIM '2008),
Braşov
726