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

net/publication/3872962

Vector controlled induction machines for stand-alone wind energy

applications

Conference Paper  in  Conference Record - IAS Annual Meeting (IEEE Industry Applications Society) · February 2000
DOI: 10.1109/IAS.2000.882069 · Source: IEEE Xplore

CITATIONS READS

63 76

4 authors, including:

R. Pena Roberto Cardenas

University of Concepción University of Chile
122 PUBLICATIONS   5,690 CITATIONS    144 PUBLICATIONS   4,495 CITATIONS   

SEE PROFILE SEE PROFILE

Jon C Clare
University of Nottingham
428 PUBLICATIONS   12,749 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Modulation of Modular Multilevel Converters View project

Vector Control of a Variable Speed Doubly-Fed Induction Machine with the Matrix Converter for Renewable Energy Generation Systems View project

All content following this page was uploaded by Roberto Cardenas on 10 March 2016.

The user has requested enhancement of the downloaded file.

 Vector Controlled Induction Machines for Stand-Alone Wind Energy Applications

R. S. Pena"), R. J. Cardenas"), G. M. Asher(2),J.C. Clare(2)

( I ) Electrical Engineering Department ( 2 ) Electrical&Electronics Engineering Department.
University of Magallanes. University of Nottingham
Punta Arenas. Chile. Nottingham. U.K.
rpe@ona.Jiumag.cl greg.asher@nottingham.ac.uk

Abstract. This paper describes the system and control (DFIG) WEC systems of Figs. land 2 respectively. Both
structures for vector controlled induction generators used for systems include a main 3-phase output, an auxiliary load for
variable speed, wind energy conversion (WEC)systems. The power-flow control (e.g. heating, irrigation pumping) and a
paper focuses on WEC systems feeding an isolated load or weak flywheel energy store to smooth power fluctuations. Note that
grid since for such systems the generated voltage and power flow
must be regulated by the WEC system itself and the control
in Fig.1 for the CIG, the elements are connected to the DC-
structures are not trivial. The structures for both cage induction power bus through which all the WEC system power must
and doubly-fed induction generators are described and their pass. In the DFIG system, the elements are shown affixed to
performance and complexity compared. Experimental results the AC lines and the pair of back-to-back vector controlled
for both systems feeding an isolated load are given and show that PWM converters are rated only for a fraction of the WEC
both systems are capable of good output voltage regulation system power [8]. As before, an auxiliary load and flywheel
whilst tracking the optimal speed for energy capture. are used for power flow and power smoothing control, and
these are connected to the AC line through appropriate
I. INTRODUCTION converters. The converter-fed drives for the auxiliary diesel
generator are not shown in Figs. 1 and 2.
The robustness, low maintenance and relative cheapness of
induction machines are well known and have long been
considered as a good choice as the electrical generator in
WEC systems. The vector or field-orientated control of
induction generators yields high dynamic torque and field
control and are thus ideal for variable-speed WEC systems. Auxiliary
Whilst vector controlled generators are now attracting interest Load
[l-51, nearly all control research has addressed the case of
grid-connected WEC systems. In these, the grid acts as a stiff qGJa--@ Flywheel
voltage source so that the generator control structures are Fig 1 WEC with Cage Induchon Machine
those of standard 4-quadrant vector drives with a sinusoidal
front-end converter.
For WEC systems feeding an isolated (or stand-alone) load
or acting as the main source on a small local grid, the WEC
system itself must regulate the output voltage and the power
flow. This is a more difficult control problem and has
attracted little research attention. The paper addresses this
problem. It is assumed that the WEC system can supply 100%
of the total load power and is augmented by a secondary
source (e.g. diesel generator). Maximum wind energy capture Flywheel
is assumed and the WEC system is thus variable speed and Fig. 2. WEC with Doubly-fed Induction Machine
driven at the optimal shaft speed [6,7]. It is further assumed
that the utility can absorb all the wind power and an auxiliary The control of these converters follows the same structure
load (e.g. auxiliary heating or irrigation schemes) is therefore as that of the wind-turbine generators with appropriate control
required. The auxiliary load itself will have a maximum transfer between the two source elements. This control
rating, and further excess power is limited through blade pitch transfer again constitutes an operational mode control and is
or stall control. These last operational mode control functions beyond the scope of this paper. The vector control structures
are not considered in the paper. for the variable speed generator, power-flow control and
The two schemes considered cover the cage induction auxiliary power smoothing are first described. Two
generator (CIG) and the doubly-fed induction generator experimental rigs implementing stand-alone cage generator

0 2000 IEEE
0-7803-6401-5/00/$10.00
1409
and DFIG systems have been constructed and results given “soft” optimal speed tracker; superior tracking can be
illustrating optimal wind-speed tracking, and power flow and obtained via a mechanical torque observer and driving the
voltage regulation The paper discusses the control complexity generator to the corresponding optimum speed [lo]. This
of the two structures. “hard” tracker however yields increased power fluctuations
and increases the rating of the flywheel drive. Note that
11. CONTROL STRUCTURES independent setting of i,* is not possible for optimum energy
extraction.
A. Cage Induction Generator For stand-alone applications, the CIG cannot control the
DC-Link voltage without varying its flux level. This is not
The control structure for the turbine fed CIG operating appropriate for this type of machine and the DC-Link voltage
under maximum energy extraction is shown in Fig. 3a. It is must be controlled by the power flow [11,12]. The DC-Link
based on a standard Indirect Rotor Flux Orientated (IRFO) Voltage is controlled by adjusting the net active power
control of the induction machine in which the dq current and current, drawn by the auxiliary load iq.aux and flywheel
voltage values are referred to the reference frame aligned to ipfi? (that imparted by an auxiliary diesel generator may also
the rotor flux. The dq currents are controlled in this reference be included here). Note that and iq-fiy will correspond to
frame and the PI controllers are designed for current (and the field-orientated torque currents of the auxiliary load and
hence torque) bandwidths of 300-500Hz. Although IRFO is flywheel if driven by vector controlled converters. If other
shown here, Stator flux orientation can be used and the choice converters are used (e.g. DC choppers) then the active
is not critical. The id* demand sets the machine flux level currents can be replaced by any quantity that represents the
which is here defined constant so that the drive is rated for control of real power. A suitable control scheme is shown in
operation below base speed and is able to develop full torque Fig. 3b. This contains a feed forward block that represents the
and over-torque. The torque current demand i,* is derived steady state relationship between (IG-IL)and iq.neland is highly
from: desirable since the dynamic relation between and Vdc is
very non-linear [12]. Note that the iy.aur* demand is passed
through a low pass filter, the scheme depicted ensuring that
T e * = K o p I ~ -r B2 u r , i, * = T , * l k i d (1)
the auxiliary load and flywheel absorb the low and high
frequency power fluctuations respectively. The iy.aux* and
whereu, is the measured or estimated generator speed, B an iq-fly* demands are fed to the appropriate vector controlled (or
equivalent friction coefficient and Kopla constant defining the other) converters as appropriate. In practice, the flywheel will
relationship between the maximum power and the shaft speed have upper and lower speed limits and the speed-dependent i,
for a given blade characteristic [9]. Equation (1) represents a limits may be easily implemented in Fig. 3b.

I -----------------(am
mr I.
Fig. 3a. CIM control schematic.

1410
 which provides the required iqr demand of Fig. 4a. The
orientation may be termed Indirect Stator Flux Orientation
(ISFO). Substituting (3) in (2a) and defining the magnetizing
current as ,i = yields:

Fig. 3b. Power flow controller with steady state feed-forward (G). 1+OS
_
Ls _dt_+imS = idr +_ _v
(4)
ds
B. Doubly-fed Induction Generator. RS RS

The control structure for the turbine fed DFIG operating so that the machine flux level (i.e. imS) may be controlled by
under maximum energy extraction is shown in Fig. 4a. Unlike i,d* as shown. Since the stator resistance is small, the setting
the CIG, the DFIG may generate the AC bus voltage directly. of i, effectively fixes V, to within 1-2%. Finally, the stator
Again, the optimal speed tracking and the absence of a fixed frequency is user-defined as q in Fig. 4a.
grid voltage impart some unusual (and novel) features to the The output frequency is thus fixed and cannot fluctuate.
system control. Most significantly, the stator-flux vector The vector control of the front-end (stator-side) converter is
orientation, standard for these machines, will not be stable, conventional [ 141 although the orientation angle should not be
especially when harmonic perturbations are present [ 131. To derived from the machine stator voltages since it is sensitive
derive a solution for the orientation and the voltage control, to even small harmonic distortions. It is proposed that the
we use the dq axis dynamic equations: orientation angle for the stator side converter be calculated as

v , ~= R,si,s+ phs +jw, hs

v, = R,i, + ph, +j(wt - @) hr
h,T= L, is + L,i, be used in which e, is the integral of the frequency demand in
Fig. 4a. Since the AC voltage is generated by the DFIG, it
where we is the dq axis rotational frequency. All other cannot be used for power-flow control. Thus the net active
variables have their normal meaning. From (2c) we note that power current ,.i is derived from the error between the real
the system will be orientated on the stator flux if we force the DFIG torque and that required for optimal tracking. The
condition:

Fig.4a. DFM control schematic.

1411
structure is shown in Fig. 4b and is analogous to that of obtained from [15]. The optimal speed (dotted line)
Fig. 3b. corresponds to the instantaneous wind speed and the real shaft
speed is seen to follow well. If the power fluctuations
corresponding to Fig. 6b need to be reduced (for example to
G reduce the rating of flywheel drive), a filter can be introduced
1. in to the iq* demand of Fig. 3a; the energy capture will of
course be reduced.

I I I I

pJ- 0
Time(s)
Fig. 4b. Control Scheme for optimum speed tracking. CIM Fig. 6a. CIM performance for step Changes in Wind Velocity

111. ILLUSTRATIVE EXPERIMENTAL RESULTS. 1 100

A. Cage Induction Generator

A 2.5kW vector controlled CIG drive rig has been

constructed and is shown in Fig. 5. The generator is driven by
a DC machine emulating wind profiles, and an auxiliary load
in the form of a DC chopper affixed to the DC link. The main
parameters of the 2.5kW system are given in Appendix IA.
The system is controlled by a Transputer network in which T2
carries out the control functions of Figs. 3a, 3b, the others I
0 20 40 M) 80 100 120 140 160 180
responsible for data gathering, PWM and user interfacing. Time(s)
r------------------------------------- Fig. 6b. CIM speed tracking performance
I

The control scheme of Fig. 3b has also been tested by

replacing the DC machine with a flywheel and energising the
DC-link with a step-up chopper. These results have been
reported in [12] and Figure 7 is illustrative; it shows the
Transputer board excellent DC-link voltage regulation to a 50% step in load
................. ...........
current iL. The quality of regulation is mainly due to the feed
T, i System Interface forward term in Fig. 3b.

670. - 2.8
E
2.1 3
Fig. 5. Stand alone CIM Experimental set-up
" 8 B
8
i 1.4
Figs. 6 test the control scheme of Fig 3a. Fig. 6a shows the
"soft" speed tracking response to a wind-speed step of 5 d s to x MI- I
- 0.7 3
8 d s , corresponding to an optimal shaft speeds of 600 and
97" respectively. The idand iq currents are also shown and 630
-- &.
-U-
0.0

are as expected. The time response of soft tracking strategy is 6201 I-0.7
0 0.2 0.4 0.6 0.8 1 1.2
dependent on the step size, but responses of 1 to 8 seconds are
T i (s)
typical. Fig. 6b shows the tracking to a typical wind profile
Fig. 7. DC link voltage control performance with flywheel operation

1412
B. Doubly-fed Induction Generator and the duty cycle 6 of the mxiliary load chopper (6 replaces
in Fig. 3b since there was no flywheel in this test).
A 7.5kW vector controlled DFIG rig has been constructed and Again the response times reflect the system inertia.
is shown in Fig. 8. The main parameters are given in
Appendix B. Again, a DC drive provides the turbine
emulation whilst the auxiliary load takes the form of a
variable duty cycle chopper. The LC filters were an attempt to
reduce the effect of the harmonics on the AC lines. It was
significant that even with heavy filtering, only the orientation
schemes of Section 2B proved successful.

"0 '5 10 15 '20 25 30 35 '40 '45"

Timecs)
Fig. loa. Speed response to a step increase in wind velocity

Fig. 8. DFIM experimental setup

The AC line voltage was controlled to 50Hz,220V; the lower

voltage being due to the converter voltage ratings limiting the
DC-link voltage to 550V. The optimal tracking control was
tested and Fig. 9 shows the speed tracking response to a real
wind profile and the corresponding rms voltage regulation.

0 1 2 3 4 5 6
' -6
Time@)
Fig. lob. Performance for operation through synchronous speed

Finally, Figs. 1 l a and 11b show the response of i,, , i , , i,,s

and 6 for step changes in load power. The effectiveness of the
power flow control is evident.

1
z50 t 1175

Fig. 9. Voltage control and speed tracking with realistic wind profile.

The real shaft speed is relatively smooth owing to the U

V
larger inertia of the wind turbine. The regulation is about 5V r
3 -
d I -

(2%), the variation deriving from neglect of R,. The power

0
flow control was tested by setting the main load to balance 0 5 IO 15 20 25 30
the wind power at =900rpm, and then introducing a step
increase in wind velocity. The corresponding final shaft
velocity being 1200rpm. Fig. 10a shows the speed response

1413
 DFIG DFIG
CIG Flywheel on Flywheel
AC bus on DC bus

Converter Rating 3 2.6 0.9

Current transducers
5 10 6

I,
I 2o
(inc. feedforward cont.)

Voltage Transducers 1 2 1
5 IO 15 20 25 30
Time(s)
Fig. 1 Ib. Magnetising current control performance for step change in load at APPENDIX
A
optimum speed.

IV. DISCUSSION
ANDCONCLUSION Experimental System Ratings for CIM:

Wind Turbine (Emulated by the wind turbine)

This paper has described and reviewed vector control
schemes for both Cage and Doubly-fed Induction Generators
operating in a stand-alone or weak grid environment. In such Power : 3.2kW
applications, the vector control strategies may not be simple, Blade radius : 2.26m
particularly for the DFIG. Wind turbine rated speed : 296 rpm
The control complexity lies in the fact that stable grid Rated wind speed : IOms-’
voltages are not available and that power flow balance Cut-in speed : 4ms-I
assumes a central importance. Inertia = 1.8kgm2.
For the CIG, the power flow balance was used to control Gear box : 2.836.
the system output (DC-Link) voltage with the generator
torque derived from the tracking of the optimal speed for Cage Induction Machine
maximum energy capture.
For the DFIG, the balance between the generated and ideal Power : 2.5kW
torque (or power) provides the mechanism for controlling the Voltage : 380V
auxiliary power flow. This structure is required because the Connection : Star
i,* demand is used for forcing the Indirect Stator Flux Rated speed : 1450rpm
Orientation required for the DFIG. Pole pairs : 4
It is of interest to compare the complexity and converter Id (rated) : 1.8A
rating of the two systems. The comparison is dependent on R,=2.1!2
the nature of the auxiliary load (which may be of any type as R ~ 1 . 7 0
long as the power flow is controllable) so for simplicity this L,= 13.32mH
load is neglected. L,,=12.9 1mH.
We can assign a converter rating of 1 to a single 3-phase to L,= 13.32mH
DC converter rated at the turbine power and further assume a Lo=12.91mH.
slip power rating of 0.3. A count of converter rating and
number of voltage and current transducers for the schemes of APPENDIX B
Figs. 1 to 4 are given in Table 1. Speed transducers are not
included since all the drives may be effectively vector Experimental System Ratings DFIM:
controlled without shaft speed sensors.
The DFIG is clearly more complex and expensive than the Wind Turbine (Emulated by the wind turbine)
CIG. However, if the flywheel speed is limited about a rated
speed, it may be advantageous to drive the flywheel with a Power : 7.5kW
DFIM whose rotor converter is fed off the reduced power DC Blade radius : 3.24m
bus. This case is also included in Table 1 (“flywheel on DC Wind turbine rated speed : 296rpm.
bus”). In this structure, the DFIG system is seen to be Rated wind speed : 1Oms-’
attractive and will be the subject of further research. Cut-in speed : 4ms-’
Inertia = 7.5 kgm2
Gear box 5.065.

1414
 Doubly-fed Induction Machine: r21 J. Sallan, E. Muljadi, M. Sanz, and C. P. Butterfield, “Control of a
self-excited inducfion generator driven by Wind”, 9” European
Conference on Power Electronics Applications. EPE’99. Lausanne.
Power : 7.5kW Switzerland 1999.
Stator voltage : 415(V) E. Levi and Y.W. Liao, “Rotor flux oriented induction machine us a
Rotor voltage : 440V DC power generator”, 9” European Conference on Power Electronics
Rated Stator current : 19A Applications. EPE’99. Lausanne, Switzerland 1999.
M. Godoy S h o e s , B. K Bose, and R. 1. Spiegel. “ Design and
Rated rotor current : 11A performance evaluation of a fuzzy-logic-based variable speed wind
Stator connection : Delta generation system”,. IEEE Trans. Industry Applications, vol. 33, pp.
Rotor connection : Star 956-965. July/AuguSt 1997.
Rated speed : 970rpm R. Jones and 1. Gilmore, “Benefits of sinusoidal rectifier.s on variable
speed wind turbines-High quality mains power from variable speed
Pole pairs : 6 wind turbines”,. Proceedings 17” BWEA Wind Energy Conference. pp.
R, : 1.06!2 339-345. 1995.
R, :0.8Cl Freris L., Wind Energy Conversion System, Prentice Hall, 1990.
“Electrical generation aspects r$ wind turbine operation ”, Proceedings
LS=0.2065H of the BWEA-D workshop, ppll-20. 1987.
L,=O.O8 10H (referred to the rotor) G. Smith and K.A. Nigim, ” Wind-energy recovery by a static
Lo=0.0664H (referred to the rotor) Scherbius induction generator. ” IEE Proceedings C, ~01.128,pp.317-
Stator-rotor turns ratio : 1.7 324, 1981.
D.L. Gourieres, Wind Power Plants, Theory and Design, Pergamon
Press, 1982.
A c KNOWLEDGMENT r101 B. Connors and W.E. J-eithhead. Investigation of a fundamental trade-
‘I

off in Tracking the C,,,, curve of variable speed wind turbines”,

The financial support provided by Fondecyt, through BWEA Conference, pp.313-319, 1993.
U11 F. Hardan, J.A.M Bleij, R. Jones and P. Bromley, “Bi-directional
project 1980688, is kindly acknowledged. The financial power control for flywheel energy store system with vector. controlled
support provided by the British Council through the induction mchine drive ”, IEE Conference publication institution of
Academic Link Programme (SAN/984/) is also electrical engineers, pp. 456-1998.
acknowledged. [ 121R. Cardenas, R. Pena, G. Asher and J. Clare, “Conrrol strategies for
energy recovery from a J’ywheel using a vector controlled induction
machine”. Accepted for publication in PESC’2000.
REFERENCES. ~ 3 1Pena R., “Vector controlled strategies for a doubly-fed induction
generafor driven by a wind turbine”, Ph. D. Thesis, University of
E.Goncalves-Marra and J. Pomilio, “Self-excited induction generator Nottingham, U.K., 1996.
controlled by a VS-PWM bidirectional converter for rural [I41 N.R Zardary and G . Joos, “Performance investigation of a current-
applications”, IEEE Trans. Industry Applications, vol. 35, pp. 877- controlled voltage regulator PWM rectifier in rotating and stationary
883. July/August 1999. frames”, IEEE Trans. On Industry Applications, vol. 42, pp396-401,
1995.

1415

View publication stats