IEEE Energy2030
Atlanta, GA USA
17-18 November, 2008
Doubly-Fed Induction Generator Control Under Voltage Sags
K. Lima*, A. Luna♠, P. Rodriguez♠, E.Watanabe*, R. Teodorescu♣ and F. Blaabjerg♣
*FEDERAL UNIVERSITY OF RIO ♠ ♣
TECHNICAL UNIVERSITY OF
DE JANEIRO
Power Electronics Laboratory Department of Electrical Engineering
Rio de Janeiro – BRAZIL
Abstract – This paper proposes a new control technique to improve
the fault-ride through capability of doubly fed induction
generators (DFIG). In such generators the appearance of severe
voltage sags at the coupling point make rise to high over currents
at the rotor/stator windings, something that makes necessary to
protect the machine as well as the rotor side power converter. As a
difference with the most extended solution, that reduces these
currents by means of the connection of a crowbar circuit, this
works intends to enhance the DFIG’s response in such conditions
without introducing extra hardware in the system. To this end, the
proposed control system feedback the stator currents as the rotor
current reference during the fault, until they are driven to their
nominal values. The feasibility of this proposal has been proven by
means of mathematical and simulation models, based on
PSCAD/EMTDC.
I. INTRODUCTION
The increasing capacity of the installed wind power
generation facilities world wide, as well as the high scale
penetration of such systems in the next future, have made
necessary to renew the grid code requirements regarding the
operation of wind farms. For the wind power case these grid
codes demand control requisites similar to those demanded in
conventional power plants, which are basically focused on
achieving a stable and reliable operation of the electricity
network.
One of these requirements is the capacity of wind power
generators to remain connected to the grid in case of severe
grid voltage sags. This feature, also known as fault ride-
through capability (FRT), is specially critical in DFIG based
wind turbines (WT). Several studies have been carried out in
order to improve the ride through capability in this kind of
WTs [1-8]. The main objective of these works has been
focused on avoiding the disconnection of the system during
the sag. Among the different proposed solutions, those based
on limiting the over currents at the machine and at the power
converter, as well as, over voltages at the DC bus of the
converters by means of auxiliary circuits stand out. However,
this kind of solution implies the installation of extra devices
in the system, something that finally increases the costs and
hinders its reliability, as the control system should became
more complex. These circuits do not only reduce the
reliability of the system, but require an extra control effort. In
addition the design of such protection is calculated
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 13, 2009 at 06:16 from IEEE Xplore. Restrictions apply.
AALBORG UNIVERSITY
CATALONIA Institute of Energy Technology
DK-9220 Aalborg - DENMARK
Barcelona – SPAIN
considering the features of the wind power system, and hence
any future change of its configuration would affect its
performance. Regarding this solution several crowbar
topologies can be selected, as it is explained in [9-12].
Other alternatives to crowbars, as the inclusion of static
keys between the stator and the electrical grid, presented by
Petterson [13], or the stator flux damping proposed by Næss
[14] have been also published. However all these techniques
require also the installation of extra active/passive elements
in the power system.
The aim of this paper is to propose a new control
technique that would permit to guarantee the controllability
of DFIG WTs during severe voltage sags near the point of
common coupling (PCC) without adding extra hardware. This
solution will be based on the feedback of the real statoric
currents to the current setpoint of the rotor converter. The
feasibility of this proposal will be analyzed in the following
sections and its performance will be tested by means of
simulation models based on PSCAD.
II. MODELLING OF THE GENERATOR UNDER FAULT
CONDITIONS
Considering the classical single cage fifth order model of a
DFIG [13] the voltage and magnetic flux of the stator can be
written as in (1), and, in an analogous way, the equations that
describe the dynamics of the voltage and magnetic flux in the
rotor as in (2).
d λds
vds = Rs ids + − ωs λqs
dt
d λqs
vqs = Rs iqs + + ωs λds (1)
dt
λds = Ls ids + Lm idr
λqs = Ls iqs + Lm iqr
Considering that the system described by (1) and (2) is
linear, something that assumes that the magnetic circuit of the
DFIG is linear, it is possible to obtain the following statoric
currents, (3)-(4), in the synchronous reference frame [14].
 d λdr III. CONTROL STRATEGIES FOR THE ROTOR SIDE
vdr = Rr idr + − ωslip λqr CONVERTER UNDER FAULTY SCENARIOS
dt
d λqs Once the reliability of (5) and (6) has been tested it is
vqr = Rr iqr + − ωslip λdr (2)
dt possible to design different control strategies for limiting the
λdr = Lr idr + Lm ids over currents in the machine during a voltage sag. In this
paper two algorithms will be presented. The first one intends
λqr = Lr iqr + Lm iqs to reduce these currents by setting an active and reactive
power equal to zero during the fault, while the other is based
( Ls s + Rs ) vds + ωs Ls vqs on a direct feedback of the measured stator currents as the
ids =
(L s 2 2
s + 2 Ls Rs s + Rs2 + ωs2 L2s ) setpoint for the current controller of the rotor side converter.
(3)
−
(L s + R s + ω L )L i
s
2
s
2
s s m dr − Rsωs Lm iqr
( L s + 2L R s + R
2 2
s s s
2
s + ωs2 L2s )

−ωs Ls vds + ( sLs + Rs ) vqs

iqs =
(s L 2 2
s + 2 sLs Rs + Rs2 + ωs2 L2s )
(4)
Rsωs Lm idr + ( s 2 Ls + sRs + ωs2 Ls ) Lm iqr
−
(s L 2 2
s + 2sLs Rs + Rs2 + ωs2 L2s )

The equations (3) and (4) can be simplified considering

that in a field oriented control system the stator flux is
aligned with the dq reference frame, and hence its quadrature
component is null. Besides, and assuming that the leakage
inductance value is low, the stator voltage vector can be Fig. 1. Accuracy of the presented simplified model compared with
considered almost aligned with the in-quadrature axis, then the full fifth order model. Stator currents.
its d component is null. In this manner the previous equations
can be rewritten as shown in (5)-(6):

ωs Lm
ids = vsq − idr (5)
Ls  s 2 + 2 s + ω 
Rs 2 Ls
 Ls 
s

 s + Rs 
 Ls  L
iqs =  vsq − m iqr (6)
Ls  s 2 + 2 s s + ωs2 
R Ls
 Ls 

The equations (5) and (6) evidence how any variation in

the stator voltage introduce oscillations in the dq components
of the stator currents in the synchronous reference frame.
This phenomenon can be specially noticed during a voltage
sag. These last equations give a simple relationship between
stator and rotor variables that permits to predict the stator
current behavior easily. If compared with the full fifth order
model the accuracy of the resulting system is quite good, as it
can be noticed in Fig.1 and Fig. 2
Fig. 2. Accuracy of the presented simplified model compared with
the full fifth order model. Rotor currents.
A. Active and Reactive Power Setpoint equal to Zero
An already existing strategy for reducing the current peaks
during a voltage sag at the PCC is based on changing the
active power setpoint to zero. This control system permits to

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 13, 2009 at 06:16 from IEEE Xplore. Restrictions apply.
 Fig. 3. Layout of the proposed control system for the rotor side converter. The position of the switch is modified from ‘1’ to ‘2’ when the fault is
detected, enabling the proposed control strategy during the sag.

reduce the stator currents since, in theory, no energy would

flow through the stator. 1 ωs
Considering that the control of the power delivery is based idr = vqs (11)
on the field oriented control (FOC) philosophy the active and Lm  s 2 + 2 Rs s + ω 2 
 Ls s 
reactive power magnitudes can be calculated thanks to (7)  
and (8).
 s + Rs 
ps = vqs iqs (7) 1  Ls 
iqr =  vqs (12)
Lm  s 2 + 2 Rs s + ω 2 
qs = −vqs ids (8)  Ls s 
 
The second term of (7) and (8), that concerns the stator
currents, can be substituted with the expressions found in (5) B. Feedback of the Stator Currents
and (6), getting the equations (9) and (10).

  s + Rs   This proposal intends also to design a control strategy for

  L  L  reducing the currents in the stator/rotor windings when a fault
Ps = vqs   s
vqs − m iqr  = 0 (9) affects the generator. The philosophy of this control, as it is
 Ls  s 2 + 2 Rs s + ωs2  Ls 
 shown in Fig. 3, is to feedback the measured stator currents
  Ls   as the set point for the current controller of the rotor side
converter when a voltage dip occurs. In this manner the
  current control system synthesizes rotor currents that generate
 ωs Lm  currents waveforms in the stator windings, with the same
Qs = −vqs  vqs − idr  = 0 (10) shape of the currents generated during the sag but in counter-
 Ls  s 2 + 2 Rs s + ωs2  Ls 
 phase.
  Ls   In this second strategy the current setpoint for the rotor
converter coincides with the measured values of the stator
currents.
Considering that the magnitude of the voltage in (9) and (10)
can be different from zero the only way to guarantee a null idr* = ids (13)
delivery of active and reactive power is to set the rotor
current reference equal to zero. This condition yields the rotor iqr* = iqs (14)
current components shown in (11) and (12).
The performance of the rotor current can be found easily
considering the conditions (13) and (14) in (5) and (6),
finding finally the expressions (15) and (16).

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 13, 2009 at 06:16 from IEEE Xplore. Restrictions apply.
 1 ωs
ids = idr = vqs (15)
Ls + Lm  s 2 + 2 Rs s + ω 2 
 Ls s 
 

 s + Rs 
1  Ls 
iqs = iqr =  vqs (16)
Ls + Lm  s 2 + 2 Rs s + ω 2 
 Ls s 
  Fig. 6. Stator currents performance under balanced sag conditions.

IV. SIMULATION RESULTS

The performances of the discussed strategies have been tested

considering a three phase balanced voltage sag, produced by
a three phase short circuit in the distribution grid. The voltage
waveforms during the sag are depicted in Fig. 4.

Fig. 7. Behavior of the electromagnetic torque during a balanced sag

As it can be noticed in Fig. 5 and in Fig. 6, both strategies

Fig. 4. Voltage at the PCC during the balanced sag produced in t =
2s and cleared after 200 ms.
The performance of the rotor and stator currents when both
strategies are implemented can be compared thanks to the
simulation plots depicted in Fig. 5 and Fig. 6. The first plot
shows the response of the rotor currents during the sag, while
the second one in focused on the stator current behavior.
permit to reduce the currents in the generator’s windings
during the fault. However the proposed algorithm produces
minor oscillations in the current waveforms if compared with
the simulation response obtained with the zero power set
point strategy. As a consequence, in Fig. 7 the magnitude of
the electromagnetic torque ripple in the first strategy is lower
also. These results were in line with the models presented
previously.
Besides, in this part it is worth to point out that in this
particular case the natural oscillation of the system has the
same frequency of the grid, ωs. This occurs because the
voltage sag is balanced and hence it contains a single positive
sequence component. As it will be discussed as follows this
oscillation would be different if the negative sequence is not
null. Because of that, this analysis has been extended to
unbalanced voltage sags as well. The importance of this kind
of disturbance lies in the fact that the voltage in the stator
contains a negative sequence component during the sag. The
testing signal in this case is plotted in Fig. 8, and corresponds
to a C type voltage dip.

From Fig. 9 to Fig. 11 the behavior of the stator/rotor currents

Fig. 5. Rotor currents performance under balanced sag conditions
In Fig. 7, the behavior of the electromagnetic torque during
the sag is shown as well.
and the electromagnetic torque in the DFIG during the fault
described in Fig. 8 are shown.
As in the previous case Fig. 9 and Fig. 10 stand out how the
currents in the generator windings are strongly reduced
during the fault.

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 13, 2009 at 06:16 from IEEE Xplore. Restrictions apply.

Fig. 8. Voltage at the PCC during the unbalanced sag produced in t =
2s and cleared after 200 ms.
Fig. 9. Rotor currents performance under unbalanced sag conditions
Fig. 10. Stator currents performance under unbalanced sag
conditions
Fig. 11. Behavior of the electromagnetic torque during an unbalanced
sag.
Moreover, the fluctuations of both variables in these graphs
are a combination of the natural and forced response of the
system, containing a composition of frequencies equal to ωs
and 2ωs. As it can be seen in the figures, these oscillations
finally have a unique frequency of 2ωs, as the natural
response of the system trends to zero. This phenomenon can
be seen with greater clarity in Fig. 9 and Fig. 11.
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 13, 2009 at 06:16 from IEEE Xplore. Restrictions apply.
V. CONCLUSIONS
In this paper the modeling of a DFIG, considering the
behavior of the generator when transients in the stator voltage
occur, was developed. Thanks to this model, that permits to
predict the performance of a DFIG under faulty scenarios, a
novel control strategy for the rotor side controllers, oriented
to enhance its response during severe voltage sags, was
proposed. This strategy is based on using the measured stator
current values as the setpoint for the rotor current controller
during the fault. As it has been demonstrated, in this way, it is
possible to synthesize a current in the stator in opposition to
the currents generated during the fault, preventing thus the
stator/rotor windings from suffering overcurrents, with no
need of using crowbar circuits.
The performance of this strategy was tested and compared
with another control algorithm, already presented in previous
works, which makes the active and reactive power references
equal to zero during the fault. The analytical equations and
the simulation models based on PSCAD/EMTDC have
shown that the fluctuations in the stator and rotor currents, as
well as in the electromagnetic torque were reduced by the
half, when using the new proposed strategy in case of severe
balanced and unbalanced voltage sags at the PCC.
Therefore the results presented in this paper, show that it is
possible to control the stability of a DFIG during severe
contingencies in the power network, without the need of
external auxiliary circuits.
VI. ACKNOWLEDGEMENT
. This work was supported by the Ministerio de Educacion y
Ciencia of Spain under Project ENE2008-06588-C04-
03/ALT and ENE2008-06841-C02-01/ALT and by the CNPq
under numbers 473294/2004-4 and 141079/2005-3 and
FAPERJ.
.
VII. REFERENCES
[1] S. Seman et. al. “Ride-Through Analysis of Doubly Fed Induction
Wind-Power Generator Under Unsymmetrical Network Disturbance”.
IEEE Transactions on Power Systems. Volume 21, Nº. 4, Nov. 2006.
page. 1782 – 1789.
[2] C. Klumpner and F. Blaabjerg. “Experimental Evaluation of Ride-
Through Capabilities for a Matrix Converter Under Short Power
Interruptions”. IEEE Transaction on Industrial Electronics. Vol. 49, nº
2, April, 2002. Page(s): 315 – 324.
[3] F. M. Hughes et. al.“A Power System Stabilizer for DFIG-based wind
generation”. IEEE Transaction on Power Systems. Vol. 21, Issue 2,
2006. Page(s):1558 – 0690.
[4] F. M. Hughes et. al .“Control of DFIG-based wind generation for
power network support”. IEEE Transaction on Power Systems. Vol.
20, Issue 4, 2005. Page(s):1958 – 1966.
[5] B. Xie et. al. “Study of fault ride-through for DFIG based wind
turbines”. IEEE International Conference on
 Electric Utility Deregulation, Restructuring and Power Technologies,
2004. (DRPT 2004). Vol. 1, Issue , 5-8 April 2004 Page(s): 411 –
416.
[6] J. Morren et. al .“Ride through of wind turbine with doubly-fed
induction generator during a voltage dip”. IEEE Transaction on
Energy Conversion. Vol. 20, nº 2, 2005. Page(s):435 – 441.
[7] A. Petersson et. al. “A DFIG wind-turbine ride-through system
influence on the energy production”. In: Nordic Wind Power
conference, Chalmers University of technology, Göteborg, 2004.
[8] D. Xiang et. al. “Control of a doubly-fed induction generator in a
wind turbine during grid fault ride-through”. IEEE Transaction on
Energy Conversion. Vol. 21, nº 3, 2006. Page(s):652 – 662.
[9] I. Erlich et. al. “Dynamic Behavior of DFIG-base Wind Turbines
during Grid Faults”. Power Conversion Conference. April 2007.
Nagoya. Page(s):1195 – 1200.
[10] J. Morren, S.W.H. de Haan. “Short-Circuit Current of Wind Turbines
with Doubly Fed Induction Generator”. IEEE Transaction on Energy
Conversion. Vol. 22, Issue 1, March 2007. Page(s):174 – 180
[11] M. Rodriquez, G. Abad, I. Sarasola, A. Gilabert. “Crowbar control
algorithms for doubly fed induction generator during voltage dips”.
2005 European Conference on Power Electronics and
Applications.11-14 Sept. 2005 Page(s):10 pp.
[12] Morren, J.; de Haan, S.W.H. “Short-circuit current of wind turbines
with doubly fed induction generator”. IEEE Transaction on Energy
Conversion. Volume 22, nº 1, March 2007 Page(s):174 - 180.
[13] W. Leonhard, Control of Electrical Drives, Springer, 3rd Edition,
New York, 2001.
[14] F.K.A. Lima and E.H. Watanabe, “Parallel Connection of Doubly-Fed
Induction Generator in Wind Generation”, in 9º Congresso Brasileiro
de Eletrônica de Potência, vol. 09, pp. 631-639, 2007.

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 13, 2009 at 06:16 from IEEE Xplore. Restrictions apply.