800 IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 19, NO.

4, DECEMBER 2004
Comparison of the Response of Doubly Fed and Fixed-Speed Induction
Generator Wind Turbines to Changes in Network Frequency
Janaka Ekanayake and Nick Jenkins
AbstractSynchronous and xed-speed induction generators
release the kinetic energy of their rotating mass when the power
system frequency is reduced. In the case of doubly fed induction
generator (DFIG)-based wind turbines, their control system
operates to apply a restraining torque to the rotor according to a
predetermined curve with respect to the rotor speed. This control
system is not based on the power system frequency and there is
negligible contribution to the inertia of the power system. A DFIG
control system was modied to introduce inertia response to the
DFIG wind turbine. Simulations were used to show that with the
proposed control system, the DFIG wind turbine can supply con-
siderably greater kinetic energy than a xed-speed wind turbine.
Index TermsDoubly fed, xed speed, frequency response, in-
ertia, wind turbines.
I. INTRODUCTION
I
N recent years, there has been increasing interest in
doubly fed induction generator (DFIG)-based wind tur-
bines and many of the large wind turbines that are now
commercially available are of this type. Operating a large
number of DFIG-based wind turbines displaces conventional
synchronous generators and reduces the system inertia. This
is due to the fact that the DFIG control system decouples
the mechanical and electrical systems, thus preventing the
generator from responding to system frequency changes. This
is undesirable when there are a large number of DFIG wind
turbines operating, especially in periods of low load and
on smaller power systems (e.g., Great Britain and Ireland).
The frequency of a power system with low inertia will, of
course, change rapidly for abrupt variations in generation or
load. In this case, additional frequency response ancillary
services must be provided to ensure frequency limits are not
exceeded [1].
In this paper, a supplementary control loop is added to the
DFIG controller to reintroduce inertia response. Simulations
on an electromagnetic transient package (EMTDC/PDCAD)
were carried out to compare the inertia response of a
xed-speed induction generator and a DFIG with the addi-
tional control loop. It is shown that the greater variations in
rotor speed available from the DFIG wind turbine lead to
considerably more kinetic energy being released to support
the power system frequency.
Manuscript received December 9, 2003. Paper no. PESL-00146-2003.
The authors are with the University of Peradeniya, Peradeniya 20400, Sri
Lanka; Manchester Center for Electrical Energy (MCEE), University of Man-
chester Institute of Technology (UMIST), Manchester M60 1QD, U.K.; and also
with the Tyndall Center for Climate Change Research, Norwich NR4 7TJ, U.K.
Digital Object Identier 10.1109/TEC.2004.827712
II. INERTIA RESPONSE FROM FIXED-SPEED INDUCTION
GENERATOR-BASED WIND TURBINES
In the case of synchronous or induction generators, energy
is released automatically from the rotating mass for a drop in
network frequency.
For a change in frequency from to , assuming that the
speed of the rotor changes proportionally from to , the
kinetic energy released by the machine is given by
(1)
(2)
where
From (2) and substituting for , we can obtain
(3)
In order to investigate the response of a xed speed induction
generator for system frequency changes, simulations were car-
ried out using PSCADfor a 2-MW, 690-Vxed-speed induction
generator connected to an innite busbar through a 0.69/11-kV
transformer. Fig. 1 shows the response obtained when the fre-
quency of the innite busbar was reduced exponentially from50
to 49.75 Hz. For simplicity and due to the lack of data, the gener-
ator drive train was modeled as a lumped inertia of 2 MW/MVA
with a mechanical damping of 0.01 p.u. A constant wind torque
was applied.
For a 2-MW generator and a sudden frequency change from
50 to 49.75 Hz, (2) indicates the kinetic energy of the rotating
mass that will be released is 40 kW. This is conrmed by the
area under the output power curve of Fig. 1.
III. DFIG CONTROL
A full description of a DFIG control system is given in [2].
Fig. 2 shows that part of the control system, which is used to
extract maximum power from the wind.
Any change in rotor speed results in a change in the set
torque or torque demand . The torque demand is trans-
lated into a rotor current demand and compared with the actual
rotor current to obtain the rotor-injected voltage. This injected
voltage acts upon the machine and changes the speed of the wind
turbine.
If another quantity varies (other than the wind torque), the
control system acts to restore the speed back to its operating
point. This control action is shown in Fig. 3. As described
0885-8969/04$20.00 2004 IEEE
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 19, NO. 4, DECEMBER 2004 801
Fig. 1. Response of a xed-speed induction generator to a drop in network
frequency (all quantities are in per unit).
Fig. 2. DFIG control for maximum power extraction.
Fig. 3. Response of a DFIG wind turbine to a drop in network frequency (all
quantities are in per unit).
above, the DFIG control system acts against the system fre-
quency change, thus responding with a negligible inertia
response.
Fig. 4. Supplementary control loop for inertia response.
IV. DFIG CONTROLLER MODIFIED FOR INERTIA RESPONSE
By taking the derivative of the kinetic energy available at any
speed , the power that can be extracted from a rotating mass
can be obtained as
(4)
The inertia constant His dened as in the usual
notation. Substituting for in (4) and (5) we get
(5)
If and are per-unit quantities of power and speed, then
(5) can be written as
(6)
Therefore, the per-unit torque is given by
(7)
The controller shown in Fig. 2 was modied by adding a
signal corresponding to (7) to the set torque (torque demand).
As the system frequency drops, the set point torque is increased
slowing the rotor and extracting kinetic energy. In order to min-
imize the impact of this supplementary control on mechanical
drive train loads, the rate of change of power injection was mod-
ied by adding a rst-order lter after the input. Delay on
this lter was shown not only to reduce the rate of increase of the
electromagnetic (EM) torque but also reduced the magnitude of
the peak torque. The modied controller with the supplemen-
tary control loop for inertia control is shown in Fig. 4.
V. DFIG INERTIA RESPONSE
Fig. 5 shows the power generated from the DFIG wind tur-
bine, change in the rotor speed, and the EM torque of the gen-
erator for an exponential change in the system frequency from
50 to 49.75 Hz. With the additional control loop as shown in
Fig. 4, kinetic energy of 700 kW was released.
802 IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 19, NO. 4, DECEMBER 2004
Fig. 5. Response of a DFIG with supplementary control loop to a drop-in
network frequency (all quantities are in per unit).
VI. CONCLUSION
Synchronous generators and xed-speed asynchronous
generators release kinetic energy from their rotating mass for
changes in the system frequency. This is not automatically
true for converter-connected plant and converter-controlled
plant such as DFIGs. However, during drops in network system
frequency, the DFIG speed can be driven to a lower value using
its controller, thus introducing an inertia response. A supple-
mentary control loop was introduced to a DFIG controller to
obtain inertia response. Estimates shows that with the given
frequency change and with the natural choice of supplementary
controller gain (2H), a DFIG can supply 700 kW of inertia
while a xed-speed machine of similar rating can only provide
40 kW.
REFERENCES
[1] I. A. Erinmez, D. O. Bickers, G. F. Wood, and W. W. Hung, NGC expe-
rience with frequency control in England and WalesProvision of fre-
quency response by generators, in Proc. IEEE Power Eng. Soc. Winter
Meeting, New York, Jan. 31Feb. 4, 1999.
[2] J. B. Ekanayake, L. Holdsworth, and N. Jenkins, Control of doubly fed
induction generator (DFIG) wind turbine, Proc. Inst. Elect. Eng., Power
Eng., vol. 17, no. 1, pp. 2832, Feb. 2003.