Académique Documents
Professionnel Documents
Culture Documents
Abstract – The paper realizes the analysis, modelling and Control of the DFIG is more complicated than the
control of the doubly-fed induction generator (DFIG) for control of a standard induction machine. In order to control
wind turbines. Different control methods are investigated the DFIG the rotor current is controlled by a power
with the objective of eliminating the influence of the back electronic converter. One common way of controlling the
electromotive force (EMF), which is that of, in control rotor current is dy means of field-oriented (vector) control.
terminology, a load disturbance, on the rotor current. Several vector control schemes for the DFIG have been
Keywords: Doubly-fed induction generator, wind turbine, proposed. One common way is to control the rotor current
wind energy, current control with stator-flux orientation or with air-gap-flux orientation
[2]. If the stator resistance can be considered small, stator-
I. INTRODUCTION flux orientation gives in principle orientation also with the
stator voltage (grid-flux orientation) [4].The stability of the
Some methods of conventional power production DFIG was investigated analitically, showing that the
shall be replaced in the future by improving the efficiency dynamics of the DFIG have poorly damped eigenvalues
of electricity use, conversion to renewable forms of energy (poles) with a corresponding natural frequency near the line
and other environmentally acceptable electricity production frequency, and, also, that the system is unstable for certain
technologies. One of the solution in this case is the wind operating conditions, at least for a stator-flux-oriented
power use. Wind turbines (WTs) can either operate at fixed system. These poorly damped poles influence the rotor
speed or variable speed. For a fixed-speed wind turbine the current dynamics through the back electromotive force
generator is directly connected to the electrical grid. For a (EMF).
variable-speed wind turbine the generator is controlled by
power electronic equipment. There are several reasons for The flux oscillations can be damped in some
using variable-speed operation of wind turbines; among different ways: reducing the bandwidth of the current
those are possibilities to reduce stresses of the mechanical controllers, introducing a flux differentiation compensation
structure, acoustic noise reduction and the possibility to or an extra converter that substitutes the Y point of the
control active and reactive power. Most of the major wind stator winding, the last two solutions being the best.
turbine manufactures are developing new larger wind
turbines in the 3-to-5-MW range [1]. These large wind The response of wind turbines to grid disturbances
turbines are all based on variable-speed operation with is an important issue, especially since the rated power of
pitch control using a direct-driven synchronous generator wind-turbine installations steadily increases. New grid
(without gear-box) or a doubly-fed induction generator codes will require WTs and wind farms to ride through
(DFIG). Today doubly-fed induction generators are voltage sags, meaning that the normal power production
commonly used by the wind turbine industry for larger should be re-initiated once the nominal grid voltage has
wind turbines. been recovered. These grid codes will influence the choise
of electrical system in future WTs, which has initiated
The major advantage of the doubly-fed induction industrial research in order to comply. Today, the DFIG
generator, which has made it popular, is that the power WT will be disconnected from the grid when large voltage
electronic equipment only has to handle a fraction (20- sags appear in the grid. The DFIG system of today, has a
30%) of the total system power. This means that the losses crowbar in the rotor circuit, which at large grid
in the power electronic equipment can be reduced in disturbances has to shortcircuit the rotor circuit in order to
comparison to power electronic equipment that has to protect the converter. This leads to that the turbine must be
handle the total system power as for a direct-driven disconnected from the grid, after a large voltage sag.
synchronous generator, apart from the cost saving of using
a smaller converter.
213
The paper is structured as follows: first are
presented the mathematical models of the DFIG system,
then the field orientation, the control of machine-side
converter and finally the control of grid-side converter,
some conclusions and references.
dΨss
v ss = R si ss + (1)
dt
dΨRs
v =R i +
s
R
s
R R − jωr ΨRs (2)
dt
Fig.2. DFIG system with a back-to-back converter
where superscript s indicates stator coordinates. The model
With the machine-side converter it is possible to can also be described in synchronous coordinates as:
control the torque and the speed of the DFIG and also the
power factor at the stator terminals, while the main dΨs
objective for the grid-side converter is to keep the dc-link vs = R sis + + jω1 Ψs (3)
voltage constant. The speed-torque characteristics of the dt
DFIG system can be seen in figure 3. As also seen in the
figure, the DFIG can operate both in motor and genrator dΨR
vR = R RiR + + jω 2 ΨR (4)
operation with a rotor-speed range of ± ∆ω around the dt
max
r
synchronous speed, ω1. A tipycal application for DFIG is
where the notations are well known.
wind turbines, since they operate in a limited speed range
of approximately ±30%.
214
The stator flux, the rotor flux and This means that the damping of the grid filter is
electromechanical torque are given by: given by:
Ψs = L M (i s + i R ) (5) 1
G f ( jω) = (11)
L2f ω 2 + R f2
ΨR = (L M + L σ )i R + L M i s = Ψs + L σ i R (6)
If Lfω » Rf, the gain can be approximated as
Te = 3n p Im[Ψs i ∗R ] (7) G f ( jω) ≈ 1 /(L f ω).
where LM is the magnetizing inductance, Lσ is the leakage C. DC-LINK MODEL
inductance and np is the number of pole pairs. Finally, the
mechanical dynamics of the induction machine are The energy, Wdc, stored in the dc-link capacitor,
described by: Cdc, is given by:
J dω r 1
= Te − Ts (8) Wdc = 2
C dc v dc (12)
n p dt 2
where J is the inertia and Ts is the shaft torque. The where vdc is the dc-link voltage. In fig. 6 an equivalent
quantities and parameters of the Γ model relate to the Park circuit of the dc-link model, where the definition of the
model (or the T representation) as follows: power flow through the grid-side converter (GSC) and the
machine-side converter (MSC) can be seen.
ir L sγ + L m
v R = γv r ; i R = ; ΨR = γΨr ; γ =
γ Lm
R R = γ 2 R r ; L σ = γL sλ + γ 2 L rλ ; L M = γL m
B. GRID-FILTER MODEL
In figure 5 the equivalent circuit of the grid filter Fig.6. DC-link model
in stator coordinates can be seen.
The energy in the dc-link capacitor is given by:
dWdc 1 d 2
= C dc v dc = − Pf − Pr (13)
dt 2 dt
This means that the dc-link voltage will vary as:
dv dc
Fig.5. Grid-filter model in stator coordinates C dc v dc = − Pf − Pr (14)
dt
The model in sychronous coordinates is:
which means that Pf = - Pr for a constant dc-link voltage.
di
E g = −(R f + jω1 L f )i f − L f f + v f (9) The total model of the DFIG system, presented in
dt fig.4 can be summarized in synchronous coordinate, as:
215
di f D1. STATOR-FLUX ORIENTATION
Lf = v f − (R f + jω1 L f )i f − E g (17)
dt For a stator-flux-oriented system the synchronous
angle θ1 is defined as:
dv dc
C dc v dc = − Pf − Pr (18)
θ1 = ∠Ψss (24)
dt
[
Te = 3n p Im Ψs i ∗R ] (22)
synchronous angle and its estimate. This means that for
perfect field orientation, i.e. θ1 = θ̂1 , that Ψs = ψ s , i.e.
Pr = 3 Re[ v i ] ∗
(23) the space vector of the flux is real valued.
R R
voltage, Eg.
jθ g
D. FIELD ORIENTATION E sg jE g e
Ψ =s
g =− (26)
In order to control the rotor current of a DFIG by jω g ωg
means of vector control, the reference frame has to be
aligned with a flux linkage. One common way is to control where ωg is the frequency of the grid voltage and θg is the
the rotor currents with stator–flux orientation or with air- corresponding angle. The grid voltage can be transformed
gap-flux orientation. If the stator resistance is considered to to synchronous coordinates as:
be small, stator-flux orientation. Pure stator-voltage
~
orientation can be done without any significant error. In ˆ
E g = E sg e − jθ1 = jE g e jθ1 (27)
this paper, the stator-voltage orientation will be referred to
as grid-flux orientation, then the machine is aligned with a
virtual grid flux. Fig.7 shows an example of the space where Eg is the grid voltage magnitude. The grid-flux
vectors of the grid voltage and the stator flux. As illustrated orientation is equal to the stator-flux orientation in the
by the figure there is only a small angular difference steady state, if the stator resistance is neglected, since then:
between the grid-voltage and stator-flux space vectors in
the stator-flux reference frame compared to the grid-flux dΨs
reference frame.
vs = E g = R sis + + jω1 Ψs ≈ jω1 Ψs (28)
dt
and ω1 = ωg.
Fig.7. Space-vector diagram of grid voltage and stator flux. E1. CURRENT CONTROL
a) Stator-flux orientation. b) Grid-flux orientation.
216
It is common to control the rotor current with 1
either stator-flux orientation or grid-flux orientation. In G ( p) = (35)
order to derive the rotor-current control law, it is pL σ + R R + R s + R a
advantageous to eliminate is and ΨR from (3) and (4),
which yields: where the controller is an ordinally PI controller, using an
internal model control (IMC) [3]:
dΨs R s
v s = −R s i R + + + jω1 Ψs (29) ki
dt L M F(p) = k p + (36)
p
di R dΨs
v R = (R R + jω 2 L σ )i R + L σ + So, k p = α c L̂ σ ; k i = α c ( R̂ R + R̂ s + R a ) (37)
dt dt
+ jω 2 Ψs = (R R + R s + jω 2 L σ )i R + (30) where αc is closed-loop bandwidth of the current dynamics,
giving:
di R
+ Lσ +E
dt p
G cl (p) = (38)
p + αc
R
E = v s − s + jω r Ψs (31)
LM
where E is the back EMF. An estimate of the whole back
EMF, Ê , will be used:
v R = v′R + ( jω
ˆ 2 L̂ σ − R a ) + k E Ê
(32)
= k p e + k i ∫ edt + ( jω
ˆ 2 L̂ σ − R a )i R + k E Ê
where "^" indicates an estimated quantity. A coefficient kE Fig.8. Block diagram of the current control system. The
is introduced in order to make the control law more dashed box is the model for the doubly-fed induction
general: generator
217
E4. CONTROL OF GRID-SIDE CONVERTER
α s −1 k
F(p) = G (p) = k p + i =
p p
(41)
Ĵα B α Fig.11. DC-link voltage control loop
= s + a s
np p
α w −1 α Ĉ α G
where F(p) = G (p) = − w dc − w a (42)
whereαs is the desired closed-loop bandwidth of the speed- p 6E g ,nom p
control loop, Ba is an "active damping" term and the
notation "^" indicates an estimated quantity. Fig. 10 shows and αw is the bandwidth of the dc-link voltage control-loop.
a block diagram of the speed control system.
− 2α w p
G PW (p) = (43)
C dc (p + α w ) 2
III. CONCLUSIONS
218