Methods For Reduction of Voltage Unbalance in Weak Grids Connected To Wind Plants
Methods For Reduction of Voltage Unbalance in Weak Grids Connected To Wind Plants
Methods For Reduction of Voltage Unbalance in Weak Grids Connected To Wind Plants
A double fed induction generator with converter Fig.1. Double Fed Asynchronous Machine Model.
and control is shown in Fig.1. The main task for the
back-to-back converter control is to regulate the 2 SYMMETRICAL COMPONENTS IN
active and the reactive power of the induction THE TIME DOMAIN
machine. The converter can thus reduce the The symmetrical components are in the following
mechanical stresses on the wind turbine and the text expressed in the time domain. This is done to
fluctuations in the delivered power. The PWM express the power flow under unbalanced
converter connected to the power grid can also be conditions analytical. By introducing the operator
used for other tasks, it may be controlled as a a=ej120=1120, the sequence voltages can be
STATCOM or as an active filter to reduce written:
unbalance in the grid voltages.
v a0 (t ) 1 1 1 v a (t )
+ 1
The capacitor, the PWM inverter and the
v a (t ) = 3 1 a a 2 vb (t )
transformer may also function as a STATCOM as
indicated by the block in Fig .1. Then the back-to- v a (t ) 1 a 2 a vc (t )
back inverter may be seen as a STATCOM when Eq.1
designing the control system for handle unbalance
in the power grid. A method to measure unbalanced The sequence voltages can be transferred back to
conditions by using a synchronous rotation frame phase voltages as follows:
for each phase together with symmetrical
components theory is used. Lasseter and Hochgraf v a (t ) 1 1 1 v a0 (t )
v (t ) = 1 a 2
have developed this measurement method [1]. In
b a v a+ (t )
this project the method is implemented in a
PSCAD/EMTDC simulation of the case indicated vc (t ) 1 a a 2 v a (t )
in Fig. 1 and the resulting power flow is compared
Eq.2
to an analytical analyses of the symmetrical
components in the time domain.
Eq.1 and Eq.2 can also be used for the currents.
illustrated in Fig.1. From this it can be concluded
By using Eq.2 and the fact that each sequence that it is not necessary to compensate for the zero-
component represents a symmetrical system the sequence components.
power flow may be computed as follows:
p (t ) = i a v a + i b v b + i c v c
3 FROM AC-QUANTITIES TO DC-
( )(
= i a+ + i a + i a0 v a+ + v a + v a0 )
+ (i +
+ i b + i b0 )(v +
+ v b + v b0 ) QUANTITIES
To eliminate the need for three phase symmetry, a
b b
+ (i +
+ i
+ i 0
)(v +
+ v c + v c0 )
single-phase synchronous frame transform is used.
a c c c
p (t )
+0
= v a+ i a0 + v b+ i b0 + v c+ i c0 = 0
p (t )
+
= v a i a+ + v b i b+ + v c i c+
= 3V
(
I + cos 2 t + u + i+ )
p (t )
= v a i a + v b i b + v c i c
= 3V
(
I cos u i )
p (t )
0
= v a i a0 + v b i b0 + v c i c0 = 0
p (t )
0+
= v a0 i a+ + v b0 i b+ + v c0 i c = 0 Fig. 2. The Synchronously Rotating qd-axes for
Phase A.
p (t ) From [1], the Vq* is defined as:
0
= v a0 i a + v b0 i b + v c0 i c = 0
Vq* = v(t ) * 2 cos(t )
p (t )
Vq* = V sin (t + ) * 2 cos(t )
00
= v a0 i a0 + v b0 i b0 + v c0 i c0
= 3V 0
I 0
{cos ( 0
0
) + cos (2 t + 0
+ 0
)}
Vq* = V * 2 * 1 * [sin (t + + t ) + sin (t + t )]
u i u i
Eq.4 2
Vq = V * [sin (2t + ) + sin ( )]
*
6 CONTROL SYSTEM
Fig.4. The Power System Represented as To find the dynamical behaviour of the negative-
Sequence Systems. sequence voltage system a regulation system, as
illustrated in Fig.6, was implemented in
The STATCOM was simulated with a pulse-width PSCAD/EMTDC.
modulated converter. The switches in the converter
were replaced with voltage sources. By doing this
the simulation speed will be improved and the
harmonic distortion from the switches are
cancelled. The converter model is designed to
respond on the capacitor voltage and thus maintain
its dynamic behaviour. An illustration of the
converter model is shown in Fig.5.
Fig.6. STATCOM Negative-Sequence Voltage
Regulation.
Fig.7. The Compensation Current as a Function The amplitude gain and phase shift was found for
of the Negative-Sequence Source Voltage each permutation signal. The gain and phase shift
was plotted in a bode diagram as shown in Fig.9. It
The negative-sequence voltage control loop is can be established, from the bode plot, that the
shown in Fig.8. The measured negative-sequence amplitude gain is almost constant to the signal
voltage is compared to the negative-sequence reach 10Hz. After 10Hz the gain will begin to
reference voltage, and then the resulting error is decrease. When the permutation is in the 60Hz area
feed into a proportional controller. The ordered the positive sequence voltage will begin to oscillate
negative-sequence voltage to the STATCOM is with the negative-sequence current. The response
given from the P-controller. This negative-sequence between the negative and positive system causes
voltage is transferred together with the other the increasing gain in the 60Hz area. This means
sequence-voltages, which is not controlled, to the that the positive and negative system cannot be
ordered dq-voltages. The dq-voltages are treated as decoupled systems in the 60Hz area. Thus
transferred to phase voltages and fed into the operating in this area should be avoided. After the
STATCOM model. The STATCOM model will 60Hz area the gain will continue to decrease to it
give out the ordered phase voltages to the power reach the 120Hz area. In this area the permutation
system. The responding load voltage are measured signal have the same frequency as the noise from
and transferred to sequence voltages. the dq-transformation. This causes the disturbance
seen in the gain and phase shift plots at 120Hz.
7 RESULTS
In Fig.12 the negative-sequence components, at t1
(without compensating) and t2 (with
compensating), are illustrated as vector
components. As seen from the figure the line
current is relatively small at t1, when injecting a
Fig.10. The dc-bus Voltage Control System.
large current from the STATCOM, at t2, the line
current will be dominated by the STATCOM
The dc-bus reference voltage is compared to the
current. The impedance between the converter and
actual dc-bus voltage, the error is fed into a P-
the power system is almost inductive, due to this
controller. The Pref is given from the P-controller.
the injected current will be nearly 90 shifted to the
Pref is compared to the actual power flowing in the
load voltage. The line current, which is dominated
STATCOM. The error is fed into a lead-lag
by the STATCOM current, will then lay almost 90 seen that the dc-bus voltage will have a second
after the source voltage. The line impedance is harmonic ripple. The capacitor current shown in the
almost purely inductive and then the line voltage 6th graph causes this ripple. The current will have
will lay 90 after the negative-sequence line an average value near zero. The small average value
current. The source voltage will be the same at t1 of the current will compensate for the losses in the
and t2, thus the load voltage has to change when the capacitor.
line voltage changes. Since the line current is
almost perpendicular to the source voltage the
negative-sequence line voltage will lay parallel to
the source voltage. Thus the negative-sequence load
voltage will decrease in amplitude and be parallel to
the source voltage. The voltage drop over the line
impedance will increase if the line impedance
increases. Then the needed compensation current
from the STATCOM to decrease the negative-
sequence load voltage will be less. This means that
the needed STATCOM current to decrease the
negative-sequence load voltage depends on the line
impedance.