Reactive Power Control Strategy of DFIG Wind Farms

for Regulating Voltage of Power Grid

Jingjing Zhai Haoming Liu
School of Electric Power Engineering College of Energy and Electrical Engineering
Nanjing Institute of Technology Hohai University
Nanjing, China Nanjing, China
[email protected] [email protected]

Abstract—If a wind farm is weakly connected to a power grid, As mentioned, the voltage of a weakly-connected power grid
then the voltage of the connection point fluctuates frequently to wind farms is usually unstable, thus taking the reactive
due to the changeable wind speed. The active and reactive power power from DFIGs into account will make the compensation
of a doubly-fed induction generator (DFIG) can be decoupling more flexible.
controlled and the grid-side converter (GSC) of a DFIG can also
generate some reactive power by adjusting the power factor, There are mainly two reactive power control modes for a
thus a DFIG can be considered as a reactive power resource to DFIG wind turbine, one is power factor control mode and the
stabilize the voltage of the connection bus. Based on the power other is voltage control mode [4]. Many papers have focused
relationship of a DFIG, the up and down reactive power on the reactive power characteristics of DFIG wind turbine.
limitations of DFIG stator and GSC are analyzed. Then a The research contents cover the dynamic reactive power limits
reactive power control strategy of a DFIG wind farm is of doubly-fed wind turbine, the use of reactive power
proposed, in which, a certain number of DFIGs are selected to regulation capability to improve the local voltage stability, and
support reactive power to the power grid when the voltage of the the reactive power support for the grid [5]. The voltage
connection point drops. The control strategy aims at bringing stability of power system is mainly dependent on the balance
the reactive power capability of DFIG into play and cutting of reactive power, as a result, the doubly-fed wind farm should
down the investments in the reactive power compensation be equipped with corresponding voltage control schemes
devices which are used less. The simulation model of a grid-
according to the reactive power compensation situation [6]-[7].
connected DFIG wind farm is developed on the
Based on zoning plans, reference [8] proposed the automatic
PSCAD/EMTDC platform, and the simulation results
demonstrate the effectiveness of the control strategy proposed. voltage control strategy of wind farm, which combined the
DFIG voltage control with other voltage regulation measures
Index Terms--Doubly-fed induction generator (DFIG); reactive such as switching capacitors. According to the optimal
power control; voltage control; power limitation. secondary voltage control theory, reference [9] suggested that
the reactive power regulation capability would be better
played when the key node voltage was controlled. These
I. INTRODUCTION
papers take all the wind turbines in the wind farm into account
Nowadays more and more wind farms are connected into when allocating the reactive power, which may not only
the power grid. Long-distance transmission lines are usually increases the power loss of wind turbines but also makes the
required for the connection because most wind farms are dispatch and control of the total wind farm more difficult.
located at the remote areas or off shores. The output active
power of wind farms is variable and intermittent due to the This paper proposes a more flexible reactive power control
changeable wind speed, which threats the voltage stability of strategy. The up and down limitations on the reactive power of
local power grids. More reactive power is demanded to a DFIG are deduced based on the power relationship. When
maintain the voltage when it drops. Since a wind farm is the voltage of the connection bus drops due to some load
composed of many wind turbines, the operation and control disturbances in the power grid, a certain number of DFIG
modes of the wind turbines have a great impact on the voltage wind turbines are selected according to the up and down
stability of local power grids. The doubly-fed induction limitations to inject the demanded reactive power into the grid,
generator (DFIG) is widely used in wind farms because it has helping to recover the voltage. Not all the wind turbines are
many advantages, one of which is the decoupled control of involved in this control strategy, the number is dependent on
active and reactive power [1]. Besides, both the stator and the the reactive power needed by the power grid and is determined
grid-side converter (GSC) of a DFIG can inject reactive power dynamically. Finally, the proposed strategy is verified by
into the grid to help to maintain the fluctuant voltage [2][3]. simulations on the PSCAD/EMTDC platform.

978-1-4799-6415-4/14/$31.00 ©2014 IEEE

 II. REACTIVE POWER CHARACTERISTIC OF A DFIG WIND For a low wind speed, the grid-side converter does not
TURBINE make full use of its capacity. When the grid requires extra
reactive power, the grid-side converter can be adjusted into a
A. Power Relationship of a DFIG Wind Turbine non-unity power factor to meet the requirement. Assuming the
The topology of a DFIG is shown in Fig. 1. The stator of a maximum capacity of the grid-side converter is Sc max , there is
DFIG is connected to the power grid directly, while the rotor Pc2 + Qc2 ≤ Sc2max . The reactive power that the grid-side
is connected to the grid through two back-to-back pulse width
modulation (PWM) converters, i.e. rotor-side converter and converter can generate or absorb is presented as:
grid-side converter. The grid-side converter usually works at
the unity power factor of 1 and is in charge of maintaining a − Sc2max − Pc2 ≤ Qc ≤ Sc2max − Pc2 (3)
constant DC-link voltage for the rotor-side converter. The
decoupling control of active and reactive power of the DFIG is Combining (2) and (3), the reactive power limitations of
achieved by adjusting the rotor’s current and voltage through the grid-side converter can be illustrated as:
the rotor-side converter.
⎧Q
⎪ c max = Sc max − ( sPs )
2 2

Ps , Qs Pg , Qg ⎨ (4)
PT
⎪⎩Qc min = − Sc2max − ( sPs ) 2
Pc , Qc
C. Reactive Power Limitations of the DFIG Stator
Based on the orientation of the grid voltage vector, the
Pr , Qr rotor current can be expressed as [10]:

Figure 1. Topology of a DFIG ⎧ Ls

⎪ird = L isd
In Fig. 1, PT is the input mechanical power from the wind ⎪ m
⎨ (5)
turbine, Ps and Qs are the output active and reactive power ⎪i = Ls i − us
generated by the DFIG, Pc and Qc are the active and reactive ⎪⎩ rq Lm sq Lm w
power to the grid-side converter from the DFIG, Pg and Qg
where ird and irq are the d- and q-axis components of the
are the active and reactive power to the grid from the DFIG, Pr
and Qr are the active and reactive power to the rotor winding rotor current, isd and isq are the d- and q-axis components of
from the rotor-side converter. Assuming that the input the stator current, Ls is the stator inductance, Lm is the
mechanical power is completely converted into the electro- mutual inductance of stator and rotor, us is the stator voltage,
magnetic power, the power relationship of DFIG can be
presented as: and ω is the synchronous angular velocity.

PT = Ps − Pr (1) It is well known that ird2 + irq2 = ir2 , so there is

2 2
The difference between the stator speed and the ⎛ Ls ⎞ ⎛ Ls us ⎞
⎜ isd ⎟ + ⎜ isq − ⎟ = ir ≤ I r max
2 2
synchronous speed results in the slip power of the rotor, which (6)
is named as the rotor power Pr . Therefore the power L
⎝ m ⎠ ⎝ m L Lm w ⎠
relationship can also be expressed as:
where I r max is the maximum current of the rotor-side
⎧ Ps = PT (1 − s ) converter.
⎨ (2)
⎩ Pr = sPT (1 − s )
Similarly, the stator current can be written as:

isd2 + isq2 = is2 ≤ I s2max (7)

B. Reactive Power Limitations of the Grid-side Converter
The rotor-side and grid-side converters only transfer active where I s max is the maximum current of the stator.
power, while the reactive power Qc and Qr are decoupled. If
the power loss is neglected, there is Pc = Pr . Since the grid- The active and reactive power of DFIG stator can be
presented as [10]:
side converter usually works at the unity power factor of 1,
there is Qc = 0 . However, Qr is divided into two parts, one ⎧ 3
part flows into the rotor, and the other part is transferred to the ⎪⎪ Ps = 2 us isd
stator by a certain percentage (slip ratio). Therefore, the ⎨ (8)
capacity of the rotor-side converter is larger than that of the
grid-side converter. As a result, when analyzing the reactive
⎪Q = 3 u i
power limitations of the grid-side converter, only the capacity
⎪⎩ s 2 s sq
of itself is considered.
 Then, of common coupling (PCC) as the reference point. The
voltage or power factor at PCC is detected and is compared
⎧ 2 Ps with the reference to calculate the reactive power demanded
⎪isd = 3u by the grid. Based on the demanded reactive power and the
⎪ s reactive power limitations of each DFIG wind turbine, the
⎨ (9)
number of DFIGs that take part in the reactive power support
⎪i = 2Qs is determined. The control mode of these DFIGs is
⎪⎩ sq 3us consequently changed from unity power factor control to
voltage control.
Combine (9) with (6) and (7), there are
The overall reactive power control strategy is shown in Fig.
⎧ ⎛
2
3us2 ⎞ ⎛ 3us Lm I r max ⎞
2
3, where U N is the rated voltage, U pcc is the measured
⎪ Ps + ⎜ Qs +
2
⎟ ≤⎜ ⎟ voltage, i.e., the voltage at PCC, ΔU % is the voltage
⎪ ⎝ 2 wLs ⎠ ⎝ 2 Ls ⎠ deviation, Qref is the demanded reactive power by the grid,
⎨ 2
(10)
⎪ 2 ⎛ 3us I s max ⎞
i
and Qref is the reactive power allocated to unit i.
⎪ Ps + Qs ≤ ⎜
2
⎟
⎩ ⎝ 2 ⎠
1
Qref
According to (10), the reactive power limitations of DFIG
stator are shown in Fig. 2, where the thick solid line is the
boundary values of reactive power.
U pcc
n
Qref
Ps

Qref ΔU U N − U pcc UN
ΔU % = *100%
UN

Figure 3. Reactive power control strategy

Qs min − 3u 2 wLs
2
Qs max Qs
s The control process is illustrated in detail as follows:

Figure 2. Reactive power limitations of DFIG stator 1) Monitor the real-time active power of each DFIG wind
turbine, and calculate the reactive power limitations of each
turbine according to (11). The reactive power limitation of the
D. Reactive Power Limitations of the DFIG n
In order to establish the magnetic field, the DFIG stator wind farm is calculated by Qwf max = ∑ Qig max .
needs to absorb reactive power from the grid, which leads to i =1

the asymmetric reactive power limitations in Fig. 2. That is to 2) Detect the voltage at PCC U pcc and compare it with the
say, the maximum capacitive reactive power of DFIG stator is
less than the maximum inductive reactive power. Assuming rated voltage U N . The reactive power demanded by the grid
the DFIG is equipped with a bank of capacitors with the Qref is obtained based on the voltage deviation ΔU % .
capacitance C = 3 2 wLs and considering them as a whole, the According to the reactive power limitations of each wind
reactive power limitations of DFIG will be shifted to the right turbine, choose the minimum number of units and allocate the
direction and a symmetrical one can be obtained as: demanded reactive power Qref to these units evenly. If
⎧ Qwf max < Qref , then all the units will take part in the voltage
3us Lm I r max 2
Q
⎪ g max = ( ) − Ps
2
+ S 2
c max − ( sPs ) 2
regulation, and the pitch angle of each wind turbine should be
⎪ 2 Ls adjusted to decrease the active power and enlarge the reactive
⎨ (11) power limitations until Q’wfmax≥Qref is satisfied.
⎪Q 3us Lm I r max 2
⎪ g min = − ( 2 L ) − Ps2 − Sc2max − ( sPs ) 2 i
3) The reactive power allocated to unit i Qref is allocated
⎩ s
again between each unit’s stator and grid-side converter, and
i
III. REACTIVE POWER CONTROL STRATEGY OF DFIG the stator has a priority. That means if Qref is in the thick-
i
WIND FARMS solid-line region shown in Fig. 2, the stator will support Qref
Since a DFIG wind turbine has the capability of absorbing i
alone, if not, Qref will be generated by the stator together with
and generating reactive power, a DFIG wind farm thus can not
the grid-side converter.
only generate active power, but also generate or absorb the
reactive power to stabilize the grid voltage. The reactive 4) The control modes of the involved units are switched to
power is mainly controlled in two modes, i.e. power factor generate the reactive power and maintain the voltage. When
control and voltage control. Both the two modes take the point finished, return to step 1).
 IV. CASE STUDY Case 2: Assuming that the wind speed is always 13 m/s,
An example case is implemented on the PSCAD/EMTDC and a 50 Mvar reactive load is connected to the 6 kV bus from
platform, as shown in Fig. 4. A 100MW wind farm is 2 s to 4 s. In this case, the wind farm works at the full power
connected to a 220 kV power grid through overhead lines after and can hardly generate any reactive power due to the rated
two-stage boosts. The wind farm is simplified to have only power limitation. The simulation results are shown in Fig. 6. It
three wind turbines G1, G2 and G3, their capacities are 30 can be seen that if the wind farm works at the unity power
MW, 30 MW and 40 MW, respectively. Capacities of the factor of 1, the voltage at PCC drops severely to 0.955 p.u.
rotor-side and grid-side converters are both 30% of each unit’s when Load2 changes. If the proposed control strategy is
total capacity. Two loads, Load1 and Load2, are connected to applied, the three units G1, G2 and G3 are all picked on since
the 10 kV and 6 kV buses. the demanded reactive power is bigger than the reactive power
limitation of the wind farm. The pitch angles of the three units
are all adjusted from 0° to 3.5°, sacrificing a certain active
power of the wind farm, that is from 1 p.u. to 0.914 p.u., and
enlarging the reactive power limitation of the wind farm.
During the voltage drop, the wind farm provides 0.4 p.u.
reactive power to the grid, helping the voltage return to normal.

voltage p.u
Figure 4. An example system for simulation

To verify the reactive power control strategy proposed,

two cases that demonstrate small and big reactive power
disturbances are simulated.
(a) Grid voltage
Case1: Assuming that the wind speed is always 10 m/s,
and the 80 MVA Load1 at the 10 kV bus increases by 50%
from 2 s to 4 s. The simulation results are shown in Fig. 5. It
can be seen that if the wind farm works at the unity power
factor of 1, the voltage at PCC drops to 0.984 p.u. when Load1
changes. If the proposed control strategy is applied, G1 will be
picked on and then switched to the voltage control mode.
Reactive power of 0.24 p.u. is injected to the power grid by
G1, therefore, the voltage at PCC hardly changes during the
load disturbance. The other two units G2 and G3 always work (b) Output reactive power of the wind farm
at the unity power factor of 1 from beginning to end. Figure 6. Reactive power control during a large load disturbance

In Case 2, the wind farm works at the full power. Once a

large load disturbance occurs, dynamic reactive power
compensation devices such as SVG should have been installed
to provide reactive power in real time. However, in Case 2,
thanks to the reactive power control strategy proposed, the
wind farm just sacrifices a certain active power to provide
reactive power, no dynamic reactive power compensation
device is required, which is more economical.

(a) Grid voltage V. CONCLUSIONS

This paper analyzes the reactive power limitations of a
DFIG wind turbine quantitatively and proposes a more
flexible reactive power control strategy of a grid-connected
DFIG wind farm. Simulation results show that the control
strategy proposed can wonderfully suppress the voltage
fluctuations at PCC caused by load changes. This control
strategy intends to pick on as few DFIG wind turbines as
possible to take part in the voltage regulation and generate the
demanded reactive power by the grid. If the maximum
reactive power generated by the wind farm still cannot satisfy
(b) Output power of wind turbines the grid’s demand, then the wind farm will sacrifice a certain
Figure 5. Reactive power control during small load disturbance
active power to generate more reactive power. The strategy
can be applied when load disturbances occur. Short-circuit
faults are not taken into account since fault ride-through Systems Research, 2009, 7(9): 1705-1716.
control does not belong to the research scope of this paper. [5] Kayikci M, Milanovic J V. Reactive power control strategies for DFIG-
based plants [J]. IEEE Transactions on Energy Conversion, 2007, 22(2):
The strategy make wind farms involved in the voltage 389-396.
regulation, thus provides more choices for the grid in real-time [6] Fang J, Li G, Liang X, et al. An optimal control strategy for reactive
scheduling. Dynamic grouping of units and optimal allocation power in wind farms consisting of VSCF DFIG wind turbine generator
of reactive power among units need further study. systems [C] // Electric Utility Deregulation and Restructuring and
Power Technologies (DRPT), 2011 4th International Conference on.
IEEE, 2011: 1709-1715.
REFERENCES [7] Amaris H, Alonso M. Coordinated reactive power management in
[1] Fernandez L M, Garcia C A, Jurado F. Comparative study on the power networks with wind turbines and FACTS devices [J]. Energy
performance of control systems for doubly fed induction generator Conversion and Management, 2011, 52(7): 2575-2586.
(DFIG) wind turbines operating with power regulation [J]. Energy, [8] Qiao Ying, Lu Zongxiang, Xu Fei. Coordinative strategy for automatic
2008, 5(6): 1438-1452. voltage control of wind farms with doubly-fed induction generators [J].
[2] Tapia G, Tapia A, Ostolaza J X. Proportional-integral regulator-based Automation of Electric Power Systems, 2010, 34(5): 96-100. (in
approach to wind farm reactive power management for secondary Chinese).
voltage control [J]. IEEE Transactions on Energy Conversion, 2007, [9] Ko H S, Bruey S, Jatskevich J, et al. A pi control of DFIG-based wind
22(2): 488-498. farm for voltage regulation at remote location [C] // Power Engineering
[3] El Moursi M, Joos G, Abbey C. A secondary voltage control strategy for Society General Meeting, 2007. IEEE, 2007: 1-6.
transmission level interconnection of wind generation [J]. IEEE [10] Rezaei E, Tabesh A, Ebrahimi M. Dynamic Model and Control of DFIG
Transactions on Power Electronics, 2008, 23(3): 1178-1190. Wind Energy Systems Based on Power Transfer Matrix [J]. IEEE
[4] El Moursi M, Joos G. Optimal tracking secondary voltage control for Transactions on Power Delivery, 2012 ,vol.27, no.3, pp:1485-1493.
the DFIG wind turbines and compensator devices [J]. Electric Power