Content uploaded by Hongchao Gao
Author content
All content in this area was uploaded by Hongchao Gao on Aug 18, 2016
Content may be subject to copyright.
Coordinated Reactive Power Control of OLTC and WTGs for Improved
Steady State Voltage Profile
Gao, Hong Chao Ahn, Seon-Ju Choi, Joon-Ho
Department of Electrical Engineering Department of Electrical Engineering Department of Electrical Engineering
Chonnam National University Chonnam National University Chonnam National University
gaohrb@qq.com sjahn@jnu.ac.kr joono@chonnam.ac.kr
Abstract
With the penetration of the renewable energy of wind
farm increasing rapidly in the past decades, reactive
power control of wind farms has become a crucial
issue. This paper proposes a quadratic programming
based optimization method of reactive power
distribution between wind turbines to improve the
steady state voltage profile while meet the point of
common coupling (PCC) reactive power demand. In
addition, a comprehensive modelling of wind farm
and power flow analysis is presented. Comparison
between the conventional average distribution
principle and optimal distribution is analyzed. With
the simulation, it is shown that with optimum of
reactive power distribution between wind turbine
generators (WTGs), voltage profile can be improved
meanwhile the number of on load tap changes (OLTC)
can be decreased.
Keywords
Wind farm, power flow, reactive power control,
OLTC, quadratic programming
1. INTRODUCTION
Due to the deterioration of the global climate, new
renewable energy in the energy market is gradually
showing advantage. The global wind energy market
was worth $130 billion in 2013 and $165.5 billion in
2014. Therefore, many related to wind farm grid
issues become the focus of the study. However, the
issue of reactive power control of wind farm is
observed and proposed by many researchers.
With the increasing integration of wind power plants,
grid utilities require extended reactive power supply
capability, not only during voltage dips, but also in
steady state operation [Hoan V. Pham et al, 2014].
The transmission system operators (TSOs) in
different countries commonly define the grid code
control requirements for WF. The reactive power
requirements can formulated in terms of power factor
or reactive power reference at the PCC. Grid code
compliance for distribution and transmission
connection is an important consideration in wind
farm construction and voltage/reactive power control
is a necessary element in achieving such compliance.
When it comes to the reactive power control of wind
farm, many relevant issues worth being taken into
consideration, such as, power factor requirement,
dynamic voltage support requirement, low/high
voltage ride through and reactive power
compensation devices.
In addition, according to the different research topic,
the model of wind farm adopted is also different. For
example, wind farm generators are often modeled as
one equivalent generator driven by a single
equivalent wind turbine, when the effect of wind
farm transmission system is studied: 1) wind farm to
damp the sub-synchronous oscillation of a power
system; 2) secondary voltage control combined with
wind farms and synchronous generators. And wind
farm is often represented by an exact number of wind
turbine generators, when the topic of wind farm
controller design is studied: 1) the coordination of
wind farm and static VAR compensator
(STATCOM); 2) allocation of reactive power of wind
farm generators to satisfy the PCC requirements.
There are a lot of distribution principles of the
allocation of reactive power between WTGs, such as
even distribution, proportional distribution and so on.
However, because of the wind distribution, wake
effect and other effects, the wind turbines in the same
wind farm does not own same operating state at the
same time. So simply adopting the even distribution
principle or the proportional sharing principles is not
the optimal use of reactive power supplied by the
wind generators. Based on this consideration,
reactive power optimal allocation will be presented in
detail in this paper. The problem of reactive power
allocation between WTGs is formulated as an
optimization problem subject to restrictions. The
principal objective function is in order to make the
terminal voltage of each WTG most effectively close
to the wind farm average voltage value through the
allocation and control of reactive power between
each WTG.
In recent years, an ever-increasing research attention
has been paid to the solution of the optimal reactive
power dispatch problem (ORPDP) based on the
application of a variety of heuristic optimization
algorithms such as genetic algorithm [I. J. Fang et al,
2011], particle swarm optimization [V. S. Pappala,
2010], evolutionary programming [Q. H. Wu et al,
1995] and differential evolution [M. Varadarajan et al,
2008]. However, these technologies easily suffer
from the partial stagnation or premature convergence
and genetic operators have to be carefully selected or
developed.
The objective of this paper is to introduce a quadratic
programming-based controller for optimal control of
reactive power allocation between WTGs, which the
voltage profile is improved and the number of on-
load tap changes of tap-changing transformer is
effectively decreased. This paper is organized as
follows. A detailed wind farm model is presented in
section 2. In section 3, the proposed optimization
method of allocation of reactive power between
WTGs and its formulation is introduced. Simulation
results and analysis are demonstrated in section 4.
And the conclusion is given in Section 5.
2. WIND FARM MODEL
2.1 Wind farm configuration
To avoid wake effects, the wind farm can be made
longer and more stretched out. Choice of wind
configuration should depend on wind data and
meteorological environment. However, this is not
taken into consideration in this paper. Therefore, the
configuration is selected based on the radial layout as
shown in Figure 1. Recommended spacing is 5-9
rotor diameters separating towers within a row and 3-
5 diameters between rows [Gilbert M. Masters, 2004].
Six diameters are used here. A higher number of
turbines mean longer cables and higher loss,
therefore, each radial connects eight wind turbines
which has the rated power of 3.6 MW and rotor
diameter 104m. Besides that, an offshore platform is
necessary for the transformer to step up the voltage
for high voltage transmission.
33kV XPLE 400
33kV XPLE 400
33kV XPLE 400
33kV XPLE 400
0.8427
km
12.56 km
7.41 km
2.26 km
Onshore
Bus 43
XLPE 150 kV
20 km
Bus 10
Bus 17 Bus 25 Bus 33 Bus 41
Bus 18 Bus 26 Bus 34
PLATFORM
TRANSFORMER
33/150
Bus 1
Bus 42
Grid Bus 44
Slack bus
33kV XPLE 400
Column1~column8
Row1~row5
33 kV XPLE 95
33kV XPLE 95
33 kV XPLE120 33 kV XPLE 95
33 kV XPLE 240
33 kV XPLE 240 33 kV XPLE 120
Bus 2
Bus 9
33 kV XPLE 95
33kV XPLE 95
33 kV XPLE120 33 kV XPLE 95
33 kV XPLE 240
33 kV XPLE 240 33 kV XPLE 120
33 kV XPLE 95
33kV XPLE 95
33 kV XPLE120 33 kV XPLE95
33 kV XPLE 240
33 kV XPLE 240 33 kV XPLE 120
33 kV XPLE 95
33kV XPLE 95
33 kV XPLE120 33 kV XPLE 95
33 kV XPLE 240
33 kV XPLE 240 33 kV XPLE 120
33 kV XPLE 95
33kV XPLE 95
33 kV XPLE120 33 kV XPLE 95
33 kV XPLE 240
33 kV XPLE 240 33 kV XPLE 120
Fig. 1 Wind Farm Configuration
2.2 Size of cables
Larger conductor cross-section gives less losses and
the power rating is higher. In this study, the
maximum generated by each radial is 28.8 MW and
WTGs absorb or produce maximum 13.92 MVar.
Therefore, 33 kV XPLE submarine cables [Randi
Aardal Flo, 2009] delivered by ABB are adopted here.
Three different sizes of cables are divided in each
radial based on different power rating. The 400
cable are connected between the platform and the
first wind turbine. The following two cables from the
first WT must have a cross section of 240 and
120 cables are adopted for the wind turbines
from the third one to sixth one. The last three cables
adopt the 95 cables. The cable data for the
cables between WTs in each radial and the cables
between the offshore platform and the radials are
given as shown in Table 1 and Table 2.
Cross-
section
Power
rating
[MW]
R
[ohm]
L
[ohm]
C
[
95
15.0
0.20225
0.11643
0.13483
120
18.6
0.16854
0.10857
0.15169
240
29.3
0.08427
0.09796
0.19383
Table 1 Cable data between the wind turbines
Bus
number
Cable
length
[km]
R
[ohm]
L
[ohm]
C
[
1-2/1- 34
12.56
0.7536
1.38104
3.5168
1-10/1-26
7.41
0.4460
0.81477
2.0748
1-18
2.26
0.1356
0.24845
0.6328
Table 2 Cable data between the platform and radial
2.3 Transmission system and transformer
Between 2002 and 2009, many famous offshore wind
power projects advanced to commissioning, such as
Horns Rev of Denmark, North Hoyle of the UK,
according to the different wind farm sizes and
offshore distances, 20 km submarine high voltage
alternative current (HVAC) transmission cable is
decided in this study. The cable parameters of 150
kV XPLE submarine is given in appendix. The size
of wind farm we proposed is 144 MW, therefore, a
160 MVA tap-changing transformer
(150/33 ) is installed at
offshore platform to set up the voltage for
transmission and regulate voltage.
Besides that, the high capacitance of submarine cable
results in the need of reactive compensation. An
estimate of the reactive power produced by the 150
kV HVAC can be taken into consideration according
to equation (1) [M. J. da R. B. Marques, 2010].
Compensation only at the onshore end is possible, but
adding compensation both end can greatly improve
the transmission distance as well as the current
profile along the HVAC link, and as a result,
transmission loss can be reduced. Thus,
approximately 60% of reactive power is compensated
at both the onshore and offshore.
(1)
2.4 grid impedance
The estimation of grid impedance has an effect on the
stability of the system. The grid impedance can be
calculated as equation (2) and (3) [Stefan Lundberg,
2006].
(2)
(3)
Where, is the line-to-line voltage of the grid;
is the XR-ratio of the grid and is the short
circuit ratio (the short circuit power of the grid divide
by the rated power of the wind farm).
3. PROPOSED CONTROL STRATEGY
3.1 Proposed optimization method
The purpose of the reactive power control is to adjust
the reactive power generated to meet the reactive
reference at the PCC, as shown in Figure 2. As the
previous presentation, there are many distribution
principles of reactive power between WTGs, such as
even distribution, proportional distribution. However,
because of the wind distribution, wake effect and
other effects, the WTs in the same wind farm may
have different operating state. Voltage profile of
WTG buses is mainly dependent on the P and Q of
each WTG, thus steady state voltage profile of WTGs
can be improved by optimal reactive allocation which
can reduce the number of switching operations of
OLTC and/or shunt capacitor/reactors. Power loss
can be reduced by optimizing the reactive power
distribution between WTs, but the effect of reactive
power allocation is small.
Grid Bus 44
Slack bus
Onshore
bus 43
PCC
PCC
V
*
PCC
V
PI
controller Distribution
principle
giref
Q
giref
Q
giref
Q
giref
Q
giref
Q
Fig. 1 The proposed reactive power control
With the purpose of improving the voltage profile,
the optimization problem is formulated as equation
(4), which is used to make the terminal voltage of
each WTG most effectively closed to the wind farm
average voltage value through the allocation and
control of reactive power between each WTG.
Equations (5)-(7) are constrains.
(4)
Subject to:
(5)
(6)
(7)
Where, is the voltage of bus , is the average
value of the voltages of all the WTG bus, and
is
the reactive power reference value of WF.
3.2 Quadratic programming and formulation
The general format of quadratic programming is as
equation (8).
(8)
Which subject to:
(9)
(10)
(11)
, is the matrix and vector in linear inequality
constraints respectively; is the matrix and
vector in linear equality constraints severally; and
is used to present lower bounds and upper
bounds of variable. Then words about algorithm
choosing are that ‘trust-region reflective’ handles
problems with only bounds or only linear equality
constraints, however, ‘interior-point-convex’ handles
only convex problems. Thus, ‘active-set’ algorithm is
chosen here.
Therefore, the previous problem is solved by
formulation of parameters which is necessary. The
effort is presented as shown in equations (12) - (17).
(12)
(13)
(14)
Where is the sensitivity between bus voltage and
reactive power which can be obtained through the
inverse Jacobian of power flow.
And the inequality constrains:
(15)
(16)
Finally, the equality constraints:
(17)
(18)
Besides that, the flow chart of control strategy based
on the optimal reactive power allocation principle is
shown in Fig. 3.
NO
YES
Update active power Pg_input, reactive power
Qg_input and reactive power reference Q_ref
At the PCC.
Update busdata and linedata.
Solution is best?
Reset
tap position
YES
0.95<Vm<1.05?
NO
Initialization
Power flow calculation
Solve the problem of reactive power allocation
with optimization method
Update busdata with the optimal
reactive power allocation
Fig. 3 The flow chart of proposed control strategy
with optimization of reactive power between WTGs
4. TEST RESULTS
According to the previous discussion and analysis,
two cases will be tested for comparison purpose as
follows. And the reactive power reference at PCC
which will be changed every 75 min as shown in Fig.
4.
Case 1: Even distribution principle for the reactive
power allocation is tested, and the controller
calculates every 15 min.
Case 2: Optimal reactive power allocation between
WTGs is tested, and the controller calculates the
optimal settings every 15 min.
Fig. 4 The reactive power reference at the PCC
Fig. 5 Voltage profile with even distribution principle
Fig. 6 Voltage profile with proposed principle
Fig. 5 shows the voltage profile of each WTG which
is simulated based on case 1. It shows that the
voltages are kept between 0.95 pu. and 1.05 pu. with
the control of tap-changing transformer. Fig. 6 shows
the voltage profile of each WTG which is tested
following case 2. It obviously presents that the
voltages of WTGs with the optimal reactive power
distribution principle are closer to the average value
comparing to the ones with even reactive power
distribution principle.
Every 15 minutes, the reactive power will be supplied
by the WTGs should be allocated between WTGs
following reactive power reference at the PCC. The
voltages obtained through power flow calculation
1
1.01
1.02
1.03
1.04
1.05
1.06
0 5 10
voltage per unit
wind turbine number
voltage profile row 1
row2
row3
1
1.01
1.02
1.03
1.04
1.05
1.06
0 5 10
voltage per unit
wind turbine number
voltage profile row1
row2
row3
will be checked whether they are within the range or
not.
Fig. 7 Tap position of the OLTC at PCC
If any voltage is out of the limits, tap-changing
transformer will change its tap position to reduce or
increase the voltage until both the maximum voltage
and minimum voltage being reached allowed range.
During 24h, tap positions are observed and the results
are given in Fig. 7. The blue line and red line shows
the tap position change based on case 1and case 2
respectively and the test points have been marked by
asterisks and circles. According to the test results, tap
position is changed 73 times with even reactive
power allocation during one day, however, if the
optimization of reactive power allocation is
implemented, the operation number of tap-changing
transformer which is used to regulate the voltages of
WTGs will be reduced to 58 times. It means that
proposed optimal reactive power allocation can
improve the voltage profile meanwhile the operation
number of OLTC can be reduced.
5. CONCLUSION
In this paper, an optimization method based on
quadratic programming for reactive power allocation
between WTGs is successfully implemented. In order
to do this analysis, a detailed wind farm model is
presented. This control strategy which coordinates
the tap-changing transformer guarantees not only the
meet of reactive power requirement at the PCC, but
also the improvement of the voltage profile of WTGs.
Test results demonstrate that based on the optimal
allocation between WTGs, the tap-changing
transformer effectively regulate the voltage of WTGs
within the limits meanwhile the number of switching
operation is reduced compared to the conventional
even-distribution control.
Future research work is aim to take the wake
effect into consideration, through the forecast of the
wind speed and distribution, more efficient reactive
power allocation distribution is explored in order to
avoid unnecessary tap changes.
Appendix
GE 3.6MW wind generator:
Generator rating=4 MVA, ,
, , ,
hub height=73.5m, rotor diameter=104m.
150 kV XPLE submarine:
R=0.039, L=0.120, C=0.190.
Acknowledgement
This work was supported by the National Research
Foundation of Korea (NRF) grant funded by the
Korea government (MSIP) (No. 2010-0028509).
References
Gilbert M. Masters, Renewable and Efficient Electric
Power System, 2004: p351- 354.
Hoan V. Pham, 2014, Online Optimal Control of
Reactive Sources in Wind Power Plants, IEEE
Transactions on Sustainable Energy, IEEE, 2014:
p608-616.
I.J. Fang, G. Li, X. Liang, and M. Martinez, 2011, An
Optimal Control Strategy for Reactive Power in
Wind Farms Consisting of VSCF DFIG wind turbine
generator systems, Electric Utility Deregulation and
Restructuring and Power Technologies, 2011: p1709-
1715.
M. Varadarajan, K. S. Swarup, Network Loss
Minimization with Voltage Security Using
Differential Evolution, Electric power system, 2008:
p815-823.
Miguel Jorge da Rocha Barros Marques, 2010,
Steady State Analysis of the Interconnection of
Offshore Energy Parks, master thesis, 2010: p 63-64.
Q. H. Wu, J. T. Ma, 1995, Power System Optimal
Reactive Power Dispatch Using Evolutionary
Programming, IEEE trans. Power System, IEEE,
1995:p1243-1249.
Randi Aardal Flo, Configuration of large offshore
wind farms, master thesis, 2009: p46-47.
Stefan Lundberg, 2006, Wind Farm Configuration
and Energy Efficiency Studies, Ph. D. thesis, 2006:
43-44.
V. S. Pappala, W. Nakawiro, and I. Erlich, 2010,
Predictive Optimal Control of Wind Farm Reactive
Sources, IEEE PES Transmission and Distribution
Conference, IEEE, 2010: p1-7.
