Journal of International Council on Electrical Engineering Vol. 3, No. 4, pp.

340~347, 2013
http://dx.doi.org/10.5370/JICEE.2013.3.4.340

Effective Voltage Control by SVR to Reduce the Capacity of SVC

using Solar Radiation Information with Real Time Simulator
Shinya Sekizaki, Mutsumi Aoki*, Hiroyuki Ukai*, Shunsuke Sasaki**,
Takaya Shigetou**, Weihua Wang*** and Jean Blanger***
Abstract With the increasing the number of Photovoltaic generators (PV) connected to distribution
system (DS), several concerns such as rise and sudden change of voltage on distribution line are growing
in Japan. Step Voltage Regulator (SVR) is well known as the one of voltage control equipment used in
current DS. Meanwhile, SVR cannot regulate rapid voltage change because SVR has time delay against
variation of voltage. In contrast, Static Var Compensator (SVC) is the effective device to control voltage
changed rapidly. However, since the cost of SVC with large capacity is expensive, it is important to
reduce the capacity of SVC in order to increase the introduction of SVC into power system. From this
background, the novel control method of SVR using solar radiation information to reduce the capacity of
SVC is proposed in this paper. The effectiveness of the proposed method is confirmed by numerical
simulation with real time simulator.
Keywords: PV, SVR, SVC, Real-time simulation, Hardware-in-the-loop

voltage deviation occurs temporarily when voltage on

distribution line varies faster than changing tap position due
to PV. Therefore, it is indispensable to improve control
method of SVR in the future so that SVR can manage
voltage properly in DS with many PVs.
On the other hand, in order to control voltage fluctuated
by PV, SVC is effective device for voltage regulation that
injects reactive power into DS. However, since SVC is
expensive, the introduction of many SVCs with large
capacity into power system is difficult. From the point of
view of cost, it is necessary to reduce the capacity of SVC
for introduction into DS. Hence, authors focused on SVR
and attempted to make SVR control voltage changed
rapidly. If SVR is employed for voltage regulation; however,
voltage fluctuation caused by PV must be controlled even
though SVR has time delay. Accordingly, new control
method of SVR which can manage voltage in case of many
PVs is proposed in this paper. By using proposed method,
since not only SVC but SVR also has function to
compensate rapid voltage fluctuation, reactive power output
of SVC can be reduced. This means that it is possible to
reduce the capacity of SVC. In order to simulate the
behavior of controller in actual DS, authors use RT-LAB
which is real time simulator and implement proposed
controller into DSP board. Authors confirmed that proposed
controller of SVR suppresses voltage deviation and reduces
the capacity of SVC rather than conventional control
method of SVR.

1. Introduction
Recently, the number of PVs connected to power system
is increasing in Japan to decrease the greenhouse gas
emissions. In particular, Japanese government focuses on
roof-top type PV; therefore, the government mainly plans to
introduce PVs into residential housing in DS. However,
since reverse power flow from PV into DS causes the rise
and sudden change of voltage on distribution line in
accordance with weather condition, it becomes more
important to maintain voltage stable within adequate range.
On current DS in Japan, some devices to control voltage
including SVR have been already installed. Since voltage
management using SVR is low-cost and simple, SVR is
generally employed. SVR is one of the transformers which
has several taps and can control its secondary voltage by
changing tap position. By contrast, SVR does not assume
the case in which a lot of PVs are introduced and cannot
control voltage changed rapidly by PV. This is because that
SVR has mechanical components such as tap changer and
needs several seconds to change tap position. From this,

Corresponding Author: Power system laboratory, Nagoya Institute of

Technology, Japan (21513505@stn.nitech.ac.jp)
*
Power system laboratory, Nagoya Institute of Technology, Japan
(s.sekizaki.845@nitech.jp)
** Distribution dept., Customer servise division, Chubu Electric Power
Co., Inc., Japan
*** Opal-RT Technologies, Montreal, Canada
Received: September 19, 2013; Accepted: September 23, 2013

340

Shinya Sekizaki, Mutsumi Aoki, Hiroyuki Ukai, Shunsuke Sasaki, Takaya Shigetou, Weihua Wang and Jean Blanger

SVR

Substation

Voltage

As mentioned above, SVR has time delay for operation.

This delay of operation may cause voltage deviation when
output of large amount of PVs changes rapidly. Tsettle can
become shorter to decrease time delay of SVR, but SVR
must need Tope. Hence, to prevent voltage deviation can be
achieved by just operating SVR in advance before rapid
voltage change.

End of line

Adequate range

Compensation

B) Lack of voltage information

Current SVR monitors self-terminal voltage (VSVR). With
no PVs, only VSVR is needed and SVR can manage voltage
properly. When a large amount of PVs are connected to DS;
however, VSVR is insufficient to manage voltage on
distribution line. This is because that voltage profile on
distribution line becomes complex due to PV and SVR
cannot know voltage profile by just observing VSVR. In order
to prevent this problem, it is necessary more information
such as voltage at remote node obtained by exchanging
information with voltage sensor (VS).

Fig. 1. Voltage on distribution line controlled by SVR.

SVR
VSVR

Line Voltage
Drop

Tap change

DEAD
BAND

Vdb

Bias
Line
Adequate range

Time
Tsettle+Tope

2.3 Proposed control method of SVR

Fig. 2. Voltage regulated by SVR.

In 2.2, some problems caused by PV have been explained.

In order to prevent these issues, the novel control method of
SVR using solar radiation sensor, VS and communication is
proposed.

2. Voltage control by SVR

2.1 Conventional control method of SVR

A) Prediction voltage changed by PV

SVR is installed into DS to manage voltage. Voltage on

distribution line controlled by SVR is shown in Fig. 1. In
Fig. 1, voltage drops with increasing of current, and is
compensated by SVR to keep within adequate range.
Algorithm of conventional SVR is shown in Fig. 2 in detail.
SVR must control voltage on entire distribution line within
adequate range, so SVR has dead band whose width is Vdb.
SVR monitors its secondary voltage (VSVR) and changes
its tap position when VSVR deviates from dead band for
settling time (Tsettle) to maintain voltage within dead band
and adequate range. However, several seconds (Tope) are
necessary from beginning operation of SVR to the end of
operation.

R1+jX1
V0

Ri+jXi
i-1

Rn+jXn
n

n-1

Substation

P1+jQ1

Pi+jQi
Pi-1+jQi-1

Pn+jQn
Pn-1+jQn-1

Fig. 3. Distribution system example.

To prevent voltage deviation caused by time delay of
SVR, proposed SVR presumes voltage change in future by
means of solar radiation which is obtained by solar
radiation sensor located at SVR. Voltage change by PV can
be calculated by voltage sensitivity against solar radiation.
Voltage sensitivity is explained by using Fig. 3 which is DS
model example. In Fig. 3, load at node i is Pi+jQi and
impedance between node i-1 and i is Ri + jXi. Variation of
voltage at node i can be calculated by[1]

2.2 Problems caused by PV

In conventional DS until now, SVR controls voltage
properly because voltage fluctuation by PV is small.
However, with many PVs, several issues may occur in
future. These issues will be explained as follows.
A) Time delay of SVR

341

Effective Voltage Control by SVR to Reduce the Capacity of SVC using Solar Radiation Information with Real Time Simulator

Vi

R P + X Q
k

k =1

l =k

k =1

l =k

(VLOW) of adequate range as following.

(1)

Tap =

V0

(2)

Pl = Pload _ l + PPV _ l

where Pl consists of Pload_l which is variation of load

and PPV_l which is variation of PVs output. Since the load
change (Pload, Ql) is slower than output change of PV
(PPV), voltage change by PV is approximated by
i

Vi

R P
k

k =1

PV _ l

l =k

(b)Voltage deviation by rapid output change of PV

By using (4), voltage change due to PV can be calculated.
In particular, the maximum and minimum voltage after
changing by PV are obtained by

(3)

V0

Vi

R P
k =1

l =k

V0

PVMAX _ l

dV
S = i S
dS

(5)

where Tap is the variation of tap from current position.

If PPV can be assumed to be in proportion to solar

radiation, (3) can be expressed as
i

+ 1 (V SVR V LOW or V sensor V LOW )

1 (V SVR VUP or V sensor VUP )
otherwise
0

(6)

V MIN _ PV _ i = V nowi

(7)

Voltage

(4)

dVi
(1 S now )
dS
dV
i S now
dS

V MAX _ PV _ i = V nowi +

Within adequate range?

presence

In (4), PPVMAX_l is the capacity of PV at node l and S is

standardized solar radiation. PPVMAX_l in DS can be known
to system operator and S may be considered as same in DS.
Therefore, voltage change by PV can be estimated by solar
radiation information.

VsecTope
Vsec
Time

1[s]

Tope

Fig. 4. Prediction of voltage changed by PV.

B) Communication between SVR and VS

In order to manage voltage on whole distribution line, it
is effective for SVR to communicate with VS located at
remote node. Therefore, SVR gets voltage information from
VS. Since voltage at the end of distribution line changes
largely rather than at upper node, VS is located at the end of
distribution line. Moreover, if several SVRs exist in DS,
upper SVR obtained voltage information from lower SVR.
An upper SVR manages voltage from secondary voltage of
upper SVR to primary voltage of lower SVR. On the other
hand, lower SVR manages voltage from own secondary
voltage to the end of line.

where Vnow and Snow are current voltage at node i and

current S, respectively.
By using (6) and (7), the maximum and minimum
voltage after changing of solar radiation can be calculated.
However, solar radiation changes randomly and does not
change always maximally or minimally. Too considering of
voltage fluctuation largely by PV causes SVR to increase
the operation and deteriorate. Therefore, it is indispensable
to predict voltage change precisely. Hence, the actual
voltage variation data observed by SVR or VS is used. SVR
calculates voltage change value each 1[s] (Vsec) at own
node and VS node. Next, voltage after Tope from presence is
calculated like Fig. 4. SVR predicts voltage after variation
by PV and changes its tap position if voltage will deviate
from adequate range. In contrast, in order to prevent the
increase of operation of SVR, the maximum and minimum
voltages after change by PV are considered as follows.

C) Tap change algorithm

By using voltage prediction and communication, SVR
decides whether to change tap position or not. Requirement
of changing tap position is following.
(a)Voltage deviation at sensor node
Voltage is monitored at secondary side of SVR (VSVR) and
at sensor node (Vsensor). SVR changes its tap position when
VSVR and Vsensor deviate upper voltage (VUP) or lower voltage

VSVR + Tope VSVR _ sec = VSVR _ presume

Vsensor + Tope Vsensor _ sec = Vsensor _ presume

342

(8)

Shinya Sekizaki, Mutsumi Aoki, Hiroyuki Ukai, Shunsuke Sasaki, Takaya Shigetou, Weihua Wang and Jean Blanger

required time-step, as well as multiple sampling rates for

the systems with different dynamics. On the other hand,
the controller of the SVR is programmed with programming
language C and implemented on DSP board. The TargetPC performs simulation in real-time and SVR controller
accepts measured data in DS from the Target-PC and
calculates and sends the reference signal (tap position of
SVR) to the Target-PC.

+ 1 (VMIN _ PV _ SVR VLOW and VSVR _ presume VLOW )

+ 1 (V
and Vsensor_ presume VLOW )
MIN _ PV _ sensor VLOW

Tap = 1 (VMAX _ PV _ SVR VUP and VSVR _ presume VUP )

(9)
1 (V
and Vsensor_ presume VUP )
MAX _ PV _ sensor VUP

otherwise
0

where VSVR_sec and Vsensor_sec is Vsec at the secondary

voltage of SVR and VS, respectively. By using (9), voltage
deviation caused by the delay of SVR can be prevented. In
proposed control method of SVR, since SVR can
compensate rapid voltage change caused by PV, the
capacity of SVC can be reduced.

3.2 Simulation model

Voltage changes largely by PV in DS with long line, so it
is difficult to control voltage in this kind of DS. Hence, to
assume rigorous situation, DS model used for simulation
has long line and needs 2 SVRs to compensate voltage drop.
DS model is shown in Fig. 6. This model has 13 nodes and
loads are connected to several nodes like Fig. 6. Model
parameters are shown on Table 1. On table 1, R and X
indicate line resistance and reactance respectively between
upper and lower node. Max load is the capacity of load
connected to lower node and PF represents power factor of
load. PVs connect to nodes where loads exist and the
capacity of PV is as same as the capacity of load.

3. Case study
3.1 Simulation using RT-LAB
Distribution system model
Real time computing
TargetPC
Host
PC

Substation
1

Programming

Language C

Reference signal
(Tap position)

Measured
data

SVR1

SVR2

10

11

12

13

Load & PV

Fig. 6. Distribution system model.

SVR
controller
Programming

Table 1 Parameters of model

DSP board

Node #

Fig. 5. RT-LAB environment.

Up
1
2
3
5
6
7
9
10
11
12

In order to confirm the effectiveness of proposed

method, a hardware-in-the-loop simulation (HIL) is used.
In this HIL setup, a real-time simulation of a distribution
system (Fig. 6) including an SVR is performed on the real
time simulator (RT-LABTM) [2]. The simulator is
connected by analog and digital inputs and outputs to a
Digital Signal Processor (DSP) board used to implement
proposed controller into SVR. The HIL simulation
environment is shown in Fig. 5. The DS model is developed
graphically in MATLAB/SimulinkTM [3] and with
SimPowerSystems on the Host-PC, then the model is
compiled with RT-LAB software and ARTEMIS realtime plug-in for SimPowerSystems, and the code is
uploaded into the Target-PC. The RT-LAB allows
parallel computation of such a complex model on multiple
processors to ensure the calculation completed within the

Line impedance
Low
2
3
4
6
7
8
10
11
12
13

Total

R[]
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
2.0

X[]
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
2.0

Max load[kVA]
(PF=0.98))
250
250
250
250
250
250
250
250
250
250
2500

3.3 Controller settings

Since DS model has 2 SVRs like Fig. 6, 2 controllers of
SVRs on different DSP board are employed. System
structure composes of 3 components like Fig. 7; 1) DS
model in RT-LAB, 2) Controller of SVR1 on DSP board, 3)

343

Effective Voltage Control by SVR to Reduce the Capacity of SVC using Solar Radiation Information with Real Time Simulator

Controller of SVR2 on DSP board. Communication

between SVRs and VS is performed in RT-LAB and
controllers on DSP boards just calculate reference signal
(reference of tap position of SVR) using voltage and solar
radiation.

active/reactive power like Fig. 8. In contrast, each PV

changes its output like Fig. 9 which is made based on actual
solar radiation data. Solar radiation data which changes
rapidly and largely like Fig. 9 is selected to confirm the
validity of proposed method even though voltage has
sudden fluctuation. The control method of SVC is set as
simple which controls self-terminal voltage near rated
voltage by using the dead band. Parameters of simulation
are shown on Table 2. Parameter values on Table 2 are ones
actually used in DS in Japan. Voltage is evaluated whether
voltage remains within adequate range or not. Adequate
range is from 101[V] to 107[V] in low voltage transformed
from high voltage by pole transformer whose ratio is
105/6600 in whole system. Authors determine the adequate
range considering the voltage drop at the low distribution
line and a lead-in wire. Simulations are performed in 4
cases like Table 3. First, simulation is performed in
conventional control of SVR without SVC in case1. Next,
the proposed control method of SVR is used without SVC
in case 2. In case 3, the conventional method with SVC
having large capacity is employed. Finally, the proposed
method with small SVC is simulated to confirm the
reduction of SVCs capacity rather than conventional
method

Distribution system model in RT-LAB

Substation

SVR1

SVR2

VS

Communication

Communication

Voltage
Solar radiation

Reference signal
(Tap position)

Voltage
Solar radiation

Reference signal
(Tap position)

SVR2 Controller
on DSP Board

SVR1 Controller
on DSP Board

Fig. 7. Model and 2 controllers.

3.4 Simulation conditions

Active power

Reactive power

Load[kW, kvar]

250
200

Table 2. Parameters of simulation

150

Simulation period[min]
Adequate range[V]
(Low voltage)
Sending voltage at substation[V]
PV position
Pole transformer ratio
Vbasis[V]
Vdb[V]

100
50
0
0

12

15

Time[min]

Conventional
SVR

Fig. 8. Each load pattern.

Output of PV[kW]

250

Tap range
Tset[s]

200
Proposed SVR

150
100

SVC

50

Tap range
Rated
voltage[V]
Dead band[V]
Location

0
0

12

15

15
101107
6600
At all over the load
105/6600
6600
99
6450/6600,
6600/6600,
6750/6600
SVR1, SVR3:60
SVR2, SVR4:90
6450/6600,
6600/6600,
6750/6600
6600
50
Node #13

Table 3. Simulation cases

Time[min]

Case
1
2
3
4

Fig. 9. Each output of PV pattern.

Each load pattern is shown in Fig. 8. All loads changes its

344

Control of SVR
Conventional
Proposed
Conventional
Proposed

Capacity of SVC [kvar]

0
0
100
50

Shinya Sekizaki, Mutsumi Aoki, Hiroyuki Ukai, Shunsuke Sasaki, Takaya Shigetou, Weihua Wang and Jean Blanger

110

Voltage[V]

3.5 Simulation results

A) Conventional method without SVC (case 1)
Simulation result in case1 is shown in Fig.10. Fig.10
shows the voltage at several nodes transformed into low
voltage by pole transformer. In Fig.10, voltage at node #9
and #13 deviate from adequate range due to rapid voltage
change and the delay of SVR. To prevent voltage deviation,
it is indispensable to compensate time delay of SVR.

108
104
102

Fig. 10. Simulation results in case 1.

110

Voltage[V]

108

#8

#9

#13

106
104
102
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time[min]

Fig. 11. Simulation results in case 2.

Voltage[V]

108

#8

#9

#13

106
104
102
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time[min]

Fig. 12. Voltage in case 3.

Output of SVC[kvar]

D) Proposed method with small SVC (case 4)

Voltage in case 4 is shown in Fig. 14. By using the
proposed control method of SVR, voltage deviation is
prevented perfectly even though there is small SVC whose
capacity is 50 [kvar] like in Fig.15, smaller than in case 3.
E) Comparison of all cases
Voltage deviation and the number of operations of SVR
in 4 cases are shown on Table 4. On Table 4, case 4 is the
best control method of SVR that prevents voltage deviation
and deterioration of SVR.

150
100
50
0
-50
-100
-150
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time[min]

Fig. 13. Output of SVC in case 3.

110

Voltage[V]

108

Table 4. Comparison of 4 cases

3
Large
Small
0
4

Delay

106

110

Small
0
6

#13

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time[min]

C) Conventional method with large SVC (case 3)

Voltage in case 3 is shown in Fig. 12. With large SVC
whose capacity is 100 [kvar], voltage deviation decreases in
comparison with in case 1. However, voltage deviation is
not prevented perfectly because output of SVC is restricted
by its capacity like Fig. 13. For example, voltage deviation
is occurred at 3[min] in Fig. 12 when output of SVC
reaches its capacity as 100 [kvar] in Fig. 13.

Large
0
4

#9

100

B) Proposed method without SVC (case 2)

Simulation result in case2 is shown in Fig. 11. In Fig. 11,
voltage deviation is almost prevented even though there is
rapid voltage change because SVR changes its tap position
in advance before output of PV changes. Like this, the
proposed SVR can compensate rapid voltage change.
However, the number of operation of SVR increases to
prevent voltage deviation and this makes SVR deteriorate.
To prevent this issue, SVC is needed as complementary
device that makes voltage fluctuation by PV smaller.

Case
SVC
Voltage deviation
The number of
SVR1
operations
SVR2

#8

4
Small
Non
0
2

#8

#9

#13

106
104
102
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time[min]

Fig. 14. Voltage in case 4.

345

Output of SVC[kvar]

Effective Voltage Control by SVR to Reduce the Capacity of SVC using Solar Radiation Information with Real Time Simulator

100

system. Dr. Sekizaki is a Member of IEEJ.

50
0

Mutsumi Aoki received the B.S., M.S.,

and Ph.D. degrees in Electrical and
Computer Engineering from Nagoya
Institute of Technology, Japan, in 1992,
1994, 1999, respectively. He was a
Researcher in TOENEC Corp. from
1994 to 2002. He is currently an
Associate Professor in the Graduate School of Engineering,
Nagoya institute of Technology. His research interest is in
the areas of power quality in distribution system. Dr. Aoki
is a Member of IEEJ.

-50
-100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time[min]

Fig. 15. Output of SVC in case 4.

4. Conclusion
Current SVR manages voltage properly with no PVs.
However, since SVR has time delay, SVR may not be able
to manage voltage properly when large amount of PVs are
connected to distribution system due to rapid voltage
change. Therefore, it is necessary to improve control
method of SVR. Furthermore, improving delay of SVR
means that SVR mitigates burden of SVC to control voltage
and the capacity of SVC can become smaller. This leads to
reduction of the cost and prompts the introduction of SVC
into distribution system. In proposed method, SVR
communicates with voltage sensor and monitors solar
radiation. Proposed SVR predicts voltage change by PV
and changes its tap position in advance before output of PV
changes, so SVR can compensate rapid voltage change and
reduce the capacity of SVC. The effectiveness of proposed
controller of SVR is confirmed by real time simulation with
DSP board.

Hiroyuki Ukai received the B.A.Sc.,

M.S., and Ph.D. degrees in engineering
from Nagoya Institute of Technology,
1977, 1979, and 1994, respectively. He
is currently a Professor in the Graduate
of Engineering, Nagoya Institute of
Technology, Japan. His current research
interests include control theory of distributed parameter
systems and control applications to power systems, robotics,
and mechanical systems. Dr. Ukai is a Member of IEEJ,
SICE, and IEEE.

Takaya Shigetou received B.S. and

M.S. degrees in engineering at University of Tokyo, Japan, in 1995 and 1997,
respectively. He is currently engaged in
development of operation and control
system for distribution lines at Chubu
Electric Power Co., Inc. Mr. Shigetou is

References
[1]

J. Kondoh, H. Aki, H. Yamaguchi, A.Murata, I, Ishii:

Voltage Regulation in Distribution Systems by
Hierarchically Cooperative Control, T. IEE Japan,
Vol.126B, No.10, pp.994-1002 (2006) (in Japanese)
[2] Reference to www.opal-rt.com
[3] Reference to www.mathworks.com

a member of IEEJ.

Shunsuke Sasaki graduated from Hakodate National College of Technology,

Japan, in 2004. He is currently engaged
in development of operation and control
system for distribution lines at Chubu
Electric Power Co., Inc.

Shinya Sekizaki received the B.S.,

M.S., and Ph.D. degrees in engineering
from Nagoya Institute of Technology,
Japan, in 2009, 2011, 2013, respectively.
He is specially appointed assistant
professor in the Institute of Engineering,
Hiroshima University, from October
2013. His research interests voltage control in distribution

346

Shinya Sekizaki, Mutsumi Aoki, Hiroyuki Ukai, Shunsuke Sasaki, Takaya Shigetou, Weihua Wang and Jean Blanger

Weihua Wang is a simulation specialist

from Opal-RT technologies, Canada,
and is currently managing the Asiapacific technical center in Beijing,
China. He received his M.Sc in
Electrical Engineering from University
of New Brunswick, Canada, and
received his B.Sc from Beijing University of Technology,
China. He has deployed many complex Hardware-in-loop
(HIL) projects for electrical applications since 2009. His
current interest lies in simulation for multi-level converters
and FACTS.

Jean Blanger As CEO and co-founder

of Opal-RT Technologies Inc., Mr.
Blanger defines the vision and corporate direction of the company and
steers its efforts in research and
development. Prior to cofounding OpalRT in 1997, he worked at Hydro
Quebec for 25 years, where he was one of the main design
engineers of the 765-kV James Bay transmission system
and the real-time simulator used to design this very large
transmission system. In 1978, he led the commercialization
of Hydro-Quebec's simulation services around the world,
designing and promoting Hydro-Quebec's electric network
simulators.

347