Paper accepted for presentation at the 2011 IEEE Trondheim PowerTech

Optimal Operation Scheduling of Pumped

Storage Hydro Power Plant in Power System
with a Large Penetration of Photovoltaic
Generation Using Genetic Algorithm
Ryota Aihara, Akihiko Yokoyama, Member, IEEE, Fumitoshi Nomiyama, and Narifumi Kosugi


Abstract--In recent years, a substantial amount of photovoltaic
(PV) generations have been installed in Japanese power systems.
However, the power output from the PV is random and
intermittent in nature. Therefore, the PV generation poses many
challenges to system operation. To evaluate impacts of the
behavior of PV on power supply reliability, we have developed
power supply reliability evaluation model considering a large
integration of PV generation into power system. As a result, the
power supply reliability is getting worse as the PV penetration
increases. To mitigate this issue, we have proposed that pumped
storage hydro power plant (PSHPP) is used to improve the
reliability. However, its operation may increase the power system
operational cost. In this paper a new method for scheduling the
effective operating pattern for PSHPP that makes it possible to
improve both reliability and economy is presented.
Index Terms--Genetic Algorithm, Monte Carlo Simulation,
Optimal Scheduling, Photovoltaic Generation, Power Supply
Reliability, Pumped Storage Hydro Power Plant, Surplus Power
Problem, Tabu Search.

I. INTRODUCTION

HE role of renewable energy resources is remarkable in

Low Carbon Society. In recent years, a large amount of
photovoltaic (PV) and wind power generations have been
installed in power systems around the world. However, Japan
is not an appropriate site to install a large amount of wind
power generations. Therefore, the Japanese government aims
at installing a large amount of PV in power system.
Since the power output from the PV is random and
intermittent in nature, the PV generation poses many
challenges to the power system operation. When the load
demand is very small and at the same time the PV generates a
large amount of power, the balance between power supply and
demand cannot be maintained. This problem is called surplus
power problem". In addition to the surplus power problem, it
Ryota Aihara is with the Department of Electrical Engineering, The
University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
(email: aihara@syl.t.u-tokyo.ac.jp).
Akihiko Yokoyama is with the Department of Advanced Energy, The
University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8561, Japan.
Fumitoshi Nomiyama and Narifumi Kosugi are with the Power System
Engineering Group, Research Laboratory, Kyushu Electric Power Company,
Inc., Fukuoka, 815-0032, Japan.

978-1-4244-8417-1/11/$26.00 2011

is also necessary to consider the intermittent and uncertain

output characteristic of PV.
Therefore, various kinds of counter measures have been
considered when a substantial amount of PV generation is
installed into a power system. Though Battery Energy Storage
System (BESS) could be considered as a solution of these
issues, their high cost is preventing them from being
considered as a solution in most situations. This paper
proposes the effective use of Pumped Storage Hydro Power
Plant (PSHPP). PSHPP is installed originally for load leveling
within one day. The energy is stored in the form of water
pumped up from a lower reservoir to an upper reservoir. The
PSHPP is usually operated as a generator in the daytime and a
pump in the nighttime. A new operation scheduling of PSHPP
is proposed to solve this problem in this paper. It will be
operated with the proposed pattern solving the surplus power
problem caused by the PV generation. The authors have
already proposed how the PSHPP operation pattern is changed
effectively to improve the power system reliability in case of a
large integration of PV into the power system [1]. However,
the scheduling method for obtaining the optimal PSHPP
operation pattern that makes it possible to improve both
reliability and economy has not been developed yet. In this
paper, we propose a new method for planning PSHPP
operation pattern considering the surplus power problem,
power supply reliability and fuel cost reduction, which is
represented by pareto optimal solutions. The power supply
reliability and the fuel cost are estimated for each PSHPP
operation plan by using Monte Carlo simulation.
II. POWER SUPPLY RELIABILITY EVALUATION MODEL
A. Pumped Storage Hydro Power Plant
In PSHPP, the energy is stored in upper reservoir in the
form of water. Therefore, it should be operated within the
capacity of the upper reservoir. In this paper, the amount of
the stored water in the upper reservoir is considered as energy.
It can be operated 6 to 8 hours continuously when the
operation starts from the full capacity of the upper reservoir.
Therefore, the capacity of the upper reservoir is assumed to be
7-hour operation of the rated power at generator mode.

In general, there are two types of PSHPP, fixed speed one

and adjustable speed one. The fixed speed PSHPP cannot
adjust the input power when it operates as a pump. The
adjustable speed one can adjust the input power. It is coming
into the limelight because of its adjustable characteristics in
the pump mode [2]. Therefore, the PSHPP is assumed to be an
adjustable speed PSHPP in this paper.

The optimization problem for one-week period is then

formulated as follows:
Minimize:

168 N

{u F ( P )  u (1  u
i
t

i
t

i
t 1

)S i}

(2)

t 1 i 1

Subject to:

Lt

 PN  PH  PVt

(3)

i 1

i
i
Pmin
d Pt i d Pmax
      (4)

B. Reliability Evaluation Model

The proposed evaluation method of the power supply
reliability is summarized as a flowchart in Fig.1. In this paper,
it is assumed that the operation schedule of PSHPP is planned
for a week. The simulation period is a week and its sampling
time is 1 hour. First, based on the available data, the
generation schedule for 7 days is determined. The available
data consist of load demand given by power system model,
theoretical PV generation output and PSHPP operation pattern.
The generation schedule is determined by solving
optimization problem, which minimizes the fuel cost by using
Dynamic Programming (DP). Next, Monte Carlo simulation is
carried out considering uncertainties of load demand, PV
output and generator failure. During the period of the
simulation, when the power supply and demand imbalance
occurs, the thermal generator outputs are adjusted and PSHPP
is operated to avoid the imbalance. If it cannot be avoided,
this period is regarded as supply interruption. We assume that
both supply shortage and surplus power supply are defined as
the supply interruption.
1) Generation Scheduling
As mentioned before, generation scheduling is found by
solving an optimization problem in which the fuel cost is
minimized. In this paper, it is assumed that nuclear and hydro
power plants are always operated at the rated power output.
As a result, only the fuel cost of thermal plants is considered
here. The fuel cost of a thermal plant is given in (1).

Fi

ai Pi2  bi Pi  ci

(1)
where a , b and c are fuel cost coefficients of generator i. Pi is
power output of generator i.
i

To solve the optimization problem, Dynamic Programming

(DP) [3] is used in this research. The merit order method is
used in the unit commitment. The merit order is determined by
using the fuel cost of generator at the rated output. The
combination of committed generators is searched sequentially
in ascending order of Ci described by (5).

Ci

F i ( P i max )
P i max

(5)

For simplification, the shutdown cost is not considered

here. The PV output is assumed to be theoretical in generation
scheduling because power system operator cannot predict its
output precisely at this time. The theoretical power output of
PV is shown in Fig. 2.

PV Output [%]

Fig. 1. Flowchart of Power Supply Reliability Calculation

where is the total fuel cost of thermal power plants for one
week. N is the number of thermal power plants. Pti is power
output of generator i at time t. uti is state variable of generator
i at time t (1: committed 0: stopped). Si is startup cost of
generator i. Lt is load demand at time t. PN is power output of
nuclear power plant. PH is power output of hydro power plant.
PVt is theoretical power output of PV at time t. Pimin is
minimum output of generator i. Pimax is maximum output of
generator i.

100
90
80
70
60
50
40
30
20
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time

Fig. 2. Theoretical PV Output

2) Monte Carlo Simulation

In this paper, we consider three types of probabilistic
uncertainty concerning load demand, power output of PV and
generator failure.
The load demand model with a randomly fluctuating
component is derived from (6). Ldt is load demand at time t
considering probabilistic fluctuation. Lgt is load demand at
time t given by the power system model and F is the random

Generator Failure

Generators Re-dispatch

Weather

Average

Standard Deviation [%]

Sunny

Cloudy

0.5

20

Rainy

0.1

0.03

PV Output [%]

G5

Supply Demand

G4

Load Demand

Load Demand

G2
G1

G1

G2

G4

G4

PSHPP

Second Step

Time t

Time t+1

Fig. 4. Generator Re-dispatching

The weather change is simply modeled at the probability of

1/3 at time t. It is assumed that the weather is independent
between before and after time t. The theoretical PV outputs
for three kinds of weathers are shown in Fig. 3.
100
90
80
70
60
50
40
30
20
10
0

G1

G1

TABLE I
RANDOM CHARACTERISTIC OF WEATHER CHANGES

Maximum Output

G3

where PVdt is PV output at time t considering probabilistic

fluctuation, PVgt is PV output at time t given by the theoretical
output and F is the random number based on the normal
distribution considering weather changes shown in Table 1.

G2

(7)

Load Demand

PVgt u F

PVdt

G4

The power output of PV generation is also modeled with a

randomly fluctuating component.

Supply
Shortage

Supply Demand

Emergency
Control

First Step

Maximum Output

(6)

Load Demand

L dt Lgt u F

unavailable generator. The tripped generator is restored at the

restoration rate after time t+1.

G2

number based on the normal distribution where the average is

1 and the standard deviation is 3%.

4) Power Supply Reliability Calculation

The power supply reliability is evaluated through LOLP
(Loss of Load Probability) index which can be computed from
(9). LOSS is the total supply interruption period. C is the
number of trial. TIME is the simulation period. The trial is
repeated until 10,000 times so that LOLP is converged to the
steady-state value.

LOLP

Sunny
Cloudy

LOSS
C u TIME

(9)

Rainy

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time

Fig. 3. PV Output for Three Kinds of Weather

The generator failure occurs during Monte Carlo

simulation at the failure rate described by (8). In (8), MTBFi
is the mean time between failure of generator i and MTTRi is
the mean time to repair of generator i

Oi

1
, Pi
MTBFi

1
MTTRi

(8)

3) Generator Re-dispatch
During Monte Carlo simulation, if the generator failure
occurs at time t, the other generators are re-dispatched
immediately. Figure 4 shows how the generators are redispatched when generator G3 is tripped at time t, for example.
First, Generator G1, G2 and G4 are re-dispatched immediately.
If imbalance of power supply and demand cannot be avoided
by the re-dispatching, PSHPP is operated immediately to
avoid it as an emergency control. In the next time t+1, the
remaining generators are committed to compensate for the

5) Weekly Fuel Cost Calculation

The weekly fuel cost is calculated at each trial during
Monte Carlo simulation. If the power output of PV and load
demand is constant, the weekly fuel cost is not changed from
the cost which determined by DP. However, the PV output
and the load demand could fluctuate randomly in Monte Carlo
simulation. Therefore, the load dispatch of generators is
changed and its operational cost rises or falls at each trial. In
this paper, the weekly fuel cost is defined as the mean value of
fuel costs of all trials excluding the trial where the generator
failure occurs. This is because the fuel cost increase is the
unexpected factor due to the generator failure. In other words,
the outage cost is not determined uniquely.
III. OPTIMAL OPERATION SCHEDULING OF PSHPP
A. Pareto Optimal Solutions
In general, improvement of power system reliability and
reduction of operational cost are competing. Optimization of
the PSHPP scheduling has no unique solution. This type of
problem is a multi-objective optimization problem. In this
paper, pareto optimal solutions are obtained.

B. Algorithm Overview
An algorithm overview of the proposed method is shown
in Fig. 5.

GA and TS are combined to obtain pareto optimal PSHPP

operation patterns. Tabu search is a local search technique
which enhances the performance using memory structures
named Tabu List. Once a potential solution has been
determined, it is marked as "taboo" so that the algorithm does
not visit that possibility repeatedly. The parameters of the
algorithm are summarized in Table 2.
TABLE II
PARAMETERS OF ALGORITHM
Repeat Generations C max

5000

No. of Genes at Each Genaraion

64
Pareto Preservation Strategy
Roulette Selection

Selection
GA

Uniform Crossover
Mutation

Reproduction
Mutation Rate m

TS

0.05

Repeat Times of Tabu Search

10

No. of Neighborhood Solutions

64

Length of Tabu List

0 25% 50% 75% 100%

-100%

Output of PSHPP

Pump
Mode Stop Generator Mode

C. Algorithm Details
1) Quantization of PSHPP operation pattern
The operation pattern of PSHPP is quantized by 6 steps at
each time t. Figure 6 shows quantization of PSHPP operation
pattern.

Time

Gene

Fig. 6. Quantization of PSHPP Operation Pattern

2) Initial Solutions in GA
The initial genes of the PSHPP operation patterns are
generated by random number at each time period, which are
quantized as mentioned above. If the solution can not satisfy
the capacity constraint of the PSHPP, the solution is rejected
and another solution is generated.
Fig. 5. Algorithm Overview

In this study, PSHPP operation schedule is determined by

Genetic Algorithm (GA) and Tabu Search (TS). GA is one of
the evolutionary algorithms, which generates solutions to an
optimization problem using techniques inspired by natural
evolution, such as inheritance, mutation, selection and
crossover. It generates multiple solutions simultaneously at
each generation, and it is good at global search. However GA
is not good at local search. Therefore, this paper proposes that

3) Evaluation in GA
Generated genes are evaluated by Power Supply Reliability
Evaluation Model described in section II. The evaluation
indexes are power supply reliability and the fuel cost.
4) Determine the Pareto Optimal Solutions
The pareto optimal solutions are determined from
evaluated solutions. Solutions are labeled by pareto-ranking
method [4]. The ranking is numbered sequentially from
solutions closer to pareto optimal solutions

5) Selection and Reproduction in GA

To re-produce the descendants, we use pareto preserve
strategy and uniform crossover. The pareto optimal solutions
are inherited to the next generation as elite whose ranking is
the best. The other descendants are produced by uniform
crossover whose parents are selected by roulette selection.
The parent genes are arranged in descending order of their
ranking value. The probability Pn at which the nth rank gene is
chosen is determined by (10). g is no. of genes at each
generation. Reproduction is implemented through uniform
crossover, whereby each component is randomly selected
from either parent with equal probability. Uniform crossover
has a good characteristic of being unbiased with respect to the
reproduction of genes [5]. Mutation is conducted at the rate m
to prevent the solutions from going into a local optimal
solution.
g  n 1
Pn
g
(10)
k

B. Simulation Condition
Figure 8 depicts two different load conditions for two
seasons, i.e. summer (August) and spring (May). The system
peak load is 3200 MW. In the simulation, the amount of PV
penetration is set to be 0MW and 1000MW.

6) Tabu Search
TS is conducted whenever GA is repeated 100 times. TS
can optimize only single objective function, TS is repeated
twice from each pareto optimal solutions. First, selecting one
of the pareto solution, TS is carried out from it. The objective
function of TS is LOLP improvement. Second, starting from
the same pareto solution, TS is conducted with objective
function for cost reduction. Figure 7 shows the searching from
pareto optimal solutions. This process is repeated until all
pareto optimal solutions are searched by TS and the pareto
optimal solutions are revised. Then the process goes back to
the GA. It is repeated until the number of generations reaches
Cmax.

A. Pareto Optimal Solutions

The pareto optimal solutions obtained by the proposed
optimization method are shown in Figs. 9 and 10.

Power Demand [MW]

3000
2500
2000
1500
1000
500

Summer (August)
Spring (May)

0
Sa t

Tu e
Day

We d

Thu

Fr i

V. SIMULATION RESULTS

3.7

Weekly Fuel Cost [Million $]

3.6

:Tabu Search Direction for LOLP Improvement

:Tabu Search Direction for Cost Reduction

PV 0MW

3.5
3.4
3.3
3.2
3.1
3

PV0 MW
PV1000 MW

2.9
0

0.05

0.1

0.15

0.2

0.25

0.15

0.2

0.25

LOLP

Fig. 9. Pareto Optimal Solutions (in Summer)

1.7

PV 0MW

1.6

Power Supply Reliability

Better

Fig. 7. TS Direction from Pareto Optimal Solutions

IV. SIMULATION CONDITION

A. Simulation Model
The simulation is carried out on the modified IEEE 24-bus
Reliability Test System (RTS) [6] with the generator data
shown in the appendix. The system is modified by adding a
300MW PSHPP in addition to the existing hydro power plants.
Since this paper focuses on the power supply reliability
evaluation, failures in transmission network are not considered.

Weekly Fuel Cost [Million $]

Better

Mon

Fig. 8. Load Demand Curve

k 1

Weekly Fuel Cost of Thermal Power Plants

Sun

1.5

1.4

1.3

1.2
PV0 MW
PV1000 MW

1.1
0

0.05

0.1
LOLP

Fig. 10. Pareto Optimal Solutions (in Spring)

B. Operation Schedule of PSHPP

The obtained operation schedule of PSHPP is shown in
Figs. 11 to 14. The best solutions concerning LOLP and the
fuel cost under each condition are discussed here.

Capacity of Upper Reservoir [MWh]

Best LOLP

Best Cost

2000
1500
1000
500
0
Sa t

Sun

Mon

Tue
Day

Wed

Thu

Fr i

Fig. 11. PSHPP Operation Pattern (in Summer with PV 0MW)

Capacity of Upper Reservoir [MWh]

Best LOLP

Best Cost

2000
1500
1000
500
0
Sa t

Sun

Mon

Tue
Day

Wed

Thu

Fr i

Fig. 12. PSHPP Operation Pattern (in Summer with PV 1000MW)

Capacity of Upper Reservoir [MWh]

Best LOLP

Best Cost

2000
1500
1000
500
0
Sa t

Sun

Mon

Tue
Day

Wed

Thu

Fig. 13. PSHPP Operation Pattern (in Spring with PV 0MW)

Fr i

Best LOLP
Capacity of Upper Reservoir [MWh]

In both summer and spring seasons, it can be seen that the

pareto optimal solutions are located in a narrow range of CostLOLP space when PV penetration is 0MW. The pareto
optimal solutions do not include various kinds of operations of
PSHPP because there is no PV output. On the contrary, the
pareto optimal solutions includes various kinds of operation
patterns when PV penetration is 1000MW in both summer and
spring seasons.

Best Cost

2000
1500
1000
500
0
Sa t

Sun

Mon

Tue
Day

Wed

Thu

Fr i

Fig. 14. PSHPP Operation Pattern (in Spring with PV 1000MW)

In summer, the PSHPP is operated as a pump in the

nighttime and as a generator in the daytime on weekday to
supply the heavy summer load when PV penetration is 0MW.
The role of the PSHPP in this case is load leveling within a
week. It is the conventional role of PSHPP in the power
system. Compared with the operation patterns between LOLP
best case and cost best case, they are not so much difference.
The near optimal operation pattern is determined uniquely in
summer season when PV penetration is 0MW. In contrast,
when the PV penetration increases to 1000 MW in summer,
the PSHPP is operated as a pump in the daytime and as a
generator in the nighttime in the LOLP best case in spite of no
surplus power. However, PSHPP is hardly operated in the cost
best case.
In spring, the PSHPP is hardly operated when the PV
penetration is 0MW because the operation of the PSHPP does
not contribute much to the fuel cost reduction. When the PV
penetration increases to 1000 MW, PSHPP is hardly operated
in the cost best case, either. The PSHPP has to be operated as
a pump in the daytime both on the weekday and weekend in
the LOLP best case. This pattern is similar to that of summer
pattern.
C. Weekly Generation Schedule
The weekly thermal generation schedules for the PSHPP
operation patterns are shown in Figs. 15 and 16. The cost best
case in summer season when PV penetration is 0MW and the
LOLP best case in spring season when PV penetration is
1000MW are discussed here. In Fig. 15, the PSHPP pushes up
the load demand in the nighttime and pushes it down in the
daytime on weekdays. In Fig. 16, the PSHPP pushes up the
load demand in the daytime and pushes it down in the
nighttime. The deducted load demand shown by solid black
line falls below the minimum output of thermal plants in the
daytime, which shows that the surplus power occurs. However,
in this case, the PSHPP can be operated for both PV output
decrease and increase by emergency operation. It is concluded
that modification of the operation pattern of the PSHPP is
effective for the surplus power problem.

7
Maximum Output of Thermal

Minimum Output of Thermal

Hydro

Nuclear

Load Demand (w/o Pump)

Load Demand (with Pump)

TABLE III
GENERATOR DATA OF IEEE RTS (MODIFIED)

Power Output / Load Demand [MW]

4000

Cap. of
each
MTBF [h] MTTR [h]
Unit[MW]
12
5000
20

Generator
Group

Type

No. of
Unit

Thermal (Oil)

Thermal (Oil)

20

3000

50

Hydro

50

5000

40

Pumped Storage Hydro

300

Thermal (Coal)

76

3926

48

Thermal (Oil)

100

3926

48

Thermal (Coal)

155

4110

89

Thermal (Oil)

197

4110

89

Thermal (Coal)

350

5541

111

10

Nuclear

400

7000

168

3500
3000
2500
2000
1500
1000

TABLE IV
THERMAL POWER PLANT DATA

500
0
Sa t

Sun

Mo n

Tu e

We d

Th u

Day

Fig. 15. Generation Schedule (in Summer with PV 0MW, for Best Cost)
Maximum Output of Thermal
Hydro
Load Demand (w/o PV, w/o Pump)
Load Demand (with PV, w/o Pump)

Type

Fr i

Minimum Output of Thermal

Nuclear
Load Demand (with PV, with Pump)

No. of
Maximum
Minimum
Merit
Unit Output [MW] Output [MW] Order

Cost Curve [$/hr]

F(P) = a*P^2+b*P+c
b
c

Startup
Cost [$]

Thermal (Oil)

12

16-20

0.0646

16.8899

49.7108

Thermal (Oil)

20

10

21-24

0.025

43.5

200

181.6416
13.356

Thermal (Coal)

76

38

6-9

0.0533

9.2374

164.3492

1111.4208

Thermal (Oil)

100

50

13-15

0.0224

19.7128

287.4982

1511.8992

Thermal (Coal)

155

77.5

2-5

0.0067

10.2202

207.1786

1777.1544

Thermal (Oil)

197

98.5

10-12

0.0081

20.2909

378.5918

2070.18

Thermal (Coal)

350

175

0.0032

10.102

350.6398

8331.9264

Power Output / Load Demand [MW]

4000

VIII. REFERENCES

3500

[1]

3000
2500
2000

[2]

1500
1000
500

[3]

0
Sa t

Sun

Mo n

Tu e

We d

Th u

Fr i

Day

Fig. 16. Generation Schedule (in Spring with PV 1000MW, for Best LOLP)

VI. CONCLUSION
This paper presents a new method for determining an
optimal PSHPP operation pattern which makes it possible to
improve both reliability and economy in the power systems
with a large integration of PV. The simulation results show
that the proposed method is very effective in terms of the
reliability of power system operation. The total fuel cost of
thermal power plants increases in order to secure the power
supply reliability using operation of the PSHPP in power
system with a large penetration of PV. In this paper, the hot
reserve capacity is not considered in scheduling of thermal
power plants. In the future work, it is necessary to develop a
cooperative scheduling method that optimizes both PSHPP
operation and hot reserve capacity of thermal power plants.
VII. APPENDIX
The generator data of IEEE RTS is tabulated in Table 3.
The cost curve and the startup cost of thermal power plants
are shown in Table 4.

[4]
[5]
[6]

R. Aihara, A. Yokoyama, F. Nomiyama and N. Kosugi, Impact of

Operational Scheduling of Pumped Storage Power Plant Considering
Excess Energy and Reduction of Fuel Cost on Power Supply Reliability
in a Power System with a Large Penetration of Photovoltaic
Generations, in Proc. 2010 International Conference on Power System
Technology, Hangzhou, China, Oct. 2010.
T. Kuwabara, A. Shibuya, H. Furuta, E. Kita and K. Mitsuhashi, Design
and dynamic response characteristics of 400 MW adjustable speed
pumped storage unit for Ohkawachi Power Station, IEEE Transactions
on Energy Conversion, vol. 11, No. 2, 1996.
Allen J.Wood and Bruce F.Wollenberg, Power Generation, Operation,
and Control, John Wiley& Son, 1983
X. Gandibleux, M. Sevaux, K. Srensen and V. T'Kindt, Metaheuristics
for multiobjective optimization, Springer, 2004.
D. Srinivasan and A.Tettamanzi, Heuristics-guided evolutionary
approach to multiobjective generation scheduling, IEE Proceedings
Generation, Transmission Distribution, vol. 143, No. 6, 1996.
Reliability Test System Task Force of the Application of Probability
Methods subcommittee, IEEE RELIABILITY TEST SYSTEM, IEEE
Transactions on Power Apparatus and Systems, vol. PAS-98, No. 6,
1979.

IX. BIOGRAPHIES
Ryota Aihara was born in Tokyo, Japan, on
September 20, 1984. He received B.S. Eng. from
Sophia University, Tokyo, Japan in 2009 and M.S
from The University of Tokyo in 2011. Currently
he is pursuing the Ph.D. degree at The University
of Tokyo. He is a student member of IEEJ.

8
Akihiko Yokoyama was born in Osaka, Japan, on
October 9, 1956. He received B.S., M.S. and Dr.
Eng. from The University of Tokyo, Tokyo, Japan
in 1979, 1981 and 1984, respectively. He has been
with Department of Electrical Engineering, The
University of Tokyo since 1984 and currently a
professor in charge of Power System Engineering.
He is a member of IEEJ, IEEE and CIGRE.

Fumitoshi Nomiyama was born in 1969. He

received his B.Eng., M.Eng. degrees in electrical
engineering from Doshisha University in 1992 and
1994 respectively. He joined Kyushu Electric
Power Company Inc. in April 1994.

Narifumi Kosugi was born in 1970. He received

his B.Eng., M.Eng. degrees in electrical
engineering from Osaka University in 1993 and
1995 respectively. He joined Kyushu Electric
Power Company Inc. in April 1995.