Upload
rajender-yadav
View
217
Download
0
Embed Size (px)
Citation preview
8/10/2019 06019279
1/8
1
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 andintermittent 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 andintermittent 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: [email protected]).
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.
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 powerproblem 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.
Optimal Operation Scheduling of Pumped
Storage Hydro Power Plant in Power System
with a Large Penetration of Photovoltaic
Generation Using Genetic AlgorithmRyota Aihara, Akihiko Yokoyama,Member, IEEE, Fumitoshi Nomiyama, and Narifumi Kosugi
T
Paper accepted for presentation at the 2011 IEEE Trondheim PowerTech
978-1-4244-8417-1/11/$26.00 2011
8/10/2019 06019279
2/8
2
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 anadjustable speed PSHPP in this paper.
Fig. 1. Flowchart of Power Supply Reliability Calculation
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 asthe 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).
iiiiii cPbPaF 2 (1)
where ai
, bi
and ci
are fuel cost coefficients of generator i.Pi
ispower output of generator i.
The optimization problem for one-week period is then
formulated as follows:
Minimize:
})1()({ 1
168
1 1
ii
tt
N
i
i
t
i
t
ii
t SuuPFu
(2)
Subject to:
tHN
N
i
i
tt PVPPPL 1
(3)
ii
t
i PPP maxmin (4)
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 iat time t. utiis state variable of generator
i at time t (1: committed 0: stopped). Si
is startup cost ofgenerator i.Ltis load demand at time t.PNis power output of
nuclear power plant.PHis 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.
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 Cidescribed by (5).
max
max )(i
iii
P
PFC (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.
0
10
20
30
40
50
60
70
80
90
100
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
PVOu
tput[%]
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 tgiven by the power system model and Fis the random
8/10/2019 06019279
3/8
3
number based on the normal distribution where the average is
1 and the standard deviation is 3%.
FLL gtdt (6)
The power output of PV generation is also modeled with a
randomly fluctuating component.
FPVPV gtdt (7)
where PVdt is PV output at time t considering probabilistic
fluctuation,PVgtis PV output at time tgiven by the theoretical
output and F is the random number based on the normal
distribution considering weather changes shown in Table 1.
TABLEI
RANDOM CHARACTERISTIC OF WEATHER CHANGESWeather Average Standard Deviation [%]
Sunny 1 3
Cloudy 0.5 20
Rainy 0.1 0.03
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.
0
10
20
30
40
50
60
70
80
90
100
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
PVOutput[%]
Time
Sunny
Cloudy
Rainy
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 iand MTTRiis
the mean time to repair of generatori
iiMTTR
1,MTBF
1 ii (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 re-
dispatched 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
unavailable generator. The tripped generator is restored at the
restoration rateafter time t+1.
Time t
G1
LoadDemand
G2
G3
G4
G1
LoadDemand
G2
G4
Supply Demand
MaximumOutput
Supply
Shortage
G1
LoadDemand
G2
G4
MaximumOutput
Emergency
Control
PSHPP
F ir st Step Second S tep
Generators Re-dispatch
G1
LoadDemand
G2
G5
G4
Time t+1
Supply Demand
Generator Failure
Fig. 4. Generator Re-dispatching
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.
TIMEC
LOSSLOLP
(9)
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.
8/10/2019 06019279
4/8
4
B. Algorithm Overview
An algorithm overview of the proposed method is shown
in Fig. 5.
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
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 thealgorithm are summarized in Table 2.
TABLEII
PARAMETERS OF ALGORITHM
Repeat GenerationsCmax 5000
No. of Genes at Each Genaraion 64
Mutation Rate m 0.05
Repeat Times of Tabu Search 10
No. of Neighborhood Solutions 64
Length of Tabu List 8
GA
TS
Pareto P reservation Strategy
Roulette Selection
Uniform Crossover
Mutation
Selection
Reproduction
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.
OutputofP
SHPP
Time
Pump
Mode
Gene
ratorMode
Stop
100%
-100%
0
75%
50%
25%
5 4 4 3 1 0 0 1 1 5 4
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.
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
8/10/2019 06019279
5/8
5
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 probabilityPnat 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.
(10)
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 tothe GA. It is repeated until the number of generations reaches
Cmax.
Power Supply Reliability
WeeklyFuelCostofThermalPowerPlants
Better
Better
:Tabu Search Direction for LOLP Improvement:Tabu Search Direction for Cost Reduction
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 reliabilityevaluation, failures in transmission network are not considered.
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.
0
500
1000
1500
2000
2500
3000
Sa t Sun Mon Tue Wed Thu F r i
PowerDemand[MW]
Day
Summer (August)
Spring (May)
Fig. 8. Load Demand Curve
V. SIMULATION RESULTS
A. Pareto Optimal Solutions
The pareto optimal solutions obtained by the proposed
optimization method are shown in Figs. 9 and 10.
PV 0MW
2.9
3
3.1
3.2
3.3
3.4
3.5
3.6
3.7
0 0.05 0.1 0.15 0.2 0.25
WeeklyFuelCo
st[Million$]
LOLP
PV0 MW
PV1000 MW
Fig. 9. Pareto Optimal Solutions (in Summer)
PV 0MW
1.1
1.2
1.3
1.4
1.5
1.6
1.7
0 0.05 0.1 0.15 0.2 0.25
WeeklyFuelCost[Million$]
LOLP
PV0 MW
PV1000 MW
Fig. 10. Pareto Optimal Solutions (in Spring)
g
k
n
k
ngP
1
1
8/10/2019 06019279
6/8
6
In both summer and spring seasons, it can be seen that the
pareto optimal solutions are located in a narrow range of Cost-
LOLP 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 operationpatterns when PV penetration is 1000MW in both summer and
spring seasons.
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.
0
500
1000
1500
2000
Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperReservoir
[MWh]
Day
Best LOLP Best Cost
Fig. 11. PSHPP Operation Pattern (in Summer with PV 0MW)
0
500
1000
1500
2000
Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperReservoir[MWh]
Day
Best LOLP Best Cost
Fig. 12. PSHPP Operation Pattern (in Summer with PV 1000MW)
0
500
1000
1500
2000
Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperReservoir[MWh
]
Day
Best LOLP Best Cost
Fig. 13. PSHPP Operation Pattern (in Spring with PV 0MW)
0
500
1000
1500
2000
Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperR
eservoir[MWh]
Day
Best LOLP Best Cost
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 powersystem. 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.
8/10/2019 06019279
7/8
7
0
500
1000
1500
2000
2500
3000
3500
4000
S a t S u n Mo n T u e We d T h u F r i
PowerOutput/LoadDemand[MW]
Day
Maximum Output of Thermal Minimum Output of Thermal
Hydro Nuclear
Load Demand (w/o Pump) Load Demand (with Pump)
Fig. 15. Generation Schedule (in Summer with PV 0MW, for Best Cost)
0
500
1000
1500
2000
2500
3000
3500
4000
S a t S u n Mo n T u e We d T h u F r i
Po
werOutput/LoadDemand[MW]
Day
Maximum Output of Thermal Minimum Output of Thermal
Hydro Nuclear
Load Demand (w/o PV, w/o P ump) Load Demand (with PV, with Pump)
Load Demand (with PV, w/o Pump)
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.
TABLEIII
GENERATOR DATA OF IEEERTS(MODIFIED)
1 Thermal (Oil) 5 12 5000 20
2 Thermal (Oil) 4 20 3000 50
3 Hydro 4 50 5000 40
4 Pumped Storage Hydro 1 300 - -
5 Thermal (Coal) 4 76 3926 48
6 Thermal (Oil) 3 100 3926 48
7 Thermal (Coal) 4 155 4110 89
8 Thermal (Oil) 3 197 4110 89
9 Thermal (Coal) 1 350 5541 111
10 Nuclear 2 400 7000 168
Cap. of
each
Unit[MW]
MTBF [h] MTTR [h]Generator
Group Type
No. of
Unit
TABLEIV
THERMAL POWER PLANT DATA
a b c
Thermal (Oil) 5 12 6 16-20 0 .0646 16.8899 49.7108 181.6416
Thermal (Oil) 4 20 10 21-24 0.025 43.5 200 13.356
Thermal (Coal) 4 76 38 6-9 0.0533 9.2374 164.3492 1111.4208
Therma l (Oil) 3 100 50 13-15 0. 0224 19.7128 287.4982 1511.8992
The rma l (Coa l) 4 155 77. 5 2-5 0. 0067 10.2202 207.1786 1777.1544
Therma l (Oil) 3 197 98. 5 10-12 0. 0081 20.2909 378.5918 2070. 18
Thermal (Coal) 1 350 175 1 0.0032 10.102 350.6398 8331.9264
Startup
Cost [$]
Minimum
Output [MW]Type
No. of
Unit
Maximum
Output [MW]
Cost Curve [$/hr]
F(P) = a*P^2+b*P+cMerit
Order
VIII. REFERENCES
[1] 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, inProc. 2010 International Conference on Power System
Technology, Hangzhou, China, Oct. 2010.
[2] 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.
[3]
Allen J.Wood and Bruce F.Wollenberg, Power Generation, Operation,and Control, John Wiley& Son, 1983
[4] X. Gandibleux, M. Sevaux, K. Srensen and V. T'Kindt, Metaheuristics
for multiobjective optimization, Springer, 2004.
[5] D. Srinivasan and A.Tettamanzi, Heuristics-guided evolutionary
approach to multiobjective generation scheduling, IEE Proceedings
Generation, Transmission Distribution, vol. 143, No. 6, 1996.
[6] 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. fromSophia 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.
8/10/2019 06019279
8/8
8
Akihiko Yokoyamawas 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 receivedhis 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.