Академический Документы
Профессиональный Документы
Культура Документы
THE
I. INTRODUCTION
This work was supported by funding from the PA DCED BFTDA and
Westinghouse. R.J. Kerestes, G.F. Reed, A.R. Sparacino, are with the
Department of Electrical & Computer Engineering and the Power & Energy
Initiative, in the Swanson School of Engineering at the University of
Pittsburgh, Pittsburgh, PA 15210 USA (e-mails: ars88@pitt.edu,
zts6@pitt.edu, reed5@pitt.edu,).
(a)
(b)
Fig. 1. Explanation of peak shaving and load leveling. (a) Daily Load Profile
with the Application of Peak Shaving, and (b) a Daily Load Profile with the
Application of Load Leveling [1]
Fig. 2. Discharge Time at Rated Power vs. System Power Rating for Grid
Level Energy Storage [2]
Pload Pi 0
(2)
i 1
(3)
i 1
F1 a1 b1 P1 c1 P12
F2 a 2 b2 P2 c 2 P22
(4)
Fn a n bn Pn c n Pn2
2c1 P1 b1
2c 2 P2 b2
1
(5)
2c n Pn bn
P1 P2 ... Pn Pload .
The solution to (5) in matrix form is given in the following
equation
FT F1 F2 ... Fn Fi ( Pi )
(1)
i 1
where F1, F2,, Fn are the cost functions for each of the n
generators and P1, P2,,Pn are the respective output power
values in MW. Most input-output curves are given in terms of
input heat (Hi) with respect to that source of generation's
output power in MW. In the case in which the input-output
curves are given in terms of input heat with respect to output
power, the fuel cost must be multiplied by these equations to
get the cost functions. These functions are constrained by
minimum and maximum output power and minimum up and
down times. The quadratic equations are used with Lagrange
multipliers to find the most economical solution to the
problem.
P1 2c1
P 0
2
Pn 0
1
0
2c 2
0
1
0
0
2c n
1
1
1
1
0
b1
b
2
bn
Pload
(6)
Fig. 6: Economic Dispatch Problem with Coal, Oil and Gas Thermal
Generators and the Integration of Energy Storage.
(7)
The load that was used for this study was taken from [19].
The load is broken up into a residential, a commercial and a
street lighting sector. This load is shown in Fig. 7.
Daily Demand Curve
2000
Lighting Load
Residential Load
Commercial Load
1800
1600
1400
1200
Power(MW)
1000
800
600
400
200
0
10
15
20
Hour
2.9 hrs
1 p.u.
case the batteries had 120 MW delivered to each unit for six
hours and then 57.6 MW for one hour to charge the batteries
to their full MWh capacity. The charging profile for the NaS
case was mimicked for the pumped storage case.
The energy that was stored by the storage units was then
allocated using the dynamic programming method of
optimization that was discussed in Section IV. The goal of the
optimization was to minimize the average change in system
demand per hour while maintaining a maximum change less
than or equal to 300 MW
Fig. 10: Discharging Profiles for 90% and 45% Depth of Discharge.
400
3.655 hrs
Fig. 9: Discharging Profile for 45% Depth of Discharge
200
-200
-400
-600
0
10
15
20
Hour
energy storage system. The effect that the energy storage had
on the system load is illustrated in Fig. 12. The case in which
the energy storage is discharged twice daily at 45% depth of
discharge achieved the desired result of decreasing the
maximum change in system load to less than or equal to 300
MW. The 90% depth of discharge case actually provided a
flatter load profile overall, however, the maximum change in
the system load was greater than 300 MW and therefore could
not be used to remove the gas turbine generators from the
system.
Daily Demand Curve With No Storage
Power(MW)
2000
Lighting Load
Residential Load
Commercial Load
1000
10
15
20
Hour
Daily Demand Curve Using Storage With 45% Depth of Discharge
1000
Power(MW)
Power in MW
Power(MW)
2000
1000
10
15
20
Hour
Daily Demand Curve Using Storage With 90% Depth of Discharge
800
600
2000
400
1000
200
1200
10
15
20
Hour
(8)
15
20
1200
1000
800
600
400
1
7.92
0
Pcoal 0.003124
0
0
.
0096
1
7
.
97
oil
1
1
0 Pload [k ] Pstor [k ]
10
Power in MW
1400
1
1
1
0 Pload [ k ]
Hour
Fig. 12: Top: System Load without Energy Storage; Middle: System Load
with 45% Depth of Discharge Storage; Bottom: System Load with 90% Depth
of Discharge Storage
200
(9)
10
15
20
Hour
Fig. 13: Top: Hourly Generator Operating Points for Daily Load Profile
without Load Leveling; Bottom: : Hourly Generator Operating Points for
Daily Load Profile with Load Leveling
Due to the fact that the curves used in (7) were from 1984
some calculations had to be made to update them to the year
2011. Data dating back to 1991 was used to calculate the
inflation cost for coal, oil and gas. The results of the
calculations are given in (10), (11) and (12) as follows [20].
The units are in dollars per dollar.
Coal Price
Inflation:
Oil Price
Inflation:
Gas Price
Inflation:
(10)
(11)
(12)
Coal
Generation
Oil
Generation
Gas
Generation
Total
Generation
Without Storage
With Storage
$685,547
$770,127
$246,838
$261,360
$162,120
$0
$1,094,398
$1,031,487
The cost of both the NaS battery storage facility and the
pumped hydro storage facility had to be calculated to check if
these energy storage systems are a feasible solution for load
leveling. The calculation for the NaS battery storage is given
in (13) and the pumped hydro storage facility is given in (14).
The prices used in this calculation were taken from [5,10].
$1500 1000KW 120 MW
6 batteries $900,000,000
1 KW
1 MW 1battery
(13)
600MW $390,000,000
1600 MW
$1 in 1979
(14)
VI. CONCLUSION
The potential economic benefit that grid level energy
storage can provide is quite clear from this study. This can be
seen by comparing the economic dispatch for the case in
which there is no energy storage with the economic dispatch
for the case where there is energy storage. There is the
potential to save millions and even billions of dollars.
When used for load leveling, pumped storage shares nearly
all of the benefits that batteries have without the
comparatively large initial capital investment. Using pumped
storage as a form of energy storage is relatively cheap and
pumped storage has a massive maximum power capacity that
can range all the way up to the multi-gigawatt level. It
provides a response time that can be as fast as seconds. With
new developments in pumped storage such as the variable
speed pumped storage unit the charging and discharging rate
can be controlled. The money that can be saved by using
pumped storage for the application of load leveling can yield a
return on investment in a relatively short amount of time. The
main downfall that pumped storage has is geographic
limitations. However, there is a great deal of locations where
new construction is underway.
Currently pumped storage is the only form of large grid
level energy storage that is economically feasible for the
application of load leveling. The locations that can support the
installation of a pumped storage facility should be maximized
in order to provide the greatest economic gain possible.
Although at the present moment battery technology is not
where it needs to be to provide this economic gain, the
economic benefit that it can potentially have is clear as long as
its efficiency and cycle life are increased and its cost is
decreased. New battery types and chemical compositions
should also be explored so the best possible battery option is
found for the use of grid level applications.
VII. ACKNOWLEDGEMENT
The electric power and energy research group for grid
infrastructure (EPERGI) would like to extend a special thanks
to the Commonwealth of PA Ben Franklin Technology
Development Authority (BFTDA) and Westinghouse for their
support of this work.
VIII. REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
IX. BIOGRAPHIES
Robert Kerestes (M2011) was born in Pittsburgh Pennsylvania. He served
in the United States Navy from 1998 to 2002 as an interior communication
electrician on board the U.S.S. Constellation. From 2002 to 2006 Robert
served in the United States Naval reserves as an electrician in the construction
battalion. He went on to attend the Community College of Allegheny County
from 2005 to 2007 where he later transferred to the University of Pittsburgh.
In 2010 Robert graduated Magna Cum Laude from the University of
Pittsburgh with a bachelors degree in electrical engineering and with a
concentration in electrical power systems. Robert will graduate with a
masters degree in electrical engineering in the fall of 2011 with a
concentration in electrical power systems and intentions of pursuing a PhD in
the same field. Robert was awarded the first ever Siemens T&D Service
Solutions Graduate Power and Energy Scholarship in September of 2011.
Gregory F. Reed (M1985) received his B.S. in Electrical Engineering from
Gannon University, Erie PA; his M. Eng. in Electric Power Engineering from
Rensselaer Polytechnic Institute, Troy NY; and his Ph.D. in Electrical
Engineering from the University of Pittsburgh, Pittsburgh PA. He is the
director of the Power and Energy Initiative in the Swanson School of
Engineering, associate director of the Center for Energy, and associate
professor in the Electrical and Computer Engineering Department at the
University of Pittsburgh. He has 25 years of electric power industry
experience, including utility, manufacturing, and consulting at Consolidated
Edison Co. of NY, Mitsubishi Electric, and KEMA Inc.
Adam R. Sparacino (M2009) was born in New Castle, Pennsylvania.
Currently, he is finishing his Masters degree in electrical engineering from
the University of Pittsburgh with a concentration in electric power
engineering. He received his B.S. in Electrical Engineering from the
University of Pittsburgh, Pittsburgh PA. Adams research interests include