Analysis of the Cost per Kilowatt Hour to Store Electricity. IEEE

Analysis of the Cost per Kilowatt Hour to Store Electricity. IEEE
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
529
Analysis of the Cost per Kilowatt Hour
to Store Electricity
Piyasak Poonpun, Student Member, IEEE, and Ward T. Jewell, Fellow, IEEE
Abstract—This paper presents a cost analysis of grid-connected
electric energy storage. Various energy storage technologies are
considered in the analysis. Life-cycle cost analysis is used. The results are presented in terms of the cost added to electricity stored
and discharged, in US dollar per kilowatt hour. Results are compared with wholesale and retail electricity costs and with the cost
of conventional pumped hydro storage.
Index Terms—Batteries, economic analysis, energy storage,
flywheels.
NOMENCLATURE
Annual storage unit replacement cost (US$/kWh).
Annualized capital cost (US$/year).
Annual energy production of storage system
(kWh/year).
ARC
Total annual replacement cost (US$/year).
BOP
Total cost for balance of plant (US$).
BOPU Unit cost for balance of plant (US$/kWh).
C
Number of charge/discharge cycles in life of
storage.
COE
Cost added by storing electricity (US$/kWh).
CRF
Capital recovery factor.
D
Annual operating days for storage unit (days per
year).
eff
Efficiency
F
Future value of replacement cost (US$/kWh).
Length of each discharge cycle (h).
HO
ir
Annual interest rate (%).
n
Number of charge/discharge cycles per day.
Fixed operation and maintenance cost (US$/kW·
OMf
year).
OMC Total annual fixed operation and maintenance cost
(US$/year).
P
Rated power output capacity of energy storage system (kW).
PCS
Total cost for power electronic (US$).
PCSU Unit cost for power electronic (US$/kW).
r
Replacement period (year).
SUC
Total cost for storage units (US$).
A
AC
AEP
Manuscript received February 6, 2007; revised July 25, 2007. Paper no. TEC00027-2007.
The authors are with the Wichita State University, Wichita, KS 67260 USA
(e-mail: [email protected]; [email protected]).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TEC.2007.914157
SUCU Unit cost for storage units (US$/kWh).
TCC Total capital cost (US$).
y
Lifetime of energy storage (year).
I. INTRODUCTION
HERE are about 90 GW of electric energy storage, almost
all pumped hydro, operating in the world today [1], which
is 3% of total generating capacity. New pumped hydro installations are limited by availability of sites. Siting concerns are
reduced if other storage technologies are located in smaller units
on the distribution system. This concept is known as distributed
energy storage (DES).
The DES may, in the future, be more important, and be present
in much higher penetrations than distributed generation (DG)
[2]. The DES technologies may include batteries, flywheels,
and electrochemical capacitors (“super” or “ultra” capacitors),
of which batteries and flywheels appear to be the most promising for bulk storage. Compressed air energy storage and pumped
hydro storage are usually large and have special siting needs,
and superconducting magnetic energy storage are short-duration
devices used for uninterruptible power supplies and other power
quality support, making them less suitable for the DES.
The benefits of electricity storage are well known and include
the following [2]:
1) Support of renewables: Storage can reduce fluctuations
in wind and photovoltaic (PV) output, and allows sale of
renewable energy at high-value times.
2) Reliability and power quality: Storage will allow loads to
operate through outages.
3) Reactive power control, power factor correction, and voltage control: Power electronic interfaces provide the ability
to rapidly vary reactive as well as active power.
4) Load leveling: Storage is charged during light-load periods, using low-cost energy from base-load plants, and
discharged during high-load times, when the energy value
is higher. The benefits are improved load factor, deferred
generation expansion, and reduced purchase at peak times
and generation by peaking units.
5) Load following: Storage with power electronic interfaces
can follow load changes very rapidly, reducing the need
for generating units to follow load.
6) Bulk energy management: Bulk power transfers can be
delayed by storing the energy until it is needed, or until its
value increases.
7) Spinning reserve: Because of its ability to rapidly change
the output, storage with power electronic interfaces can
act as spinning reserve, reducing the need for conventional
spinning reserve units.
T
0885-8969/$25.00 © 2008 IEEE
530
8) Deferral of new transmission capacity: Properly located
storage units can be charged during off-peak times, reducing peak loading of transmission lines and effectively
increasing transmission capacity.
9) Deferral of new generating capacity: Fewer peaking units
are needed when storage reduces peak demand.
10) Support of distributed generation: Storage allows the DG,
such as microturbines and fuel cells, to be operated at constant output at its highest efficiency, reducing fuel use and
emissions. Discharging DES during peak demand times
also reduces the needed capacity of the DG.
11) System stability: Power and frequency oscillations can be
damped by rapidly varying the real and reactive output
of storage. The improved stability margin is obtained by
electronic controls for the DES.
12) Automatic generation control: Energy stored on a system
can be used to minimize area control error. The benefits are
easier compliance with North American Electric Reliability Corporation (NERC) standards (C1–C4) and reduced
mechanical wear on cycling units.
13) Black start capability: Stored energy can be used to start
an isolated generating unit.
14) Reduced fuel use: Use of less-efficient peaking units is reduced by charging storage with energy from more-efficient
base load-generating units. Because peaking units often
burn natural gas, this also offers natural gas conservation benefits. Also, by improving the system power factor,
losses will be reduced, and there is a concomitant reduction of energy use.
15) Environmental benefits: Reduced fuel use results in reduced emissions and natural gas conservation.
16) Increased efficiency and reduced maintenance of generating units: Load following by storage units allows prime
movers to be operated at more constant and efficient set
points, increasing their efficiency, maintenance intervals,
and useful life.
17) Increased availability of generating units: During peak
periods, charged energy storage added to available generation increases total system capacity.
While electricity storage is understood to have these strong
technical merits, it is generally thought of as too expensive to
be used in high penetrations. The costs of storage technologies,
however, are dropping, and cost/benefit analyses have shown
that it is economically justified in some cases [3].
This paper presents the development of a new technique for a
simple economic feasibility evaluation of small energy storage
facilities. Such facilities would be used, for example, for renewable energy or distributed storage. The technique calculates the
cost added to each unit of energy [in kilowatt hour (kWh)] that
is stored, and later, returned to the grid.
Results for several commercially available types of energy
storage, including conventional pumped hydro, are, then, presented. The results are compared with existing and forecast
electricity prices, and other issues that might affect the feasibility of small storage units are discussed. Conclusions are, then,
presented on the possible future use of small energy storage
systems.
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
II. DISTRIBUTED ENERGY STORAGE COSTS
There are two costs to consider for an electricity storage
system. The energy cost is the cost of storage elements, e.g.,
pumped hydro reservoirs or batteries. The energy cost is expressed in cost per unit of stored energy, US$ per kilowatt hour,
for example. This is not to be confused with the conventional
cost of purchasing a unit of electricity, which has identical units.
The energy cost for storage is the cost of the devices that actually store the energy, which can be charged and discharged
many times. The energy rating of a storage system is the total
energy that the system can store.
The other cost of energy storage is the power cost. This would
include the rotating synchronous machines in a pumped hydro
unit, or the power electronic rectifier/inverters in a battery storage system. The power cost is expressed in cost per unit of power,
US$ per kW, for example. The power rating is the instantaneous
capacity of the storage unit. It determines how quickly the storage system can be charged or discharged. The two costs, power
and energy, in combine, give the total initial capital cost of a
storage unit.
The economics of large pumped hydro units are analyzed with
production costing techniques. For smaller applications such as
flywheels, battery energy storage units, DG, and renewable energy applications, it is useful to have a simpler technique of
estimating the economics of storage units. One such method
is to convert the energy, power, installation, and operating and
maintenance costs of a storage unit to the cost added to a unit of
electricity stored. This cost is, then, added to the conventional
electricity price to determine a total price for stored electricity. For example, if electricity is generated at US$0.05/kWh,
and a particular storage system adds US$0.10/kWh, then, the
total price of that unit of stored electricity is US$(0.05 +
0.10)/kWh = US$0.15/kWh. This can, then, be compared with
potential additional value of storing electricity, such as shifting
wind-generated electricity from off- to on-peak, for an initial
estimate of the feasibility of energy storage.
This paper presents a technique to convert the installed and annual costs of energy storage to the cost added to each stored unit
of electricity. The technique was developed for applications such
as distributed storage and renewable energy. The most promising technologies for such applications are flywheels and various
types of batteries. All use a power electronic rectifier–inverter
interface, so that the technique assumes such an interface.
III. CALCULATION OF COST ADDED TO STORE ELECTRICITY
The total energy discharged annually by an energy storage
system is referred to as annual energy production (AEP), which
can be written as
AEP = P ∗ n∗ H0 ∗ D.
(1)
The annual fixed operation and maintenance cost in US$ per
year is
OMC = OMf ∗ P.
(2)
The TCC for the energy storage system consists of three
components: the total (power) cost of power electronic rectifier/
POONPUN AND JEWELL: ANALYSIS OF THE COST PER KILOWATT HOUR TO STORE ELECTRICITY
inverters, the total (energy) cost for storage units, and the TCC
for the balance of plant.
The total cost for the power electronics in US$ is
PCS = PCSU∗ P.
531
TABLE I
INPUTS AND OUTPUTS OF THE CALCULATIONS
(3)
The total cost for storage units in US$ can be obtained by
SUCU∗ P ∗ HO
eff
energy (kWh) out during discharge
eff =
.
energy (kWh) in during charge
SUC =
(4)
(5)
The total cost for the balance of plant in US$ is
BOP = BOPU∗ P ∗ H0 .
(6)
The TCC, which is the sum of the total costs for the power
electronics, storage units, and balance of plant, is
TCC = PCS + SUC + BOP.
(7)
The annualized capital cost is, then,
AC = TCC∗ CRF.
(8)
The CRF [4] is given as
CRF =
ir (1 + ir )y
.
(1 + ir )y − 1
(9)
When batteries are used as the storage element, they may have
to be replaced one or more times during the life of the plant.
This cost is annualized (US$ per kilowatt hour) [4] as
A = F ∗ [(1 + ir )−r + (1 + ir )−2r + · · ·]∗ CRF.
(10)
The number of terms in the factor of the previous equation is
equal to the number of times batteries are replaced during the
life of the system. Thus, the equation shown, with two terms, is
for batteries being replaced twice during the plant life.
Battery life is the fixed number of charge/discharge cycles.
The replacement period in years can, then, be calculated as
follows
C
(11)
r= ∗ .
n D
The annual battery replacement cost, then, is
A∗ P ∗ Ho
.
(12)
eff
Finally, the cost added to a unit (in kilowatt hour) of electricity
stored is
(AC + OMC + ARC)
COE =
.
(13)
(P ∗ n∗ HO ∗ D)
ARC =
Table I summarizes the inputs needed to do these calculations
and the outputs of the calculations.
IV. CASE STUDIES
A. Assumptions
In the case studies presented in this section, systems are assumed to operate either 250 or 100 d/year. Systems operating all
year will operate about 250 d/year, the approximate number of
weekdays minus holidays in a year. Systems designed to operate
only during peak use seasons are assumed to operate 100 days,
or 20 weeks, per year.
The length of the discharge cycle depends on the application. In this paper, 8 h [5] is assumed for generation applications. Generation applications are designed to charge overnight
and discharge during the day. Thus, storage for generation is
assumed to charge and discharge one time during each 24-h
period. For transmission and distribution (T&D) applications,
storage discharges during morning and afternoon peak periods
and is charged at other times. A 4 h [5] discharge time is assumed for T&D storage systems, and these systems charge and
discharge twice during each 24-h period.
Rated output capacity for generation applications ranges from
10 to 1000 MW [5]. Capacity for transmission/distribution applications is between 100 kW and 2 MW [5].
This paper assumes that the annual interest rate for financing
the storage system is 7.7% [6]. Inflation and escalation rates are
not considered in this analysis.
B. Storage Systems and Technologies
The most promising commercial or near-commercial battery technologies are considered in this analysis: lead acid
(LA), valve-regulated LA (VRLA), sodium sulfur (Na/S),
zinc/bromine (Zn/Br), and vanadium redox (VB). Flywheels,
which are commercially available for power quality applications
and are now being demonstrated for frequency regulation [7],
are the most promising nonbattery storage technology, and these
are also considered for T&D applications.
532
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
TABLE II
ADJUSTED CASE VALUES FOR GENERATION APPLICATIONS [5], [8]
Fig. 1. Added cost (COE) vs. discharge time, 10 MW generation application
operating 250 d/year.
TABLE III
ADJUSTED CASE VALUES FOR T&D APPLICATIONS [5], [8]
Fig. 2. Added cost (COE) vs. discharge time, 10 MW generation application
operating 100 d/year.
These storage technologies are all commercially available.
Manufacturers of each provided price quotes and performance
information in 2007 for bulk storage applications. These 2007
prices are presented in Tables II and III. Table II shows the data
for generation applications, and Table III presents the values for
T&D applications.
V. RESULTS
The technique developed in Section III of this paper is applied
to the case study values and assumptions from Section IV. The
resulting costs, which are the costs added by the various storage
technologies to each kilowatt hour of electricity that is stored,
Fig. 3. Added cost (COE) vs. discharge time, 2.5 MW T&D application
operating 250 d/year.
are plotted in Figs. 1–4. Each figure shows the cost added as
the actual charge/discharge times are varied. For generation
applications (see Figs. 1 and 2), the systems are designed for
8 h discharge, and thus, the lowest cost is seen at 8 h, because at
shorter times, available capacity goes unused. Similarly, Figs. 3
and 4 are for T&D applications, designed for 4 h discharge time,
and the lowest cost is at 4 h.
The data for generation applications are for systems operating
year-round (250 d/year, Fig. 1), and operating only during the
POONPUN AND JEWELL: ANALYSIS OF THE COST PER KILOWATT HOUR TO STORE ELECTRICITY
TABLE IV
COST ADDED TO COST OF ELECTRICITY USING SYSTEM AS DESIGNED
Fig. 4. Added cost (COE) vs. discharge time, 2.5 MW T&D application
operating 100 d/year.
peak summer season (100 d/year, Fig. 2). Cost are substantially
higher for systems operating only part of the year, because the
system fixed costs are spread over a much lower number of total
kilowatt hour stored.
Data are also presented for T&D applications operating
250 d/year (see Fig. 3) and 100 d/year (see Fig. 4). Higher costs
are seen, similar to generation results, for the systems operating
fewer days per year.
Cost for battery generation applications are increased relative
to T&D applications because of the difference in storage time.
The same number of total kilowatt hour is stored each day by
each kilowatt of generation or T&D storage capacity, but the
generation system needs twice the energy storage:
Generation:
T&D:
1 kW × 8 h × 1 cycle/d = 8 kWh/d
Energy rating: 8 kWh
1 kW × 4 h × 2 cycles/d = 8 kWh/d
Energy rating: 4 kWh.
The T&D storage system cost, however, is increased by more
frequent replacement of batteries. Battery life is measured in
charge/discharge cycles; so, the life, converted to years, of a
battery in T&D service will be half that of a battery in generation
service. For example, a battery with 1500 cycle life:
Generation: 1500 cycles/1 cycle/d = 1500 d
T&D:
1500 cycles/2 cycles/d = 750 d.
Table IV summarizes the results for all systems operating as
designed.
VI. INTERPRETATION OF RESULTS
With installed capacity equal to 3% of the total installed electric generating capacity in the world, pumped hydro storage,
and its added cost to electricity of US$0.05/kWh, can be con-
533
TABLE V
US ELECTRICITY PRICES
sidered economically feasible for many applications. Indeed,
many more pumped hydro units would be built if there were
suitable locations for them; pumped hydro is limited to those
sites with locations for two large reservoirs at significantly different heights.
For battery and flywheel systems operating 250 d/year,
Table IV shows that at the present day prices, such systems can
store energy for about 3–12 times the cost of pumped hydro.
The costs for systems operating 100 d/year are three–seven
times the cost of pumped hydro.
Table V presents recent values for peak wholesale and average
retail electricity prices in the US [9], [10]. Pumped hydro’s
added US$0.05/kWh cost is less than one-third of the wholesale
peak, and almost equal to the average industrial retail price.
The lowest cost battery system studied adds a cost to electricity
approximately equal to the average retail price, while the highest
cost system adds a cost of four times the wholesale peak price.
Energy generated or purchased off-peak, when prices are lower,
stored in a pumped hydro system, and used at peak times, will
have a total price less than the peak purchase price. This is
probably not yet the case with battery and flywheel systems. If
prices in the US, however, were not capped, the wholesale peak
price would be much higher, and other storage technologies
would be more feasible. In Australia, for example, wholesale
prices are capped at AU$10/kWh.
VII. CONCLUSION
A battery storage system designed to operate 250 d/year with
one 8 h charge/discharge cycle per day adds US$0.18–0.64 to
the cost of electricity at 2006 prices. A system that charges and
discharges twice a day on a 4 h cycle adds about US$0.07–0.57.
The lowest value is less than recent US wholesale peak prices
and comparable to recent US average retail electricity prices.
The high value is much higher than average wholesale or retail
prices. Such costs will be justified for some applications, but
difficult to justify for many others.
Similarly, a battery storage system designed to operate only
during peak seasons, 100 d/year, on one 8 h cycle, adds
US$0.42–0.86 to each kilowatt hour stored. A system that
charges and discharges twice a day on a 4 h cycle adds about
US$0.20–0.64 to each kilowatt hour stored.
These costs are between 3 and 12 times the cost of conventional pumped hydro storage. Pumped hydro capacity now
equals 3% of the total world generating capacity, evidence of its
economic viability. Suitable sites for new pumped hydro facilities are, however, limited. If the costs of battery and flywheel
technologies continue to decrease, then, their operating costs
will someday approach that of new pumped hydro.
534
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 23, NO. 2, JUNE 2008
If wholesale electricity price caps in the US are raised or
removed, those prices at times will be many times higher than
they are today. Storage of low-cost off-peak energy for use at
on-peak times would become more economical because of the
higher savings achieved.
Twenty-two US states have passed renewable portfolio standards [11] and more are enacted each year. Wind and solar
generation produce only when the resource is available, which
may not be when the electricity is needed. This makes penetrations exceeding 10%–15% impractical [12]. Storage allows the
energy to be used when needed, allowing higher penetrations of
renewables and greatly increasing the value of renewable energy
generated off-peak.
From the comparison with wholesale electricity prices, it appears that the cost of electricity storage systems needs to drop
significantly before it will be useful for widespread load-leveling
use. Additional cost of energy, however, is only one component
of the economic justification for storage. Other issues, such as
deferral of transmission and generation facilities, and all others
listed in Section I must be considered in a complete economic
analysis. Market design and how markets will treat stored energy
must also be considered.
The costs added to stored electricity are highly dependent
on the system’s design parameters: number of discharge cycles
per day and number of operating days per year. Deviation in
operation from design values, as indicated by the high (left end)
values in Figs. 1–4, greatly increases operating costs. Great care
should be taken in designing such systems to optimize operating
costs.
Replacement period of batteries also has a crucial effect on
stored energy costs. The system design should insure that the
replacement period is proportional to the life of the power electronic (power conversion system) and balance of plant. Operation and maintenance costs are much less significant than capital
and replacement costs.
The technique used to evaluate storage costs in this paper
was implemented in a spreadsheet, which is available from the
authors.
REFERENCES
[1] Pumped Hydro Storage. (2006). Electricity Storage Association,
[Online]. Available: http://www.electricitystorage.org/tech/technologies_
technologies_pumpedhydro.htm
[2] W. Jewell, P. Gomatom, L. Bam, and R. Kharel. (Jul. 2004). Evaluation of Distributed Electric Energy Storage and Generation, Final
Report for PSERC Project T-21. PSERC Publication 04–25, Power
Syst. Eng. Res. Center [Online]. Available: www.pserc.org/cgi-pserc/
getbig/publicatio/reports/2004report/jewell_der_final_report_2004.pdf
[3] EPRI-DOE Handbook of Energy Storage for Transmission and Distribution Applications, EPRI, 2003.
[4] M. R. Lindberg, “Engineering economic analysis,” in Mechanical Engineering Review Manual, 8th ed. San Carlos, CA: Professional Publications, 1990, ch. 2, pp. 2–3.
[5] S. M. Schoenung and W. V. Hassenzahl, “Long- vs. short-term energy
storage technologies analysis: A life-cycle cost study,” Sandia Natl. Lab.,
Albuquerque, NM, Sandia Rep. SAND2003-2783, 2003.
[6] Energy Information Administration (2006). Annual energy outlook 2006
with projection to 2030. Report No. DOE/EIA-0383. [Online]. Available:
http://www.eia.doe.gov/oiaf/aeo/
[7] S. Blankenship, “Flywheel prototype to be demonstrated for frequency
regulation/grid stability,” Power Eng., vol. 109, no. 4, p. 46, Apr. 2005.
[8] Information from price quotes and performance data provided by energy
storage device manufacturers.
[9] Energy Information Administration (Nov. 2006). Wholesale day
ahead prices at selected hubs, peak. [Online]. Available: http://www.
eia.doe.gov/cneaf/electricity/wholesale/wholesale.html
[10] Energy Information Administration (Oct. 2006). “Average Retail
Price of Electricity to Ultimate Customers by End-Use Sector,”
Electric Power Annual with data for 2005. [Online]. Available:
http://www.eig.gov/cneaf/electricity/epa/epat7p4.html
[11] States with Renewable Portfolio Standards (May 2006). Pew center
on global climate change. [Online]. Available: www.pewclimate.org/
what_s_being_done/in_the_states/rps.cfm
[12] W. Jewell, R. Ramakumar, and S. Hill, “A study of dispersed photovoltaic
generation on the PSO system,” IEEE Trans. Energy Convers., vol. 3,
no. 3, pp. 473–478, Sep. 1988.
Piyasak Poonpun (S’06) received the Bachelor’s degree from the King Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand, in 1997, and the
M.S. degree in electrical engineering in 2006 from
Wichita State University, Wichita, KS, where he is
currently working toward the Ph.D. degree.
He is currently a Graduate Research Assistant at
Wichita State University.
Ward T. Jewell (M’77–F’03) received the B.S.E.E.
degree from Oklahoma State University, Stillwater, in
1979, the M.S.E.E. degree from Michigan State University, East Lansing, in 1980, and the Ph.D. degree
from Oklahoma State University, in 1986.
He has been with Wichita State University, Wichita, KS, since 1987, where he is currently a Professor of Electrical Engineering. He is the Site Director at the Power System Engineering Research
Center (PSerc), Wichita State University. His current
research interests include electric power quality and
advanced energy technologies.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement