March 19, 2000

ENERGY STORAGE TECHNOLOGIES

FOR UTILITY SCALE INTERMITTENT RENEWABLE ENERGY SYSTEMS

Alfred J. Cavallo, Consultant

289 Western Way, Princeton, NJ USA 08540

ABSTRACT In current conventional utility systems, fossil fuel or nuclear power plants must back up
intermittent renewable electricity generators if the utility grid is to have acceptable reliability. For low
levels of intermittent power, this can be achieved without difficulty or added expense. For high levels this
reliance on other generators becomes explicit and increasingly expensive, and transmission capacity also
becomes an issue. An attractive alternative to the conventional approach is one that relies on bulk or utility
scale storage of intermittent electricity to provide system reliability. A comparison of the currently
available storage technologies shows that the most cost effective and environmentally acceptable is a
compressed air energy storage (CAES) system and especially its more advanced derivatives (CASH, CAES
with Humidification and CAESSI, CAES with Steam Injection). For wind energy systems, both short term
and seasonal storage are technically and economically feasible.

Background
It is increasingly obvious that anthropogenic global climate change must be taken seriously, and that
carbon emissions must be reduced drastically if serious harm to the Earth’s biosphere is to be avoided.
This certainly will involve dramatically reducing or eliminating the use of carbon based fuels to generate
electricity and utilizing nuclear or solar energy systems instead. While nuclear power can technically be
used to replace fossil fuels, it has important disadvantages, most significantly the irreducible finite
possibility of a catastrophic accident. It would thus be desirable to limit its use to a minimum and satisfy
most or all of our electricity needs using solar power. However, if solar generated electricity is to be a
credible alternative it must have technical characteristics equal to those of fossil or nuclear power; that is, it
must be easily utilized in modern industrial state, and its cost must be reasonable. Since renewable
resources are generally diffuse, remote from major demand centers, and intermittent, the issues of
transmission and storage must be addressed; fortunately, it is easy to demonstrate that an affordable
technical solution to the challenge of electricity storage already exists.
The need for utility scale storage is illustrated by the current status of wind energy, which now supplies
about 15 percent of Denmark’s electricity. Wind turbines are by far the lowest cost and most successful
source of renewable electrical energy available today. This is due both to the superb quality of the turbines
developed over the past decade, as well as to far-sighted and effective public policy that mandates a
justifiably high price for wind electricity. These same policies do have some negative effects, however,
which up to now have not impeded the rapid increase of wind electricity onto the grid. Utilities, in most
cases, are forced to absorb the costs of transmission line and substation reinforcement and of insuring
overall system reliability. Given low wind turbine capacity factor (25-30 percent), transmission has already
become an issue in some areas, while system reliability is increasingly a problem as wind penetration
grows above 10 percent of average electricity demand, as it has in Denmark. In addition, a new system of
balancing charges designed to insure that supply balances demand in deregulated markets threatens to
penalize wind quite strongly [1].
One way to resolve these issues to the advantage of wind and other intermittent energy is to include
storage on the system in a way that recognizes the wind/storage plant as a unified entity: that is, the output
of the total system should be classified as renewable energy. This will resolve immediate transmission and
reliability issues as well as allowing wind in the not too distant future to supply up to about 80 percent of
total electricity demand [2].

Comparison Of Storage Technologies

Pumped storage, batteries, superconducting magnet energy storage, flywheel energy storage,
regenerative fuel cell storage and compressed air energy storage could be considered for bulk power
storage; a cost comparison [3,4 ] of these is listed in Table 1. The critical parameters for these systems are

1
March 19, 2000

the cost for power output (plant capital cost, $/kW) and the cost of energy storage capacity, given as the
cost per hour of operation at full output power (storage capital cost, $/kWhop). The systems are compared
with a 50 hour reservoir size which would allow intermittent wind energy to be transformed to baseload
power for a wind regime with a wind speed autocorrelation time of about 8 hours. Based on a wind
plant/CAES system simulation that included the wind speed autocorrelation time [5], this reservoir size is
reasonably adequate for short term baseload operation but far below what is necessary for seasonal storage,
so that the comparison understates the advantages of CAES.
To put this problem in perspective, it is useful to compare the energy density of a typical fossil fuel to
alternative storage media. Fuel oil has an energy density of about 38 GJ m-3; for comparison, a 25 kG
magnetic field has an energy density of 10 MJ m-3, a cubic meter of water at a height of 100 m, 1 MJ m-3, a
rechargeable gell cell battery, about 240 MJ m-3, compressed air (80 Bar), 8 MJ m-3 and a rechargeable fuel
cell (Innogy PLC), 120 MJ m-3. Fossil fuels have an immense advantage on this basis alone. If one also
considers their very low cost and ease of transportation and utilization, the advantages of fossil fuels would
appear to be overwhelming. Yet intermittent renewable energy, with a properly chosen storage system, can
in fact be competitive, both technically and economically as defined below, with fossil and nuclear
systems.
Technical competitiveness means that intermittent renewable energy systems with storage must have the
same forced outage and scheduled outage rates, as well as all other measures of power quality, as the best
fossil fuel or nuclear systems. Economic competitiveness means that the electricity market must be
designed so that cost of electricity delivered is affordable for consumers and profitable for producers and
equipment manufacturers. Given the advantages of fossil fuel systems (low installed capital costs,
relatively low fuel costs, lack of any cost assigned to the damage done by mining, transportation or burning
fuels), it is not realistic to assume that renewable energy can compete as the markets presently function.
However, with the excellent renewable energy technologies now available and the increased understanding
and awareness of the dangers of the alternatives, it is clear that the rules by which markets currently operate
must be adjusted to allow renewable energy to supply a much larger fraction of the demand.

Pumped Storage. Pumped storage (Table 1), with an overall efficiency of about 75 percent, is widely
used around the world. However, it is only economical in large installations (1000 MW), and the
aboveground reservoir has a significant environmental impact due to its size and dynamic behavior. In
addition, many regions do not have any suitable sites for storage reservoirs or have sites only in areas
where there is strong opposition to such a facility. Finally, the installed capital cost of aboveground
pumped storage is much higher than for a CAES system; seasonal storage, which requires two to three
hundred hours of storage capacity, is not economical. And while the environmental impact of an
underground pumped storage reservoir is minimal, the cost is high.

Battery Storage. Battery storage, with an overall efficiency of 75 percent, is also a possible candidate
(Table 1). While the plant capital cost ($/kW) is low, the storage capital cost is quite high, and the total
installed capital cost, even for advanced batteries, is extravagant. In addition, the volume of materials
needed for a utility scale facility raises environmental issues that are difficult to overcome. Certainly the
use of lead acid batteries, even in advanced systems, would be out of the question. Furthermore, the battery
system cost does not include the replacement cost. Clearly, the use of batteries in a utility scale storage
system is not realistic.

Superconducting Magnets. Large scale superconducting magnet energy storage systems (90 percent
efficiency) are still under development (Table 1), while small scale systems used for short term dropout
protection on critical equipment like computers are already deployed. Again, the very high storage capital
cost for these systems ($300/kWhop) makes these systems economically impractical for utility scale
systems. Also, the environmental impact of large solenoids and their associated unconfined magnetic fields
might be a problem.

Flywheels. Flywheels (70 percent efficiency) have long been used to store energy in rotating machinery,
and larger flywheels using advanced materials are under development. Once again, their very high storage
capital cost ($300/kWhop) indicates that while such systems may be useful in special applications like
automobiles, bulk electricity storage using flywheels is highly impractical (Table 1).

2
March 19, 2000

Regenerative Fuel Cells. The newest storage technology is based on the recently developed regenerative
fuel cell [4]. To charge the system, electrical energy is converted into chemical energy in two electrolytic
solutions in the fuel cell and pumped into storage tanks; during discharge the process is reversed. Fuel cell
electrodes function as a cation selective membrane; these are made of carbon fiber, a low cost and high
strength material. The electrolytes are concentrated solutions of sodium bromide and sodium polysulphide,
which are readily available commercially. System lifetime is estimated to be greater than 15 years; overall
system efficiency is about 65 percent.
The technology has many advantages. The system is modular so that it can be easily expanded and
easily repaired; tens or hundreds of modules are linked in series and parallel. Storage capacity is separately
adjustable from power output. The response time of the system is less than 3 seconds, so that applications
such as spinning reserve, load leveling, and distributed generation (peak shaving) are feasible.
Costs listed in Table 1 are based on the first large-scale system to be built, a 15 MW, 120 MWh facility
to be constructed in Wales (UK), and are expected to drop as more experience is gained; it already appears
to be competitive with battery storage systems. This appears to be a promising technology for certain
applications, but one that is likely to remain significantly (factor of 2-3) more expensive than CAES, even
with large reductions in the Cell plant and storage capital costs and with high fossil fuel costs for the CAES
system. This is a consequence of the relatively low capacity factor at which storage systems operate and
the much higher plant and storage capital costs of the Regenerative Cell system compared to the CAES
system.

Compressed Air. Compressed air energy storage (CAES) was invented in Germany in 1949, and a 290
MW CAES facility has been operating reliably at Huntorf since 1978. In the USA a more modern plant has
been in operation since 1991 at the Alabama Electric Cooperative in Macintosh, Alabama USA [6,7].
CAES is based on gas turbine (or jet engine) technology that has advanced enormously over the past
decade. A turbine is in principle a simple machine consisting of a compressor, a combustor and an
expander. A turbine extracts energy from a fuel in a simple thermodynamic Joule cycle of isentropic
compression of the input gas, isobaric heating of the compressed gas in the combustor, and isentropic
expansion and isobaric rejection of heat in the expander. Modern single cycle combustion turbines have an
efficiency of between 30 and 40 percent, so that about 60 to 70 percent of the output of the expander is
used to drive the compressor [8]; it is this energy (less any losses) that is available for storage in a CAES
system, as explained below.
CAES can be understood as interrupting this thermodynamic cycle; instead of injecting the compressed
gas directly into the combustor, it is stored in an underground reservoir. When power is needed, high-
pressure gas is withdrawn from the reservoir and the remainder of cycle completed.
A CAES system in its simplest form consists of a compressor, a turboexpander (a combustor and
expander), a generator and an underground storage volume such as a solution mined cavern in a salt
deposit, a capped porous rock formation such as a depleted gas reservoir, or a hardrock cavern or
abandoned mine. To charge the reservoir, power is supplied to a compressor which pumps air at a pressure
of about 80 bar into the underground storage reservoir. When power is needed the high pressure air is
withdrawn from the cavern and supplied with fuel to the turboexpander to generate electricity.
This system has many important advantages. Power generation is based on gas turbines, which are
simple, reliable, and inexpensive. The storage medium is air, which is readily available and free. The
turboexpander, which does not drive the compressor, has a very high ramp rate, so that the system can be
brought on line and respond to system changes very quickly; in addition, the heat rate is constant over a
wide range of output power. The compressor charging system is completely independent from the generator
and can be sized to match the wind resource and wind turbine array. In the US, geological surveys have
indicated that suitable underground conditions for CAES systems are found over about 80 percent of the
country, including those areas with good wind resources. Finally, the environmental impact of the
underground storage volume is minimal.
Two parameters characterize CAES system efficiency, a heat rate (HR, Btu/kWh) and an energy ratio
(ER), or the ratio of the input energy to the output energy. Since CAES uses fuel itself, the energy output is
greater than the energy input, and ER is less than 1. For example, the AEC CAES plant the HR is 4100
Btu/kWh and the ER is 0.82.
There are several ways to enhance the efficiency of the simple CAES system. The most direct is to
preheat the compressed air in a recuperator before it enters the turboexpander to extract additional energy

3
March 19, 2000

from the exhaust gas. This lowers the heat rate and reduces operating costs; it has been implemented in the
Alabama Electric Cooperative CAES plant.
One possible improvement [9] in the basic system is to generate steam, then inject it with the
compressed air from the storage volume into the turboexpander where it is heated and expanded. This type
of plant (CAESSI, CAES with steam injection) has a higher heat rate but a lower energy ratio, as well a
significantly lower installed capital cost.
Another possibility is to heat and humidify air from the storage volume in an air saturator, then pass it
through a recuperator before injection into the turboexpander. This plant (CASH, or CAES with
Humidification) has a lower installed capital cost and a lower heat rate than CAESSI, and appears to be the
most economical version of the CAES concept. Compared to the basic CAES system, storage volume is
reduced by a factor of 1.67 using CASH, resulting in a significant reduction in the total system installed
capital cost. A CASH plant might have an HR of 5000 Btu/kWh and an ER of 0.5.
The wind resource can vary significantly during the course of the year, in many cases being much better
in the winter or spring than in the summer, so that a system with a seasonal energy storage capacity would
be a great advantage. Costs of seasonal storage using a CAES or CASH system are compared in Table 2
with a 250 hour storage reservoir [10] CASH systems have an installed capital cost that is about 20 percent
lower than CAES systems for both solution mined caverns or porous rock reservoirs. Both systems are
technically and economically feasible possibilities for bulk electrical seasonal energy storage.

Recommendations
In order to insure that CAES systems can easily be adapted to wind industry needs, several issues must
be addressed. These are CAES siting potential, demonstration plants, and possible new regulations.

CAES Siting Potential. The first issue is CAES siting potential. Geological surveys have been done in
the USA that have identified regions where storage reservoir based on solution mined salt caverns, porous
rock or hardrock caverns could be located. Favorable underground conditions for one or more of these
types of reservoirs have been identified over more that 85% of the continental US, indicating that CAES
systems are indeed a reasonable bulk electricity storage option there. Similar results might be expected in
Europe, India or China, for example, but this needs to be documented.

Demonstration Plant. A demonstration plant that combines a CAES system with the appropriate number
of wind turbines should be built. One possible configuration would be a 200 MW wind/CAES baseload
facility; this would require about 575 MW of installed wind turbine capacity coupled to a CAES system
with a 225 MW compressor charge rate and a 150 MW discharge rate. Since the required amount of
nameplate wind turbine capacity already exists in several locations, only the CAES plant and any required
transmission upgrade needs to be financed at an estimated cost of $75 million ($500/kW).
While CAES plants are already operational in Germany and the USA, they typically are coupled to a
baseload rather than to an intermittent power plant. A demonstration plant would serve to resolve the
details of the control system that coupled intermittent wind energy to the high power compressors. Most
importantly, such a project would overcome the reluctance of a utility or company to be the first to build a
new type of installation by underwriting the risk inevitably associated with a unique type of effort.

Renewable Energy/Fossil Fuel Combination Plants. It is critical that the total integrated system
consisting of the renewable energy source and storage system, including those storage technologies that use
fossil fuels directly, be considered a renewable energy supplier. For example, following the guidelines in
PURPA (Public Utilities Regulatory Policies Act) in the USA, a power plant may be considered to be a
renewable energy facility (a qualifying facility) provided that the fossil fuel energy input is limited to 25
percent of the total annual energy input.
An illustration of this approach is given by the Luz solar thermal power plants[11] in California. These
use natural gas or fuel oil to generate steam in parallel with sun-tracking parabolic trough solar
concentrators. In this fashion, the plant could generate power reliably at times of maximum demand and
thus capture a premium price for its output. The Luz Company was forced into bankruptcy by low natural
gas prices in 1991; however, their plants were the largest and most economical solar electric technology
developed to date.

4
March 19, 2000

Using PURPA as a model, legislation allowing fossil fuel/renewable energy hybrid plants to be
considered as renewable energy facilities should be enacted elsewhere.

Conclusion
The basic CAES system can be considered a reliable and demonstrated technology and is ideally suited
for use with current wind turbine arrays; more advanced CAES concepts appear to have even greater
promise for use with large high penetration intermittent wind energy systems. A comparison with
alternative storage technologies indicates that CAES systems are to be preferred by a large margin both on
environmental and on economic grounds. With the appropriate governmental actions with regard to site
surveys, demonstration plants and legislation, CAES systems should be an important part of the effort to
reduce and finally eliminate our dependence on fossil fuels.

Table 1. Storage Plant Installed Capital Cost*

(from R. Schainker, EPRI, 1996 Power Gen. Conf., Orlando, FL, USA and

5
March 19, 2000

Innogy (Regenerative Fuel Cell))

Storage Plant Capital Storage Capital Hours (b) Installed COE (e)
Technology Cost Cost (full power) Capital Cost ($/kWh)
$/kW $/kWhop (ICC)
$/kW
CAES (a)
>110 MW 390 1 50 440 0.0613
(Large) 530 2 50 630 0.0675
50MW (Small)
Pumped Hydro
Conventional 1100 10 50 1600 0.119
(1000 MW)
Underground 1200 50 50 3700 0.187
(2000 MW)
Battery
(Target) (c) 120 170 50 8620 0.347
Lead Acid 120 100 50 5120 0.233
Advanced
Superconductin
g Magnet 1000 120 300 50 15120 0.5484
MW (Target)
Flywheel
(Target) 150 300 50 15150 0.565
100 MW
Regenerative
Fuel Cell 1500 150 50 9000 0.370
(15 MW) (d)
*Costs for the Fuel Cell are in 2000 Dollars, and all others in 1994 Dollars. According to the US
Department of Commerce, there has been a negligible change in Producer Prices from 1994 through 2000.
Thus, the quoted 1994 Dollar figures have not been adjusted.
(a) This capital cost is for the reservoir capacity per hour of full power plant operation, and is based on a
solution mined salt cavern storage reservoir and basic CAES cycle. CASH and CAESSI systems and
porous rock storage reservoirs have significant technical and economic advantages for wind energy
applications.
(b) Based on a wind speed autocorrelation time of 8 hours and baseload operation.
(c) Battery cost does not include battery replacement.
(d) Proprietary System; information from Innogy Technology Ventures, Ltd; costs are approximate based
on a 15 MW, 120 MWh system, and an exchange rate of 1.5 USD/UK Pound.
(e) COE (Cost of Electricity) comparisons are computed as follows:
For CAES Systems: COE ($/kwh) = ICC*CCR/(8766*CF) + EC*ER + HR*FC, where the heat rate (HR)
is 4500 Btu/kWh, the Fuel Charge is $5/mmBtu, the Energy Ratio (ER) is 0.49, the cost of electricity used
to charge the reservoir (EC) is $0.05, the capacity factor (CF) is 0.35 and the capital charge rate is 0.1;
For Other Systems: COE = ICC*CCR/(8766*CF) + EC/Efficiency

6
March 19, 2000

Table 2. Comparison of CAES and CASH for Seasonal Storage

Storage Plant Capital Cost Storage Capital Hours* Installed Capital

Technology $/kW Cost (full power) Cost (ICC)
$/kWhop $/kW
CAES 390 1 250 640
Solution Mined
Cavern
CAES 390 0.5 250 515
Porous Rock
CASH 350 1 250 511
Solution Mined
Cavern
CASH 350 0.5 250 431
Porous Rock
*Based on wind class 3 (350 W/m2) summer and wind class 5 (550 W/m2) spring wind regime.

References
1
Milborrow, D., 2000, “Trading Rules Trap Wind in the Balance,” Windpower Monthly, Vol. 16, pp 40-
43.
2
Cavallo, A. 1995, “High Capacity Factor Wind Energy Systems,” J. Solar Energy Eng., Vol. 117, pp 137-
143.
3
Schainker, R., 1996, private communication, Electric Power Research Institute, Palo Alto, CA, presented
at the PowerGen Conference, Orlando, FL USA.
4
Innogy PLC, 2001, Innogy Technology Ventures Ltd., Harwell International Business Center, Harwell,
Didcot OX11 0QA (www.innogy.com).
5
Cavallo, A., 1996, Storage System Size as a Function of Wind Speed Autocorrelation time for a Wind
Energy Baseload System, Proceedings of the European Wind Conference, Goeteborg, Sweden, pp 476-
479.
6
Schainker, R.B., Mehta, B. and Pollak, R., 1993, Overview of CAES Technology, Proceedings of the
American Power Conference, Chicago, IL, Illinois Institute of Technology, pp 992-997.
7
Ter-Garzarin, A, Energy Storage for Power Systems, Chapter 7, IEEE, London, UK, Peter Pergrinus Ltd.,
Redwood Books, Trowbridge, Wiltshire, UK.
8
Obert, E.F., “Thermodynamics,” McGraw-Hill, New York, London, Toronto, pp 478-490.
9
Nakhamkin, M., Swensen, E., Abitante, P, Schainker, R and Pollak, R., 1993, Technical and Economic
Characteristics of Compressed Air Energy Storage Concepts with Air Humidification, Proceedings of the
American Power Conference Chicago, IL, Illinois Institute of Technology, pp 1004-1009.
10
Cavallo, A., and Keck, M., 1995, Cost Effictive Seasonal Storage of Wind Energy, SED-Vol 16, Wind
Energy, Editors, W.D. Musial, S.M. Hock, E. Berg, Book No. H00926-1995, pp 119-125.
11
De Laquill III, P., Kearney, D., Geyer, M., and Diver, R. Solar Thermal Electric Technology,” 1993,
Renewable Energy: Sources for Fuels and Electricity, T.B. Johannson, H. Kelly, A.K. Reddy and R.H.
Williams, eds., Island Press, Washington, DC.