Energy Storage

Ann. Rev. Energy. |934.9.| 29
Copyright © |98¢by Annual Reviews Inc. All rights reserved
1Ï1VLÀ bÅLV/L1
Bent Sorensen
Institute of Mathematics, Physics and their Applications in Research,
Education and Society, Roskilde University Center, P.O. Box 260,
Roskilde, Denmark
INTRODUCTION
The demand for energy varies with time. One reason for this is that certain
human activities must be performed at defnite times. If such activities
require the conversion of energy, they lead to more or less rigorous time
patterns of energy demand. This is also true for energy demand not
associated with any activity, such as that for space conditioning. The
production of suitable energy forms from primary sources, however, may
most conveniently follow a diferent pattern from that of demand. If
variable, renewable energy sources are used, the level of energy production
cannot be controlled.
Storing energy for later use is a way of coping with these problems.
Diferent situations lead to the desirability of having a variety of storage
options, difering according to the form of energy needed, the quanti ties of
energy to be stored, the duration of storage, and the rate at which energy
can be fed into or extracted from storage.
Traditional energy-supply systems based, for example, on wood, coal, or
oil involve storage of fuel. Usually, storage takes place before the main
conversion step, but in a number of cases, primary conversion steps are
performed before storage (for instance, the conversion of wood to charcoal,
or the refning of crude oil).
A reference energy-storage system is the flled oil tank (Figure 1). Its
energy density is 42 MJ/kg or 37 GJ/m3. Losse� during storage are zero.
Rates of flling and emptying are very high, and the transfers are practically
loss free. Storage may take place at ordinary temperatures and pressures,
and the energy form and quality is essentially unchanged, no matter what
the duration of storage. Very few alternative energy-storage systems can
match these specifcations.
0362-1626/84/101(001$02.00
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Quick links to online content
Further
ANNUAL
REVIEWS
2 SØREMSEM
The applications of energy storage may be roughly divided according to
whether they are directed at the demand (the "load") or the supply system.
In the frst case, the time variation of the load is altered to suit the fow of
produced energy better. In the second case, intermissions in the source fow
are flled, and the output from the energy-supply system is smoothed.
Load management is usually undertaken for reasons of economy or
e
f
ciency. Here, the role of energy storage is to increase the load (by adding
energy to the store) when there is a surplus of energy relative to demand or
when such a surplus can be produced at low cost. The store can be drawn on
when the generating capacity is insuffcient to meet demand, or when
energy from the store is cheaper than directly produced energy. Load
leveling by using energy stores reduces the maximum capacity require­
ments for direct energy conversion equipment. For example, heat and cold
stores in dwellings may reduce the peak ratings of boilers and air
conditioners, vehicle energy stores may reduce engine size, and energy
storage in electric utilities may reduce or eliminate the need for peak-load
generating units such as gas turbines or diesel sets (using the most expensive
kinds of fuel).
Figure 2 shows typical heat and electric power load duration curves, i.e.
the fractions of time during which particular power levels are required,
ordered according to declining power levels. The efect of load leveling
4
� -
³
.Figure I The rate at which gasoline is transferred into the tank of the author's car
corresponds to an energy fow of about 30 MW.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
T
N
o
w
�
O
+
O
C
o
0.¿ 0.& 0b
0.0
Load dura1 |on cutvf
(1rac11on of 1!m÷J
ENERGY STRAGE 3
Figure Z Annual load duration curves for
heat (space heating and hot water) and
electric power (domestic, commercial and
industrial) for an industralized region with
temperate climate. The efect of energy
storage is shown by the striped areas.
Reprinted with permission from Jensen &
Sfrensen, Fundamentals of Energy Storage,
copyright !VbJ by John Wiley & Sons, New
York (1).
application of energy storage is to move part of the peak in the load
duration curve down into the valley.
For variable energy sources, such as direct solar or wind energy,
increasing energy storage capacity will make the conversion-plus-storage
system behave more and more like a fuel-based supply system. Figure 3
shows typical power duration curves for solar thermal collectors and wind
power generators, i.e. duration curves for the production of power from
these systems. Here, adding storage fattens the duration curve in the
#
Ê
E
�
Û.5
Frsc1íon m time duínQ
wích per exceem P
Figure J Annual power duration curves for
solar thermal collector and wind power gene­
rator located at favorable sites in
D
enmark
(a) without and (b) with energy storage
facility. The effect of adding storage is il­
lustrated by the striped areas. Reprinted
with permission from Jensen & Sfrensen,
Fundamentals of Energy Storage, copyright
1983 by John Wiley &Sons, New York (1).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
4 S0RENSEN
middle by absorbing power when production exceeds the load and
releasing it again in the opposite situation.
Short-term storage may be needed with variable energy sources entirely
for stabilizing reasons, if the excursions of production are unacceptable to
the users or to the energy transmission system. One example is the hot­
water temperature of a solar thermal system, which in the absence of even a
short-term heat store might exhibit unacceptable excursions. Another
example is the fickering of lamps and interference with telecommuni­
cations that may be caused by wind turbine electric power systems in the
absence of short-term storage. Long-term storage may be needed if the
seasonal variation of energy production is diferent from that of the load (as
it is, for example, when solar collectors are used to satisfy space heating
needs).
SYSTEM ASPECTS
An energy system may consist of one or more energy conversion devices,
routes of primary energy supply, energy storage and transmission com­
ponents, and interfaces to the fnal users of energy-demanding services. The
loads satisfed by a system may be those of a building, a village, a county, a
nation, or even a regon covering several countries. A total energy system
would be one satisfying all types ofloads, but a system of more limited scope
(e.g. an electric utility system) may also be sufciently closed to allow a
separate system analysis. Very few energy systems are completely closed;
most involve energy fows across their boundaries. This does not invalidate
the system approach so long as these energy exchanges can be counted and
treated as external parameters in modeling and assessing system
performance.
System analysis of present or future energy systems allows for evaluation
according to technical, economic, and social criteria, including, for
example, efciency, dependability, supply security, safety, and democracy in
control structure. Using such analyses in planning energy systems can help
to ensure optimal use of resources, as well as fexibility under future changes
in external and internal conditions. Energy storage components turn out to
be of particular importance in ensuring the adaptability of an energy
system, as I illustrate below.
Energy systems may include hybrid and combination design features. A
hybrid system is one with two or more diferent conversion systems (with or
without storage components) serving a common load. Examples of hybrid
systems are heating systems with (active or passive) solar collectors and
wood stoves, district-heating systems with seasonal heat stores (fed by
waste burning) and heat pumps, and vehicles with both compressed air
stores and fuel-based engines. The advantage of a hybrid system may be
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
ENERGY STORAGE 5
supply security (like that provided by adding a heat pump to a district­
heating system that has enough storage to' cover heat demands during
average but not particularly cold winters) or better utilization of standard
systems (e.g. the fuel-based engine of a vehicle may have a much lower
rating when an energy storage device with high-power capability is
included). The basic issue here is a trade-of between the cost reduction
associated with smaller capacity requirements and the i ncreasing cost
associated with adding multiple systems.
In combined systems one conversion device (with or without storage) can
satisfy two or more types of loads. An example is cogeneration or combined
heat and power production (CHP). Both are combined production systems
for electricity and heat. The term "cogeneration" is most ofen used for
industrial systems (in which process steam is usually the primary product
and electric power the residual one); "CHP" most often refers to electric
utility systems whose residual heat is used for district heating.
Systems of the kinds described above, as well as integrated systems with
several components of one or the other kind, may be characterized by a
system efciency, defned as the amount of free energy or exergy (1) input
needed for the best system that can be imagined, divided by the actual input.
The systems compared must give the same free-energy outputs in energy
forms suited for serving the same fnal purposes, and the energy must be
delivered at identical levels of power. The free-energy concept is used in
order to be able to deal with combined systems having output-energy forms
of difering quality. The system efciency is meaningful only when averaged
over time periods large enough to absorb the delays in the system, notably
those associated with energy storage. Otherwise, inputs and outputs may
have a
r
bitrary relationships. For example, a system with seasonal energy
storage should be assessed on an annual basis, with respect to effciency.
The incorporation of power levels in the defnition of efciency is
important, because otherwise it would often be easy to increase long-term
average efciencies if energy delivery could be spread out in an arbitrary
manner.
The system efciency for national energy-supply systems of indus­
trial
I
zed countries is at present about 25%, based on energy delivered to the
fnal user, and perhaps below 5% if based on the services derived from the
delivered energy. These efciencies are based on comparison with alterna­
tives using presently known technology only (2).
Individual energy-storage components of energy systems may be charac­
terized by their energy and power densities. Table I gives a survey of energy
densities. Maximum power densities are ofen more design specifc. For
example, they span three orders of magnitude for battery systems currently
studied (1).
Let me now illustrate some system aspects on the basis of specifc system
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
6 S0RENSEN
structures. My frst example is the energy system of a highly industrialized
and service/information-oriented society. Figure 4(a) shows what might be
the present energy system of such a society. The system is based on three
grid systems: an electric-power grid of systemwide coverage, a central
natural gas pipeline system, and a number ofIocal district-heating pipeline
systems. The electric grid has a central loop (the high-voltage grid) and a
number of local (lower-voltage) grids. The natural gas distribution system
difers from most current ones by not having a fnely meshed local part, but
only the main distribution system, which leads to large industrial users and
to local CHP plants. The reason for reserving natural gas for these uses is
that the advantage of combined heat and power production can be
1abIc1 Storage capacity
Energy storage form
Conventional fuels
Crude oil
Coal
Dry wood
Synthetic fuels
Hydrogen, gas
Hydrogen, liquid
Hydrogen, metal hydride
Methanol
Ethanol
Thermal-low quality
Water, 100°C ¯ 40°C
Rocks, 100°C ¯ 40°C
Iron, 100°C ¯ 40°C
Thermal-high quality
Rocks, e.g. 400°C ..20°C
Iron, e.g. 400°
(
....20°C
Inorganic salts, heat of fusion > 300°C
Mechanical
Pumped hydro, lO-m head
Compressed air
Flywheels, steel
Flywheels, advanced
Electrochemical
Lead¯acìd
Nickel-cadmium
Advanced batteries
Energy density
42,000
32,000
15,00
120,00
120,000
2,00-9,000
21,00
28,000
250
4050
�30
�1 60
�lo
>30
30-120
>20
6
40-140
�350
>40
37,00
42,000
10,00
10
8,700
5,00-15,00
17,000
22,000
250
100-140
�230
�430
�800
>30
�
15
240-950
>10
100900
�350
>30
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
(a)
Coal
~·'�
/
.
+
+
EMERGY STORAGE 7
Electric
� v ehicles
["
Natural gas - --,- �.
_¸
�
Õ1!6r u5t5 O\
natural gas
(b)
W¡^Q 0Þ6J__ l ctnîJðl
photovoltaic converters ¯
� �
&
`
\ Other uses of

¯ nyd¡cgeo
Elec tricity grid
¬

··¬¬¬
Hydrogen pipeline
-------. District heating line
� Fuel cell unit
[ Combined heat and power (CHP) plant
• t!vC11͹ battery S\OJt
o Community |Eç1 store
O IOC\V¡CUð\ users
•
do. WIî| SC\ðJ 0CU6£\Ot5
Figure 4 Outline of fossil fuel and renewable energy based energy systems with grid
distribution, (a) The mainly fossil system, shown schematically, has just one central
distribution loop and three local grid systems, (b) A possible future system based on renewable
energy sources is depicted similarly, A key to symbols is at bottom,
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
8 S0RENSEN
obtained with smaller environmental impacts than if coal or oil were
burned locally in the CHP plants. Some of the local systems have
community heat stores, so that their CHP plants can follow the electricity
demand and still meet the time-displaced heat load. One community has a
fuel-cell CHP plant rather than a combustion-type plant. A few individual
users have solar heat installations (rooftop collectors), but no storage,
because present hot-water storage technology is much more economical
with community-size stores (because losses are proportional to surface
area, not to volume). Natural gas is also stored, but centrally (for safety
among other reasons), e.g. in underground caverns.
In this system, central coal-fred CHP plants make up for the defcit in
electric power and sell some heat to a nearby district-heating system. Such
systems are common in Denmark and a few other countries. They are not as
efcient as the local CHP plants, because a large power plant produces
more associated heat than can be used close by. Thus, some ofthe potential
heat production cannot be used; furthermore, transmission losses in long
district-heating lines (up to 50 km in Denmark) may be signifcant. On the
other hand, the advantage of the large plant is that it can uSe the cheapest
fuels; at least this is the present view, because, it is claimed, sitting large
power plants away from population centers allows pollution limits to be
relaxed. Recent discussions oflong�range transport of power-plant efuents
with acidifying impacts (among others) indicate that this view may not hold.
These considerations strengthen the case for using natural gas in de­
centralized CHP plants. The likelihood of future reductions in heat
demands, as buildings become increasingly better insulated, works against
systems based on district-heating lines. The reason is that the cost of the
district-heating lines would become shared among fewer units of delivered
energy, and hence the price per unit of heat would go up. However, there are
other factors to consider, such as the reduction in the temperatures used in
district-heating systems that would allow low-temperature systems (e.g.
solar thermal) to be attached to the heat grid. With lower temperatures (say
50°C instead of 90°-110°C, as is customary today), cheaper plastic pipes­
which would not be able to withstand the higher temperatures-could be
used, thus again altering the economic calculation.
Community hot-water stores in connection with CHP plants have been
used in Denmark for a number of years. If there is an electricity surplus-as
there might be in an all-CHP system, for example, on a cold winter night-it
could be used in two possible ways: (a) Convert surplus power to heat with
a heat pump (placed at the community CHP plant and having a higher
coefcient of performance than small units for individual dwellings). (b) Use
surplus power to charge batteries for la\er automotive use in electric
vehicles. This possibility is indicated for one of the local grids in Figure 4(a).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
LMLRGY S1OKAGE 9
If fuels such as coal and natural gas disappear from the scene at some
future stage (e.g. owing to environmental concern), the system shown i n
Figure 4(a) i s fexible enough to adapt to new energy sources. The main
reason for this is that district-heating lines are not energy-source specifc,
unlike natural gas grids extending to individual houses. Natural gas
pipelines suited for future hydrogen transmission are coming into use, but
this option seems of i nterest only for the main distribution grid, because the
use of hydrogen as a fuel in individual households is not expected to be
feasible (for safety reasons).
Figure 4(b) shows a future version of the system considered in Figure 4(a).
Much more of the energy is now produced locally, from solar collectors
(thermal or photovoltaic) and semicentral plants such as wind converter
arrays or large photo voltaic i nstallations (3). The variable production of
these energy sources is dealt with by central facilities to convert surplus
electric power into hydrogen and to store the hydrogen. (Again, safety
considerations forbid local storage.) The local CHP plants used when the
renewable and variable sources are insuffcient may be fuel cells, to
maximize efciency. Heat stores make up for insuffcient heat production
from the renewable energy systems; if the stores are also low, the CHP
plants may be used with heat pumps (or directly as heat plants, if this
situation is too rare to warrant additional investment). If the area incl udes
agricultural activities, an alternative to hydrogen may be methane (the
main constituent of natural gas) derived from biogas by the removal of
carbon dioxide, for which simple processes have been developed (4).
As the density of settlement and of energy use decreases, the grid systems
one by one become uneconomical: frst district-heating lines, then gas
pipelines, and fnally electricity-carrying grids. In remote areas, away from
all of the grid systems, energy storage is a means of securing stand-alone
systems with a high degree of dependability. Many such areas are now
served by diesel electric generators, boilers, and motors, all depending on
fuels that must be transported to the areas. The diesel sets have failure rates
so high that it is customary to install capacity equal to as much as twice the
peak demand, if all loads are considered essential. It is clear that storage
systems could be viable alternatives.
A more versatile stand-alone system for a rural area is i llustrated in
Figure 5. The primary energy sources are all renewable: wind power, solar
radiation, and biomass, such as residues from agricultural activities (mainly
nonligneous material) or ligneous material from silviculture (wood fuel).
¸ The-ind turbines and solar collectors accommodate the energy demand
when they can, and the biomass-based systems provided backup through a
fuel-cell system with waste-heat utilization. The gaseous fuel is producer gas
(mainly hydrogen), from the ligneous fraction of the biomass, or biogas
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
10
SØRENSEM
(methane and carbon dioxide, of which the latter may be removed before
reaching the fuel cell) from the nonligneous biomass. The intent of this
system is to avoid storagein the wind and solar parts of the system because
such storage is at present very expensive. (If hydroelectric power systems
are locally available, they can be operated in combination with the variable
power systems, and the need for energy storage may in the way be avoided.)
Instead, the biomass fuel may be inexpensively stored before conversion to
gas, and the rate of gasifcation and the following fuel-cell conversion can
occur as dictated by the load minus the solar and wind energy
contributions.
The above examples illustrate but a few of the many combinations in
which energy storage can dramatically change the performance of the
system. In some cases, the distinction between storage and backup systems
is unclear. For example, in pumped hydroelectric storage, water is pumped
upward when there is surplus power from nonhydroelectric units and let
down through turbines when the non hydroelectric units alone are unable
to accommodate the demand. However, a conventional hydroelectric
power system with elevated reservoirs may function in precisely the same
way as the store, but without any upward pumping. This is possible if the
hydroelectric system normally covers a fraction of the load such that it may
be reduced when nonhydroelectric units generate a surplus, and if the
turbine capacity is large enough for the hydroelectric system to take over
any defcit in production by the other units. Thus, the hydroelectric system
is operated as backup but really is a store (5).
Wi
nd ****��
**
��
nonoII
conversion
Solar r a
d ia tiOn ~~� µµgg
Ligneous Gasifier
biomass store

Non-Ii gneous
biomass store
r
./ Fuel
ceII
, EIec1rí¢·!y
¬'^� Heat
Figure J Stand-alone energy system based on renewable sources of energy. Only the biomass
sources can be stored before conversion. They go through two conversion steps, being
converted first to gas and then to electric power and heat in a fuel-cell unit(s). Local
distribution systems for power, heat, and gas (in case the fuel cells are small, decentralized
units) may be used, although no connection to a central grid system is assumed. Most likely,
however, the fuel cells are located where the gasifers and digesters are.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
EMERGY STORAGE 11
INDIVIDUAL STORAGE CONCEPTS
This section gives a brief account of the status of most currently
contemplated energy-storage concepts.
Mechanical Energy Storage
Gravitational energy storage with elevated water reservoirs, mentioned
earlier, is a fully developed technology, used both for backup operation (5)
and (with two-way turbines for upward pumping and downward generat­
ing modes) for storage. The use of hydroelectric storage is limited only by
the availability of suitable reservoirs and by competing concerns, such as
preservation of natural environments and mountain-area wildlife habitats.
The scope for expanded use of hydroelectric storage thus may be fairly
limited. A number of short-term stores are being considered or have been
constructed in mountainous regions throughout the world, but only in
sparsely populated areas (such as Norway) has long-term (seasonal) storage
of such a nature been considered feasible. Alternatives with an underground
lower level and lake or sea water as the upper compartment have been
considered (6, 7). Only when natural cavities or very cheap excavation
methods can be used, however, are underground pumped hydroelectric
systems economically feasible for energy storage at present.
The energy density of a typical pU!lped hydroelectric storage system is
very low, as shown in Table 1. The expense of underground excavation
therefore suggests that other storage forms using rock caverns and having
higher energy densities than pumped hydroelectric storage should be
considered. The elastic energy of compressed gases oters such a possibility,
if the cavern can be made to withstand pressures 40-70 times that of
ambient air. Figure 6 illustrates some compressed-gas storage concepts. At
present, only one full-scale compressed-gas store is being operated (8). It
uses air and a salt dome excavation of the kind illustrated in Figure 6(a).
This system is now operated in a hybrid mode, with natural gas burners to
boost the pressure before the hot air is used to regenerate electricity in a
turbine/generator. A second-generation compressed-air storage system is
illustrated in Figure 7. By recovery of the heat generated by compression, it
avoids use of fossil fuels to preheat the air. If the scheme is to work in normal
situations of storing and retrieving energy at different times, heat stores
should be added. Schemes of this type are contemplated in a number of
countries for load-leveling operation where they are calculated to be viable.
The currently operating plant is only marginally viable, since it returns just
1.2 units of power for an input of 1 unit of electric power plus 1.6 units of
fuel. However, the addition of a heat exchanger is expected to raise the
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
12 S0RENSEN
overall conversion effciency to that of a base-load power plant (8). Since the
cost of the compressed-air plant with a saIt cavern storage compartment is
considerably lower than that of a base-load power plant, this storage
concept looks promising.
Because salt caverns are found near the surface only in modest numbers,
the more expensive rock cavern possibility illustrated in Figure 6(b) has also
been contemplated, as has the use of aquifers, illustrated in Figure 6(c).
Aquifers are layers of high water permeability, bounded by non permeable
layers. They are less costly to use than rock cavities, but their precise
locations are hard to determine before drilling, so costly exploratory boring
is considered necessary (9).
Kinetic energy storage has been used throughout history, as inertia in
linear as well as rotational motion. The potter's wheel (Figure 8) is a
fywheel with a disc of constant thickness, while fywheels used in early
steam engines were thin-rim constructions. However, the best utilization of
material strength is obtained with a disc of variable thickness, for which the
stress is the same everywhere (1). This is the kind of shape shown in Figure 9.
An alternative would be to maintain constant thickness but to use layers
made of diferent materials to obtain constant strength as a function of
distance from the center.
$uttace
(4I
|ÞI
PermeabIe
|ayer
WB\J$
ïcI
Figure 0 Underground compressed air storage (a) in salt cavity, (b) in rock bed, and (c) in
encapsulated aquifer. Reprinted with permission from Jensen & Serensen, Fundamentals of
Energy Storage, copyright 1983 by John Wiley & Sons, New York ( 1).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
LOOlCDl
¬
Proc8S5
heoI
Aìr ìOlet
I�
EMERGY STORAGE 13
Ll00lt|0
mOlOr!
QenerO\Or
H-
Alr Oull0l
exchOnger
bÛ"L
YClv05
to cOvi\y
tígurr / Compressed-air storage facility with heat recuperation and no fuel input.
Compressors are denoted by Cand turbines T. The subscript H stands for high pressure, L for
low pressure. Reprinted with permission from Jensen & Sorensen, Fundamentals ofEnergy
ðl0r0gr, copyright 1983 by John Wiley & Sons, New York ( 1).
Flywheel storage based on advanced materials technology is still in the
development stage. Advantages include rapid access to stored energy, high
power output, small cycle losses, and energy densities considerably higher
than for compressed gases. New magnetic suspension techniques, using
permanent magnets, and operation in a near vacuum may remove friction
problems, so that a charged fywheel can be lef in motion for several
months without any signifcant loss (10). Problems in the development of
fywheel concepts include high weight (in vehicular applications), high cost
for the advanced concepts, and safety considerations. For stationary
applications, the safety under failure conditions can be highly improved by
underground horizontal installation. Typical rotational speeds would be
three to fve revolutions per second. For short-term storage of electrical or
mechanical energy, fywheel technology appears quite promising, but its
economic viability cannot yet be precisely assessed.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
14 S0RENSEN
Figure ò Potter's wheel (from 1 1).
Electromagnetic Energy Storage
Apart from truly small-scale energy storage in capacitors, interest in
electromagnetic energy storage is concentrated on superconducting mag­
netic storage. Energy densities of about 100 MJ 1m ¯ are envisaged. The cycle
efciency will be not 100% but rather around 85%, owing to the energy
required to keep the store at cryogenic temperatures (the value of which
depends on the materials used). Only type-II superconductors can be used,
because the superconducting property of type-I superconductors disap­
pears in the presence of even modest-sized magnetic fel
d
s.
Superconducting storage systems with capacities of about 109 J are in use
for special purposes, such as at the European Particle Accelerator Center
(CERN) in Geneva, where pulsed acceleration would otherwise require
Figure 9 Constant stress fywheel shape. Reprinted with permission from Jensen &S0rensen,
Fundamentals of Energy Storage, copyright 1983 by John Wiley & Sons, New York (1).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
EMERGY STORAGE 15
Iigure l0 Superconducting magnetic energy storage at CERN, Geneva.
excessive power input to electromagnets (Figure 10). Utility applications
are in the research stage; systems with capacities up to around 1014 J are
being considered (12). Such systems would be sited underground to permit
easy shielding of the magnets.
Storage of radiation energy is in principle performed in lasers. Although
this type of storage is not considered for general use, it might play a role if
radiation is used as an energy transmission technology. There are no such
plans at present ( 13).
Organic Fuels
Biomass, stored as living plants, harvested plants, and fossilized organic
material, constitutes the main fuel of the present energy-supply system. All
of these biomass fuels can be stored before conversion. Plant material such
as wood is the dominating source of energy in the rural Third World
(Figure 1 1). It can be stored as wished except for tree species prone to attack
by boring insects (14). Fossil fuels dominate the energy picture in urban and
industrialized regions all over the world. The main concerns are rarely over
energy storage (although there are risks associated with liquefed gas
storage and oil depots, which may catch fre), but rather over the question of
supply of these fuels from geographically concentrated sources, and over
the environmental impacts of burning fossil fuels.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
16 S0RENSEN
In using nonfossilized biomass sources for energy purposes one should
always keep in mind that biomass is also a store of nutrients and potential
raw materials for several industries, and that biomass is frst of all part of
the global ecological system. Any use of biomass sources should therefore
be part of a carefully managed ecosystem.
The sources of biomass for storage would then not be natural vegetation,
but rather agriculture, silviculture, and animal husbandry, i.e. managed
ecosystems. There is signifcant scope for using biomass in a renewable
mode, for energy conversion as well as storage, and several such schemes
are economically viable at present. However, combined production offood,
timber, pulp, and bioenergy would shift the emphasis away from easily
storable biomass components for the energy sector (wood, grains, etc), and
would place increased emphasis on residues from food production serving
as feedstocks for energy and raw material uses. The direct storage of such
residues may be less convenient, and primary conversion to more versatile
fuels (such as charcoal, compressed pellets, or alcohols) could improve the
possibilities for transport and long-term storage of biomass for energy
purposes. Biomass collection and transportation costs are reduced by
combining energy production schemes with agriculture or forestry.
Schemes involving such primary conversion before storage are being
investigated on a fairly large scale in several parts of the world. Some seem
Figure 1I Masai woman just returned with a newly collected bundle of frewood (Great Rif
Valley, East Africa).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
ENERGY STORAGE 17
technically and economically viable or nearly viable, while others need
further development. Figure 12 gives an overview of techniques for
converting diferent types of primary biomass sources to fuels that may
replace present fossil fuels in stationary and nonstationary applications.
Alcohol fermentation is probably the most common conversion tech­
nique, if alcohol for drinking is included. However, the energy balance of
the standard production technique is poor, owing to heat requirements in
the distillation process. Economic viability as an energy product therefore
depends on fnding alternative ways of separating water and ethanol, or
hybrid schemes in which other methods replace the last and most energy­
expensive distillation step. Such schemes have been demonstrated on a
laboratory scale (15-17). Environmental efects of ethanol fermentation
include those caused by acids or enzymes used in the initial hydrolysis
process for cellulosic materials (cf Figure 12). In the combustion step,
nitrous oxides, aromatics, and aldehydes are formed in larger quantities
than in gasoline combustion (18). However, the high octane number of
ethanol makes lead or other antiknock agents superfuous.
The other main road from biomass to alcohols is gasifcation by heating
in oxygen-poor environments. The producer gas formed in this way (cf
CELL-FREE
CULTURES
I
Hvdron prouction
HYDROGEN
ANIMAL SUGAR-RICH
MANURE MATERIAL
I
EX1fa1lon
I
SUGAA
/
� ::=[
HEAT 7
METHANE DILUTE AL0DHDL
BIOGAS
I
Distillation
I
ETHANOL
THE BIOHEMICAL ROUTES
CELLULOSIC MATERIAL

WOOD
I
-
-
-
-
-
Fiber crushing
w-
/
HVdrOIYSiS __
LIGNIN
'-W/�y
PRODUCER GAS
HIGH-TEMPERATURE
I
HEAT
Syntheis
1-"
SYNGAS METHANOL
THE THERMOHEMICAL ROUTES
Figure I2 Main routes for conversion of biomass to synthetic fuel. Reprinted with permission
from Jensen & Sorensen, Fundamentals ofEnergy Storage, copyright 1983 by John Wiley &
Sons (1).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
18 SØREMSEM
Figure 12) may be cleaned and piped to users, or converted to a liquid fuel
such as methanol by a shift reaction (forming hydrogen) followed by a
synthesis reaction (1). Current methanol production is based on natural gas
and not on wood gasifcation (19). Prototype wood-to-methanol conver­
sion plants are underway in countries with easy access to wood scrap from
forestry operations (e.g. Canada, Sweden).
Economic viability is already established for a set of fermentation
processes usually denoted "anaerobic digestion." They use a wet process to
convert biomass to a mixture of methane and carbon dioxide ("biogas"),
from which the carbon dioxide is easily removed before compression (4).
Compressed biogas is used for automotive purposes, as illustrated in Figure
13, while direct biogas uses include process heat (e.g. for cooking) and
electric power generation (e.g. using diesel engines). Early biogas instal­
lations used animal manure as feedstock, but many other biomass sources
are suitable, including straw (20) and other agricultural residues, with or
without manure complements. Inhibiting agents include metal saits,
antibiotics, soluble sulfdes, and high concentrations of ammonia (21). The
cultures of fermenting bacteria require an initial period of adjustment to a
given composition of biomass feedstock, and if the feedstock composition
varies over time, diminished gas production may result. While simple one­
compartment digesters with labor-intensive operation requirements are
Figure IJ Diesel tractor using compressed biogas (photo courtesy of D. Stewart, Invermay
Agricultural Research Center, New Zealand).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
ENERGY STORAGE 19
common in China, India, and Pakistan, fully automated biogas plants have
been constructed in countries such as New Zealand, Australia, the United
States, and Denmark. There are hardly any environmental hazards, and the
residues from digester operation have high value as fertilizer and contain
strongly reduced pathogen populations (21). For this reason, anaerobic
digestion is used in sewage treatment.
As a stored fuel, even compressed biogas is of course less versatile than
liquid alcohols, and this is the reason for trying to develop these alternatives
despite the economic viability of biogas.
Hydrogen is a storable fuel, which, as mentioned previously, may arise as
the result of biomass gasifcation. However, it can also be produced directly
from electric power (by electrolysis) or by thermally induced reactions, and
perhaps in the future by direct photolysis in suitable membrane systems.
Safety problems are higher than those of methane gas, but hydrogen holds
the potential for large-scale storage in facilities located away from
population centers, and it may play an important role in future energy
systems, as discussed previously.
Electrochemical Energy Storage
Battery energy storage and regeneration of electric energy may be
characterized as having medium energy-storage densities (Tables 1 and 2).
Some battery types have good power densities, but all have lifetimes shorter
than most alternative storage devices. Versatility in design, for instance by
modifying specifcations according to type of application, and in particular
the possibility of small unit sizes, has led to widespread use of batteries for
remote and stand-alone systems, such as fashlights, watches, computers,
security and emergency power supply, and automobile starters. Lead-acid
batteries are most common, although alkaline electrolyte batteries are
being used for .purposes for which high charge retention and long life are
important.
Major areas of battery development are for automotive and power utility
uses. For stationary applications, low cost is more important than weight,
and here lead-acid batteries are emphasized, pariicularly since cycle life
with new, tubular cell designs has been increased from 300 to nearl y 2,000
cycles in recent years (23). Figure 14 shows part of such an installation used
in connection with a photovoltaic/diesel power system in a region not
connected to any central utility grid. The system is not economically viable
at current fuel prices.
Alkaline electrolyte batteries have energy densities up to 45 Wh/kg (as
compared with 35 Wh/kg for the best lead-acid batteries). Their power
densities are 200-600 W /kg (for nickel-cadmium batteries) and 65-90 W /kg
(for nickel-iron batteries), as compared to 20175 W /kg for lead-acid
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
N
0
Table 2 Characteristics of selected batteries (likely achievable performance)"
V
Lead- Nickel- Nickel- Nickel- Iron- Zinc- Zinc- Sodium- Lithium-
�
�
Characteristic acid cadmium iron znc air air chlorine sulfur sulfur
m
Z
V
m
Electrolyte H2SO4 KOH KOH KOH KOH KOH ZnClz Beta- LiCljLilj
Z
A1203 KI
Voltage (V)
Open circuit 2.05 1.35 1.37 1.71 1.27 1.65 2.12 2.1-1.8 1.9-1.4
Discharge at 2-hr rate 1.9 1. 2 1.2 1.6 0.7 1.2 1.85 1.7-1.4 1.3-1.0
Energy efciency, charge-discharge (%) 75 70 <60 75 4 55 65 70-75 75
Specifc energy (Wh-kg)
I-hr rate 24 28 40 70 50 80 120 120 -140
5-hr rate 40 30 55 75 80 10 150 140
Energy density (Wh-dm3) I-hr rate 70 60 100 140 80 80 180 170
Specifc power (W /kg)
Peak 120 30 4 400 60 100 280 240 20
Sustained 25 140 220 20 50 120 140
Life (to 80% discharge) cycles 50b 200 200 350 200 10 -10 200 20
Recharge time (hr) 5- 8 4-7 47 3-6 4-5 5-8 5 7-8 5
Operating temperature (DC) -20 -30 10 -30 0 0
0 300 430-50
+50 +50 50 +40 +50 +60
Existing Under development
Ambient temperature High temperature
"Source: Based on Jensen & Tofeld (22).
b Conventional cells.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
ENERGY STORAGE 21
batteries. Nickel-iron batteries have the best currently available cycle life
(up to 5,000 cycles) (23).
Electric vehicles (buses, vans, passenger cars) for urban transport are in
limited use in a number of countries, and large development programs are
in progress in the United States, Japan, and Europe (1). In recent years their
range has increased from 50 km to over 100 km, and speed and acceleration
characteristics have been improved considerably in cases where they are
important. The cost of an electric vehicle is n
�
higher than that of a fuel­
based one, beyond what limited production series can account for, and the
energy cost is nearly at the break-even point. The reason for this is the better
energy utilization of electric motors as compared with fuel engines (around
90% efciency as compared with 30% for the best diesel automobile
engines). This offsets the power station losses during production of
electricity for battery charging (the losses are around 60% of the fuel heating
value for pure power-generating fossil power plants, plus 5% transmission
losses relative to the fuel input) and almost all of the battery losses (cycle
losses of 20%-30%).
Hybrid concepts such as fuel-battery, compressed gas-battery, and fy­
wheel-battery combinations are being investigated. They are typically
expensive, but savings in battery or engine rating can be expected if the
other storage system can handle all high-power-level situations.
High-temperature batteries and other special types of batteries are being
Figure I4 Lead-acid battery storage for ûsolar photovoitaic system in a Saudi village.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
22 S0RENSEN
constructed. As shown in Table 2, they would be suited for specifc
applications, such as in cases where both energy and power densities must
be large. It is too early to consider the economic viability ofthese concepts.
The main obstacle to automotive uses of batteries is inconvenience
caused by battery weight and space requirements, and driving performance
diferent from that of fuel-based vehicles. Environmental benefts for urban
use include the displacement of air pollution and toxic waste accumulation
from the area traversed by the vehicle to power plants and battery
manufacturing and recycling facilities, which may be located in remote
areas.
Chemical Energy Storage
Phase changes and chemical reactions ofer possibilities for energy storage,
if reactants capable of engaging in exothermic reactions can be separated
and stored. Some of the reaction schemes that have been investigated are
methane-to-hydrogen reactions (operating at high temperatures) (24),
hydrogenation of carbohydrates (25), detachment of ammonia from
ammoniated salts, and metal hydride-to-hydrogen reactions (26). For low­
temperature heat storage, incongruent melting (i.e. melting combined with
dissociation) of salt hydrates has been extensively studied over a number of
years (27). Table 3 gives latent heats of some such reactions, and Figure 15
shows the variations in stored energy around the phase-change tempera­
ture for the most commonly used substance, Glauber's salt with extra water
added (28). Despite development eforts, the system is not in widespread use
today. It would ofer possibilities for seasonal storage of low-temperature
heat in connection with solar heating systems, which, however, are at
present most abundant in climatic regions with little need for seasonal
storage. Solar thermal systems for use in regions with a strong seasonal
anticorrelation between heating needs and solar radiation are not econo­
mically viable at present, and the addition of a seasonal heat store to the
system would not remedy the cost imbalance.
1ablc3 Characteristics of salt hydrates
Hydrate
CaClz'6HzO
Na2S04' 10H20
NazCO, 'lOH20
NazHP04 ' 1 2HzO
NazHP04' 7HzO
NaZSZ03 • 5HzO
Ba(OH)2' 8H20
Incongruent melting point,
TmeC)
2º
32
33
35
48
48
78
Specifc latent heat,
AH(MJm-3)
281
342
360
2Û5
302
346
655
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
EMERGY STORAGl 23
Salt hydrates also form the basis for storage in the "chemical heat pump"
concept. The energy associated with crystal water incorporation and
dissociation is controlled by the type of scheme shown in Figure 16,
involving a closed system with the possibility of isolating water in a
container separate from that of the salt. Early systems used sulphuric acid,
which caused severe corrosion problems and rendered the system un­
suitable for household heat-storage applications. Present development
efforts are concentrated on salts such as sodium sulphide, and two
prototype systems have been operated in Sweden (30). The storage capacity
is 312 kJ per mole, for Na2S· 5H20. Combined solar thermal and chemical
heat pump systems for residential application currently ofer no economic
advantage over solar thermal systems with hot-water storage.
Þ 20 40 b0 0 00
00~~�~�~�
ZÛ
b00
J60
b0
¬
E
kWh
¬
40
Æ
-;
\0
30
50
Tempersture,"C
Figure IJ Heat-storage capacity as function of temperature for Glauber's salt-water
mixtures (29).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
24 S0RENSEN
Heat
load
T ; 65"C
T< 55·C
F low diretion dUring
disharg
Water vapour at
_
presure-lo Nm-2 (0.01 atm.)
r=
Valve
Na,S
T -60·C
A
�
Charge
T-l0·C
B
Soil
evaporator
Figure 16 Schematic picture of solar thermal system with a chemical heat pump as energy
storage facility. Connection to the store can be switched between solar collector and building
heat load. Reprinted with permission from Jensen & S0rensen, Fundamentals oj Energy
Storage, copyright 1983 by John Wiley & Sons (1).
Heat-Capacity Storage
The simplest way of storing heat-also the only way currently being used
on a noticeable scale-is by the plain heat capacity ("sensible heat") of
materials such as water, rock, soil, or metals. Table 4 gives a few typical
examples of storage energy densities. Traditional building styles in many
parts of the world use buildings' thermal masses as mediums of diurnal heat
storage for temperature leveling. Active solar thermal systems use heat
stores, which are most often hot-water systems. Small hot-water stores are
also routinely installed in connection with fuel-based boilers with the
purpose of decreasing the number of starts and stops and the associated
preheating losses in chimneys.
However, heat-capacity stores lose energy to the surroundings, and
medium-term storage is possible only with insulated containers. Long-term
storage is not feasible with any amount of insulation for small systems, but
for larger systems it may become practical, owing to the improved volume
(storage capacity) to surface (heat loss) ratio (3). Communal hot-water
systems are at present used and considered economically viable for district­
heating systems based on CHP plants. The scheme is to let combined
production levels be determined by the electric power load (and thereby
avoid storage of electric energy), and to use the heat store to compensate for
the usual mismatch between power and hcat loads. This works only when
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Table 4 Heat capacities of various materials'
Materal
Solids
Sodium chloride
Iron (cast)
Rock (granite)
Bricks
Earth (dry)
Liquids
Water
Oil ("thermal")
Sodium
Diethylene glycol
Temperature interval
(0C)
< 80
< IS00
< 1700
O-lOO
-S0-330
98-880
-10-240
Mass specifc heat
(kJ kg- 1 0C-I)
0.92
0.46
0.79
0.84
0. 79
4.2
2.4
1 . 3
2.8
Volume specifc heat
(MJ m-3 °C-1)
2.0
3.6
2.2
1.4
1 .0
4.2
1 .9
1 .3
2.9
• All quantities have some temprature dependence, Standard atmospheric pressure has been assumed, that is, all heat capacities are C;s,
b Less for granulates with air-flled voids.
' At 10°C.
d At 70°C.
Heat conductivity
(Wm- 1 0C- 1)
9
b
"
70-34d
2,7b
0.6
1 .0
0,6
0.1
85'-60d
�
�
o
>
'
�
t
V
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
26
SØRENSEM
the total loads are consistent with CHP output, and usually these plants
must be operated as pure power plants at intervals during summer, and
occasionally as pure heat plants on cold winter nights. Sometimes the hot­
water stores are also used to cover heat loads during electric-power peak­
load periods (a few hours a day), in order to run the CHP plant as a pure
power plant and thereby to gain a few percent in conversion efciency (31).
The other application of communal hot-water stores is for solar thermal
systems in high-latitude areas. Figure 17 shows the results of operating such
a system, with 61 0m3 of water and 1 20 m2 of concentrating solar collectors,
to heat an ofce building with 500 m2 of foor area in Studsvik, Sweden (33).
The store stretches six meters underground, with foating insulation on top.
Temperature stratifcation is seen to play an important role and may be
said to replace downward insulation. In most such schemes the heat
capacity of the surrounding dry soil contributes to the storage capacity. A
larger system of 104 m3 of hot water serving 2600 m2 foor area in 55 single­
family dwellings has been operated in Linkoping, Sweden, since 1979. The
experience so far has indicated a wide scope for applications of such heat­
storage systems, which are relatively inexpensive and may be operated with
a variety of heat sources, once district-heating lines have been installed.
Heat storage in industrial processes is also gaining in reputation. Steam
stores have operated for years, and total energy planning with heat
cascading (at declining temperatures) is becoming common in energy­
intensive industries. Since heat at diferent temperatures is used several
places in many production lines, but not necessarily continuously, the scope
is considerable for short-term storage at a number of operating tempera­
tures (34).
Storage in connection with process-heat supply is also of importance for
0.sçnacq|oq
uu- Varch

5
¤ �� U � U��
30 40 b0 60 70 30 40
LHðtg| rg
Varch¯Auqost
b0 ¤O 70
Figure I / Temperature profles in Studsvik hot-water store, during discharge, August to
March (A to M, left), and during charge, March to August (M to A, right) (32).
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
ENERGY STORAGE 27
solar cookers, being introduced in many low-latitude countries. The time
displacement between solar collection and normal meal preparation hours
makes a storage system operating at temperatures of lOoo-300DC essential
(35).
Cold storage by heat capacity and phase-change energy has been used in
both industry (dairy product treatment) and households (air conditioning),
normally based on ice formation. Community-size ice-storage systems
seem to ofer the best possibility for economic viability (36). Individual ice­
storage refrigeration systems were in widespread use before the develop­
ment of refrigerators based on thermodynamic cycles. They are at present
uneconomic because of ice distribution costs. With community-size
systems the problem is that separate district heating and cooling lines
would be required, because of the hot-water demand during the same
periods when space cooling is needed.
CONCLUSION
This brief survey of energy-storage concepts should convey the existence of
an active feld of research, development, and commercialization, in which
diferent storage concepts are in very diferent stages along the path toward
viability-economic and otherwise. Areas not covered in the survey include
plutonium production, which may well be considered a way of storing
energy for future use, provided that the plutonium is stockpiled. One reason
for the broad interest in energy storage is undoubtedly its potential role as
an important component in future energy systems, regardless of whether
one fancies models of future energy supply based on renewable or nuclear
energy sources.
Let me conclude with a general remark on the cost structure of energy­
storage systems. For systems such as the hydrogen-storage systems
outlined in Figure 4, one can identify three cost components (3) : (a) the
input or store charging device, which in this case may be an electrolysis
unit ; (b) the energy-storage device, which in this case is a hydrogen
container or a hydrogen-flled underground cavity ; and (c) the output or
extraction device, which in the hydrogen case may be a fuel-cell unit. For
some other systems, two or more of these components are combined. (For
pumped hydroelectric storage the input and output device may be a
common, two-way turbine, while the storage unit is the reservoir ; for a
battery, all three devices are combined). The corresponding three cost
components do not scale in the same way when the size of the system is
modifed. The input device should be rated to the maximum surplus energy
production that one wishes to absorb. The output device should be rated as
the maximum supply defcit expected to be covered by the store. Only the
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
28 SØRFMSFM
storing component scales with the capacity of storage, so that if the
hydrogen store is changed from short-term to long-term capacity, the
hydrogen cavern must be enlarged, but the input and output devices (and
hence their costs) may remain unaltered.
The altered cost structure that addition of energy storage can induce in
the system may have profound consequences. In urban applications, the
average power level of an automobile is typically ten times lower than the
maximum power rating. If acceleration is performed by a storage device, the
main engine rating may then be reduced by almost 90%, because the peak
loads are so infrequent that they do not change the average power level
much when spread over a slow storage charging process. Community
district-heating systems with central heat stores will allow individual
buildings to install all kinds of heat-producing devices, including solar
collectors, particularly if the distribution temperature is low. Local
conversion systems, such as wind or photovoltaic devices, can-by
including storage-become independent of backup fuel supplies (to be
transported over a distance) and common grid systems. This opens up new
possibilities in decentralizing the energy-supply structure. Decentralization
may improve stability and supply security ; however, it still may be
worthwhile to install local grids for power, gas, and heat, to increase the
reliability of the decentralized supply system.
Literature Cited
1 . Jensen, J., Slrensen, B. 1984. Funda­
mentals ofEnergy Storage. New York :
Wiley. 345 pp.
2. Slrensen, B. 1 981 . Renewable energy
planning for Denmark and other coun­
tries. Energy 6 : 293-303 ; 1982. An
American energy future. Energy 7 : 783-
99
3. Slrensen, B. 1979. Renewable Energy.
London : Academic. 683 pp.
4. Stewart, D., McLeod, R. 1980. Pro­
ducing Fuels on the Farm. New Z. J.
Agric., Sept., pp. 9-24
5. Slrensen, B. 1981. A combined wind and
hydro power system. Energy Policy, Mar.,
pp. 51-55
6. Fassenden, R. 1921. Sci. Am. 1 24: 348
7. Blomquist, C, Tam, S., Frigo, A. 1979.
Proc. 14th Intersoc. Energy Conversion
Eng. Conf, pp. 405-1 3. Washington DC:
Am. Cher. Soc.
8. Lehmann, J. 1981. Air Storage Gas
Turbine Power Plants. Proc. Int. Conf
Energy Storage, Brighton, England, pp.
327-36. Cranfield, England : Br. Hydro­
dyn. Res. Assoc. Fluid Eng.
9. Adolphsen, W., Mahan, J., Schmid, E.,
Weinstein, K. 1979. See Ref. 7, pp. 452-
54. Washington DC: Am. Cher. Soc.
10. Millner, A. 1979. Technol. Rev., Nov., pp.
32-0
1 1 . Thomsen, A. 1 865. N aturkrtfterne i men­
neskets tjeneste (The forces of nature
in the service of Man). Copenhagen :
Philipsen
1 2. Laquer, H. 1976. Energy Storage, Com­
pression and Switching, ed. W. Bostick,
Y. Nardi, Q.Zucker, pp. 279-305. New
York : Plenum
13. Pound, R. 1980. Science 208 : 494-95
14. Nkonoki, S., Slrensen, B. 1984. A rural
energy study in Tanzania : The case of
Bundilya village. Natural Resources
Forum 8 : 51-62
1 5. Ladisch, M., Dyck, K. 1979. Science
205 : 898-900
16. Hartline, F. 1979. Science 206 : 205-6
17. Apace Res. Ltd. 1982. Tech. Inf Bull.,
Rep. PA/l l 1/1. Hawkesburyg New South
Wales
18. Hespanol, I. 1979. Energia (Sao Paulo),
5 : (Nov./Dec.)
19. Wise, D. 1 981 . Solar Energy, 27 : 1 59-78
20. Mardon, C 1982. High-Rate Thermo-
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
philic Digestion of Cellulosic Wastes.
Presented at 5th Austr. Biotechnol.
Conf., Sydney, August
21 . Stafford, D., Hawkes, D., Horton, R.
1981. Methane Production from Waste
Organic Matter. Boca Raton, Fla: CRC
22. Jensen, J., Tofeld, B. 1981 . Advanced
Batteries for Energy Storage. See Ref 8,
pp. 205-16
23. Kordesch, K. 1 980. Electrochemical
Energy Storage. In Energy Storage, ed. J.
Silverman, pp. 8-34. Oxford : Pergamon
24. Harth, R., Range, J., Boltendahl, U. 1981.
Eva-Adam System: A Method of Energy
Transportation by Reversible Chemical
Reactions. In Energy Storage and Trans­
portation, ed. G. Beghi, pp. 358-74. Dor­
drechl, Netherlands : Reidel
25. Cacciola, G., Giodano, N., Restuccia, G.
1 981 . The Catalytic Reversible (De)
Hydrogenation of Cyclohexane. See Ref.
8, pp. 73-89
26. Buchner, H. 1980. Thermal Energy
Storage using Metal Hydrides. See Ref.
23, pp. 327-34
27. Telkes, M. 1976. In Critical Materials
Problems in Energy Production, ed. C.
Stein, pp. 440-63. New York : Academic
28. Biswas, D. 1977. Solar Energy, 19 : 99-
100
ENERGY STORAGE 29
29. Furbo, S. 1982. Commun. No. l l0from
Thermal Insulation Lab., Tech. Univ.
Lyngby, Denmark
30. Bakken, K. 1981 . System Tepidus : High
Capacity Thermochemical Storage/Heat
Pump. See Ref. 8, pp. 23-28
31 . Jensen, J. 1 981 . Improving the overall
energy efciency in cities and communi­
ties by the introduction of integrated
heat, power and transport systems. Proc.
lEA Int. New Energy Conservation Tech­
nologies Conf, p. 2981, ed. J. Millhom,
E. Willis. Berlin : Springer-Verlag
32. Roseen, R. 1978. Central Solar Heat
Station in Studsvik. Rep. ET-78/77. AB
Atomenerg, Studsvik, Sweden
33. Margen, P. 1 980. Sunworld 4 : 1 28-34
34. Kreider, J. 1979. Medium and High Tem­
perature Solar Processes. New York :
Academic
35. Hall, L., Swet, C., Temanson, L. 1979.
SunIl, Proc. ISES Conf, New Delhi l9/ò,
pp. J5óóº. London : Pergamon
36. Kirkpatrick, D., Masoero, M., Socolow,
R., Taylor, T. 1981. A Unique Low­
energy Air-conditioning System using
Naturally1rozen Ice. Rep. Ctf. Energy &
Environ. Studies. Princeton, NJ :
Princeton Univ.
A
n
n
u
.

R
e
v
.

E
n
e
r
g
y
.

1
9
8
4
.
9
:
1
-
2
9
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
r
j
o
u
r
n
a
l
s
.
a
n
n
u
a
l
r
e
v
i
e
w
s
.
o
r
g
b
y

T
e
x
a
s

T
e
c
h

U
n
i
v
e
r
s
i
t
y

-

L
u
b
b
o
c
k

o
n

0
6
/
1
5
/
0
9
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.