Energy storage devices

This is about methods of storing relatively large amounts of captured energy in some form that can be transported to where it is needed. It does not cover fossil fuels because they are primary energy sources. It is only concerned with energy that has been generated by a primary source and then stored until it is needed. Hence it covers batteries, cylinders of hydrogen and similar devices.

Electricity storage

The straight fact is that there is no known efficient, large scale way of storing electricity. The best of the current methods are pumped hydro, which has excellent capacity but relatively low efficiency, and flow batteries, which have lower capacities but better efficiency. With the exception of carbon-carbon batteries, which are really capacitors, all the other methods of storing bulk electricity are types of secondary cell.

Much of this storage is needed to cope with fluctuations in renewable energy supply. This applies principally to wind, solar and tidal sources: other forms of renewable energy can provide power as consistently as conventional, non-renewable sources. The storage requirement can be reduced remarkably cheaply by the widespread adoption of dynamic demand management.

Secondary cells, often referred to as batteries, store the electricity put into them until its needed. Batteries are fairly efficient: you typically get back around 85% of the energy you put into them if you discharge them immediately. However, some types self-discharge if they are left. The self-discharge rate of some types, such as Nickel Metal-Hydride (NiMH), is high enough to make them unsuitable as long term stores. After a year a fully charged NiMH cell will be flat and may have lost so much charge that it has destroyed itself.

Carbon-carbon batteries

This is really a new form of capacitor that can store electric charge in quantities that until recently has only been possible in a battery. It offers good energy density, reasonable cost and fairly environmentally friendly constituents. Ultracapacitors Challenge the Battery seems to be a reasonable review. However, it suggests that current devices are best suited for short term storage, such as capturing energy from regenerative braking in a hybrid vehicle so it can be used during the next burst of acceleration. Another likely application is in backup power supplies, e.g uninterruptable power supplies, but further development could produce devices that can replace lithium-polymer cells.

Flow batteries

Flow batteries are rather like a cross between a fuel cell and a secondary cell with a liquid electrolyte. Like fuel cells, the current generation involves a proton flow across a semi-permeable membrane. Like a secondary cell, a reversible redox chemical reaction is the proton source but, unlike most secondary cells, the reactants are dissolved in the electrolyte. This gives the possibility of using a relatively small reaction cell and storing the reactants in large storage tanks. During charge and discharge operations the reactants are pumped through the cell.

The technology is currently being exploited by VRB Power. Their system uses a vanadium redox reaction. In this system vanadium ions are dissolved in dilute sulphuric acid. It consists of two electrolyte tanks, each containing active vanadium species in different oxidation states. One tank, feeding one side of the cell, contains the V(IV)/V(V) redox couple and the other tank, feeding the other side of the cell contains the V(II)/(III) redox couple. The cell voltage at full charge is 1.6v. The overall charge/discharge efficiency is 80% and, because the same chemical species is used in both halves of the cell, leakage through the membrane doesn't degrade the cell because each recharge completely regenerates the material on both sides. Its main disadvantage is that its energy density is only half that of lead-acid batteries but this is offset because the capacity of a fixed installation is limited only by the tank size and a vehicle could potentially be "refuelled" by changing spent tank contents for fresh rather than recharging it. The system seems to be quite scalable: the VRB Power gallery shows systems with capacities ranging from 10 kWh to 2 MWh. The original 1 MWh installation at King Island, Australia uses 70,000 litres of vanadium sulphate to buffer the output from a 2.5 MW wind farm. This has been operating since late 2003.


Relatively small amounts of electicity can be stored using flywneels. A hollow steel cylinder with an attached genemotor set is supported on very low friction bearings and spun up in an evacuated chamber to minimise friction. Energy is stored by spinning up the flywheel with the genemotor and extracted by using the genemotor as a generator. Currently, flywheels can store 2-25 kWh efficiently for periods of minutes to hours. They can be optimised for low output over a long period or for large outputs over very short periods. The main use at present is to provide a bridging UPS service while an emergency diesel generator set is brought online.

The JET fusion reactor at Culham draws up to 1000 MW during each test shot, which can last for up to 20 seconds. Of this, 800 MW is provided by two 775 ton, 9 m diameter flywheels. When spun up to 225 rpm, each stores 1 MWh and can generate 400 MW at full output. By contrast, the motor that spins up the flywheel is only 8.8 MW. Thus each flywheel can discharge its stored energy about 45 times as fast as it accumulates it. This energy buffering prevents a test shot from disrupting the National Grid. Switching is achieved by energising windings on the flywheel: the surrounding static windings are connected to the JET torus via a bank of rectifiers with no intervening switches. Most systems with this storage capacity would be implemented as a flywheel farm rather than a pair of very large units. Presumably JET's current switching requirement rules out the farm approach.

Sources: the Electricity Storage Association and EFDA-JET, Focus on: power supplies.


Cars have been operated from electric storage batteries for over 100 years. The favourite battery to date has been the lead-acid battery. Its simple and well-understood technology, retains its charge for a long time and is reasonably efficient. However, it has a number of disadvantages:


Lithium polymer batteries offer several advantages, though at a higher cost. They need relatively complex charging circuits and must be protected against deep discharge or they will destroy themselves. However, they offer much higher energy densities than any other current battery technology and don't contain spillable liquids. In a crash they will deform rather than cracking or shattering but will burn fiercely if the cell container is pierced or the charging system malfunctions. Their energy efficiency through a charge/discharge cycle is about 85%, which compares very well with that achievable with a fuel cell-based hydrogen cycle. See Hydrogen economy revisited.

Li-poly cells are already a viable storage system for use in solar powered aircraft and motor gliders. Their price and capacity have now reached the point where they are viable power sources for top-of-the-range electric road vehicles.

LithiumIronPhosphate batteries

These are essentially a riff on Lithium-polymer technology. Size for size they have a somewhat reduced capacity but are rather safer and more robust. They do not have the same tendancy to catch fire if mistreated and may survive more charging cycles.

Nickel-Cadmium batteries

The biggest battery storage unit yet built, as of 2006, is a 1300 tonne assembly of NiCd rechargeable cells which can deliver 40 megawatts for 15 minutes. This is enough to provide emergency power for the 90,000 people in Fairbanks, Alaska while a backup generator comes online. However, this is a long way short of providing Fairbanks whole overnight requirements, especially in winter.

Pumped Hydro schemes

In 2003 there was over 90 GW of pumped storage in operation world wide, accounting for about 3% of global generation capacity. At that time it was the most widespread energy storage system in use on power networks.

The idea here is that during periods of low demand the generating turbines are used as motors to pump water into a raised reservoir, which is often a dam in an upland valley. During peak demand periods this water runs back down the penstocks and drives the generators to supply electricity. The main problems are that geography determines the location and storage capacity of the plant and that an efficient pump makes an inefficient generator and vice versa. In addition they have high capital costs and long construction times.

Installations typically have generation capacities in the 1000 - 2500 MW range and use hydraulic heads of anything from 33m to 1200m, but around 500m is common. The time a plant can run at full capacity depends on the reservoir size and the designed generation capacity. Current installations run times range from 5 to 153 hours, with the typical time being 10 to 20 hours. (Source: the Electricity Storage Association)

Non-electric energy storage

This section lists the known ways of storing the output of primary energy sources that isn't in the form of electricity.

Compressed gas storage

Compressed gas reservoirs can store similar amounts of energy to pumped hydro plants and, like them, are intended to handle peak loads. They store surplus power by pumping air into a reservoir. The energy is retrieved by using the compressed gas to spin turbines that drive generators. Advantages are fast start-up, easily controllable output and lossless storage. The reservoirs are likely to be underground for safety. They could be large special-purpose buried tanks, which are much cheaper than equally safe above-ground tanks would be, as well as disused mines or depleted oil wells and aquifers. Water-filled storage structures, if they are deep enough underground to provide the required hydrostatic pressure and are topped by gastight rocks, are preferred because these provide more or less constant gas pressure.

The favoured designs use the high pressure gas to feed the combustion chambers of a Closed Cycle Gas Turbine (CCGT). This raises the efficiency well above that of a conventional CCGT generator by eliminating the compressor, which absorbs 50% to 65% of the power output by the turbine, and allows 85% of the energy used to compress the gas into the reservoir to be recovered. However, this approach makes the system dependent on fossil fuel and is not emission-free. The overall energy efficiency after taking the fossil fuel burn into account is unstated, but if we guess the efficiency of a conventional CCGT as 45%, the compressorless CCGT might reach 90%, giving an overall efficiency of about 70%.

An additional siting restriction for combustion boosted designs is that the aquifer must not contain reducing agents such as iron pyrites, which will use up oxygen in the compressed air and impact efficiency, and their use must not affect the extraction of irrigation or drinking water.

Plants with the capacity to generate 580 MWh to 4300 MWh (300 MW for 2 hours to 250 MW for 16 hours) are typical of those currently installed or likely to be built in the short term. (Sources: New Scientist, 29 Sep 2007, p46 and the Electricity Storage Association)

Cryogen - Liquid air storage

The idea is to use excess generation capacity to cool and compress dry air to form a liquid, known as cryogen, and store it in well-insulated tanks at ambient atmospheric pressure. The equipment needed to do this is well-understood, reliable, and has been in routine use for well over 50 years.

The energy is extracted by pumping the cryogen through a heat-exchanger and using the resulting high-pressure gas to drive a turbo-generator. The cold exhaust from the generator can be captured, stored and reused to make more cryogen or to provide industrial cooling. The overall efficiency is said to be 50% if the heat-exchanger uses ambient air as its heat source or up to 70% if it uses industrial waste heat.

These plants would be relatively small and safe since they don't involve high pressure storage systems and all stages of the process operate at low temperatures. Plant costs are claimed to be about $1000/kw of generating capacity, which is 25% of the equivalent batteries. A commercial scale 3.5 MW pilot plant should be in operation by late 2012 and scaled up to 8-10 MW by 2014

This energy storage method is being developed by Highview Power Storage.

The Dearman Engine Company are taking a different approach which is aimed at smaller capacity systems. They are developing a reciprocating engine that sprays liquid air into the cylinders where it is gassified in the heated cylinder. Costs, maintenance and lifetime are all comparable to an internal combustion engine and an overall system efficiency of 70% is projected, though it was around 26% in late 2012. Like the Highview system, the Dearman project can be used as a renewable energy storage plant or, if the air liquifaction stage is separated from the Dearman Engine, it can be used to generate power from waste heat at a remote location or to run a vehicle.

The main drawback of both systems is that anything more than short term energy storage requires either a very strong storage vessel or the constant application of energy to keep the liquid air cool: pure cryogenic storage relies on the liquid air's evaporation to keep the remaining liquid cool, so the tank will slowly empty over time regardless of whether the system is in use or not. In this respect it is no different from liquid hydrogen-based energy storage except that liquid air can be stored at 78K rather than 33K. This reduces the cost of cryogenic air storage by comparison with cryogenic hydrogen storage.


Hydrogen is just a way to store large amounts of energy in a fairly easily transportable form. See the hydrogen economy revisited for more detail.

"Hydrogen battery"

This is another name for hydrogen on demand.

Metallic energy storage

Recent work by Solomon Labinov and Dave Beach at the Oak Ridge Laboratory in Tennessee (New Scientist, 22 October 2005 page 35) has shown that powdered metal is a viable and non-polluting way of storing and releasing energy where and when it is required.

50 nanometer particles of iron, aluminium and boron can be safely stored and transported at room temperature if coated with a thin layer of oxide, but will burn when heated to 250°C. The peak combustion temperature is a relatively cool 800°C, so they could be burnt in a conventional aluminium block piston engine or used to run a Stirling engine. Combustion speed can be controlled over a wide range (a few milliseconds to a second or two) by forming clusters from the nanoparticles. If the burnt particles are retrieved from the engine's exhaust stream they can be recycled by reducing the particles at 425°C in a stream of hydrogen and then burning them again. This process is non-polluting. Combustion consumes oxygen but produces no waste products at all. The combustion temperature is too low to form NOx pollutants and the hydrogen reduction process just releases water. Unlike burning hydrogen, the water would be all released at recycling plants, so it could be easily collected and discharged back into rivers without any release of water vapour into the air.

A major attraction is that the energy density of this system is high. All three metals give a higher energy per unit volume value than petrol. The energy density per unit weight of iron is worse than petrol but it is cheap: boron and aluminium are expensive storage media but could store much more energy in a given tank size and weight.

I think this is one to watch as a possible way of powering road vehicles and aircraft.


Methanol is another way to store large amounts of energy as an easily transportable liquid. See Methanol for more detail.