Sunday, 16 October 2016

An overview of energy storage - where we are now.

We know we need energy storage, but what does that actually mean? How much is enough? What are our options? This is a large subject but I am going to try to give a birds eye view of the main issues because Transition Cambridge Energy Group are going to discuss this soon. I am primarily concerned here with electricity, though heat is also useful. There are three parts to this post: a bit of economics, a run through of the technologies and some examples of (mostly) existing installations. Northern Ireland now has grid scale examples of both compressed air storage and flywheels while England has a pilot liquid air energy storage system and some inter-seasonal heat storage. Battery storage is important but it is not the only game in town.

Making money from energy storage

There are several ways that energy storage can be worth something:
  • By providing services such as voltage regulation or participating in the capacity market. Payment may be based on being available rather than how much is delivered.
  • By allowing us to use energy that would otherwise be wasted - so the 'fuel' is free.
  • By trading on a price differential.
    • Consumers can buy energy when it is cheap to use later.
    • Generators can store energy when it is cheap so as to get a better price later
It is sometimes possible to 'stack' revenues, for example trading on price differentials and also providing balancing services.

Balancing service providers must be 3 MW or larger, but aggregated supply can qualify.
Balancing services require very fast and flexible response (such as flywheels, batteries and hydro-electric pumped storage). They are normally supplied by large installations (at least 3 MW). However, I have heard of at least one company working on a scheme to aggregate many homes with battery storage so as to supply balancing services and make additional revenue for home owners. I was told this would take a few more months to set up.

Capacity market is for 2 MW and up. Storage systems are disadvantaged by shorter contracts.
The UK grid has a capacity market. To provide this service you have to guarantee to provide capacity (or reduce demand) at times when the system is stressed. In return you get regular payments, regardless of whether your capacity was actually needed. Energy storage systems can theoretically take part but they must be at least 2 MW and they are at a disadvantage because they can only get 4 year contracts whereas fossil fuel generation can get 15 year contracts. Longer contracts mean more guaranteed revenue. In the last auction the price offered was £18/year/kW. So a 2 MW system would get £36,000/year.

Energy that would be wasted - storage integrated with generation
At the moment there is not a huge amount of energy that is wasted, because we do not generate more than we need. However, sometimes the grid is not strong enough to carry power from where it is generated to where it will be used and we end up 'paying wind farms to switch off' (constraint payments). This problem will increase as we get more renewable energy unless we spend a lot more money on the grid. However using storage we can save the power for later when the grid is less constrained. This requires storage close to the generator so that the grid can handle it.

Integrating storage with generation reduces network costs
In practice there are economic advantages to having the storage directly integrated with the generator. Otherwise the storage facility has to pay the grid twice: once for power import and again for export. Anesco was the first to fit a commercial battery storage system in a PV farm in the UK: a 250 kWh system in Dorset. They recently announced a 1.2 MWh system at Oxcroft. Camborne Energy Storage is fast catching up using Tesla batteries. Notrees is an American example of battery systems integrated with a wind farm. Also General Electric is offering wind turbines with batteries already inside them, for smoothing out fluctuations in wind and delivering more reliable power over 15-60 minute timescales. When storage is integrated with the grid it also enables trading on a price differential.

Energy subsidies and consumer prices are fixed.
Most consumers pay a fixed price all day and all year. Also small renewable generators (like Reach) usually have a fixed price power purchase agreement. All the subsidies that have been implemented (Feed in Tariffs, Renewable Obligation Certificates, Contract for Difference) are also fixed price, regardless of time of day, season or weather. However there are still ways to take advantage of price differentials on the market.

TRIAD payments for generating at times of peak demand
One of the more obscure is the National Grid's system of TRIAD payments designed to reduce peak system demand. At the end of the winter, three half hour periods when demand was very high are selected and suppliers are charged according to their load during those periods. Generators connected to the distribution system act as negative demand and can obtain TRIAD payments for the power they exported at the right time. The payments are usually around £30/kW. At that rate a storage system supplying 250 kW during all three TRIAD periods would get £22,500. Anesco has been successful in obtaining these payments for its battery storage system integrated with PV.

Generators can get power purchase agreements that track the market
On the normal wholesale market there is little price differential between seasons though there are differences through the day. You can view historical data here. Generators get paid by suppliers via a power purchase agreement and this can have any terms the parties are prepared to agree on. EDF offers a range of contracts that track the market to a greater or lesser extent.

For consumers, time of use tariffs can promote battery storage
For consumers, many other countries have time of use tariffs with peak time electricity much more expensive than at other times. California is leading the way by requiring all electricity companies to offer a time of use tariff by 2019. Time of use tariffs are already common in Australia and here is some advice for Australian consumers about how time of use tariffs can be leveraged with battery storage. We could do the same here - and it will be easy to switch tariffs when we have smart meters.

Inter-seasonal storage needs to be 200-300 times cheaper to be economic.
Trading on price differences through the day is practical but trading on price differences though the year is not. You can use a battery every day for 10 years but inter-seasonal storage is only used once every year - stored in summer for winter. This means the storage has to 200-300 times cheaper (per kWh stored) or the price differential has to be 200-300 times larger At the moment the only plausible technology for inter-seasonal storage is as low grade heat (less than 100C), because this can be stored relatively cheaply.

Hydrogen (or other) gases.
You can use electricity to make hydrogen, for example by electrolysis of water. You can also make hydrogen by 'cracking' methane - if you store the carbon safely afterwards this is low carbon energy. Hydrogen can be used to fuel a gas turbine or fuel cell and make electricity later - or burnt directly in a gas boiler for heat. You can make hydrogen by hydrolysis with around 70% efficiency and convert this back into electricity using a fuel cell with 40%-60% efficiency. So overall the round trip efficiency is at best about 40%.

The other problem with hydrogen is it has a very low density (high volume) which makes storage difficult. Compressing it uses energy so reduces the efficiency. None the less, there are many engineers who think hydrogen would make an excellent energy carrier and storage medium (see Wikipedia Hydrogen Economy). There is a lot of development going on in Germany on this topic.

Alternatively, if you have a source of carbon (such as a biomass power station) you can reform hydrogen to make methane, which is easier to store and we already have infrastructure for it. Again, we can use it for heat or to make more electricity. Another possibility is to use nitrogen from the air to make ammonia, which can be used as a fuel or as a chemical feedstock. However these additional steps lower the efficiency.

Even with methane, however, serious amounts of storage need serious volumes - our main gas storage is a depleted gas field under the North Sea. In Germany gas is stored in underground salt caverns - unfortunately we do not have many of these.

In general gas storage has low efficiency and requires convenient geology for large volumes.

Batteries are another kind of chemical storage - the electrolyte in the battery stores the energy. Batteries come in a huge range of types and sizes from hearing aid batteries providing milliwatts to grid scale providing megawatts. Most batteries we see are encapsulated and the electrolyte is often a gel rather than a liquid. However it is also possible to make flow batteries, where the electrolyte is a liquid that can be stored externally. There is a flow battery project in construction at the moment for a project on the Isle of Gigha in Scotland.

Batteries are efficient but still fairly expensive. The biggest battery systems available now are grid scale battery farms of up to about 10 MW, useful for grid balancing.

Hydro-electric pumped storage is quite old technology. The UK has four large power stations in Wales and Scotland totalling 2.6 GW and 21 GWh storage (Cruachan, Dinorwig, Ffestinog and Glyn Phonwy). They use gravity to store energy, by pumping water uphill. To reclaim the energy you let it flow down again through turbines. Hydro-electric power is fast response and efficient but you have to have or be able to construct two enormous lakes at different heights with a pipe between. There is an assessment of European potential here. The longer you are prepared to make your pipe the bigger the potential. With a limit of 20 km the EU potential has been assessed at 11 TWh. That would be quite useful - the EU uses about 8 TWh/day - but 20km is a very long pipe and these facilities are very expensive to build.

There are other possibilities. One suggestion is a 'gravity railway'. You drive the train uphill to store energy and allow it to roll downhill to regenerate. This also works out rather expensive.

Heat - a candidate for inter-seasonal storage - and cold.
Most of our power stations use heat to make electricity - the heat is used to expand gases to drive the turbines. Concentrating Solar Power (CSP) stations use solar power to heat up a thermal store - usually molten salt. With a big enough store the power station can run through the night without topping up. Unfortunately, CSP needs direct sunlight (to focus sunlight from a wide area onto the thermal store) and the UK's cloudy weather means CSP is pretty useless here. CSP is common in Spain and California!

A while ago I was impressed by Isentropic - a startup working on pumped heat energy storage. To make electricity from heat you need a difference in temperature: Isentropic uses two thermal stores, one cold and one hot. You charge it up by using refrigeration technology to move heat from the cold store to the hot store. I am not sure why this technology is not more successful. A demo project was announced but they have gone bust and are in administration.

However, we are starting to see installations storing heat for heating buildings. With a big enough thermal store this can be inter-seasonal storage. ICAX is one company working in this area and Tesco has one of their first installations. Underground heat stores are impractical for existing buildings but plausible to build into new ones.

As well as storing heat it is also possible to store cold. For example you can make blocks of ice to use later for air conditioning. This works well at a building level.

Compressed Air Energy Storage (CAES) - the difficulty being how to handle the heat.
You can use power to compress air and then let it expand through turbines to get the energy back. The main problem is that when you compress the air it gets very hot and when it expands it gets very cold. If you don't store the heat as well you lose a lot of energy. This can make the difference between 40% efficiency and 80% efficiency. The first CAES power station was built in Huntorf, Germany in 1978. It could generate 320 MW for 2 hours. The air was stored in pre-existing salt caverns.

Salt caverns are incredibly useful as they are large and gas tight. It would be nice to have more of them. However, as long as we have a suitable layer of salt we can make the caverns - the CAES facility being built at Larne in Northern Ireland will use engineered caverns in a salt layer 800m below ground.

Liquid air energy storage (LAES) is an extreme version of CAES which involves cooling as well. Highview has a pilot running with this. It uses waste heat from a power station to keep the gas warm as it expands.

Flywheels are very good for storing smallish amounts of energy with incredibly rapid response. You get them in buses for regenerative braking. They are also increasingly being used for grid balancing services.

Where we are now - real installations (mostly) in the UK

The table below has some examples of installations in the UK. In each I give three important numbers: power, capacity and efficiency. Capacity is measured in Watt-hours. This is not the same as how fast you can charge it up or deliver the power which is measured in Watts. I get very cross when I see an article on a storage system which only gives one of these numbers - you need both. For example, the a typical domestic sized battery to go with PV could store 5 kWh which is about half an average household daily use. It can deliver power at up to 4 kW, which is enough for a kettle (3 kW) but not an electric power shower (10 kW).

The Rough gas storage facility - by far our biggest - can store enough gas to supply 14 days worth of winter electricity use but the maximum supply rate is only 10% of our peak demand. In practice most gas from this facility is used for heating, not electricity.

Dinorwig, our largest hydro-electric pumped storage station can supply 3% of peak demand for a few hours.

The table is sorted by decreasing capacity and has three sections - the top one is in GW, the next one in MW and the last one is kW. To give you a better idea of what these numbers mean I have also included two rows for typical consumption - one for the UK as a whole and the other for a typical home.

Rough gas storage
(a depleted gas field under the North Sea)
(3.3 billion m³ storage, can be delivered at a rate of 41 million m³/day). This can be used for heat or to power gas power stations.
35,000 GWh
14,000 GWh
18 GW
7.2 GW

Starting with gas (so not round trip)
95% for heat
50% conversion from gas to electricity
UK winter day consumption.1,000 GWhUp to 60 GWN/A
Dinorwig Power station (Wales)
Hydroelectric pumped storage station.
9 GWh1.7 GW75%
Larne (Northern Ireland)
Compressed air in engineered caverns.
2 GWh0.330 GWCAES can be 40% to 80% efficient depending on how heat is handled.
Highview Viridor Pre-commercial demonstrator
Liquefied air (stored at cryogenic temperatures).

15 MWhUp to
5 MW
60% standalone.
More than 70% with a source of waste heat
Tesla deal with Southern California Edison Grid scale battery storage.80 MWh20 MW90% (?)
Rhode (Northern Ireland) flywheel for grid stability.? probably only enough for a few minutes20 MW90%
Oxcroft (England) batteries integrated wtih PV farm (announced in March).1.2 MWh1 MW90% (?)
Electric car: Tesla Model S85D85 kWhUp to 310 kW delivered but only for brief periods.
Charge at up to 16 kW from a domestic supply.
80% to 90% (?)
Average home daily consumption9 kWhUp to 10kW (electric power shower)N/A
Litre of diesel fuel.
NB. Synthetic diesel is hard to make.
9.9 kWh
4 kWh
Depends on equipmentSimilar to gas storage. - 95% for heat and maybe 40% for conversion to electricity
LiON battery to go with your home PV panels 5 kWh4 kW95%

The Renewable Energy Association's report: Energy Storage in the UK has a complete list of energy storage projects in various stages from Announced to Operational.

Efficiency may or may not be important. It is not so important if the energy stored would otherwise go to waste. For example if you have excess or PV power that can't be delivered to the grid for some reason, then the cost of the energy is zero so losing half of it has no financial loss. However, there is usually a limit to the energy available so efficiency also affects how much you can store and sell later. That affects your rate of return on investment.

Other important factors for storage include energy density. For electric vehicles energy density is critical in terms of both weight and volume but for buildings the main constraint is volume. For home use you probably want it to fit in a cupboard, for an office it should fit in a room. For grid size storage, you want it to fit in a warehouse. A litre of diesel fuel stores as much energy as a domestic LiON battery system the size of a small cupboard.

Lifetime may be a factor too. Batteries can typically only handle 2000-5000 cycles before the charging capacity is significantly reduced. They are usually assumed to last at least 10 years. Hence to be financially viable they need to pay for themselves in 10 years.

Where should we go from here?
Personally I think we should encourage storage by tariffs in several different ways:
  • Charging for rapid fluctuations in generation would encourage generators to integrate storage, enough for at least half an hour.
  • We should introduce variable tariff rates for both generators and consumers. As a minimum there should be a premium price for winter peak times.
The integrated storage is likely to be a mix of batteries, compressed air and flywheels.

I doubt that inter seasonal storage will ever be practical on a large scale. Therefore, I recommend we limit the proportion of PV that can be installed so that we do not get excess power in summer time. Wind is better suited to the UK use pattern (see Does the UK need solar electricity?)

It ought to be possible to possible to develop gas storage enough to keep us going through days and weeks of low wind - just one or two more depleted oil fields would be enough. However to make this economic we must have a carbon price to make synthetic gas - either hydrogen or methane - cheaper than fossil fuels.


  1. An excellent summary. I liked Isentropic's technology too. It's not clear what happened to the Midlands pilot project. I assume financing difficulties, but possibly getting the tech to work. Towards the end they were offering CAES for gas plants as something that should be financially viable, and the heat pump was on the back-burner.

    You say ARES (gravity train) works out expensive, which does seem plausible, but I've not seen any actual numbers. Have you?

  2. This article says $55 million for the first of a kind which would be 12.5 MWh, 50 MW. That means $1,100/kW but $4,400/kWh

  3. Here is another flow battery option that looks useful. This is iron/zinc, so much less toxic than others, and the largest size is 1.4 MW, 4.2 MWh - useful for deployment on the grid. However I can't see anything about actual installations.

  4. A 49 MW battery storage system is to be built in Cumbria.
    The article does not say how many MWh though. It will be used for grid balancing.