There are already many ways to store energy, but adapting them to the needs of certain renewable technologies – in a cost effective way – is challenging. The CSP industry for example is keeping its fingers crossed that molten salts (or some other thermal storage medium) will become cheap enough to improve the technology's proposition.

Further afield and much research is being done into new battery technologies and even fuel cells (useful for transportation; backup power and other small scale applications; but thus far still hasn't been able to scale up in a commercially viable way).

Numerous other technologies such as super capacitors and flywheels have also been proposed as ways to store excess energy, generated when the resource is strong but demand is weak (think the potential of wind power at night). Most of these technologies however, remain in the experimental stage.

Almost every vision of an intelligent grid requires the inclusion of some aspect of energy storage as part of its concept, yet energy storage has been a very small player in the traditional grid.

There are a couple of methods, though, that have already been operating non-stop for decades, reliably balancing between nightly production and daytime demand – pumped hydro storage (PHS) and compressed air energy storage (CAES).

PHS has been around since the 1890s. There is nearly 40GW of PHS capacity in the EU and another 20 GW in the USA. Expansion of PHS, however, is limited due to the shortage of suitable reservoirs. Thus it doesn't offer a solution to the incorporation of more renewables into the grid.

CAES, on the other hand, may offer some possibilities. There is an abundance of suitable storage locations throughout much of the world, and proven, reliable technology exists that can rapidly scale to the sizes required.

In CAES, electricity is used to force air under pressure into a cavern. To extract it, operators heat the compressed air with natural gas, and then push it through turbines to generate electricity. Like pumped hydroelectricity, this method is limited by geography. And its use of natural gas produces emissions that undermine some of the benefits of turning to renewable sources. But if this reduces the amount of power needed from high-intensity fossil fuel peaking plants, it could have some benefits.

“We need a series of large-scale, bulk energy storage facilities strategically located around [a] country, operating at transmission-level voltages, not just for renewable energy but for arbitrage and general ancillary services,” says Jason Makansi, Executive Director of the Coalition to Advance Renewable Energy through Bulk Storage (CAREBS). “All you have to do is look at what natural gas storage has done for the natural gas markets: over a 15 year period it has stabilised a highly volatile commodity.”

CAES concept

CAES is a relatively simple concept: separating out the compression and expansion functions of a combustion turbine (CT). Compression typically consumes about 2/3 of the energy of a CT, with the remainder going to spin the generator. Since they are more expensive to operate than steam turbines, CTs can't compete with the lower cost base load power stations when demand is low.

With CAES, however, combustion turbines can efficiently operate around the clock. During periods of low demand, the excess energy from the grid is used to turn the generator into a motor and which drives the compression sections of the turbine. The compressed air, instead of going into the turbine's combustion chamber is routed to an underground cavern for storage.

Later, during periods of high demand, the stored, compressed air is fed back into the turbine's combustion chamber to burn fuel and drive the generator. Since the turbine doesn't have to also drive the compressor at that time, it can generate more power for the amount of fuel burned.

From a renewable energy standpoint, CAES could help establish wind and solar on a more even footing with fossil fuels. At times when wind power generation exceeds demand or transmission capacity, as happens in areas as diverse as West Texas and Denmark, instead of dumping capacity, the power can be used to drive the compressors and store the energy for use when demand is higher.

This applies to solar, too, although in places such as California the peak solar production period corresponds with the peak demand for air conditioning on summer afternoons. But in Canada or Finland, for instance, there is a higher demand for light and heat rather than cooling. As a result, solar production and demand peaks occur at different times of the day and so CAES can store the energy to offset the time differential.

CAES requires an adequate underground cavern, so that the compressed air can be stored. There are three primary types of locations. One is mines no longer used which can be sealed off and filled with compressed air. A second option is to use depleted oil fields, pumping air into the spaces formerly filled with oil. The third, and most common, approach is to find a suitable underground salt deposit and use water to leach out the salt as a means of creating a cavern. According to the Electric Power Research Institute (EPRI), about 85% of the U.S. alone has geologic sites suitable for CAES.

Northern Europe is also replete with suitable salt deposits. For example there are already nearly 500 salt caverns currently being used for natural gas storage.

CAES studies

There are two primary designs for CAES systems – one proven and one planned – and a wide variety of alternatives. The proven category is exemplified by two existing CAES plants using a single shaft design: compressor-clutch-motor/generator-clutch-expander. This design is the focus of this article: it allows the compressor and expander to share the motor/generator, saving significantly on initial equipment costs.

The first such facility is a 290 MW E.ON-Kraftwerke plant in Huntorf, Germany. Built in 1978, it has more than three decades of successful operation, with as many as 375 compression and 450 generation starts in a single year.

The site uses two salt caverns with a combined capacity of 310,000 m³. There is roughly a 1:4 input/output ratio on the caverns, so that the 60MW compressor can fill the repositories in 8 hours at a rate of 108 kg/sec, while the 290 MW turbine operates for 2 hours at an air mass flow rate of 417 kg/sec.

The normal operating pressures in the caverns are 43 to 70 bar. Intercoolers on the compressor reduce the energy needed for compression, but this then makes it necessary to use gas to heat the air once it is removed from the caverns to drive the generator. Overall, the plant operates at about 42% efficiency.

The second such CAES plant is Power South Energy Cooperative's McIntosh plant. It went commercial in 1991 and incorporates advances that boost the efficiency by 25% over the Huntorf design. McIntosh harnesses a 19 million ft³ salt cavern with a three unit Dresser-Rand compressor trains consisting of one model MGA 4015 axial compressor and two model 4M86 centrifugal compressors delivering a combined compression power of 49 MW.

“The compressor train is about 80% efficient,” says McIntosh plant manager Lee Davis. “We have intercoolers between each stage of compression, and the expander section is also recuperated so that keeps your heat rate down. But you can't just put those numbers together without factoring in where you get the compression energy.”

The compressors generally get their electricity from a coal plant about 25 miles away, but when prices are right they might buy from the grid or a combined cycle plant. It takes about 41 hours to compress the chambers, and the generator can then operate at full capacity for 26 hours. The unit is kept on a turning gear and can go through synchronisation to 110 MW in 14 minutes. Emergency starts can be done in 9 minutes. Since the expander does not have to drive the compressor at the same time, it can efficiently operate at loads as low as 20 MW.

“The plant used to be seasonal, but now we run year round,” says Davis. “In the fall and spring we run as a backup unit or to help control the grid, and then we provide peaking power as necessary year round.”

The use varies with fuel costs and electricity sales prices. Davis says that, depending on the cost of gas, CAES is generally less expensive than a simple cycle GT and comparable to a combined cycle GT.

As a result, the McIntosh plant can cope with a large number of starts and stops.

“The starts on the machine are far more than you get on any other kind of a peaking plant usually,” says Davis. “We use the HP expander to start the compressor so you have two starts in the day. And the way the dispatcher uses us to regulate the grid, you may have multiple starts a day unlike a simple cycle unit.”

Looking ahead

With more utilities striving to reach double-digit use of renewable energy, there is increasing interest in establishing CAES plants to turn their wind farms and solar plants into fully-dispatchable energy sources.

“Energy storage could reduce the amount of peaking gas turbine capacity that would be required to compensate for wind's variability,” said Jacquelyn Bean, a Consultant at Navigant Consulting.

In January 2010, Pacific Gas & Electric received US$50 million in funding for the first stage of a 300 MW CAES plant in Kern County, California. FirstEnergy owns the rights to a CAES site at an abandoned limestone mine in Norton, Ohio with the potential of generating 2700 MW. Municipal utilities, the Federal Government and the State of Iowa are funding the Iowa State Energy Park, which would pump the air into an aquifer 3000 feet underground. Gaelectric is proposing CAES sites in Montana and Northern Ireland.

“The driving force is the massive amount of renewable energy that is expected to come on line,” says Jason Makansi of CAREBS. “These CAES plants can be built quickly and they provide an unparalleled level of flexibility for grid management.”

About:

Drew Robb is a graduate of the University of Strathclyde in Glasgow. Currently living in Los Angeles, he is a freelance writer focusing on engineering and technology.

Renewable Energy Focus, Volume 12, Issue 1, January-February 2011, Pages 18-19