By Steve Blankinship, Associate Editor
It was almost inevitable that Texas would see one of the first big grid glitches induced by a growing reliance on wind capacity.
Alabama’s McIntosh Unit 1 was an early CAES adopter. Photo: PowerSouth Energy Co-op, Harold Dubose.
Late during the afternoon of Tuesday, February 26, the Electric Reliability Council of TexasERCOTcut service to a handful of large customers in the Houston area after losing 1,400 MW of wind power over the previous three hours. During those hours wind production fell from more than 1,700 MW to 300 MW. The drop coincided with rising electricity demand as evening approached and with a weather front pushing colder temperatures into the state.
Even as the winds eased, ERCOT activated an emergency plan to curtail power to interruptible customers. This action shaved 1,100 MW of demand within 10 minutes. No other customers lost power during the declared emergency and the affected interruptible customers were fully restored after about 90 minutes.
The addition of hundreds of megawatts of wind power to the thousands of megawatts already installed in Texas and elsewhere is forcing grid planners to develop ways to compensate for increasingly large amounts of intermittent capacity. For example, just weeks before the Texas wind event, plans were announced to install 24 Wärtsilä 20V34SG reciprocating engines fueled by natural gas at a site 50 miles southwest of San Antonio. The Pearsall Plant will help compensate for the effects of wind intermittency. Besides providing peak power for South Texas Electric Cooperative’s eight member affiliates, the planned 202.5 MW facility will provide ancillary and other grid support services to ERCOT.
To help quantify the need for additional ancillary services provided by gas-fired units to follow wind variability, ERCOT recently commissioned a report by GE Energy to look at large wind penetration on its system. William O. (“Bill”) Bojorquez, vice president of system planning for ERCOT, said indications are the grid will need a 20 percent to 23 percent increase in gas-fired regulation capacity. “When wind dies off in the morning and load picks up, there might be some significant ramping requirements,” he said. The evening hours are another time when gas plants likely will be called on to regulate the grid as winds rise and loads ease.
But Bojorquez said that losing 1,400 MW of wind over three hours is not like losing a big coal or nuclear unit due to a trip. What happened in Texas last February took place over several hours and at a pace that allowed the grid operator to deal with the changing conditions. “This kind of event does not cause a big stability shock,” he said. (For more on the possible effects on baseload generation, see the sidebar on page 56.)
While every form of power has merits and liabilities, one troubling downside for wind is its inherent inability to turn on and off as needed. The Holy Grail for many wind advocates now that the resource has achieved near-mainstream status involves finding a common and low-cost way to store electricity. Achieving those goals would go a long way toward smoothing the resource’s inherent lumpiness.
To see what the future may hold, travel from West Texas 1,000 miles east to McIntosh, Ala. There, for more than 17 years, a compressed air energy storage (CAES) system has helped the PowerSouth Cooperative (formerly Alabama Electric Cooperative) meet peak and intermediate power demands. McIntosh Unit 1 began service in 1991. The 110 MW CAES unit uses air compressed and stored in an underground cavern during off-peak periods. When needed for generation, the compressed air is released and mixed with natural gas in a fired reheat high pressure and low pressure expansion turbine to generate electricity. McIntosh 1 was the second commercial-scale CAES gas plant in the world. The first, a 290 MW facility in Germany, has been running since 1978.
In a conventional gas turbine, atmospheric air is compressed to a higher pressure, fuel is added in a combustion chamber and the resulting hot high-pressure combustion gas expands through a turbine. This provides 60 percent or more of the compressor’s motive power. It also provides the balance of the power (40 percent or less) as mechanical energy to drive an electric generator.
In a CAES/gas turbine cycle, by contrast, the compression cycle is separated from the combustion and generation cycles through the use of low-cost off-peak or excess electricity. Motor-driven inter-cooled compressors provide the compressed air that is first held in storage and then released on demand to a modified steam turbine for power generation. The compressed air is heated in a heat recovery unit (HRU) coupled to the gas turbine exhaust. This can improve the economics of the CAES gas turbine operation as the fuel input can be less than 4,000 British thermal units per kilowatt-hour (Btu/kWh) delivered.
“Without its parasitic load, the hot air (gas) expander delivers about two-thirds more power with no increase in fuel consumption,” said Septimus van der Linden, president of Brulin Associates, a consultancy that focuses on gas turbines used with emerging technologies. He said the required compressed air comes at a much lower cost, which, in turn, enables lower-cost generation during high demand cycles. This reduces demand from other intermediate load systems such as gas fired thermal, combined cycle plants or even from lower capital cost simple cycle gas units.
Now imagine using wind energy as the resource to compress the air instead of relying on off-peak fossil capacity. That would mean realizing the same benefits as those achieved by the Alabama and German plants, but without the associated fossil fuel load. A wind turbine connected to a compressor rather than a generator could produce and store compressed air whenever the wind blows and do so without the need to first convert mechanical energy to electricity to operate a compressor to produce the stored air.
What’s more, around 75 percent of the United States is geologically suited for air storage caverns, including salt domes, depleted gas fields, depleted mines, hard rock mined caverns and deep saline aquifers (Figure 2). And compressed air requires 1/25th of the storage volume as pumped hydro storage to produce the same MW and time duration.
Such dispatchable wind plants essentially become hybrid generation projects, using wind energy and storage in conjunction with heat. The technology’s proponents think the hybrids will cost slightly more to build than conventional wind or fossil plants, but will offer much lower operating costs than either one due to their lower fuel consumption and higher capacity factor.
The hybrids also can be configured as baseload plants, peaking or intermediate plants or even as firming operations to improve transmission utilization of existing wind farms.
And they can be big. CAES plants 1,000 MW in size may be feasible in wind-rich areas. They also can be scaled to provide schedulable dispatch from a few hours per day to round-the-clock operation. What’s more they use relatively little fuel (or none, in the case of waste heat or other renewable sources) meaning that such plants can provide new power capacity in markets where conventional wind and fossil or nuclear plants would be difficult or impossible to build.
Just this sort of approach has caught the imagination of several developers. In West Texas, for example, a proposal has been put forward by Shell and Luminant (a unit of Future Energy Holdings) to combine 3,000 MW of wind and CAES fueled with natural gas. Similar proposals for 4,000 MW of wind have been floated by oilman T. Boone Pickens and Babcock & Brown.
A full-scale wind compression facility being developed by the Iowa Association of Municipal Utilities is planned for operation in 2011. The Iowa Stored Energy Park would use off-peak conventional generation and unsold wind energy to store air in a 3,000-foot-deep aquifer. The plant configuration calls for 268 MW of CAES generating capacity with 75 MW of available wind energy. As demand for electricity rises, the stored air would be released, heated and used to help drive expander gas turbines to make power for customers in Iowa, Minnesota and the Dakotas.
Austin Energy, the municipally-owned utility for Austin, Texas, has proposed incorporating compressed air into a broader strategy that would include wind, solar and stored heat. Mark Kapner, senior strategy engineer, termed the concept a “dispatchable hybrid solar wind power system” that would use elements of wind, solar thermal and thermal storage. If realized, the scheme would take CAES a step further by eliminating the need for fuel-fired power to compress the air as well as to pre-heat the air before expanding it in the generation mode. In place of a combustible fuel, Austin Energy’s idea is to use solar energy stored in a molten salt bed or other medium to drive a turbo expander.
“It should be considerably less expensive to integrate wind, solar-thermal and the appropriate energy storage media in a single system, than to separately design and construct wind and solar plants,” said Kapner.
Essential subsystems would include:
- Wind turbines directly coupled to air compressors, or conventional wind turbines used to power a central air compressor on the ground
- A collector system consisting of high- pressure pipes to collect compressed air from several turbines
- A 100 bar large-diameter pipeline to transport the compressed air from the wind turbines to the solar site, which would also contain the expander-generator
- A compressed air storage cavern, which could be sited either near the wind turbines, along the transmission pipeline or at the solar collector site
- Solar thermal collectors with integral thermal storage for heating the compressed air to 392 F
- A compressed air expander-driven generator.
One advantages the concept may offer over separate wind and solar generating facilities is its use of a single prime mover, dispatchable generator. Furthermore, there would be no need for electric transmission connection to the wind farm, only to the expander-generator.
One CAES proponent is General Compression of Newton, Mass. The company is developing a wind turbine compressor system that stores wind energy that later can be delivered to meet the demand cycle. In the General Compression scheme, the wind turbine powers a compressor instead of turning an electric generator. The company is now testing a full-scale prototype for its dispatchable wind power system (DWPS) and expects to verify the technology this summer. The first prototypes are planned for 2011 and the company hopes for the first commercial DWPS installations by 2012.
General Compression’s wind turbine will look like a conventional 1.5 MW wind tower, the main difference being the contents of the nacelle, which is the enclosure behind the blades that houses the machinery that turns wind into electricity. In a conventional wind turbine, the nacelle contains an electric generator and high-ratio gear box. A wind compressor’s nacelle would house a light-weight compressor array and low-ratio gear box.
When the air is released from the high pressure under which it has been compressed and stored it is almost cryogenic in adiabatic expansion, meaning it is very, very cold. Once it is released into the expander it must be heated using gas or some other heat source. In the future, waste heat from an industrial source, geothermal, solar thermal or digester gas may be used. For the first projects, at least, natural gas seems to make the most sense.
An Economic Fix for Wind?
“The pain of the wind industry is that it doesn’t get a good price for power because the utilities that buy it do it to meet renewable portfolio mandates,” said Michael Marcus, a founder of General Compression and its president.
The beauty of using CAES with wind, said Marcus, is that it works well with the existing power sector infrastructure, grid or power markets requiring no special accommodations. “It fits right in.”
The economics for combining wind and CAES may prove enticing. For example, Marcus said most peaking gas plants currently cost $800 to $1,000/kW to install. The DWPS in a natural gas hybrid peaking configuration (say, 180 MW of wind turbines with storage and a 400 MW natural gas turbine expander, dispatching 400 MW up to eight hours per day) could cost slightly more per kilowatt than the gas peaking plant, but it would burn one-third as much fuel. The wind-produced CAES/gas plant hybrid has a heat rate (expressed as British thermal units consumed to generate a kilowatt-hour) of about 4,000 Btu/kWh or less. By contrast, the heat rate of most single cycle gas turbines is around 9,000 to 10,500 Btu/kWh. A combined cycle configuration could lower that to perhaps 7,500 Btu/kWh.
“So depending on the type of plant, you are using only one-third to two-thirds of the fuel because the fuel providing the compression is free,” he said.
Septimus van der Linden said he believes that achieving 50 GW of bulk energy storage in the United States, primarily via CAES, is realistically achievable. “The stability of the grid will be enhanced and renewable energy will make an economic contribution without requiring tax incentives.”
Compressing air using wind energy could revolutionize the wind farm concept by expanding the focus from the relatively narrow perspective of turbine efficiency to the wider view of overall system productivity. Marcus said he thinks compression can substantially increase the value of power produced. “In most of the areas where we have modeled our projects, the after-tax internal rate of return is about twice those of conventional wind projects.”
To date, the industry has done a good job of converting wind to electricity, although the resource’s intermittency has been a difficult obstacle to overcome. Building long-distance transmission to bring wind resources to market has been difficult because of wind’s relatively low capacity factor, on the order of 28 percent. A key objective behind building CAES is to help wind become a more efficient user of transmission, which, in turn, should stimulate more investment in long-haul transmission.
Through CAES, said Marcus, “we are creating a storable, dispatchable product.”
Cycling Coal for Wind?
Bill Bojorquez, vice president of system planning for ERCOT, said that improvements in wind production forecasting will offer much better warning of wind changes, like the one that affected Texas in February.
Data from a soon-to-be released wind production model actually predicted last February’s event. Bojorquez said ERCOT is accelerating the process to integrate the forecast model into current tools.
“We will have more advanced notice of large wind changes and we think the accuracy of these tools will continue to improve.”
Bojorquez said he has no immediate concern over the potential to back off baseload coal plants at night. That possibility has been raised because Texas’ wind resource tends to be most robust at night.
“We are not currently into situations where we’re backing down baseload to take advantage of wind that picks up off-peak,” he said. ERCOT has enough gas units and load to use all the currently available wind resource, even at night. He said, however, that ERCOT is backing down both single- and combined-cycle gas units.
Bojorquez said that even backing down coal-fired generation would be better in terms of total production costs. That’s because the bigger problem is whether an operator can keep the coal units and ramp them back up in time for peak. That’s an eventuality that has not yet presented itself. A recently completed study by GE Energy did not identify backing down coal plants as an option. And power from nuclear plants is not backed down at all.
“If we get more nuclear we would have to look at backing down coal at night and possibly nuclear power too,” he said. “But that’s several years down the road when there may be enough load growth,” he said.SB