Taking Wind Mainstream
Given wind's intermittency, can the power grid handle much larger amounts of variable generation?
Photo: Hawaiian Electric Co.
The potential of wind power to help meet America's growing demand for electricity is staggering: According to a definitive 1993 study by the Pacific Northwest National Laboratory, areas of strong winds cover about 6 percent of the mainland states and, if exploited, could supply more than current U.S. electricity consumption. Conversely, just 0.6 percent of the land of the contiguous 48 states would have to be developed with wind turbine farms to provide 15 percent of the nation's electricity requirements. Even then, less than 5 percent of the developed land would actually be occupied by wind turbines, associated electrical equipment, and access roads. In most cases, existing land uses, such as farming and ranching, could remain as they are now.
Harnessing this potential could make an enormous contribution to reducing the United States' dependence on imported oil for power generation, as well as helping to stem an increasing dependence on imported liquefied natural gas. Adding wind power to the grid can help stabilize electricity prices, too, in the face of escalating fuel costs, with oil topping US $72/bbl and natural gas prices at all-time highs. Wind power, in contrast to oil and natural gas, is both price- competitive and price- stable , and that stability can help provide a cap on the price of electricity. And let's not forget that wind power is a whole lot cleaner than fossil fuels--be they imported oil or domestic coal.
Yet one often hears questions related to wind power's intermittent nature; unavoidably, electricity is generated only when the wind blows. Can the power grid handle massive amounts of variable production? Can wind energy be delivered where it's needed when it's needed? Can wind energy harnessed at times of low demand be stored for high-demand periods? Can new storage technologies be devised so that wind energy would become, in effect, dispatchable? The answer to all of these questions is yes, and in some cases the answers are already in practice.
Wind-energy and power-transmission technologies are already adapting to accommodate the impressive growth of wind power. Large semiconductor devices referred to collectively as power electronics are, for example, enabling wind farms to provide rapid response to fluctuations in grid frequency and voltage. This is one of many reasons why grid studies consistently estimate that the cost of integrating wind power will be low. However, integration costs will rise when one considers small power grids or high proportions of wind power in a grid. In such cases, power electronics devices can be combined with energy storage technologies that operate over a range of time scales to manage the shifts in wind power production. In fact, a growing number of innovative energy storage options are providing grid operators ways to dispatch wind power in the same way they do with thermal generating plants. Continental supergrids eventually will help, too, by distributing wind-generated power across whole regions, balancing regions where the wind happens to be blowing with those that may be becalmed, while simultaneously spreading the burden of providing backup power.
What follows is a taste of the technology and policy strategies that are already helping to give wind power new strategic importance, and which will be critical to sustaining its growth in coming decades.
Electronic "Shock Absorbers"
Hawaii provides a perfect laboratory for demonstrating future technologies for integrating wind power with the grid. Each island must be self-sufficient in electrical energy, because there are no large regional power systems to provide backup, as on the mainland. When wind generation reaches 510 percent of one of these networks' total generating capacity, sudden changes in wind speed can result in a loss of power generation great enough to trigger automatic load shedding, disrupting consumers. This is the challenge that now faces Hawaiian Electric Co. (HECO), headquartered in Honolulu, as it adds enough wind-energy installations on its Big Island grid to bring the total wind generating capacity to 30 MW. The proportion of the Big Island's load demand supplied by wind energy could soon exceed 15 percent during peak demand hours and 30 percent during hours of minimum load.
The current approach to preventing such wind-triggered load-shedding is to add conventional power capacity to the system. As an alternative, HECO has launched a proof-of-concept demonstration combining power electronics and energy storage technology, which we call the Electronic Shock Absorber (ESA). HECO's first ESA demo unit, installed at the Lalamilo Wind Farm on the Big Island, has been operating since early January.
The ESA is programmed to respond in three scenarios. It absorbs power briefly when it detects a sharp increase in the instantaneous output of the Lalamilo Wind Farm, such as one caused by a strong gust of wind. Conversely, a lull in the wind lasting a few seconds will cause the ESA to inject power into the system. More gradual deviations from the average power output that exceed specified limits can also prompt the ESA to absorb or release power accordingly.
Importantly, the ESA can regulate reactive power--the product of current on a transmission line that is alternating out of phase with its voltage. Reactive power is consumed by the energetic fields in transmission lines; too much or too little reactive power can cause a line's voltage to spike or sag, respectively. By regulating reactive power, ESA can compensate for voltage changes on the grid to improve both power quality and system stability. Commercial power electronics built into wind turbines or in stand-alone wind-farm-scale units do provide such reactive-power support, but the ESA's energy storage capacity means it can do that much more.
The ESA consists of an inverter connected in shunt to the power line (that is, it branches off the line) that interfaces between the power line's alternating current and a direct-current energy storage component. For energy storage, HECO selected ultracapacitors, which store energy electrostatically by polarizing an electrolytic solution between two highly porous conductors. Thanks to the ultracapacitors' compact design, an ESA capable of matching 1525 percent of an average wind farm's output for 15 seconds to 1 minute may be small enough to mount on a truck trailer. The ESAs should also require little or no maintenance, and they have a much more favorable cycle life than electrochemical batteries. The demonstration unit at Lalamilo has a 1-MVA continuous rating, but it can provide 330 percent more reactive power for up to 2 seconds. To serve larger wind farms, modular units resembling the ESA at Lalamilo can be ganged together, accommodating 232 MVA ratings at system voltage.
Since its commissioning in January, the demo ESA has performed as designed and has provided significant stabilization to the HECO grid. It has counteracted power and voltage fluctuations from the wind farm, as well as fluctuations on the grid caused by conventional generators or load interruptions. Further tuning of the device continues as we seek to optimize the ESA's response to a variety of conditions, particularly those that would be encountered on grids with greater wind contribution and larger wind farms.
Whereas power electronics with some storage capability like the ESA can handle fluctuations on a subseconds-to-minutes time frame, other solutions can help accommodate wind power's minutes-to-hourly fluctuations. Today, such fluctuation is generally accommodated in much the same way as fluctuating demand from consumers is handled: by ramping conventional power plants up and down. This, of course, means burning fuel--a solution that requires the maintenance of extra power-generating capacity. The cost of such backup, while currently negligible, will rise as the percentage of wind energy on the power grid multiplies--a situation that is already imminent in Hawaii. Energy storage offers a backup solution that is potentially less costly, as well as being truly renewable.
HECO is evaluating the potential of adding what is known as a pumped hydro system, in which energy is stored by pumping water from a lower reservoir to a higher one using excess generating capacity during off-peak hours. Then, during peak hours, the water is released and flows back down through a hydroelectric turbine, generating additional power as needed. This type of system could smooth out hour-to-hour fluctuations at a wind farm. Recognizing this potential, HECO is evaluating the installation of 2050 MWh of pumped hydro storage capacity to accept energy from three major wind farms; the storage will compensate for variable winds by storing energy available during off-peak hours for use during periods of high demand.
Batteries present another option for backing up wind without firing up oil- or gas-fueled plants. Whereas locations with suitable terrain for pumped hydro storage are limited, batteries can be placed almost anywhere. Strategically placed battery installations could help smooth power supply from a wind farm while also easing power management concerns for the transmission grid: wind power produced outside of peak consumption hours can be delivered to battery installations over the power lines during off-peak hours, when the lines have spare capacity.
Consider an energy storage project under way at the New York Power Authority in the New York City area. This project, in which HECO is participating, uses a battery to reduce peak demand on the heavily loaded Long Island grid. Specifically, a 1.2-MW sodium sulfur battery with a storage capacity of 7.2 MWh, manufactured by Japan-based NGK Insulators, is being installed at the Long Island Bus Refueling Station. By charging this battery during off-peak hours, the customer can run energy-intensive fuel compressors during midday hours without pulling power over the congested grid, thereby avoiding peak demand charges.
Ultimately, the largest single cost associated with integrating large amounts of wind power into a utility grid will occur not over seconds or hours, but on a time scale of days or longer. Better coordination of power system resources can help reduce the costs associated with this uncertainty.
Improved day-ahead wind forecasting is helping to reduce costs by better anticipating and preparing for wind power supply, thereby helping to avoid the need to fire up a system's most expensive generating plants. New analytical methods are being applied, including artificial-neural-networks technology and physical wind-flow models.
The California Independent System Operator (ISO) is so confident in the algorithms it developed for hourly and daily wind generation forecasting that it has incorporated them into the California power market. The California ISO's Participating Intermittent Resources Program allows wind farms to bid into the California electricity market without incurring the ISO's 10-minute imbalance charges--penalties for deviations between energy promised and energy delivered. Wind producers that use the California ISO's wind forecasts to schedule their energy deliveries are charged instead for their monthly net deviation. Because many of the short-term over- and under-production figures cancel out over this time, the monthly charges tend to be small.
Coordinating wind power and hydropower is another option for handling daily and monthly variation. Where hydropower stations are available, utilities are increasingly using them as a form of energy storage and to buffer fluctuations in wind generation. Hydro-rich Sweden and Norway already provide a partial backup for wind-rich Denmark and northern Germany. A somewhat similar pattern is found in the U.S. Pacific Northwest, where the Bonneville Power Administration (BPA) supplies huge amounts of electricity to California and the entire U.S. West from one of the world's premier hydropower complexes, the Federal Columbia River Power System.
BPA's Network Wind Integration Service is part of a major BPA program designed to meet growing demand in the Pacific Northwest. The Columbia River hydro system experiences significant seasonal variation, owing to the region's dry summers. By integrating significant amounts of wind power into its system, BPA hopes to conserve its hydro resources and, in the same stroke, transform wind energy into dispatchable generation capacity for peak-load periods. Wind-farm operators help finance the system by paying BPA an integration fee of $4.50/MWh.
The realization of wind power's potential in the United States will require bold efforts to coordinate power distribution on a continental scale. That's because our most promising area for development by far is the upper Midwest. Harnessing this energy resource will require massive new transmission facilities to carry the electricity from primarily remote areas to urban load centers. Eventually, one can envision a massive high-voltage dc (HVDC) transmission "spine" taking shape along the windy corridor stretching from the Dakotas to Texas, with branches extending east and west toward distant cities. Similar concepts are already gaining credence in Europe. For example, Dublin, Ireland-based wind farm developer Airtricity and Swiss engineering giant ABB are promoting the concept of an HVDC undersea supergrid running from Spain to the Baltic Sea. Such long-term grid connections could serve to further smooth out the fluctuating supply of wind energy, since the wind is usually blowing somewhere .
As such new technology--not only extensive HVDC but also electronic shock absorbers and battery and hydropower storage--enables large-scale integration of wind farms with utility grids, wind energy will increasingly displace fossil fuels for power generation. Part of this displacement will take place organically, as new technologies are introduced and the cost of producing wind energy continues to decline. But rising concerns over energy prices and resource security are likely to significantly accelerate the trend.
What is required is appropriate planning--on the R&D side, to continue reducing the cost of the technologies described above, and on the transmission side, to intelligently integrate these technologies into grids as new wind farms come online. Doing so will promote maximum penetration of this clean, renewable resource.
About the Author
Karl Stahlkopf is senior vice president and chief technology officer at Hawaiian Electric Co.