Power electronics and exotic energy storage devices are making wind power steady enough to compete with conventional electricity sources
In this season of discontent in the electricity business, only wind power seems to stand out as a global success story. While petroleum prices were convulsing in response to war and labor strife, and nuclear plants were stoking controversy in the Middle East and Asia, wind turbines were quietly becoming the fastest-growing energy source in the world. They now provide more than 31 000 MW of power, a total that has swelled by almost 30 percent in scarcely a year’s time and that keeps more than 200 million tons of carbon dioxide out of the atmosphere every year. Wind power’s ascendance has been so stunning that advocates are now rallying around an idea that would have seemed preposterous just a couple of years ago: that the wind could supply 12 percent of the world’s electrical demand by 2020.
Impressive as the gains have been, it isn’t quite clear yet that the wind can blow away the developed world’s fossil-fuel dependence. One of the most important reasons is that clean, renewable wind power comes with a serious hitch: while conventional power plants yield a steady stream of electricity, wind turbines often ply turbulent gusts and therefore spit out an irregular stream of electricity that is tough for power grids to swallow.
Now, though, high-tech solutions are at hand. Systems based on advanced power-electronics and energy storage devices are massaging and managing power flows from wind turbines, enabling them to contribute mightily to electricity grids without putting those grids at risk. Not only are the technologies making wind power more palatable to grid operators, they are even making it possible for engineers to finally harness wind energy’s tremendous potential in wind-swept, remote locales.
Perhaps nowhere is this potential so evident as in the state of Hawaii, whose isolated power grids could not otherwise risk taking full advantage of the archipelago’s abundant, renewable resource. In fact, with its lush, endless trade winds and growing commitment to wind power, Hawaii’s Big Island is emerging as a laboratory of the future of the technology. As wind power becomes a steadier and more reliable resource, it could help wean power producers all over the state from their dependence on costly imported oil.
But, for now, says Karl Stahlkopf, chief technology officer at Hawaii Electric Co., in Honolulu, even the existing wind farms on the Big Island—putting out just
10 MW, the equivalent of four state-of-the-art wind turbines—make grid controllers hop on days when the palm fronds fly.
The utility and its contractors plan to build what Stahlkopf calls an ”electronic shock absorber” to buffer the island’s power grids against the wind’s worst behavior. It’s a development that engineers elsewhere are following closely.
The reason is that the solutions to integrating modest levels of wind power on small, isolated grids today may foreshadow the installation of truly large-scale wind power in mainland networks five or 10 years from now. ”What’s happening in the Hawaiian Islands is a peek at the future,” says Bob Zavadil, an expert on wind power at the Arlington, Va.-based power systems analysis firm Electrotek Concepts Inc. ”They’re on the leading edge.” And that’s true not only of the technology but also of the new legal and regulatory conventions between utilities and independent power producers that will be needed before wind energy can truly thrive.
Reactive Power 101
Back on the mainland, wind farms have grown to dozens of turbines and hundreds of megawatts—rivaling the size of conventional power plants. To pave the way for installations like those, engineers had to grapple with the tendency of wind turbines to introduce voltage instability into electrical grids. That tendency follows from the intermittent nature of wind-generated electricity, which waxes and wanes in irregular episodes unrelated to the predictable daily ebb and flow of consumer demand.
To understand the instability issue—and its solution—you’ll need a few basics on the nature of power flows on utility grids. First, recall that these flows consist of both active and reactive power. The difference between them arises from the fact that the wave of alternating current on a power grid almost always leads or lags the voltage wave. In short, inductors, such as the coils in motors or transformers, cause current to lag voltage; capacitors cause current to lead voltage.
Active power is the familiar watts consumed by light bulbs and toasters; it is the product of voltage and the component of current that is in phase with the voltage. Reactive power, measured in volt-amperes reactive, or VARs, is the product of voltage and the out-of-phase component of alternating current. This out-of-phase power is consumed by energetic electric or magnetic fields—for example, in coils in inductive loads, such as motors, and also in the utility’s own transformers. Transmission lines themselves are also inductive. In the case of wind-generated power, this line inductance can be a critical factor, because the turbines are usually far from population centers, strung out across windy plains or even areas out at sea.
Utilities have to pay close attention to reactive power because it determines, along with active power, how the voltage declines, or ”sags,” on an electric network. Utilities must keep voltages very close to rated values because when they fall too low, motors and other inductive loads draw too much current, overheating the equipment and possibly damaging it and other utility gear as well.
Declining voltage on a network is a function of the consumption of both active and reactive power. In other words, if a big load consumes lots of reactive power, it will cause the voltage to sag intolerably unless that reactive load is adequately compensated for. Because these reactive loads are almost always inductive, causing current to lag voltage, utilities must feed them with reactive power in which the reverse is true—with current that leads voltage.
For this they have several options. They can supply reactive power right from the generators themselves, by producing current waveforms that lead voltage. They can also install banks of capacitors close to big inductive loads. As the load rises, relays switch on the capacitor banks to add reactive power.
So what does all this have to do with wind power? Many modern wind turbines generate electricity with an induction generator, also known as an asynchronous generator. Conventional, fossil-fueled power plants, on the other hand, are built around synchronous generators. There’s a big difference between the two kinds of generators: a synchronous unit can provide both active and reactive power, as noted above. An induction generator, on the other hand, provides active power but consumes reactive power. This characteristic, coupled with the wind’s intermittence, plays havoc with utility grids.
”As the wind comes and goes, the voltage goes down and up right along with it,” explains Bud Kehrli, manager of transmission and distribution planning for the power equipment maker American Superconductor Corp., in Middleton, Wis. You could handle the problem by switching capacitor banks in or out, but it takes lots of nimble switching to keep up with the ever-varying gusts. And switching a big block of capacitance in or out can swing the voltage up or down a few percent or more, a variation that is felt as an abrupt change in torque on the turbines’ gearboxes. Slamming a gearbox like that repeatedly is a bad idea because it soon wears it out, and replacing one costs about US $200 000 in parts, labor, and other expenses.
These are exactly the problems that power semiconductors are solving from Andalucia to Abilene. Consider the Foote Creek Rim wind project in southeastern Wyoming, one of the biggest in the United States. The installation, which feeds grids operated by the utility giant PacifiCorp, in Portland, Ore., includes 183 induction turbines installed over the last four years, which generate up to 135 MW—enough for 25 000 households.
In choppy winds, the switching capacitor banks were often a step behind the wind, leaving an excess or shortage of reactive power on the local 230-kV transmission line, a major conduit through which PacifiCorp moves power from large coal-fired plants in Wyoming to customers in six western states. As a result, voltage on the line would spike up or down by as much as 5 or 6 percent.
Until this year, PacifiCorp stabilized those lines by brute force: it simply idled dozens of wind turbines at Foote Creek, throwing away cheap, clean energy in order to ship its coal-fired power. But this past February, the company turned to power electronics for a more satisfying solution, installing a $1.5 million system of silicon switches designed and built by American Superconductor to dynamically adjust reactive power at Foote Creek.
The new system is called D-VAR, for dynamic volt-amperes reactive. It can provide up to 8 mega-VARs of reactive power continuously (or 24 MVARs momentarily). Best of all, it can inject this reactive power in increments of as little as 0.027 MVAR, and it can turn on a dime, going from, say, -8 MVARs to +8 MVARs in less than one electrical cycle, or 1/60 of a second. At Foote Creek, the D-VAR system also controls nine banks of capacitors, totaling 50 MVARs, switching these capacitors on and off and interjecting its own MVARs as necessary to provide a smooth and essentially continuous range of reactive compensation up to 58 MVARs [see illustration].
The secret of the D-VAR’s success is the insulated-gate bipolar transistor (IGBT), which was invented at about the same time, 20 years ago, at General Electric Co. (Fairfield, Conn.) and at the former RCA Corp. IGBTs are like high-power cousins of the transistors in your computer or cellphone, and they function in the D-VAR system as rapid power valves, opening and closing many times per cycle to produce an alternating current waveform that mimics the effects of an ultraprecise, adjustable capacitor bank.
Inside the D-VAR is a direct current bus whose voltage of 750 V is maintained by transforming and rectifying voltage from the external bus fed by the wind turbines. The IGBTs, controlled by microprocessors, switch between the dc bus and the external bus at a pace fast enough to establish a carrier frequency at 3 kHz.
Switching in elaborate, precisely timed patterns, the IGBTs produce sine waves at 60 Hz by means of a standard technique called pulse-width modulation. The longer the IGBT switches dwell in the ”on” state, the closer the sine wave gets to an amplitude peak. And the longer the IGBTs are off, the closer the wave is to zero amplitude. Before the sine waves are transmitted on to the external bus, they are filtered to remove the 3-kHz carrier wave, producing smooth 60-Hz waves.
By adjusting the on-off patterns of the IGBTs, the system controller can make the current sine wave lead or lag the voltage by any degree desired. However, D-VAR is set up so that it produces a waveform in which the current sine wave leads the voltage sine wave by 90 degrees, thus mimicking the effects of a capacitor bank. To vary the amount of reactive power the D-VAR is supplying—the ”capacitance,” you could call it—the controller simply adjusts the amplitude of the current.
Given the commitment to induction generators on the part of heavy hitters like Japan’s Mitsubishi Heavy Industries Ltd. (Tokyo), substations will provide a ready U.S. market for IGBTs and other power semiconductors from American Superconductor and rivals ABB Ltd. (Zurich) and Siemens AG (Munich) for years to come.
But in Europe, wind farm operators are taking a different approach to harnessing the wind—and to handling wind turbines’ thirst for reactive power.
The impetus had nothing to do with reactive power. Developers were just trying to reduce the mechanical strain on large, fixed-speed turbines, whose rotors can stretch wider than the wingspan of a jumbo jet. Their solution involves building the silicon switches right into the turbine.
What makes it all possible is a shift in the technology of the turbines themselves: instead of induction generators, wind farm operators are now more and more favoring variable-speed generators that can absorb the force of a gust by speeding up. IGBTs and other power electronics in the turbine convert the variable-frequency ”wild” ac from the generator into a steady sine wave synchronized to the grid (50 or 60 Hz, depending on where you are).
The beauty of this approach is that the same electronics can be programmed to simultaneously convert some of the turbine’s active power into reactive power, as the network demands.
The idea has been slower to catch on in the United States, where GE Wind Energy, in Tehachapi, Calif., has deftly defended patents on variable-speed turbines that will be on the books through 2011.
Whether the power electronics are inside the turbine or in a substation, however, they are being used in new ways as wind farms become larger, more critical components of power systems. ”Utilities are generally lining up to require that wind farms be able to provide dynamic reactive compensation, much as a conventional generator would be able to do,” says American Superconductor’s Kehrli.
That means being able to pitch in and help restore stability during a disturbance or some other crisis on the grid. In the past, when a generator failure or momentary short-circuit roiled a network, wind farms would automatically disconnect themselves. But now, most grid operators ”want the turbines to ride through the disturbance,” says Chuck McGowin, manager for wind power technology at the Electric Power Research Institute, a utility-funded R&D consortium based in Palo Alto, Calif. ”It’s a big shift.”
The movement began in Europe, where, since this past January, for instance, the German utility E.ON Netz GmbH, in Bayreuth, has mandated a ”ride-through” for the more than 5 GW of wind farms in its territory. That is, the turbines must continue to operate during ”faults” when the voltage sags precipitously or the frequency deviates from 50 or 60 Hz. In recent months, utilities across the United States have begun to jump on the bandwagon. Suddenly, wind farms and turbine developers must alter their designs and operating procedures, particularly for voltage faults.
According to Craig Quist, a principal engineer at PacifiCorp, most wind turbines are programmed to disconnect themselves from the grid if voltage drops by 30 percent for 50 milliseconds—just a few cycles in the power wave. Once again, the ultimate solution may lie in power electronics. Substation semiconductors or bolstered on-board electronics can help wind farms ride through such events by generating extra reactive power to hold up the local line voltage.
No more trouble in paradise
Riding through voltage faults is just one of the challenges confronting Hawaii’s Big Island, where Hawaii Electric and independent developers are considering proposals to boost wind power generation to 30-40 MW on a 150-MW grid. If it happens, the Big Island will be tapping wind power in the same proportion as Denmark, which gets about 20 percent of its power from the wind. But the Hawaiians will be doing it on a comparatively tiny grid lacking the stabilizing embrace of a continent-wide power grid.
The main problem is that bigger wind farms would amplify the impact of the wind gusts that are already taxing Hawaii’s grid controllers. A 20-MW wind farm, for example, could surge by over 2 MW in just two seconds—much faster than the island’s oil-fired generators could ramp down or up to balance supply and demand. Even momentary sporadic mismatches between supply and demand are unacceptable because they would alter frequency and voltage on the network, putting equipment at risk and causing lights to flicker.
To make sure that doesn’t happen, Hawaii Electric airlifted Karl Stahlkopf to the island last year. He was a natural for the job, having led the development of a variety of power-electronics-based systems as vice president for power delivery and transmission at the Electric Power Research Institute. The giant shock absorber that Stahlkopf envisions for the Big Island would mediate between the power grid and the turbines, and it would combine power electronics with an advanced energy storage device, such as an ultracapacitor or a battery.
Slamming a gearbox repeatedly is a bad idea, because replacing one costs about $200 000
When the turbines are going full bore, Stahlkopf explains, the power electronics will divert some power into the storage system, drawing it out again when the wind dips. If the line voltage drops in a fault, the power electronics will dig deeper into the storage reservoir to generate reactive power and prop up the line. For the project, Hawaii Electric, the state, and the U.S. Department of Energy are now considering various storage technologies and capacities, and Stahlkopf reports that a shock-absorbing system could be on-line by the end of next year.
Meanwhile, wind farm developer Apollo Energy Corp. (Foster City, Calif.) is negotiating with Hawaii Electric to double its 9.8-MW Big Island wind farm at South Point and it has proposed to connect the added capacity to the grid via a shock absorber similar in concept to Stahlkopf’s.
The heart of Apollo’s system would be 30 flow batteries from Menomonee Falls, Wis.-based ZBB Energy Corp. Flow batteries are a hybrid between electrochemical batteries and fuel cells. They use pumps to circulate a pair of electrolytes past an ion-exchange membrane similar to the ones employed in many fuel cells. Ions pass across the membrane from one electrolyte to the other to charge and discharge the battery.
The flow batteries in the system Apollo envisions could back up a 20-MW wind farm for several minutes. It wouldn’t be cheap; building and operating the battery system could cost $1.8 million over its projected 17-year life span. But by keeping the wind farm going during the 200-plus hours each year when erratic winds would otherwise force operators to shut down some turbines, ZBB and its partners insist that the battery storage system would more than pay for itself—to the tune of $5.4 million over its lifetime.
Hawaii’s wind dreamers are already looking beyond these grid-stabilizing systems to larger storage devices that could negate the ultimate challenge facing utilities with lots of wind power: keeping the juice flowing during the hours, days, or even weeks when the air is not. On average, wind turbines produce their rated power only 20 to 30 percent of the time. Really big storage systems, capable of storing at least tens of megawatt-hours of energy, could make wind power almost as ”dispatchable” as that from a fossil-fuel plant.
Several years ago, Apollo Energy proposed installing a 10- to 11-MWh flow battery on the Big Island. And Stahlkopf has launched a research collaboration with local water authorities to test a lower-tech, but potentially more capacious, option. The idea is to piggyback on the Big Island’s existing infrastructure of water mains and reservoirs to fashion a pumped-storage system that would push water to higher-altitude reservoirs when the wind was strong, and then let it fall through hydroelectric turbines to produce power when the air was still.
Rewriting the rule books
With power semiconductors and other technologies rapidly minimizing wind power’s technical shortcomings, the hurdles increasingly seem to be contractual and regulatory. Utilities such as PacifiCorp, which had to foot the bill for the D-VAR installation that put an end to the voltage instabilities at Foote Creek, are rewriting their contracts to put the burden of integrating wind farms on their developers. ”We got stung on the first wind farm, and we’re not going to get stung again,” declares PacifiCorp’s Quist.
Lacking detailed standards for wind farm behavior, negotiations between utilities and wind farm developers can get ugly—even in paradise. On the Big Island, Hawaii Electric and Apollo Energy have been at odds for four years over Apollo’s proposal to install newer, larger turbines at its wind farm. Its game plan to add the flow batteries didn’t help matters. Apollo contends the battery system will make the wind installation a ”firm” source of power—in other words, one that can dispatch power to help shore up the network in a crisis. State regulations generally specify that utilities must pay more for firm power.
Stahlkopf replies that Hawaii Electric isn’t yet convinced that the flow battery will adequately protect the Big Island’s grid. But he is confident that the two sides will find common ground in their negotiations.
Whatever agreement ultimately emerges, it is likely to push wind energy that much closer to the mainstream. And when regulations catch up to technology, wind energy—clean, renewable, and unaffected by geopolitical conflict—should come into its own as a steady energy source for these turbulent times.