Utilities working to harness the wind must also become better meteorologists if they are to grow their renewable energy portfolios and meet a possible federal mandate.
Toward this end, the nation's largest wind power provider, Xcel Energy, yesterday teamed with two major research facilities, the National Center for Atmospheric Research (NCAR) and the National Renewable Energy Laboratory (NREL), to improve day-to-day forecasts at its wind farms.
If the 18-month project is successful, wind forecasting companies around the world might use the new tools to make utilities more secure in relying on this notoriously fickle renewable resource.
"Accurately forecasting our wind power generation will allow Xcel Energy to reliably bring on more wind energy and reduce costs at the same time," said Eric Pierce, the company's managing director of energy trading.
Other utilities may soon have the same needs if Congress passes renewable portfolio standard legislation being introduced by Democratic leaders to ramp up clean electricity production, as many states have already done.
Today's forecast: variable winds and a 20% error rate
Pierce explained that, today, Xcel Energy's wind forecasts typically have a 20 percent error rate, a number that is unacceptable if the company is to more than double its wind production by 2020, as planned.
To avoid service disruption, utilities must match electricity supplies with customer demand at any given moment, night or day. "We are not like the phone company," said company spokesman Tom Henley. "They can just give you a busy signal." Too much electricity can overload the system, and too little brings power outages.
But wind, unlike traditional fossil fuels, must be consumed while the breeze is blowing. Researchers and companies are now developing better batteries and energy storage technologies to help solve such variability issues, but this ability is not available today.
In its absence, utilities try to foresee at least a day in advance when gales will blow or when the air will stagnate. The latter is the more risky scenario. If a slow day catches a utility off-guard, it is forced to fire up backup energy generation, which for Xcel Energy is usually natural gas but for other companies could be coal or nuclear power.
If this can't happen quickly enough, the utility buys electricity from a pricey "spot" market, an expensive proposition for customers. Pierce estimated that if the wind prediction project goes as planned, it could result in $10 million to $12 million in savings for Xcel's customers over two years.
In the longer term, he said, better prediction will allow utilities to maintain or increase their reliability while integrating more wind power.
But even with perfect predictions, that doesn't change the fact that sometimes the wind just doesn't blow. Xcel Energy plans to back up the 800 megawatts of wind it will develop in Colorado over the next seven years with an equal capacity to generate electricity with natural gas, said Henley.
For now, this means that along with more wind comes the need to at least build, if not necessarily always use, some new fossil fuel-powered facilities as a backup.
Underestimate the weather and you could get brownouts
Michael Goggin, an electric industry analyst with the American Wind Energy Association, disagreed, however, that such a one-for-one approach was necessary. He said that there are many other operational measures that a utility might take to even out renewable electricity ebbs and flows, adding that utilities are already well-equipped to deal with all kinds of supply and demand fluctuations.
But on a cold night in February, three years ago, Xcel Energy was not prepared, said William Mahoney, the project's director at NCAR. Overnight temperatures were unexpectedly frigid, and rolling brownouts ensued.
While the cause of the electricity imbalance was temperature, not wind, the incident brought Xcel Energy in touch with NCAR for weather consultation. "Everything we do is affected by weather," said Pierce. The company returned to NCAR last March when it realized that its wind prediction was not up to par.
To develop the new system, NCAR will rely on its cutting-edge computer models, blended with the latest data sets, taken at all scales -- locally, regionally and globally and from the wind farms themselves. Because wind is affected by everything from topography to atmospheric conditions and even the number of leaves on trees, it is one of the ultimate forecasting challenges.
The forecasts will be made for winds at the heights of the turbine blades themselves -- usually 200 to 400 feet -- where winds blow stronger than at the ground-level stations that are relied upon today.
NREL will then develop tailored formulas for converting wind speeds and directions to the actual energy output from the turbines.
"Up until recently, wind farms have been treated as black boxes," said Mahoney. Soon, that will no longer be the case, he hopes.