Wednesday, April 28, 2010

Predicting Wind Power's Growth -- an Art That Needs More Science

Weather's whims -- and the ability to predict them -- have become another hot button in the electric power industry's debates over the future of wind generation in the United States.

Wind generators now deliver about 2 percent of the nation's electric power. That could grow to 20 percent in 15 or 20 years, according to government studies -- but not without a substantial improvement in the speed and quality of weather and wind forecasting, experts agree.

None of the current computer models of the atmosphere is up to the task, according to the Pacific Northwest National Laboratory. "Current practices will not be adequate" as wind and solar generation expands, because the variability of these sources may significantly affect grid operations, the North American Electric Reliability Corp. (NERC) said. A "dramatic" improvement in forecasting is required, said Spain's Iberdrola Renewables, the second-largest U.S. wind power producer.

"Wind integration isn't a question of 'Can we?' It's economics, and who pays has to be determined," said Pascal Storck, vice president of 3TIER, the Seattle-based wind analytics firm. "This is a big, big issue."

Who pays has become another bone in the fight between wind generators and many established utilities throughout the grid. The question of whether wind generators should bear the expense of improved forecasting -- and whether the federal government should make a major investment in better weather monitoring and prediction -- is one issue.

Tied to it is a larger issue with a much higher price tag: If wind generators' forecasts miss the mark, and grid operators have to call for more costly coal- or natural gas-fired generators at the last minute to make up the gap, should the entire bill be sent to the wind companies?

Comments from the industry are flowing into the Federal Energy Regulatory Commission in response to its inquiry into regulations, practices or rates that may be discriminating against wind power or restricting its integration into the grid (FERC docket RM10-11). FERC's initiative gains in importance if congressional climate and energy legislation fails to pass.

Wrestling with an inexact science

"Anyone who has been caught in an un-forecasted downpour without an umbrella can attest that weather forecasting is far from an exact science," said Westar Energy Inc. in the FERC proceeding.

"Forecasts are simply expectations, not foreknowledge. Reliability of the electric system cannot be assured by reliance on forecasts," and improved forecasting will not eliminate the need for grid operators to hold backup generation in reserve to deal with sudden wind shifts, said Westar. The wind industry should bear that burden, it said.

"Between 10 and 30 knots, a wind forecast error of one knot may affect electric output by as much as ten percent," said MidAmerican Energy Holdings Co. Wind generators should be held accountable for the differences between their forecasts and the power they actually deliver, as coal- and gas-fired generators are. That is the surest way to improve forecasting accuracy, MidAmerican said.

3TIER's Storck said that wind assessments have been steadily improving. His company has been making forecasts for 10 years, and their accuracy has increased more than 30 percent over that time, thanks in large part to better data from wind farms.

WindLogics, a St. Paul, Minn.-based wind analysis firm, reported last year that a new forecasting system developed for Xcel Energy subsidiary Northern States Power had achieved an average error rate on hour-ahead forecasts of 8 percent, about half of the comparable National Weather Service figure. The system includes high-resolution wind models, generators' data on wind speed and power output, and advanced statistical analysis.

More data, better weather and atmospheric models, and more powerful computer runs are the paths to the next generation of forecasting systems. Public-private partnerships between private forecasting firms and the National Oceanic and Atmospheric Administration and the National Center for Atmospheric Research are crucial, industry officials say.

"We are all pretty much on the same page on how to move wind forecasting to the next level in the U.S.," said Mark Ahlstrom, CEO of WindLogics. The division of labor is settled, wind industry officials say. The federal government should generate weather modeling. Wind generators should use that to estimate their power output.

Could better forecasts pay for themselves?

"The government is tackling the hard pieces -- making improvements to the basic weather forecasting models and putting in better observational systems that allow them to make better forecasts. The private sector turns this into wind energy forecasts. There is great synergy between the federal agencies and us," Ahlstrom said.

Iberdrola said traditional government weather analysis can be strengthened by feeding in data from the Federal Aviation Administration's Automated Weather Observing System network, particularly if this system is beefed up to make more frequent observations.

Ahlstrom said a current goal is to improve high-resolution, rapid-refresh depictions of weather movements and wind speeds a few hundred meters above the ground. Currently, weather models are generally refreshed every three to six hours, at a coarse resolution. Hourly updates are needed, he said.

How far these efforts will go is not clear, some industry experts say. "We believe current practices can be improved substantially through science-based forecasting," said the Pacific Northwest National Laboratory. "The primary barrier to these additional measures is funding."

Wind advocates say the best should not be the enemy of the good. Analysis shows "that forecasts do not necessarily need to be perfect to be useful," said the National Renewable Energy Laboratory. For example, NREL said that by forecasting the delivery of 10 gigawatts of wind generation in the Electric Reliability Council of Texas, the state's grid could reduce electricity generation costs by $180 million a year, compared to costs if that much wind power is added to the grid without forecasting and scheduling. The advance notice of when wind power is expected makes the grid more efficient, according to NREL.

Despite the attention that weather and wind modeling is getting, some wind generators emphasize that the human element remains crucial. "It's becoming more and more clear that computer models can take us so far, but it's tough to beat the human eye and human mind in being able to see patterns developing, especially what's going to happen in the next few hours," Ahlstrom said.

Prediction: sharp gusts of controversy

A skilled control room operator can interpret radar displays of approaching thunderstorms better than computer models, said NERC. Indeed, it would make sense for control rooms in wind areas to have specialists on hand to assimilate and analyze wind conditions.

Since last year, Iberdrola Renewables has used meteorologists to interpret vendor forecasts. "Where there is a clear financial benefit to producing a better forecast than currently available [from vendors], Iberdrola Renewables produces its own forecasts," the company said.

The questions put to the industry by FERC's staff in its wind integration inquiry show the commission's interest in improving forecasts as a way to help wind power compete.

Better forecasts could help grid operators anticipate more accurately when large up-and-down ramps are about to arrive. Earlier, more accurate warnings could lead to the use of more economically efficient backup reserves, thereby ensuring that power rates "are just and reasonable," FERC said.

In parts of the country like the Southeast, where utilities control the transmission grid, power is typically scheduled an hour ahead, and changes in wind schedules aren't allowed within the hour except for emergencies, FERC said. Transmission schedules for wind generators are typically set 20 or 30 minutes before each hour, so forecasts may be 90 minutes old by the time the hour is up.

Would scheduling power within hour intervals, coupled with better forecasts, help level the playing field for wind? Is holding to the current practices imposing unjust and unreasonable charges on renewable generators? FERC asked. These are not simple questions, and opposing sides of the wind dispute are braced for an expensive fight. The answers will continue to stir sharp gusts of controversy.

No comments:

Post a Comment