Wind Forecasting Key for More Efficient Turbines

Bill Mahoney with the National Center for Atmospheric Research (NCAR) said partnerships between utilities and applied meteorology firms can help minimize that obstacle.

Mahoney is program director with NCAR, which is working with Xcel Energy to provide detailed, real-time wind analysis specific to each of the company's 47 wind farms.

Utilities with wind portfolios of more than 10 percent or so are quickly learning how important accurate wind forecasts can be for wind generation, as well as how challenging it can be to build a reliable forecast model, Mahoney said.

Eric Pierce, managing director of energy trading at Xcel, said the models and forecasts will update themselves every 3 hours.

Xcel chose to work with NCAR after hearing about the wind forecasts the laboratory had provided to the U.S. Department of Defense, the Federal Aviation Administration and the U.S. Department of Homeland Security, Pierce said.

With a growing wind portfolio, Pierce said more accurate wind data is crucial for Xcel.

"Our portfolio continues to grow to meet renewable portfolio standards in Colorado, Minnesota and Texas. By 2020, we will have an estimated 7,400 megawatts (MW) of wind," Pierce said, adding the company has 2,850 MW of wind today.

Mahoney said that seemingly insignificant factors, such as whether there are leaves on the trees or snow on the ground in the area around a wind farm, can have a measurable impact on the amount of electricity the wind turbines can generate.

"Wind is one of the most difficult things to forecast," Mahoney said. "Changes in wind speed and directions, are impacted by the topography."

NCAR's high-resolution wind forecasting technologies can take these factors into consideration, he said.

In the data NCAR will provide to Xcel, global, regional and local data sets will be used. These data sets will be constantly updated with real-time wind measurements, he said.

"It will essentially be a learning system that uses the data it collects to improve its forecasts. The system will constantly tune itself," he said.

With this information, a utility can balance its load more efficiently, he said. With the trend of more states writing requirements for wind generation into their renewable portfolio standards, knowing when the wind resource can be relied upon and when another form of generation will have to be used instead, becomes more important.

"If you're expecting the wind to blow 25 mph all day and generate 1,000 MW, and that wind drops off or doesn't occur, you have to make up for that with another plant, and that is expensive to throttle up those other, more traditional generation sources," he said.

Mark Ahlstrom, CEO of WindLogics, an applied meteorology group, said private enterprise firms have provided this kind of data to utilities for years. WindLogics has worked with utilities, including Xcel, providing wind analysis for more than 300 wind energy projects.

Ahlstrom said measuring wind resources at a given site is done in two different ways: pre-construction assessments and operational forecasting.

In pre-construction assessments, available historical data is paired up with wind measurements from a meteorological tower to find out what the long-term energy capacity factor of the area is, Ahlstrom said.

"Once it's built, then your issue is different. You have to use real-time operational forecasting, running weather models through learning systems to figure out a power forecast for that wind plant," Ahlstrom said.

Many utilities use several different models from a variety of sources to provide the most accurate predictions of how wind will behave in the future, he said. Many different factors must be considered, he said.

For example, wind direction is an important factor because wind farms perform differently when wind comes from a different direction. The placement of the turbines becomes important here because the wind has to move around other turbines, he said.

Mahoney said he hopes NCAR's techniques will become a benefit to the rest of the renewable energy community, not just Xcel. He said he believes the technology transfer will help the entire industry, as well as electricity consumers.

"We will be advertising our results, positive and negative, so that the community can know what is available," he said.

Jeff Postelwait is the online editor at Power Engineering International. In that capacity, he manages news content for Power Engineering's website and edits the magazine's weekly electronic newsletter. He is based in Tulsa, Oklahoma.

This article was originally published by Power Engineering International and was reprinted with permission.