You and I might be able to ignore the nightly weather forecast. But for power companies, grid operators, and energy traders, forecasts have never been more important.
As more renewable energy comes on-grid, forecasts of future wind and solar conditions could be the difference between a reliable power system and one that’s expensive and unpredictable.
Small fluctuations can make a big difference. In Denmark—which gets about 20% of its power from wind—a one-meter-per-second change could mean a 450MW shift in national output, according to one estimate. That is the equivalent of a whole coal-fired power station.
"If you are a utility or power provider, if you know how much energy you’ll have from wind, you can shut down your fossil energy," says Iheb Triki, business development manager at Natural Power France, one of several consultancies offering wind power forecasts. "You can reduce your reserve costs, and your imbalance charges, and you can improve your planned dispatch scheduling."
By "dispatch," Triki means generating plants that can be switched on and off when needed. If grid operators can run back-up power stations optimally—and turn off plants, rather than running them on standby—they can reduce both carbon emissions and costs significantly.
According to the Bulletin of the American Meteorological Society, better wind forecasting could cut U.S. energy costs by between $1.6 billion and $4.4 billion, depending on the level of production. Those savings come from reducing the need for back-up power, better integration of renewable energy with the grid, and reducing the need to turn to spot-markets, where energy prices are highest.
Such numbers are not lost on the government. Several U.S. agencies now have projects to improve forecasting. The Department of Energy and the National Oceanic and Atmospheric Administration are working with two private companies on the $6 million, year-long Wind Forecast Improvement Project, which is focused on atmospheric conditions well above ground. At the moment, forecasts tend to extrapolate wind patterns at the top of turbines based on on-the-ground data, which can introduce a lot of inaccuracy, studies have shown.
The National Center for Atmospheric Research (NCAR) is also developing new models and systems. Since 2009, it has been working with Xcel Energy, which claims to have saved $6 million last year from improved forecasting. NCAR’s system uses data from satellites, aircraft, weather radars, weather stations, and turbine-mounted sensors. It then generates simulations of future atmospheric conditions, combining those with turbine performance data to produce actual power forecasts.
"Wind is challenging because of the impacts it can have on our operations due to its intermittency," says Eric Pierce, Xcel Energy’s managing director of energy trading operations. "This new forecasting system will enable us to harness wind far more effectively while saving millions of dollars for our customers."
Natural Power has signed up 20 wind farms in the U.K. for its ForeSite service, which combines data from a meteorological forecaster, Meteoblue, with two years of production figures for each site. It can produce forecasts for between 30 minutes and five days ahead at various levels of accuracy.
Forecasts are also useful for deciding on the best times to shut down turbines, either for maintenance, or because of extreme weather that might damage equipment, Triki notes. Last week, 165 MPH gusts in Scotland caused a 100-meter-high turbine to fall to the ground and catch fire. Better forewarning might help avoid such accidents.
"There have been a number of events in the last few months to warrant a better forecasting service," says Alex Woodward, Natural Power’s marketing manager.