As an energy source, the wind holds a lot of promise. It doesn’t pollute, it doesn’t cause climate change, and it doesn’t use or contaminate water.
But wind has a huge drawback: it is unreliable.
More than two dozen states have mandated renewable energy standards stating that a certain percentage of electricity must be generated by renewable sources like wind, and the nation may soon have a similar RES. But in order to meet electricity demand, power generators will have to be able to predict how much power the wind will generate an hour, a day or a year from now.
Stepping in to meet that need is the evolving technology of wind power forecasting.
Several business and government labs have started working on ways to make wind power forecasting more accurate to ensure a steady overall power supply.
GE Energy, the largest U.S. supplier of wind turbines, and WSI Corporation, a Weather Channel company, announced a new partnership Tuesday to try to incorporate wind and turbine data into wind forecasting technology.
In December, Argonne National Laboratory in Illinois hired Portuguese company INESC Porto to develop a wind forecasting model for the whole country.
Accurate wind forecasting could reduce the amount of fossil-fuel-based power people use. It's a power load-balancing act, explains Ken Pennock, forecasting business manager at AWS Truewind. When the wind is producing power, utilities can dial back the amount of coal and oil they burn, but they need to know when and how much.
“Historically, the power supply that is putting energy on the grid is a solid stream of resource. You burn coal to make electricity, and when you need more, you burn more coal,” Pennock says.
“In the case of wind, it’s a variable resource. We need to be able to predict when the wind will blow and when the wind turbines will spin and produce energy to put that power on the grid. If we have renewable energy like wind-producing power, we need to ramp down the fossil fuel sources so we don’t have too much power on the grid.
"When wind power goes down, we need to make sure other sources get spun back up in order to prevent blackouts and brownouts.”
An accurate forecast allows grid operators to maximize the use of clean wind energy and increase efficiency, says Kristin Larson, manager of wind energy forecasting at 3Tier, a renewable energy forecasting company.
“Wind forecasting can help the people who run the electricity grid choose between different resources. The more accurate the forecast is, the more efficiently the grid can be run and the more wind can be integrated into it," she says.
Like weather forecasting, wind forecasting uses ground measurements and computer models. “It’s a matter of predicting each volume of air and how each piece of energy, heat and moisture moves in that air. We have to do that for the entire globe to get the amount of wind at one point on the globe 80 meters in the air,” (the height of a wind turbine), says Pennock.
Data of the actual wind speed and the amount of energy produced is then used to refine the models. Say that one turbine was predicted to have a wind velocity of 10 meters per second, creating 2 megawatts of power. The data might show that, actually, its wind speed was 9 meters per second and produced 1.8 megawatts of power. These real numbers could be fed back into the model to help it more accurately predict the wind power generated in similar weather conditions in the future.
Just as in real estate, it’s all about location, location, location.
Wind forecasting in the flatter Plains states, like Nebraska or Kansas, is fairly accurate, but forecasting in less uniform landscapes with fickle weather like the Rocky Mountains is much less so. A report published by the North American Electric Reliability Corporation said that for a system of 10,000 MW of wind power, the 12-hour wind forecast could be off by as little as 20% or as much as 100%. Pennock says the numbers are getting smaller as wind forecasting improves, with most predictions off by no more than 20%.
Specific forecasting is more accurate. For instance, a client may only want forewarning of “ramp events” – instances in which the wind speed surges or plummets in a short time period, perhaps an hour.
Accuracy is also greater for larger regions. Combining the wind forecasts for all the wind farms in Texas creates a fairly accurate prediction for the region, because estimates that are too high for some farms would be balanced by low estimates for others.