Wind Farms To Receive Future Wind Forecasts 57
An anonymous reader writes "If the US plans to develop wind farms across the country they need a better way to predict the wind direction and the duration. NCAR (National Center for Atmospheric Research) is looking to do just that. In December, NCAR signed an agreement with Xcel Energy to develop a wind prediction system for the company's wind energy farms in Colorado, Minnesota, and Texas. Experimental forecasts may start as early as May. At present, most wind forecasts rely heavily on statistical forecasting methods, since the numerical weather forecast products available from operational centers are produced with coarse-grid, larger-scale models. The RTFDDA system, however, is designed to provide a birds-eye view of local weather for small areas of special interest, like wind farms, through a multiple level downscaling algorithm." I hope that decentralized weather-data gathering stations (like many people have feeding data to The Weather Underground) would be useful for this purpose.
Why? (Score:5, Interesting)
So what will they do with the wind forecast?
It's not like they turn off the windmills when there is no wind. The ones that cannot be rotated gain no benefit from knowing the direction. And the ones that can be rotated rotate automatically based on the current wind, not in anticipation of a future wind.
Perhaps I should be turning off my solar panels at night, or on cloudy days. (hmm... Actually, there are diodes to do that to prevent the panels from consuming stored energy)
They could be their own prediction system (Score:4, Interesting)
If windmills were distributed widely enough, they could produce their own prediction data. i.e. windmill A, downwind from windmill B, should use windmill B's output as a predictor. Or, more broadly, windmill A would rely on a set of windmills within a given radius for prediction data. This would require a much larger distribution than will be available any time soon, but would probably be the best way to go in the long term. P2P wind power prediction. Me likey.
Re:Wind prediction is common in Germany (Score:3, Interesting)
The state of prediction in Europe isn't as good as you might think. Nor is it in corporate America. That's a big reason NCAR is getting involved. Several of us are working on how to make WRF or a similar weather model work to forecast at the appropriate heights where the turbine blades work. AND at the resolutions required.
One wind prediction company I'm aware of here claims to have resolutions as fine as 250m for the whole US. What I know of the model they're using, however, suggests that the model isn't stable at that high a resolution. Getting models to behave as we adapt them for new tasks isn't always easy...