Front detection using wind park data and NWP

The precise forecasting of ramps in production of windparks is a problem that is not satifsfactorily solved. This is of particular interest because ramps contribute to a major part to the forecasting error in power production, in particular for offshore wind parks.

Since ramps are often due to fronts passing a location, we developed a method for the correction of front speed and -direction using a combination of wind park meteorological measurements and numerical weather prediction (NWP). On one hand we use conventional methods like the Canny algorithm for NWP data, on the other hand, we use data from a collection of wind parks to determine a passing front. By the front speed, and the relative location of wind parks, the front speed is computed and a correction can  be applied to downstream wind parks.

The results can be  validated and a corresponding error measure can be computed on the basis of measured and numerical data. Our method shall be implemented into a proprietary forecast system with the goal of an automatized detection and correction mechanism.