A New Algorithm for Ultra-Short Term Wind Prediction Based on Dynamic Relative Turbine Locations

Abstract

This paper is focusing on ultra-short term wind prediction, using the wind speed of so called “leading turbine” as a reference to calculate the distribution wind speed in a wind farm, where the leading turbine is the first one to contact the wind determined by relative location of every turbine in wind farm. Then wake effect, time-lag effect and dynamic relative locations algorithms are introduced and integrated into the calculation to get every input wind speed of remained wind turbines. The illustrative cases take a wind farm in China as study subject to forecast wind speed of every wind turbine based on actual captured data of “leading turbine”. Then the comparison demonstrates difference of selected turbine's time-varying wind speed and power output in wind farm estimated by wake effect factor and time-lag effect factor. The conclusion reflects feature that wake effect will reduce power output and time-lag effect will smooth power output. At last we transform the integrated wind speed prediction results into the total power output data, and compare it with actual captured output power data of wind farm, the result shows that the accurate and rapid calculation method based on dynamic relative location of turbines in wind farm is authentic and effective.