Short-term wind speed forecasting method based on local integrated study

Abstract

The invention discloses a short-term wind speed forecasting method based on local integrated study. The method comprises steps as follows: step one, K wind speed samples most similar to a measured wind speed sample point are found out with a K-nearest neighbor algorithm; step two, a corresponding wind speed predicted value of a local sample of each wind speed sample point is obtained; step three, a plurality of wind speed predicted values obtained in the step two are fused by adopting certain fusion strategies to obtain a predicted value of a current wind speed point. The short-term wind speed forecasting method based on the local integrated study is proposed firstly, and the complexity and the diversity of wind speed sample change are considered; mathematical theory knowledge of the local integrated study algorithm is provided; difference of different wind speed samples is effectively obtained, different wind speed forecasting models are established according to different wind speed sample points, and compared with an overall wind speed forecasting model, a local wind speed forecasting model and a whole integration wind speed forecasting model, the forecasting error rate is reduced by more than 10%.

Description

technical field [0001] The invention belongs to the technical field of wind power generation, in particular to a short-term wind speed forecast method based on local learning. Background technique [0002] Due to the continuous increase of energy consumption and the increasingly severe environmental pollution, many countries have carried out large-scale exploration of new energy sources. Among them, wind energy, as an important renewable resource, has attracted more and more attention from all over the world. With the maturity of wind power generation technology, wind power supply and demand must be balanced when wind power is connected to the grid. However, wind power has strong fluctuations and uncertainties, which have a major impact on the dispatching method, stability and peak frequency regulation of the power grid. Accurate forecasting of wind power is helpful for power grid dispatching departments to timely grasp the output power of wind farms and adjust dispatching ...

AI Technical Summary

Problems solved by technology

However, local learning has a strong learning ability, which will cause over-fitting problems, making the generalization ability of the algorithm poor

Therefore, using the local learning algorithm to make predictions, the prediction results cannot meet the actual requirements.

Images