Wind power plant short-term wind speed prediction method and system and electronic equipment

Abstract

The invention discloses a wind power plant short-term wind speed prediction method and system and electronic equipment. The wind speed sequence is decomposed into a series of modal components with different center frequencies, the volatility and randomness of a wind speed sequence can be reduced, a first prediction wind speed is obtained by using modal components, a feature set is constructed by using the wind speed sequence, the feature set can assist a prediction model in analyzing the wind speed sequence, the model is prevented from falling into local optimum according to the features, andthe generalization is improved; the superposed first prediction wind speed is added into the training set, the related information of the first prediction wind speed is reserved, and the feature information of the wind speed sequence is fully utilized. According to the invention, the final prediction wind speed can reserve the prediction precision among different models, so that the model with accurate prediction is allocated with greater weight, and the prediction accuracy of the fused model is improved.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to a short-term wind speed prediction method, system and electronic equipment of a wind farm. Background technique [0002] In the context of global energy transformation gradually becoming a trend, wind power as an important part of power market reform and the rapid development of information technology, the research on short-term wind speed prediction has more important significance. Accurate wind speed prediction is not only helpful for assisting wind farms to control power quality, optimize dispatching operation management, and give full play to the potential of the system to accept wind power, but also for regional power generation planning and overall coordination, and cooperate with energy storage equipment such as thermal storage electric boilers to track wind power generation , It is of great strategic significance to reduce the generation of large-scale "abandoned wind", im...

AI Technical Summary

Problems solved by technology

Different models use different data feature information, resulting in different prediction accuracy between wind speed prediction models, but the prediction accuracy is often limited

The commonly used average weighting method has the problem that the model with superior predictive performance is covered up, and the feature information cannot be effectively used, resulting in the inability to fully utilize the data information and affecting its predictive ability.

Images