基于IFP-Informer模型的超短期风速预测

Ultra-short-term Wind Speed Prediction Based on IFP-Informer Model

针对传统模型预测精度较差、忽略风速混沌特性与模态分量内部隐含关联的问题,提出了结合模态分解、特征筛选、Informer的超短期风速预测模型。首先,采用改进的完全自适应噪声集合经验模态分解将风速原始序列分解为模态分量,同时对具有混沌特性的风速原始序列进行相空间重构,得到风速混沌序列;然后,通过相关性快速过滤算法对模态分量与混沌序列进行特征筛选,实现模态分量的优选;最后,将优选的模态分量输入到Informer模型,输出未来时刻风速。以某风场为研究对象,通过仿真计算验证了预测模型的准确性和有效性。

In response to the problem of poor prediction accuracy,ignoring the chaotic characteristics of wind speed and the implicit correlation within modal components of traditional models.This article proposes an ultra short term wind speed prediction model that combines modal decomposition,feature selection,and Informer.Firstly,an improved fully adaptive noise set empirical mode decomposition is used to decompose the original wind speed sequence into modal components.At the same time,the chaotic wind speed original sequence is reconstructed in phase space to obtain the wind chaotic speed sequence.Subsequently,a fast correlation based filtering algorithm is used to filter the features of modal components and chaotic sequences,achieving optimal selection of modal components.Finally,the selected modal components are input into the Informer model to output future wind speeds.The accuracy and effectiveness of the prediction model were verified through simulation calculations,by using the wind field in the Northeast region as the research object.