基于样本熵聚类分解-倒置Transformer模型的光伏功率预测OA

[Objectives]Accurate prediction of photovoltaic power is of great significance for the safe,stable,and economic operation of the power grid.To improve the accuracy of photovoltaic power prediction,this study proposes a photovoltaic power prediction model based on sample entropy clustering decomposition-inverted Transformer.[Methods]Firsty,a sample entropy clustering decomposition method is proposed to optimize the sample entropy reconstruction process by introducing a hierarchical clustering algorithm and the silhouette coefficient to construct an adaptive assessment system.Secondly,the optimized variational mode decomposition algorithm is used to perform targeted secondary decomposition of the clustered components with noise interference.Then,the decomposed components are predicted separately using the inverted Transformer model to deeply mine the correlations among variables in long time series.Finally,the prediction results of each component are superimposed to obtain the final prediction results.[Results]Case analysis shows that the proposed method improves coefficient of determination by 0.7%,3.1%and 3.2%on sunny,cloudy,and rainy days,respectively,compared with the benchmark iTransformer model.The coefficient of determination improves by 4.2%,4.7%and 4.6%for different prediction sequence lengths,which fully verifies the superiority and feasibility of the proposed model.[Conclusions]The proposed method effectively solves the problem of low prediction accuracy caused by the weak characterization of highly fluctuating and intermittent photovoltaic power data in traditional models,as well as the insufficient capability for multivariate long-term time series modeling,and significantly improves the prediction accuracy.