可再生能源2026，Vol.44Issue(6)：780-789,10.
基于HHO-SVMD-ASCSSA-LSTM的风电短期出力预测
Short term wind power prediction based on HHO-SVMD-ASCSSA-LSTM
摘要
Abstract
To improve the accuracy of short-term wind power output prediction,this paper proposes a combined prediction model that uses improved successive variational mode decomposition(SVMD)to extract wind power features and then employs an improved long short-term memory(LSTM)neural network.The model first selects strongly correlated parameters such as wind speed and wind direction as input samples,and conducts a correlation analysis on historical wind power output and 11 types of meteorological parameters.Furthermore,the Harris Hawks Optimization(HHO)algorithm is used to optimize the balance parameters of SVMD,decomposing the historical wind power output data into 6 subsequences with weak nonlinearity.Subsequently,the improved sparrow search algorithm(ASCSSA)is introduced to optimize parameters of the LSTM model,including the number of hidden units,training epochs,and initial learning rate,and regression predictions are performed on each SVMD subsequence and sample data respectively.Finally,the final prediction result of wind power output is obtained by superimposing the predicted values of each subsequence.The research shows that the improved sparrow search algorithm(ASCSSA)integrating Cauchy mutation and spiral strategy has high solution accuracy and fast convergence speed in solving multi-dimensional nonlinear optimization problems.The output prediction results based on three sets of actual cases indicate that the proposed combined model can significantly improve prediction accuracy without sacrificing the applicability of the original LSTM model as much as possible.关键词
连续变分模态分解/改进麻雀算法/长短期记忆神经网络/风电短期出力预测Key words
successive variational mode decomposition/improved sparrow search algorithm/long short-term memory neural network/short-term wind power output prediction分类
能源科技引用本文复制引用
梁兴,胡卓,彭钰璋,周明捷,邓飞,李培生..基于HHO-SVMD-ASCSSA-LSTM的风电短期出力预测[J].可再生能源,2026,44(6):780-789,10.基金项目
江西省教育厅科技项目(GJJ211941) （GJJ211941）
国家自然科学基金项目(51969017). （51969017）
