Computer Engineering and Applications ›› 2021, Vol. 57 ›› Issue (22): 288-294.DOI: 10.3778/j.issn.1002-8331.2007-0087

• Engineering and Applications • Previous Articles     Next Articles

Wind Farm Ultra-Short-Term Wind Speed Prediction Based on EEMDSE-ILSTM

YI Lingzhi, WANG Shitong, YI Fang, DENG Dong, YI Zhimin, JIANG Peng   

  1. 1.College of Automation and Electronics Information, Xiangtan University, Hunan Province Engineering Research Center for Multi-energy Collaborative Control Technology, Xiangtan, Hunan 411105, China
    2.Hunan Province Cooperative Innovation Center for Wind Power Equipment and Energy Conversion, Xiangtan, Hunan 411105, China
    3.Xiangdian Wind Energy Co., Ltd., Xiangtan, Hunan 411105, China
  • Online:2021-11-15 Published:2021-11-16



  1. 1.湘潭大学 自动化与电子信息学院,湖南省多能源协同控制技术工程研究中心,湖南 湘潭 411105
    2.湖南省风电装备与能源变换协同创新中心,湖南 湘潭 411105
    3.湘电风能有限公司,湖南 湘潭 411105


The depletion of nonrenewable resources pushes forward the development of new energy. Wind power, as the main form of wind energy utilization, has been widely promoted. However, the non-linear, non-stationary and time-series characteristics of wind speed have adverse effects on the wind turbine generator itself and power system. Therefore, accurate wind speed prediction has become the key issue to be solved. In this paper, an EEMDSE-ILSTM wind speed prediction model is proposed based on the combination prediction method. In this model, Ensemble Empirical Mode Decomposition(EEMD) is used to decompose the wind speed data into several component data sets, and each component is filtered by sample entropy to simplify the data. Improved whale algorithm is combined with  Long Short-Term Memory(LSTM) to generate appropriate model prediction parameters without supervision. The model predicts each decomposing data in turn and adds up the results to get the final prediction value. The simulation results show that the model has better prediction accuracy and generalization performance compared with other methods.

Key words: Ensemble Empirical Mode Decomposition(EEMD), sample entropy, wind speed prediction, Improved Whale Optimization Algorithm(IWOA), Long Short-Term Memory(LSTM)


不可再生资源的枯竭推动着新能源的发展,风电作为目前风能利用的主要形式得到了大面积推广。但风速非线性、非平稳性、时序性的特点对风机本身和电力系统都会产生不利的影响,因此精准的风速预测已经成为亟待解决的关键课题。基于组合预测方法,提出了一种EEMDSE—ILSTM风速预测模型。该模型利用集合经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)将风速数据分解为若干个分量数据集,并通过样本熵对各分量进行筛选以简化数据。将改进的鲸鱼算法与长短期记忆网络(Long Short-Term Memory,LSTM)结合,无监督生成合适的模型预测参数。在预测时依次对每个分量数据预测并将结果累加获得最终预测值。仿真结果表明,该模型与其他方法比较,显示出较好的预测精度和泛化性能。

关键词: 集合经验模态分解(EEMD), 样本熵, 风速预测, 改进的鲸鱼优化算法(IWOA), 长短期记忆网络(LSTM)