为了提高风电功率短期预测精度，本文提出了一种基于ICEEMDAN-SE-MSGJO-LSTM-EC模型的短期风电功率预测模型。首先，通过ICEEMDAN对原始风功率信号进行分解并通过样本熵计算熵值相近的分量相加重构。其次，建立MSGJO-LSTM预测模型，通过改进金豺优化算法（MSGJO）优化LSTM网络参数，对各模态分量进行预测。最后，通过对各模态分量预测结果进行误差修正（EC）并将所有模态预测结果相加得到最终预测结果。以新疆某风电场为例，采用本文所提预测模型进行仿真分析，试验结果表明本文基于ICEEMDAN-SE-MSGJO-LSTM-EC的预测模型预测精度更高。

In order to improve the accuracy of short\|term wind power prediction, this paper proposes a short\|term wind power prediction model based on the ICEEMDAN\|SE\|MSGJO\|LSTM\|EC model. Firstly, the original wind power signal is decomposed using intrinsic computing expressive empirical mode decomposition with adaptive noise (ICEEMDAN) and reconstructed by adding components with similar entropy values through sample entropy calculation; Secondly, establish an multi\|strategy golden jackal optimization (MSGJO)\|long short\|term memory network (LSTM) prediction model, optimize the LSTM network parameters through the improved MSGJO, and predict various modal components; Finally, error correction (EC) is applied to the prediction results of each modal component and all modal prediction results are added to obtain the final prediction result. Taking a wind farm in Xinjiang as an example, simulation analysis was conducted using the prediction model proposed in this paper. The experimental results showed that the prediction model based on ICEEMDAN\|SE\|MSGJO\|LSTM\|EC proposed in this paper has higher prediction accuracy.

云南省基础研究重点项目（202301AS070055）