可再生能源2025，Vol.43Issue(7)：902-910,9.
基于CEEMDAN-PCA-BiLSTM-LSTNet的短期风电功率组合预测模型
A short-term wind power combination forecasting model based on CEEMDAN-PCA-BiLSTM-LSTNet
摘要
Abstract
Improving the accuracy of wind power forecasting is of great significance for the safe and stable operation of the power grid.To address this issue,this paper proposes a short-term wind power combination forecasting model based on CEEMDAN-PCA-BiLSTM-LSTNet.Firstly,the original wind power data and numerical weather prediction data(NWP)are decomposed and dimensionally reduced using the adaptive noise ensemble empirical mode decomposition(CEEMDAN)and principal component analysis(PCA)methods,respectively,for data preprocessing.Secondly,a bidirectional long short-term memory network(BiLSTM)is trained to predict each component obtained from the decomposition and then superimpose to obtain preliminary prediction results.Simultaneously,the comprehensive meteorological factors extracted after dimensionality reduction are used to train the long short-term time series neural network(LSTNet)to obtain the numerical weather prediction results.Finally,an error weight matrix is constructed based on information entropy theory,and the numerical weather prediction results are used to perform combined weighted correction on the preliminary prediction results.Experimental results demonstrate that the combined forecasting model using different types of raw data and mechanism models can effectively capture the spatiotemporal characteristics of wind power,exhibiting higher prediction accuracy compared to existing methods,thus validating the effectiveness of the model.关键词
数值天气预报/风电功率预测/经验模态分解/主成分分析/双向长短时记忆神经网络/长短期时间序列神经网络Key words
numerical weather prediction/wind power prediction/CEEMDAN/principal component analysis/bi-directional long short-term memory neural networks/long-and short-term time series network分类
能源科技引用本文复制引用
沈海波,王凌梓,邓力源,程贤良,吴慧军..基于CEEMDAN-PCA-BiLSTM-LSTNet的短期风电功率组合预测模型[J].可再生能源,2025,43(7):902-910,9.基金项目
国家自然科学基金项目(41875118) （41875118）
国家自然科学基金青年基金项目(41805047). （41805047）
