基于VMD-BIGRU与误差修正的风电功率预测  

Wind Power Prediction Based on VMD-BIGRU and Error Correction

作　　者：付波[1] 聂兴宇 赵熙临[1] 王诗雯 FU Bo;NIE Xingyu;ZHAO Xilao;WANG Shiwen(School of Electrical and Electronic Engineering,Hubei Univ.of Tech.,Wuhan 430068,China;Tangshan Power Supply Company,State Grid Jibei Electric Power Co.,Ltd.,Tangshan 063000,China)

机构地区：[1]湖北工业大学电气与电子工程学院,湖北武汉430068 [2]国网冀北电力有限公司唐山供电公司,河北唐山063000

出　　处：《湖北工业大学学报》2025年第2期17-21,28,共6页Journal of Hubei University of Technology

基　　金：湖北省重点研发计划项目(2021BAA193)。

摘　　要：针对风能的波动性和间歇性问题,提出了一种基于VMD-BIGRU与误差修正的风电功率预测模型。首先,使用变分模态分解(VMD)与样本熵(SE)结合方法分解风电时间序列;然后,使用双向门控循环单元(BIGRU)模型来预测风电的每个子序列并叠加得到初步预测结果;最后,构建基于时间卷积网络(TCN)的误差修正模型,对初步预测结果进一步优化。利用风电场的实际风电数据进行分析,结果表明所提方法具有更高的预测性能。Aiming at the volatility and intermittency of wind energy,a wind power prediction model based on VMD BIGRU and error correction is proposed.First,the time series of wind power are decomposed by using the method of variational mode decomposition(VMD)and sample entropy(SE);Then,BIGRU model is used to predict each subsequence of wind power,and the preliminary prediction results are obtained by superposition;Finally,an error correction model based on time convolution network(TCN)is constructed to further optimize the prediction results of the model.The analysis of the actual wind power data of wind farms shows that the proposed method has higher prediction performance.

关 键 词：变分模态分解 双向门控循环单元 时间卷积网络 风电功率预测 误差修正 

分 类 号：TM614[电气工程—电力系统及自动化]