现代电力2025，Vol.42Issue(2)：201-208,8.DOI:10.19725/j.cnki.1007-2322.2023.0025
基于图卷积神经网络-双向门控循环单元及注意力机制的风电功率短期预测模型
A Short-term Wind Power Prediction Model Based on Graph Convolutional Neural Network-bidirectional Gated Recurrent Unit and Attention Mechanism
摘要
Abstract
Accurate prediction of wind power is of great signi-ficance to the stable operation of power systems.To address the issue that traditional combinatorial models face the challenges in fully exploring the potential dependencies among variables and exhibit low accuracy in wind power prediction when deal-ing with high-dimension and largescale data,a model,incorpor-ating a graph convolutional neural network-bidirectional gated cyclic unit and an attention mechanism,is proposed for short-term wind power prediction.The model takes numerical weath-er prediction(NWP)data and wind power historical data as in-put.It initially employs Pearson correlation analysis to filter features,followed by the utilization of graph convolutional neural network(GCN)with the help of residual connection to mine spatial feature relationships via network and graph.Sub-sequently,a bidirectional gated recurrent unit(BiGRU)is em-ployed to mine the time-series features of historical data.Fi-nally,an attention mechanism(AM)is introduced to assign weights to achieve short-term wind power prediction.The ex-perimental results demonstrate that this method exhibits superi-or prediction accuracy in both single-step and multi-step predic-tion compared to other methods.关键词
风电功率预测/混合深度神经网络/图卷积神经网络/双向门控循环单元/注意力机制Key words
wind power prediction/hybrid deep neural net-works/graph convolutional neural networks/bidirectional gated recurrent units/attention mechanism分类
信息技术与安全科学引用本文复制引用
张光昊,张新燕,王朋凯..基于图卷积神经网络-双向门控循环单元及注意力机制的风电功率短期预测模型[J].现代电力,2025,42(2):201-208,8.基金项目
国家自然科学基金项目(51667018) （51667018）
新疆维吾尔自治区自然科学基金项目(2021D01C044).Project Supported by National Natural Science Foundation of China(51667018) （2021D01C044）
Natural Science Foundation of Xinjiang Uygur Autonomous Region(2021D01C044). （2021D01C044）
