基于改进双重压缩和激励与多头特征注意力机制的电-热负荷协同预测  

Collaborative forecasting of electricity-thermal load based on improved dual squeeze and excitation and multi-head feature attention mechanism

作　　者：余强 韩静娴 杨子梁 宋济东 杨德昌[1] 齐海杰 于芃 YU Qiang;HAN Jingxian;YANG Ziliang;SONG Jidong;YANG Dechang;QI Haijie;YU Peng(College of Information and Electrical Engineering,China Agricultural University,Beijing 100083,China;State Grid Smart Grid Research Institute Co.,Ltd.,Beijing 102200,China;State Grid Shandong Electric Power Research Institute,Jinan 250003,China)

机构地区：[1]中国农业大学信息与电气工程学院,北京100083 [2]国网智能电网研究院有限公司,北京102200 [3]国网山东省电力公司电力科学研究院,山东济南250003

出　　处：《电力自动化设备》2025年第3期201-208,共8页Electric Power Automation Equipment

基　　金：国家电网有限公司科技项目(SGSDDK00PDJS2250114);国家自然科学基金资助项目(52377127)。

摘　　要：综合能源系统中负荷多样且存在耦合,为提升负荷预测精度,提出一种基于改进双重注意力机制的分组卷积神经网络-门控循环单元短期电-热负荷协同预测模型。通过改进的压缩和激励注意力为各输入通道加权,再对其进行分组卷积;利用多头特征注意力对卷积结果进行赋权,并利用输入门控循环单元模型对负荷进行预测。算例仿真结果表明,所提模型的平均绝对百分比误差均低于3%。Loads are diverse and coupled in the integrated energy system,in order to improve the accuracy of load forecasting,a short-term electricity-thermal load collaborative forecasting model based on an im-proved dual attention mechanism and group convolutional neural network-gated recurrent unit is proposed.Each input channel is weighted by the improved squeeze and excitation attention mechanism,and grouped for the convolution.The convolution results are weighted by the multi-head feature attention mechanism and the load is forecasted by the input gated recurrent unit model.The example simulative results show that the mean absolute percentage errors of the proposed model are less than 3%.

关 键 词：综合能源系统 负荷预测 分组卷积神经网络 门控循环单元 改进的压缩和激励注意力机制 多头特征注意力机制 

分 类 号：TK019[动力工程及工程热物理]