基于改进BILSTM/BIGRU的多特征短期负荷预测  

Multi-Feature Short-term Load Forecasting Based on Improved BILSTM/BIGRU

作　　者：王昊[1] 王树东[1] 唐伟强[1] WANG Hao;WANG Shudong;TANG Weiqiang(Lanzhou University of Technology,Lanzhou 730050)

机构地区：[1]兰州理工大学,兰州730050

出　　处：《计算机与数字工程》2025年第3期755-759,864,共6页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:62063018)资助。

摘　　要：针对传统神经网络在多输入特征下预测时间较长且精度欠佳的问题,论文提出了一种基于深度双向策略改进的长短期记忆神经网络与门控循环单元神经网络相结合的短期负荷预测模型。该模型采用自适应噪声完整集成经验模态算法将负荷数据进行分解,降低负荷数据复杂度;利用互信息主成分分析法提取原始多维输入变量,降低主成分因子;然后通过改进鲸鱼优化算法对构建模型进行寻参优化。以中国某地区的负荷数据作为算例,将论文所构建模型与其它模型进行了对比分析,预测结果表明,论文所构建的模型能够缩短预测的时间,提高负荷预测的精度。Pointing at the issue that conventional neural network has long expectation time and destitute precision beneath multi-input highlights,this paper proposes a short-term load forecast model based on a deep bidirectional strategy made strides long short term memory(LSTM)neural network combined with a gated recurrent unit(GRU)neural network.The model embraces the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)algorithm to break down the load information to diminish the complexity of the load information.This paper extricates the initial multi-dimensional input variables utilizing the mutual information principal component analysis(MIPCA)strategy and diminishes the central component calculate.Then the constructed model is optimized by improved whale optimization algorithm(IWOA).Taking the stack information in a certain locale of China as an illustration,the demonstrate developed in this paper is compared with other models.The forecast comes about appear that the show developed in this paper can abbreviate the figure.

关 键 词：负荷预测 深度双向策略 改进鲸鱼优化算法 长短期记忆神经网络 门控循坏单元神经网络 

分 类 号：TM715[电气工程—电力系统及自动化]