基于VMD-GRU-EC的工业用户短期负荷预测方法  

Short-term Load Forecasting Method for Industrial Users Based on VMD-GRU-EC

作　　者：蒋建香 杨苹[1] 官裕达 杨康 刘璐瑶 Jiang Jianxiang;Yang Ping;Guan Yuda;Yang Kang;Liu Luyao(Key Laboratory of Clean Energy Technology of Guangdong Province,South China University of Technology,Guangzhou Guangdong 510641,China)

机构地区：[1]华南理工大学广东省绿色能源技术重点实验室,广东广州510641

出　　处：《电气自动化》2023年第5期34-37,41,共5页Electrical Automation

基　　金：广东省重点领域研发计划项目(2021B0101230003)。

摘　　要：工业用户作为城市用电主体,具有负荷组成结构复杂,易受生产计划等因素影响产生冲击性负荷的特点。而传统的负荷预测方法难以处理工业负荷的非线性,且未充分利用负荷残差序列信息,导致负荷预测效果不佳。为此,在采用传统的变分模态分解-门控循环网络预测方法得到初始预测值的基础上,进一步考虑增加基于门控循环网络的误差修正方法对初始预测值进行修正,从而更好地适应工业负荷的多变和突变性。算例分析表明,所提出的短期负荷预测方法可有效提升的工业用户负荷预测准确率,为工业用户在生产运行和参与电力市场交易等提供强有力的决策支撑。As the main body of urban electricity consumption,industrial users have the characteristics of complex load composition structure and are easily affected by production plans and other factors to generate impulsive loads.However,traditional load forecasting methods are difficult to handle the nonlinearity of industrial loads and do not fully utilize the residual sequence information of loads,resulting in poor load forecasting performance.Therefore,on the basis of using the traditional variational modal decomposition gated recurrent network prediction method to obtain the initial prediction value,further consideration was given to adding an error correction method based on gated recurrent network to correct the initial prediction value,in order to better adapt to the variability and mutation of industrial loads.The example analysis shows that the proposed short-term load forecasting method can effectively improve the accuracy of industrial user load forecasting,providing strong decision-making support for industrial users in production operation and participating in electricity market transactions.

关 键 词：冲击性负荷 变分模态分解 门限循环网络 误差修正 负荷预测 

分 类 号：TM743[电气工程—电力系统及自动化]