基于VMD和IDBO-ELM的短期光伏发电功率预测模型  

Short-Term Photovoltaic Power Prediction Model Based on VMD and IDBO-ELM

作　　者：冯广锦 王玉林 卢文博 Feng Guangjin;Wang Yulin;Lu Wenbo(School of Electrical and Control Engineering,Liaoning University of Engineering and Technology,Huludao Liaoning 125105,China)

机构地区：[1]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105

出　　处：《现代工业经济和信息化》2025年第1期257-259,共3页Modern Industrial Economy and Informationization

基　　金：国家大学生创新训练计划(202410147011)。

摘　　要：提出了一种基于变分模态分解(VMD)和改进蜣螂优化算法优化极限学习机(IDBO-ELM)的短期光伏发电功率预测模型。该模型首先利用VMD将原始光伏发电数据分解,得到不同时间尺度的信号成分并提取特征。然后通过极限学习机对各子序列进行建模预测,同时利用IDBO优化ELM的输入层权值和隐含层偏置。最后通过实验验证模型的实用性和有效性。A short-term PV power prediction model based on variational modal decomposition(VMD)and improved dung beetle optimization algorithm optimized extreme learning machine(IDBO-ELM)is proposed.The model first utilizes the decomposition of raw PV power generation data to obtain signal components at different time scales and extract features.Then the subseries are modeled and predicted by the limit learning machine,while the optimized input layer weights and implicit layer bias are utilized.Finally,the practical validity of the model is verified through experiments.

关 键 词：短期光伏负荷预测 变分模态分解 改进蜣螂优化算法 极限学习机 

分 类 号：TM615[电气工程—电力系统及自动化]