面向电力需求分析的多源数据处理算法设计  

Design of multi⁃source data processing algorithm for power demand analysis

作　　者：凌荣光 许巍 严若婧 王妍 梁玉洁 LING Rongguang;XU Wei;YAN Ruojing;WANG Yan;LIANG Yujie(Ningbo Power Supply Company,State Grid Zhejiang Electric Power Co.,Ltd.,Ningbo 315010,China)

机构地区：[1]国网浙江省电力有限公司宁波供电公司,浙江宁波315010

出　　处：《电子设计工程》2025年第5期147-151,共5页Electronic Design Engineering

摘　　要：针对电力需求端复杂度提高所带来的电力负荷预测精度差的问题,文中基于电力系统数供平台的多源数据提出了一种面向电力需求分析的负荷预测算法,以此辅助提高电网的运行控制水平。该算法利用SVM算法提取输入数据特征,对于SVM算法参数较为复杂的缺点,通过引入GWO算法对SVM的初始化参数进行优化。利用LightGBM模型在搜索和分割领域的性能优势,对数据特征展开训练并输出负荷预测结果,同时基于Stacking集成学习框架,将所提算法进行融合,从而有效提升了算法的运算效率及预测精度。在实验结果中,所提算法的计算效率较原算法有显著提升,迭代次数减少了近50%,性能参数也在对比算法中处于领先,体现了算法性能的优越性。In response to the phenomenon of poor accuracy in power load forecasting caused by the increased complexity of the power demand side,this paper proposes a load forecasting algorithm for power demand analysis based on multi-source data from the power system data supply platform,to assist in improving the operational control level of the power grid.This algorithm utilizes the SVM algorithm to extract input data features.To address the complex parameters of the SVM algorithm,the GWO algorithm is introduced to optimize the initialization parameters of the SVM.Utilizing the performance advantages of the LightGBM model in search and segmentation,train data features and output load prediction results.At the same time,based on the Stacking integrated learning framework,the proposed algorithm is fused,effectively improving the computational efficiency and prediction accuracy of the algorithm.In the experimental results,the computational efficiency of the proposed algorithm is significantly improved compared to the original algorithm,with a reduction of nearly 50%in iteration times.The performance parameters are leading in the comparison of the algorithms,reflecting the superiority of the algorithm’s performance.

关 键 词：多源数据分析 电力负荷预测 支持向量机 灰狼优化算法 LightGBM 集成学习框架 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]