基于近邻传播聚类算法的集中监控信号智能分析方法  

Intelligent Analysis Method of Power Grid Monitoring Signals Based on Affinity Propagation Clustering

作　　者：裴震 高雪冰 PEI Zhen;GAO Xuebing(State Grid Fujian Electric Power Co.,Ltd.,Zhangzhou Power Supply Company,Zhangzhou 363000,China)

机构地区：[1]国网福建省电力有限公司漳州供电公司,福建漳州363000

出　　处：《电工技术》2025年第3期173-176,共4页Electric Engineering

基　　金：国网福建省电力有限公司科技项目“基于数据机理混合驱动的电网监控信息信号智能分析技术研究”(编号B313-50230003)。

摘　　要：随着电网规模不断扩大,为提升电网运行效率,供电公司将变电站的电网监控信号进行集中管理。虽然集中处理后电网的信号数据分析效率将显著提高,但同时也存在监控站点无效信号过多、信号分类不明析等问题。为此建立了电网监控信息信号智能分析系统,基于近邻传播聚类对监控信号进行智能分析。算例表明,该方法能准确对电网监控信号的级别和类型进行分类,并统计相应的数量,提高了电网监控的工作效率。With the continuous expansion of the power grid scale,in order to improve the efficiency of power grid operation,power supply companies centralize the monitoring signals of substations for centralized management.After centralized processing,the efficiency of signal data analysis for the power grid will be significantly improved.However,there are also challenges such as too many invalid signals at monitoring sites and unclear signal classification.In order to effectively address these challenges,this article establishes an intelligent analysis system for power grid monitoring information signals,and conducts intelligent analysis on monitoring signals based on nearest neighbor propagation clustering.Case studies show that the method proposed in this article can accurately classify the levels and types of power grid monitoring signals,and count the corresponding quantities,thereby improving the efficiency of power grid monitoring work.

关 键 词：配电网 监控信号 近邻传播聚类 独热编码 

分 类 号：TM73[电气工程—电力系统及自动化]