基于多源数据融合的变电站设备缺陷识别技术研究  

Research on substation equipment defect identification technology based on multi-source data fusion

作　　者：曾宏宇 杨冰 许悦 汪杨凯 杨威 Zeng Hongyu;Yang Bing;Xu Yue;Wang Yangkai;Yang Wei(Maintenance Company of State Grid Hubei Electric Power Co.,Ltd.,Hubei Wuhan,430051,China)

机构地区：[1]国网湖北省电力有限公司检修公司,湖北武汉430051

出　　处：《机械设计与制造工程》2025年第2期75-80,共6页Machine Design and Manufacturing Engineering

摘　　要：为了确保变电站设备缺陷识别效果,降低识别误差和能耗,提出基于多源数据融合的变电站设备缺陷识别方法。通过不同类型传感器采集各类信息,结合支持向量机,搜索最优最大间隔超平面特点,将数据融合结果作为支持向量机输入,借助支持向量机分类方式,完成多种类别的变电站设备缺陷识别。实验结果表明,所提方法的多源数据融合结果更接近实际结果,对数据融合结果进行缺陷识别未出现错误识别,且不同缺陷的识别误差均未超过2%,能够有效确保缺陷识别效果,降低识别能耗。In order to ensure the effect of substation equipment defect identification and reduce identification error and energy consumption,a substation equipment defect identification method based on multi-source data fusion is proposed.Through different types of sensors to collect information,it combines with support vector machine,searches the characteristics of the optimal maximum interval hyperplane,takes the data fusion results as the input of support vector machine,and completes the defect identification of various types of substation equipment with the help of support vector machine classification.The experimental results show that the multi-source data fusion results of the proposed method are closer to the actual results,the data fusion results are obtained for defect identification,there is no error identification,and the identification errors of different defects are not more than 2%,which can effectively ensure the defect identification effect and reduce the identification energy consumption.

关 键 词：多源数据融合 变电站设备 缺陷识别 支持向量机 

分 类 号：TM764[电气工程—电力系统及自动化]