变电站远程故障信息的图形化挖掘与仿真  

Graphical Mining and Simulation of Remote Fault Information in Transformer Substation

作　　者：徐勇 张永杰 姚毅杰 赵晨阳 XU Yong;ZHANG Yong-jie;YAO Yi-jie;ZHAO Chen-yang(Yangzhou Power Supply Branch,State Grid Jiangsu Electric Power Co.,Ltd,Yangzhou Jiangsu 225000,China;Ultra High Voltage Branch,State Grid Jiangsu Electric Power Co.,Ltd,Jiangsu Nanjing 211000,China)

机构地区：[1]国网江苏省电力有限公司扬州供电分公司,江苏扬州225000 [2]国网江苏省电力有限公司超高压分公司,江苏南京211100

出　　处：《计算机仿真》2025年第2期509-513,共5页Computer Simulation

基　　金：国家电网有限公司科技项目(5501-202332120A-0-1-01)。

摘　　要：变电站是复杂而精密的系统,易受外界干扰。此期间产生的故障信息包含设备状态、时间和故障类型等多维且复杂的数据,难以快速确定故障情况。为此,提出变电站远程故障信息图形化挖掘方法。采集变电站设备状态信息,根据采集信息输出变电站图形化结果;利用人工卷积神经网络模型分类变电站远程故障信息数据,并在故障位置标注设备状态、故障类型以及故障时间等信息;使用数据挖掘方法对变电站远程图形化故障信息实施深度挖掘,通过属性约简找出故障之间关联关系,完成故障信息的图形化挖掘,获取变电站相关故障模式。仿真结果表明,通过图形化展示,可以同时展示多个维度的信息,并帮助用户理解不同维度之间的关联性和影响;利用所提方法开展故障信息图形化挖掘时,挖掘效果较好、精度较高。Substation is a complex and sophisticated system,which is susceptible to external interference.The fault information generated during this period contains multidimensional and complex data such as equipment status,time and fault type,which makes it difficult to quickly determine the fault situation.For this reason,a graphical mining method for remote fault information of substations is proposed.We collected substation equipment state information,and output substation graphical results according to the collected information;used an artificial convolutional neural network model to classify substation remote fault information data,and labeled the fault location with equipment state,fault type,and fault time;used data mining methods to implement deep mining of substation remote graphical fault information,and found out the correlation relationship between faults through attribute simplification to complete the graphical mining of fault information,obtained the substation-related fault mode.Simulation results show that the graphical display can simultaneously display information in multiple dimensions and help users understand the correlation and influence between different dimensions;when the proposed method is used to carry out graphical mining of fault information,the mining effect is better and the accuracy is higher.

关 键 词：变电站 远程故障信息 图形化 挖掘方法故障标注 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]