基于数据稀疏特征的架空电力线路故障可视化运检技术  

Visual Operation Inspection Technology of Overhead Power Line Fault Based on Data Sparse Characteristics

作　　者：刘晓晶 陈显达[1] 曹帅 陈楠 王婷婷 LIU Xiaojing;CHEN Xianda;CAO Shuai;CHEN Nan;WANG Tingting(Ji’nan power supply company,State Grid Corporation of China,Ji’nan 250000,China;Shandong Electric Power Transmission&Transformation Engineering Company,State Grid Corporation of China,Ji’nan 250000,China)

机构地区：[1]国网济南供电公司,山东济南250000 [2]山东送变电工程有限公司,山东济南250000

出　　处：《测试技术学报》2023年第2期135-139,164,共6页Journal of Test and Measurement Technology

摘　　要：针对架空电力线路易出现故障、存在的安全隐患大等问题,提出一种基于数据稀疏特征的架空电力线路故障可视化运检技术。利用稀疏数据矩阵求解稀疏系数,重建矩阵并恢复数据,明确恢复前后信息差距程度,计算稀疏数据的中心距离特征,凭借贡献度决定是否提取稀疏数据的某项特征,分别在已知和未知线路行波速度两种情况下,分析发生故障处行波距离电力线路起终点间的运行时间,利用电力线模和电力零模分量,计算单相接地和线路故障距离,计算求解电力线路故障位置以及线路两端电阻值,通过散点图呈现线路故障的具体信息。经仿真实验数据分析证明,所提技术可挖掘到深层信息,故障运检准确度高,误差小,可适用性强。Aiming at the problems that overhead power lines are prone to faults and have great potential safety hazards,a visual operation inspection technology for overhead power line faults based on sparse data characteristics is proposed.Use the sparse data matrix to solve the sparse coefficient,reconstruct the matrix and recover the data,clarify the information gap before and after recovery,calculate the center distance feature of the sparse data,determine whether to extract a feature of the sparse data by virtue of the contribution degree,analyze the running time between the traveling wave at the fault and the starting and ending points of the power line under the conditions of known and unknown line speed,and use the power line mode and power zero mode components,Calculate the distance between single-phase grounding and line fault,calculate and solve the fault location of power line and the resistance value at both ends of the line,and present the specific information of line fault through scatter diagram.The analysis of simulation experiment data shows that the proposed technology can mine deep information,with high accuracy,small error and strong applicability.

关 键 词：数据稀疏特征 电力线路运检 数据可视化 故障检测 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]