摘要: |
发生单相接地故障时,在零模电压与其差分构成的平面上相邻点之间的欧氏距离大于发生虚幻接地情况时的欧氏距离,借此原理实现虚幻接地的识别。谐振接地系统线路发生接地故障时,健全线路的暂态零模电流波形之间有很强的相似性,而故障线路与健全线路的暂态零模电流波形之间差异显著。取故障后1/4工频周期作为分析时窗,采用分形盒维数和相对能量熵2个测度来定量表征故障线路和健全线路零模电流波形形态差别进行选线,并采用K-means聚类算法实现对线路的故障辨识。仿真分析表明,所提选线方法能够可靠地选出故障线路。将大量的历史故障数据作为样本映射到故障线路辨识平面,可提高其故障选线可靠性。 |
关键词: 谐振接地系统 虚幻接地 故障选线 分形盒维数 相对熵 K-means聚类分析 |
DOI: |
分类号: |
基金项目:云南省重点新产品开发科技计划项目(2011BA004) |
|
Unreal grounding identification and comprehensive line selection for resonant grounding system |
SHU Hongchun
|
Faculty of Electric Power Engineering,Kunming University of Science and Technology,Kunming 650051,China
|
Abstract: |
The Euclidean distance between two adjacent points on the plane formed by zero-mode voltage and its differential during a single-phase grounding is larger than that during an unreal grounding,which is adopted to identify the unreal grounding. When a grounding fault occurs in a resonant grounding system,the transient zero-mode current of a healthy line is very similar to that of another healthy line,but very different from that of the faulty line. The first quarter of post-fault power-frequency cycle is taken as the analytical time-window,the fractal box dimension and relative energy entropy are used to represent the difference of transient zero-mode current between healthy and faulty lines for selecting the faulty line,and the K-means clustering algorithm is adopted to identify the fault. Simulative analysis shows that,the proposed method can effectively select the faulty line. The reliability of faulty line selection is improved if massive historical fault data are projected onto the faulty line identification plan as samples. |
Key words: resonant grounding system unreal grounding faulty line selection fractal box dimension relative entropy K-means clustering analysis |