基于波形伸缩变换和模糊K均值聚类的配电线路单相接地选线研究
本文关键词: 配电网 暂态零序电流 伸缩变换 相平面 模糊K均值聚类 出处:《福州大学》2014年硕士论文 论文类型:学位论文
【摘要】:我国配电网普遍采用小电流接地方式,即中性点不接地方式和中性点经消弧线圈接地方式。在小电流接地方式中,单相接地故障的概率很高,一般占总故障的80%以上。当小电流接地系统发生单相接地故障时,由于接地电流较小,且故障后三相电压仍保持对称关系,不会影响负荷的持续供电,在一定程度上可以提高电力系统的供电可靠性。由于小电流接地系统发生单相接地故障的情况复杂多变、故障电流较微弱,且常伴有间歇性电弧,以及受线路结构参数、互感器非线性特性、电磁干扰等影响,故障线路零序电流的频谱特性、能量分布和衰减特性有着很大的差异,一直缺乏可靠的单相接地故障选线方法,其接地选线问题仍未得到很好解决。本文总结了配电线路单相接地故障选线方法,分析了稳态量选线方法的不足之处,说明利用暂态量进行单相接地故障选线的优势。通过对谐振接地系统暂、稳态零序电流的分析,指出非故障线路间暂态零序电流波形的相似性强于故障线路与非故障线路暂态零序电流波形的相似性,基于此提出一种基于波形伸缩变换和模糊K均值聚类的配电线路单相接地选线新方法。选取故障后半个工频周期的暂态零序电流进行处理,运用相平面分析的方法对暂态零序电流作分段相平面变换,用另一种形态来表征暂态零序电流波形;然后利用能够综合反应幅值和极性信息的欧氏距离来构造各线路的特征量,并形成特征矩阵。结合系统各馈线只需分为故障线路和非故障线路两类的特点,利用模糊K均值聚类的方法将特征矩阵分为两类,则自成一类的线路即为故障线路。在暂态零序电流信号的处理上,本文首次提出了暂态零序电流波形伸缩变换理论,伸缩变换处理较好地解决了非故障线路间暂态零序电流数值型属性大小不一的影响,提高了非故障线路间暂态零序电流波形的相似度,提高了选线的裕度。利用ATP电磁暂态仿真软件搭建10kV谐振接地系统的仿真模型,对不同接地电阻、不同接地相角、不同故障线路、不同故障位置、不同接地类型等单相接地故障情况进行大量的仿真,仿真结果表明本文所提方法均能正确选线。综合考虑工程应用中可能存在的影响因素,文中对零序电流互感器、消弧线圈补偿度、电磁噪声、采样不同步、间歇性电弧接地等可能存在的影响因素进行仿真,仿真结果表明该方法能在以上干扰条件下正确选线。大量仿真表明,本文所提单相接地故障选线方法可靠性高、选线裕度高、适应性强,具有一定的工程应用价值。
[Abstract]:In our distribution network, small current grounding is widely used, that is, neutral ungrounded mode and neutral point grounding through arc suppression coil. In the low current grounding mode, the probability of single-phase grounding fault is very high. Generally speaking, it accounts for more than 80% of the total faults. When a single-phase grounding fault occurs in a small current earthing system, the three-phase voltage still maintains a symmetrical relationship because of the small grounding current, which will not affect the continuous power supply of the load. To a certain extent, the reliability of power supply can be improved. Because of the complex and changeable condition of single-phase grounding fault in small current grounding system, the fault current is weak, and it is often accompanied by intermittent arc, as well as the structural parameters of the line. Due to the influence of nonlinear characteristics of transformer, electromagnetic interference and so on, the spectrum characteristics, energy distribution and attenuation characteristics of zero sequence current of fault lines are very different, and there is a lack of reliable method of fault line selection for single-phase grounding. The problem of grounding line selection has not been solved well. This paper summarizes the method of single-phase grounding fault line selection for distribution lines, and analyzes the shortcomings of the steady-state line selection method. The advantages of single phase grounding fault line selection using transient parameters are explained. Through the analysis of transient and steady zero sequence current of resonant grounding system, It is pointed out that the similarity of transient zero-sequence current waveforms between non-fault lines is stronger than that between fault lines and non-fault lines. Based on this, a new method of single-phase grounding line selection for distribution line based on waveform expansion transformation and fuzzy K-means clustering is proposed. The transient zero sequence current of half power frequency period after fault is selected for processing. The phase plane analysis method is used to transform the transient zero-sequence current into a piecewise phase-plane transform, and another form is used to characterize the transient zero-sequence current waveform, and then the Euclidean distance, which can synthesize the response amplitude and the polar information, is used to construct the characteristic quantities of each line. According to the characteristics of each feeder of the system, it can be divided into two categories by using fuzzy K-means clustering method, which only need to be divided into fault line and non-fault line, and the method of fuzzy K-means clustering is used to divide the feature matrix into two categories. In the processing of transient zero-sequence current signal, the theory of transient zero-sequence current waveform scaling is proposed for the first time. The scalable transformation can solve the influence of different values of transient zero-sequence current between non-fault lines, and improve the similarity of transient zero-sequence current waveform between non-fault lines. The simulation model of 10 kV resonant grounding system is built by using ATP electromagnetic transient simulation software. For different grounding resistance, different grounding phase angle, different fault line, different fault location, different grounding resistance, different grounding phase angle, different fault line, different fault location, different grounding resistance, different grounding phase angle, different fault line, different fault location, A large number of simulations have been carried out on single-phase grounding faults such as different grounding types. The simulation results show that all the methods proposed in this paper can correctly select lines. Considering the possible influence factors in engineering application, the zero sequence current transformer is discussed in this paper. The influence factors, such as compensation degree of arc suppression coil, electromagnetic noise, asynchronous sampling and intermittent arc grounding, are simulated. The simulation results show that the method can correctly select the line under the above interference conditions. The method proposed in this paper has the advantages of high reliability, high margin of line selection, strong adaptability, and has certain engineering application value.
【学位授予单位】:福州大学
【学位级别】:硕士
【学位授予年份】:2014
【分类号】:TM862
【相似文献】
相关期刊论文 前7条
1 张宏伟;卢晓荣;宋执环;;基于K均值聚类的熔炼过程烟雾检测[J];中国科技论文;2012年01期
2 王建宇;张峰;周献中;史迎春;骆文;;利用小波变换和K均值聚类实现字幕区域分割[J];计算机辅助设计与图形学学报;2006年10期
3 宁凯;;基于K均值聚类的编组站分类方法研究[J];铁道运营技术;2011年03期
4 李霞;邵春福;曹鹏;;基于快速K均值聚类的经济水平与货运量模型[J];吉林大学学报(工学版);2008年05期
5 龙飞飞;王琼;宋阳;李伟;;基于K均值聚类对Q345R钢低温拉伸的声发射信号分析[J];无损检测;2013年09期
6 李云;鲍鸿;;语音分组识别技术的研究[J];广东工业大学学报;2014年02期
7 ;[J];;年期
相关会议论文 前2条
1 吴波;王勇;王伟;王景翰;;K均值聚类分析法在绿色摇头丸分类中的应用[A];公共安全中的化学问题研究进展(第二卷)[C];2011年
2 季铎;蔡东风;郑伟;于水;;基于相似序列的增量式K均值聚类算法[A];第四届全国信息检索与内容安全学术会议论文集(上)[C];2008年
相关硕士学位论文 前8条
1 戴少颖;基于K均值聚类算法的科技馆展使用情况分析[D];云南大学;2015年
2 李力;面向大数据的云搜索引擎设计及并行K均值聚类算法研究[D];重庆大学;2015年
3 郑新桃;基于波形伸缩变换和模糊K均值聚类的配电线路单相接地选线研究[D];福州大学;2014年
4 朱曦;基于改进K均值聚类的证券时间序列奇异点研究[D];昆明理工大学;2013年
5 张f ;K均值聚类在高新技术企业信用评级中的应用[D];电子科技大学;2010年
6 廖宗明;基于K均值聚类与支持向量机的电力系统短期负荷预测[D];华南理工大学;2012年
7 李希;基于K均值聚类和BP神经网络的耐火材料损伤模式识别[D];武汉科技大学;2012年
8 都兴朔;结合人工蜂群算法的调和K均值聚类算法的研究[D];东北师范大学;2012年
,本文编号:1552190
本文链接:https://www.wllwen.com/kejilunwen/dianlilw/1552190.html