配电网络故障选线与定位研究
发布时间：2018-08-02 18:08
【摘要】：配电线路单相接地故障稳态电流微弱,故障特征不明显,故障电弧不稳定,故障检测较困难。配电网还具有“闭环结构,开环运行”的特点,分支众多、结构复杂,且拓扑结构经常随着系统不同的运行方式而发生变化,即使准确检测到了故障并获得了故障距离,也很难确定故障所在分支或者是故障点的确切位置。另外,架空线发生导线坠地、树木接触架空馈线引起的高阻故障,使得故障检测在实际运行中的操作变得非常困难。加之线路通常只在变电站侧有量测点,使配电线路保护和定位大多只能依靠单端算法。行波法测距亦面临着受馈线数量多、各馈线长短相差很大,负荷干扰等因素影响而造成的首波头不易可靠标定,以及由于行波的折返射增多、传播过程分析复杂而造成的后续波头不易有效识别等难题。上述因素共同决定了配电网故障选线和故障定位的特殊性、复杂性和艰巨性,如何准确、迅速地识别故障并确定故障位置,长期以来一直是未能有效解决的难题。随着用户对供电可靠性和电能质量要求持续提升,配电线路故障快速检测和迅速修复的重要性日益增强,工程现场亟需可靠、有效的故障选线和定位技术。 本文利用小波分析原理提取各馈线零序电流最高频带下的小波系数,根据故障线路小波系数与健全线路小波系数相位相反的原理构建配网选线新方法。在直配线路发生单相接地故障时,由于对波形性质分析较为复杂困难,利用人工神经网络强大的学习和泛化能力,将前三个波头的极性和二、三波头与首波头的时差作为输入属性,实现故障定位。由于多分支辐射配网节点的影响,行波波头的识别更为困难,而频谱分布的不同能够体现故障位置,据此探寻辐射配网的故障定位方法。
[Abstract]:The steady-state current of single-phase grounding fault of distribution line is weak, the fault characteristic is not obvious, the fault arc is unstable, and the fault detection is difficult. The distribution network also has the characteristics of "closed loop structure, open loop operation", many branches, complex structure, and the topology structure often changes with different operation modes of the system, even if the fault is detected accurately and the fault distance is obtained. It is also difficult to determine the exact location of the fault branch or fault point. In addition, the high resistance fault caused by the overhead wire falling and the tree contact the overhead feeder makes it very difficult to operate the fault detection in the actual operation. In addition, the distribution line protection and location can only rely on single terminal algorithm. Traveling wave method is also faced with a large number of feeder lines, the length of each feeder is very different, the first wave head is not easy to be reliably calibrated due to factors such as load interference, and because of the increase of traveling wave reentry, It is difficult to identify the subsequent wave head due to the complex analysis of the propagation process. The above factors together determine the particularity, complexity and difficulty of fault line selection and fault location in distribution network. How to identify faults accurately and quickly and determine fault location is a difficult problem that can not be effectively solved for a long time. With the continuous improvement of power supply reliability and power quality requirements, the importance of rapid fault detection and rapid repair of distribution lines is increasing, and reliable and effective fault line selection and location techniques are urgently needed in the engineering site. In this paper, wavelet coefficients are extracted under the highest frequency band of zero sequence current of each feeder by using the principle of wavelet analysis, and a new method of distribution network line selection is constructed according to the principle that the phase of fault line wavelet coefficient is opposite to that of sound line wavelet coefficient. In the case of single-phase grounding fault in direct distribution line, because of the complex analysis of waveform properties, the polarity and two of the first three wave heads are combined with the powerful learning and generalization ability of artificial neural network. The time difference between the three wave head and the first wave head is used as input attribute to realize fault location. It is more difficult to identify the traveling wave head because of the influence of the multi-branch radiation distribution network node, and the different spectrum distribution can reflect the fault location, so the fault location method of the radiation distribution network is explored.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM862
[Abstract]:The steady-state current of single-phase grounding fault of distribution line is weak, the fault characteristic is not obvious, the fault arc is unstable, and the fault detection is difficult. The distribution network also has the characteristics of "closed loop structure, open loop operation", many branches, complex structure, and the topology structure often changes with different operation modes of the system, even if the fault is detected accurately and the fault distance is obtained. It is also difficult to determine the exact location of the fault branch or fault point. In addition, the high resistance fault caused by the overhead wire falling and the tree contact the overhead feeder makes it very difficult to operate the fault detection in the actual operation. In addition, the distribution line protection and location can only rely on single terminal algorithm. Traveling wave method is also faced with a large number of feeder lines, the length of each feeder is very different, the first wave head is not easy to be reliably calibrated due to factors such as load interference, and because of the increase of traveling wave reentry, It is difficult to identify the subsequent wave head due to the complex analysis of the propagation process. The above factors together determine the particularity, complexity and difficulty of fault line selection and fault location in distribution network. How to identify faults accurately and quickly and determine fault location is a difficult problem that can not be effectively solved for a long time. With the continuous improvement of power supply reliability and power quality requirements, the importance of rapid fault detection and rapid repair of distribution lines is increasing, and reliable and effective fault line selection and location techniques are urgently needed in the engineering site. In this paper, wavelet coefficients are extracted under the highest frequency band of zero sequence current of each feeder by using the principle of wavelet analysis, and a new method of distribution network line selection is constructed according to the principle that the phase of fault line wavelet coefficient is opposite to that of sound line wavelet coefficient. In the case of single-phase grounding fault in direct distribution line, because of the complex analysis of waveform properties, the polarity and two of the first three wave heads are combined with the powerful learning and generalization ability of artificial neural network. The time difference between the three wave head and the first wave head is used as input attribute to realize fault location. It is more difficult to identify the traveling wave head because of the influence of the multi-branch radiation distribution network node, and the different spectrum distribution can reflect the fault location, so the fault location method of the radiation distribution network is explored.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM862
【参考文献】
相关期刊论文 前10条
1 束洪春,肖白;配电网单相电弧接地故障选线暂态分析法[J];电力系统自动化;2002年21期
2 吴振升,杨学昌,曹振，
本文编号：2160255
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2160255.html
