基于时频能谱研究的配电网单相接地故障定位方法-location method of single-phase ground fault in distribution network based on time-frequency energy spectrum research.docxVIP

摘要为保障电力系统安全、稳定、经济运行，我国中低压配电网主要采用小电流 接地系统，此系统中各类故障发生的概率以单相接地故障最高。单相接地故障发 生后，系统能继续运行一段时间，若不能及时找出故障位置并加以处理，很容易 扩大故障范围，导致电力系统供电可靠性降低。因此，快速准确的故障定位，对 缩短供电恢复时间、提高运行可靠性、减少经济损失等具有极其重要的意义。但 由于小电流接地系统单相接地故障特征提取困难，使得其故障选线与定位问题一 直没有得到很好的解决。基于此，本文针对小电流接地系统单相接地故障选线与 定位方法展开研究。通过对单相接地故障暂态特性的详细分析，基于小波变换良好的时频分析能 力，提出了基于零序电流暂态能量的小电流接地系统单相接地故障选线方法。利 用小波变换对各馈线零序电流分解、重构，计算各尺度下重构信号的能量值，选 取能量最大的频带作为特征能量频带，通过改进暂态能量的计算方法，缩小健全 线路之间的暂态能量差异，放大故障线路和健全线路之间的暂态能量差异，从而 以线路能量最大为原则来定位故障线路。仿真结果表明，该故障选线方法在不同 故障条件下均能正确选线，缩短了选线时间，克服了零序电流信号倒相给选线带 来的影响，具有良好的适应性。在正确选择故障线路的基础上，针对故障点相邻两侧暂态零序电流波形的差 异，结合相似度计算方法，提出了基于时频能谱相似度的故障区段定位方法。对 故障线路各检测点暂态零序电流信号进行小波包分解，计算信号在各频带下的能 量值，并将各频带的能量值等时间间隔划分，得到各检测点信号精细的时频能谱 矩阵，进一步计算各相邻检测点之间的能谱相似度，以时频能谱相似度最小的区 域作为故障区域，实现故障定位。仿真结果验证了本文所提故障区段定位方法的 可行性，结果表明其克服了接地电阻、故障初始角等因素的影响。关键词：配电网；单相接地故障；故障选线；故障定位；时频分析；小波变换； 相似度计算AbstractTo ensure the security and stability of economical power system operation in China, small current grounding system is mainly used in lo w voltage distribution network of China. The existing literatures have shown that the single-phase grounding fault is the most prone to various types of power system fault. When the line fault occurs, the system operation stay stable for a period of time, but it is easy to expanding the scope of the fault and reducing power supply reliability of the power system if the fault position may not be found in time. In many cases, it is desirable to have a protective relay operate as fast as possible in order to control damage to major equipment and preserve system stability. Therefore, locating the fault position quickly and accurately has extremely important significance to shorten the power recovery time, reduce the economic loss and improve security of the grid. Because the fault feature extracting is difficult, the problem of fault line selection and fault position location in distribution network has not been effectively solved. In this case, the paper proposes a procedure based on the time-frequency spectrum analysis of wavelet tra