基于时频能量相似度逼近的配电线路故障定位方法

Distribution line fault location method based on time-frequency energy similarity approximation

旨在解决10 kV配电线路尤其是乡村及远郊区段配电线路故障点精确定位问题,以降低线路故障维修人力物力消耗,减少故障停电损失。故障线路所有监测点中故障点的信号能量最强,且为同一套故障信号波形,因此从某监测点出发,基于信号时频能量相似度向增大的方向逼近,总可找到故障点,从而实现故障点定位。据此提出基于改进的时频相似度方法:首先合理布局适当监测点在线监测线路故障暂态电压信号,在线判别出故障类型及上下游区域;其次分上下游区域对各监测点信号分别作S变换处理成时频能量矩阵,经矩阵相似度分析逐步判别出故障支线、故障区段及故障紧邻端点;最后基于紧邻端点的电压暂态信号幅值偏差累计值与端点间距离的线性关系计算故障点位。仿真试验结果表明:相比传统技术定位,该方法具有准确性提高、全系统故障定位抗干扰能力增强、无须建立“历史故障样本库”、无须制定“相似度阈值”等优点。

This paper aims to the solution of accurate fault point location of 10 kV distribution lines, especially in rural and outer suburbs, so as to reduce the consumptions of manpower and material resources for line fault maintenance as well as the loss of fault outage of power customers. The signal energy of fault point is the strongest among all monitoring points of the faulty line, and it is the same set of fault signal waveforms.Therefore, starting from a monitoring point, the fault point can always be found and the fault point can be located based on the similarity of signal time-frequency energy approaching the direction of increasing. This paper proposes a method based on improved time-frequency similarities. Firstly, appropriate monitoring points are reasonably arranged to on-line monitor the transient voltage signal of line fault, and the fault type and the section of upstream or downstream are identified online. Secondly, the signals of each monitoring point are transformed to time-frequency energy matrix by S-transform in the upstream and downstream sections. Fault branch, fault section and adjacent fault endpoints are gradually recognized by matrix similarity analysis. Finally,the fault point is located based on the linear relationship between the cumulative values of the voltage transient signal amplitude deviation adjacent to the endpoints and the distance between the endpoints. The simulation results show that the location accuracy of this method is higher than that of traditional technology. The antiinterference ability of fault location of the whole system is enhanced, and there is no need to establish "historical fault sample database" and "similarity threshold".