基于神经网络的输电线路行波故障测距方法

Traveling Wave Fault Location Method for Transmission Line Based on ANN

针对单端行波测距法难以区分故障点反射波和对端母线反射波的问题,提出一种利用神经网络非线性拟合能力实现故障定位的测距方法。首先利用相模变换获取故障反向行波,对反向行波进行小波模极大值变换;然后提取模极大值突变时刻和对应极性,从中选取3个有效时刻构成满足时差、波速与线路长度关系的一组"有效时差对";最终配合极性信息构造神经网络的输入样本,对网络进行训练和测试后,可利用该网络输出故障距离,从而实现故障定位。仿真结果表明,该方法适用性强,有较高的可靠性和精确度。

Aiming at the difficult problem of identification of the travelling wave reflected from the fault point or the remote-end bus for the single terminal fault location, a fault location method was proposed by using the nonlinear fitting ability of neural network. The method firstly utilizes phase mode transformation to obtain the fault reverse travelling wave and it was processed by the wavelet modulus＂ maximum transformation. Then the time of sudden change point for modulus maximum and the corresponding polarity were extracted. Three effective moments were selected to form ＂a set of effective time difference＂ for meeting a relationship of time difference, wave speed and line length. Finally, input sam- ples of neural network were formed by adding polarity information. The fault distance was obtained by the output of neu- tral network model, which is trained and tested. The simulation results show that the method is widely applicable and ef- ficient in fault location with higher accuracy.