摘要
故障行波零模分量与线模分量到达线路端测距装置处的时间差与故障距离之间呈非线性关系,反向传播BP神经网络具有良好的非线性逼近拟合能力,可以利用BP神经网络对这种关系进行拟合以达到故障测距的目的。以小波分解各个频带下模量传输时间差为输入特征量、故障距离作为输出特征量,对神经网络进行训练,形成故障测距神经网络模型。仿真表明该基于模量传输时间差的特高压直流输电线路单端故障测距神经网络算法精度较高。
There is a nonlinear relationship between fault distance and the propagation time gap between zero mode and aerial mode, and back propagation （BP）neural network with the ability of nonlinear curve-fitting is selected to fit this re- lationship to realize fault location . The propagation time gap in different frequency bands of wavelet decomposition as inputs and the fault distance as outputs are used to train the neural network, and thus a faul location model using neural network is established. Simulation results indicate that the proposed single-ended fault location method useing neural network for ultra high voltage direct current（UHVDC） transmission line based on propagation time gap has higher accu- racies.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2016年第8期20-24,共5页
Proceedings of the CSU-EPSA
基金
国家自然科学基金资助项目(51267008)
昆明理工大学校人培基金资助项目(kkz3201304019)
关键词
特高压直流
行波
神经网络
小波分解
故障测距
ultra high voltage direct current （UHVDC）
traveling wave
neural network
wavelet decomposition
faultlocation
