摘要
现有行波测距装置可以同时对线路雷击和普通短路故障进行检测与定位,但其不能将两者区分开来,为此提出了一种基于小波能量谱和神经网络理论的输电线路雷击与短路故障的识别方法。首先,利用小波变换将故障测距装置采集到的各种电流行波信号分解为不同频带的重构信号,并计算信号在各个频带内的能量,提取小波能量谱,然后构造信号的小波能量分布特征向量,将其作为BP神经网络分类器的输入,最终实现雷击与短路故障的识别。仿真结果显示,该方法在不同故障相角和过渡电阻的情况下均能达到满意的识别正确率,是一种有效的线路雷击与普通短路故障识别方法。
To identify two transient signals, this paper presents a wavelet energy spectrum and neural network method based on algorithm to distinguish lightning strike and fault on transmission line. Firstly, the wavelet transform was used to analyze the transient signal collected by the traveling wave location device, and the wavelet energy of different frequency bands of the signal was calculated so as to extract the wavelet energy spectrum. Secondly, the wavelet energy spectrum based eigenveetor was formed, which was taken as the input of the Bp neural network. Finally, the identification between lightning strike and fault was realized through the analysis of the neural network.. Lightning strikes at different points on the 500kV transmission line were simulated using the Electromagnetic Transient Program （EMTDC）. Db4 wavelet transform was used to analyze the transient signal. Extensive simulation results show that the proposed method can achieve satisfactory precision of identification under different fault angles and transient resistance conditions, so it is an effective method to identify lightning strike and fault on transmission line.
出处
《高电压技术》
EI
CAS
CSCD
北大核心
2007年第10期64-68,共5页
High Voltage Engineering
基金
四川省重点科技攻关项目(02GG021-025)。~~
关键词
输电线路
小波能量谱
神经网络
雷击
短路故障
行波
识别
transmission line
wavelet energy spectrum
neural network
lightning
fault
traveling wave
identification
