首页 | 本学科首页   官方微博 | 高级检索  
     

基于多小波包和人工神经网络的电力系统故障类型识别
引用本文:李东敏,刘志刚,苏玉香,蔡军.基于多小波包和人工神经网络的电力系统故障类型识别[J].电力自动化设备,2009,29(1).
作者姓名:李东敏  刘志刚  苏玉香  蔡军
作者单位:西南交通大学电气化自动化研究所,四川,成都,610031
基金项目:教育部霍英东青年教师教育基金,四川省杰出青年基金 
摘    要:采用PSCAD/EMTDC仿真500 kV高压输电线路不同工况下的故障.先对采集到的故障电流信号进行适当的多小波包分解,并计算各频带的能量,然后构造信号的多小波包特征向量,并以此向量作为训练样本对BP神经网络进行训练,当输电线路发生故障时,将提取的故障电流信号的多小波包能量特征向量输入训练好的BP神经网络,即可实现故障类型的识别.仿真结果表明采用多小波包提取的故障电流特征量比采用传统小渡包提取的特征量信息更丰富,对人工神经网络的训练效果更好,网络识别精度具有明显优势.

关 键 词:电力系统  多小波包  传统小渡包  BP神经网络  故障类型识别

Fault recognition based on multi-wavelet packet and artificial neural network
LI Dongmin LIU Zhigang SU Yuxiang CAI Jun.Fault recognition based on multi-wavelet packet and artificial neural network[J].Electric Power Automation Equipment,2009,29(1).
Authors:LI Dongmin LIU Zhigang SU Yuxiang CAI Jun
Affiliation:Institute of Electrification & Automation;Southwest Jiaotong University;Chengdu 610031;China
Abstract:Different faults of 500 kV transmission line are simulated with PSCAD/EMTDC.Appropriate multi-wavelet packet decomposition is applied to the sampled fault current signals,the energy of each fault current frequency band is calculated and the multi-wavelet packet eigenvectors of the current signals are constructed,which are taken as samples to train the BP neural network.When a transmission line fault oc- curs,the energy eigenvectors of the sampled fault currents are extracted and then input into the trained ...
Keywords:power system  multi-wavelet packet  traditional wavelet packet  BP neural network  fault recognition  
本文献已被 CNKI 万方数据 等数据库收录!
点击此处可从《电力自动化设备》浏览原始摘要信息
点击此处可从《电力自动化设备》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号