摘要
为提高电网故障诊断的准确率和速度,提出一种将小波分时灰度矩与概率神经网络相结合的电网故障诊断方法,通过对小波灰度矩进行时间上的划分,计算得到故障发生后电流在不同时刻的灰度矩的值,从而得到小波系数随时间的变化情况;以小波分时灰度矩作为概率神经网络的输入,诊断结果作为输出,实现对电网故障的自动诊断,利用PSCAD/EMTDC对电网不同类型的故障进行了仿真,采用连续小波变换对电网发生短路故障后的暂态信息进行分析,提取其灰度矩信息,利用概率神经网络进行了故障识别。仿真结果表明,小波分时灰度矩具有较强的细节表现能力,可作为电网故障的故障特征,与概率神经网络相结合可有效地实现对电网故障的自动识别。
A novel fault diagnosis method combined with wavelet time-division gray moment and probability neural network（PNN） is presented to improve the accuracy and speed of fault diagnosis of power system.Fault current wavelet gray moment is divided into time-division gray moments to indicate the details of the wavelet coefficients change over time.Time-division gray moments are used as the input of the PNN,and the output of the result can achieve automatic fault diagnosis of power.Different types of faults are simulated with PSCAD/EMTDC,continuous wavelet transformer is used to the transient information analysis,and their time-division gray moments are calculated to be the input of the PNN,which the fault diagnosis is achieved.Simulation results indicate that time-division gray moment,with strong performance capacity of details,can be an effective fault characteristic to diagnose the fault of power network.
出处
《重庆邮电大学学报（自然科学版）》
北大核心
2012年第1期121-126,共6页
Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)
基金
四川省教育厅青年基金(11ZB100)~~
关键词
小波分时灰度矩
故障特征
概率神经网络
故障诊断
wavelet time-division gray moment
fault feature
probability neural network（PNN）
fault diagnosis
