| 一种特征提取新方法在PQD识别中的应用 |
| |
| 引用本文： | 张新闻,周春艳. 一种特征提取新方法在PQD识别中的应用[J]. 宁夏大学学报(自然科学版), 2010, 31(3): 243-246,252 |
| |
| 作者姓名： | 张新闻 周春艳 |
| |
| 作者单位： | 北方民族大学电气信息工程学院,宁夏银川,750021 |
| |
| 基金项目： | 北方民族大学科研基金资助项目 |
| |
| 摘 要： | 从时域和频域出发,提出一种Hilbert变换和Fourier变换相结合的方法,快速识别电能质量扰动.该方法先利用Hilbert变换提取扰动信号的包络,根据包络线提取扰动特征,将5种扰动从常见的8类扰动中分离,对未识别3类扰动的包络线进行FFT变换,再从频域中提取特征对其分类.借助Matlab随机生成800个样本,此方法在无噪声和30 dB高斯白噪声条件下平均分类准确率分别达到了99.7%和99.1%.算法在特征提取耗费时间上是采用DB4小波的特征提取方法的25.7%.
|
| 关 键 词： | 电能质量扰动 Hilbert变换 包络线 特征提取 Fourier变换 |
A New Method of Feature Extraction Applied in Recognition of PQD |
| |
| Zhang Xinwen,Zhou Chunyan. A New Method of Feature Extraction Applied in Recognition of PQD[J]. Journal of Ningxia University(Natural Science Edition), 2010, 31(3): 243-246,252 |
| |
| Authors: | Zhang Xinwen Zhou Chunyan |
| |
| Affiliation: | (School of Information Engineering,Northern China University for Nationality,Yinchuan 750021,China) |
| |
| Abstract: | A new method of feature extraction is presented for recognizing disturbances by means of Hilbert transform combined with Fourier transform.Those envelope features of the disturbance signal are extracted by Hilbert transform,and then 5 disturbances are classified from others.The left disturbances are recognized by the features extracting from Fourier transform of envelope.800 PQD disturbance samples for simulations of identification were carried on via the Matlab software.The correct rates of identification reached 99.7% without noise and 99.1% with 30 dB white Gaussian noise respectively.Especially,the time consumed on feature extraction is 25.7% of the Wavelet transform when same recognition algorithm is used. |
| |
| Keywords: | power quality disturbances Hilbert transform envelope feature extraction Fourier transform |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
