基于经验小波变换的变压器振动信号特征提取

发布时间：2018-02-24 15:41

本文关键词： 变压器振动信号特征提取经验小波变换 Hilbert谱时频表示特征矢量　出处：《电力系统自动化》2017年20期 　论文类型：期刊论文

【摘要】：为有效提取变压器振动信号特征,提出了一种基于经验小波变换(EWT)的信号特征提取方法。首先利用EWT方法将不同工况的变压器振动信号分别分解为若干经验小波函数(EWF)分量;然后计算各分量Hilbert谱,通过时频表示直观反映不同工况变压器振动信号的频率特征信息;最后计算不同工况振动信号各EWF分量与原信号的相关系数,并提取相关度高的分量,根据其能量构建信号的特征矢量,实现对不同工况变压器振动信号特征提取的量化处理。仿真试验表明,该方法能有效提取变压器振动信号特征,且根据提取的特征矢量能够正确识别变压器绕组所属的不同工况。
[Abstract]:In order to extract the characteristic of transformer vibration signal effectively, a signal feature extraction method based on empirical wavelet transform (EWT) is proposed. Firstly, the transformer vibration signal under different working conditions is decomposed into several empirical wavelet function components by EWT method. Then the Hilbert spectrum of each component is calculated, and the frequency characteristic information of transformer vibration signal under different working conditions is directly reflected by time-frequency representation. Finally, the correlation coefficient between each EWF component of vibration signal under different working conditions and the original signal is calculated, and the components with high correlation degree are extracted. The characteristic vector of the signal is constructed according to its energy, and the quantization processing of feature extraction of transformer vibration signal under different working conditions is realized. The simulation results show that this method can effectively extract the characteristics of transformer vibration signal. According to the extracted feature vectors, the transformer windings can be correctly identified under different working conditions.
【作者单位】：华北电力大学电气与电子工程学院;
【基金】：国家重点研发计划资助项目(2016YFB0901200) 中央高校基本科研业务费专项资金资助项目(2017XS013)~~
【分类号】：TM41

【相似文献】