GIS设备刀闸振动信号噪声分离  

Noise Separation from Vibration Signal of GIS Isolation Switch

作　　者：范展滔[1] 吴小刚 吕耀棠 陈兴望 FAN Zhantao;WU Xiaogang;LÜYaotang;CHEN Xingwang(China Southern Power Grid Company Limited,Guangzhou 510623,China)

机构地区：[1]中国南方电网有限责任公司,广州510623

出　　处：《电力系统及其自动化学报》2023年第7期29-35,共7页Proceedings of the CSU-EPSA

基　　金：国家自然科学基金资助项目(62203321);中国博士后科学基金资助项目(2021M692390)。

摘　　要：气体绝缘封闭式组合电器刀闸动作产生的振动可用于刀闸状态二次确认。针对为减小现场各类振动干扰对二次确认结果的影响,需将噪声分离并去除的问题,本文提出一种改进经验小波算法。该方法通过计算信号频谱包络线来突出能量集中的频段以避免峰值过于密集;在此基础上提出频谱阈值分割法以分割频谱,并构建经验小波函数用以将信号分解为多个经验模态;根据刀闸振动信号特征,筛选经验模态将噪声去除。实验表明,改进经验小波算法对刀闸振动信号和噪声的分离效果较好。The vibration signal generated by the on-off action of a gas insulated switchgear(GIS)isolation switch can be used for the secondary confirmation of the isolation switch status.To reduce the impact of various vibration distur⁃bances on the secondary confirmation results,the noise should be separated and removed.Aimed at this problem,an improved empirical wavelet algorithm is proposed in this paper,which highlights the frequency band with concentrated energy by calculating the signal spectrum envelope,thus avoiding the over-dense peak values.On this basis,a spec⁃trum threshold segmentation method is put forward to segment the frequency spectrum,and empirical wavelet functions are constructed to decompose the signal into multiple empirical modes.Finally,the noise is removed by filtering the em⁃pirical modes based on the characteristics of the isolation switch’s vibration signal.Experimental results show that the improved empirical wavelet algorithm has a good effect on separating noises from the isolation switch’s vibration signal.

关 键 词：振动信号 气体绝缘封闭式组合电器刀闸 噪声分离 经验小波变换 

分 类 号：TM734[电气工程—电力系统及自动化]