摘要
避雷器泄漏电流是分析避雷器工作运行状态的重要特征之一,但极易受到混叠噪声干扰,因此作者提出一种将滑动窗口和奇异谱分析(Singular Spectrum Analysis,SSA)相结合的泄漏电流信号降噪算法。该方法利用滑动窗口跟踪信号之间的相似度的优势,提取信号片段特征,并将SSA与层次聚类(Hierarchical Clustering,HC)相结合,根据信号片段自身特性驱动重组信号和去除噪声,最后通过时序组合得到降噪后的泄漏电流信号。结果表明,与SSA相比,文中提出的降噪算法SNR和RMSE分别为40.183 5dB和0.342 1,能更准确、完整地去除噪声,并采用该方法对实际泄漏电流进行降噪处理,降噪效果良好。
The leakage current of lightning arrester is one of the important characteristics to analyze the operating state of lightning arrester,but it is easily disturbed by aliasing noise.Therefore,this paper proposed a leakage current signal denoising algorithm combining sliding window and singular spectrum analysis(SSA).In this method,the characteristics of signal fragments were extracted by using the similarity between sliding window tracking signals,and SSA was combined with Hierarchical Clustering(HC)to drive signal recombination and noise removal according to the characteristics of signal fragments.Finally,the denoised leakage current signal was obtained by time series combination.The experimental results show that,compared with SSA,the SNR and RMSE proposed in this paper are 40.183 5 dB and 0.342 1,respectively,which can remove the noise more accurately and completely,and the denoising effect of this method is good for the actual leakage current.
作者
齐庆周
石英
徐腊梅
谢长君
周申培
QI Qing-zhou;SHI Ying;XU La-mei;XIE Chang-jun;ZHOU Shen-pei(School of Automation,Wuhan University of Technology,Wuhan 430070,China)
出处
《武汉理工大学学报》
CAS
2021年第6期76-82,共7页
Journal of Wuhan University of Technology
基金
国家重点研发项目(2020YFB1506802)。
关键词
避雷器
泄漏电流信号
滑动窗
奇异谱分析
层次聚类
arrester
leakage current signal
sliding window
singular spectrum analysis
hierarchical clustering
