Cable Defect Localization Method Based on Subspace Decomposition
王尉军 1毛先胤 2李波 2邹雕 2李原3
扫码查看
点击上方二维码区域，可以放大扫码查看
作者信息
1. 贵州电网有限责任公司贵阳供电局,贵州省贵阳市 550004
2. 贵州电网有限责任公司电力科学研究院,贵州省贵阳市 550004
3. 四川大学电气工程学院,四川省成都市 610065
折叠
摘要
频域反射法(frequency domain reflectometry,FDR)是目前电缆缺陷定位的有效方法之一.针对传统FDR的缺陷定位效果受窗函数和干扰项影响大的问题,该文提出一种基于子空间分解的电缆缺陷定位方法,该方法不需要选择窗函数,并且滤除了干扰信号的子空间,因此拥有更好的缺陷定位效果.首先,利用传输线理论研究电缆的反射系数谱,证实了反射系数谱可用于定位电缆缺陷.然后,介绍了子空间分解技术、改进的贝叶斯信息准则(Bayesian information criterion,BIC)及基于密度聚类算法的原理,通过估计反射系数谱中各复指数衰减振荡函数的参数,并滤除相关干扰,提出了新的电缆缺陷定位谱图.最后,根据所提方法,对200m仿真电缆模型与500m真实电缆开展缺陷定位.结果表明,该定位谱图可准确地定位电缆中缺陷,并且干扰项较少,较大程度地提高FDR方法对电缆缺陷的定位正确率.
Abstract
Frequency domain reflection(FDR)is one of the effective methods for cable defect localization.To solve the problem of the defect location result of traditional FDR being greatly affected by window function and interference term,this paper proposes a cable defect location method based on subspace decomposition.In the proposed method,choosing the window function is unnecessary,and the subspace of the interference signal is eliminated.As a result,the proposed method can obtain an excellent defect localization effect.Firstly,power cable's reflection coefficient spectrum is studied using transmission line theory,which confirms that the reflection coefficient spectrum can be used to locate cable defects.Then,the subspace decomposition technology,improved Bayesian Information Criterion(BIC)and density-based spatial clustering of applications with noise are introduced.By estimating the parameters of the complex Exponential decay oscillation function in the reflection coefficient spectrum and eliminating the interferences,the new cable defect location spectrogram is proposed.Finally,based on the proposed method,defect localization experiments are performed on a 200m simulated cable model and a 500m actual cable.The results show that the proposed defect location spectrogram can accurately locate the defect in the power cable with rare interference terms,which can greatly improve the accuracy of the FDR method in locating cable defects.
关键词
电力电缆/反射系数谱/缺陷定位/子空间分解/BIC准则/DBSCAN聚类
Key words
power cable/reflection coefficient spectrum/defect localization/subspace decomposition/BIC criterion/DBSCAN
维普
万方数据