.png){kind=icon}
返回
A Cable Defect Localization Method Based on Multiresolution Singular Value Decomposition and Stochastic Subspace Identification
.png){kind=icon}
PRE
AI
DOI：10.1109/TIM.2024.3476603.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Frequency-domain reflection (FDR) stands as an effective technique for localizing cable defects. To address the issue of the significant influence of window functions and interference problems on the defect location accuracy of traditional FDR, this article proposes a cable defect location method based on multiresolution singular value decomposition (MRSVD) and stochastic subspace identification (SSI). Notably, this method eliminates the need for window function selection and effectively filters out interference signal subspaces, thereby enhancing defect location accuracy. First, the reflection coefficient spectrum (RCS) of the power cable is studied by using transmission line theory, which confirms that the RCS can be used to locate cable defects. Then, the principles of MRSVD, SSI, and density-based spatial clustering of applications with noise (DBSCAN) algorithm are introduced. By estimating the parameters of complex exponential decay oscillation functions in the RCS and removing relevant interference, a new cable fault location spectrum is proposed. Finally, the proposed method is applied to fault location in a 200-m simulated cable model and a 500-m real cable. The results demonstrate that the proposed fault location spectrum accurately locates faults in the cable with fewer interference problems, significantly improving the accuracy of the FDR method in locating cable defects.
Keyword:
Defect localization
multiresolution singular value decomposition (MRSVD)
power cable
reflection coefficient spectrum (RCS)
stochastic subspace identification (SSI)
Defect localization
multiresolution singular value decomposition (MRSVD)
power cable
reflection coefficient spectrum (RCS)
stochastic subspace identification (SSI)
期刊
IF:
5.9
论文数:
2.0W
被引数:
5.8W
机构
引用论文
暂无论文信息