Fault Localization Method for Submarine Observation Network Based on Swin-Transformer Fault Information Mining
The submarine observation network is often affected by the marine environment and human factors,leading to insulation dam-age in the photoelectric composite cable and the formation of electrical fault points in contact with seawater.Accurately locating electri-cal fault points is crucial for improving the reliability of power and information transmission in the underwater observation network.Firstly,a transmission model is established according to the transmission structure of the submarine observation network,and the tran-sient currents propagating from the electrical fault points to the observation points are derived.Subsequently,continuous wavelet trans-formation is applied to extract intrinsic correlational features between transient currents and fault points.Finally,a neural network,Swin-Transformer,is utilized to explore the matching relationship between intrinsic correlational features and fault distances,thereby locating the electrical fault points.The research results indicate that under the conditions of the sample test set of intrinsic correlational features,the positioning error of electrical fault points in photoelectric composite cables in the length of 160 km is less than 400 m.This method provides an effective approach for locating electrical fault points in the submarine observation network of long-distance pho-toelectric composite cables.
submarine observation networkphotoelectric composite cableelectrical fault pointstransient currentSwin-Transform-erfault point localization
万方数据
维普
