Identifiability Improvement for Minor Defect in Cables Using FMCW and Adaptive Density-Based Clustering

During the process of defect identification in cables, the conventional frequency domain reflectometry (FDR) method directly determines the defect position through the localization spectrum, exhibiting a low accuracy in minor defect detections. To address the issue, by introducing the density-based spatial clustering of applications with noise (DBSCAN) method into the method using frequency modulated continuous wave (FMCW), in this paper, a combined method based on FMCW-DBSCAN is proposed. Firstly, with the injection of FMCW signal into cable’s head, the beat signal with characteristic information of defects and the diagnostic function are both obtained. Then, through DBSCAN, the adaptive parameters are automatically determined based on the diagnostic data. Defects can be intuitively identified according to the data categories after clustering, which enhances the recognizability of those indistinct defect peaks through peak extraction. Finally, the effectiveness of the method is verified by simulation, experiments on the 1100m laboratory cable, and field tests of 110kV cables. It is indicated that the proposed method can significantly improve the identifiability of minor defects in cables, with the localization error in field tests within 0.84%.