Comparison of Location Accuracy between Frequency Domain Reflectometry and Line Resonance Analysis for Power Cables

• Published in: 2020 IEEE International Conference on High Voltage Engineering and Application (ICHVE) pp. 1 - 4
• Main Authors: Chen, Lu, Zhou, Yunjie, Zhang, Wei, Ye, Ting, Hu, Yuxiao, Xu, Yang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 06.09.2020
• Subjects: cable joint; fault; Fault location; frequency; Impedance; Instruments; line resonance analysis; Power cable insulation; Power cables; Reflection coefficient; Resonant frequency
• ISSN: 2474-3852
• DOI: 10.1109/ICHVE49031.2020.9279663

Frequency domain reflectometry (FDR) and line resonance analysis (LIRA) are both considered as effective methods in the detection of faults of power cables. FDR measures the spectrum of the reflection coefficient of cables, while LIRA measures the complex input impedance of cables over a broad frequency range, normally from 1 to 100 MHz. In this paper, to compare the location accuracy of FDR and LIRA in underground power cables, FDR tests designed by the authors and LIRA tests using a commercial instrument are conducted on a three-phase 10-kV single-core cross-linked polyethylene (XLPE) cable system, which is 180 m in length and has two cable joints. Results show that both LIRA and FDR can locate the points of the cable joints. However, regarding the impedance discontinuities that are far away from the test port, the LIRA method shows higher location accuracy compared with the FDR method. In addition, to investigate the abilities of the two methods to identify faulty joints, a flashover fault is simulated on one of the cable joints in phase A. Then, the obtained results indicate that both the reflection coefficient spectrum and the input impedance spectrum of phase A are smaller than those of the normal phases. After time-frequency transformation, the peaks of the faulty and non-faulty cable joints in phase A become lower than those in the normal phases, both in the FDR and LIRA methods. Overall, neither method can identify the faulty joint by comparing the curve of the fault phase with that of the normal phase.