Identification of Insulation Defects Based on Chaotic Analysis of Partial Discharge in HVDC Superconducting Cable

Abstract

High-voltage direct current (HVDC) technology is considered to have some important advantages over traditional high-voltage alternating current, such as higher overall efficiency and smaller power losses for long-distance transmission. In addition, applications of superconducting cables in dc electric power networks may realize real zero impedance, and the economic and technical advantages could be maximized. Therefore, many research institutes have tried to develop advanced superconducting cables for HVDC grids with higher reliability, by considering insulation diagnosis in order to avoid unexpected failures. As one of the plausible diagnostic methods for power cables applied to the ac grid, the detection of partial discharges (PDs) taking place inside the apparatus has been widely investigated. With regard to the related PD pattern analysis, a phase resolved PD analysis (PRPDA), which was first developed in the early 1970s, accounts for the phase information of the applied ac voltage. In 2001, we also proposed a method for pattern recognition, i.e., chaotic analysis of PD (CAPD), that considers three normalized parameters obtained from the values between two consecutive PD pulses: amplitude difference (Pt), occurring time difference (Tt), and applied voltage difference (Vt). However, none of the proposed methods of pattern analysis can be employed for PD under dc stress. Therefore, in this paper, we propose a modified CAPD for the related pattern recognition of possible defects inside a joint box and termination of an HVDC superconducting cable. PDs are produced from four artificial defects and are then detected by a self-designed and fabricated sensor, for which the analysis was performed based on our newly modified CAPD.