Identification of insulation defects by modified Chaotic Analysis of partial discharge under DC stress

Abstract

HVDC technology has been considered to give several advantages over the traditional HVAC such as higher overall efficiency, higher current carrying capability and lower power losses for long-distance transmission. However, it has been reported that HVDC power apparatus have experienced sometimes vital failures bringing about heavy damages or inconveniences to the customers. Therefore, many research institutes have tried to develop any advanced power apparatus employable to the HVDC grids with higher reliability including the diagnosis system to avoid the unexpected failures. As one of the plausible diagnostic methods for the AC power apparatus, the detection of PD (Partial Discharges) taking place inside the apparatus has been widely investigated. With regards to the related PD pattern analysis, PRPD (Phase Resolved Partial Discharge) Analysis was firstly developed in the early 1970s taking the phase information of the AC applied voltage into account. In 2001, we also proposed a method, CAPD (Chaotic Analysis of PD), considering three normalized parameters obtained from the values between two consecutive PD pulses: Magnitude difference (Pt), Occurring time difference (Tt) and Applied voltage difference (Vt). However, all the methods for PD analysis considering AC applied voltage proposed cannot be employed for the PDs under DC stress. Therefore, in this paper, a modified CAPD is proposed for the analysis of PDs occurring from the possible defects inside the apparatus under service at DC. For the preliminary investigation, experiments have been done by using four artificial defects fabricated imitating vital defects formed during the manufacturing process. PDs from them are detected by our chip type sensor and then analyzed, for which relevant hardware and software is under development for the commercial purpose.