3차원 Multilateration을 이용한 가스절연변압기 내부 절연결함 위치 판별기법 개발

Abstract

Since the number of service failure of Gas insulated apparatus under service in power grid is increased, any reliable and relevant insulation diagnosis technologies are highly required to be developed. Many a investigation has been precedently carried out to detect the partial discharge signals, for which various methods have been investigated and then reported: acoustic signal detection, laser polarization detection with Pokels cell, chemical decomposition detection, and electromagnetic wave detection. However, they seem to be inappropriate for the detection and analysis of partial discharge signals and their related pattern recognitions. Besides, it is likely necessary to establish a credible method to find the location of the presence of the insulation defects that could be introduced into Gas insulated apparatus. In this work, our interest is focused to develop fault location method which could be, in near future, employed to GITr (Gas Insulated Transformer) since its application has been gradually expanded in Korean power grid. For this purpose, UHF sensors has been designed and fabricated to detect the partial discharge signals and then our own method has been developed to discriminate the surrounding noises and the estimation of the fault location has been carried out by use of 3D Multilateration. Experimental set-up consists of GITr Mock-up of 170kV-22.5kV with 1/3 down scaled and instrumentation facilities to detect the partial discharges from the artificial defects introduced into the mock-up. For a moment, the main part of the transformer winding was removed to avoid the reflection and diffraction of the PD signals from this structure. Furthermore, 5 algorithms (3 our own and 2 commercially available) have been investigated and then the results have been compared to obtain the proper one for our mock-up.