Study Of Inner Couplers Detecting Partial Discharge In Gas-insulated Substations

The on-line partial discharge monitoring technique of power apparatus is very important to find insulation faults in time and increase the reliability of power supply. Therefore, the study of the on-line partial discharge monitoring technique of power apparatus is of practical values. As to gas-insulated substations, on one hand, PD detection is one of the most basic ways in diagnosis of insulation in GIS. On the other hand, it is very difficult to detecting PD in GIS. The thesis is devoted to developing the detection of PD using couplers built in GIS from the aspect of electrical measurement.On-line monitor of PD signals in GIS is a comprehensive research project, which relates to the theories and knowledge of many different areas, such as microwave theory, electrical measurement and digital signal processing theory. The outlines of thesis are as follows:For the configuration of GIS and the characteristic of partial discharge, the thesis analyses the electromagnetic waves excitated by the PD signal and the reason of a coupler built in GIS inducing PD signal in terms of the electromagnetic theory and antenna theory. The author frames a GIS model and designs a common disc coupler and a unique loop coupler to detect PD signals.Theory analysis and practical measurements of response characteristics are done about two couplers. From the aspect of capacitive coupling, the frequency response characteristics of two sensors are assessed in terms of square wave response. At the same time, the amplitude-frequency relationship characteristics are calculated on the base of antenna model. Lastly the practical measurement results validate that the couplers designed is excellent for the partial discharge signals detecting.On the base of broadband loop coupler, some basic experiments are carried out under different conditions of insulation and pressure on the base of broadband loop coupler. And some useful rules are concluded from these experiments. The PD signals on different frequency scopes are surveyed, and the thesis explores anti-noise narrowband detection of partial discharge signals. For the first time the thesis brings forward the choosing-frequency detection of PD ground on the radio frequency communication. In order to solve the problem that noises such as corona cannot be removed by routine electrical measurement methods, the ultra high frequency band of the signal is chosen for the PD detection. The results of experiments proved that theUHF method detecting PD signals has the ability of resisting corona in air.Wavelet analysis is used to process the signal of PD in GIS. Since the ultra high frequency energy is weak and the white noise is in existence, threshold method has been put forward to reject the white noise on the base of the wavelet packet decomposition.