| Gas Insulated Switchgear (GIS) is widely used in power system for its compact structure, small footprint and easy to maintain. It could lead to large area blackout and huge losses to the national economy once GIS fails. So internal partial discharge (PD) detection of GIS insulation faults timely is significantly important to guaranteeing safety and reliability of GIS. Aiming at existing problems in GIS PD detection, a set of GIS PD ultrasonic detection system based on optical fiber transmission is designed, combined with photoelectric conversion technology, optical fiber transmission technology, digital signal processing technology and virtual instrument technology.A set of GIS PD detection hardware system based on optical fiber transmission is developed. Amplifying and filtering circuit designed in the paper effectively improved the SNR of collected signal and removed interfering signal on the basis of extracting discharge signals accurately. The remote collection module adopting optical pulse synchronization trigger signal solved simultaneous acquisition problem of multiple signals problem. Digital signal transmitted through optical fiber communication systems with high anti-interference ability realized accurate signal transmission and electrical isolation of high-voltage side and low high-voltage side. The system realized data sending and receiving between host computer and lower computer by UDP-protocol, and improved the expansion capability.The corresponding software system to hardware system and interactive interface were developed based on LabVIEW virtual instrument technology. The system can realize data acquisition, data processing, data storage, data query, etc. The data acquisition module realized data communication by UDP network protocol. The data processing module can realize filtering processing of collected ultrasonic signals by wavelet analysis and extracts the partial discharge data by threshold setting.The GIS partial discharge test platform is set up, by which four kinds of typical defects in GIS operation were detected, and then got typical discharge characteristics of different defect models. Using fractal theory, fractal parameters of different defect model were calculated, and the parameters are applied to clustering recognition as characteristic quantity by fuzzy C-means. The results show that the use of extracted fractal parameters combined with fuzzy C-means algorithm can recognize four types of discharge defects effectively, which is a new method for pattern recognition for GIS PD.Relying on the national 863 scientific research project, the author conducted a series of site detection of GIS PD in substations of Xi’an, Shijiazhuang and Baoding, and found several PD defects. Experimental research and site applications have verified the effectiveness of the system. |
PDF Full Text Request