Development And Application Of UHF Sensor For GIS

With the development of Chinese economy,Gas Insulated Switchgear(GIS)is more and more widely used in power systems because of its advantages of small floor area and high operational reliability.However,due to some defects caused by the manufacturing or handling of GIS equipment,different types of partial discharge may gradually occur in GIS,and some partial discharge types may cause damage to the internal insulation of electrical equipment and eventually lead to serious power failure.Therefore,how to more accurately detect the partial discharge signal and judge the type of partial discharge are the key problems in the field of online monitoring of power equipment.There are many methods for partial discharge detection in GIS,such as the chemical detection method,ultrasonic detection method,etc.among them,the ultra-high frequency method(UHF)has become the main method for partial discharge detection in GIS because of its high sensitivity and strong anti-interference.As the core of the UHF detection method,the development and application of the UHF sensor have always been the focus of research.Firstly,this paper discusses and studies the principle of partial discharge in GIS and the propagation characteristics of electromagnetic waves generated by partial discharge in waveguide systems,which provides a theoretical basis for the application of the UHF detection method.According to the characteristics of Hilbert curve,this paper designs UHF antenna suitable for GIS by HFSS software,and optimizes antenna parameters such as antenna size,conductor width,dielectric plate thickness and feed point position,and makes GIS UHF sensor that meets the design requirements.According to different partial discharge generation mechanisms,four typical GIS partial discharge defect models are designed,and the corresponding simulation platform is built,and the designed UHF sensor is used to test and record the data of different partial discharge signals.To better analyze the experimental data,this paper constructs the Phase Resolved Partial Discharge(PRPD)spectrum of the data.By analyzing PRPD spectra,this paper summarizes the characteristics of different partial discharge defects and extracts 18-dimensional statistical eigenvalues according to the differences of different partial discharges in PRPD spectra.In this paper,a support vector machine(SVM)is used to classify and identify the eigenvalues of different partial discharges.The results show that the partial discharge signals measured by the Hilbert UHF sensor can be effectively classified by the support vector machine.The paper also carries out on-line monitoring of GIS experimental platform and GIS equipment in actual operation in the power supply section of Nanning Railway Bureau,which also proves the effectiveness of the UHF sensor designed in this paper.