| Gas insulated switchgear(GIS),as an important electrical equipment,is used in high and super voltage field worldwide because of its compact structure,small area occupied,high reliability and security when compared with conventional opened electricity substation device.However,due to its closed structure design,it will bring difficulties in fault detection and positioning,complex maintenance,long power failure time after the accident and large area power failure problems.So the traditional method to examine and repair regularly cannot meet the requirements of the device.It is significant to design a device which can record and analysis the work state of the GIS.In this article,a method to sample and analysis GIS current data was designed through the study of GIS trip/close coil current waveform.First,the system control AD converter AD7276 to sample 8 ways GIS mechanical action digital current signals by using FPGA’s high speed data processing capability.And record the data into the SDRAM buffer by using FPGA’s parallel processing capability in the meanwhile.After sampling GIS signal,the system will import the data from SDRAM to SD card through SPI bus controller.Then transfer the current data from SD card to PC device.Use Matlab toolbox to filter the current wave and find the characteristic value.Finally,put the characteristic value into the trained BP neural network to realize the fault classification.Moreover,this article has designed and developed a FPGA core system board,and built a sample and record test platform.It also realizes a fault classify test platform by using BP neural network based on genetic algorithm optimization.Experimental results show that this GIS mechanical action sample and fault analysis system is accurate in measurement,convenient in carrying and stability in operation. |