The Research Of High-Speed Sampling And Processing System Monitoring GIS Partial Discharge Signal

This paper introduces a data acquisition system based on DSP and aimed to collect GIS partial discharge signals.The introduction includs the hardware design and reliazation, and embedded software program. Then it proves the function of the system by collecting sine wave signals and GIS partial discharge signals. In the later part of the paper, it introduces how to transform algorithm in MATLAB to embedded C code directly and run in CCS platform. Then it validates the practicality of this algorithm by processing the partial discharge signals which were collected.The data acquisition system uses distributed architecture and can be devided into three parts: one DSP module, four AD modules and one main module. DSP module is the master part of the system and DSP chip is TI’s TMS320C6713B with high-speed, high-performance and floating-point. Its external interfaces include SDRAM, FLASH, network interface and server chips, MCU, etc.. DSP communicates with server through Ethernet, to receive and implement the sampling instruction, configuration instruction, initialization instruction, algorithm instruction and so on,which realized the control and data transmition; AD module is the sampling module. High-speed ADC is in charge of the sampling and its external interfaces include SRAM, CPLD, etc.; Main module provides the power supply and, in accordance with custom-designed bus, provides a bus slots for connecting DSP module and AD modules stably.DSP establishes communication with server through Ethernet. There are two kinds of communications in system: TCP communication and UDP communication. The server uses UDP to send sample instruction and makes sure that all the acquisition system would accept the instruction and execute it at the same time. The server also uses TCP to send other instructions including initaliztion, reset, configuration, transportation and so on.In order to reduce the pressure of data transfer and storage, it is needed to realize the algorithm in DSP, such as narrow-band interference filter. We took advantage of the own function in MATLAB to realize the algorithm and then transformed it to embedded C code by Embedded MATLAB. We ran it directly in CCS platform and greatly reduce the development pressure. Then, we respectively drawed the time domain and frequency domain map of the GIS partial discharge signal before and after processing, and showed the practicality of the algorithm clearly.