Data Acquisition System Of Vacuum Degree Of Vacuum Interrupter Based On DSP

The breaking and insulating ability of vacuum interrupter is determined by vacuum degree.Once the vacuum degree of a vacuum switch exceeds the max value permitted, the vacuumswitch would fail to break correctly and directly endanger the safety of the power system. Theresult of conventional measurements for vacuum interrupter is accurate only within a certainrange of vacuum pressure and vulnerable to be interfered. In order to obtain accurate data, thepulsed magnetron discharge method and the improved potential of shielding case method waschoosed.BP neural network was used to distribute and integrate the redundancy data whichwould improve the accuracy. It could provide data support to vacuum interrupter life predictionand management expert system.BP neural network is a powerful information processing system and has highly nonlinearfitting capacity which can be mapped arbitrarily complex nonlinear relationship. Its learningrule is simple and achieve computer easyly, also it has strong robustness and memory capacity.That is the system requirements. According to the offline detection method of vacuuminterrupter and the vacuum changes characteristics, DSP was choosen as the data acquisitionand processing platform and BP neural network was choosen as the data fusion method. Dualsensor information which was processed by the DSP and fused by BP neural network.Then theindividual sensor datas were fused by the global BP neural network. So data fusion system ofvacuum degree was setted up which was based on the DSP.To meet the requirement of high-speed of BP neural network, DSP F2812was used for thelarge amounts of data acquisition and processing.An external memory and eCAN transmissioncircuit were designed for the data storage and transmission. The precision logarithmic amplifierLOG104was adopted with process sensor signal to meet the input voltage requirement of DSP.The fusion datas were communicated by the USB-CAN bus adapter then the data wascommunicated with PC and analyzed by the visual of VB page. The hardware foundation wasprovided for data fusion and transmission of vacuum.Finally the system programs were written. The experiment of signal which was acquiredand processed and BP neural network data fusion Matlab simulation then process the voltagesignal corresponding to vacuum value. Last combined with the host computer.The data fusionmethod of BP neural network could effectively improve the consistency of measurements,reduce the impact of the external environment changing on the measurement results which hasbeen showed in the experiment.