论文部分内容阅读
为了实现电晕放电辐射信号的自动化监测与识别,以PXIe-5185数字化仪为主要硬件,采用虚拟仪器技术和神经网络算法研制了电晕放电辐射信号监测与识别系统。该系统模仿示波器设计操控界面,具有信号采集、数据处理、频域分析、数据存储、报告生成等功能。系统综合采用触发电平限制、数字滤波和神经网络识别3种于信号处理,可有效降低数据存储压力。试验结果表明,BP神经网络可有效识别与区分电晕放电和火花放电,在改变电极、增加电晕放电样本复杂性的情况下,对电晕放电的识别率仍然可达87%。监测系统为进一步研究电晕放电信号的特性奠定了基础,其信号处理流程可推广应用于其他脉冲信号监测领域。
In order to realize the automatic monitoring and recognition of corona discharge radiation signals, the PXIe-5185 digitizer is used as the main hardware. The corona discharge radiation signal monitoring and identification system is developed by using virtual instrument technology and neural network algorithm. The system simulates the oscilloscope design control interface, with signal acquisition, data processing, frequency domain analysis, data storage, report generation and other functions. System integration using the trigger level limit, digital filtering and neural network identification three kinds of signal processing, can effectively reduce the pressure on data storage. The experimental results show that the BP neural network can effectively identify and distinguish between corona discharge and spark discharge, and the recognition rate of corona discharge can still reach 87% when changing the electrodes and increasing the complexity of the corona discharge samples. The monitoring system lays the foundation for further research on the characteristics of corona discharge signals, and the signal processing flow can be widely applied to other pulse signal monitoring fields.