电晕放电辐射信号的探测系统与识别方法

发布时间：2018-04-01 15:30

本文选题：电晕放电　切入点：火花放电　出处：《高电压技术》2014年09期

【摘要】：准确探测电晕放电辐射信号对精确定位电晕放电源、确保高压输电线安全稳定运行具有重要意义。为此,设计了基于差分降噪原理的探测系统,分别对背景噪声、电晕放电和火花放电信号进行了测试。实验结果及其时频分析表明:该差分模块能很好地将0~100 MHz和600~1 000 MHz范围内的广播、电视、通讯信号以及空间电磁干扰信号滤除,在复杂电磁场环境下探测到静电放电信号。火花放电与电晕放电的频谱特征相似,为排除火花放电干扰,基于学习向量化(LVQ)神经网络算法,对2者的时域波形进行了特征提取与模式识别,实现了静电放电类型的判别,且判别准确率达95%。因此,利用探测系统和提出的LVQ神经网络模式识别方法能够有效探测和识别电晕放电辐射信号,为高压输电线的电晕放电监测以及电晕放电目标定位提供参考。
[Abstract]:Accurate detection of corona discharge radiation signal is of great significance to accurately locate corona discharge power supply and ensure the safe and stable operation of high voltage transmission line.Therefore, a detection system based on differential noise reduction principle is designed, and the background noise, corona discharge and spark discharge signal are tested respectively.The experimental results and time-frequency analysis show that the differential module can filter out the radio, television, communication and space electromagnetic interference (EMI) signals in the range of 0 ~ 100 MHz and 600 ~ 1 000 MHz, and detect electrostatic discharge signals in a complex electromagnetic field environment.The spectrum characteristics of spark discharge and corona discharge are similar. In order to eliminate the spark discharge interference, the feature extraction and pattern recognition of the time domain waveforms of the two subjects are carried out based on the learning vector quantization (LVQ) neural network algorithm, and the type of electrostatic discharge is distinguished.The accuracy of discrimination was 95%.Therefore, the detection system and the proposed LVQ neural network pattern recognition method can effectively detect and identify corona discharge radiation signals, and provide a reference for corona discharge monitoring and corona discharge target location of high-voltage transmission lines.
【作者单位】：军械工程学院静电与电磁防护研究所;
【基金】：国家自然科学基金(61172035)~~
【分类号】：TM732

【共引文献】