摘要
对108次/s高速采样率获取的超宽带局放脉冲波形-时间序列,使用等效时频法提取了局部放电脉冲群的波形特征参数,并研究了模糊C均值聚类算法的特性及其在局部放电脉冲波形特征向量参数处理中的应用.基于GIS的局部放电试验数据处理表明:由等效时频特征提取和模糊C均值聚类分析组成的局部放电脉冲群快速分类技术可以对多个局部放电源构成的脉冲群进行准确分类.这为研制基于单个脉冲的局部放电超宽带检测与模式识别系统提供了实用的脉冲群快速分类技术.
A novel partial discharge (PD) detection and pattern recognition system are developed with ultra-wideband (UWB) based on a single pulse shape, where a fast grouping technique for PD pulse sequence is adopted. Equivalent time-frequency method (ETFM) is used to extract fea- tures for PD pulse shape-time sequence detected by UWB with a sampling rate of 10^8 per second. And the application of fuzzy C-means clustering (FCM) in fast grouping to PD pulse sequence is investigated based on feature parameters of pulse shapes. The analysis result of noised PD date of a gas insulated switch (GIS) shows that the fast grouping technique conducts a good separation for multi-PD pulses sequence which provides a feasible and practical fast grouping technique for PD pulse sequence to develop the multi-PDs detection and pattern recognition system.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2008年第8期1021-1025,1060,共6页
Journal of Xi'an Jiaotong University
基金
河北省自然科学基金资助项目(F2007000636)
关键词
局部放电
脉冲波形-时间序列
等效时频
模糊C均值聚类
partial discharge
pulse shape-time sequence
equivalent time-frequency
fuzzy C-means clustering