摘要
进行局部放电(Partial Discharge,PD)在线监测时,为解决传统降噪算法难以提取湮没在强白噪声干扰中的多态性局部放电脉冲的问题,提出一种小波变换(WT)与数学形态学(MM)相融合的方法。该算法在传统小波硬阈值降噪方法的基础上加入形态滤波环节,对小波分解系数进行滤波处理。仿真分析以强白噪声干扰下的多态性局部放电信号为研究对象,采用降噪后信号的信噪比和均方误差为降噪性能指标,分析比较该方法与其他传统降噪方法的降噪性能,对实测局部放电信号研究结果表明该方法与传统的小波阈值方法相比,能够更有效地抑制强噪声干扰并较为完整地提取局部放电脉冲信号,且对母小波选取的依赖性低、计算量小、便于实现。将该方法运用于局部放电实测信号的处理,取得了良好的降噪效果。
In order to solve the denoising problem of multi-mode partial discharge （PD） pulses overwhelmed in white noises in PD on-line monitoring, a kind of novel denoising method which combines wavelet transform （WT） with mathematical morphology （MM） is designed in this paper. This method processes wavelet coefficients by adding a mathematical morphology filtering unit to traditional wavelet hard threshold filtering. Multi-mode PD simulation signals under high-energy white noises is studied, denoising performance is measured by examining the signal noise ratio （SNR） and mean square error （MSE）. Compared with traditional WT, the result shows that this method is able to remove high-energy white noises while preserving complete features of PD pulses with less computational complexity and less dependence on the selection of mother wavelet. This method also performs well on on-site PD signals.
出处
《陕西电力》
2013年第6期1-5,75,共6页
Shanxi Electric Power
基金
国家重点基础研究发展计划项目(973项目)资助(2009CB724500)
中央高校基本科研业务费专项资金资助(201120702020002)
关键词
局部放电
白噪声
小波变换
数学形态学
在线监测
partial discharge
white noise
wavelet transform
mathematical morphology
on line monitoring
