摘要
在高精密的振动信号检测中,信号源的工作环境受到不特定的外部环境干扰,使得高精度信号检测存在高频低频同时噪声干扰的特征。传统的检测方法在去噪过程中忽略了这种特征对信号中高频成分的干扰,导致振动信号去噪检测效果不好。提出基于经验模态分解(EMD)和离散小波变换(DWT)的变压器振动信号去噪检测方法。采用的在EMD域加窗处理方法仅降低初始的IMF(经验模式分解)中的噪声而不是完全丢弃这些分量,比较完整地保留振动信号中的高频成分。将EMD处理后获得的信号进行DWT域变换,采用一种基于自适应软阈值降噪技术,这两种算法的结合不仅保留了DWT的去噪性能,同时保留EMD处理中含有强噪声IMF分量的高频成分,能精确检测原始振动信号。仿真结果表明,该方法信噪比高,结果与原始信号更加一致,为变压器故障信号检测提供了参考。
Based on empirical mode decomposition （EMD） and discrete wavelet transform （DWT）, we proposed a transformer vibration signal denoising method. The method of adding window in the EMD domain only reduces the noise in the initial of the empirical mode decomposition （IMF） rather than completely rejectes these component and relatively intact vibration signal of high frequency components. The EMD processed signal of DWT domain transform was received, and an adaptive soft threshold noise reduction technology was used. The combination of these two kinds of algorithms retained not only the denoising performance of DWT, but also the EMD with strong noise contained in the IMF component of high frequency components, which can accurately detect the original vibration signal. The simulation results show that the method is of high signal-to-noise ratio and more consistent results with original signal.
出处
《计算机仿真》
CSCD
北大核心
2016年第3期401-404,441,共5页
Computer Simulation
