基于多层联合降噪的信号处理方法

针对旋挖钻机钻杆振动信号具有非线性、非平稳的特点,以及多源振动耦合影响,提出了一种基于多层联合信号降噪方法,对振动信号进行降噪处理。首先,采用局域均值分解(local mean decomposition, LMD),得到一系列乘积函数(product functions, PF),根据计算得出的相关系数,挑选出含噪声成分最多的PF分量,舍弃残余分量,实现第一层降噪;其次,利用小波阈值降噪(wavelet threshold denoising, WTD),对挑选分量实现了第二层降噪;最后,将WTD重构信号设为奇异值分解(singular value decomposition, SVD)的前置处理单元,实现第三层降噪。基于MATLAB仿真实验与轴承数据降噪实验,分别使用EMD-SVD、LMD-SVD两种算法对目标信号进行降噪处理,LMD-WTD-SVD方法可以提高信噪比,并对比波形图与频谱图结果表明,此方法是一种更有效的降噪方法。

Research on Signal Processing Method Based on Multi-layer Joint Denoising

In view of the non-linear and non-stationary vibration signals of rotary drilling RIGS and the coupling effects of multi-source vibration, a noise reduction method based on multi-layer joint signals was proposed to reduce the noise of vibration signals. Firstly, local mean decomposition (LMD) is adopted to obtain a series of product functions (PF). According to the calculated correlation coefficient, the PF component with the most noise components is selected and the residual component is discarded. Realize the first layer noise reduction; Secondly, wavelet threshold denoising (WTD) is used to achieve the second level of noise reduction for the selected components. Finally, the reconstructed WTD signal is set as the preprocessing unit of Singular Value Decomposition (SVD) to achieve the third layer noise reduction.Based on MATLAB simulation experiment and bearing data noise reduction experiment, two algorithms, EMD -SVD and LMD -SVD, are used to de-noise the target signal. LMD -SVD -WTD method can improve the signal-to-noise ratio, and by comparing waveform and spectrum, it can be seen that this method is a more effective method to de-noise.