轴系轮廓重构信号的CEEMDAN—小波阈值去噪方法

CEEMDAN-wavelet Thresholding Denoising Method for Axial System Profile Reconstruction Signals

针对轴系回转过程中动不平衡引起的位移误差信号含噪问题,提出自适应噪声完备集合经验模态分解(CEEMDAN)—小波阈值去噪的轴系轮廓重构模型。采用CEEMDAN对位移信号进行分解得到各阶本证模态函数(IMF),采用相关性分析提取含噪的IMF分量,并对其进行小波阈值去噪,与经验模态分析EMD—小波阈值去噪方法进行比较,最后将提纯后的信号进行重构。模拟仿真表明,去噪后的信号不仅保持了原有信号的特征,并且有效去除了噪声。将去噪后的信号输入到重构轮廓模型进行试验,结果表明,去噪后分离的单个截面回转误差准确度提高了0.05μm,圆度误差准确度提高了0.0703μm。

The adaptive noise complete set empirical mode decomposition(CEEMDAN)-wavelet threshold denoising model is proposed to reconstruct the shaft system profile for the problem of noisy displacement error signal caused by dynamic unbalance during the shaft system rotation.CEEMDAN is used to decompose the displacement signal to obtain each order of the intrinsic modal function(IMF),then correlation analysis is used to extract the noisy IMF components,wavelet threshold denoising is applied to them,and compared with EMD-wavelet threshold denoising method by empirical mode analysis,finally the purified signal is reconstructed.Simulation results show that the denoised signal not only maintains the characteristics of the original signal but also effectively removes the noise.The results show that the accuracy of the individual cross-sectional rotation error after noise removal is improved by 0.05μm and the accuracy of roundness error is improved by 0.11μm.