摘要
为了优化EEMD算法的去噪效果,采用一种归一化指标来自适应优化EEMD的去噪效果。该方法对信号进行迭代EEMD分解,运用敏感IMF选取方法,自适应选取每次EEMD分解得到的敏感IMF来重构信号,并通过该归一化指标来评价去噪效果并确定EEMD中的迭代次数,得到优化的去噪信号。再对该去噪信号进行MED滤波,最后进行包络谱分析,再与轴承理论上的特征频率进行比对,从而完成故障诊断。用模拟轴承故障信号与实测信号验证了该方法的可行性。
The improved adaptive EEMD de-noising method based on sensitive IMF selection and normalized index optimization is proposed. This method is a kind of iterative EEMD decomposition, the sensitive IMF is adaptively selected at each EEMD iteration by using the sensitive IMF selection method, and then used to reconstruct the signal. Then the de-nosing effect of each iteration is evaluated and the number of optimal iterations of the EEMD decomposition is determind based on the normalized index, and the optimal de-noised signal is obtained. MED filtering is then performed on the denoised signal. At last, the failure frequency is extracted by envelope spectrum analysis. The feasibility of this method is verified by analogue signals and measured signals.
作者
邹朋
王会杰
ZOU Peng;WANG Hui-jie(College of Mechanical Engineering, Chongqing University, Chongqing 400044, China)
出处
《测控技术》
2019年第3期47-51,共5页
Measurement & Control Technology
基金
国家自然科学基金资助项目(51675064)
