基于参数自适应VMD的滚动轴承故障特征提取

Fault Feature Extraction of Rolling Bearing Based on Parameter Adaptive VMD

针对提取的滚动轴承故障特征信号易受复杂工作环境的影响以及变分模态分解(variational mode decomposition,VMD)参数依赖人为经验选择的问题,课题组提出了一种基于参数自适应的VMD的滚动轴承故障特征提取方法。首先,以原始信号经过VMD后的固有模态函数(intrinsic mode function,IMF)的包络谱熵作为适应度函数,采用猎豹优化(cheetah optimizer,CO)算法对分解阶数k、惩罚因子α进行自适应寻优;其次,基于峭度准则对各IMF分量进行重构;然后,对重构信号进行Hilbert包络谱分析从而提取故障特征,并通过滚动轴承故障仿真信号和实验信号经VMD和SSA-VMD处理结果对比验证可行性。研究结果表明:该方法相比于经典VMD所得故障特征更为准确;在参数寻优时间方面CO算法相比麻雀搜索算法(sparrow search algorithm,SSA)提升了65%。课题组的研究具有一定工程应用价值。

Aiming at the problem that the fault feature signal extraction of rolling bearing was easily affected by complex working environment and the parameters of variational mode decomposition(VMD)were selected by human experience,a method of fault feature extraction of rolling bearing based on parameter adaptive VMD was proposed.Firstly,the envelope spectral entropy of the intrinsic mode function(IMF)of the original signal after VMD was used as the fitness function,and the cheetah optimizer(CO)algorithm was used to optimize the decomposition order k and penalty factorαadaptively;Secondly,the IMF components were reconstructed based on the kurtosis criterion;Then,Hilbert envelope spectrum analysis was performed on the reconstructed signal to extract fault features,and feasibility was verified through simulation signal and experimental signal.The research results show that this method is more accurate in extracting fault characteristics compared to classical VMD;In terms of parameter optimization,CO algorithm has increased by 65%compared to SSA.The research has certain engineering application value.