采用改进HVD与Lempel-Ziv复杂性测度的滚动轴承早期损伤程度评估方法

Assessment Method of Early Damage Severity for Rolling Bearing Based on the Improved HVD and Lempel-Ziv Index

针对滚动轴承早期故障特征微弱难以定量检测的问题,提出一种基于希尔伯特振动分解(HVD)和Lempel-Ziv复杂性测度(LZC)的滚动轴承内外圈损伤程度评估方法。该方法利用HVD将滚动轴承故障信号分解为幅值由大到小、不同瞬时频率的HVD初始分量,基于峭度分析从各HVD初始分量中选出最优分量并重构该故障信号,采用均值符号化方法计算重构信号的LZC指标。关于HVD的端点效应问题,采用镜像延拓方法对其进行改进,仿真分析的结果表明,镜像延拓有效地改进了其端点效应的问题,提高了故障信号的分解精度;在此基础上,将HVDLZC方法应用于不同损伤程度的滚动轴承内外圈故障信号,结果表明该方法能够解除信号非线性、非平稳性干扰,快速提取信号的故障振动模式分量,给出内外圈各损伤状态的HVD-LZC特征区间;采用由遗传算法进行参数寻优的支持向量机(GA-SVM)方法对3种不同损伤状态的内外圈信号的HVD-LZC特征进行测试分析,结果表明HVD-LZC指标在区分轴承内外圈的不同损伤程度上具有优越性,从而为滚动轴承内外圈损伤程度的评估提供了一种有效方法。

A fault severity assessment method is proposed to solve the problem that early fault signals of rolling bearings are weak and difficult to be quantitatively measured,and the method uses Hilbert vibration decomposition（HVD）combined with Lempel-Ziv complexity（LZC）.HVD is adopted to decompose fault signals of rolling bearings into initial HVD components ordered from large to small and with different instantaneous frequencies.The best initial HVD component is selected based on the kurtosis index analysis,and then the component is used to reconstruct the fault signals.The normalized LZC index of the reconstructed signals is calculated using the mean symbolic method.The mirror extension method is adopted to deal with endpointeffect of HVD.Simulation results show that the mirror extension method effectively improves the endpoint effect of HVD,and the decomposition accuracy is obviously increased.The HVD-LZC method is used to analyze fault signals of the inner and outer rings of rolling bearings with different level of defect severity,and the results show that the method eliminates the nonstationary and non-linear interferences.As a result,the main fault vibration modes are extracted and the HVD-LZC characteristic intervals for different damage state are estimated; The characteristics of HVD-LZC for three kinds of fault severities of rolling bearing are modeled using the GA-SVM based on RBF kernel.Experimental results show that the HVD-LZC method recognizes different levels of fault severity of the inner and outer rings of rolling bearings.Hence,this study provides an effective fault severity assessment method for rolling bearings.