变转速下滚动轴承的阶频谱相关域微弱故障特征增强与提取  

Enhancement and Extraction of Weak Fault Features in Order-Frequency Spectral Correlation Domain for Rolling Bearings Under Variable Speed Conditions

作　　者：王冉 黄裕春 张军武 余亮 WANG Ran;HUANG Yu-chun;ZHANG Jun-wu;YU Liang(Logistics Engineering College,Shanghai Maritime University,Shanghai 201306,China;State Key Laboratory of Mechanical System and Vibration,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海海事大学物流工程学院,上海201306 [2]上海交通大学机械系统与振动国家重点实验室,上海200240

出　　处：《测控技术》2023年第3期79-86,92,共9页Measurement & Control Technology

基　　金：国家自然科学基金(51505277,12074254);国家重点研发计划项目(2019YFB2004600);上海市自然科学基金资助项目(21ZR1434100)。

摘　　要：滚动轴承是旋转机械常用且故障率较高的部件之一,其故障的及时发现,对于设备安全、稳定运行具有重要意义。滚动轴承的早期故障特征十分微弱,容易被强背景噪声干扰所掩盖。同时,滚动轴承往往在变转速工况下运行,故障特征的时变特性导致特征提取较为困难。针对上述问题,提出一种变转速下滚动轴承的阶频谱相关(OFSC)域微弱故障特征增强与提取方法。首先,利用变转速下滚动轴承故障信号的角度时间域循环平稳特性,将故障信号转换到阶频谱相关域。然后,采用鲁棒主成分分析(RPCA)的低秩稀疏分解方法,将轴承振动信号的阶频谱相关矩阵分解为表征轴承故障特征的稀疏成分,并去除表征噪声的低秩成分,进一步提高稀疏分量的分辨率。最后对分解出的稀疏分量构建增强包络阶次谱(EEOS)来检测滚动轴承的故障特征。仿真和实验分析验证了该方法对于变转速工况轴承微弱故障特征增强和提取的有效性和鲁棒性。Rolling bearing is one of the common and high failure rate parts of rotating machine,and its faulty detection is essential for the safe and stable operation of the equipment.The initial fault characteristics of rolling bearings are very weak,which can be easily masked with strong background noise interference.Meanwhile,rolling bearings often operate under variable speed conditions,and the time-varying characteristics of fault features make the feature extraction more difficult.To address the above problems,a method that extracts weak fault features of rolling bearings based on the order-frequency spectral correlation(OFSC) under variable speed conditions is proposed.Firstly,the rolling bearing fault signals are converted to the OFSC domain by using the angular/time cyclostationarity of the rolling bearing fault signal under variabe speed.Then,the low-rank and sparse decomposition method of the robust principal component analysis(RPCA) is used to decompose the OFSC matrix of the bearing vibration signals.Where the sparse components characterizing the rolling bearings fault features are obtained,and the low-rank components characterizing the noise are removed to further improve the resolution of the sparse components.Finally,the enhanced envelope order spectrum(EEOS) is constructed for the sparse components to detect the fault characteristics of rolling bearings.Simulated and experimental analysis verify the effectiveness and robustness of the proposed method to enhance and extract the weak fault features of bearings with variable speed conditions.

关 键 词：滚动轴承 变转速工况 角度时间域循环平稳 阶频谱相关 鲁棒主成分分析 

分 类 号：TH212[机械工程—机械制造及自动化] TP29[自动化与计算机技术—检测技术与自动化装置]