基于多信号和改进经验傅里叶分解的故障特征提取方法  

Fault feature extraction method based on multi-signals improved empirical Fourier decomposition method

作　　者：朱丹宸 何伟 朱群伟 ZHU Danchen;HE Wei;ZHU Qunwei(Department of Electro-mechanics,Naval Petty Officer Academy,Bengbu 233012,China;College of Mechanical Engineering,Jiujiang Vocational and Technical College,Jiujiang 332007,China;Guangzhou Bureau,Naval Equipment Department,Guangzhou 510220,China)

机构地区：[1]海军士官学校机电系,安徽蚌埠233012 [2]九江职业技术学院机械工程学院,江西九江332007 [3]海军装备部驻广州地区军事代表局,广东广州510220

出　　处：《机电工程》2024年第10期1853-1864,共12页Journal of Mechanical & Electrical Engineering

基　　金：国家自然科学基金资助项目(52201362)。

摘　　要：针对滚动轴承故障信号受设备多结构和复杂传递路径干扰,故障诊断准确性受到影响这一难题,提出了一种基于多信号改进经验傅里叶分解(MS-IEFD)的轴承故障特征提取方法。首先,为了充分利用多信号中的故障特征,利用改进经验傅里叶分解处理了两个不同测点测得的故障信号,设定各阶模态信号与原始信号的相关系数阈值为0.1,并以此为依据确定了信号分解的最佳个数;然后,提出了加权的谐波显著性指标对初始的频带划分进行了优化,避免了信号过分解,减少了带宽过窄的无效频带,以此指标最大值为准,确定了最优模态分量;其次,借助互相关分析的优势,分析了两信号的最优模态分量以进一步增强信号的特征成分,借助快速傅里叶变换准确提取了滚动轴承的故障特征,判断了轴承故障类型;最后,利用MS-IEFD方法对仿真和实验信号进行了分析,仿真分析时构造了信噪比为-10 dB和-15 dB的信号,用以模拟不同测点信号的情况,实验分析时利用实验台测得了不同测点处的滚动轴承振动信号。研究结果表明:MS-IEFD方法能从强背景干扰中准确提取出滚动轴承故障特征,为准确判断滚动轴承故障类型提供依据;与变分模态分解(VMD)等方法相比较,可进一步突出MS-IEFD方法在弱特征提取方面的有效性。To address the problem that the bearings fault signals are usually contaminated by strong background interference due to multiple structures and complex transmission path,which affects accurate fault feature extraction of the bearings.A multi-signals improved empirical Fourier decomposition(MS-IEFD)method was proposed.Firstly,to fully utilize the fault characteristics in multiple signals,the improved empirical Fourier decomposition was applied to two signals acquired from different measuring points,the optimal decomposition number was determined based on the correlation coefficients threshold 0.1 between each modal functions and the original signal.Secondly,the weighted harmonics significant index was constructed to optimize the initial band division results,it reduced the invalid frequency bands with too narrow bandwidth and determined the optimal modal components.Then,with the advantage of cross-correlation analysis,the optimal modal functions of the two signals were analyzed and the fault signature was further enhanced.With the help of fast Fourier transform,the fault features of bearings were accurately extracted and the type of bearing fault was judged.Finally,the MS-IEFD method was applied to the simulation and experimental signals.Two simulation signals with the signal-to-noise ratios of-10 dB and-15 dB were constructed to simulate the signals at different measurement points,the bearing vibration signals at different measurement points were selected for the experimental analysis.The research results show that the MS-IEFD method can effectively extract the fault characteristics of rolling element bearing from strong background interference.When comparing with methods such as the variational mode decomposition(VMD),the effectiveness of the MS-IEFD method is further highlighted.

关 键 词：滚动轴承 故障诊断 多信号 频带划分 互相关谱 多信号改进经验傅里叶分解 变分模态分解 

分 类 号：TH133.3[机械工程—机械制造及自动化]