基于LLE-TVF和PE-MOMEDA的货车轴承复合故障诊断方法  

Composite fault diagnosis method of truck bearings based on LLE-TVF and PE-MOMEDA

作　　者：杨兴宽 孙祯 周素霞 石珮廷 杨延峰 YANG Xingkuan;SUN Zhen;ZHOU Suxia;SHI Peiting;YANG Yanfeng(Institute of Metals and Chemistry,China Academy of Railway Sciences Group Co.,Ltd.,Beijing 100081,China;College of Mechatronic and Vehicle Engineering,Beijing Architecture University,Beijing 100044,China)

机构地区：[1]中国铁道科学研究院集团有限公司金属及化学研究所,北京100081 [2]北京建筑大学机电与车辆工程学院,北京100044

出　　处：《振动与冲击》2024年第23期56-65,74,共11页Journal of Vibration and Shock

基　　金：中国国家铁路集团有限公司科技研究开发计划、重点项目(铁路货车轮轴修程修制改革技术研究)(N2022J050);中国铁道科学研究院集团有限公司科研项目、重点项目(货车段修轮轴质量自动检测技术研究)(2022YJ208)。

摘　　要：考虑到在嘈杂的噪音环境中,货车滚动轴承的复合故障特征相对模糊,并且各个故障特征之间的相互影响导致了复合故障特征的有效区分困难,提出了基于李雅普诺夫指数(largest Lyapunov exponents,LLE)的全变分滤波(total variation filtering,TVF)和加入排列熵(permutation entropy,PE)的多点最优最小熵解卷积(multipoint optimal minimum entropy deconvolution adjusted,MOMEDA)的滚动轴承复合故障诊断方法。首先对复合故障信号进行噪声分析,根据信号存在噪声大小来确定李雅普诺夫指数和信号混沌性,同时将惩罚设置为合适的正则化参数从而实现对复杂环境噪声的自适应降噪,然后通过排列熵改变MOMEDA中的滤波器长度对故障信号进行解卷积运算,分离出不同的故障特征,对信号作傅里叶变换提取故障特征频率,最后利用Teager能量算子增强解卷积后的故障冲击信号,实现滚动轴承复合故障的精确判别。通过将此方法应用于仿真信号模拟滚动轴承复合故障以及实际货车轴承复合故障进行验证,结果表明此方法可以实现复合故障特征的准确分离,成功识别出故障类型。Here,Considering composite fault features of truck bearings being relatively fuzzy in noisy environment,and effects among various fault features making it difficult to effectively distinguish them,a rolling bearing composite fault diagnosis method based on largest Lyapunov exponents(LLE)total variation filtering(TVF)and permutation entropy(PE)-added multi-point optimal minimum entropy deconvolution adjusted(MOMEDA)was proposed.Firstly,noise analysis was performed for composite fault signal to determine Lyapunov exponent and signal chaos based on the magnitude of noise existing in signal.Meanwhile,the penalty was set as an appropriate regularization parameter to realize adaptive denoising of complex environmental noise.Then,the fault signal was deconvoluted by changing the filter length in MOMEDA with permutation entropy to separate different fault features.Fault feature frequencies were extracted through the fault signal’s Fourier transform.Finally,Teager energy operator was used to enhance the deconvoluted fault impact signal,and realize accurate discrimination of rolling bearing composite faults.This method was used to recognize bearing composite faults in simulation signals and actual truck bearing composite faults for verification,the results showed that the proposed method can accurately separate characteristics of truck bearing composite faults and successfully recognize fault types.

关 键 词：货车轴承 故障诊断 故障监测 特征提取 李雅普诺夫函数 

分 类 号：U279.3[机械工程—车辆工程]