基于VMD-ICMSE和半监督判别SOINN L-Isomap的滚动轴承故障诊断  

A rolling bearing fault diagnosis method based on VMD-ICMSE and semi-supervised discriminant SOINN L-Isomap

作　　者：戚晓利[1] 王振亚[1] 吴保林 叶绪丹 潘紫微[1] QI Xiaoli;WANG Zhenya;WU Baolin;YE Xudan;PAN Ziwei(School of Mechanical Engineering,Anhui University of Technology,Ma'anshan 243032,China)

机构地区：[1]安徽工业大学机械工程学院

出　　处：《振动与冲击》2020年第4期252-260,共9页Journal of Vibration and Shock

基　　金：国家自然科学基金(51505002);安徽省自然科学基金(1808085ME152);安徽省高校自然科学研究重点项目(KJ2017A053);研究生创新研究基金(2017012)

摘　　要：针对从滚动轴承非线性、非平稳振动信号中提取故障特征困难的问题,提出一种基于半监督判别自组织增量学习神经网络界标点的等度规映射(SSDSL-Isomap)的滚动轴承故障诊断方法。利用基于变分模态分解的改进复合多尺度样本熵(VMD-ICMSE)从复杂域提取振动信号的故障特征,构建高维故障特征集;采用SSDSL-Isomap方法对高维故障特征集进行维数约简,提取出利于识别的低维、敏感故障特征子集;应用粒子群优化极限学习机(PSO-ELM)分类器对低维故障特征进行故障识别,判别故障类型。VMD-ICMSE方法集成了VMD自适应分解非线性信号与ICMSE衡量时间序列复杂性程度的优势,提高故障特征提取能力;SSDSL-Isomap方法综合了全局流形结构、半监督型双约束图构建以及SOINN界标点选取的优点,增强故障分类能力。调心球轴承故障诊断实验分析结果表明,该方法对实验数据的故障识别率达到100%。For the difficulty that extracting fault features from the nonlinear and non-stationary vibration signals of rolling bearings,a fault diagnosis method of rolling bearing based on semi-supervised discriminant self-organizing incremental neural network landmark Isomap(SSDSL-Isomap)was proposed.Firstly,the fault features of vibration signals were extracted from the complex domain by using the improved composite multiscale sample entropy based on variational mode decomposition(VMD-ICMSE),and the high-dimensional fault feature set was constructed.Secondly,the SSDSL-Isomap method was used to reduce the dimension of the high-dimensional fault feature set,and the low-dimensional and sensitive feature subset was extracted.Finally,low-dimensional features were input into a particle swarm optimization extreme learning machine(PSO-ELM)classifier to recognize fault types.The VMD-ICMSE method integrates the advantages of VMD in adaptive nonlinear signals decomposition and ICMSE to measure the complexity of time series,and improves the ability of fault feature extraction.The SSDSL-Isomap method considers global manifold structure information,semi-supervised double-constraint graph construction and SOINN landmark selection;therefore,it enhances the ability of fault classification.Experimental results of fault diagnosis on self-aligning ball bearings show that the identifying rate of the proposed method is 100%.

关 键 词：故障诊断 滚动轴承 SSDSL-Isomap 变分模态分解(VMD) 改进复合多尺度熵(ICMSE) 粒子群优化极限学习机(PSO-ELM) 

分 类 号：TH165.3[机械工程—机械制造及自动化] TN911.7[电子电信—通信与信息系统]