基于ISGMD-CMAE的滚动轴承 复合故障特征提取方法  

Compound Fault Feature Extraction Method for Rolling Bearing Based on ISGMD-CMAE

作　　者：李颖 于家奇 吴仕虎 巴鹏[1] 马小英 LI Ying;YU Jiaqi;WU Shihu;BA Peng;MA Xiaoying(Shenyang Ligong University,Shenyang 110159,China;Tianmen Vocational College,Tianmen 431700,China)

机构地区：[1]沈阳理工大学机械工程学院,沈阳110159 [2]天门职业学院职业技术教育中心,湖北天门431700

出　　处：《沈阳理工大学学报》2025年第3期81-89,共9页Journal of Shenyang Ligong University

基　　金：国家自然科学基金项目(51934002);辽宁省教育厅青年科技人才“育苗”项目(LJKZ0259);辽宁省属本科高校基本科研业务费专项资金资助项目(LJ212410144039)。

摘　　要：滚动轴承复合故障信号具有非线性和不确定性的特点,且信号中含有噪声,直接提取故障特征存在困难,为此提出一种基于改进辛几何模态分解(ISGMD)和复合多尺度注意熵(CMAE)的滚动轴承复合故障特征提取方法(ISGMD-CMAE)。针对辛几何模态分解(SGMD)方法中分解信号分量过多,导致信号特征过于分散,无法进行有效提取的问题,采用聚类算法对信号分量进行处理,依据相关系数和峭度构成的综合评价指标筛选分量重构信号,以突出故障特征;针对多尺度注意炳(MAE)方法在提取时序信号过程中会造成信息损失的问题,采用摘值稳定性较好的CMAE方法准确全面地提取故障信号。实验结果表明,本文提出的ISGMD-CMAE方法能够精准地对滚动轴承复合故障特征进行提取,为滚动轴承故障诊断提供了一种新思路。The compound fault signal of rolling bearings exhibits nonlinear and uncertain characteristics,often accompanied by noise,which makes harder the direct extraction of fault features.To address the challenge of extracting compound fault features in rolling bearings,a novel method is proposed,which integrates improved symplectic geometric mode decomposition(ISGMD)with complex multi-scale attention entropy(CMAE).To address the issue of excessive signal component decomposition in symplectic geometry mode decomposition(SGMD)method,which leads to overly dispersed signal features and hinder effective extraction,clustering algorithms is employed to process the signal components.The components are then filtered and the signal is reconstructed based on a comprehensive evaluation index composed of correlation coefficient and kurtosis,in order to highlight fault features.To solve the problem of information loss during the extraction of time-series signals using multi-scale attention entropy(MAE)method,adopt CMAE method is adopted, which has better entropy stability,to accurately and comprehensively extract fault signals.Experimental data analysis demonstrates that the ISGMD-CMAE method can effectively extract thecompound fault features of rolling bearings, offering an innovative approach in the field of fault feature extraction of rolling bearing.

关 键 词：滚动轴承 故障特征 辛几何模态分解 多尺度注意熵 

分 类 号：TH133.33[机械工程—机械制造及自动化] TP2[自动化与计算机技术—检测技术与自动化装置]