多域特征提取和MD-MTS的滚动轴承故障诊断方法  

Rolling Bearing Fault Diagnosis Approach Based on Multi-Domain Features and MD-MTS

作　　者：彭宅铭 程龙生[1] 詹君 姚启峰 PENG Zhai-ming;CHENG Long-sheng;ZHAN Jun;YAO Qi-feng(School of Economic&Management,Nanjing University of Science&Technology,Jiangsu Nanjing 210094,China)

机构地区：[1]南京理工大学经济管理学院,江苏南京210094

出　　处：《机械设计与制造》2022年第2期152-156,共5页Machinery Design & Manufacture

摘　　要：为及时提取滚动轴承的有效故障特征,准确识别其故障状态,提出一种多域特征提取和多维马田系统(MD-MTS)相结合的故障诊断方法。该方法主要包括三个方面的内容:振动信号的多域特征提取、基于MD-MTS的故障诊断模型构建和实验验证。首先,利用统计分析、快速傅里叶变换(FFT)、Hilbert变换和改进的经验模态分解(EMD)等方法提取故障信息,构建初始特征集。然后,根据多域特征构建MD-MTS实现滚动轴承多故障状态的识别。最后,通过实验和比较对所提出的模型进行了全面评估学习。结果表明,该方法能够有效地检测轴承故障,对滚动轴承的不同故障状态具有较高的诊断精度。To timely extract the effective fault characteristics of rolling bearing and accurately identify the fault state,a novel method of rolling bearing fault diagnosis based on multi-dimensional Mahalanobis-Taguchi System(MD-MTS)with multi-domain features was proposed. This method mainly consisted of three steps:Multi-domain feature extraction,MD-MTS fault diagnosis model construction and experimental verification. Firstly,statistical analysis,Fast Fourier Transformation(FFT),Hilbert transform and improved Empirical Mode Decomposition(EMD)have applied to extract and construct the feature set. Secondly,the MD-MTS model was constructed by multi-domain features to implement the identification of multiple fault conditions. Finally,the proposed model was comprehensively evaluated through experiments and comparisons. The findings show that the proposed method can effectively detect bearing faults and have high diagnostic accuracy in different working conditions of rolling bearings.

关 键 词：滚动轴承 多域特征 多维马田系统(MD-MTS) 故障诊断 

分 类 号：TH16[机械工程—机械制造及自动化] H133.33[语言文字—汉语]