基于MCKD-NMD的滚动轴承故障诊断  

Research on Fault Diagnosis Method of Rolling Bearing Based on MCKD-NMD

作　　者：张志宇 章翔峰[1] 姜宏[1] ZHANG Zhiyu;ZHANG Xiangfeng;JIANG Hong(School of Mechanical Engineering,Xinjiang University,Urumqi 830047,China)

机构地区：[1]新疆大学机械工程学院,乌鲁木齐830047

出　　处：《组合机床与自动化加工技术》2023年第6期93-96,101,共5页Modular Machine Tool & Automatic Manufacturing Technique

摘　　要：针对滚动轴承故障信号具有复杂非线性且极易受到强烈噪声干扰导致的故障特征难以提取的问题,提出了一种非线性模式分解(nonlinear mode decomposition,NMD)算法和最大相关峭度解卷积(maximum correlated kurtosis deconvolution,MCKD)相结合的滚动轴承故障诊断方法。首先,利用MCKD对原始信号中的噪声进行了滤波和处理,实现了噪声的初步动态降噪并且进一步提高原始信号中的相关峭度值;其次,利用小波分解提取出时频脊线;再次,利用NMD算法分解时频脊线并筛选出其中的有效分量;最后,采用包络谱分析来提取出故障特征。结果表明,该算法所提取的分量仅包括转频以及轴承故障特征频率,说明该算法对提取该类轴承故障特征的有效性,可为轴承故障早期诊断方法的研究提供参考。In the case of rolling bearing failure,the vibration signal has complex nonlinear characteristics,and is susceptible to strong noise interference makes the fault characteristics difficult to extract,for the characteristics of the vibration signal proposed a nonlinear mode decomposition(NMD)algorithm and maximum correlated kurtosis deconvolution(MCKD)is proposed.Combined rolling bearing fault diagnosis method.First of all,the noise in some original vibration signals is filtered and processed by MCKD,which realizes the initial dynamic noise reduction of noise and further improves the relevant steepness values in the original signal,and then uses wavelet decomposition to extract the time frequency ridge,and then uses the NMD algorithm to decompose the time frequency ridge and screen out the effective components,and finally uses envelope spectral analysis to extract the fault characteristics.Simulation theory and experiments have shown that this treatment method can more effectively extract and analyze the bearing fault characteristics that may be damaged by vibration and noise interference.

关 键 词：故障诊断 滚动轴承 非线性模式分解 最大相关峭度解卷积 

分 类 号：TH133.3[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]