小波降噪和局域均值分解的齿轮故障特征提取技术
The Feature Extraction of Gear Fault Using Wavelet Noise Reduction and Local Mean Decomposition
魏永合 1牛保国 1刘雪丽 1赵旭宁 1李曙光2
作者信息
- 1. 沈阳理工大学机械工程学院,沈阳110159
- 2. 霍林郭勒职业技术学校,内蒙古 通辽 029200
- 折叠
摘要
针对齿轮系统非线性、非平稳性特点及传统时频分析方法的局限性,提出一种将小波和局域均值分解(Local mean decomposition,LMD)相结合进行齿轮故障特征提取的方法.该方法将原始信号通过小波分解再重构进行处理,以降低噪声的干扰,然后对重构信号进行LMD分解,并且对分解后所得到的乘积函数(PF)分量进行筛选.对筛选后的乘积函数进行包络谱分析,提取其故障特征进行研究.结果表明,两者相结合是一种很有效的故障特征提取方法,减弱了噪声对信号的干扰,可以实现对其振动信号故障特征的提取和诊断.
Abstract
Considering the limitations of traditional time-frequency analysis method and in view of the gear system nonlinear and non-stationary characteristics,a method is put forward to combine the wavelet and the local mean decomposition (Local mean decomposition,LMD) for gear fault feature extraction.In this approach,firstly,in order to reduce noise interference,the wavelet is applied to decompose and reconstruct the original signal.Then,LMD method is used to decompose the reconstructed signal for Product Functions(PF).The interrelated PF is adopted to envelope spectrum analysis.Finally,the fault features are extracted.Simulation results show that the combination is an effective method for fault feature extraction,which can reduce the signal interference of noise,realize the extraction and the diagnosis of fault vibration signal feature.
关键词
小波/局域均值分解(LMD)/齿轮故障/特征提取Key words
the wavelet/local mean decomposition(LMD)/gear fault/feature extraction引用本文复制引用
出版年
2016
NETL
NSTL
维普
万方数据