局部波动特征分解及其在齿轮包络分析中的应用

Local oscillatory-characteristic decomposition and its application to gear envelope analysis

提出了一种新的自适应非平稳信号分析方法——局部波动特征分解(Local Oscillatory-Characteristic Decomposition,LOD),该方法以信号本身的时间尺度特征为基础,并采用微分、坐标域变换、分段线性变换等运算手段将信号分解为一系列瞬时频率具有物理意义的单一波动分量(Mono-Oscillatory Component,MOC),非常适合于处理多分量信号。在详细说明LOD分解原理的基础上,通过对仿真信号的分析将LOD和经验模态分解(Empirical Mode Decomposition,EMD)进行了对比,结果表明了LOD的优越性。同时,针对多分量调制的齿轮故障振动信号在包络分析中的特点,将LOD应用于齿轮故障振动信号的分析,对齿轮实验信号和实际信号的分析结果表明,LOD可以有效地应用于齿轮的包络分析。

A new self-adaptive nonstationary signal analysis method named local oscillatory-characteristic decomposition（LOD）is proposed.This method is based on time-scale characteristics of signal itself,and it uses kinds of operations such as differential,coordinates domain transformation and piecewise linear transformation to decompose the signal into a series of mono-oscillatory components（MOC）whose instantaneous frequency has physical meanings,and thus especially suitable for multi-component signal processing.On the basis of illustrating the decomposition principle of LOD in detail,the LOD is compared with the empirical mode decomposition（EMD）by analyzing the simulated signal.The results show the advantages of LOD.Meanwhile,taking account of the characteristics of multi-component modulated gear fault vibration signal in envelope analysis,the LOD is applied to the gear fault vibration signals analysis.The analytical results from experimental gear signal and actual gear signal demonstrate that the LOD apply to gear envelope analysis effectively.