局部波动特征分解及其在滚动轴承故障诊断中的应用研究

提出了一种新的自适应时频分析方法--局部波动特征分解（Local oscillatory-characteristic decomposition，LOD），该方法以信号本身的局部波动特征为基础，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，通过仿真信号将LOD、经验模态分解（Empirical mode decomposition，EMD）和局部均值分解（Local mean decomposition，LMD）进行了对比分析，结果表明了LOD 的优越性。同时，针对滚动轴承故障振动信号的多分量调制特点，将LOD应用于滚动轴承故障诊断，对滚动轴承实验信号进行了分析，结果表明LOD可以有效地提取滚动轴承故障振动信号的特征。

Local Oscillatory-Characteristic Decomposition and Its Application to Roller Bearing Fault Diagnosis

A new self-adaptive time-frequency analysis method named local oscillatory-characteristic decomposition (LOD) is proposed. This method is based on local oscillatory characteristics of signal itself, and it uses the operations including differential, coordinates domain transform and piecewise linear transform to decompose the signal into a series of mono-oscillatory components (MOC) which instantaneous frequency has physical meanings, and thus especially suitable for processing the multi-component signals. On the basis of illustrating the decomposition principle of LOD in detail, the LOD is compared with the empirical mode decomposition (EMD) and Local mean decomposition (LMD) by analyzing the simulated signals, and the results show the superiorities of LOD. Meanwhile, aiming at the multi-component modulated feature of roller bearing fault vibration signals, the LOD is applied to the roller bearing fault diagnosis. The analytical results from the experimental roller bearing signals demonstrate that the LOD can extract the fault characteristics of roller bearing fault vibration signals effectively.