基于信号经验模态分解与分集合并的齿轮故障诊断

Gear fault diagnosis based on Empirical Mode Decomposition andSignal diversity merging

摘要:

为了准确地进行齿轮故障诊断,结合信号经验模态分解与分集合并,提出了一种新的故障诊断方法。首先,运用经验模态分解对齿轮振动信号进行分解得到若干个分量；其次,根据分量的峭度大小以及相邻分量的峭度是否接近,筛选、合成有效分量；然后,运用等增益分集合并技术对有效分量进行合并,即将其包络进行叠加；接着,使用快速傅立叶变换得到信号包络和的频率谱；最后,根据该频率谱进行故障诊断。通过对仿真信号和齿轮断齿故障振动信号的分析,验证了方法的有效性。

Abstract:

In order to accurately diagnose gear faults, a new fault diagnosis method is proposed by combining signal empirical mode decomposition and diversity combination. First, use empirical mode decomposition to decompose the gear vibration signal to obtain several components; secondly, filter and synthesize effective components according to the kurtosis of the components and whether the kurtosis of adjacent components are close; then, use equal gain diversity combining technology Add the signal envelopes of the effective components; then, use the fast Fourier transform to obtain the envelope and the frequency spectrum of the signal; finally, perform fault diagnosis based on the frequency spectrum. The effectiveness of the method is verified by analyzing the simulation signal and the vibration signal of the gear broken tooth fault.