基于LMD的能量算子解调方法及其在故障特征信号提取中的应用

摘要

局域均值分解是将多分量凋频调幅信号分解为一系列单分最调频调幅信号的有效时频分析方法.为提取故障信号的特征,提出了基于局域均值分解的能量算子解凋方法,局域均值分解将复杂的多分量信号分解为若干个乘积函数的线性组合,再通过能量算子解凋方法可求取每个乘积函数的幅频信息,从而可进一步获取故障信号的时频分布或提取其故障特征.为提高分析精度,提出了特征趋势正弦函数数据延拓方法以有效克服端点效应的影响.实验信号的分析结果表明,所提出的基于局域均值分解的能量算子解调方法能有效提取机械故障振动信号的特征.%Local mean decomposition (LMD) is an effective time-frequency for decomposing a multi-component AM-FM signal into a number of mono-component AM-FM signals. An energy operator demodulating approach based on LMD was proposed for extracting characteristics of a faulty signal. LMD was used to decompose a complex multicomponent AM-FM signal into a linear combination of a finite number of production functions ( PF), then the energy operator demodulation was applied to extract amplitudes and frequencies of each PF so as to obtain the time-frequency distribution of the faulty signal and extract the fault characteristics. In order to improve the analysis precision, a data extension method based on trend character of sine function was proposed to overcome end effects. The application results showed that the proposed energy operator demodulating approach based on LMD can extract characteristics of mechanical fault vibration signals effectively.