基于LMD和局域时频熵的旋转机械故障诊断方法

提出了一种基于局部均值分解和局域时频熵的旋转机械故障诊断方法。以旋转机械作为研究对象,利用LMD方法分解旋转机械振动信号,对分解得到的各乘积函数进行Hilbert变换,得到振动信号的时频分布。为了定量描述振动信号能量的时频分布情况,提出了局域时频熵的概念,根据旋转机械故障的频谱特征,将整个时频平面划分为若干时频段,计算时频段的局域时频熵,以局域时频熵作为旋转机械故障特征,实现旋转机械故障特征提取。基于局域时频熵进行故障特征提取可以细致地反映各时频区域能量分布的差别,同时可以减小计算量,提高运算速度。仿真与实验结果表明,该方法能有效地应用于旋转机械故障诊断中。

Method of Rotating Machinery Fault Diagnosis Based on LMD and Local Time-frequency Entropy

A method based on LMD (Local Mean Decomposition) and local time-frequency entropy in rotating machinery fault diagnosis is proposed. Aiming at bearing, the vibrating signal is decomposed into PFs (Product Functions) by LMD, and then Hilbert transformation is applied to every PF to get time-frequency distribution. The local time-frequency entropy is introduced to study the energy in time-frequency distribution quantitatively. In detail, according to the spectrum characteristic of bearing fault, the whole time-frequency plane is divided into some segments, and whose entropies are calculated to extract the fault feature of the bearing. By the method of fault feature extraction based on local time-frequency entropy, differences among the segments could be reflected in large. Also the computational complexity is reduced at the same time. The results of simulation and experiment are presented to verify the theory analysis.