基于Infogram的重复瞬态特征提取与旋转机械故障诊断

Repetitive Transient Extraction and Fault Diagnosis of Rotating Machinery Using Infogram

针对旋转机械常见故障类型与检测问题,考虑到旋转机械故障振动信号的重复瞬态、冲击和循环平稳特性,提出基于Infogram的重复瞬态特征提取与旋转机械故障诊断方法。不同于谱峭度、快速谱峭度图以及包络谱,Infogram不仅能够同时在时域和频域捕获重复瞬态特征,而且以热力学的平衡态分离思想为基础能够测量信号平方包络和平方包络谱的负熵,因此它对冲击和循环平稳较为敏感,有助于实现旋转机械故障诊断。通过对仿真和电机轴承实验数据分别用提出方法和集总经验模式分解提取故障征兆特征,结果表明提出方法优于集总经验模式分解方法,能够实现旋转机械关键零部件的故障诊断。

In order to solve the common fault types and detection problems of rotating machinery, and in consideration of the repetitive transient, shock and cyclostationarity of vibration signals, a method based on Infogram to extract repetitive transients and diagnose the faults of rotating machinery is proposed. Different from spectral kurtosis, fast kurtogram and envelope spectrum, Infogram can not only capture the signatures of repetitive transients in both time and frequency domains, but also measure the negentropy of the squared envelope and squared envelope spectrum based on the ideas of departures from a state of equilibrium. Therefore, it is sensitive to impulsiveness and cyclostationarity, which helps to diagnose local faults of rotating machinery. The fault symptom features were extracted by the proposed method and the lumped empirical mode decomposition,respectively. Simulation and experimental data indicate that the proposed method is better than ensemble empirical mode decomposition when extracting repetitive transients and can achieve fault diagnosis of key mechanical components.