频率辅助信号结合EMD的旋转机械故障诊断

Rotating Machinery Fault Diagnosis Using Frequency Auxiliary Signal and EMD

摘要:

针对旋转机械的故障自动诊断问题，提出一种基于经验模态分解(EMD)和频率辅助信号(FAS)的故障诊断方法。首先，利用滤波器移除非故障分量，通过实验采集各种故障下的特征频率，构建故障模型。然后，在实时故障诊断中，对光电位移传感器采集到的机械振动信号进行频谱分析，当主频接近一个特定故障的特征频率时，根据该特征频率构建一个FAS，并将其与振动信号进行叠加。接着，对叠加后的信号进行EMD，根据能量准则选择出主固有模态函数(IMF)。最后，通过三次样条插值法获得主IMF信号的包络，并获得包络谱的中心频率，以此对故障进行诊断。实验结果表明，提出的方法能够解决EMD的模态混叠问题，同时对故障的并发情况具有鲁棒性。

Abstract:

For the issues that the automatic fault diagnosis of rotating machinery, a fault diagnosis method based on empirical mode decomposition (EMD) and frequency auxiliary signal (FAS) was proposed. First of all, the filter was used to remove the nonbearing fault component, and through the experimental acquisition of the frequency of various failures, a fault model was built. Then, the mechanical vibration signal collected by the photoelectric displacement sensor was spectrum analyzed. When the main frequency is close to the characteristic frequency of a specific fault, a FAS was constructed according to the characteristic frequency and superimposed with the vibration signal. Then, the EMD operation was performed on the superimposed signal, and the main intrinsic mode function (IMF) was selected according to the energy criterion. Finally, the envelope of the main IMF signal was obtained by the cubic spline interpolation method, and the center frequency of the envelope spectrum was obtained to diagnose the fault. The experimental results show that the proposed method can solve the problem of modal aliasing of EMD and is robust to the concurrency of the fault.