基于 HVD 的振动故障诊断方法研究

摘要

For the nonlinear and non-stationary of rotating machinery vibration fault diagnosis , in order to diagnose faults accurately, a vibration signals time-frequency approach based on Hilbert vibration decomposition ( HVD ) method is introduced.Then HVD marginal spectrum and the three dimensional spectrum is proposed to further analyze the signals . And compared with the empirical mode decomposition ( EMD ) , the HVD shows a higher precision of decomposition , overcoming the defects of mode aliasing and amplitude distortion .Meanwhile , the HVD marginal spectrum and the three dimensional spectrum can show the signal of each frequency changes more clearly .In this paper , the simulation signal and the actual rotor vibration signal are researched by experiment , which clearly demonstrates the fault characteristics and achieves good analytical results , proving the effectiveness of HVD and its marginal spectrum and three dimensional spectrum in dealing with nonlinear and non-stationary problems .%由于旋转机械振动信号具有非线性、非平稳性，为了准确诊断故障，介绍了一种基于希尔伯特振动分解（ Hil-bert Vibration Decomposition ，HVD）的振动信号时频分析方法，并在此基础提出了HVD边际谱和时频三维谱用于振动信号的进一步分析。通过与经验模态分解（ Empirical Mode Decomposition ，EMD）方法对比发现，HVD方法克服了模态混叠和幅值失真缺陷，具有更高的分解精度，而且其边际谱和时频三维谱更加清晰地展示出信号各频率的变化情况。对仿真信号及转子实际振动信号进行实验研究，清晰地展现了故障特征，取得了良好的分析效果，证明了HVD方法及其边际谱、三维谱在处理非线性、非平稳问题上的有效性。