基于EEMD和变尺度随机共振的轴承故障诊断

Bearing Fault Diagnosis Based on EEMD and STSR

提出了一种基于集合平均经验模式分解(EEMD)和变尺度随机共振(STSR)的滚动轴承故障提取方法。首先通过EEMD对含噪振动信号进行自适应抗混分解,得到不同频带的本征模态函数(IMF);然后将不同频带的IMF作为双稳系统的输入,通过变步长数值算法和调节非线性双稳系统的结构参数来提取微弱低频故障特征信号;最后运用切片双谱对双稳系统的输出进行后处理。仿真分析验证了STSR的特性,通过对强噪声背景下的滚动轴承实测信号分析表明,该方法充分利用高斯白噪声,能有效提取滚动轴承微弱故障特征。

A fault feature extraction method of rolling bearing based on ensemble empirical mode decomposition （EEMD） and scale-transformation stochastic resonance（STSR） is proposed.Firstly,the vibration signal with noise is adaptively anti-aliasing decomposed by EEMD to conduct intrinsic mode function（IMF） of different frequency bands.Then making the IMFs as the input of bi-stable system,the low frequency fault features signal is extracted by the step-changed numberical algorithm and the adjustment of the bi-stable system parameters. Finally,slice bi-spectrum is adopted to postprocess the output of the bi-stable system.Simulation analysis is performed to prove the characteristics of STSR,the analysis on measured signal of the rolling bearing in strong background noise shows that the approach can extract the weak fault features of rolling bearing with the full use of Gaussian white noise successfully.