自适应变分模式提取的轴承故障诊断方法

Bearing fault diagnosis method based on adaptive variational mode extraction

变分模式提取(variational mode extraction, VME)能够从复杂信号中提取特定的模式分量,但其在轴承故障诊断中的应用潜力受到初始中心频率和平衡参数的影响。因此,为了克服VME在轴承故障诊断应用中超参数的设置问题,深入探究VME模型中心频率迭代更新过程,发现中心频率收敛趋势现象并通过理论证明其存在性,由此提出中心频率定位策略,可自适应地确定目标中心频率。为了最大化匹配故障信息,构造基于故障特征幅值比的平衡参数优化策略,能够优化目标分量的带宽。以上中心频率定位策略和平衡参数优化策略,构成自适应变分模式提取的故障诊断方法,该方法能够在无需预设初始中心频率及平衡参数的情况下自适应提取故障相关分量。仿真和两个试验案例分析结果验证所提方法在轴承故障诊断领域相比于连续变分模式分解、经验模式分解和快速谱峭度方法更具有效性和优越性。

Variational mode extraction(VME)can extract specific mode components from complex signals,but its application potential in bearing fault diagnosis was affected by initial center frequency and balance parameters.Here,in order to overcome the problem of setting super parameters in application of VME in bearing fault diagnosis,deeply explore iterative updating process of VME model center frequency,discover the convergence trend of center frequency,and prove its existence through theory,thus a center frequency positioning strategy which could adaptively determine the target center frequency was proposed.To maximize matching of fault information,a balance parametric optimization strategy based on fault feature amplitude ratio was constructed to optimize bandwidth of target component.The above center frequency positioning strategy and balance parametric optimization strategy could constitute an adaptive variational mode extraction fault diagnosis method,it could adaptively extract fault-related components without presetting initial center frequency and balance parameters.Simulations and two test case studies showed that the proposed method is more effective and superior to methods of continuous variational mode decomposition,empirical mode decomposition,and fast spectral kurtosis in field of bearing fault diagnosis.