基于循环脉冲谱的动车轴箱轴承故障诊断方法

Fault Diagnosis Method for EMU Axle Box Bearings Based on Cyclic Pulse Spectrum

针对动车轴箱轴承故障冲击的脉冲性与周期性,提出基于归一窗S变换时频切片的循环脉冲谱分析方法。采用能量归一化窗函数对故障轴承振动信号进行归一化窗S变换并建立其能量随频率的变化关系,根据能量峰值与峭度最大原则在故障冲击共振频率点处进行时频切片。引入可变循环窗对时频切片序列的周期脉冲进行分离;采用可变循环窗内整体脉冲峰值矩的变异系数对故障信号的循环脉冲度进行表征,得到故障信号循环脉冲谱。通过仿真与动车轴箱轴承故障诊断实例表明,提出的循环脉冲谱不仅能够在强噪声干扰下对轴承复合故障进行准确诊断,而且对故障冲击共振频带中心的选择具有较好的鲁棒性,避免了对故障冲击共振频带的滤波解调与复杂的高次循环统计分析的繁琐步骤,计算简单高效,具有较好的工程适用性。

In response to the impulsivity and periodicity of fault impact of EMU axle box bearing, a cyclic pulse spectrum analysis method was developed based on the time-frequency slice of Stransform. In this paper, the energy normalized window function was used to perform normalized window S transform on the vibration signal of the faulty bearing, and the relationship of energy change with frequency was established. Along the time axis, the time-frequency distribution was sliced at the frequency corresponding to the maximum energy, thus to obtain the slice time series at the center of impact frequency. A variable cycle window was used to separate the periodic pulses in the slice time series. Then the variation coefficient of the overall peak pulse moment in the moving variable window was used to characterize the degree of cyclic pulses to obtain the cyclic pulse spectrum of the fault signal. The simulation and experiments show that the proposed cyclic pulse spectrum method can not only accurately diagnose the weak early faults and compound faults of EMU axle box bearing under strong interference noise, but also has good robustness to the selection of impact resonance frequency. This method avoids the filtering demodulation of resonance frequency band and complex high-order cycle statistical analysis, with good engineering applicability in the field of rotating machinery local fault diagnosis under complex working conditions.