基于振动特征值的动车组齿轮箱故障检测

故障检测与状态评估对保障动车组齿轮箱的正常运行具有重要意义。通过对振动信号的特征提取方法与判定进行研究，对比不同故障模式与试验工况下的齿轮箱振动有效值、峰值、标准差、偏度、峭度等振动特征值差异。结果表明:动车组齿轮箱的故障模式直接影响振动特征值；有量纲振动特征值受运行转速影响较明显，运行转速越高，相应特征值变化越明显；无量纲振动特征值受轴承故障模拟形式的影响较大，与运行转速关系较小。综合对比各种故障状态下振动特征值的敏感度，有量纲特征值比无量纲特征值对故障更加敏感。

Fault Detection of Electric Multiple Unit Gearbox Based on Vibration Eigenvalue

Fault detection and condition evaluation are of great significance to ensure the normal operation of EMU gearbox. By studying the feature extraction method and judgment of vibration signal, the differences of vibration eigenvalues such as effective value, peak value, standard deviation, skewness and kurtosis of gearbox vibration under different fault modes and test conditions are compared. The results show that the fault mode of the gearbox of EMU directly affects the vibration eigenvalue; The characteristic value of dimensional vibration is obviously affected by the running speed. The change of the characteristic value is evident under the high running speed condition. The dimensionless vibration eigenvalue is greatly affected by the bearing fault simulation form, but little affected by the running speed. By comprehensively comparing the sensitivity of vibration eigenvalues under various fault state, dimensional eigenvalues are more sensitive to fault than dimensionless eigenvalues