基于广义复合多尺度熵的滚动轴承故障状态表征

Fault status characterization of rolling bearing based on generalized composite multiscale entropy

针对滚动轴承故障振动时间序列非线性的特点,采用广义复合多尺度熵(GCMSE)对故障状态进行表征分析。对白噪声和1/f噪声进行分析,并与多尺度熵(MSE)对比后,发现熵值在高尺度因子下的振荡现象被有效抑制。在此基础上,分析不同载荷和损伤直径条件下SKF6205-2RS型深沟球轴承正常和单点损伤故障数据,发现随着尺度因子的增加,正常状态的熵值先上升后下降再趋于平稳,内圈、外圈和滚动体故障的熵值则单调递减。不同故障状态的熵值受载荷的影响较小,对损伤直径的变化呈现不同敏感程度。研究结果表明,GCMSE能够在克服MSE不足的基础上,有效表征不同故障状态的演化过程。

According to the nonlinear characteristics of the vibration time series of rolling bearings,GCMSE is applied to characterize failure stages.By analyzing white noise and 1/f noise,and comparing with MSE,it is found that the oscillation of entropy is suppressed under high scale factor,and the validity of the proposed method is proved.Then the method is applied to analysis normal and single point damage fault data of SKF6205-2RS deep groove ball bearing.It is discovered that the entropy distribution presents obviously different between normal status and fault status,under different load and damage diameter conditions.With scale factor rise,entropy of normal status first increases,next descends,then stabilizes again,but the entropy of outer-race faults,inner-race ones and rolling element ones present monotonously decreases.Entropy of different fault status are affected less by the load,and are sensitive to the change of damage diameter.The results show that GCMSE not only is feasible to overcome disadvantages of MSE,but also can effectively characterized evolution process of different fault status.