Damaged motor bearing may cause its stator current to generate the corresponding current harmonics.The bearing fault characteristic frequencies exist in the current harmonic frequencies. In order to effectively evaluate the complexity of the stator current signal( current harmonic generation probability), EEMD sample entropy is used to decompose the stator current into several intrinsic mode components, then the sample entropy of each component is calculated.comparison, in the stator current signal complexity as-sessment, it is gained that the effects of EEMD sample extropy are better than ones of the sample entropy. The EEMD sample entro-py variation is increase-decrease-increase trend which is consistented with the stator current variation trend when bearing damage gradual y increases. According to the above conclusions, this method can be used for monitoring and anticipating the motor bearing condition in enclosed structure and this signal source can be applied to the intel igent fault identification.%电机轴承损伤会导致电机定子电流产生相应的电流谐波，电流谐波频率包含轴承故障特征频率。为了有效评估定子电流信号的复杂性(即电流谐波出现概率)，采用总体平均经验模态分解( EEMD)结合样本熵来实现，该方法先用EEMD将定子电流信号分解为若干个内禀模态分量，再计算分量的样本熵。通过比较得出在评估损伤轴承定子电流信号复杂性时EEMD样本熵的效果较样本熵更好，并且 EEMD样本熵增大-减小-增大的变化趋势与轴承损伤逐渐加大时定子电流的变化趋势一致。根据上述结论该方法可应用于封闭结构中电机轴承运行状态的监测和预判，也可以作为智能故障识别的信号源。
展开▼
