滚动轴承疲劳故障在线诊断的研究

通过对滚动轴承振动信号的在线监测提取出对疲劳故障敏感的参数：峭度、功率谱故障频带能量值、小波包故障频带能量值．选择足够的具有代表性的样本数据训练神经网络，用训练好的神经网络进行在线诊断，可以得出轴承发生疲劳故障的程度，再经过共振解调法诊断出轴承具体损伤的元件，实验表明本方法对滚动轴承的疲劳故障能正确诊断。该监测和诊断方法对其他设备的监测和诊断也有重要的意义。

On-line fatigue fault diagnosis of rolling element bearings

A fatigue fault diagnosis method based on on-line monitoring of vibration signal of rolling element bearings is given. Some characteristic parameters which are sensitive to fatigue fault are extracted, and enough representative data are chose to train the neutral network. The fatigue degree of rolling element bearings can be diagnosed using the trained neutral network. Further, the specific fault element of rolling element bearings can be ascertained by resonance demodulation. The experimental results show that fatigue fault of rolling element bearings can be diagnosed accurately. The method is also significant for monitoring and fault diagnosis of other machines.