基于梳状滤波器和二维灰度图的滚动轴承故障诊断

Fault Diagnosis of Rolling Bearing Based on Comb Filter and Gray Level Images

通过设计梳状滤波器和构造二维灰度图,保留振动信号的时域特征,解决了传统信号时域平均方法的滤波性能受信号周期估计误差和采样截断误差影响的问题,为提取振动信号的时域特征提供新的方法。推导了梳状滤波器的数学模型,给出了二维灰度图的构造方法,并通过仿真试验和应用实例验证了该方法的正确性。

The performance of traditional time domain averaging method is greatly influenced by the error of estimating period and sampling signal. A design method for comb filter is proposed,it is demonstrated that the traditional time domain averaging is just a special case of the general comb filter and its performance is influenced by the error of estimating period only. On this basis,a new signal analysis technique is presented capable of extracting the periodic wave, form having inexactly known period from noisy discrete - time signal. In the end, the validity of the approach is confirmed by analyzing the synthetic signal and the vibration signal produced by multiple point defects in a rolling dement bearing.