广义形态滤波和VMD分解的滚动轴承故障诊断

Rolling bearing fault diagnosis of generalized morphological filtering and VMD decomposition

滚动轴承早期故障信号的特征提取通常都会受到强噪声的影响,难以得到有效信息,因此提出了广义形态滤波和变分模态分解(variational mode decomposition,VMD)相结合的方法。首先利用不同结构元素构成的广义形态滤波去除故障信号中噪声信息的影响,突出故障信号的冲击特征;其次运用VMD对处理过后的信号进行分解,得到若干个模态分量;然后由峭度准则选取故障信息最为丰富的模态分量进行包络谱分析,获取滚动轴承的故障特征频率。通过对仿真信号和实测数据的处理结果表明,方法对噪声的去除效果显著,获取故障特征频率也十分明显,能够作为滚动轴承故障诊断的有效方法。

The extraction of bearing fault signal is usually influenced by strong noise in early time,the effective information is difficult to obtain,thus the combination of generalized morphological filter and VMD( variational mode decomposition,VMD) is proposed. Firstly,the influence of noise information in fault signal is removed by the generalized morphological filtering of different structural elements,and the impact characteristics of fault signal are highlighted. Secondly,the processing signal is decomposed by VMD,a series of model component are obtained. According to kurtosis criterion selected the model component that has abundant fault information to analysis envelope spectrum,the rolling bearing fault characteristic frequency is obtained. The processing results through the simulation signal and the measured data show that the proposed method is effective to eliminate the noise and the fault characteristic frequency is also obvious,so it is an effective method of rolling bearing fault diagnosis.