基于ITD-MOMEDA联合降噪的滚动轴承故障诊断研究

Fault diagnosis of rolling bearing based on ITD-MOMEDA combined noise reduction

在滚动轴承的实际运行过程中,其故障信号往往容易淹没于强背景噪声中,使其故障的类型难以得到识别,针对这一问题,提出了一种基于固有时间尺度分析(ITD)和多点最优调整的最小熵解卷积(MOMEDA)相结合的联合降噪方法,并将其应用于滚动轴承的故障诊断中。首先,用ITD算法对滚动轴承故障原始信号进行了分解,得到了多个固有旋转分量(PRC);其次,依据相关系数和峭度原则,挑选出了与原信号相关度较大的PRC分量,并对其进行了重构;然后,采用MOMEDA算法对重构信号进行了进一步降噪处理,完成了有用信号与噪声信号的分离;最后,对降噪后的信号进行了包络解调分析,提取出了故障特征频率,诊断出了轴承故障具体位置;此外,为了验证该方法的有效性,通过ITD与局域均值分解(LMD)、MOMEDA与最大相关峭度解卷积(MCKD)算法对仿真信号进行了对比分析,并对轴承外圈进行了实例分析。研究结果表明:相比于ITD-MCKD方法,基于ITD-MOMEDA联合降噪方法的故障诊断准确率提高4.3%,能更有效地去除强噪声,并成功地检测出轴承的故障类型。

Fault signal of rolling bearing was easily submerged in strong background noise during actual operation,which made it difficult to identify the fault type.Aiming at these problems,a joint noise reduction method based on intrinsic time-scale decomposition(ITD)and multipoint optimal minimum entropy deconvolution adjusted(MOMEDA)was proposed,and applied to the fault diagnosis of rolling bearing.Firstly,the ITD algorithm was used to decompose the original signal of the rolling bearing fault to obtain multiple proper rotation components(PRC);Secondly,according to the principle of correlation coefficient and kurtosis,the PRC that had a greater correlation with the original signal was selected for reconstruction;Then,MOMEDA algorithm was used to further denoise the reconstructed signal to separate the useful signal from the noise signal.Finally,the envelope demodulation analysis of the signal was performed to extract the fault characteristic frequency and diagnose the specific location of the bearing fault.In addition,in order to verify the effectiveness of the method,the simulation signals were compared and analyzed by ITD and local mean value decomposition(LMD),MOMEDA and maximum correlation kurtosis deconvolution(MCKD),and the analysis of the outer ring instance was presented.The results indicate that the diagnosis acuracy of the joint noise reduction method based on ITD-MOMEDA is 4.3% higher than the ITD-MCKD diagnosis accuracy,which can more effectively remove strong noise and successfully detect the type of bearing failure.