CQTBEW算法及其在轴承早期故障诊断中的应用

CQTBEW Algorithm and its Application in Bearing Early Fault Diagnosis

滚动轴承出现局部剥落、点蚀等故障时,会产生周期性的振动冲击信号,通常对此信号进行分析即可诊断出故障的严重程度以及出现的位置。但是,设备工作时往往伴随着较大的噪声,因此冲击信号,尤其是早期故障的振动冲击信号,很容易被噪声淹没。针对此问题,本文提出一种基于恒Q变换与二进制能量权重的CQTBEW算法,首先将振动冲击信号进行恒Q变换分析,获得时频谱矩阵;其次对矩阵进行频率分段处理,在时间轴设置滑移窗口,筛选局部极值并二值化时频谱,向时域叠加获得能量时域信号,进行功率谱分析诊断获得特征频率;最后进行仿真信号与实验信号的分析验证。结果表明,该方法具有可行性。

Periodic vibration and shock signal will be generated when rolling bearing has local spalling,pitting and other faults.Usually,the severity and location of the fault can be diagnosed by analyzing the signal.However,the equipment operation is often accompanied by large noise,so the shock signal,especially the vibration shock signal of early fault,is easy to be submerged by noise.In order to solve this problem,this paper proposes a constant Q transform binary energy weight(CQTBEW)algorithm based on constant Q transform and binary energy weight.Firstly,the vibration shock signal is analyzed by constant Q transform to obtain the time-frequency spectrum matrix.Secondly,the matrix is processed by frequency segmentation.The sliding window is set in the time axis,and the local extreme value is filtered and binarized.The energy time domain signal is obtained by superimposing the time domain signal,and the characteristic frequency is obtained by spectrum analysis diagnosis.Finally,the simulation signal and experimental signal are analyzed and verified,the results show that the CQTBEW method is feasible.