According to the great challenge that the weak signal of early fault and mutational composition including pulse bring to fault collection, dinoising, prediction, the filtering characteristics and time-frequency of Morlet wavelet transform are analyzed, and a feature extraction method based on parameter opfimized Morlet wavelet is proposed.Minimum Shannon entropy and cycle detection method of singular value decomposition are used to optimize shape parameter β and scale factor a of Morlet wavelet. Finally, choses the optimal Morlet wavelet filter core to test and analyze early condition of CNC machine main shaft bearing, so as to complete early condition monitoring and fault prediction of CNC machine tool spindle bearing. The experimental results and analysis results of bearing signals show that the actual frequency values is 232.7Hz and it is close to fault permission frequency. So speculates that property of bearing outer ring is reduced, so as to complete the fault prediction. This research is helpful to extract and test weak signal of mechanical mutation failure, and it is very meaningful for fault diagnosis, prediction, and fault injection in the future.%针对故障早期状态信号的微弱与包含脉冲突变成分对故障信号采集、 去噪、 预测等带来的极大挑战,分析了Morlet小波变换的滤波特性及其时频分辨率,提出了基于参数优化的Morlet小波变化的故障特征提取办法.利用最小Shannon熵方法和奇异值分解的周期检测方法分别对Morlet小波的形状参数 β 和尺度因子a进行优化.最终选择了最优Morlet小波作为滤波内核,对轴承早期状态进行震动检测与分析,从而完成对数控机床主轴轴承的早期状态监测与故障预测.仿真试验和实际应用的结果表明,实际f=232.7Hz,接近故障通过频235.6Hz,推测轴承外圈出现性能下降,完成故障预测.该研究有助于对机械突变故障信号的微弱信号检测和提取,对数控装备的故障诊断、预测及未来故障注入,BIT有重要研究意义.
展开▼
